SPECIAL PUBLICATION

of CARNEGIE MUSEUM OF NATURAL HISTORY

CONTRIBUTIONS IN QUATERNARY
VERTEBRATE PALEONTOLOGY:

A VOLUME IN MEMORIAL TO
JOHN E. GUILDAY

edited by

HUGH H. GENOWAYS

Curator of Mammals

MARY R. DAWSON

Chief Curator, Earth Sciences
and

Curator of Vertebrate Fossils

NUMBER 8

PITTSBURGH, 1984

SPECIAL PUBLICATION OL CARNEGIE MUSEUM OF NATURAL HISTORY

Number 8, pages 1-538

Issued 11 September 1984
Price: $56.00 a copy

Robert M. West, Director

Editorial Staff: Hugh H. Genoways, Editor ; Duane A. Schlitter, Associate Editor;
Stephen L. Williams, Associate Editor; Marion A. Burgwin, Editorial Assistant;
Mary Ann Schmidt, Technical Assistant.

© 1984 by the Trustees of Carnegie Institute, all rights reserved.

ISBN 0-935868-07-0

Library of Congress card number 84-70729

CARNEGIE MUSEUM OF NATURAL HISTORY, 4400 FORBES AVENUE
PITTSBURGH, PENNSYLVANIA 15213

CONTENTS

Introduction 1

Mary R. Dawson and Hugh H. Genoways, Section of Vertebrate Fossils, Carnegie Museum of
Natural History, 4400 Forbes Avenue, Pittsburgh, PA 15213; Section of Mammals, Car-
negie Museum of Natural History, 4400 Forbes Avenue, Pittsburgh, PA 15213.

John Edward Guilday (1925-1982) 3

Mary R. Dawson, Section of Vertebrate Fossils, Carnegie Museum of Natural History, 4400 Forbes
Avenue, Pittsburgh, PA 15213.

Mid-Wisconsinan and Mid-Holocene Herpetofaunas of Eastern North America: A Study in Minimal

Contrast 14

Leslie P. Fay, The Museum, Michigan State University, East Lansing, MI 48824-1045.

Herpetofaunas of the Duck Creek and Williams Local Faunas (Pleistocene: Ilhnoian) of Kansas 20

J. Alan Holman, The Museum, Michigan State University, East Lansing, MI 48824-1045.

Pleistocene Birds from Cumberland Cave, Maryland 39

Pierce Brodkorb and Cecile Mourer-Chauvire, Department of Zoology, University of Florida,
Gainesville, FL 3261 1; Centre de Paleontologie stratigraphique et Paleoecologie de l’Uni-
versite Claude Bernard, 69622 Villeurbanne Cedex, France.

A Very Large Enigmatic Owl (Aves: Strigidae) from the Late Pleistocene at Ladds, Georgia 44

Storrs L. Olson, Department of Vertebrate Zoology, National Museum of Natural History, Smith-
sonian Institution, Washington, DC 20560.

A Middle Pleistocene (Late Irvingtonian) Avifauna from Payne Creek, Central Florida 47

David W. Steadman, Department of Vertebrate Zoology, National Museum of Natural History,
Smithsonian Institution, Washington, DC 20560.

An Evaluation of the Fossil Curlew Palnumenius victima L. Miller (Aves: Scolopacidae) 53

Storrs L. Olson, Department of Vertebrate Zoology, National Museum of Natural History, Smith-
sonian Institution, Washington, DC 20560.

Phylogeny and Paleobiogeography of Short-tailed Shrews (Genus Blarina) 56

Cheri A. Jones, Jerry R. Choate, and Hugh H. Genoways, Museum of the High Plains, Fort Hays
State University, Hays, KS 67601; Section of Mammals, Carnegie Museum of Natural
History, 4400 Forbes Avenue, Pittsburgh, PA 15213.

Armadillos in North American Late Pleistocene Contexts 149

Walter Klippel and Paul W. Parmalee, Department of Anthropology, University of Tennessee,
Knoxville, TN 37916; Frank H. McClung Museum, University of Tennessee, Knoxville,

TN 37916.

A Pleistocene Occurrence of Geomys (Rodentia: Geomyidae) in West Virginia 161

Frederick Grady, Department of Paleobiology, National Museum of Natural History, Smithsonian
Institution, Washington, DC 20560.

Neotoma in the Late Pleistocene of New Mexico and Chihuahua 164

Arthur H. Harris, Laboratory for Environmental Biology, University of Texas at El Paso, El Paso,

TX 79968.

iii

The Evolution of Cotton Rat Body Mass 179

Robert A. Martin, Department of Biological and Allied Health Sciences, Fairleigh Dickinson Uni-
versity, Madison, NJ 07940.

Biostratigraphy and Biogeography of Quaternary Microtine Rodents from Northern Yukon Territory,

Eastern Beringia 184

Richard E. Morlan, Archaeologial Survey of Canada, National Museum of Man, Ottawa, Ontario,

K1A 0M8, Canada.

New Arvicolines (Mammalia: Rodentia) from the Blancan of Kansas and Nebraska 200

Richard J. Zakrzewski, Sternberg Memorial Museum and Department of Earth Sciences, Fort Hays
State University, Hays, KS 67601.

Geographic Differentiation in the Rancholabrean Dire Wolf (Canis dints Leidy) in North America. . . 218
Bjorn Kurten, Department of Geology, University of Helsinki, 00170 Helsinki 17, Finland.

Use and Abuse of Dogs 228

Elizabeth S. Wing, Florida State Museum, University of Florida, Gainesville, FL 3261 1.

Time of Extinction and Nature of Adaptation of the Noble Marten, Martes nobi/is 233

Donald K. Grayson, Department of Anthropology and Burke Memorial Museum, University of
Washington, Seattle, WA 98195.

Lava Blisters as Carnivore Traps 241

John A. White, H. Gregory McDonald. Elaine Anderson, and James M. Soiset, Idaho Museum of
Natural History, Idaho State University, Pocatello, ID 83209; Vertebrate Paleontology,
Royal Ontario Museum, 100 Queen’s Park, Toronto, Ontario M5C 2C6, Canada; 730
Magnolia Street, Denver, CO 80220; 1284 South Fisher Avenue, Blackfoot, ID 83221.

Review of the Small Carnivores of North America During the Last 3.5 Million Years 257

Elaine Anderson, 730 Magnolia Street, Denver, CO 80220.

The Pleistocene Dung Blanket of Bechan Cave, Utah 267

Owen K. Davis, Larry Agenbroad, Paul S. Martin, and Jim J Mead, Department of Geosciences,
University of Arizona, Tucson, AZ 85721; Department of Geology, Northern Arizona
University, Flagstaff, AZ 86011; Department of Geosciences, University of Arizona,
Tucson, AZ 85721; Center for the Study of Early Man, University of Maine, Orono, ME
04473.

Pleistocene Tapirs in the Eastern United States 283

Clayton E. Ray and Albert E. Sanders, Department of Paleobiology, National Museum of Natural
History, Smithsonian Institution, Washington, DC 20560; Department of Natural Sci-
ences, Charleston Museum, Charleston, SC 29403.

Sangamona : The Furtive Deer 316

Charles S. Churcher, Department ofZoology, University of Toronto, Toronto, Ontario, M5S 1 Al,
Canada and Department of Vertebrate Palaeontology, Royal Ontario Museum, Toronto,
Ontario M5S 2C6, Canada.

Geological Setting and Preliminary Faunal Report for the St-Elzear Cave, Quebec, Canada 332

Pierre LaSalle, Ministere de 1’Energie et des Ressources, 1620 Blvd. de l’Entente, Quebec City,
Quebec G1S 4N6, Canada.

IV

Meadowcroft Rockshelter and the Pleistocene/Holocene Transition in Southwestern Pennsylvania . . . 347
J. M. Adovasio, J. Donahue, R. C. Carlisle, K. Cushman, R. Stuckenrath, and P. Wiegman,
Department of Anthropology, University of Pittsburgh, Pittsburgh, PA 15260; Depart-
ment of Geology and Planetary Sciences, University of Pittsburgh, Pittsburgh, PA 1 5260;

3262 Cripple Creek Trail, Boulder, CO 80303; Radiation Biology Laboratory, Smithsonian

Institution, Washington, DC 20560; Western Pennsylvania Conservancy, 316 Fourth
Avenue, Pittsburgh, PA 15222.

Historical Biogeography of Florida Pleistocene Mammals 370

S. David Webb and Kenneth T. Wilkins, Florida State Museum, University of Florida, Gainesville,

FL 3261 1.

Two Irvingtonian (Medial Pleistocene) Vertebrate Faunas from North-Central Kansas 384

Ralph Eshelman and Michael Hager, Calvert Marine Museum, Solomons, MD 20688; Museum
of the Rockies, Montana State University, Bozeman, MT 59719.

Paleoecology of a Late Wisconsinan/Holocene Micromammal Sequence in Peccary Cave, Northwestern

Arkansas 405

Holmes A. Semken, Jr., Department of Geology, University of Iowa, Iowa City, IA 52242.

Late Pleistocene and Early Recent Mammals from Fowlkes Cave, Southern Culberson County,

Texas 432

Walter W. Dalquest and Frederick B. Stangl, Jr., Department of Biology, Midwestern State Uni-
versity, Wichita Falls, TX 76308.

A Late Pleistocene Mammalian Fauna from Cueva Quebrada, Val Verde County, Texas 456

Ernest L. Lundelius, Jr., Department of Geological Sciences, University of Texas at Austin, Austin,

TX 78712.

Alaskan Megabucks, Megabulls, and Megarams: The Issue of Pleistocene Gigantism 482

R. Dale Guthrie, Institute of Arctic Biology, University of Alaska, Fairbanks, AK 99701.

Quaternary Marine and Land Mammals and Their Paleoenvironmental Implications — Some Examples

from Northern North America 511

C. R. Harington, National Museum of Natural Sciences, National Museums of Canada, Ottawa
K1A 0M8, Canada.

Phyletic Trends and Evolutionary Rates 526

Larry D. Martin, Museum of Natural History and Department of Systematics and Ecology, Uni-
versity of Kansas, Lawrence, KS 66045.

v

INTRODUCTION

Mary R. Dawson and Hugh H. Genoways

The loss of a good friend and outstanding col-
league, as John Guilday certainly was to us, tends
to leave the survivors with mixed feelings of help-
lessness and “let’s do something.” Fortunately, fol-
lowing the example set time and again by John him-
self, our feelings of wanting to take action soon won
out and this volume, devoted to vertebrates from
the Quaternary of North America, is the result. This
broad subject, also set by John, was his consuming
scientific interest for the past forty years. His own
work, especially on the Quaternary of eastern North
America, contributed greatly to understanding of
this complex, dynamic interval of time, as can be
appreciated fully by reading his bibliography.

Once we had decided on a memorial volume as
an appropriate tribute, we wrote to Quaternary spe-
cialists in North America and Europe, asking for
their participation in the volume. The response to
our letters was very encouraging. The people who
know John’s work best were eager to contribute to
a volume in memory of our remarkable friend. Let-
ters of acceptance came in, many accompanied by
written tributes to the man we are honoring. Among
the expressions of individuals’ appreciation for John,
are these good examples. “I am honored to be asked
to contribute as I admired and valued John greatly.”
“I greatly appreciate the opportunity to honor John
Guilday’s memory by writing a chapter for his book.”
“As you know John and I had been friends for a
good many years and have co-authored several pa-
pers; I was pleased to learn such a volume honoring
his memory is already in the planning stages.” “I
did, indeed, know that John Guilday had passed
away. Sad, not only because of the unparalleled
quality of his work on Pleistocene and Holocene
North American faunas, but also, on a more per-
sonal level, because he was one of my personal he-
ros, and I have very few.” Such responses indicated
clearly the appropriate nature of our endeavor.

Contributions cover a wide scope of studies on
Quaternary faunas, ranging from faunal studies,
through paleoecological and taphonomic ap-
proaches, to theoretical discussions. The papers are
arranged in a more-or-less systematic order, with
lower vertebrates heading the list. Theoretical pa-
pers bring up the rear, not out of disrespect but
because they tend to summarize evidence from all

systematic groups. We thank all the contributors for
their willingness to participate and their cooperation
and promptness in providing us with well prepared
manuscripts. We are glad to have been able to as-
semble this fine collection of papers devoted to the
Quaternary. Our only regret is that the volume is
unable to contain a paper by John Guilday.

Many of John’s accomplishments were recog-
nized by his peers during his lifetime. Bjorn Kurten
and Elaine Anderson, for example, chose to dedicate
their “Pleistocene Mammals of North America” to
John and Alice. Four species were named for John
during his lifetime; one fossil amphibian, Crypto-
branchus guildayi Holman, 1977; and three fossil
rodents, Leptotomus gui/davi Black, 1971, Proneo-
fiber guildayi Hibbard, 1977, and Microtus guildayi
van der Meulen, 1978. Recognition continues to
come to John posthumously. In this volume, we find
additions to John’s taxonomic spectrum with a new
owl, Otus guildayi, by Brodkorb and Mourer-Chau-
vire, a new genus of arvicoline rodent, Guildayomys,
dedicated by Zakrzewski, and a new subspecies of
dire wolf, Canis dirus guildayi, provided by Kurten.
Promised by Richard Lund, a long time friend of
John, is a new order of Paleozoic fishes.

One gauge of a person’s accomplishments is fre-
quency of citation of his work. Here John rates high-
ly. One typical instance is in the chapter, “Terres-
trial vertebrate faunas” in the recently published
volume 1 of “Late Quaternary Environments of the
United States” in which 18 papers with John as
senior author are cited.

A further posthumous recognition was the award
to John by the Society of American Archaeology of
the Fryxell Award for Interdisciplinary Research.
The tribute with the award appraised this aspect of
his scientific endeavors; “His own research interests,
involving especially cave and fissure-deposited fos-
sil assemblages, led him to publish a number of
seminal studies on Pleistocene-Holocene faunal
changes, which supplied crucial environmental data
on the backgrounds of ancient human communities
in the Northeast. Additionally, he graciously made
his skills available to solving a range of archaeolog-
ically generated problems, from the identification of
the occasional puzzling specimen to the analysis and
interpretation of excavated collections as large as

1

2

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

that from Meadowcroft Rockshelter. The quality of
his work, whether on his own projects or those of
others, set standards for paleontological research and
interdisciplinary cooperation that apply continent-
wide.

“Guilday contributed significantly to a wide range
of scientific problem domains that are shared be-
tween archaeology and paleontology: human be-
havior as manifested in patterns of butchering and
food processing; microclimatic information derived
from archaeological faunal assemblages; tapho-

nomic processes in caves and rockshelters; the dis-
crimination of human from animal contributions to
site bone assemblages; relationships between chang-
ing faunal assemblages and other environmental in-
dicators such as fossil plant associations; patterns
and causes of faunal extinctions; macroclimatic
change and its effects on faunal distributions.”

So our memorial volume joins other acknowl-
edgements of the life and contributions of a truly
remarkable man.

JOHN EDWARD GUILDAY (1925-1982)

Mary R. Dawson

On November 1 7, 1 982, John Guilday, our friend
and colleague to whom this volume is dedicated,
lost his fight against medical complications follow-
ing pneumonia. This was one of John’s few defeats,
for his life was marked by numerous victories, fre-
quently against odds that would have overwhelmed
most mortals. John’s scientific dedication was to the
Quaternary. His contributions to the understanding
of this complex, relatively recent interval of earth
history were outstanding. But John was more than
a dedicated scientist, he was a full, rich human being.
Robert J. Gangewere, editor of Carnegie Magazine,
expressed this well shortly after John’s death, when
he wrote, “In a world in which so many people are
claimed to be unique, John Guilday truly was. Part
of his gift was in being like everyone else while en-
during the most disabling effects of polio. At the
same time he achieved international esteem as a
paleontologist in the Section of Vertebrate Fossils.
The fact that he wrote beautifully, in a remarkable,
poetic way, was the only fact that the general public
may have known about him. But those who knew
him personally, knew immediately that he was un-
forgettable, and that his story was one that would
be with them for life.”

An active, supportive family and early interests
in natural history formed a basic part of the back-
ground for John’s development. He was born in
Pittsburgh, Pennsylvania, on 23 October 1925, first
child of John W. and Margaret McKinley Guilday.
His sister Marjorie, one and one-half years younger,
was his childhood buddy and frequent co-conspir-
ator as well as his lifelong friend. To his second
sister, Patricia, sixteen years younger, John was an
example and idol. John’s bent for natural history
expressed itself through collecting moths and similar
boyhood activities.

A strong influence on John, as a teen-ager, was
his acquaintance with the noted comparative anat-
omist S. Harmsted Chubb. John, living with his
family near New York City in the late 1930s and
early 1940s, met Dr. Chubb during a visit to the
American Museum of Natural History. As John
wrote later, “My first inkling that a bone might be
something other than a passive Halloween prop came
from the comparative anatomist S. Harmstead
Chubb. Dr. Chubb was almost 80 when I first met

him, a small, happy man with bright-blue eyes and
a most elegant, snow-white, impeccably trimmed
Edwardian beard. . . . After we had come to know
one another Dr. Chubb, brimming with an enthu-
siasm that belied his years, would take me from case
to case, pointing out small features that only served
to emphasize the care and accuracy that went into
the making of such skeletal mounts.” Thus, John’s
interest in osteology was fired. He worked as a vol-
unteer preparator in the Osteology section of the
American Museum of Natural History in 1941 and

1942. John and the family moved back to Pitts-
burgh, and John served as a volunteer in the Section
of Mammals, Carnegie Museum, in 1942 and 1943.
He kept in close contact with Dr. Chubb, who con-
tinued to be a strong influence, writing to John in

1 943, “Now as we look ahead toward college it seems
to me that neither the medical or veterinary course
would suit your purpose best. I would rather advise
broader scientific study and then specializing in Zo-
ology and Comparative Anatomy.”

But college had to wait. In 1944 John entered the
United States Army. He had hoped for assignment
to the medical corps, but the army chose the infantry
for him. He was in the 329th “Buckshot” Infantry
Regiment of the 83rd Division. Even during his
army service, John’s interests in natural history
reigned supreme. In March 1944, he wrote to J.
Kenneth Doutt, Curator of Mammals at Carnegie
Museum, from Camp Wheeler, Georgia: “This
afternoon we were having an outdoor lecture on
military courtesy. I’m afraid I didn’t pay much at-
tention to the speaker though because three gray
squirrels were playing tag in a tree right above him.
They seem to be very small in comparison with the
ones I used to see in New York.” To Caroline Hep-
penstall, then Assistant in the Section of Mammals,
Carnegie Museum, he wrote in July, “I’m afraid
from the way things are going so far I’ll get very
little time for collecting mammals, but I’ll try my
best.” John’s service included time in Luxembourg
and Belgium. His regiment fought through the bitter
winter of 1944-1945 in the Battles of the Ardennes
and the Bulge. Through all this stress, John kept up
his interests in natural history. In field notes sent to
Caroline Heppenstall from Europe there is this en-
try: “Tohogne, Prov. Namur, Belgium, Feb. 1 , 1 945.

3

4

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

John Edward Guilday (1925-1982)

1984

DAWSON — GUILDAY OBITUARY

5

While passing through a meadow covered with a
snow, during the first warm spell of the season found
a dead Microtus (?). Measurements— The circum-
stances at the time would not permit the taking of
any type of measurements. Sex (?) dissection im-
possible.” This masterpiece of understatement de-
scribed John’s preoccupations shortly after he took
part in some of the major battles of World War II.
In July 1945, after the war in the European theater
had ended and John was in Austria, he wrote to
Caroline Heppenstall about collecting birds but
complained, “Can’t make many skins though be-
cause a 30 caliber really chews them up . . . .” Such
were the difficulties of being an infantryman with
interests in collecting specimens. Many of the ani-
mals John collected in Europe are in the collections
of Carnegie Museum’s Section of Mammals.

After discharge from the army, John returned to
Pittsburgh and began his undergraduate work in the
Zoology Department at the University of Pitts-
burgh, from which he received his Bachelor of Sci-
ence degree in 1950. While still an undergraduate,
John worked as a biological technician for the Penn-
sylvania Game Commission, working out of Car-
negie Museum between 1 947 and 1 950 on the Penn-
sylvania Mammal Survey. In 1950 he worked also
as a laboratory assistant in the Section of Vertebrate
Fossils. The year 1950 was marked by one of the
best things John ever did — he married Alice Mekeel.
John met Alice in the museum office, where she
worked as secretary to the director. Appropriately
enough, John and Alice spent their honeymoon on
a fossil collecting expedition to Utah, Montana, and
British Columbia with the museum’s vertebrate pa-
leontologist, J. LeRoy Kay. In 1951, John began
graduate work in zoology at the University of Pitts-
burgh, and joined the curatorial ranks in the mu-
seum as Assistant Curator of Comparative Anato-
my. His work included research on the rice rat from
archaeological sites in West Virginia and popular
articles for Carnegie Museum. In many of his sci-
entific and writing projects, Caroline Heppenstall
was his valued mentor.

In the early fall of 1952, when his daughter Kath-
leen was just ten months old, John was struck down
with polio. As John’s long time friend and associate
Allen McCrady wrote, at this time John began “. . .
the most remarkable portion of his career. Although
he had little use of his hands and arms, with Alice’s
unfailing aid and some clever effort-saving devices,
John began turning out an incredible amount of
quality research from his basement lab.” John’s di-

rect interests in Quaternary mammals went back at
least to 1947 when he took part in a museum ex-
pedition to the vicinity of the town of New Paris in
Bedford County, Pennsylvania. The nearly com-
plete skeleton of the extinct eastern wapiti, Cervus
canadensis canadensis, with a flint arrowhead
embedded in its neck was excavated from a sinkhole
in the Helderberg limestone by John and colleagues
from the Pittsburgh Grotto of the National Speleo-
logical Society and the Explorers Club of Pitts-
burgh. So when John’s health allowed him to return
to scholarly pursuits, it was natural that his interests
should become channeled toward animal remains
from archaeological sites and thence to the earlier
parts of the Quaternary. His studies included care-
ful, reasoned analyses of butchering techniques, with
results that have helped many others to determine
what are, and are not, signs of human activity related
to bones. Very important in John’s productivity were
the results of a return, in August 1956, to the New
Paris region by a group from the museum (Roy Kay)
and the Pittsburgh Grotto (including Allen Mc-
Crady, Ralph Bossart, John Leppla). Later joined
by Harold Hamilton and A. C. Lloyd, these people
and a group of enthusiastic volunteers worked on
the excavation of Sinkhole No. 4, or Lloyd’s Rock
Hole, until sterile cave-derived sediments were en-
countered in March 1963. The paper resulting from
this work, co-authored with Paul Martin and Allen
McCrady and published in 1964, demonstrated cli-
matic changes from cool taiga parkland through sub-
sequent warming, and gave, by study of plant and
animal remains, a clear picture of the late Pleisto-
cene environment of non-glaciated Pennsylvania.
The work of John, his co-workers, and dedicated
field crews firmly established the nature of late Pleis-
tocene biota and climates in the eastern United
States, contributing decisively where there had been
almost a void. As John wrote in a review article in
1971, “As late as 1958, Martin stated, ‘Rather than
in eastern North America, the main full-glacial refu-
gia for tundra mammals and birds lay in unglaciated
Alaska.’ This was based upon the almost total lack
of vertebrate paleontological evidence. Since that
time, accelerating research, especially in cave sites,
has changed this picture; and tundra forms, as well
as those characteristic of all northern ecotopes, have
been recovered from at least some part of the Ap-
palachian system.” Later in the same review John
wrote, “It is frustrating to review the history of the
Appalachian mammal fauna from the time of the
Wisconsinan maximum to modern times because

6

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

of the rapidity of new discoveries and the knowledge
that many blanks in the paleontological record will
be filled in within the next few decades. Research
activities within the past 10 years have shown that
these deposits exist in such numbers in areas of karst
that they can be profitably studied at more than a
regional level. Just how refined, geographically, this
can be done remains to be seen; but each year brings
additional reports of sites as cavers become more
aware of the possibilities.” Much of the refinement,
of course, came from the continuing work of John
Guilday. Site after site of Quaternary mammals came
under John’s scrutiny, and his thoughtful analyses
were published. Robinson Cave, Tennessee, Welsh
Cave, Kentucky, Clark’s Cave, Virginia, Baker Bluff,
Tennessee, yielded bones to John’s crews and in-
formation to the studies of John and his co-workers.
Cumberland Cave, Maryland, a pre-Wisconsinan
site, was reopened by Allen McCrady and Harold
Hamilton, John's two closest “field lieutenants.”
Remains from the pre-Wisconsinan Hanover Quar-
ry Fissure of Adams County, Pennsylvania, occu-
pied John as one of the last projects he undertook.
Throughout his career, John also retained his in-
terest in and activity on archaeological sites. The
Buffalo site in West Virginia was a big undertaking,
with more than 61,000 bones identified and ana-
lyzed; the Meadowcroft Rock Shelter was another
nearly Herculean task, with around 300,000 bones
and teeth to be identified. During this time of re-
search productivity, John was also an active con-
tributor to other museum activities.

In 1962 he was promoted to Associate Curator of
Comparative Anatomy, and in 1964 joined the Sec-
tion of Vertebrate Fossils. His contributions to the
educational and exhibition projects of the museum
were always valuable.

John’s productivity is known; the full impact of
his work is difficult to assess so soon, but it can be
anticipated at least partly from a study of his bib-
liography. Never one to be daunted by investiga-
tions avoided by others, John undertook and com-
pleted difficult studies of variation. His scientific
contributions include analyses of variation in Bla-
rina and, surely even more difficult, of dental vari-
ation in three species of Microtus. Bob Gangewere,
though not a scientist himself, had the proper per-
spective to summarize John’s contributions very
well: “After surviving some of the bitterest fighting
of World War II unharmed, he returned to Pitts-
burgh to resume his career in the museum and con-
tracted polio. The ravages of this disease left him

permanently disabled, and his periodic use of the
iron lung in his home was essential to his health.
His visits to the museum were extremely rare, and
many of his colleagues never laid eyes on him.
Nevertheless, with the constant help of his wife Al-
ice, his scientific work at home prospered, and as a
paleontologist his reputation spread around the
world. The museum directors under whom John
served, and all of his closest associates, knew that
John Guilday was one of the research strengths of
Carnegie Museum of Natural History.”

A glance at John’s bibliography shows another
aspect of his productivity, its extension to popular-
izations of science. He contributed frequently to
journals such as The Netherworld News, Pennsyl-
vania Game News, Carnegie Magazine, and The
Explorer. Enticing titles such as “New Paris Sinks,”
“Scien-terrific Names,” “A Rat is a Rat is a Rat?,”
“The Sturgeon Surgeon,” “The Faunagraph Rec-
ord” (how I groaned when he told me that one!),
“Confessions of an Aging Paleontologist” make it
almost impossible not to seek out and read the ar-
ticles. The popularizations contained good science
so well written and interestingly presented that any
reader would finish an article by Guilday richer in
knowledge as well as understanding.

One of my favorites was John’s “Ask the Ex-
perts . . . ,” published in 1970 in The Explorer. I was
involved in the case, as it fell to me to go and look
at a very large bone in a public park in Martinsburg,
Pennsylvania. John had seen photographs of this
enormous specimen, more than six feet across, and
began his odyssey about a specimen that he first
imagined to be a giant muskox. As Associate Di-
rector James Swauger and I set off to see the spec-
imen, John stayed home thinking about it, as he
later wrote: “Meanwhile, Guilday is pouring over
the pictures with a ten-power magnifying glass, his
feet lightly touching the floor now and again. But
still no details. Assume that we are looking at the
skull from the top, then the horns sweep first down,
then out and forward in good ovibovine fashion, in
which case, given the relative dimensions of modern
members of that subfamily, the beast must have
stood no less than eight feet high at the shoulder to
accommodate such a tremendous skull! If, on the
other hand, the skull is upside down, then the horns
would sweep in the opposite manner, somewhat
reminiscent of a bison. Now there are bison known
from early Pleistocene deposits that were very large
indeed. Bison latifrons had a horn span of ten feet,
but the horns were not as massive and the skull itself

1984

DAWSON — GUILDAY OBITUARY

7

much smaller than whatever it is in the park shelter
in Martinsburg. Assignment to the Bison latifrons
group would create additional problems if the skull
had originated at Frankstown Cave. The fauna from
that cave does not appear to be as old as these fossil
bison are known to be. If it were indeed a bison,
shoulder height would then be, counting hump and
all, truly as high as that of an elephant.

“Guilday spends all night turning restlessly, goad-
ed on all sides by great blue oxen of Bunyanesque
proportion, mentally writing and rewriting descrip-
tions of what must surely be a new genus and species.
Preposterobos'. Preposterobos incredibilousl First
thing to do. Photograph Pittsburgh Pleistocene peo-
ple standing beside skull, bearing small placard EAT
YOUR HEART OUT, CLAYTON! (Dr. Clayton E.
Ray, U.S. National Museum, foremost authority on
fossil muskoxen). But Walter Mitty thoughts start
muddying up this euphoric dream. Suppose a truck
hits it between now and tomorrow? What condition
is it in? Can we get permission to borrow the spec-
imen and study it? Must get some sleep. But sleep
comes fitfully to him whose bed vibrates to the col-
lective hooves of multi-ton muskoxen — especially
scientifically undescribed ones — who roll their mas-
sive eyeballs as they deftly hook up and carry off
one’s bedsheets as they thunder through the room.

“The Swauger-Dawson expedition returns from a
first-hand assessment of the situation and dull thuds
can be heard throughout the paleo section as every-
one returns to earth. Everyone except Curator Ber-
man, whose focus is upon Paleozoic reptiles and
who is more inclined to excitement over things with
scales and forked tongues.

“‘It’ turns out to be the base of a skull of a large
whale donated to the town of Martinsburg by a pa-
triotic citizen to form part of a sailors’ memorial in
the park. The mighty horn cores are but the rem-
nants of the glenoid fossae — which afford articula-
tion for the lower jaw.” John goes on with, “The
moral of this story being never to show two pale-
ontologists, one from Michigan and one from west-
ern Pennsylvania, any part of a whale skull from
high in the Appalachian Mountains. It gets them
too excited.”

John’s popular writing was a marvelous combi-
nation of science, good style, and a gently satiric
poking fun at himself and others. His abilities along
this line became recognized not only in Pittsburgh
and to his immediate colleagues but also to a wider
audience. Bob Gangewere noted, in a postscript to
John’s “Conjure Bones”: “An editor would be re-

miss if he did not point out, at least once, that the
writing of Johnny Guilday is very special indeed,
and belongs in a class of popular writing about nat-
ural history subjects that few authors ever attain.
The combination of scientific acumen and poetic
insight is remarkable, as more than one reader has
noted in the past few years. In fact, a museum ad-
ministrator from Florida once observed that John’s
essay defining a natural history museum is the best
thing he ever read on the subject, and is required
reading for his college students in a museum studies
course. That essay is ‘The Long Journey’.”

In “Caves, Bones and Ice Age Owls” John ex-
plained his science in terms full of meaning for sci-
entists and laymen alike: “Science proceeds in no
set direction, neither forward toward an unknown
future, nor backward to our ultimate beginnings, but
radially toward that horizon separating the known
from the unknown that stretches unbroken, like the
rim of the sky, there in whatever direction we turn.
So that research directed toward one point becomes
a part of a circuitous whole that expands not by
increments, but by ripples from the common center
upon which we stand, the whole and its parts in-
extricably mixed. And, that is why we pick up little
bones.”

Very little known except to close friends was John’s
very considerable ability as an artist. Only Alice can
explain just how he did it, but John produced Christ-
mas cards, an illustrated Book of Owls for McCrady,
Book of Rabbits for Dawson, Book of Birds and
Butterflies for Harry and Mary Clench, all aptly cap-
tioned by Alice and Kathleen. He made tiles, and
his watercolors were masterpieces of oriental-style
beauty.

John’s poetry ranged from whimsical to the se-
rious. The lighter side of his writing is well illus-
trated by his Thoughts on Migration :

Oh, little feathered dinosaur.

Kin to croc and pterosaur.

Tell us, oh continent-spanning babbler.

Were you the original Archosaur Traveler?

John Guilday’s friends and colleagues gathered
together to memorialize him on 1 9 November 1 982.
M. Graham Netting, Director Emeritus of Carnegie
Museum, under whom John had worked for many
years, shared his impression of John — excellent sci-
entist, keen, sensitive observer; a man, if he once
had faults had had them burned away by polio.
Leonard Krishtalka, fellow Associate Curator in
Vertebrate Fossils, named John, “A man for all rea-

8 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 8

Water Shrew. An original painting in color on tile by John E. Guilday. John’s feelings about shrews are expressed by the following two
quotations from his own writings: “They are nervous, high-strung little beasts with the energy of junior executives.” “A shrew’s head
looks as if it had freshly emerged from a pencil sharpener. In life its rubbery little trunk is constantly bobbing, testing, probing; in death
the little snout sticks stiffly forward, the trade mark of a shrew.”

sons.” Bob Gangewere concluded his tribute to John
with a lovely, apt summation of the man: “In many
ways John Guilday’s life was rich and normal, al-
though circumscribed. Alice and John had reason
to be proud of their daughter, Kathleen, and their
home was full, full of the everyday business of their
lives, and of the scientific work. In the basement
John had the evidence, the information, that he
needed. To the untutored eye an incredible collec-
tion of miniature bones from ages past — to him, a
world of meaning about past life, past environ-
ments.

BIBLIOGRAPHY

Compiled by Alice Guilday

1948 1949

The fallen monarch. Carnegie Mag., 22:46-49. Winter food of Pennsylvania mink. Pennsylvania Game News,

Little brown bats copulating in winter. J. Mamm., 29:416-417. 20(9): 12, 32.

“For many years John was a contributor to Car-
negie Magazine, where the poetic side of his sci-
entific mind found a natural outlet. Readers of his
prose knew that word for word he was one of the
best writers about nature they were ever likely to
encounter. In 1981 he published his series of ob-
servations about the months of the year. His last
thought about November, the month in which he
died, was, ‘It snows again, and deep, and the stars
draw closer.’”

1984

DAWSON -GUILDAY OBITUARY

9

1950

Winter fetus in the little brown bat, Myotis lucifugus. J. Mamm.,
31:96-97.

1951

What are fossils? Carnegie Mag., 25:352-354.

Sexual dimorphism in the pelvic girdle of Microtus pennsylvan-
icus. J. Mamm., 32:216-217.

1952

Rhinoceroses— a family tree. Camegie Mag., 26:388-390.

An occurrence of the rice rat ( Oryzomys ) in West Virginia. J.
Mamm., 33:253-255 (with W. J. Mayer-Oakes).

1955

Some biological aspects of archeology. Archeological Newsletter,
10:5-7.

Animal remains from the Globe Hill site 46Hk34-l. Appendix
1. Pp. 27-28, in The Globe Hill shell heap (site 46Hk34-l),
Hancock County, West Va. (W. J. Mayer-Oakes), West Vir-
ginia Archeological Soc., Inc., Publ. Ser., 3:1-34.

Animal remains from an Indian village site, Indiana County,
Pennsylvania. Pennsylvania Archaeologist, 25:142-147.

1956

Archeological evidence of the fisher in West Virginia. J. Mamm.,
37:287.

[Letter to the editor]. Netherworld News, 4:31 and 4:68. [Re-
printed in Speleo Digest 1956, Pittsburgh Grotto Press, Natl.
Speleol. Soc., Pittsburgh, PA.]

New Paris sinks. Netherworld News, 4:85-86. [Reprinted in Spe-
leo Digest 1956, Pittsburgh Grotto Press, Natl. Speleol. Soc.,
Pittsburgh, PA.]

Fossils and New Paris. Netherworld News, 4:113-114. [Reprint-
ed in Speleo Digest 1956, Pittsburgh Grotto Press, Natl.
Speleol. Soc., Pittsburgh, PA.]

Pennsylvania mammals— a history. Pennsylvania Game News,
27(12):20-26.

1957

Pennsylvania’s biggest game — the whale. Pennsylvania Game
News, 28(4):26-28. [Reprinted, pp. 253-255 in Pennsyl-
vania Game News Treasury, Pennsylvania Game Commis-
sion, Harrisburg, 528 pp., 1979.]

[Reviews of] Mammals of the air, land and waters of the world,
by G. G. Goodwin; and Living mammals of the world, by
L T. Sanderson. Camegie Mag., 31:212-213.

Scien-terrific names. Pennsylvania Game News, 28(8):3 1—33.

[Field trip log for July] June 18. Netherworld News, 5:82.

[Letter to the editor]. Netherworld News, 5:89.

A bat’s wing. Netherworld News, 5:175-176. [Reprinted in Pow-
dermill Nature Reserve Educational Release No. 10, Car-
negie Mus., 7 March 1958; and in Speleo Digest 1957, Pitts-
burgh Grotto Press, Natl. Speleol. Soc., Pittsburgh, PA.]

The masked shrew at Powdermill Nature Reserve. Powdermill
Nature Reserve Educational Release No. 9, Camegie Mus.,
1 November 1957.

Individual and geographic variation in Blarina brevicauda from
Pennsylvania. Ann. Camegie Mus., 35:41-68.

1958

The mastodon — forest elephant. Pennsylvania Game News, 29(7):
15-17.

Price tag on Penn’s Woods? Pennsylvania Game News, 29(9):
42-43.

The prehistoric distribution of the opossum. J. Mamm., 39:39-
43.

Errata. Netherworld News, 6:28.

Report from New Paris. Netherworld News, 6:31. [Reprinted in
Speleo Digest 1958, Pittsburgh Grotto Press, Natl. Speleol.
Soc., Pittsburgh, PA.]

Educated guessing. Netherworld News, 6:39-42.

[Review of] The geology of the Hidden Valley Boy Scout area.
Perry County, Pennsylvania, by J. T. Miller. Netherworld
News, 6:9 1 .

A rat is a rat is a rat! Netherworld News, 6:97-99. [Reprinted in
Speleo Digest 1958, Pittsburgh Grotto Press, Natl. Speleol.
Soc., Pittsburgh, PA.]

So what’s a peccary? Netherworld News, 6:165-167. [Reprinted
in Speleo Digest 1958, Pittsburgh Grotto Press, Natl. Speleol.
Soc., Pittsburgh, PA.]

And some like it cold. Netherworld News, 6:252-254.

Reports from New Pans. Netherworld News, 6:22-23 (with N.
D. Richmond).

A Recent fissure deposit in Bedford County, Pennsylvania. Ann.
Camegie Mus., 35:127-138 (with M. S. Bender).

Glacier bear. Camegie Mag., 32:185-189 (with O. Agathon, se-
nior author).

[Review of] Bones for the archaeologist, by I. W. Cornwall. Amer.
Antiquity, 23:441 .

Time and Powdermill. Powdermill Nature Reserve Educational
Release No. 16, Camegie Mus., 11 September 1958. [Re-
printed in Camegie Mag., 34:10; and Camegie Mag., 52:31].

1959

[Review of] The great chain of life, by J. W. Krutch. Camegie
Mag., 33:32-33.

The flintlock forest. Camegie Mag., 33:263-266, 269.

Box score— New Paris. Netherworld News, 7:55.

Now that you’ve got 'em — . Netherworld News, 7:96-98.

[Review of] Soils for the archaeologist, by I. W. Cornwall. Neth-
erworld News, 7:140-141.

Slurp-slurp-slurp. Netherworld News, 7:199-200.

Say — A-a-a-a-n. Pennsylvania Game News, 30( 1 2):45— 46.

Muskrats at the Reserve. Powdermill Nature Reserve Educa-
tional Release No. 23, Camegie Mus., 28 August 1959.

1960

Mouse bones. Netherworld News, 8:3-5. [Reprinted in Nether-
world News, 17:140-143.]

Of microtines and men. Netherworld News, 8:127-129. [Re-
printed in Speleo Digest 1960, Pittsburgh Grotto Press, Natl.
Speleol. Soc., Pittsburgh, PA.]

Kangaroo rats. All-Pets Mag., 3 1(1): 19-20.

The Fannin pedigree. Camegie Mag., 34:307-309, 311.

. . . there’s science in dump rooting. Pennsylvania Angler, 29(1 1 ):
14-15.

Late Pleistocene records of the yellow-cheeked vole, Microtus
xanthognathus Leach. Ann. Camegie Mus., 35:315-330 (with
M. S. Bender).

10

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

April and Powdermill. Powdermill Nature Reserve Educational
Release No. 28, Carnegie Mus., 13 April 1960. [Reprinted
in Carnegie Mag., 41:1 13-1 14; and Carnegie Mag., 52:7.]

1961

Vertebrate remains from the Varner Site (36-Gr-l). Pennsylvania
Archaeologist, 31:119-124.

Prehistoric record of Scalopus from western Pennsylvania. J.
Mamm., 42:117-118.

Plecotus from the Pleistocene of Pennsylvania. J. Mamm., 42:
402-403.

Abnormal lower 3rd molar in Odocoileus. J. Mamm., 42:551-
553.

The sky dance. Pennsylvania Game News, 32(4):44— 45.
Jaguar— old style. Netherworld News, 9:41-43. [Reprinted in
Speleo Digest 1961, Pittsburgh Grotto Press, Natl. Speleol.
Soc., Pittsburgh, PA.]

Bones from Field House Cave, Pendleton County, West Virginia.
Netherworld News, 9:93.

Bones from Jones Quarry Cave, 1 mile NE of Falling Waters,
Berkeley County, West Virginia. Netherworld News, 9:94.
“Because it’s there.” Netherworld News, 9:105.

New cave organism discovered. Netherworld News, 9:127-128.
[Reprinted in Speleo Digest, 1961, Pittsburgh Grotto Press,
Natl. Speleol. Soc., Pittsburgh, PA.]

The pterosaur and the lemming. Netherworld News, 9: 1 42-144.
The Ten Commandments of caving. Netherworld News, 9:197.
[Reprinted in Speleo Digest, 1961, Pittsburgh Grotto Press,
Natl. Speleol. Soc., Pittsburgh, PA.]

The rice rat riddle. SPAAC Speaks, 1:23-25.

The collared lemming (Dicrostonyx) from the Pennsylvania Pleis-
tocene. Proc. Biol. Soc. Washington, 74:249-250 (with J. K.
Doutt).

1962

Notes on Pleistocene vertebrates from Wythe County, Virginia.
Ann. Carnegie Mus., 36:77-86.

The Pleistocene local fauna of the Natural Chimneys, Augusta
County, Virginia. Ann. Carnegie Mus., 36:87-122.

Bird remains from Pennsylvania archaeological sites. Sandpiper
(Meadville, PA), 4:75-80.

The headhunters. Netherworld News, 10:53-54. [Reprinted in
Speleo Digest 1962, Pittsburgh Grotto Press, Natl. Speleol.
Soc., Pittsburgh, PA.]

The comings and goings of squirrels. Netherworld News, 1 0:67—
70.

The case of the missing possam [sic]. SPAAC Speaks, 2:27-30.
Refuse and the archaeologist. Compost Science, 3(3):20-21.
Supernumerary molars of Otocyon. J. Mamm., 43:455-462.
The deer— of 350 years ago. Pennsylvania Game News, 33(12):
10-13.

Portraits of the season. Pennsylvania Game News, 33:(1)47; (2)42;
(3)58; (4)45; (5)11; (6)6; (7)25; (8)26; (9)11; (10)19; (11)7;
(12)9. [Reprinted in Carnegie Mag. See 1981.]

The sturgeon surgeon. Pennsylvania Angler, 3 1 (2):9.

Bone refuse from the Bristol Hills Site, Can 29-3. Manuscript in
files of The Rochester Mus. Arts and Sciences, Rochester,
NY.

The woodland jumping mouse and its kin. Powdermill Nature
Reserve Educational Release No. 42, Carnegie Mus., 9 No-
vember 1962.

Aboriginal butchering techniques at the Eschelman Site (36 La
12), Lancaster County, Pennsylvania. Pennsylvania Archae-
ologist, 32:59-83 (with P. W. Parmalee and D. P. Tanner).
Animal remains from the Quaker State Rockshelter (36 Ve 27),
Venango County, Pennsylvania. Pennsylvania Archaeolo-
gist, 32:131-137 (with D. P. Tanner).

1963

Bone refuse from the Oakfield Site, Genesee County, New York.

Pennsylvania Archaeologist, 33:12-15.

Evidence for buffalo in prehistoric Pennsylvania. Pennsylvania
Archaeologist, 33:135-139.

The cup-and-pin game. Pennsylvania Archaeologist, 33:159-163.
Pleistocene zoogeography of the collared lemming ( Dicrostonyx ).
Evolution, 17:194-197. [Reprinted (1) pp. 509-512, in
Readings in mammalogy (J. K. Jones, Jr. and S. Anderson,
eds.), Mus. Nat. Hist., Univ. Kansas Monogr., 2:1-586, 1970;
and (2) pp. 547-550, in Selected readings in mammalogy (J.
K. Jones, Jr., S. Anderson, and R. S. Hoffmann, eds.) Mus.
Nat. Hist., Univ. Kansas Monogr., 5:1-640, 1976.]

The faunagraph record. Carnegie Mag., 37:233-237.

Rabbits is rabbits. Pennsylvania Game News, 34(5): 30—3 1 .

Of rod and gun. Pennsylvania Game News, 34(9): 15-1 6.

The lemming— New Paris. Pp. 2-72 — 2-73, in Speleo Digest
1961 (H. L. Black and A. P. Haarr, eds), Pittsburgh Grotto
Press, Natl. Speleol. Soc., Pittsburgh, PA, 409 pp.

Time. Netherworld News, 1 1:39-40.

Final report— late Pleistocene fauna of New Paris No. 4 (Lloyds
Rock Hole). Netherworld News, 11:153-162 (with A. D.
McCrady, senior author). [Reprinted in Speleo Digest 1963,
Pittsburgh Grotto Press, Natl. Speleol. Soc., Pittsburgh, PA.]
The Clarksville deer— a case history. Amer. Antiquity, 29:109-
1 1 1 (with A. D. McCrady).

1964

New Paris No. 4: A Pleistocene cave deposit in Bedford County,
Pennsylvania. Bull. Natl. Speleol. Soc., 26:121-194 (with P.
S. Martin and A. D. McCrady).

Summary of artifacts, Pp. 9-1 1, in The Footer Site (Can 29-3),
mimeographed archaeological site report produced by A. J.
Parker, 62 Hinkleyville Road, Spencerport, NY, 22 pp.
First record of the least shrew for Powdermill and for West-
moreland County. Powdermill Nature Reserve Educational
Release No. 50, Carnegie Mus., 5 October 1964.

The Cumberland caper. Netherworld News, 12:159-162.

1965

Don’t tell the kids. Carnegie Mag., 39:191-194.

The Ice Age Heffalump trap. Canadian Audubon, 27:1 10-1 14.
Animal remains from the Sheep Rock Shelter (36 Hu 1), Hun-
tingdon County, Pennsylvania. Pennsylvania Archaeologist,
35:34-49 (with P. W. Parmalee).

Quaternary mammals of North America. Pp. 509-525, in The
Quaternary of the United States (H. E. Wright, Jr. and D.
D. Frey, eds.) Princeton Univ. Press, 922 pp. (with C. W.
Hibbard, senior author, C. E. Ray, D. E. Savage, and D. W.
Taylor).

Vertebrate remains from the Mount Carbon Site, (46-Fa-7), Fay-
ette County, West Virginia. The West Virginia Archeologist,
18:1-14 (with D. P. Tanner).

1984

DAWSON — GUILDAY OBITUARY

1 1

1966

Rangifer antler from an Ohio bog. J. Mamra., 47:325-326.

The bone breccia of Bootlegger Sink, York County, Pa. Ann.
Carnegie Mus., 38:145-163 (with H. W. Hamilton and A.
D. McCrady).

Armadillo remains from Tennessee and West Virginia caves.

Bull. Natl. Speleol. Soc. 28:183-184 (with A. D. McCrady).
Lower third molar in Procyon. J. Mamm., 47:149-151 (with P.
W. Parmalee).

Mammals of Pennsylvania. Pennsylvania Game Commission,
Harrisburg, 273 pp. (with J. K. Doutt and C. A. Heppenstall,
senior authors).

A Recent record of porcupine from Tennessee. J. Tennessee Acad.

Sci., 6 1 (3):8 1 —82 (with P. W. Parmalee, senior author).

The vertebrate fauna. Appendix ii, in Mound City revisited, by
J. A. Brown and R. S. Baby. Report of the Ohio Historical
Society to the National Parks Service, June 1966, 103 pp.
(with D. P. Tanner).

Animal remains from the Westmoreland-Barber Site (40 Mi- 1 1 ),
Marion County, Tennessee. Appendix A. Pp. 138-145, in
Westmoreland-Barber Site (40 Mi- 1 1 ), Nickajack Reservoir,
season 1 1, by C. H. Faulkner and J. B. Graham. Dept. An-
thropology, Univ. Tennessee, Knoxville, 150 pp. (with D.
P. Tanner).

Powdermill September. Powdermill Nature Reserve Educational
Release No. 66, Carnegie Mus., 21 September 1966. [Re-
printed in Carnegie Mag., 42:224; and Carnegie Mag., 52:
14.]

1967

The climatic significance of the Hosterman’s Pit local fauna,
Centre County, Pennsylvania. Amer. Antiquity, 32:231-232.
Notes on the Pleistocene big brown bat ( Eptesicus grandis
(Brown)). Ann. Carnegie Mus., 39:105-114.

Differential extinction during late-Pleistocene and Recent times.
Pp. 121-140, in Pleistocene extinctions— the search for a
cause (P. S. Martin and H. E. Wright, Jr., eds.), Yale Univ.
Press, New Haven, 453 pp.

Evolution’s test tube. [Review of] The life of the cave, by C. E.

Mohr and T. L. Poulson. Carnegie Mag. 41:195-197.
Vertebrate remains from the Banshee Hole, Cumberland County,
Tennessee. Speleotype, 2:44-45. [Reprinted in Speleo Digest
1967, Natl. Speleol. Soc.]

What skull is that? Pennsylvania Game News, 38(4):9— 1 1.
Trout fishing. Netherworld News, 15:188-192.

Small mammal remains from Jaguar Cave, Lemhi County, Idaho.
Tebiwa, J. Idaho State Univ. Mus., 10:26-36 (with E. K.
Adam).

Animal remains from Homed Owl Cave, Albany County, Wy-
oming. Contrib. Geol., Univ. Wyoming, 6:97-99 (with H.
W. Hamilton and E. K. Adam).

A new Peromyscus (Rodentia: Cricetidae) from the Pleistocene
of Maryland. Ann. Carnegie Mus., 39:91-103 (with C. O.
Handley, Jr.).

Extinct Florida spectacled bear Tremarctos floridanus (Gidley)
from central Tennessee. Bull. Natl. Speleol. Soc., 29:149-
162 (with D. C. Irving).

A bestiary for Pleistocene biologists. Pp. 1-62, in Pleistocene
extinctions— the search for a cause (P. S. Martin and H. E.
Wright, Jr., eds.), Yale Univ. Press, New Haven, 453 pp.
(with P. S. Martin, senior author).

1968

Grizzly bears from eastern North America. Amer. Midland Nat..
79:247-250.

Archaeological evidence of caribou from New York and Mas-
sachusetts. J. Mamm., 49:344-345.

Pleistocene zoogeography of the lemming Dicrostonyx, a reeval-
uation. Univ. Colorado Studies Ser. Earth Sci., 6:61-71.

Cope’s mouse. Turtox News, 46:333-335.

[Review of] Recent mammals of the world, by S. Anderson and
J. K. Jones, Jr. Ecology, 49:190.

Vertebrate remains from the Fairchance Mound (46 Mr 13),
Marshall County, West Virginia. West Virginia Archeologist,
21:41-54 (with D. P. Tanner).

The rains of May. Powdermill Nature Reserve Educational Re-
lease No. 76, Carnegie Mus., 29 May 1968.

[Review of] Quaternary extinction of large mammals, by D. I.
Axelrod. J. Paleontol., 42:1319.

1969

Faunal remains from Dutchess Quarry Cave No. 1. Pp. 17-19,
in The archeology of Dutchess Quarry Cave, Orange County,
New York (R. E. Funk, G. R. Walters, and W. F. Ehlers,
Jr.), Pennsylvania Archaeologist, 39:7-22.

Small mammal remains from the Wasden Site (Owl Cave),
Bonneville County, Idaho. Tebiwa, J. Idaho State Univ. Mus.,
12:47-57.

A possible caribou-Paleo-Indian association from Dutchess
Quarry Cave, Orange County, New York. Bull. New York
State Archeological Assoc., 45:24-29.

This planet’s diversity of life. [Review of] Vertebrates, by C. J.
McCoy, Jr. Carnegie Mag., 43:1 17.

[Review of] Pleistocene mammals of Europe, by B. Kurten. Ecol-
ogy, 50:531.

Mastodon and Paleo-Indian in West Virginia. West Virginia Ar-
cheologist. 22:1-3 (with D. S Berman).

The Pleistocene vertebrate fauna of Robinson Cave, Overton
County, Tennessee. Palaeovertebrata, 2(2):25— 7 5 (with H.
Hamilton and A. D. McCrady).

Pleistocene and Recent vertebrate faunas from Crankshaft Cave,
Missouri. Illinois State Mus. Repts. Investigations, 14:1-37
(with P. W. Parmalee and R. D. Oesch, senior authors).

1970

Animal remains from archaeological excavations at Fort Ligo-
nier. Ann. Carnegie Mus., 42:177-186. [Reprinted, pp. 121 —
132, in Experimental archeology (D. Ingersoll, J. E. Yellon,
and W. McDonald, eds.), Columbia Univ. Press, New York,
423 pp.]

Ask the experts. The Explorer, 1 2(4):20— 2 1 .

Leviathan. [Review of] The year of the whale, by V. Scheffer.
Carnegie Mag., 44:59-61.

1971

The Pleistocene history of the Appalachian mammal fauna. Pp.
232-262, in The distributional history of the biota of the
Southern Appalachians, Part III: Vertebrates (P. C. Holt,
ed.), Virginia Polytechnic Inst, and State Univ.. Res. Div.
Monogr., 4: 1-306.

Two letters to Steve Callen. Pp. 43-44, in Discovery of a mas-
todon tooth— Part V. Karst Kaver, Monongahela Grotto,
Natl. Speleol. Soc., 5:39-44.

12

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Biological and archeological analysis of bones from a 17th Cen-
tury Indian village (46 Pu 31), Putnam County, West Vir-
ginia. West Virginia Geol. and Econ. Surv., Rept. Archeo-
logical Investigations, 4:1-64.

J. LeRoy Kay. Carnegie Mag., 45:296.

Big game of the Pleistocene. Pp. 46-52, in North American big
game (R. C. Alberts, ed.), Boone and Crockett Club, Pitts-
burgh, 403 pp.

Penn wood folk. Pennsylvania Game News, 42:15-16.

The Welsh Cave peccaries {Platygonus) and associated fauna,
Kentucky Pleistocene. Ann. Carnegie Mus., 43:249-320 (with
H. W. Hamilton and A. D. McCrady).

Thirteen-lined ground squirrel, prairie chicken and other verte-
brates from an archeological site in northeastern Arkansas.
Amer. Midland Nat., 86:227-229 (with P. W. Parmalee).

1972

[Introduction to] Mice, men and mastodons, by R. S. Mills. The
Explorer, 14(2):9-12.

Archaeological evidence of Scalopus aquaticus in the Upper Ohio
Valley. J. Mamm., 53:905-907.

Mouse flowers. The Explorer, 14(4): 19-21.

Bear Sink— a pitch. Netherworld News, 20:4-6.

Jaguar ( Panthera onca) remains from Big Bone Cave, Tennessee
and east central North America. Bull. Natl. Speleol. Soc.,
34:1-14 (with H. McGinnis).

Quaternary periglacial records of voles of the genus Phenacomys
Memam (Cricetidae: Rodentia). Quat. Res., 2:170-175 (with
P. W. Parmalee).

First record of Cervalces scotti Ledekker from the Pleistocene of
Ohio. Amer. Midland Nat., 88:255 (with R. S. Mills, senior
author).

1973

Confessions of an aging paleontologist. The Explorer, 15:25-26.

The late Pleistocene small mammals of Eagle Cave, Pendleton
County, West Virginia. Ann. Carnegie Mus., 44:45-58 (with
H. W. Hamilton).

1974

Arctic life in the South (letter to editor). Geotimes, 19:8.

Caribou (Rangifer tarandus L.) from the Pleistocene of Tennes-
see. J. Tennessee Acad. Sci., 50:109-1 12 (with H. W. Ham-
ilton and P. W. Parmalee).

[Review of] The children of pride, edited by R. M. Myers. Car-
negie Mag., 49:39.

1975

Faunal remains from the Zebree Site. Appendix I. Pp. 228-234,
in Report of excavations at the Zebree Site 1969 (D. F.
Morse), Research Rept. No. 4, Arkansas Archeological Surv.,
Fayetteville, 246 pp. (with P. W. Parmalee).

Extinct peccary ( Platygonus compressus LeConte) from a central
Kentucky cave. Bull. Natl. Speleol. Soc., 37:83-87 (with R.
C. Wilson, senior author, and J. A. Branstetter, junior au-
thor).

1976

Appalachian bone caves. Pp. 88-103, in Geology and biology of
Pennsylvania caves (W. B. White, ed.), Pennsylvania Geol.
Surv., General Geol. Rept., 66:1-103.

Owls and Ice Age caves. Netherworld News, 24:86-89.

First records of the giant beaver ( Cast oroides ohioensis) from
eastern Tennessee. J. Tennessee Acad. Sci., 51:87-88 (with
P. W. Parmalee and A. E. Bogan, senior authors).

1977

Sabertooth cat, Smilodon floridanus (Leidy), and associated fau-
na from a Tennessee cave (40 Dv 40), the First American
Bank site. J. Tennessee Acad. Sci., 52:84-94.

The long journey. Carnegie Mag., 5 1 (2):76— 79.

The pigeons of Meadowcroft. Carnegie Mag., 51(1 0):33— 37.

Mysterious mouse. Rustlings from the Grove. Newsletter, Grove
Heritage Assoc. (Glenview, IL), 9:5.

The Clark’s Cave bone deposit and the late Pleistocene paleo-
ecology of the central Appalachian Mountains of Virginia.
Bull. Carnegie Mus. Nat. Hist., 2:1-87 (with P. W. Parmalee
and H. W. Hamilton).

[Selection, pp. 70-77, in] Meadowcroft Rockshelter: retrospect
1976, by J. M. Adovasio, J. D. Gunn, J. Donahue, and R.
Stuckenrath. Pennsylvania Archaeologist, 47:(2-3): 1-93.

1978

[Review ol] A guide to the measurement of animal bones from
archaeological sites, by A. von den Dnesch. Amer. Anthro-
pologist, 80: 180.

Ecological significance of displaced boreal mammals in West
Virginia caves. J. Mamm., 59:176-181 (with H. W. Ham-
ilton).

The Baker Bluff cave deposit, Tennessee, and the late Pleistocene
faunal gradient. Bull. Carnegie Mus. Nat. Hist., 1 1:1-67 (with
H. W. Hamilton, E. Anderson, and P. W. Parmalee).

The Pleistocene mammalian fauna of Harrodsburg Crevice,
Monroe County, Indiana. Bull. Natl. Speleol. Soc., 40:64-
75 (with P. W. Parmalee and P. J. Munson, senior authors).

Meadowcroft Rockshelter. Pp. 1 40- 1 80, in Early man in America
(A. L. Bryan, ed.), Occas. Papers, Dept. Anthropology, Univ.
Alberta 1:1-327 (with J. M. Adovasio, J. D. Gunn, J. Don-
ahue, R. Stuckenrath, senior authors, and K. Lord, junior
author).

1979

Eastern North American Pleistocene Ochotona (Lagomorpha:
Mammalia). Ann. Carnegie Mus., 48:435-444.

The star-nosed mole/A Polish connection. Carnegie Mag., 53(1):
18-25.

Section of Vertebrate Fossils, Carnegie Museum of Natural His-
tory. Carnegie Mag., 53(8): 18-27 (with E. Hill).

Pleistocene and Recent vertebrate remains from Savage Cave
(15L011), Kentucky. Pp. 5-10, in Western Kentucky Spe-
leol. Surv. Ann. Rept. 1979 (J. E. Mylroie, ed.), 84 pp. (with
P. W. Parmalee).

Meadowcroft Rockshelter— retrospect 1979: part 1. North Amer.
Archeologist, 1( 1):3— 44 (with J. M. Adovasio, J. D. Gunn,
J. Donahue, R. Stuckenrath, senior authors, and K. Lord,
junior author).

1980

Vertebrate faunal remains from Meadowcroft Rockshelter
(36WH297), Washington County, Pennsylvania. Manu-
script on file, Dept. Anthropology, Univ. Pittsburgh, 140 pp.

1984

DAWSON -GUILDAY OBITUARY

13

(with P. W. Parmalee and R. C. Wilson). [To be published
in Meadowcroft Rockshelter and the archaeology of the Cross
Creek drainage, J. M. Adovasio et al., Univ. Pittsburgh Press.]
Meadowcroft Rockshelter— retrospect 1977: part 2. North Amer.
Archaeologist, 1(2):99— 1 37 (with J. M. Adovasio, J. D. Gunn,

J. Donahue, R. Stuckenrath, senior authors, K. Lord, and

K. Volman, junior authors).

Yes Virginia, it really is that old: a reply to Haynes and Mead.
Amer. Antiq., 45:588-595 (with J. M. Adovasio, J. D. Gunn,
J. Donahue, R. Stuckenrath, senior authors, and, K. Volman,
junior author).

Cavern de Saint-Elzear-de-Bonaventure, rapport preliminaire sur
ies fouilles de 1977 et 1978, Ministere de 1’Energie et des
Ressources, Direction Generate de la Recherche Geolgique
et Minerale, Direction de la Geologie, 3 1 pp. (with P. LaSalle,
senior author).

Additional comments on the Pleistocene mammalian fauna of
Harrodsburg Crevice, Monroe County, Indiana. Bull. Natl.
Speleol. Soc., 42:78-79 (with P. J. Munson and P. W. Par-
malee, senior authors).

1981

[Review of] Pleistocene mammals of North America, by B. Kur-
ten and E. Anderson. J. Vert. Paleontoh, 1:241.

Conjure bones. Carnegie Mag., 55(5):4-9.

February. Carnegie Mag., 55(2):43.

March. Carnegie Mag., 55(3):43.

April. Carnegie Mag., 55{4):41.

May. Carnegie Mag., 55(5):43.

June, July, August. Carnegie Mag., 55(6):33.

September. Carnegie Mag., 55(7):40.

October. Carnegie Mag., 55(8):41.

November. Carnegie Mag., 55(9):49.

December and January. Carnegie Mag., 55(1 0):32.

1982

Dental variation in Microtus xanthognathus, M. chrotorrhinus,
and M. pennsylvanicus (Rodentia, Mammalia). Ann. Car-
negie Mus.. 51(1 1 ):2 1 1-230.

Appalachia 11,000-12,000 years ago: a biological review. Ar-
chaeology of Eastern North Amer., 10:22-26.

Carnegie’s cornice — science. Carnegie Mag., 56( 1 ): 1 6—23.

Caves, bones and Ice Age owls. Carnegie Mag., 56(3):20-29.
[Introduction to] The scientist’s notebook: Kayapo, by D. A.
Posey. Carnegie Mag., 56(4): 18.

[Introduction to] The scientist’s notebook: Endangered species
and the museum, by C. J. McCoy. Carnegie Mag., 56(5):28.

“by Golley, J. W.” Verbatim, 9(2): 17-1 8.

An introduction to the Meadowcroft/Cross Creek archaeological
project: 1973-1982. Pp. 1-30, in Meadowcroft— collected
papers on the archaeology of Meadowcroft Rockshelter and
the Cross Creek drainage (R. C. Carlisle and J. M. Adovasio,
eds.) (Dept. Anthropol., Univ. Pittsburgh) 47th Ann. Mtg.,
Soc. Amer. Archaeol., Minneapolis, Minnesota, 14-17 April
1982, 270 pp. (with R. C. Carlisle, J. M. Adovasio, J. Don-
ahue, and P. Wiegman, senior authors).

Vertebrate faunal remains from Meadowcroft Rockshelter,
Washington County, Pennsylvania: summary and interpre-
tation. Pp. 163-174, in Meadowcroft— collected papers on
the archaeology of Meadowcroft Rockshelter and the Cross
Creek drainage (R. C. Carlisle and J. M. Adovasio, eds.)
(Dept. Anthropol., Univ. Pittsburgh) 47th Ann. Mtg., Soc.
Amer. Archaeol., Minneapolis, Minnesota, 14-17 April 1982,
270 pp. (with P. W. Parmalee).

1983

[Introduction to] The scientist’s notebook: The birds of Socorro,
by K. C. Parkes. Carnegie Mag., 56(7):31.

Terrestrial vertebrate faunas. Pp. 31 1-353, in The Late Pleis-
tocene (S. C. Porter, ed.), Vol. 1 of, Late-Quaternary envi-
ronments of the United States (H. E. Wright, Jr., ed.), Univ.
Minnesota Press, Minneapolis (with E. L. Lundelius, Jr., R.
W. Graham, E. Anderson, senior authors, J. A. Holman, D.
Steadman, and S. D. Webb, junior authors).

In Press

Pleistocene extinction and environmental change: case study of
the Appalachians. In Quaternary extinctions: A prehistoric
revolution (P. S. Martin and R. G. Klein, eds.). Univ. Ari-
zona Press.

The physiographic provinces of Pennsylvania. In Species of spe-
cial concern in Pennsylvania (H. H. Genoways and F. J.
Brenner, eds.).

Paleoenvironmental reconstruction at Meadowcroft Rockshelter,
Washington County, Pennsylvania. In Environment and
extinctions: man in late glacial North America (J. I. Mead
and D. J. Meltzer, eds.). Peopling of the Americas Series,
Center for the Study of Early Man, Univ. Maine, Orono
(with J. M. Adovasio, R. C. Carlisle, K. Cushman, J. Don-
ahue, senior authors, W. C. Johnson, K. Lord, P. W. Par-
malee, R. Stuckenrath and P. Weigman, junior authors).

Address: Section of Vertebrate Fossils, Carnegie Museum of Natural History, 4400 Forbes Ave., Pittsburgh,
Pennsylvania 15213.

MID-WISCONSINAN AND MID-HOLOCENE HERPETOFAUNAS OF
EASTERN NORTH AMERICA: A STUDY IN MINIMAL CONTRAST

Leslie P. Fay

ABSTRACT

Two herpetofaunas (Strait Canyon, Virginia — mid-Wisconsin-
an, and St. Elzear, Quebec — mid-Holocene) provide the first view
of non-glacial Quaternary environments in eastern North Amer-
ica. Six amphibians and one snake were recovered from St. El-
zear, nine amphibians and 16 reptiles from Strait Canyon. All
herpetological members of these local faunas presently occur at
or near the fossil localities, indicating conditions nearly identical
to the current climates for western Virginia 30,000 years ago and

about 5,000 years ago in eastern Quebec. No Altonian Stadial
(Strait Canyon) or Hypsithermal (St. Elzear) effects are evident
in the herpetofaunas, and no apparent similarity to the steepened
mammalian faunal gradient of eastern North America is revealed.
Non-glacial herpetofaunas may have been nearly identical to the
present distribution, with only minor range adjustments in re-
sponse to changing climate. Ambystoma laterale is reported from
the fossil record for the first time.

INTRODUCTION

Quaternary vertebrates of eastern North America,
outside of Florida, are known to us mostly through
the work of John E. Guilday and colleagues. Field
work by Carnegie Museum of Natural History crews
for over two decades has produced a number of
important local faunas from the Appalachian re-
gion; especially those of late Wisconsinan to early
Holocene age. In these faunas, Guilday delineated
a mammalian faunal gradient (Guilday et al., 1978)
that is steeper than the present interglacial mammal
distribution. With the kind permission and encour-
agement of John Guilday, I have begun to test for
any steepened gradient in the herptiles associated
with these mammalian local faunas. Two small her-
petofaunas, from Caverne de Saint Elzear de Bon-
aventure, Gaspe Peninsula, Quebec (mid-Holo-
cene), and Strait Canyon Fissure, Highland County,
Virginia (mid-Wisconsinan), were examined first
because the mammals had not yet been reported in
detail and because no interglacial or verified inter-
stadial herpetofaunas have previously been reported
from eastern North America.

Caverne de Saint Elzear is located near the village
of Saint Elzear de Bonaventure on the south coast
of the Gaspe Peninsula, Quebec, in upland boreal
forest. The fossils, originally discovered by local res-

idents, were excavated from a talus deposit within
the cave by Carnegie Museum of Natural History
and Quebec Natural Resources Ministry personnel
in 1977, 1978, and 1 979 (LaSalleand Guilday, 1980).
Age of the local fauna is considered to be mid-Ho-
locene (about 5,000 y.b.p.), based on stratigraphic
and faunal evidence. The mammals strongly resem-
ble the late Wisconsinan local faunas of the central
Appalachians, providing contrasts useful for region-
al paleoclimatic analysis (J. E. Guilday, in litt., 29
March 1982).

Strait Canyon Fissure local fauna, Highland
County, Virginia (38°28' 1 4"N, 79°30'42"W) was ex-
cavated from a fissure fill matrix in 1968 by a field
party of the U.S. National Museum. This is the first
pre-Woodfordian Wisconsinan fauna from the Ap-
palachians. A carbon- 14 date of 29,870 + 1800/
— 1400 y.b.p. (GX-7017-A) places the fossils near
the Altonian Stadial-Farmdalian Interstadial
boundary of Illinois terminology (Frye and Will-
man, 1960) or within the Plum Point Interstadial
of Ontario nomenclature (Dreimanis and Karrow,
1972).

St. Elzear and Strait Canyon fossils are reposited
at the Carnegie Museum of Natural History.

ACKNOWLEDGMENTS

John Guilday launched me on this project and I regret not
being able to report these findings to him. Alice M. Guilday has
cheerfully provided unpublished information and helped me sort
out stratigraphic provenance. J. Alan Holman and Thomas C.

LaDuke have aided my passage into the maze that is fossil herp-
tile identification. This paper was improved by Holman’s thor-
ough review. I thank them all for their guidance and encourage-
ment.

14

1984

FAY-EASTERN HERPETOFAUNAS

15

FAUNAL LISTS

Strait Canyon, Virginia

Class Amphibia
Order Caudata

Family Ambystomatidae
Ambystoma maculatum (Shaw)

Family Plethodontidae
Desmognathus Ifuscus Rahnesque
Plethodon glutinosus (Green)

Order Salientia
Family Bufomdae
Bufo americanus Holbrook
Bufo woodhousei fowleri Flinckley
Family Ranidae

Rana catesbeiana Shaw
Rana clamitans Latreille
Rana cf. R. pipiens Schreber
Rana sylvatica LeConte
Class Reptilia
Order Chelonia
Family Emydidae
Pseudemys, sp. indet.

Terrapene Carolina (Linne)

Order Squamata
Family Scincidae
lEumeces
Family Colubridae

Coluber or Masticophis
Diadophis punctatus (Linne)

Elaphe guttata (Linne)

Elaphe obsoleta (Say)

Lampropeltis calligaster (Harlan)
Lampropeltis getulus (Linne)
Lampropeltis triangulum (Lacepede)
Opheodrys vernalis (Harlan)

Nerodia sipedon (Linne)

Storeria loccipitomaculata (Storer)
Thamnophis sirtalis (Linne)

Family Viperidae
Agkistrodon contortrix (Linne)

Crotalus horridus Linne

Saint Elzear, Quebec

Class Amphibia
Order Caudata

Family Ambystomatidae
Ambystoma laterale Hallowed
Ambystoma maculatum (Shaw)

Order Salientia
Family Bufonidae
Bufo americanus Holbrook
Family Hylidae
Hyla crucifer Wied
Family Ranidae
Rana pipiens Schreber
Rana sylvatica LeConte
Class Reptilia
Order Squamata
Family Colubridae

Thamnophis sirtalis (Linne)

SPECIES ACCOUNTS

Following is an annotated list of herptiles occur-
ring in the Saint Elzear and Strait Canyon local fau-
nas. Conant (1975) and Logier and Toner (1961)
were consulted for modern distribution informa-
tion. Listed taxa presently occupy the area of the
fossil localities unless otherwise stated in the re-
marks.

Class Amphibia
Order Caudata
Family Ambystomatidae

Ambystoma laterale Hallowell
Blue-spotted Salamander

Occurrence. — St. Elzear.

Remarks. — Ambystomatid salamanders may be
distinguished by length-width ratios derived from
measurements of trunk vertebrae (Tihen, 1958). All
St. Elzear ambystomatid salamanders fit the mac-
ulatum- group, which includes the two most north-
easterly distributed members of the family -A. lat-
erale and A. maculatum. According to the vertebral

ratios (centrum length/anterior centrum width), two
taxa are present. By means of comparison with mod-
ern skeletal material, the shorter vertebrae (<2.65)
are assigned to A. laterale , the longer to A. macu-
latum. This is the first report of Ambystoma laterale
from the fossil record.

Ambystoma maculatum (Shaw)

Spotted Salamander

Occurrence. — St. Elzear and Strait Canyon.

Remarks. — The method used to distinguish A.
maculatum is described above. All Strait Canyon
ambystomatid vertebrae fall within the range of this
species.

Family Plethodontidae

Desmognathus ?fuscus Rafinesque
Dusky Salamander

Occurrence. — Strait Canyon.

Remarks. — One salamander vertebra has the
opisthocoelous condition of Desmognathus (Soler,

16

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

1950) and very closely resembles D.fuscus in form
and size. A degree of uncertainty remains because
D. ochrophaeus vertebrae are quite similar to those
of D. fuscus, making identification tenuous based
on one fossil specimen.

Plethodon glutinosus (Green)

Slimy Salamander

Occurrence. — Strait Canyon.

Remarks. — One plethodontid vertebra is indis-
tinguishable from comparative material of P. glu-
tinosus.

Order Salientia
Family Bufonidae

Bufo americanus Holbrook
American Toad

Occurrence. — St. Elzear and Strait Canyon.

Remarks. — Ilia of B. americanus may be distin-
guished from B. woodhousei by the wider base and
more anterior deflection of the dorsal protuberance
in the former (Holman, 1967, 1977; Wilson, 1975).

Bufo woodhousei fowleri Hinckley
Fowler’s Toad

Occurrence. — Strait Canyon.

Remarks.— See B. americanus remarks for spe-
cific determination. In addition to present distri-
bution, subspecific assignment is possible because
the dorsal protuberance is relatively lower in ilia
of the eastern form, B. w. fowleri, than in the western
subspecies, B. w. woodhousei (Eshelman, 1975; Hol-
man, 1977).

Family Hylidae

Hyla crucifer Wied
Spring Peeper

Occurrence. — St. Elzear.

Remarks. — Ilia of H. crucifer may be distin-
guished from other hylid frogs by a combination of
ilial characteristics (Holman, 1967) and by small
size.

Family Ranidae

Rana catesbeiana Shaw
Bullfrog

Occurrence. — Strait Canyon.

Remarks.— R. catesbeiana and R. clamitans ilia
have a steeper posterodorsal slope of the shaft as it
merges with the dorsal acetabular expansion than
do ilia in a group comprised of R. palustris, R. pip-

iens, and R. sylvatica (Holman, 1967). R. catesbei-
ana is separated from R. clamitans by a more rugged
(Holman, 1967) and often slightly concave vastus
prominence.

Rana clamitans Latreille
Green Frog

Occurrence. — Strait Canyon.

Remarks. — Distinguishing features for R. clami-
tans are discussed with R. catesbeiana.

Rana pipiens Schreber

Northern Leopard Frog

Occurrence. — St. Elzear and Strait Canyon.

Remarks. — According to Holman ( 1 967, pp. 1 58—
159): “ Rana pipiens and Rana palustris . . . may be
separated from R. sylvatica in that the prominence
for the origin of the vastus externus head of the
triceps femoris muscle (terminology of Holman,
1965) is larger, less produced, and less roughened
than in R. sylvatica. Rana pipiens has a somewhat
steeper slope of the posterodorsal border of the ilial
shaft into the dorsal acetabular expansion than in
. . . R. palustris .” In the Strait Canyon faunal list,
tentative identification as R. pipiens is made be-
cause no method of separating the ilia of the leopard
frog complex (R. berlandieri, R. blairi, R. pipiens,
and R. utricularia) has been devised (Holman, 1 977).
Therefore, I will only make firm specific assign-
ments within the pipiens- group when modern dis-
tribution warrants. Leopard frogs do not presently
occupy the Strait Canyon area, but R. pipiens (as a
disjunct record) and R. utricularia occur in Mary-
land and eastern Virginia.

Rana sylvatica LeConte
Wood Frog

Occurrence. — St. Elzear and Strait Canyon.

Remarks. — Identification of R. sylvatica by Hol-
man’s (1967) criteria is discussed with R. pipiens.

Class Reptilia
Order Chelonia
Family Emydidae

Pseudemys, sp. indet.

Occurrence. — Strait Canyon.

Remarks. — Several carapace fragments belong to
the genus Pseudemys. The material is in somewhat
poor condition, rendering identification difficult.
Also, no specimens of P. rubiventris, the species
which most closely approaches Strait Canyon today.

1984

FAY- EASTERN HERPETOFAUNAS

17

were available for comparison. No Pseudemys cur-
rently occupy the fossil locality, although several
forms inhabit the Piedmont and coastal plain.

Terrapene Carolina (Linne)

Eastern Box Turtle

Occurrence. — Strait Canyon.

Remarks.— A left epiplastral is identical to com-
parative material of T. Carolina.

Order Squamata
Family Scincidae

lEumeces

Skink

Occurrence. — Strait Canyon.

Remarks.— A small jaw fragment comparable to
the genus Eumeces was recovered. No further iden-
tification was attempted.

Family Colubridae
Coluber or Masticophis

Occurrence. — Strait Canyon.

Remarks.— Snakes of the subfamily Colubrinae
are best distinguished by a combination of charac-
ters, including morphology, size, and ratios of linear
measurements of vertebrae (Auffenberg, 1963; Hol-
man, 1981; Meylan, 1982). Coluber and Mastico-
phis have very long, relatively large vertebrae. The
two genera have roughly similar habits and mor-
phology, making them difficult to distinguish. C.
constrictor lives in the area today, and thus might
be considered the more likely choice.

Diadophis punctatus (Linne)

Ringneck Snake

Occurrence. — Strait Canyon.

Remarks. — This diminutive snake is distinct from
other colubrids, except Carphophis and Rhadinaea ,
with vertebral lengths less than 2 mm. The haemal
keel of Rhadinaea is distinctly different from the
other two genera, and Diadophis can be distin-
guished from Carphophis by a slightly higher neural
spine and more vaulted neural arch on the former
(Holman, 1977).

Elaphe guttata (Linne)

Corn Snake

Occurrence. — Strait Canyon.

Remarks.— Elaphe and Lampropeltis specimens
are best sorted by measuring ratios of linear dimen-
sions of vertebrae and cross-checking the resulting

groups with comparative material to verify identity
of morphological features. Results of this analysis
for all Quaternary Appalachian herpetofaunas will
be reported elsewhere (Fay, in preparation). Nu-
merical and gross morphological approaches are both
used because each method is imperfect when applied
to such variable objects as snake vertebrae. Varia-
tion within individuals as well as within taxa often
exceeds the norms established with comparative
material.

Elaphe obsoleta (Say)

Rat Snake

Occurrence. — Strait Canyon.

Remarks.— See E. guttata remarks.

Lampropeltis calligaster (Harlan)
Kingsnake

Occurrence.— Strait Canyon.

Remarks. — See E. guttata remarks.

Lampropeltis getulus (Linne)

Eastern Kingsnake

Occurrence. — Strait Canyon.

Remarks.— See E. guttata remarks.

Lampropeltis triangulum (Lacepede)
Milksnake

Occurrence. — Strait Canyon.

Remarks.— See E. guttata remarks.

Opheodrys vernal is (Harlan)

Smooth Green Snake

Occurrence. — Strait Canyon.

Remarks. — Opheodrys vertebrae have higher
neural spines than do Diadophis and Carphophis
(Holman, 1 977) and tend to be slightly larger as well.
The fossils appear identical to O. vernalis.

Nerodia sipedon (Linne)

Water Snake

Occurrence. — Strait Canyon.

Remarks. — Snakes of the subfamily Natricinae
possess laterally flattened hypopophyses on all ver-
tebrae (Holman, 1981). Nerodia vertebrae are rel-
atively shorter and wider than Thamnophis verte-
brae (Holman, 1981). N. sipedon may be
distinguished from its congeners by relatively low
neural spines (Holman, 1972).

18

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Storeria loccipitomaculata (Storer)
Red-bellied Snake

Occurrence. — Strait Canyon.

Remarks. — Holman (1962) discusses criteria for
separating three genera of small natricines: Storeria,
Tropidoc/onion, and Virginia. I cannot satisfactorily
discern S. dekayi from S. occipitomacu/ata, but the
fossils more nearly resemble the latter. Fossil Sto-
reria are not often identified to species level in the
literature; most are listed as Storeria cf. S. dekayi
(Holman, 1981). Both species currently inhabit the
Strait Canyon area.

Thamnophis sirtalis (Linne)

Garter Snake

Occurrence. — St. Elzear and Strait Canyon.

Remarks. — Holman (1962) describes distinguish-
ing features for Thamnophis vertebrae.

Family Viperidae

Agkistrodon contortrix (Linne)
Copperhead

Occurrence. — Strait Canyon.

Remarks. --North American vipers (subfamily
Crotalinae) have distinctively elongate hypopo-
physes which are rounded in cross-section. Agkis-
trodon vertebrae are distinguished from other North
American crotalines on the basis of large, single co-
tylar foraminae, often recessed in deep pits. Crotalus
and Sistrurus have smaller, often multiple cotylar
foraminae (Holman, 1982). A. contortrix usually
have slightly smaller cotylar foraminae, less vaulted
neural arches, and lower neural spines than do A.
piscivorus (Holman, 1982).

Crotalus horridus Linne
Timber Rattlesnake

Occurrence. — Strait Canyon.

Remarks. — C. horridus has the lowest neural
spines of any eastern North American crotalid and
usually the least vaulted neural arches as well (Hol-
man, 1982).

DISCUSSION

Strait Canyon

Although all 25 taxa are not found at Strait Can-
yon today, this herpetofauna does coexist 150 km
to the east on the Piedmont, with all but four of the
mammals from the local fauna as well. Strait Can-
yon, like St. Elzear, has a slightly more boreal mam-
mal and bird complement with herptiles which cur-
rently occupy the region. More northerly distributed
herptiles do not appear in the Strait Canyon local
fauna to mirror the steepened mammalian faunal
gradient.

St. Elzear

The seven species comprise a harmonious fauna
which can be found at St. Elzear today. Bufo ameri-
canus is by far the most common herptile (87% of
255 MNI), perhaps signifying a hibernaculum mor-
tality. Ambystoma laterale is reported from the fos-
sil record for the first time, possibly due to the fact
that St. Elzear is the first fossil locality within A.
laterale range. The herptiles have a less boreal char-
acter than the St. Elzear mammals (total 34 species),
which include four taxa that do not range as far

southeast as the Gaspe Peninsula today. This steep-
ened gradient cannot be reflected in the herptiles,
as there are no “more boreal” taxa than those which
currently occupy the fossil locality.

Interglacial/Interstadial Herpetofaunas

These first two herpetofaunas reported from non-
glacial intervals in eastern North America suggest
near-equivalence with present distributions. No in-
fluence of terminal Altonian events (presumably
cooler than present) are illustrated by the Virginia
fauna and Hypsithermal effects are not noted in the
Quebec fauna. At least some herptiles were sentitive
to climatic change in Irvingtonian time, as indicated
by northem/prairie species (especially Elaphe vul-
pina) in Appalachian local faunas (Holman, 1977,
1982; Fay, unpublished manuscript). Verification of
the possibility that non-glacial herpetofaunas have
remained the same throughout the RancholaBrean
(that is, resuming the “modern” pattern after each
glacial pulse) must wait for the discovery of more
interglacial and interstadial local faunas.

1984

FAY-EASTERN HERPETOFAUNAS

19

LITERATURE CITED

Auffenberg, W. 1963. The fossil snakes of Florida. Tulane
Studies Zool., 10:131-216.

Conant, R. 1975. A field guide to reptiles and amphibians of
eastern North America. Houghton Mifflin Co., Boston, 429

pp.

Dreimanis, A., and P. F. Karrow. 1972. Glacial history of the
Great Lakes-St. Lawrence region, the classification of the
Wisconsin Stage, and its correlation. Inti. Geol. Congr., Sec.
12: Quat. Geol., pp. 5-15.

Eshelman, R. E. 1975. Geology and paleontology of the early
Pleistocene (late Blancan) White Rock fauna from north-
central Kansas. Mus. Paleont. Univ. Michigan, Papers Pa-
leont., 13:1-60.

Frye, J. C., and H. B. Willman. 1960. Classification of the
Wisconsinan Stage in the Lake Michigan glacial lobe. Illinois
State Geol. Surv. Circ. 285:1-16.

Guilday, J. E., H. W. Hamilton, E. Anderson, and P. W.
Parmalee. 1978. The Baker BluffCave deposit, Tennessee,
and the late Pleistocene faunal gradient. Bull. Carnegie Mus.
Nat. Hist., 1 1:1-67.

Holman, J. A. 1962. A Texas Pleistocene herpetofauna. Co-
peia, 1 962(2):25 5— 26 1 .

. 1965. Early Miocene anurans from Florida. Quart. J.

Florida Acad. Sci., 28:68-82.

. 1967. A Pleistocene herpetofauna from Ladds, Georgia.

Bull. Georgia Acad. Sci., 25:154-166.

. 1972. Amphibians and reptiles. In Early Pleistocene

preglacial and glacial rocks and faunas of northcentral Ne-

braska (M. F. Skinner, et ah). Bull. American Mus. Nat.
Hist., 148:55-71.

. 1977. The Pleistocene (Kansan) herpetofauna of Cum-
berland Cave, Maryland. Ann. Carnegie Mus., 46:157-172.

. 1981. A review of North American Pleistocene snakes.

Publ. Mus., Michigan State Univ., Paleont. Ser., 1:263-306.

. 1982. The Pleistocene (Kansan) herpetofauna of Trout

Cave, Maryland. Ann. Carnegie Mus., 51:391-404.

LaSalle, P., and J. E. Guilday. 1980. Caverne de Saint Elzear
de Bonaventure: Rapport preliminaire sur les fouilles de
1977 et 1978. Ministere de l’Energie et des Ressources, Gouv.
du Quebec, 29 pp.

Logier, E. B. S., and G. C. Toner. 1961. Check list of the
amphibians and reptiles of Canada and Alaska. Life Sci. Div.
Contr., Royal Ontario Mus., 53:1-92.

Meylan, P. A. 1982. The squamate reptiles of the Inglis IA
fauna (Irvingtonian: Citrus County, Florida). Bull. Florida
State Mus., Biol. Sci. 27:1 1 1-195.

Soler, E. I. 1950. On the status of the family Desmognathidae
(Amphibia, Caudata). Univ. Kansas Sci. Bull., 33:459-480.

Tihen, J. A. 1958. Comments on osteology and phylogeny of
ambystomatid salamanders. Bull. Florida State Mus., 3:1-
50.

Wilson, V. V. 1975. The systematics and paleoecology of two
late Pleistocene herpetofaunas from the southeastern United
States. Unpublished Ph.D. dissert., Michigan State Univ.,
East Lansing, 67 pp.

Address: The Museum, Michigan State University, East Lansing, Michigan 48824-1045.

HERPETOFAUNAS OF THE DUCK CREEK AND WILLIAMS LOCAL
FAUNAS (PLEISTOCENE: ILLINOIAN) OF KANSAS

J. Alan Holman

ABSTRACT

The Duck Creek and Williams local faunas of west-central and
central Kansas are considered to represent the Illinoian stage of
the Pleistocene based on previous faunal and stratigraphic stud-
ies. The Duck Creek fauna yielded one salamander, three an-
urans, and one snake. The Williams fauna yielded one salaman-
der, four anurans, four turtles, and 13 snakes, and is the largest
Illinoian herpetofauna known. The major habitat indicated by
both herpetofaunas is a pond or slow-moving stream. The pres-
ence of prairie flats and hillsides as well as woodland is also

indicated. None of the amphibians and reptiles in either fauna
are extinct. The presence of three extralimital forms (Rana syl-
vatica, Emydoidea blandingii, and Elaphe vulpina) that occur
mainly to the northeast of the areas today may indicate somewhat
cooler, moister conditions. But a full glacial climate or a domi-
nant coniferous forest community is not indicated by the rest of
the herpetological species, all of which can be found in the area
today. The topics of “disharmonious” Pleistocene faunas and
“mosaic” Pleistocene communities are addressed.

INTRODUCTION

Herpetofaunas from the Illinoian stage of the
Pleistocene are quite rare. In fact, only one large
herpetofauna is now known from an unquestionably
Illinoian deposit (Sandahl local fauna, McPherson
County, Kansas; Holman, 1971). Thus, the recent
availability for study of two Illinoian herpetofaunas
from west-central and central Kansas offers the op-

portunity to add considerably to the knowledge of
the amphibians and reptiles of this stage of the Pleis-
tocene. These faunas are especially important in the
light of the statement (Holman, 1980: 133) that evi-
dence from the study of the Sandahl herpetofauna
did not support the classical idea of a cooler, moister
climate for the Illinoian age in Kansas.

ACKNOWLEDGMENTS

I should here like to acknowledge Dr. Richard J. Zakrzewski
of the Sternberg Memorial Museum of Fort Hays State College
for the privilege of studying the Duck Creek fauna fossils and

Dr. Gerald Smith of the Museum of Paleontology of the Uni-
versity of Michigan for the privilege of studying the Williams
fauna material. Rosemarie Attilio made the drawings.

THE DUCK CREEK LOCAL FAUNA

The Duck Creek local fauna near the Smoky Hill
River, Ellis County, Kansas (northeast quarter of
section 33, T 15 S, R 16 W) was previously studied
by Kolb et al. (1975), mollusks; McMullen (1975
and 1978) and Zakrzewski and Maxfield (1971),
mammals. Based on its stratigraphic location in the
Pfeifer Terrace, and on a comparison with other
Pleistocene plains faunas, an Illinoian age was sug-
gested (McMullen, 1978). Five collecting localities
(designated by Arabic numerals 1--5) were defined.
Different lithologies in these localities suggest dif-
ferent depositional mechanisms. Localities 1 and 2
were in coarse sand with sandy silt lenses that sug-
gested a stream channel. Locality 3 is a clayey de-
posit that suggested quiet water such as an oxbow
lake. Localities 4 and 5 are silt deposits with sandy
silt lenses, suggesting a backwater area where current

action was slow and where deposition of fine par-
ticles might have taken place.

Class Amphibia Linnaeus, 1758
Order Caudata Oppel, 1811
Family Ambystomatidae Hallowed, 1858

Ambystoma tigrinum (Green, 1825)

Material. — Locality 1, vertebra FHSVP 2824; Locality 3, right
femur FHSVP 2903; Locality 4, vertebra FHSVP 2905.

Remarks. — Tihen (1958) and Holman ( 1 969) have
discussed the identification of vertebrae of Ambys-
toma tigrinum. The large right femur (FHSVP 2903)
is indistinguishable from those in modern skeletons
of A. tigrinum. Today in Kansas, adult A. tigrinum
spends much of its time beneath the ground in bur-
rows of other animals (Collins, 1974:26). From a

20

1984

HOLMAN -HERPETOFAUNAS FROM KANSAS

21

c

Fig. 1. — A) Rana sylvatica, left ilium FHSVP 2859 from Duck Creek local fauna; B) Acris crepitans, right ilium UM 81732 from
Williams local fauna; C) Bufo woodhousei woodhousei, right cranial crest UM 81733 from Williams local fauna; D) Bufo woodhousei
woodhousei, right ilium UM 81733 from Williams local fauna. Each line equals 2 mm.

taphonomic standpoint it is interesting to note that
A. tigrinum bones were found in both high energy
(Locality 1) and low energy (Localities 3 and 4) sit-
uations. This species occurs in Ellis County, Kansas,
today (Collins, 1974: map p. 26).

Order Anura Dumeril, 1807
Family Bufonidae Fitzinger, 1826

Bufo woodhousei woodhousei Girard, 1854

Material. — Locality 1, a left and a right ilium FHSVP 2857-
2858; Locality 2, right ilium FHSVP 2897; Locality 5, two right
ilia FHSVP 2922-2933.

Remarks. — Holman (1971 ) and Tihen (1962) have
shown how elements of Bufo woodhousei wood-
housei may be identified at the subspecific level.
This subspecies has also been reported from the
Sandahl Illinoian fauna of McPherson County, Kan-

sas (Holman, 1971). According to Collins (1974:56)
B. w. woodhousei prefers lowlands and sandy areas
and is generally the only toad found on the flood
plains of the larger streams and rivers in Kansas
today. Taphonomically it is of interest that bones
of this species are also found in both high energy
(Localities 1 and 2) and low energy (Locality 5) sit-
uations. This subspecies occurs in Ellis County,
Kansas, today (Collins, 1974: map p. 55).

Family Ranidae Bonaparte, 1831
Rana sylvatica Le Conte, 1825

Material — Locality 1, left ilium FHSVP 2859 (Fig. la).

Remarks. — The ilium of Rana sylvatica is quite
distinct from those of many other Rana species in
having a very distinct dorsal protuberance on the

22

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

anterodorsal part of its lateral surface near the ace-
tabular fossa. In fact, this prominence is rounded
and somewhat similar to those of tree toads of the
genus Hyla. It is noteworthy that several authors
have commented on the toad-like gait and short-
leggedness of Rana sylvatica, especially in northern
populations. Moreover, the ilia of Rana sylvatica,
although of small size, do not show the perforate
condition of the acetabular area as do immature
individuals of larger species of Rana. Rana sylvatica
is not found in Kansas today, but mainly occurs to
the northeast (Fig. 1 1 ). The nearest population today
is an isolated one in the Ozark Region of south-west
Missouri and north-west Arkansas. Today, the wood
frog occurs in wooded areas (Wright and Wright,
1 949). The significance of this record will be detailed
in the Discussion section.

Rana pipiens complex Schreber, 1782

Material. — Locality 1, left and right ilia FHSVP 2865 and
2860, 2862, sacral vertebra FHSVP 2847; Locality 2, right ilia
FHSVP 2898, 2861; Locality 4, left ilia FHSVP 2910, 291 1;
Locality 5, left ilium FHSVP 2924, sacral vertebra FHSVP 2919.

Remarks. — Holman (1971) discusses characters
that distinguish the ilia of the Rana pipiens complex
from other species of Rana. Collins (1974:78) states
that R. pipiens is found in every aquatic situation
in the state today. Taphonomically it is noted that

the fossil R. pipiens was found in both high energy
(Localities 1 and 2) and low energy (Localities 4 and
5) sedimentary situations. Collins (1974: map p. 77)
shows that R. pipiens occurs in Ellis County, Kan-
sas, today, although he realizes that some workers
refer these populations to a distinct species, Rana
blairi.

Class Reptilia Laurenti, 1768
Order Squamata Oppel, 1811
Family Colubridae Oppel, 1811
Subfamily Colubrinae Cope, 1893

Lampropeltis calligaster (Harlan, 1827)
Material. — Locality 1, trunk vertebra FHSVP 2822.

Remarks. — This species has a very well-defined
hemal keel and subcentral ridge, but these processes
are not as well-defined as in Lampropeltis getulus.
Lampropeltis calligaster may be distinguished from
L. triangulum in that it has a more vaulted neural
arch and a more distinct hemal keel. Today, this
species is found slightly to the east and south of Ellis
County in Barton County, Kansas (Collins, 1974:
map p. 183). This snake inhabits a variety of areas
in Kansas today from rocky hillsides with open
woods to prairie grasslands and often retreats into
the burrows of other animals (Collins, 1974: 185).

THE WILLIAMS LOCAL LAUNA

The Williams local fauna of the Great Bend Prai-
rie Area, Rice County, Kansas (north-east corner of
south-east 'A section 21, T 18 S, R 7 W) was pre-
viously studied by Hall ( 1 972), mollusks; McMullen
(1975), shrews; Lundberg (1975), catfishes; Preston
(1979), turtles. It is considered to represent the II-
linoian age of the Pleistocene based on its fauna and
stratigraphy. The fauna was recovered from a bor-
row pit by field workers and friends of the late C.
W. Hibbard. This is one of the largest fossil her-
petofaunas known from the central plains, and is
the largest Illinoian herpetofauna known. The mode
of accumulation of the fossils in the Williams fauna
does not appear to be unlike that of the Duck Creek
fauna.

Class Amphibia Linnaeus, 1758
Order Caudata Oppel, 1811
Family Ambystomatidae Hallowed, 1858

Ambystoma tigrinum (Green, 1825)

Materia!. — Partial right maxilla and distal left humerus
UM81731.

Remarks. — This species, also recorded and dis-
cussed in the previous fauna, has not been recorded
from Rice County, Kansas, today, but according to
Collins (1974: map p. 26) should occur there.

Family Hylidae Hallowed, 1857
Acris crepitans Baird, 1854

Material. — One left and two right ilia UM 81732 (Fig. lb).

Remarks.— These ilia have the dorsal protuber-
ances anterior to the anterior edge of the acetabular
cup, and two of them are complete enough to have
an ilial shaft ridge. These characters in combination
are diagnostic for the genus Acris. The above fossils
are indistinguishable from those of modern Acris

1984

HOLMAN -HERPETOFAUNAS FROM KANSAS

23

Fig. 2. — A) Rana catesbeiana, right ilium UM 81734; B) Rana pipiens, left ilium in lateral view UM 81735; C) Rana pipiens. left ilium
in medial view UM 81735. All from Williams local fauna. Each line equals 2 mm.

crepitans, a species that occurs in Rice County, Kan-
sas, today (Collins, 1974: map p. 57). The preferred
habitat is said to be the muddy beach-like edges of
small, shallow streams and ponds in Kansas today
(Collins, 1974:58).

Family Bufonidae Fitzinger, 1826

Bufo woodhousei woodhousei Girard, 1854

Material. — Right cranial crest and one left and three right ilia
UM 81733 (Fig. lc, d).

Remarks.-— The comments on the identification
and habitat of this subspecies that were made in the
Duck Creek fauna section apply here. Bufo w. wood-
housei has been recorded from Rice County, Kansas,
today (Collins, 1974:55).

Family Ranidae Bonaparte, 1831
Rana catesbeiana Shaw, 1802
Material. — One left and three right ilia UM 81734 (Fig. 2a).

Remarks. — The ilia of Rana catesbeiana are dis-
tinguishable from those of other species of Rana in
the middle United States on the basis of (1) the
steeper slope of the posterior part of the ilial blade
into the dorsal acetabular expansion, (2) its large
size, and (3) the very porous area of the bone near
the acetabular border. Condition 3 is particularly
marked in small individuals. The bullfrog is re-

stricted to permanent lakes, rivers, streams, and
swamps today where deep water is available (Col-
lins, 1974:71). This species occurs in Rice County,
Kansas, today (Collins, 1974:71).

Rana pipiens complex Schreber, 1782

Material. — Sixteen left and 15 nght ilia UM 81735 (Fig. 2b,
c).

Remarks. — The comments that were made on the
identification and habitat of this form in the Duck
Creek section apply here. Collins (1974: map p. 77)
has recorded R. pipiens from counties surrounding
Rice County, Kansas, but there are no actual records
from Rice County. Certainly this form occurs in the
area today.

Class Reptilia Laurenti, 1768
Order Testudines Batsch, 1788

Family Chelydridae Swainson, 1839

Chelydra serpentina (Linnaeus, 1758)

Material. — Right femur, head of left femur, left ilium, left pubic
fragment, nuchal fragment UM 60084.

Remarks. — This material was previously reported
by Preston (1979:28). The snapping turtle is not
listed as occurring in Rice County, Kansas, today,
but according to the distribution shown in Collins
(1974: map p. 87) should occur in the area. This

24

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

species is said to occur in every aquatic situation,
but to prefer water with a soft mud bottom, abun-
dant pond-edge vegetation, and numerous sunken
logs and branches (Collins, 1974:87).

Family Testudinidae Gray, 1825
Emydoidea blandingii (Holbrook, 1838)

Material.— Right 8th and left 1 1th peripherals and additional
fragments UM 60089.

Remarks. — This species was previously identified
and reported from the Williams local fauna by Pres-
ton (1979:31). Emydoidea blandingii does not occur
in Kansas today, the closest records being in south-
central Nebraska (Fig. 12). The significance of this
occurrence in the Kansas fossil record will be de-
tailed in the Discussion Section of this paper.

Chry’semys picta Schneider, 1783

Material. — Nuchals and epiplastra UM 60085, carapace parts
UM 60086, plastral parts UM 60087, and 7th cervical vertebra
UM 60577.

Remarks. —This material was previously identi-
fied and reported by Preston (1979:37). Although it
is not listed as occurring in Rice County, Kansas,
today, the subspecies Chrysemys picta belli un-
doubtedly is present there (Collins, 1974: map p.
107). The painted turtle lives in slow-moving, shal-
low streams and rivers, and shallow ponds and lakes
with soft bottoms in Kansas today (Collins, 1974:
107-108).

Pseudemys scripta Schoepff, 1792

Material. — Peripheral fragments and an epiplastron UM 60088.

Remarks.— This material was previously identi-
fied and reported by Preston (1979:35). This species
is not recorded from Rice County, Kansas, today,
but according to Collins (1974: map p. 110) the
subspecies Pseudemys scripta e/egans should occur
in the area. This turtle is said to occur in about every
permanent body of water throughout its range in
Kansas today (Collins, 1974:1 10).

Order Squamata Oppel, 1811

Family Colubridae Oppel, 1811
Subfamily Xenodontidae Cope, 1893

Heterodon cf. Heterodon platyrhinos Latreille, 1 802

Material. — Four anterior trunk vertebrae and one fragmental
middle trunk vertebra UM 81736.

Remarks. — Most discussion about the identifi-
cation of Heterodon on the basis of vertebral re-

mains has been on the basis of middle trunk ver-
tebrae. Unfortunately, the relatively complete
Heterodon vertebrae from the Williams fauna are
anterior trunk vertebrae and do not appear to be
specifically diagnostic. Nevertheless, these verte-
brae are quite similar in size to those of a modern
Heterodon platyrhinos with a total length of 70 cm
(27.55 inches), thus I am tentatively referring the
vertebrae to this species. Conant (1975) lists the
typical ranges in total length as 16-25 inches for
Heterodon nasicus and 20-33 inches in H. platy-
rhinos. Heterodon platyrhinos has not been recorded
from Rice County, Kansas, today, but according to
Collins (1974: map p. 163) it should occur there.
This snake is said to be common along valleys of
major rivers in the Great Bend Prairie Area of Kan-
sas today and it is said to prefer sandy areas (Collins,
1974:163).

Subfamily Colubrinae Cope, 1895

Compared to vertebrae of the Subfamily Natri-
cinae in the Williams fauna, vertebrae of the
Subfamily Colubrinae were rare (Table 1 ). Two frag-
mentary vertebrae UM 81737 are assigned to Col-
ubrinae gen. et sp. indet.

Coluber constrictor Linnaeus, 1758

Material.— Trunk vertebra UM 60256 (Fig. 3).

Remarks. — Several authors (see summary in Hol-
man, 1981) have been unable to distinguish the ver-
tebrae of Coluber and Masticophis. Nevertheless,
based on the smaller size of the specimen and the
relatively small size of the neural canal, and based
on the modern geographic ranges of the two genera
(Collins, 1974: maps pp. 169 and 173) I am cau-
tiously referring this specimen to Coluber constric-
tor. The subspecies Coluber constrictor flaviventris
occurs in Rice County, Kansas, today (Collins, 1974:
map p. 169). This snake occurs in open grassland
and prairies in Kansas today (Collins, 1974: map p.
170).

Elaphe vulpina (Baird and Girard, 1853)

Material — An associated partial skeleton (Fig. 4) consisting of
a cervical vertebra, 55 trunk vertebrae, 5 caudal vertebrae, and
5 ribs UM 6023 1 ; also 3 trunk vertebrae from another individual
or individuals UM 81738.

Remarks. — The associated skeleton was collected
by the late C. W. Hibbard and his field party on 8
June 1969. Holman (1982) provided an osteological
definition of Elaphe vulpina. Diagnostic characters

1984

HOLMAN-HERPETOFAUNAS FROM KANSAS

25

Table 1 . — In ferred habitat and abundance of herpetological species of the Duck Creek and Williams Local Faunas of the Pleistocene
(Illinoian) of Kansas. Habitats based on Collins (1974). Abbreviations: F = frequent, S = seasonal, I = infrequent.

Species

Habitat

Fossil abundance

Prairie pond
or slow stream

Pond or
stream edge

Prairie flats
or hillsides

Woodland-

prairie

edge

Woodland

Minimum
number of
individuals

Total number of
salamander or
snake vertebrae

Duck Creek Fauna

Amby stoma tigrinum

s

s

F

F

F

i

2

Bufo w. woodhousei

s

F

F

3

Rana sylvatica

I

1

Rana pipiens complex

s

F

I

4

Lampropeltis calligaster

I

F

F

1

1

Williams Fauna

Ambystoma tigrinum

s

S

F

F

F

1

Cranial only

Acris crepitans

s

F

2

Bufo w. woodhousei

s

F

F

3

Rana catesbeiana

F

F

3

Rana pipiens complex

s

F

I

16

Chelydra serpentina

F

S

1

Emydoidea blandingii

F

S

1

Chrysemys picta

F

s

2

Pseudemys scripta

F

s

i

Heterodon platyrhinos

I

F

F

i

4

Colubrinae indet.

i

2

Coluber constrictor

F

F

1

i

1

Elaphe vulpina

I

F

2

63

Lampropeltis triangulum

F

F

F

i

1

Pituophis melanoleucus

F

F

i

!

Sonora episcopa

F

i

1

Tantilla sp.

F

i

3

Natricinae indet.

i

131

Nerodia sipedon

F

F

i

21

Regina grahami

F

F

i

1

Thamnophis radix

F

F

i

21

Thamnophis sp.

F

i

26

Tropidoclonion lineatum

F

F

i

3

of the trunk vertebra are as follow: neural spine
usually slightly longer than high, neural arch vault-
ed; condyle not enlarged; ventral processes of cen-
trum gracile. Holman (1982:40, Table 2) indicates
the height of neural spines of the trunk vertebrae of
E/aphe vulpina compared with related colubrid
species.

Elaphe vulpina is well out of its present day range
in the Williams fauna. Today, the species gets no
closer to Rice County, Kansas, than the northwest-
ern part of Missouri and southeastern part of Ne-
braska (Fig. 1 3). Collins ( 1 974:24 1 ) states “It is quite
possible that the western fox snake will be found in
extreme northeast Kansas.” The fox snake had a
wider distribution in the Pleistocene than it does
today (Holman, 1981: map p. 292). The significance
of this extralimital species will be detailed in the

Discussion section of this paper. The fox snake is a
grassland and woodland edge species today (per-
sonal observation).

Lampropeltis triangulum (Lacepede, 1788)

Material — Three trunk vertebrae UM 81739 (Fig. 5).

Remarks. — This is one of the easiest snakes of
eastern and central North America to identify on
the basis of vertebral characters. It has a low neural
spine, a depressed neural arch, and only a moder-
ately distinct hemal keel. The species occurs in Rice
County, Kansas, today (Collins, 1974: map p. 188).
This snake is said to occur in rocky ledges of prairie
canyons today (Collins, 1974:189).

Pituophis melanoleucus (Daudin, 1803)
Material. — Trunk vertebra UM 60255 (Fig. 6).

26

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 3. — Coluber constrictor, trunk vertebra UM 60256 from Williams local fauna. Upper left, dorsal, right, ventral; middle left, anterior,
right, posterior; lower, lateral. The line equals 2 mm and applies to all views.

Remarks. — AufFenberg (1963:183-184) and
Van De vender and Mead (1978:472) give vertebral
characters of Pituophis melanoleucus. Pituophis
melanoleucus savi, the subspecies found in Rice
County, Kansas, today (Collins, 1974: map p. 182)
has lower neural spines than the southeastern sub-
species P. m. mugitus. The Williams fauna vertebra
appears identical to P. m. sayi in neural spine height
(Fig. 6). This snake is said to live in open grassland
as well as open woodland and woodland edge today
in Kansas (Collins, 1974:182).

Sonora episcopa (Kennicott, 1859)

Material. — Trunk vertebra UM 81740.

Remarks. —The tiny vertebrae of Sonora may be
distinguished from those of Diadophis and Tantilla
on the basis of their more prominent neural spines,
thinner hemal keels, and shorter vertebral form. Van
Devender et al. (1977:55-56) state that S. semian-
nulata vertebrae cannot be distinguished from those
of S. episcopa. The nearest S. semiannulata occurs
to the area today is in southern New Mexico (Co-
nant, 1975: map 157). Conant (1975: map 158),
indicates S. episcopa occurs in Rice County, Kansas,
today, but the more detailed map of Collins (1974:
163) does not record it from Rice County, although
there are records to the north and to the south of
Rice County. The subspecies found in Kansas today

1984

HOLMAN -HERPETOFAUN AS FROM KANSAS

27

Fig. 4 .—Elaphe vulpina, trunk vertebra from an associated partial skeleton UM 60231 from Williams local fauna. Upper left, dorsal,
right, ventral; middle left, anterior, right, posterior; lower, lateral. The line equals 2 mm and applies to all views.

is S. e. episcopa. This snake is said to occupy dry,
rocky, prairie hillsides (Collins, 1974:194).

Tantilla sp. indet.

Material— Three trunk vertebrae UM 81741.

Remarks. —This tiny snake has trunk vertebrae
with obsolete neural spines. The species of Tantilla
appear to be difficult or impossible to distinguish
from one another on the basis of vertebral charac-
ters. Two species of Tantilla occur in Kansas today
(Collins, 1974: maps pp. 195 and 197). Tantilla
gracilis occurs only in the eastern one-third of Kan-
sas, whereas Tantilla nigriceps nigriceps occurs rather

extensively in the western three-fourths of the state,
although it has never been specifically recorded from
Rice County. Tantilla gracilis is said to occupy rocky
hillsides of open prairies and woodlands in Kansas
today, whereas T. nigriceps is said to be found on
rocky hillsides of grassland prairies (Collins, 1974:
195, 197).

Subfamily Natricinae Bonaparte, 1840

Many more natricine than colubrine vertebrae oc-
cur in the Williams fauna (Table 1) and this un-
doubtedly reflects the fact that these snakes lived
nearer the site of deposition than the more terrestrial
colubrines. One hundred and thirty-one vertebrae

28

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 5 . — Lampropeltis triangulum, trunk vertebra UM 81739 from Williams local fauna. Upper left, ventral, right dorsal; middle left,
anterior, right, posterior; lower, lateral. The line equals 2 mm and applies to all views.

with diagnostic processes missing and caudal ver-
tebrae are identified as Natricinae gen. et sp. indet.
UMV.

Nerodia sipedon (Linnaeus, 1758)

Material. — Twenty-one trunk vertebrae UM 81742 (Fig. 7).

Remarks. — Holman (1967) has given vertebral
characters that enable one to distinguish this species
from related forms in eastern and central North
America. Nerodia sipedon sipedon has not been re-
corded from Rice County, Kansas, today, but ac-
cording to Collins (1974; map p. 227) should occur
there. According to Collins ( 1 974:227) the northern
water snake is found in almost every aquatic situ-

ation in Kansas today, with both high and low en-
ergy bodies of water being frequented.

Regina grahami Baird and Girard, 1853
Material. — A trunk vertebra UM 81743 (Fig. 8).

Reward. — Holman (1972) gives vertebral char-
acters that separate R. grahami from other related
species. This species is found in Rice County, Kan-
sas, today (Collins, 1974: map p. 219). This snake
lives near ponds and sluggish streams of prairie
meadows and river valleys in Kansas today (Collins,
1974:219-220).

Storeria cf. Storeria dekayi (Holbrook, 1842)

Material. — Four trunk vertebrae UM 81744.

1984

HOLMAN-HERPETOFAUNAS FROM KANSAS

29

Fig. 6 .—Pituophis melanoleucus, trunk vertebra UM 60255 from Williams local fauna. Upper left, dorsal, right, ventral; middle left,
anterior, right posterior; lower, lateral. The line equals 2 mm and applies to all views.

Remarks. — Auffenberg (1963:192) and Holman
(1962:258-259) discuss the identification of Store-
ria dekayi on the basis of vertebrae. Storeria occip-
itomaculata is a woodland species that occurs in the
eastern two tiers of counties in Kansas today (Col-
lins, 1974: map p. 217). Storeria dekayi texana is
not recorded from Rice County, Kansas, today, but

has been reported from Ellsworth County just to the
north (Collins, 1974: map p. 215). This snake gen-
erally lives near moist situations in woodland and
along woodland edges (Collins, 1974:215).

Thamnophis radix (Baird and Girard, 1853)
Material.— Twenty-one trunk vertebrae UM 81745 (Fig. 9).

30

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 7 . — Nerodia sipedon, trunk vertebra UM 81742 from Williams local fauna. Upper left, dorsal, right, ventral; middle left, anterior,
right, posterior; lower, lateral. The line equals 2 mm and applies to all views.

Remarks. — This is only the second published rec-
ord of this species as a fossil as Rogers (1982) has
listed it from an early Kansan site in Kansas. Ver-
tebrae of species of Thamnophis are difficult to iden-
tify. Nevertheless, trunk vertebrae of modern skel-
etons of T. radix (11) and T. marcianus (5) are
definitely shorter and also less gracile than those of
modern T. sirta/is (10) and T. proximus (4). More-
over, the neural spines of T. radix and T. marcianus
have more anterior and posterior overhang than in
most T. sirtalis and T. proximus. The neural spines
of available modern T. radix are somewhat higher
than those of T. marcianus. The Williams local fau-
na vertebrae appear identical to those of T. radix.
The subspecies T. radix haydeni is found in Rice

County, Kansas, today (Collins, 1974: map p. 205).
These snakes are said to prefer open grassy prairies,
particularly along the edges of streams, marshes, and
lakes in Kansas today (Collins, 1974:206).

Thamnophis sirtalis (Linnaeus, 1758) or
Thamnophis proximus (Say , 1823)

Materia!.— Twenty-six vertebrae, UM 81746 (Fig. 10).

Remarks. — I cannot distinguish whether these
more elongate Thamnophis vertebrae represent
Thamnophis sirtalis or T. proximus. Thamnophis
sirtalis parietalis (Say) is listed as occurring in Rice
County, Kansas, today (Collins, 1974: map p. 207),
and although Thamnophis sauritus occurs in neigh-

1984

HOLMAN- HERPETOFAUNAS FROM KANSAS

31

Fig. 8. — Regina grahami, trunk vertebra UM 81743 from Williams local fauna. Upper left, dorsal, right, ventral; middle left, posterior,
right, anterior; lower, lateral. The line equals 2 mm and applies to all views.

boring Stafford County to the southwest (Collins,
1974: map p. 203) it has not been reported from
Rice County, Kansas. Thamnophis sirtalis is found
in a wide number of habitats such as marshes, wet
meadows, pond-margins, woodlands, and wood-
land edges and floodplains in Kansas today (Collins,
1974:208). It is said to prefer moist situations.
Thamnophis proximus frequents the edges of
swampy marshes, lakes, streams, and rivers (Col-
lins, 1974:204).

Tropidoclonion lineatum (Hallowed, 1856)

Material. — Three trunk vertebrae UM 81747.

Remarks. — This is only the third record of this
genus and species as a fossil. It has previously been
reported from the Pleistocene of Texas (Holman,
1965) and from the Pleistocene (Illinoian) of central
Kansas (Holman, 1971). Holman (1965) gives ver-
tebral characters for this form, which has a distinc-
tive vertebral type. Tropidoclonion lineatum is not
recorded from Rice County, Kansas, today, but ac-
cording to Collins (1974: map p. 209) should be
expected to occur there. This snake is said to inhabit
the hillsides of open prairies and woodland edges
in Kansas today (Collins, 1974:209).

32

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 9 . — Thamnophis radix, trunk vertebra UM 81745 from Williams local fauna. Upper left, dorsal, right, ventral; middle left, anterior,
right, posterior; lower, lateral. The line equals 2 mm and applies to all views.

DISCUSSION

Vertebrate faunas bear on several compelling
Pleistocene problems such as ( 1 ) the time of its on-
set, (2) proper divisional terms, (3) “disharmon-
ious” assemblages, (4) paleoclimates, (5) megafaun-
al extinction, and recently (6) why no C14 ages for
extinct Pleistocene genera are younger than 10,000
yrs B.P. (Meltzer and Mead, 1983). The Illinoian
herpetofaunas reported herein relate directly to
questions 3 and 4 and thus indirectly to questions
5 and 6.

One of the most discussed aspects of North Amer-

ican Pleistocene vertebrate faunas is that many are
considered to be “disharmonious” assemblages
(Lundelius et al., 1983). The term “disharmonious”
is used to describe faunas containing animals that
would be “ecologically incompatible” today (term
used by Holman, 1976). In such faunas one is unable
to find an area on the map where all of the species
identified from that fauna could be found living
together today (“area of sympatry” of McMullen,
1975, and others). Holman (1980) uses the term
“extralimital” to refer to Pleistocene herpetological

1984

HOLMAN-HERPETOFAUNAS FROM KANSAS

33

Fig. 10 . — Thamnophis sirtalis or Thamnophis proximus, trunk vertebra UM 31746 from Williams local fauna. Upper left dorsal, right,
ventral; middle left, anterior, right, posterior; lower, lateral. The line equals 2 mm and applies to all views.

species that occur outside of their present ranges.
The degree of “disharmoniety” in Pleistocene fau-
nas varies; some faunas having only one or two
extralimital forms, and other faunas having Boreal
and Tropical animals occurring together. Dishar-
monious faunas may be found in both glacial and
interglacial deposits (Holman, 1980).

On the other hand in the British Isles (Stuart,
1982), glacial and interglacial faunas appear to con-
tain more “harmonious” and “ecologically com-
patible” assemblages, and also to reflect more ex-
treme climatic changes in glacial and interglacial

times than in North America. For instance, inter-
glacial faunas may contain macaque monkeys, spot-
ted hyaenas, African lions, African rhinos, and Af-
rican hippos, whereas glacial faunas may contain
tundra voles, arctic foxes, woolly mammoths, wool-
ly rhinos, reindeer, and musk oxen.

Explanations for disharmonious faunas in the
North American Pleistocene vary. A common hy-
pothesis (Hibbard, 1960) repeated by numerous
other workers has been called “The Pleistocene Cli-
matic Equability Model” where cooler summers
supposedly account for the presence of northern ex-

34

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 11. — Modem range of Rana sylvatica (stippling) and record of Rana sylvatica from the Duck Creek local fauna. Pleistocene, Illinoian
(circled dot).

tralimital species in the fauna, and warmer winters
supposedly account for the presence of southern ex-
tralimital species in the fauna. Other workers (Lun-
delius et al., 1983) point out that disharmonious
faunas may be associated with mixed vegetational
associations. The classic concept that glaciers pushed
northern forms down into southern communities
may be partially true, but this concept does not ex-
plain the presence of southern extralimital forms
satisfactorily.

Herpetological species are important in recon-
structing fossil communities especially because (1)
they are ectothermic and quite sensitive to ecolog-
ical changes, (2) are probably less vagile than mam-
malian species, and (3) because most herpetological
species of the Pleistocene represent extant ones whose
ecological tolerances and habitat preferences are

known. It is noteworthy here that all species of am-
phibians and reptiles from all Illinoian localities
studied in Kansas are indistinguishable from living
species.

Duck Creek Local Fauna

Based on the study of the sedimentary environ-
ment and on the molluscan fauna, Kolbet al. (1975)
state: “The presence of Pisidium compression and
Sphaerium stratinum within cross-bedded sands in-
dicates a perennial stream with some current action,
while the abundance of Va/vata tricarinata suggests
the stream was lake-like in places. The abundance
of several strictly woodland species such as Cione/la
lubrica suggests that a continuous strand of trees
bordered the stream, while valley slopes were pos-
sibly covered with grasses and scattered trees. Like-

1984

HOLMAN-HERPETOFAUNAS FROM KANSAS

35

Fig. 12. — Modem range of Emydoidea blandingii (stippling, Xs, ?s) and record of Emydoidea blandingii from the Williams local fauna.
Pleistocene, Illinoian (circled dot).

wise, cooler summers and winters than at present
are indicated for Ellis County at the time these taxa
lived by the predominance of species with northern
distribution.”

Based on the mammalian fauna of the Duck Creek
locality McMullen (1978) states “In summary, the
Duck Creek local fauna lived near a stream which
supplied the area with permanent water. Occurring
in the moist lowlands and probably bordering the
stream in places was a riparian forest. Whether the
interspersed trees were coniferous or deciduous is
unknown because forest dwelling species in the fau-
na might have inhabited either situation and paly-
nological data are lacking. The adjacent lowlands
area probably was covered with mixed tall and short
grasses.”

The small herpetofauna of the Duck Creek site
appears mainly to fit in with the above interpreta-

tions (Table 1). The only extralimital species is the
wood frog, Rana sylvatica, which occurs mainly to
the north and east of Kansas today (Fig. 11). As is
the case with the woodland mammalian species dis-
cussed by McMullen (1978), Rana sylvatica occurs
in both deciduous and coniferous forests. Never-
theless, based on the R. sylvatica record, one might
suggest the possibility of at least cooler, perhaps
moister summers, and at least the presence of some
kind of woodland cover near the major aquatic hab-
itat. But it should be pointed out that other herpe-
tological elements of the Duck Creek fauna ( Am -
bystoma tigrinum, Bufo w. woodhousei, Rana
pipiens, and Lampropeltis caliigaster) occur in the
area today, and do not indicate a climate indicative
of northeastern South Dakota as inferred by
McMullen’s (1975) discussion of the area of sym-
patry for Duck Creek fauna shrews.

36

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 13. — Modem range of Elaphe vulpina (stippling) and record of Elaphe vulpina from the Williams local fauna. Pleistocene, Illinoian
(circled dot).

Williams Local Fauna

The Williams fauna has not been interpreted as
thoroughly as has the Duck Creek fauna. Hall (1972)
gave an abstracted paper at the Michigan Academy
of Science, Arts, and Letters on the mollusks of the
Williams local fauna. McMullen (1975) discussed a
new species of shrew that also occurred in the Duck
Creek fauna. Lundberg (1975) listed catfishes, and
Preston (1979) listed turtles from the Williams fau-
na. Based on species in these papers, the paleo-
ecology appears to have been generally similar in
both the Duck Creek and Williams fauna; that is a
permanent body of water with woodland bordering
the water, at least in places.

The larger herpetofauna of the Williams fauna
indicates (based on habitats of modem Kansas forms,
Collins, 1974) some rather specific situations (Table

1). The main habitat appears to have been a pond
or the low energy part of a stream and its bank.
Thirteen of the 21 species identified from the Wil-
liams fauna occur frequently or seasonally near this
main habitat. On the other hand, 11 of 2 1 species
are typical of prairie flats or prairie hillside habitats;
and six of 21 animals would be found in prairie-
woodland edge habitats. One animal is entirely a
woodland species.

Two extralimital species, Emydoidea blandingii
and Elaphe vulpina , occur in the Williams fauna.
The nearest Emydoidea blandingii occurs to Rice
County, Kansas, today is southcentral Nebraska (Fig.
12). The nearest Elaphe vulpina occurs to Rice
County today liTTn-^outheastern Nebraska and
northeastern Missouri (Fig. lAfUThese occurrences
may indicate somewhat cooler, moister climate for

1984

HOLMAN -HERPETOFAUNAS FROM KANSAS

37

the area in the Pleistocene. On the other hand the
rest of the herpetological species in the Williams
fauna do not indicate a climate that is different from
the area today; in fact all of them are typical animals
of the region. Certainly a “full glacial” climate or
dominant coniferous forest situation is not indicated
by the Williams fauna amphibian and reptile species.
Animals of the fauna that occur conspicuously south
of coniferous forest situations today include: Pseud-
emvs scripta. Coluber constrictor, Sonora episcopa,
Tantilla sp., Regina grahami, and Tropidoclonion
lineatum.

Comment

The statement made (Holman, 1980) that her-
petological evidence from the Sandahl local fauna
(Illinoian) did not support the classical idea of a
cooler, moister climate for the Illinoian age in Kan-
sas should be somewhat modified based on Illinoian
herpetological species in both the Duck Creek and
Williams faunas. The Sandahl fauna had two turtles,
a salamander, five anurans, three lizards, and six
snakes (17 forms), all of which are extant and occur
in the area today. But the presence of the wood frog
( Rana sylvatica) in the Duck Creek fauna and Bin-
ding's turtle (Emydoidea blandingii) and the fox
snake (E lap he vulpina) in the Williams fauna pos-
sibly indicates a somewhat cooler, moister climate.

and the presence of more woodland than occurs in
the area today. But, as previously stated, it is em-
phatically clear that a “full glacial” or Boreal cli-
mate, or a dominant coniferous forest association
is not indicated by the other herpetological species.

In the past, Pleistocene mammalian faunas have
sometimes been interpreted as indicating Boreal or
even Tundra situations when the herpetofaunas from
the same localities have been similar or identical to
those found in the areas today. Certainly a problem
of interpretation exists here. Moreover, as Stuart
(1979) indicates in his study of Pleistocene occur-
rences of the European pond turtle ( Emys orbicu-
laris) in Great Britain, a certain minimum mean
summer temperature is needed in order for the tur-
tles to breed. In the future, it would seem that studies
on critical temperatures for egg-hatching of turtles
and other reptilian species would provide exceed-
ingly important information about Pleistocene sum-
mer climates in North America.

Perhaps, rather than indicating the presence of
full Boreal or Tundral vegetation and climate, the
presence of northern extralimital mammals might
best be explained by southward glacial displacement
of both mammalian and plant species. This would
form a mosaic Pleistocene community of both an-
imals and plants. Such a mosaic community could
survive as long as the climate remained equable.

LITERATURE CITED

Auffenberg, W. 1963. The fossil snakes of Florida. Tulane
Stud. Zool., 10:131-216.

Collins, J. T. 1 974. Amphibians and reptiles in Kansas. Univ.
Kansas Mus. Publ. Nat. Hist., Ed. Ser., 1:1-283.

Conant, R. 1975. A field guide to reptiles and amphibians of
eastern and central North America. Houghton Mifflin Co.,
Boston, 429 pp.

Hall, S. A. 1972. A new Illinoian molluscan faunule from
central Kansas. Abs. Michigan Acad. Sci., Arts, Letts., 76th
Ann. Meeting, Program with Abstracts, p. 6.

Hibbard, C. W. 1960. Pliocene and Pleistocene climates in
North America. Ann. Rept. Michigan Acad. Sci., Arts, Letts.,
62:5-30.

Holman, J. A. 1962. A Texas Pleistocene herpetofauna. Co-
peia, 1978:255-261.

. 1965. Pleistocene snakes from the Seymour Formation

of Texas. Copeia, 1965:102-104.

. 1967. A Pleistocene herpetofauna from Ladds, Georgia.

Bull. Georgia Acad. Sci., 25:154-166.

. 1 969. Herpetofauna of the Pleistocene Slaton local fau-
na of Texas. Southwestern Nat., 14:203-212.

. 1971. Herpetofauna of the Sandahl local fauna (Pleis-

tocene:Illinoian) of Kansas. Contrib. Mus. Paleont., Univ.
Michigan, 23:349-355.

. 1972. Herpetofauna of the Kanapolis local fauna (Pleis-
tocene: Yarmouth) of Kansas. Michigan Acad., 5:87-98.
. 1976. Paleoclimatic implications of “ecologically in-
compatible” herpetological species (late Pleistocene: south-
eastern United States). Herpetologica, 32:291-295.

. 1980. Paleoclimatic implications of Pleistocene her-
petofaunas of eastern and central North America. Trans.
Nebraska Acad. Sci., 8:131-140.

. 1981. A review of North American Pleistocene snakes.

Publ. Mus. Michigan State, Univ. Paleont. Ser., 50:261-306.

. 1982. A fossil snake (Elaphe vulpina) from a Pliocene

ash bed in Nebraska. Trans. Nebraska Acad. Sci., 10:37-42.
Kolb, K., M. E. Nelson, and R. J. Zakrzewski. 1975. The
Duck Creek molluscan fauna (Illinoian) from Ellis County,
Kansas. Trans. Kansas Acad. Sci., 78:63-74.

Lundberg, J. G. 1975. The fossil catfishes of North America.

Univ. Michigan Mus. Paleont., Pap. Paleont., 11:1-51.
Lundelius, E. L., R. W. Graham, E. Anderson, J. Guilday, J.
A. Holman, D. Steadman, and S. D. Webb. 1983. Ter-
restrial vertebrate faunas. In Late Quaternary Environments

38

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

of the United States: the Late Pleistocene, Univ. Minnesota
Press, Minneapolis, 407 pp.

McMullen, T. L. 1975. Shrews from the late Pleistocene of
Kansas with the description of a new species of Sorex. J.
Mamm., 56:316-320.

. 1978. Mammals of the Duck Creek local fauna, late

Pleistocene of Kansas. J. Mamm., 59:374-386.

Meltzer, D., and J. I. Mead. 1983. The timing of the late
Pleistocene mammalian extinctions in North American.
Quaternary Research, 19:130-135.

Preston, R. E. 1979. Late Pleistocene cold-blooded vertebrate
faunas from the mid-continental United States. I. Reptilia;
Testudines, Crocodilia. Llmv. Michigan Mus. Paleont., Pap.
Paleont., 19:1-53.

Rogers, K. L. 1982. Herpetofaunas of the Courland Canal and
Hall Ash local faunas (Pleistocene: Early Kansan) of Jewell
Co., Kansas. J. Herpetol., 16:174-177.

Stuart, A. J. 1979. Pleistocene occurrence of the European
pond tortoise ( Emys orbicularis L.) in Britain. Boreas, 8:
359-371.

Address: The Museum, Michigan State University,

. 1982. Pleistocene vertebrates in the British Isles. Long-
man, London and New York, 212 pp.

Tihen, J. A. 1958. Comments on the osteology and phylogeny
of ambystomatid salamanders. Bull. Florida State Mus., 3:
1-50.

. 1962. A review of New World fossil bufonids. Amer.

Midland Nat., 68:1-50.

Van Devender, T. R., and J. I. Mead. 1978. Early Holocene
and late Pleistocene amphibians and reptiles in Sonoran Des-
ert packrat middens. Copeia, 1978:467-475.

Van Devender, T. R., A. M. Phillips, and J. I. Mead. 1977.
Late Pleistocene reptiles and small mammals from the lower
Grand Canyon of Arizona. Southwestern Nat., 22:49-66.
Wright, A. H., and A. A. Wright. 1949. Handbook of frogs
and toads of the United States and Canada. Comstock Publ.
Co., Ithaca, New York, 640 pp.

Zakrzewski, R. J., and J. L. Maxfield. 1971. Occurrence of
Clethrionomys in the late Pleistocene of Kansas. J. Mamm..
52:620-621.

East Lansing, Michigan 48824.

PLEISTOCENE BIRDS FROM CUMBERLAND CAVE, MARYLAND

Pierce Brodkorb and Cecile Mourer-Chauvire

ABSTRACT

Seven species of birds are recorded, including a new species of The fossil record of Bonasa umbellus, Ectopistes migratorius, and

owl, Otus guildayi. The avifauna is compatible with the Illinoian Perisoreus canadensis is extended to Illinoian time,
age previously postulated on the basis of the mammalian fauna.

INTRODUCTION

Present knowledge of the Appalachian vertebrate
fauna older than 20,000 years is limited to fossils
from three or four caves— Port Kennedy in Penn-
sylvania, Cumberland Cave in Maryland, and Trout
and perhaps Rapps caves in West Virginia. The fos-
sil vertebrates from these sites include a high per-
centage of extinct genera and species, some of north-
ern and others of southern affinities. Hay (1923)
thought the vertebrates with southern affinities to
be of Sangamonian interglacial age, but the entire
fauna is now considered to date from the Illinoian
glaciation about 100,000 years ago (Guilday, 1971).
The only bird fossils previously reported from these
caves are turkey bones ( Meleagris gal/opavo ) from
Port Kennedy Cave (Mercer, 1899) and a bone of

the ruffed grouse (Bonasa umbellus) from Cumber-
land Cave (Wetmore, 1927).

Cumberland Cave is located one-half mi south of
Cornganville, Alleghany Co., in the Appalachian
Mountains of northwestern Maryland, and close to
the Pennsylvanian border. Early exploration was
carried out by J. W. Gidley of the Smithsonian In-
stitution, who found the grouse bone among the rich
non-avian vertebrate fauna (Gidley and Gazin,
1938). In 1965 further excavation was made by Al-
len D. McCrady and Harold Hamilton, under the
direction of John E. Guilday of the Carnegie Mu-
seum of Natural History (Guilday and Handley,
1967). The bird remains collected by the Carnegie
party form the subject of the present report.

SYSTEMATIC PALEONTOLOGY

The following list describes the bird remains that
we have determined to generic or specific level. The
collection also includes many indeterminate scraps,
mostly tiny fragments of small passerines.

Family Anatidae

Anas crecca Linnaeus
Green-winged or Common Teal

Material.— CM 34027, proximal end of left carpometacarpus.

This specimen agrees with A. crecca in having the
proximal end of the carpometacarpus wider than in
the Blue- winged Teal ( Anas discors Linnaeus), as
shown in Table 1 .

The Green-winged Teal is a Holarctic species that
breeds in the north and winters as far south as the
Equator. Its three subspecies are differentiated on
minor plumage differences, and we are unable to
separate them on osteological grounds.

In Europe its fossil record extends back to the
Mindel glaciation, and possibly even to the Gunz

(Mourer-Chauvire, 1975). In America the oldest
previous records are of Illinoian or Sangamonian
age in Kansas (Galbreath, 1955) and Florida (Brod-
korb, 1957, 1959; Ligon, 1965).

Family Accipitridae

Aquila chrysaetos (Linnaeus)

Golden Eagle

Material.— CM 34018, right claw core with tip and base miss-
ing.

In eagles the claw of digit I (the hind toe) is largest,
and the claw of digit II (the inner anterior toe) is
next largest. The lateral surface of an eagle’s claw is
flat, and the medial side is rounded in the three
anterior toes. The same situation would hold in the
hind toe if all four toes were directed forward.

The damaged state of the fossil makes it difficult
to determine the digit represented. Nevertheless, the
fossil agrees in size with the Golden Eagle and ex-
ceeds the largest of our series of the Bald Eagle,

39

40

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 1 Measurements (in nun) of carpometacarpus of Anas
crecca and A. discors.

Material

Width of
trochlea

Depth of
internal rim
of trochlea

CM 34027

4.2

5.0

A. c. nimia

4.3

4.9

1 8 Aleutians

A. c. carolinensis

3. 7-4.2

4. 8-5. 2

3 8, 3 2 North America

A. c. crecca

3.8

5.1

1 8 Holland

A. discors

3. 6-3. 8

4. 6-5.0

3 <5, 2 9 North America

Haliaeetus leucocephalus (Linnaeus), as shown in
Table 2. It also agrees with Aqui/a in lacking the
pronounced distal tapering of the claws in Haliaee-
tus.

The Golden Eagle is a widespread Holarctic
species. In the eastern United States it is relatively
rare, but it has bred south in the mountains to New
York and North Carolina.

In Europe this species first appears as fossil during
the Mindel stage of the Pleistocene (Mourer-Chau-
vire, 1 975) and in America during the Illinoian stage
(Ritchie, 1980).

Family Phasianidae

Bonasa umbellus (Linnaeus)

Ruffed Grouse

Material. — USNM 1 1690, distal half of left humerus. We have
not seen this specimen reported by Wetmore (1927).

The Ruffed Grouse is a bird of North Woods and
similar environments in the mountains. In the Ap-
palachians, it extends southwards into northern
Georgia. This is the oldest known occurrence of the
species. The site at Arredondo, Florida, formerly
also thought to be of Illinoian glacial age (Brodkorb,
1959), is now considered to represent the Sanga-
monian interglacial age (Webb, 1974).

Meleagris gallopavo Linnaeus
Wild Turkey

Material.— CM 34028, upper part of left coracoid.

Male turkeys are much larger than females, but
it is not possible to determine the sex of this isolated
specimen with certainty. Its measurements are as
follows: head through scapular facet, 28.4; depth of
head, 12.2 mm. Comparison with data in the latest

Table 2 .—Measurements (in mm) of largest claw core of eagles.

Material

Plantar width
where broken

Height at
same level

CM 34018

7.9

9.7

Aquila chrysaetos
canadensis
1 3, 2 9 United States
(Pennsylvania, Texas, Wyoming)

8.70

(7.3-10.0)

9.0

(8.4-9. 7)

Haliaeetus leucocephalus
washingtoniensis
2 9, Alaska

7.35

9.45

Haliaeetus leucocephalus
leucocephalus

6.70

(5. 9-6. 9)

8.40

(7.6-9. 1)

2 8, 2 9 Florida

revision (Steadman, 1980) shows that both mea-
surements of the Cumberland Cave coracoid fall
within the size range of female M. gallopavo.

The genus first appears in Blancan times (late
Pliocene or early Pleistocene) and M. gallopavo is
known in several localities of Illinoian age.

Wild turkeys formerly occurred in most of the
United States and much of Mexico. Their major
requirement seems to be trees for roosting at night.

Family Columbidae

Ectopistes migratorius (Linnaeus)
Passenger Pigeon

Material.— CM 34020, upper end of right coracoid, lacking
procoracoid process.

In both the Passenger Pigeon and the feral do-
mestic pigeon ( Columba livia) the coracoid is much
larger than in the Mourning Dove ( Zenaida ma-
croura). Measurements of the wild species are given
in Table 3. The coracoid of Ectopistes is shorter, its
proximal end is wider, and its shaft is narrower than

Table 3.— Measurements (in mm) of coracoid o/Ectopistes and
Zenaida.

Material

Clavicular facet

Humeral facet

Head

through

procoracoid

Length

Width

Length

Width

CM 34020

5.3

3.6

6.1

3.8

9.5 +

Ectopistes

4.3-5. 1

3. 2-4.0

5. 4-6.6

4.3-5. 1

9.7-9. 9

migratorius

(n = 4)

Zenaida

3. 3-4.3

2.4-2. 5

4.3-5. 1

2. 1-2.6

6. 5-7.0

macroura
(n = 20)

1984

BRODKORB AND MOURER-CHAUVIRE- CUMBERLAND CAVE BIRDS

41

in C. livia. Other differences between these two
species are also apparent. In Ectopistes the brachial
tuberosity narrows toward the tip, which is rather
pointed, whereas in C. livia the tuberosity is wider
and its tip is more rounded. Additionally, in Ec-
topistes the posterior face of the bone above the
sternal facet is much less pneumatic than in C. livia.

Exterminated 70 years ago, the Passenger Pigeon
formerly bred in the North Woods and thence south
to the Midwest and the Ohio River Valley. It win-
tered in the Southeast.

It is reported from pre-Columbian and Wiscon-
sinan age sites west to Los Angeles (Howard, 1937)
and south to southern Florida (Woolfenden, 1959;
Weigel, 1962). It is also known from sites now
throught to be of Sangamonian age, at Reddick,
Haile, and Arredondo in northern Florida (Brod-
korb, 1957, 1959; Ligon, 1965).

Family Strigidae

Otus guildayi, new species
Fig. 1

Etymology. — Named in memory of John E. Guilday, late spe-
cialist in Pleistocene and Holocene mammals of eastern North
America, who directed the excavation of the Cumberland Cave
fossils, and who sent us the bird material for identification.

Holotype.— C arnegie Museum of Natural Histo-
ry, Section of Vertebrate Fossils, CM 8040, left tar-
sometatarsus, lacking proximal end. Collected in
1965 by Allen D. McCrady and Harold Hamilton.

Type locality. - - Cumberland Cave, one-half mi
south of Corriganville, Alleghany County, in the
Appalachian Mountains of northwestern Maryland.

Age. — Middle Pleistocene, Illinoian glacial age.

Diagnosis.— A strigid owl, referable to the genus
Otus Pennant by having the tarsometatarsus of me-
dium length; its shaft moderately long and slender,
with the borders nearly parallel (without marked
proximal and distal expansion, and without marked

C

Fig. 1. — Otus guildayi , new species, CM 8040, holotype, Cum-
berland Cave, Maryland. A) anterior, B) posterior, and C) distal
views. Scale is 1 cm. Photographs by Ronald G. Wolff.

mid-length constriction); groove of middle trochlea
deep.

Similar to Recent and Pleistocene Otus asio (Lin-
naeus), but larger (see Table 4); shaft scarcely con-

Table 4.— Measurements (in mm) of tarsometatarsus of Otus.

Material

Total length

Partial length1

Distal

width

Least

width shaft

Trochlear gap2

Inner

trochlea width

Middle

trochlea width

O. guildayi, holotype

38. 563

25.4

8.2

4.2

4.2

3.8

3.2

O. a. naevius

33.25

21.90

7.17

3.50

4.1

3.65

2.80

n = 3 (2 2, 1 ?)

(32.3-34.2)

(20.6-23.2)

(6. 8-7.4)

(3. 3-3.8)

(3. 6-3. 7)

(2.5-3. 1)

O. a. asio

30.39

20.62

6.29

2.90

2.99

2.85

2.40

n = 12 (5 3, 7 2)

(29.1-31.5)

(19.1-21.8)

(5.9-6. 8)

(2.7-3. 3)

(2. 3-3. 5)

(2.3-3. 3)

(2. 0-2. 8)

1 Length from distal end of Tuberculum musculi tibialis cranialis to distal end of middle trochlea.

2 Width between tips of outer and inner trochleae.

3 Estimate; length as preserved is 31.6 mm.

42

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 5.— Measurements (in mm) of tarsometatarsus o/ Perisoreus and Cyanocitta.

Material

Total length

Proximal width

Proximal width
of shaft

Least width
of shaft

Width through
trochleae

CM 29274

36.7

5.1

2.8

1.8

(3.1)*

C. cristata bromia

33.6-37.5

4.4-4. 9

2.7-3. 2

1.5-1. 8

3. 2-3.4

3 S, 2 9 New Brunswick,

(35.34)

(4.68)

(2.98)

(1.68)

(3.32)

Maine, Pennsylvania

P. canadensis canadensis

33.0-36.0

4.0-4. 8

2. 3-2. 7

1. 1-1.6

2.7-3. 4

2 <3, 4 9 Ontario, Maine

(35.20)

(4.45)

(2.50)

(1.43)

(3.17)

P. c. pacificus

36.5

4.3

2.6

1.6

3.0

1 6 Alaska

P. c. capitalis

36.8-37.3

4.6

2.4-2. 7

1. 5-1.6

3.2-3. 3

2 8 Rocky Mts

(37.05)

(2.55)

(1.55)

(3.25)

P. infaustus ruthenus

36.3

4.9

2.8

1.6

3.4

1 adult, Karelia

* As preserved. External trochlea is damaged.

stricted in middle, and its sides nearly straight; outer
trochlea thrust plantad and mediad toward middle
trochlea (in O. asio outer trochlea thrust laterad away
from middle trochlea, leaving a relatively wider gap
between tips of the outer and inner trochleae).

Much larger than other New World species of
Otus.

Otus asio is reported from many pre-Columbian
and Wisconsinan age sites throughout the United
States. It also occurs in three Sangamonian age lo-
calities, at Reddick, Haile, and Arredondo, Florida
(Brodkorb, 1957, 1959; Ligon, 1965).

Family Corvidae

Perisoreus canadensis (Linnaeus)

Canada Jay

Material. — CM 24274, right tarsometatarsus, complete except

for damage to hypotarsus; the shaft has been broken in two and
glued together.

Tarsometatarsi of Perisoreus and the Blue Jay
( Cyanocitta cristata) are similar in general size, but
the element is distally more slender in Perisoreus ,
stouter in Cyanocitta (see Table 5).

Perisoreus is a Holarctic genus of the Northern
Coniferous Forest. In eastern North America the
southern limit of Perisoreus is in northern New En-
gland and northeastern New York state.

The only previous record of the species is from
Natural Chimneys, Virginia (Wetmore, 1962). This
locality is from late Wisconsinan to Recent age
(Guilday, 1962). The two other members of this
genus, the Palearctic P. infaustus (Linnaeus) and P.
internigrans (Thayer and Bangs), are unknown as
fossils.

DISCUSSION

Climate of the Deposit

Cumberland Cave lies within the modern geo-
graphic range of five of the species identified. They
are Anas crecca, Aqui/a chrysaetos, Bonasa umbel-
lus, Meleagris ga/lopavo, and Ectopistes migrator-
ius. Their modern distribution also extends very
much farther in all directions. Otus guildayi is as
yet unknown elsewhere. Its presumed descendant is
the screech owl, Otus asio, whose modern geograph-

ic range is similarly extensive. These six species thus
tell us nothing about any difference in climate when
the fossils were deposited.

The case of Perisoreus canadensis is more helpful.
Today its closest occurrence, about 400 mi farther
north, is in the Adirondacks Mountains. We may
thus conclude that the climate at the time of de-
position was cooler than at present, and that the
avifauna lived during a glacial age.

1984

BRODKORB AND MOURER-CHAUVIRE-CUMBERLAND CAVE BIRDS

43

Age of the Deposit

The known fossil record of the Cumberland Cave
birds gives an insight as to the age of the deposit.
Anas crecca is known in Europe from the Mindel
glaciation and possibly earlier, and its North Amer-
ican record begins with the Illinoian glaciation
Aquila chrysaetos first appears in Europe during the
Mindel stage, and during the Illinoian stage in North
America. Me/eagris gallopavo is known from sev-
eral Illinoian age localities.

Bonasa umbellus and Ectopistes migratorius are

previously known from localities of Sangamonian
interglacial age, as is Otus asio, the possible descen-
dant of Otus guildayi.

The only other fossil record of Perisoreus cana-
densis is of Wisconsinan glacial age, supposedly the
most severe of the American stages.

All the data of the paleoavifauna conform with
the hypothesis proposed by Guilday of an Illinoian
age for the Cumberland Cave deposit.

RESUME

L’avifaune de la Cumberland Cave comporte sept taxons parmi
lesquels se trouve une nouvelle espece de hibou, Otus guildayi.
Elle est compatible avec Page Illinoien indique par la faune de

mammiferes. C’est la plus ancienne apparition connue jusqu'a
present pour les especes Bonasa umbellus. Ectopistes migratorius
et Perisoreus infaustus.

LITERATURE CITED

Brodkorb, P. 1957. New passerine birds from the Pleistocene
of Reddick, Florida. J. Paleont., 31:129-138.

. 1959. The Pleistocene avifauna of Arredondo, Florida.

Bull. Florida State Mus., Biol. Sci., 4:269-291.

Galbreath, E. C. 1955. An avifauna from the Pleistocene of
central Kansas. Wilson Bull., 67:62-63.

Gidley, J. W., and C. L. Gazin. 1938. The Pleistocene verte-
brate fauna from Cumberland Cave, Maryland. Bull. U.S.
Nat. Mus., 171:1-199.

Guilday, J. E. 1 962. The Pleistocene local fauna of the Natural
Chimneys, Augusta County, Virginia. Ann. Carnegie Mus.,
36:87-122.

. 1971. The Pleistocene history of the Appalachian mam-
mal fauna. Virginia Polytech. Inst. State Univ., Research
Div. Monogr., 4:233-262.

Guilday, J. E., and C. O. Handley, Jr. 1967. A new Pero-
myscus (Rodentia : Cricetidae) from the Pleistocene of Mary-
land. Ann. Carnegie Mus., 39:91-103.

Hay, O. P. 1923. The Pleistocene of North America and its
vertebrated animals from the states east of the Mississippi
River and from the Canadian provinces east of longitude
95°. Carnegie Inst. Washington Publ., 322:1-499.

Howard, H. 1937. A Pleistocene record ofthe Passenger Pigeon
in California. Condor, 39:12-14.

Ligon, J. D. 1965. A Pleistocene avifauna from Haile, Florida.
Bull. Florida State Mus., 10:127-158 (published Jan. 3, 1966)

Mercer, H. C. 1 899. The bone cave at Port Kennedy, Pa., and
its partial excavation in 1 894, 1895, and 1 896. J. Acad. Nat.
Sci. Philadelphia, 11:269-286.

Mourer-Chauvire, C. 1975. Les oiseaux du Pleistocene moy-
en et superieur de France. Doc. Lab. Geol. Faculte Sci. Lyon,
64:1-624.

Ritchie, T. L. 1980. Two mid-Pleistocene avifaunas from Cole-
man, Florida. Bull. Florida State Mus., Biol. Sci., 26:1-36.

Steadman, D. W. 1980. A review of the osteology and pa-
leontology of turkeys (Aves: Meleagridinae). Contrib. Sci.,
Nat. Hist. Mus. Los Angeles Co., 330:131-207.

Webb, S. D. 1974. Pleistocene mammals of Florida. Univ.
Presses Florida, Gainesville, 270 pp.

Weigel, R. D. 1962. Fossil vertebrates of Vero, Florida. Florida
Geol. Survey, Special Bull., 10:1-59 pp. Published Jan., 1963.

Wetmore, A. 1927. A record of the ruffed grouse from the
Pleistocene of Maryland. Auk, 44:561.

. 1962. Notes on fossil and subfossil birds. Smithsonian

Misc. Coll., 145(2): 1-1 7.

Woolfenden, G. E. 1959. A Pleistocene avifauna from Rock
Spring, Florida. Wilson Bull., 71:183-187.

Address (Brodkorb): Department of Zoology, University of Florida, Gainesville, Florida 3261 1.

Address (Mourer-Chauvire): Centre de Paleontologie stratigraphique et Paleoecologie de l'Universite Claude

Bernard, 69622 Villeurbanne Cedex, France.

A VERY LARGE ENIGMATIC OWL (AVES: STRIGIDAE) FROM THE
LATE PLEISTOCENE AT LADDS, GEORGIA

Storrs L. Olson

ABSTRACT

An undescribed species of owl is represented in the late Pleis- Strigidae. This is most similar to the mandible in living owls of

tocene (Rancholabrean) deposits at Ladds, Georgia, by a man- the genus Strix. Paleontologists are alerted to seek more diag-

dibular symphysis that is larger than in any living species of nostic specimens so that the species may eventually be described.

INTRODUCTION

One of very few vertebrate faunas known from
the late Pleistocene (Rancholabrean) of Georgia was
reported from fissure fillings in a limestone quarry
at Ladds, Bartow County (Lipps and Ray, 1967;
Ray, 1965, 1967). Some of the avian remains from
this site were identified by Wetmore ( 1 967), of which
the most interesting record was that of a Spruce
Grouse ( Dendragapus canadensis), a species now
occurring mainly in boreal coniferous forests of
Canada. Whereas this and certain of the mammalian
species are characteristic of cooler northern regions
today, some of the other mammals have southern
affinities (Ray, 1967; Kurten and Anderson, 1980).
Although the deposits at Ladds may have been het-
erochronic, such “disharmonious” faunas are typ-

ical of many late Pleistocene fossil deposits (Lun-
delius et al., 1 983) and are believed to reflect a more
equable climate in the Pleistocene.

Among some additional material from the Ladds
deposits in the National Museum of Natural His-
tory, Smithsonian Institution (USNM), I encoun-
tered the anterior portion of a mandible of an owl
that appears to have been larger than any of the
living species of Strigiformes, although I do not con-
sider the specimen to be sufficiently diagnostic to
be named at this time. Nevertheless, it represents a
significant addition to the Pleistocene fauna of east-
ern North America that should be included in faunal
surveys and that should be looked for in other Pleis-
tocene deposits.

DESCRIPTION

The specimen (USNM 214769) is a mandibular
symphysis with unequal portions of the rami pre-
served on either side. The length of the symphysis
as preserved is 1 1.5 mm, but as some of the thin,
fragile tip is missing, this measurement would have
exceeded 1 2 mm in the intact specimen. The width
at the posterior margin of the symphysis is 15.1 mm.
The specimen is referable to the Strigidae, the sym-
physis being broader and more truncate than in the
Tytonidae, in which the symphysis is narrow and
elongate.

Compared with the three largest North American
owls ( Bubo virginianus, Nyctea scandiaca, and Strix
nebulosa ), the fossil is seen to be much larger (Fig.
1 ). The symphysis is proportionately longer and nar-
rower and the rami less divergent than in either
Bubo or Nyctea. Furthermore, the two large nutrient
foramina on the ventral surface of the tip are more
widely separated and situated more on the lateral
surfaces of the symphysis than in Bubo or Nyctea.
In these and other details the specimen is more sim-
ilar to Strix.

DISCUSSION

Howard (1933) described a presumably extinct
owl from the tar pits at Rancho La Brea, California,
as Strix brea. This was characterized as being larger
than either of the living species S. varia or S. oc-
cidentalis, but no size comparisons were made with

the panboreal Great Gray Owl, S. nebulosa, because
at that time the species was regarded as belonging
in a monotypic genus, Scotiaptex. From Table 1, a
rather confusing picture of the size of Strix brea
emerges. The measurements for the rostrum and

44

1984

OLSON-OWL FROM LADDS

45

Fig. 1. — Mandibular symphyses of owls in dorsal view (top row) and ventral view (bottow row): A) Great Gray Owl, Strix nebulosa
(USNM 289980); B) fossil specimen from Ladds Quarry, Georgia (USNM 214769); C) Snowy Owl, Nyctea scandiaca (USNM 19901);
D) Great Homed Owl, Bubo virginianus (USNM 289978). Natural size.

tibiotarsus all lie within the range of variation of a
very small sample of S. nebulosa and those for the
femur and wing elements are either somewhat
smaller or in the lower end of the size range of S.
nebulosa. On the other hand, the holotype and eight
referred tarsbmetatarsi of Strix brea are all consid-
erably larger than in S', nebulosa. The much greater
size of the mandible of the Ladds owl as compared
to S. nebulosa makes it unlikely that this specimen
is referable to S. brea, whatever the status of that
species may be.

The only extinct strigid owls of immense size be-
long to the genus Ornimegalonyx, originally known

from a single species from Quaternary cave deposits
in Cuba (see Arredondo, 1 976). Kurochkin and Mayo
(1973) alluded to the possible existence of additional
species of Ornimegalonyx and Arredondo (1982)
has recently named three as new. Unfortunately, the
mandibular symphysis has not been reported for
Ornimegalonyx. Because Ornimegalonyx is be-
lieved to be closely related to Strix and Ciccaba
(Kurochkin, personal communication), it is possible
that with better material the Ladds owl, which also
appears close to Strix, may shed some light on the
geographical origins of the Cuban birds.

Table 1 .—Measurements of the extinct owl, Strix brea (from Howard, 1933) and the mandible from Ladds, Georgia, compared with
three specimens of the extant Great Gray Owl, Strix nebulosa (from left to right USNM 502543, male ; USNM 289980, female; USNM

289429, female).

Measurements

Strix brea

Strix nebulosa

Ladds owl

Height rostrum

20.7 (n =

1)

—

20.8

20.9

—

Breadth rostrum

21.5 (n =

1)

—

25.0

23.7

—

Length humerus

1 12.5-121.3 (n =

3)

120.0

134.8

142.0

—

Length carpometacarpus

56.2-59.9 (n =

5)

58.0

66.3

68.0

—

Length femur

75.6-76.6 (n =

4)

78.9

87.2

91.6

—

Length tibiotarsus

1 12.7-120.0 (n =

10)

106.4

120.0

126.6

—

Length tarsometatarsus

63.5-68.0 (n =

9)

50.0

57.0

59.0

—

Length mandibular symphysis

—

8.1

7.7

12 +

Breadth mandibular symphysis

-

9.6

9.9

14.6

46

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

ACKNOWLEDGMENTS

Lewis Lipps encouraged me to look through the additional bird
material from Ladds quarry and supplied useful information.
Kenneth E. Campbell, Jr., John Farrand, Jr., and Clayton E. Ray

offered helpful comments on an early draft of the manuscript,
and several later drafts were criticized by David W. Steadman
The photographs are by Victor E. Krantz.

LITERATURE CITED

Arredondo, O. 1976. The great predatory birds of the Pleis-
tocene of Cuba. In Collected Papers in Avian Paleontology
Efonoring the 90th Birthday of Alexander Wetmore (S. L.
Olson, ed.), Smithsonian Contr. Paleobiol., 27:169-187.

. 1982. Los Strigiformes fosiles del Pleistoceno Cubano.

Bol. Soc. Venezolana Cien. Nat., 140:33-55.

Howard, H. 1933. A new species of owl from the Pleistocene
of Rancho La Brea, California. Condor, 35:66-69.
Kurochkin, E. N., and N. Mayo. 1973. Las lechuzas gigantes
del Pleistoceno superior de Cuba. Acad. Cien. Cuba Inst.
Geol. Resumenes, Comunicaciones, y Notas del V. Consejo
Cientifko Actas, 3:56-60.

Kurten, B., and E. Anderson. 1980. Pleistocene mammals of
North America. Columbia Univ. Press, New York, 442 pp.
Lipps, L., and C. E. Ray. 1967. The Pleistocene fossiliferous

deposit at Ladds, Bartow County, Georgia. Bull. Georgia
Acad. Sci., 25:113-119.

Lundelius, E. L., Jr., R. W. Graham, E. Anderson, J. E.
Guilday, J. A. Holman, D. W. Steadman, and S. D. Webb.
1983. Terrestrial vertebrate faunas. Pp. 311-353, in The
Late Pleistocene (S. C. Porter, ed.), vol. 1 of. Late Quaternary
environments of the United States (H. E. Wright, Jr., ed.),
Umv. Minnesota Press, Minneapolis.

Ray, C. E. 1965. A new chipmunk, Tamias aristus, from the
Pleistocene of Georgia. J. Paleont., 39:1016-1022.

. 1967. Pleistocene mammals from Ladds, Bartow Coun-
ty, Georgia. Bull. Georgia Acad. Sci., 25:120-150.
Wetmore, A. 1967. Pleistocene Aves from Ladds, Georgia.
Bull. Georgia Acad. Sci., 25:151-153.

Address: Department of Vertebrate Zoology, National Museum of Natural History, Smithsonian Institution,

Washington, D.C. 20560.

A MIDDLE PLEISTOCENE (LATE IRVINGTONIAN) AVIFAUNA FROM

PAYNE CREEK, CENTRAL FLORIDA

David W. Steadman

ABSTRACT

Ten taxa of birds are reported from a late Irvingtonian fossil
site at Payne Creek, in the Bone Valley region of central Florida.
No radiometric dates are available for this alluvial deposit, whose
age is based only upon mammalian biochronology. Unlike the
extensive marine avifauna of the underlying Bone Valley For-
mation (Hemphillian), the Payne Creek avifauna represents species
characteristic of fresh water and terrestrial habitats. Tachybaptus
dominicus is reported for the first time as a North American
fossil. This species, which no longer occurs in Florida, is another
member of the Pleistocene “Gulf Coast Savanna Corridor” fauna.
Other living species reported are Podilymbus podiceps, Colinus
virginianus, Rallus cf. R. limicola, and Zenaida macroura, each

of which still lives in Florida. Less diagnostic fossils are referred
to “Anserinae, genus and species indeterminate,” Anas cf. A.
discors or A. cyanoptera, an extinct but poorly understood species
of Bucephala, Limnodromus sp., and “Icterinae, genus and species
indeterminate.” The following described Pliocene or Pleistocene
species are regarded as of questionable validity or as being based
on inadequate material, or both: Podilymbus magnus Shufeldt
1913, P. wetmorei Storer 1976, Podiceps dixi Brodkorb 1963,
Bucephala fossilis Howard 1963, B. ossivallis Brodkorb 1955,
Colinus suilium Brodkorb 1959, and Zenaida prior Brodkorb
1969.

INTRODUCTION

The Bone Valley region of central Florida is well
known for phosphatic deposits that have produced
rich faunas of Mio-Pliocene (Hemphillian land
mammal age) vertebrates. The fossil birds of Bone
Valley were reviewed by Brodkorb (1955), but only
part of the hundreds of additional avian fossils that
have been collected since that time have ever been
mentioned briefly (Brodkorb, 1972), and an updated
review of the Hemphillian fauna is needed. This
report describes the birds of a much younger ver-
tebrate fossil fauna from the Bone Valley region of
Polk County. The site, known as Payne Creek, is an
alluvial deposit discovered and collected by John S.
Waldrop and Michael K. Frazier. No radiometric
dates are available for the Payne Creek local fauna,
but the fossil mammals (especially Ondatra cf. O.
annectens) indicate a biostratigraphic age of late Ir-
vingtonian (M. K. Frazier, personal communica-
tion). This places the Payne Creek local fauna at
approximately 4-6 million years younger than typ-
ical Hemphillian vertebrate faunas of the Bone Val-
ley region. Other Irvingtonian mammal sites in
Florida are described by Webb (1974).

In Florida and throughout North America, Ir-
vingtonian avifaunas are much rarer than those of
the late Pleistocene Rancholabrean land mammal
age. Six Irvingtonian avifaunas, including Payne

Creek, are known from Florida, whereas fossil birds
are known from at least 22 localities from only the
latest Rancholabrean (25,000-10,000 years BP) of
Florida (Lundelius et al., 1983). The Irvingtonian
sites include the Williston local fauna in Levy Coun-
ty (Holman, 1959) and the Coleman IIA local fauna
in Sumter County (Ritchie, 1 980). Steadman ( 1 980)
reviewed the fossil turkeys from these two faunas
and from three others in the Florida Irvingtonian —
Inglis IA (Citrus County), Haile XVIA (Alachua
County), and Santa Fe River IIA (Gilchrist County).

Bird bones are often fragmentary in alluvial sites,
and Payne Creek is no exception. The avian fossils
from Payne Creek include 27 specimens for which
no identification of any sort could be safely made.
The remaining 29 specimens are sufficiently diag-
nostic to allow identification at least to the familial
level (Table 1), but I have identified to species only
five of the 10 taxa of birds. This conservative treat-
ment is dictated by both the quality of the fossils
and the poorly defined nature of species-level sys-
tematics in early Pleistocene birds.

Unless noted otherwise, sequence and nomenclature follows
Brodkorb’s Catalogue of Fossil Birds, Parts 1-5 (1963-1978).
Distributions of living species are taken from the American Or-
nithologists’ Union Checklist (1957). All specimens are housed
in the collection of the Timberlane Research Organization, Lake
Wales, Florida.

47

48

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 1— Left coracoid of Tachybaptus dominicus. A, D) modem specimen (USNM 343476) in dorsal and ventral aspects (scapula
attached). B, C) fossil from Payne Creek in dorsal and ventral aspects. All specimens are 2 x .

SYSTEMATIC PALEONTOLOGY

Order Podicipediformes
Family Podicipedidae

Tachybaptus dominicus (Linnaeus)

Fig. 1, Table 2

Material. — Coracoid, lacking sternal end.

I follow Storer (1979) in separating the genus
Tachybaptus from Podiceps. This specimen cannot
be distinguished qualitatively from the coracoid in
five modern specimens of T. dominicus, with which
it also agrees in size, being much smaller than in
other living species of New World grebes or the fossil
species reported by Brodkorb (1963a), Murray
(1967), and Storer (1976).

Table 1 .—Birds of the Payne Creek local fauna, Polk County,
Florida. For each taxon, the minimum number of individuals is
one.

Tachybaptus dominicus— Least Grebe
Podilymbus podiceps— Pied-billed Grebe
Podicipedidae, genus and species indeterminate— grebe
Anserinae, genus and species indeterminate— goose
Anas cf. A. discors or A. cyanoptera— probable Blue-winged or
Cinnamon Teal

Bucephala sp.— goldeneye-type duck

Anatidae, genus and species indeterminate— duck

Colinus virginianus— Bobwhite

Rallus cf. R. limicola— probable Virginia Rail

Limnodromus sp.— dowitcher

Zenaida macroura— Mourning Dove

Ictennae, genus and species indeterminate — blackbird

This is the first fossil record of T. dominicus from
North America. The only other paleontological re-
port of the species is from the apparently Pleistocene
site of Lapa da Escrivania, Brazil (Winge, 1888:4).
T. dominicus is one of many species of vertebrates
with tropical or southwestern affinities that inhab-
ited Florida during the late Cenozoic. It is a wide-
spread species today, occurring essentially through-
out the lowland Neotropics and much of the West
Indies (Storer, 1979). The modern distribution of
T. dominicus most closely approaches Florida at its
northern limits in the Bahamas. On the North
American continent, T. dominicus occurs no farther
north than southern Texas, with accidental records
in Louisiana (Eyster, 1978). Unless T. dominicus
colonized Florida via the West Indies and spread
no further, then it seems that its distribution was
once continuous across the southern Gulf region of
the United States, lending additional support to the

Table 2 . — Measurements (in mm) of the coracoid in modern and
fossil Tachybaptus dominicus.

Specimens

Least width
of shaft

Length of
glenoid facet

Modem skeletons (N = 5)

(4 males, 1 unsexed; USNM
7034, 343476, 344901,
345750, 347811)

Payne Creek fossil

1.5-1. 8
1.8

4.0-4. 5
4.4

1984

STEADMAN -PAYNE CREEK AVIFAUNA

49

existence of a faunal connection between northern
Mexico and Florida during the Pleistocene (the “Gulf
Coast savanna corridor”; see Webb, 1974; Stead-
man, 1980; Pregill and Olson, 1981).

Podilymbus podiceps (Linnaeus)

Material. — Humeral end of scapula, distal end of humerus,
distal end of femur, proximal end of tibiotarsus, distal end of
tarsometatarsus, anterior one-third of notorium.

Measurements of the Payne Creek fossils (distal
width of humerus, 6.5 mm; distal width of femur,
9.9 mm; distal width of tarsometatarsus, 6.0 mm)
are within the ranges of late Pleistocene and modern
specimens of P. podiceps (Murray, 1967; Storer,
1976), with which they also agree qualitatively.

Podilymbus podiceps has been recorded in many
Pleistocene localities in Florida and elsewhere in
North America (Brodkorb, 1963a), although the only
other Irvingtonian record is from the Coleman IIA
local fauna (Ritchie, 1980). Three fossil species of
Podilymbus have been described— P. majusculus
Murray, 1967, of the Hagerman local fauna, Idaho
(Blancan land mammal age); P. wetmorei Storer,
1976, of the Reddick local fauna, Florida (Ran-
cholabrean land mammal age); and P. magnus Shu-
feldt, 1913, of Fossil Lake, Oregon (Rancholabrean),
which was synonymized with P. podiceps by Wet-
more (1937). Although Brodkorb (1963a:230) sug-
gested that a subspecies, P. podiceps magnus, might
be recognizable for certain fossils from Florida and
Fossil Lake, this was disproven by Storer (1976),
who showed the specimens of P. magnus to fall
within the normal range of variation of living P.
podiceps. P. wetmorei is known from two tarsome-
tatarsi and two femora that differ from those of P.
podcieps only in being stouter. Thus the status of P.
wetmorei as a distinct species is also suspicious.

Podicipedidae, genus and species indeterminate

Material. — Proximal end of humerus, ventral half of cervical
vertebra, distal end of pedal phalanx.

These fragmentary specimens are of a size com-
patible with either Podilymbus podiceps or Podiceps
auritus, each of which occurs commonly in Florida,
both living and as fossils. Because these specimens
may pertain to Podilymbus podiceps, they do not
necessarily represent a new taxon for the Payne Creek
fauna.

In reviewing the literature on osteology and sys-
tematics of Quaternary grebes of North America,
the unsatisfactory nature of Podiceps dixi Brodkorb,

1963 b, became apparent. This name was based on
a fragmentary and undiagnostic proximal end of a
carpometacarpus from the Reddick local fauna,
Florida (Rancholabrean land mammal age). The only
difference noted by Brodkorb (1963 b) between Pod-
iceps dixi and the living P. auritus was the slightly
greater size of the former. Grebes are known to be
sexually dimorphic in size (Storer, 1976), yet neither
sex nor sample size were given for the modern spec-
imens of P. auritus measured by Brodkorb (1963 b).
Even if the putative difference in size between P.
dixi and P. auritus were demonstrable from a single
fragment, I do not believe that such a specimen can
be diagnosed meaningfully as a new, extinct species.
In the absence of corroborative fossils, it is prefer-
able to regard Podiceps dixi as a synonym of P.
auritus.

Order Anseriformes
Family Anatidae

Anserinae, genus and species indeterminate

Material. — Proximal end and two distal ends of pedal pha-
langes.

These undiagnostic specimens belong to some
taxon of goose. There is little or nothing in the post-
cranial osteology of North American geese, to say
nothing of that of the pedal phalanges, to distinguish
the genera Chen, Anser, and Branta (Woolfenden,
1961). The only anserine reported from the Pleis-
tocene of Florida is Branta canadensis from several
sites of Rancholabrean age (Brodkorb, 1964) and
the Irvingtonian Coleman IIA local fauna (Ritchie,
1980). The only other Irvingtonian record of any
goose is of B. canadensis from the type Irvington
fauna in California (Wetmore, 1956:26).

Anas cf. A. discors or A. cyanoptera

Material. — Distal end of humerus, proximal half of carpometa-
carpus.

These specimens are from a teal the size of Anas
discors or the closely related and osteologically very
similar A. cyanoptera. The only other Irvingtonian
record of either species is of A. discors from the
Williston local fauna, Florida (Holman, 1959). A.
discors is a common Pleistocene species in Florida
and elsewhere, whereas the fossil record of A. cy-
anoptera is confined to the late Pleistocene of Cal-
ifornia and Oregon (Brodkorb, 1964). These specific
assignments were doubtless based on geographical
probability.

50

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Bucephala sp., ?aff. B. ossivallis or B. fossilis

Material. — Nearly complete tarsometatarsus, slightly eroded
on both ends.

This specimen closely resembles the tarsometa-
tarsi in modern diving ducks Bucephala albeola and
the larger B. clangula, but is intermediate in size
and represents an extinct form. It agrees with Bu-
cephala and differs from the related genera Mergus
and Lophodytes in all of the characters outlined by
Woolfenden (1961) and Alvarez and Olson (1978).
Its size suggests affinities to B. ossivallis Brodkorb,
1955, from the Bone Valley Formation (Hemphil-
lian land mammal age), or B. fossilis Howard, 1963,
from the Vallecito Creek local fauna, California (Ir-
vingtonian). B. ossivallis is known only from the
humeral two-thirds of a coracoid, and B. fossilis is
based upon the proximal end of a carpometacarpus
and a proximal fragment of humerus. Therefore nei-
ther is directly comparable to the Payne Creek fossil.
Brodkorb (1964) noted that the distal end of a car-
pometacarpus from the Hagerman local fauna, Ida-
ho (Blancan), listed as Bucephala sp. by Brodkorb
(1958; 1961), may possibly be referable to B. fossilis.
However, only the proximal end of the carpome-
tacarpus is known in B. fossilis. The only other pre-
Rancholabrean record of Bucephala is the living B.
albeola from the Blancan Rexroad local fauna, Kan-
sas (Wetmore, 1944).

In summary, two very poorly known named
species, Bucephala ossivallis Brodkorb, 1955, and
B. fossilis Howard, 1963, represent Hemphillian and
Irvingtonian forms of Bucephala that were inter-
mediate in size between the living B. clangula and
B. albeola. These extinct species are based on spec-
imens inferior to that from Payne Creek, which ap-
pears to represent a species of similar size. Because
of their fragmentary nature and because none of the
elements is comparable, it is impossible to say at
this time how many species of Bucephala are rep-
resented or whether the fossils are from a single
species lineage. This is, however, the only certainly
extinct species of bird represented in the Payne Creek
local fauna.

Anatidae, genus and species indeterminate

Material. — Antero-dorsal fragment of sternum, tiny distal frag-
ment of humerus, two distal ends of carpometacarpi, proximal
half of manus digit II, phalanx I.

Each of these specimens represents a small, teal-
sized anatid, but otherwise lacks diagnostic features,
even at the subfamilial level. These fossils do not
necessarily represent a new taxon for the Payne Creek

fauna because they could possibly pertain to Anas
discors or A. cyanoptera.

Order Galliformes

Family Phasianidae

Colinus virginianus Linnaeus

Material. — Distal end of ulna, middle one-third of synsacrum,
distal end of tibiotarsus.

The distal width of the tibiotarsus (5.2 mm) is the
only measurement that I could obtain from these
specimens to compare with the extensive series of
measurements of living and fossil Colinus in Hol-
man (1961). (Unfortunately, Ritchie [1980] provid-
ed means but no ranges for his measurements of
Pleistocene Colinus in Florida.) The measurement
5.2 mm is within the range of living C. virginianus
as well as the named Pleistocene species C. suilium
Brodkorb, 1959. Holman (1961) regarded C. suil-
ium as a large temporal form of C. virginianus. I
agree, and because of the lack of consistent quali-
tative differences and their substantial overlap in
size, I regard C. suilium as a synonym of C. virgini-
anus. Fossils of Colinus occur in many Pleistocene
sites in Florida (Brodkorb, 1964), including the Ir-
vingtonian sites of Williston (Holman, 1959) and
Coleman IIA (Ritchie, 1 980). There is a north-south
decline in size in living C. virginianus of the eastern
United States (Holman, 1961), so the larger fossils
probably reflect a cooler climate rather than another
species.

Order Ralliformes
Family Rallidae

Rallus cf. R. limicola Vieillot

Material. — Humeral end of coracoid.

This coracoid resembles that of R. limicola and
is qualitatively separable from that of the similarly-
sized Porzana Carolina. It is, however, too incom-
plete for unequivocal identification.

R. limicola occurs in several Pleistocene localities
in Florida (Brodkorb, 1967; Olson, 1974). Appar-
ently it was larger in late Pleistocene times than
today (Olson, 1974). From other Irvingtonian sites,
R. limicola is known only from a sternal fragment
from Vallecito Creek, California, referred question-
ably to this species by Howard (1963).

Order Charadriiformes
Family Scolopacidae

Limnodromus sp.

Material. — Distal end of humerus lacking internal condyle.

1984

STEADMAN -PAYNE CREEK AVIFAUNA

51

This fossil agrees with the humerus of L. griseus
and differs from that of L. scolopaceus in that the
ectepicondylar spur forms a more acute angle with
the shaft. It resembles that of L. scolopaceus versus
L. griseus, however, in having the ectepicondylar
spur less sharply pointed. It is within the size range
of both living species, and thus could represent either
of these or perhaps an extinct form.

L. griseus is unknown in the Pleistocene of Flor-
ida, whereas L. scolopaceus was recorded from the
Rock Spring local fauna of Rancholabrean age
(Woolfenden, 1959). Ligon (1965) reported Lim-
nodromus sp. from the Rancholabrean Haile local
fauna, Florida (listed questionably as L. scolopaceus
by Brodkorb, 1967). This is the first Irvingtonian
record for the genus.

Order Columbiformes
Family Columbidae

Zenaida macroura (Linnaeus)

Material. — Carpometacarpus lacking metacarpal III and distal
end of metacarpal II

This specimen is indistinguishable from the car-
pometacarpus of Z. macroura, a living species that

is commonly recorded as a late Pleistocene fossil in
Florida and elsewhere (Brodkorb, 1971). The only
other record of Z. macroura for the Irvingtonian is
from the Coleman IIA local fauna, Florida (Ritchie,
1 980). A Blancan specimen reported as Z. macroura
(from Rexroad, Kansas; Wetmore, 1944) was later
described by Brodkorb (1969) as a new species, Z.
prior. Confirmation of the validity and relationships
of Z. prior must await the discovery of additional
specimens.

Order Passeriformes
Family Fringillidae

Icterinae, genus and species indeterminate

Material. — Distal half of ulna.

This fossil is the size of certain species of Icterus
and Molothrus, but otherwise lacks diagnostic fea-
tures. Other pre-Rancholabrean records of icterines
are those of Euphagus cyanocephalus from the Blan-
can age Sanders local fauna, Kansas (Harrell, 1959;
listed as Big Springs Ranch local fauna by Brodkorb,
1978), and Agelaius phoeniceus and Pandanaris
floridanus from the Coleman IIA local fauna, Flor-
ida (Ritchie, 1980).

DISCUSSION

The small avifauna from Payne Creek provides
the first Irvingtonian records for Tachybaptus do-
minicus and Limnodromus sp. In addition, it has
helped to define systematic problems in Neogene
grebes and ducks, although more fossils are needed
in both cases to resolve the issues. Grebes, ducks,
and semi-aquatic birds (geese, rails, shorebirds)
dominate the Payne Creek avifauna. In fact, each
of the taxa in the Payne Creek avifauna (Table 1)
may be found today at or near some sort of fresh
water body, including even the terrestrial taxa such

as Colinus virginianus, Zenaida macroura, and the
indeterminate icterine. The Payne Creek region was
composed of a mixture of fresh water and terrestrial
habitats at the time of fossil deposition. Thus the
Irvingtonian environment at Payne Creek was very
different from that reflected by the underlying Bone
Valley Formation, where the predominance of cor-
morants, alcids, loons, and sulids reflects a marine
setting, although near-shore because of the lesser
numbers of fresh water or terrestrial birds (Brod-
korb, 1955; 1972).

ACKNOWLEDGMENTS

The fossils were borrowed through Michael K. Frazier and
John S. Waldrop. Storrs L. Olson of the National Museum of
Natural History, Smithsonian Institution, and Mary C. Me-
Kitrick and G. Scott Mills of the University of Arizona allowed
access to skeletal collections under their supervision. Michael K.

Frazier supplied information on geology, chronology, and fossil
mammals from the sites. Storrs L. Olson and Michael K. Frazier
provided beneficial comments on the manuscript and also ex-
amined the fossil of Bucephala. The photographs are by Victor
E. Krantz.

LITERATURE CITED

Alvarez, R., and S. L. Olson. 1978. A new merganser from American Ornithologists’ Union. 1957. Check-list of North
the Miocene of Virginia (Aves: Anatidae). Proc. Biol. Soc. American Birds. A.O.U., Baltimore, 5th ed., 691 pp.

Washington, 91: 522-532.

52

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Brodkorb, P. 1955. The avifauna of the Bone Valley For-
mation. Rep. Invest., Florida Geol. Surv., 14:1-57.

. 1958. Fossil birds from Idaho. Wilson Bull., 79:237—

242.

. 1961. Birds from the Pliocene of Juntura, Oregon. Quart.

J. Florida Acad. Sci., 24:169-184.

. 1963a. Catalogue of fossil birds. Part 1 (Archaeoptergy-

iformes through Ardeiformes). Bull. Florida St. Mus., Biol.
Sci., 7:179-293.

. 1963 A A new Pleistocene grebe from Florida. Quart.

J. Florida Acad. Sci., 26:53-55.

. 1 964. Catalogue of fossil birds. Part 2 (Anseriformes

through Galliformes). Bull. Florida St. Mus., Biol. Sci., 8:
195-335.

. 1967. Catalogue of fossil birds. Part 3 (Ralliformes,

Ichthyomithiformes, Charadriiformes). Bull. Florida St. Mus.,
Biol. Sci., 11:99-220.

. 1969. An ancestral mourning dove from Rexroad, Kan-
sas. Quart. J. Florida Acad. Sci., 31:173-176.

. 1971. Catalogue of fossil birds. Part 4 (Columbiformes

through Piciformes). Bull. Florida St. Mus., Biol. Sci., 15:
163-266.

. 1972. New discoveries of pliocene [sic] birds in Florida.

Proc. XVth Inter. Omith. Cong., Abstract, p. 634.

. 1978. Catalogue of fossil birds. Part 5 (Passeriformes).

Bull. Florida St. Mus., Biol. Sci., 23:139-228.

Eyster, M. B. 1978. Least Grebe in Louisiana. Louisiana Or-
nith. Soc. News, 81:3.

Harrell, B. E. 1959. Notes on fossil birds from the Pleistocene
of Kansas and Oklahoma. Proc. South Dakota Acad. Sci.,
38:103-106.

Holman, J. A. 1959. Birds and mammals from the Pleistocene
ofWilliston, Florida. Bull. Florida St. Mus., Biol. Sci., 5:1-
24.

. 1961. Osteology of living and fossil New World quails

(Aves, Galliformes). Bull. Florida St. Mus., Biol. Sci., 6:131—
233.

Howard, H. 1963. Fossil birds from the Anza-Borrego Desert.

Contrib. Sci., Los Angeles Co. Mus., 73:1-33.

Ligon, J. D. 1 965. A Pleistocene avifauna from Haile, Florida.

Bull. Florida St. Mus., Biol. Sci., 10:127-158.

Lundelius, E. L., Jr., R. W. Graham, E. Anderson, J. E.
Guilday, J. A. Holman, D. W. Steadman, and S. D. Webb.

1983. Terrestrial vertebrate faunas. Pp. 31 1-353, in The
late Pleistocene (S. C. Porter, ed.), vol. 1 of. Late Quaternary
environments of the United States (H. E. Wright, Jr., ed.),
Univ. Minnesota Press, Minneapolis.

Murray, B. G., Jr. 1967. Grebes from the late Pliocene of
North America. Condor, 69:277-288.

Olson, S. L. 1974. The Pleistocene rails of North America.
Condor, 76:169-175.

Pregill, G. K., and S. L. Olson. 1981. Zoogeography of West
Indian vertebrates in relation to Pleistocene climatic cycles.
Ann. Rev. Ecol. Syst., 12:75-98.

Ritchie, T. L. 1980. Two mid-Pleistocene avifaunas from Cole-
man, Florida. Bull. Florida St. Mus., Biol. Sci., 26:1-36.

Steadman, D. W. 1980. A review of the osteology and pa-
leontology of turkeys (Aves: Meleagridinae). Nat. Hist. Mus.
Los Angeles Co., Contrib. Sci., 330:131-207.

Storer, R. W. 1976. The Pleistocene Pied-billed Grebes (Aves:
Podicipedidae). Pp. 147-153, in Collected papers in avian
paleontology honoring the 90th birthday of Alexander Wet-
more (S. L. Olson, ed.), Smithsonian Contrib. Paleobiol., 27:
xxvi + 1-211.

. 1979. Order Podicipediformes. Pp. 149-155 in Check-
list of Birds of the World, 2nd ed. (E. Mayr and G. W. Cot-
trell), Mus. Comp. Zook, Cambridge, Massachusetts, 1:1-
547.

Webb, S. D. (ed.). 1974. Pleistocene mammals of Florida. Univ.
Presses Florida, Gainesville, 270 pp.

Wetmore, A. 1937. A record of the fossil grebe, Colymbus
parvus, from the Pliocene of California, with remarks on
other American fossils of this family. Proc. California Acad.
Sci., ser. 4, 23:195-201.

. 1944. Remains of birds from the Rexroad fauna of the

upper Pliocene of Kansas. Univ. Kansas Sci. Bull., 30:89-
105.

. 1956. A check-list of the fossil and prehistoric birds of

North America and the West Indies. Smithsonian Misc. Coll.,
1 3 1(5): 1—105.

Winge, O. 1888. Fugle fra Knoglehuler i Brasilien. E Museo
Lundii, 1 (art. 2): 1-54.

Woolfenden, G. E. 1959. A Pleistocene avifauna from Rock
Spring, Florida. Wilson Bulk, 71:183-187.

. 1961. Postcranial osteology of the waterfowl. Bulk Flor-
ida St. Mus., Biol. Sci., 6:1-129.

Address: Department of Vertebrate Zoology, National Museum of Natural History, Smithsonian Institution,
Washington, D.C. 20560.

AN EVALUATION OF THE FOSSIL CURLEW PALNUMENIUS V1CTIMA

L. MILLER (AVES: SCOLOPACIDAE)

Storrs L. Olson

ABSTRACT

The holotypical tarsometatarsus of Palnumenius victima, from which it differs in only a few details. Numenius victima is ten-
the late Pleistocene (Rancholabrean) of Nuevo Leon, Mexico, is tatively retained as a problematic taxon that may represent a

not generically separable from the extant genus Numenius and temporal or geographic form of N. americanus.

falls within the lower size range of Numenius americanus, from

INTRODUCTION

With re-examination using better comparative
material, many supposedly extinct taxa of North
American Pleistocene birds have been shown to be
synonymous with living forms, some of which, how-
ever, have retreated from North America into the
tropics (for example, Olson, 1974). The process of
re-evaluating nominal fossil taxa is still important
to an accurate assessment of the effects of the Pleis-
tocene on North American birds. In this connection
I have restudied the supposedly extinct curlew Pal-
numenius victima L. Miller ( 1 942), from late Pleis-
tocene (late Rancholabrean) deposits in San Josecito
Cave, Nuevo Leon, Mexico.

Palnumenius victima was founded solely on a
complete left tarsometatarsus (LACM (CIT) 2944),
and was diagnosed as follows (Miller, 1942:45):
“Length about four-fifths that of Numenius amer-
icana [sic]; shaft almost uniform in transverse di-
ameter throughout; outer cotyla almost the same
level as the inner; inner trochlea less elevated.” Nei-
ther this diagnosis nor the description that followed
were organized in a manner that permits one to
distinguish generic from specific characters.

RESULTS AND DISCUSSION

Miller quotes a communication from A. Wetmore
in which the latter expressed the opinion that Pal-
numenius combined characters of Numenius with
those of godwits, Limosa. Nevertheless, Miller
clearly considered Palnumenius to be closer to Nu-
menius and compared it in particular with the Long-
billed Curlew, Numenius americanus, an extant
species that occurs in Mexico in winter. My ex-
amination of the holotype of P. victima disclosed
no characters linking it with Limosa, in which, for
example, there is no closed medial hypotarsal canal
(clearly shown in Miller’s illustration of P. victima )
and in which the inner trochlea is not as medially
flared. I found no characters that will permit Pal-
numenius victima to be separated generically from
Numenius. The genus Palnumenius L. Miller 1942
therefore becomes a junior subjective synonym of
Numenius Brisson 1760. It thus remains to be de-
termined whether Numenius victima can in fact be
separated from the extant species of Numenius.

The holotype of N. victima is much larger than
in the smallest of the curlews, N. borealis and N.

minutus, and decidedly larger and more slender than
the tarsometatarsus in Numenius phaeopus or N.
tahitiensis. It is smaller than in TV. arquata or TV.
madagascariensis. The tarsometatarsus in a single
skeleton of the rare Old World species TV. tenuiros-
tris (AMNH 547) was slightly smaller than the ho-
lotype of TV. victima, with the most internal ridge of
the hypotarsus being shorter than in either TV. vic-
tima or TV. americanus.

In stating that TV. victima was smaller than TV.
americanus, Miller (1942) clearly did not have ad-
equate comparative material. In curlews, as in many
shorebirds, males are smaller than females. There
was once a rather heated debate (Oberholser, 1918;
Grinnell, 1921) about whether TV. americanus can
be divided into a large southern subspecies and
smaller northern one. Although there is overlap be-
tween the two forms, the name TV. a. parvus contin-
ues to be applied to the northern populations by
some authors (Allen, 1980:7).

The tarsometatarsus in the smallest (USNM
499444, male) of seven skeletons of TV. americanus

53

54

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

that I examined measures 72.9 mm, whereas the
holotype of N. victima measures 72 mm. Ridgway
(1919) records the tarsal length in skins of males of
N. americanus parvus as ranging from 69.8 to 81.5
mm. In the Smithsonian collections I found skins
of two males in which the tarsal length was 71.5 and
72 mm, respectively, when measured along the an-
terior face from the intercotylar knob to the distal
margin of the middle trochlea, as one would mea-
sure a skeletal specimen. Thus, the holotype of N.
victima falls within the lower size range of males of
N. americanus.

As for the qualitative characters ascribed to N.
victima, the supposed differences in the relative levels
of the cotylae and of the inner trochlea were not
apparent to me. The uniform transverse width of
the shaft would also seem to occur in N. americanus
(Fig. 1) and is affected by age in any case, as in
juveniles the proximal end is wider than in adults.
Perhaps by this. Miller (1942) was attempting to
describe the fact that the internal cotyla in N. victima
projects more abruptly from the shaft than in N.
americanus, which is the case. Another apparently
valid character of N. victima mentioned by Miller
in his description, but not in the diagnosis, is the
larger, more rounded and more distally located dis-
tal foramen, with a deeper extensor groove. The
significance of these two relatively minor points can-
not be evaluated without additional fossil and mod-
ern specimens.

The closest relative of Numenius victima is un-
doubtedly N. americanus, a species known to breed
in central and western North America from south-
ern Canada south to Utah, New Mexico, and Texas,
and east to Michigan, Illinois, Iowa, and Kansas,
although it is now absent as a breeding bird from
the eastern parts of its range. The species is migra-
tory and winters mainly from California to Texas
and south to Oaxaca, Mexico, and Guatemala. In
the breeding season it inhabits open grasslands, pas-
tures, and shrub steppe. Because San Josecito Cave
is presently situated amidst pine and live oak forest
at 2,300 m above sea level (Kurten and Anderson,
1980), the presence of Numenius at this site, even
if a migrant, would indicate a more open environ-
ment in the late Pleistocene, as do certain of the
fossil mammals (Kurten and Anderson, 1980).

It cannot be ascertained whether Numenius vic-
tima was a migrant or a resident, nor is it possible
at this point to say whether it represents a temporal
or geographic variant of N. americanus, or a dis-
tinct, closely related species. Of the two new genera

Fig. I. — Left tarsometatarsi of Numenius : A) holotype of Pal-
numenius (=Numenius) victima, anterior view; B)same, posterior
view; C) small individual of N. americanus (USNM 499444,
male), posterior view. Natural size.

and species of waterbirds proposed by Miller ( 1 942)
from San Josecito (the rail Epirallus natator, and
Palnumenius victima) neither genus is valid but nei-
ther species can be dismissed unequivocally. Else-
where I have shown the genus Epirallus to be a
synonym of Rallus, with natator being a member
of the Rallus longirostris/elegans complex (Olson,
1974) that is larger than any of the modern members
of that complex. As with N. victima, it is uncertain
whether its differences are of a specific or subspecific
nature.

Thus, of the more than 42 species of birds that
Miller ( 1 944) ultimately reported from San Josecito,
the only certainly extinct species other than raptorial
birds and scavengers are the turkey Meleagris cras-
sipes (see Rea, 1980; Steadman, 1980, for docu-
mentation of the validity of this species) and the
large roadrunner Geococcyx conklingi (the specific
validity of which has recently been questioned, how-
ever [Harris and Crews, 1983]). This supports the
idea that most of the Pleistocene extinctions among
North American birds involved large species that
were for the most part dependent upon the mam-
malian megafauna (Lundelius et al., 1983; Stead-
man and Martin, in press).

1984

OLSON -FOSSIL CURLEW

55

ACKNOWLEDGMENTS

I thank Robert McKenzie and Armando Solis of the Natural
History Museum of Los Angeles County (LACM) for providing
a cast and photographs of the holotype of Palnumenius victima.
The photograph of Numenius americanus is by Victor E. Krantz,
National Museum of Natural History, Smithsonian Institution
(USNM). I am grateful to John Farrand, Jr., for facilitating the

loan of a skeleton of Numenius tenuirostris from the American
Museum of Natural History (AMNH) and for comments on an
early draft which was also read by Kenneth E. Campbell. David
W. Steadman criticized several subsequent versions of the manu-
script.

LITERATURE CITED

Allen, J. N. 1980. The ecology and behavior of the Long-billed
Curlew in southeastern Washington. Wildlife Monogr., 73:
1-67.

Grinnell, J. 1921. Concerning the status of the supposed two
races of the Long-billed Curlew. Condor, 23:21-27.

Harris, A. H., and C. R. Crews. 1983. Conkling’s Roadrun-
ner— a subspecies of California Roadrunner? Southwestern
Nat., 28:407-412.

Kurten, B., and E. Anderson. 1980. Pleistocene mammals of
North America. Columbia Univ. Press, New York, 442 pp.

Lundelius, E. L., Jr., R. W. Graham, E. Anderson, J. E.
Guilday, J. A. Holman, D. W. Steadman, and S. D. Webb.
1983. Terrestrial vertebrate faunas. Pp. 31 1-353, in The
Late Pleistocene (S. C. Porter, ed.), vol. 1 of, Late Quaternary
environments of the United States (H. E. Wright, Jr., ed.),
Univ. Minnesota Press, Minneapolis.

Miller, L. 1942. Two new bird genera from the Pleistocene of
Mexico. Univ. California Publ. Zook, 47:43-46.

. 1944. The Pleistocene birds of San Josecito cavern,

Mexico. Univ. California Publ. Zook, 47:143-168.

Oberholser, H. C. 1918. Notes on the subspecies of Numenius
americanus Bechstein. Auk, 35:188-195.

Olson, S. L. 1974. The Pleistocene rails of North America.
Condor, 76:169-175.

Rea, A. 1980. Late Pleistocene and Holocene turkeys in the
Southwest. In Papers in avian paleontology honoring Hilde-
garde Howard (K. E. Campbell, Jr., ed.), Contr. Sck, Nat.
Hist. Mus. Los Angeles Co., 330:209-224.

Ridgway, R. 1919. Birds of North and Middle America. Part
8. Bulk U.S. Nat. Mus., 50:xvi + 1-852.

Steadman, D. W. 1980. A review of the osteology and pa-
leontology of turkeys (Aves: Meleagridinae). In Papers in
avian paleontology honoring Hildegarde Howard (K. E.
Campbell, Jr., ed.), Contr. Sci., Nat. Hist. Mus. Los Angeles
Co., 330:131-207.

Steadman, D. W., and P. S. Martin. In press. Extinction of
birds in the late Pleistocene ofNorth America. In Quaternary
extinctions (P. S. Martin and R. G. Klein, eds.), Univ. Ar-
izona Press, Tucson.

Address: Department of Vertebrate Zoology, National Museum of Natural History, Smithsonian Institution,
Washington, D.C. 20560.

PHYLOGENY AND PALEOBIOGEOGRAPHY OF SHORT-TAILED

SHREWS (GENUS BLARINA )

Cheri A. Jones, Jerry R. Choate, and Hugh H. Genoways

ABSTRACT

Dental measurements of Pleistocene and Holocene shrews of
the genus Blarina were analyzed in a multivariate assessment of
phylogeny and paleobiogeography of the genus. Blarina brevi-
cauda, or an indistinguishable precursor, apparently was the an-
cestral form. Two semispecies, brevicauda and talpoides. are rec-
ognizable early in the fossil record; these two phena probably
appeared as a result of increasing diversity' in the environment
of the late Irvingtonian. Evidently, repeated fluctuations of cli-
mate and the subsequent changes in range have not been sufficient
to cause the cessation of gene flow necessary to complete specia-
tion.

A second species, B. carolinensis, appeared in the mid-Irving-
tonian. Although its original distribution is unclear, the species

apparently evolved in temperate conditions in the south. In-
creased continentality following the Wisconsinan glaciation re-
sulted in the appearance of the species B. hylophaga, which is
restricted to the southwestern part of the range of the genus.

Sympatry of the two semispecies of B. brevicauda and/or sym-
patry of B. carolinensis and B. hylophaga detected in some of
the paleofaunas are thought to be the result of periods of more
equable climate. Post-Wisconsinan fluctuations of climate are
the probable cause of the eastward restriction of the genus and
the zones of sympatry observed between Holocene phena.

Fossils of Blarina have been found in warm-moist and cool-
moist faunas. The fossil record (late Blancan-Rancholabrean)
and paleoecology of the genus are summarized.

INTRODUCTION

Holocene short-tailed shrews of the genus Blarina
occur throughout southeastern Canada and the east-
ern half of the United States (Hall, 1981). In Canada
they range westward from the Gaspe Peninsula and
Nova Scotia to southeastern Saskatchewan. The
northernmost records are from south-central Sas-
katchewan (Nero, 1960) and western Manitoba
(Krivda, 1 957). In the United States the distribution
of the genus extends from the Atlantic coast west-
ward to eastern Texas, central Oklahoma, north-
eastern Colorado, central Nebraska, South Dakota,
and North Dakota. An isolated population occurs
in Aransas County in southern Texas (Jones and
Loomis, 1954; Hall, 1981; George et al., 1 982; Jones
et al., 1982).

Short-tailed shrews inhabit a variety of mesic hab-
itats. In Canada and the northern and eastern United
States they have been reported from bogs, swamps,
grasslands, and coniferous and deciduous woodland
by Sheldon ( 1 936), Blair ( 1 940, 1941), Smith ( 1959),
Buckner (1966), Choate (1972), Kirkland (1978),
Handley (1979), and others. Except in Aransas
County, their distribution in Texas is limited to the
pine-oak and pine forests in the northeastern part
of the state (Schmidly and Brown, 1979). In agri-
cultural areas of the Great Plains, short-tailed shrews
are confined largely to riparian or otherwise mesic
habitats and to grassy roadsides and fence rows in
which ground cover and litter are present (Geno-
ways and Choate, 1972; Choate and Fleharty, 1975).

In an analysis of six plant communities on a rem-
nant prairie in western Kansas, a significant pref-
erence was shown for the most mesic community
in which the deepest litter was found (Choate and
Fleharty, 1973). The shrews are less common or
absent in grasslands and forests that lack herbaceous
ground cover (Smith, 1959; Kirkland, 1978).

Although Blarina is ubiquitous over much of its
range, local distribution apparently is limited by
moisture. Temperature and ground cover are of sec-
ondary importance, especially when vegetation is
sparse. Temperature and friability of the soil also
are important. Short-tailed shrews are not found in
areas where the moisture content of the soil is in-
sufficient to keep the air in the soil and litter satu-
rated (Pruitt, 1953, 1959; Getz, 1961).

The first taxonomic revision of the genus Blarina
was undertaken by Merriam (1895). He recognized
three species— B. brevicauda from Canada and the
northern, central, and eastern United States; B. car-
olinensis from the southeastern United States; and
B. telmaiestes, described on the basis of one spec-
imen from Dismal Swamp, Virginia. Among the
several Holocene subspecies of these species that
have been recognized by mammalogists are B. c.
peninsulae, which was described by Merriam (1895)
from peninsular Florida, and B. b. hylophaga , which
was described by Elliot (1899) from southern Okla-
homa. The latter was placed in synonymy with B.
b. carolinensis by Jones and Glass (1960), and all

56

1984

JONES ET AL. — B LARIN A SYSTEMATICS

57

nominal taxa of short-tailed shrews, with the ex-
ception of B. telmalestes, came to be regarded as
subspecies of B. brevicauda (Hall and Kelson, 1959).

In 1972 Genoways and Choate concluded, on the
basis of multivariate analyses, that the phena B. b.
brevicauda and what they termed B. b. carolinensis
[=i?. hylophaga ] behaved as biological species where
their ranges were contiguous in southern Nebraska.
Subsequent studies of Blarina in Illinois (Ellis et al.,
1978), Virginia (Tate et al., 1980), and Tennessee
(Braun and Kennedy, 1983) likewise revealed no
evidence of intergradation between the two phena.

Recently, karyologic data were used to clarify
taxonomic relationships among taxa of Blarina.
Genoways et al. (1977) compared karyotypes of B.
brevicauda from northern Nebraska and Pennsyl-
vania and what they termed B. b. carolinensis [ =B .
hylophaga] from southern Nebraska and Kansas.
Their data supported earlier morphometric analyses
(Genoways and Choate, 1972) that suggested the
two taxa in Nebraska were distinct species. Com-
parisons of the karyotypes of B. brevicauda with
those reported for B. brevicauda talpoides (by Mey-
lan, 1967) and B. b. kirtlandi (by Lee and Zimmer-
man, 1969) revealed that these northern taxa had
the same chromosomal numbers and Robertsonian
polymorphism. This chromosomal pattern is one of
three that subsequently were interpreted to repre-
sent separate species (George et al., 1982).

The three species are B. brevicauda in the north-
ern United States and eastern Canada, B. caroli-
nensis in the southeastern United States, and B.
hylophaga in the southwestern part of the range of
the genus. A fourth chromosomal pattern from pen-
insular Florida, for which morphometric data were
lacking, tentatively is referred to as B. c. peninsulae
but might prove to represent a fourth species (George
et al., 1981, 1982).

Shrews are poorly known in the fossil record be-
fore the Pleistocene. The Tribe Blarinini, which in-
cludes the genus Blarina. is first represented in the
late Miocene of North America by Adeloblarina, a
genus close to the ancestral stock of Blarina (Re-
penning, 1967). Late Pliocene (Blancan) specimens
with Blarina-\ike dentition originally were referred
to that genus— Blarina adamsi Hibbard (1953a)
from the Fox Canyon local fauna and Blarina gidleyi
Gazin (1933) from the Hagerman local fauna. These
two species subsequently were referred to other gen-
era, adamsi to Cryptot is (Repenning, 1967; Choate,
1970) and gidleyi to Paracryptotis (Choate, 1970;
Hibbard and Bjork, 1971); however, because of the

similarity and presumed close phyletic relationship
of those shrews with Blarina, the Fox Canyon and
Hagerman faunas are included among the faunas
summarized herein.

The fossil record of Blarina extends from the
Blancan through the Rancholabrean land mammal
ages. Sites from which specimens of Blarina have
been recovered are located throughout the eastern
half of the United States. Much of their accumu-
lation in caves, fissures, and sinkholes resulted from
the activities of owls and other predators. All fossil
localities, except those in Texas, Oklahoma, and
southwestern Kansas, are within the modern range
of the genus. Sites from which fossils of Blarina have
been reported, including early Holocene sites from
the United States and Canada, are listed in Appen-
dix 1.

Because the local distributions of extant species
of Blarina apparently are limited by moisture, re-
mains of Blarina in fossil faunas have been used as
paleoecological indicators. The first paleontologist
to do so was Hibbard (1956, 1963), who interpreted
the presence of Blarina in the Mt. Scott local fauna
as evidence of “trees and shrubs, with the devel-
opment of a humus and litter cover” (Hibbard,
1963). “Stratigraphically distinct samples of the
Short-tailed Shrew . . . represent different-sized an-
imals that would be ascribed to an evolving lineage
in conventional study. Similar differences can be
found in the large northern and small southern sub-
species of Blarina brevicauda, and associated faunas
support the interpretation that the observed mor-
phological differences are due to shifts of a dine”
(Hibbard et al., 1 965). The presence of different sizes
of Blarina in the same deposit also was interpreted
as an indication of changing climate by others, in-
cluding Guilday et al. (1964) and Oesch (1967).

More detailed analyses of the paleoecology of
Pleistocene Blarina were conducted by Graham and
Semken (1976) and Graham (1976). Graham and
Semken (1976) examined the modern distribution
of the taxa B. brevicauda, B. kirtlandi (regarded
herein as a subspecies of B. brevicauda ), and B. car-
olinensis and concluded that brevicauda is separated
from kirtlandi by moisture extremes, kirtlandi from
carolinensis by temperature extremes, and brevi-
cauda from carolinensis by temperature and mois-
ture extremes. Fossils that appeared to correspond
to the three extant phena seemed sympatric in three
Pleistocene deposits examined by Graham and
Semken (Cumberland Cave, New Paris No. 4, and
Peccary Cave), and two of the phena were thought

58

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

to be sympatric in the local faunas of Crankshaft
Cave, Meyer Cave, Natural Chimneys, and Welsh
Cave. Because of the co-occurrence of other cur-
rently allopatric species in the fossil deposits, sym-
patry of different phena of Blarina was interpreted
as an indication of more equable climate rather than
climatic changes as previously suggested by Guilday
et al. (1964), Oesch (1967), and others. Subsequent
allopatry or parapatry of the three phena of Blarina
and eastward contraction of the western boundary
of distribution since the Pleistocene were attributed
to increasing continentality of the climate.

Graham (1976) studied late Wisconsinan envi-
ronmental gradients by comparing faunal compo-
sition, densities of soricid and microtine species,
and species distributions of Wisconsinan and Ho-
locene faunas. In 11 of the 12 late Wisconsinan
faunas examined, more species of shrews were pres-
ent during the late Wisconsinan than at present. The
greater species density was interpreted as indicating
“more equable climate with reduced temperature
and moisture gradients.” As in the previous paper,
Graham thought the presence of more than one phe-
non of Blarina in a fossil fauna indicated sympatric

distributions under more equable climatic condi-
tions.

An additional nominal species, B. ozarkensis, was
recognized by Graham (1972) from the Kansan de-
posit from Conard Fissure in northern Arkansas.
This taxon originally was described by Brown ( 1 908)
as a subspecies of B. brevicauda on the basis of
fossils that were said to be intermediate in size be-
tween extant B. b. brevicauda and B. b. kirtlandi.
Additional distinguishing characteristics were given
by Graham (1972) and Graham and Semken (1976).
The taxonomic status of ozarkensis is reviewed
herein.

No attempt previously has been made to relate
new information on the taxonomy of Holocene Bla-
rina to fossils of this genus. Accordingly, the objec-
tives of the present study were 1) to summarize
available information regarding the fossil record of
Blarina and 2) to identify fossils of Blarina (most
of which previously were assigned to B. brevicauda)
in an effort to shed light on the phylogeny and pa-
leobiogeography of the species brevicauda, caroli-
nensis, and hylophaga.

MATERIALS AND METHODS

A total of 672 Holocene specimens was examined. These spec-
imens are housed in the following institutions: Carnegie Museum
of Natural History (CM); Museum of Zoology, Louisiana State
University (LSUMZ); Museum of Comparative Zoology, Har-
vard University (MCZ); Museum of the High Plains, Fort Hays
State University (MHP); National Museum of Natural Sciences,
Canada (NMC); Texas Cooperative Wildlife Collections, Texas
A & M University (TCWC); Museum of Natural History, Uni-
versity of Connecticut (UCONN); National Fish and Wildlife
Laboratory, National Museum of Natural History (USNM).

Twenty-six dental and dentary measurements, as listed by Gra-
ham and Semken (1976), were taken from each Holocene spec-
imen (dental terminology after Choate, 1969, 1970): 1) length of
P4-M3; 2) length of upper molar toothrow (M1-M3); 3) labial
length of P4; 4) lingual length of P4; 5) anterior width of P4; 6)
posterior width of P4; 7) labial length of Ml; 8) lingual length of
Ml; 9) anterior width of Ml; 10) posterior width of Ml; 11)
labial length of M2; 12) lingual length of M2; 13) anterior width
of M2; 14) posterior width of M2; 15) length of M3; 16) width
of M3; 17) length of ul -m3; 18) length of lower molar toothrow
(ml-m3); 19) length of ml; 20) length of trigonid of ml; 21)
length of m2; 22) length of trigonid of m2; 23) length of m3; 24)
length of trigonid of m3; 25) width of mandibular condyle; 26)
depth of body of mandible. Hereafter, these measurements are
referred to by the corresponding numbers. Measurements 1, 2,
17, and 18 were taken with dial calipers to the nearest 0.1 mm.
Other measurements ( 1 4 of the upper teeth and eight of the lower
teeth and dentary) were taken with an ocular micrometer at 10 x

magnification and rounded to the nearest 0.1 mm. Age classes
were assigned according to condition of pelage and degree of wear
to teeth (Choate, 1972). Most of the taxa mentioned herein were
illustrated by Merriam (1895), Hibbard (1957), Repenning ( 1967),
and Graham and Semken (1976). Therefore, illustrations are not
provided herein.

Holocene specimens were assigned to 24 reference samples
(Fig. 1) representing all but one ( B . brevicauda hooperi Bole and
Moulthrop) of the 16 nominal subspecies of the three species of
Blarina currently recognized (Hall, 1981; George et al., 1981,
1982): 1 ) Autauga, Cullman, Hale, Russell, and Sumter counties,
Alabama (B. carolinensis carolinensis), 16 specimens; 2) Brevard,
Collier, Dade, Highlands, Osceola, Palm Beach, and Polk coun-
ties, Florida (B. carolinensis peninsulae), 28 specimens; 3) Craw-
ford, Early, Grady, Liberty, Randolph, Richmond, Thomas, and
Tift counties, Georgia (B. carolinensis carolinensis), 12 speci-
mens; 4) Allen, Newton, and Porter counties, Indiana ( B . brev-
icauda kirtlandi), 26 specimens; 5) Hamilton, Henry, Johnson,
Marion, and Pottawattamie counties, Iowa (B brevicauda brev-
icauda), 19 specimens; 6) Ashtabula, Athens, Erie, Fairfield,
Geauga, Hamilton, and Meigs counties, Ohio (B. brevicauda kirt-
landi), 36 specimens; 7) Dukes County, Massachusetts (B. brev-
icauda aloga), 82 specimens; 8) Nantucket County, Massachu-
setts ( B . brevicauda compacta), 79 specimens; 9) New Brunswick
and Nova Scotia, Canada; Aroostook, Penobscot, and Piscata-
quis counties, Maine ( B . brevicauda pallida), 59 specimens; 10)
Wake County, North Carolina ( B . carolinensis carolinensis), 24
specimens; 11) Great Smoky Mountains and Magnetic City

1984

JONES ET AL. —BLARINA SYSTEMATICS

59

Fig. 1. — Present geographic distribution of Blarina brevicauda (B), B. carolinensis (C), and B. hylophaga (H) (after George et al., 1982).
Reference samples are described in text.

(counties uncertain}. Buncombe, Cherokee, Haywood, Macon,
Mitchell, Transylvania, Watauga, and Yancey counties. North
Carolina; Greenville and Oconee counties, South Carolina;
Campbell, Carter, Cocke, Johnson, Sevier, and Sullivan counties,
Tennessee; and Grayson County, Virginia ( B . brevicauda chur-
chi), 79 specimens; 12) Antelope, Buffalo, Cherry, Platte, and
Washington counties, Nebraska (B. brevicauda brevicauda ), 10
specimens; 13) Warren County, New York (B. brevicauda tal-
poides), 39 specimens; 14) Lee County, Florida { B . carolinensis
shermani), 3 specimens; 15) Darlington, Dorchester, George-
town, Richland, and Williamsburg counties. South Carolina (B.

carolinensis carolinensis), 9 specimens; 1 6) Reelfoot Lake (county
uncertain), Benton, and Fayette counties, Tennessee (B. caroli-
nensis carolinensis), 5 specimens; 1 7) Ellis, Rooks, Russell, and
Trego counties, Kansas (B. hylophaga hylophaga), 44 specimens;
18) Manitoba, Canada (B. brevicauda manitobensis ), 5 speci-
mens; 19) Ontario, Canada (B. brevicauda talpoides), 8 speci-
mens; 20) New Brunswick and Quebec, Canada (B. brevicauda
angusta), 1 1 specimens; 21) Aransas County, Texas (B. hyloph-
aga plumbed), 7 specimens; 22) Cleveland, Garvin, and Murray
counties, Oklahoma (B. hylophaga hylophaga ), 7 specimens; 23)
East Baton Rouge and Red River parishes, Louisiana ( B. caro-

60

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 2. — Sites from which fossils of Blarina were examined for this study. Localities are numbered as in Appendix 1.1. Java; 2. Bartek
Brothers Quarry; 3. Angus; 4. White Rock; 5. Wathena; 6. Rezabek; 7. Kanopolis; 8. Williams; 9. Kentuck; 10. Cudahy; 1 1. Jinglebob;
12. Mt. Scott; 13. Robert; 14. Dixon; 15. Berends; 16. Doby Springs; 17. Howard Ranch; 18. Easley Ranch; 19. Vera; 20. Miller’s Cave;
21. Longhorn Cavern; 22. Felton Cave; 23. Hall’s Cave; 24. Klein Cave; 25. Schulze Cave; 26. Friesenhahn Cave; 27. Cave Without
a Name; 28. Barton Springs; 29. Cherokee Sewer Site; 30. Willard Cave; 31. Schmitt Cave; 32. Mud Creek; 33. Craigmile; 34. Garrett
Farm; 35. Pleasant Ridge; 36. Waubonsie; 37. Thurman; 38, 39. Brynjulfson Cave No. 1 and No. 2; 40. Boney Spring; 41. Trolinger
Spring; 42. Crankshaft Cave; 43. Bat Cave; 44. Zoo Cave; 45. Conard Fissure; 46. Peccary Cave; 47. Moscow Fissure; 48. Meyer Cave;
49. Catalpa Creek; 50. Welsh Cave; 51. Robinson Cave; 52. Baker Bluff Cave; 53. Guy Wilson Cave; 54. Riverside Cave; 55. New
Paris Sinkhole No. 4; 56. Hanover Quarry Fissure; 57. Bootlegger Sink; 58. Port Kennedy Cave; 59. Cumberland Cave; 60. Eagle Cave;
61. Hoffman School Cave; 62. Trout Cave; 63. Strait Canyon Fissure; 64. Natural Chimneys; 65. Back Creek Cave No. 2; 66. Clark's
Cave; 67. Ripplemead Quarry; 68. Jasper Saltpeter Cave; 69. Meadowview Cave; 70. Ladds Quarry; 71. Haile XIB; 72. Arredondo;
73. Reddick IA; 74. Williston III A; 75. Waccasassa River; 76. Inglis IA; 77. Coleman IIA; 78. Bradenton 51st Street; 79. Vero; 80.
Caveme de Saint-Elzear.

1984

JONES ET AL ,-BLARINA SYSTEM ATICS

61

V2

Fig. 3. — Analysis of measurements 3-10 of upper teeth. Range
of variation of Holocene specimens of B. b. brevicauda (A), other
subspecies of B. brevicauda (B), B. hylophaga (C), B. c. peninsulae
(D), and B. carolinensis (E).

V2

Fig. 4. — Analysis of measurements 19-24 of lower teeth. Range
of variation of Holocene specimens of B. b. brevicauda (A), other
subspecies of B. brevicauda (B), B. hylophaga (C), B. c. peninsulae
(D), and B. carolinensis (E).

Imensis minima ), 29 specimens; 24) Quebec, Canada ( B . brevi-
cauda talpoides ), 8 specimens.

A total of 1538 specimens from 82 Pliocene, Pleistocene, and
early Holocene faunas was examined (Fig. 2). The fossils are
housed in the following collections: Carnegie Museum of Natural
History (CM); Vertebrate Paleontology Collection, Central Mis-
souri State University (CMSU); Mississippi Museum of Natural
Science (FACM); Illinois State Museum (ISM); Museum of Nat-
ural History, University of Kansas (KU); Midwestern University
(MWU); Department of Geology, University of Iowa (SUI); Ver-
tebrate Paleontology Laboratory, Balcones Research Center
(TMM); Florida State Museum, University of Florida (UF); Mu-
seum of Paleontology, University of Michigan (UMMP); De-
partment of Paleobiology, Smithsonian Institution (USNM). Sites
from which these specimens were recovered are listed in Appen-
dix 1 and illustrated in Fig. 2.

Because the fossils were fragmented or fragile, length mea-
surements of toothrows generally were not taken. As many as
possible of the other dental and dentary measurements used for
Holocene specimens were taken. Age classes were assigned ac-
cording to degree of wear to teeth.

Analyses of mensural data were performed with the Statistical
Analysis System, SAS (Helwig and Council, 1979). The Proce-
dure Univariate subroutine was used to calculate univariate sta-
tistics (mean, range, standard error, and coefficient of variation)
for each character within the reference samples. Age classes and
sexes were not run separately because previous studies revealed
inconsistent patterns of variation among ages and between sexes
(Guilday, 1957; Choate, 1972; Graham and Semken, 1976;
Schmidly and Brown, 1979; French, 1981; Moncriefet al., 1982;
Braun and Kennedy, 1983). The General Linear Model subrou-
tine (Procedure GLM) was used to test for significant differences
among reference samples for each character. Subsequently, a
Duncan’s Multiple Range Test (Procedure Duncan) was em-
ployed to determine maximally non-significant subsets among
samples. To assess multivariate relationships among samples,
the Manova option of the GLM subroutine was used to compute
characteristic roots and vectors. These values then were em-
ployed in canonical (discriminant) analyses of variation within

and among reference samples (methodology described by Yates
and Schmidly, 1977; NefT and Marcus, 1980).

Because specimens with missing measurements could not be
included in multivariate analyses, and because most fossils had
missing measurements, we selected suites of measurements from
the original 26 for which sizes of fossil samples were greatest.
The suites selected were measurements 3 through 10 for upper
teeth, and 19 through 22 or 24 for lower teeth. Samples of fossils
were analyzed using the one or more of these suites of measure-
ments for which adequate numbers of specimens were available.
Accordingly, it was necessary to compute characteristic roots and
vectors of the reference samples for these suites of measurements
for use in the canonical procedure to compare Holocene and fossil
specimens.

Canonical analyses of the appropriate suites of characters for
reference samples were used to plot centroids of variation within
and among those samples. Initially, centroids were plotted to
encompass one standard deviation on either side of the mean for
each canonical variate. However, it wasjudged more informative
to use a curved line enclosing peripheral specimens of each cluster
so to encompass all known variation within each reference sample
even though overlap of clusters resulted. In order to make com-
parisons with fossils, reference samples were pooled into five
clusters (Figs. 3 and 4). Cluster A represented the range of vari-
ation among specimens of the largest Holocene subspecies, B. b.
brevicauda (samples 5 and 1 2). Remaining samples of this species
were not as clearly distinguished. These smaller eastern subspe-
cies were grouped together as “small brevicauda" and are referred
to as the talpoides semispecies. This group (samples 4, 6, 7, 8,
9, 11, 13, 18, 19, 20, and 24) is represented by cluster B. Cluster
C represented the range of variation among the individuals of B
hylophaga (samples 17, 21, and 22). Specimens of the taxon
plumbea proved to be more similar to specimens of hylophaga
than to carolinensis, as previously suggested by Schmidly and
Brown (1979) and George et al. (1981). Because of the uncertain
taxonomic status of B carolinensis peninsulae (George et al.,
1982), the range of variation among specimens from sample 2
was represented by a separate cluster (D). Cluster E represented
the range of variation among the other three subspecies of B.

62

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

carolinensis (samples 1,3, 10, 14, 15, 16, and 23). The relatively
large size of the three specimens of B. c. shermani available for
examination implies that this taxon actually might be a relict
population of B. brevicauda. Because of this uncertainty, the three
specimens were not included in Figs. 3 and 4, nor in comparison
with fossils from outside Florida.

Subsequently, fossil samples were analyzed using the corre-
sponding characteristic roots and vectors for the Flolocene sam-
ples, and were plotted on the illustrations of canonical variation

within and among Flolocene samples. This procedure enabled
identification of fossil specimens, even when fossil samples con-
sisted of two or more species, by comparison with Holocene
samples. Of course, this procedure is accurate only if variation
within and among taxa has remained essentially uniform
throughout their existence; preliminary tests revealed that this
assumption is valid. After this procedure was used to sort species
in the samples of fossils, univariate analyses of fossil samples
were conducted as described for the reference samples.

RESULTS

Univariate statistics and results of Duncan’s mul-
tiple range tests for the reference samples of Holo-
cene Blarina are shown in Table 1. Measurements
3 through 10 of the upper teeth and 19 through 24
of the lower teeth were selected in order to maximize
sample sizes for groups of fossils, as discussed in
Materials and Methods. Results of Duncan’s mul-

Table 1 . — Univariate statistics for Holocene samples for mea-
surements 3-10 of the upper teeth and 19-24 of the lower teeth.
Vertical lines represent nonsignificant subsets as indicated by
Duncan's multiple range tests. F value for each character is in-
dicated by parentheses. Species represented by each sample are
identified as B. brevicauda (B for brevicauda semispecies. T for
talpoides semispecies), B. carolinensis (C). and B. hylo-
phaga (II).

tiple range tests illustrate the geographic variability
and variability within populations that were pointed
out previously by George et al. ( 1 98 1 ) and Moncrief
et al. ( 1 982). The Duncan’s multiple range tests con-
sistently showed significant differences between
samples of B. b. brevicauda from Iowa and Nebraska
(samples 5 and 12) and samples of other taxa. Sam-
ples of eastern subspecies of B. brevicauda were
smaller, and differed significantly from samples of
B. hylophaga and B. carolinensis in only one mea-
surement (lingual length of P4); for the other 13
measurements, there was some degree of overlap
with samples of B. hylophaga and B. carolinensis.

Table 1 . — Continued.

Sam-

ID pie

Mean

2 SE

Range

CV

Sam-
1D pie

5

12

24

18

13
1 1

9

4

6

19

20
8

17

7

14
22

I

10

21

15

16

23

3

9

2.74

0.04

2

6-2.9

3.46

2.70

0.07

2

5-2.9

4.28

2.59

0.07

2

5-2.7

3.48

2.58

0.08

2

5-2.7

3.24

2.48

0.03

2

3-2.7

3.68

2.47

0.02

2

3-2.6

3.41

2.46

0.02

2

3-2.6

3.23

2.46

0.03

2

3-2.7

3.49

2.43

0.04

2

2-2.6

4.24

2.37

0.04

2

3-2.4

2.19

2.33

0.05

2

2-2.4

3.38

2.31

0.03

2

2-2.5

4.18

2.30

0.03

2

2-2.5

3.98

2.28

0.02

2

1-2.4

3.61

2.27

0.07

2

2-2.3

2.55

2.20

0.05

2

1-2.3

2.88

2.19

0.04

2

0-2.3

4.05

2.18

0.04

2

1-2.4

4.44

2.17

0.04

2

1-2.2

2.25

2.14

0.08

1

9-2.3

5.76

2.14

0.05

2

1-2.2

2.56

2.09

0.03

1

9-2.2

4.04

2.08

0.05

2

0-2.2

3.61

2.08

0.04

1

9-2.3

5.34

18

B

5

10

B

12

7

T

18

5

T

24

38

T

20

78

T

13

56

T

1 1

26

T

19

34

T

8

7

C

14

1 1

T

4

52

T

7

44

T

6

46

T

9

3

H

17

6

H

22

16

H

21

24

C

1

7

C

2

9

C

10

5

C

16

29

C

15

1 1

C

23

27

C

3

Mean

2 SE

Range

cv

N

r th of P4 (F =

68.64)

2. 12

0.04

2.0-2. 3

3.61

18

2.09

0.06

2. 0-2. 2

4.19

10

1.88

0.04

oo

xO

2.38

5

1.86

0.1 1

1.6-2. 1

8.14

7

1.86

0.06

1. 7-2.0

5.04

1 1

1.85

0.03

1. 6-2.0

5.44

38

1.85

0.02

1.6-2. 1

5.57

78

1.84

0.08

1.6-1. 9

2.74

7

1.84

0.02

1. 7-2.0

4.31

52

1.83

0.07

1.8-1. 9

3.15

3

1.82

0.04

1. 7-2.0

4.73

26

1.82

0.02

1. 7-2.0

3.96

46

1.79

0.03

1. 6-2.0

5.1 1

34

1.79

0.03

1. 6-2.0

5.23

56

1.71

0.03

1.6-1. 9

4.69

44

1.63

0.07

1. 5-1.7

5.00

6

1.61

0.03

1.6-1. 7

2.34

7

1.59

0.05

1.6-1. 9

6.27

16

1.58

0.03

1.4-1. 7

4.98

27

1.55

0.04

1.4-1. 9

6.67

24

1.54

0.05

1.5-1. 6

3.56

5

1.52

0.06

1.4-1. 6

5.48

9

1.49

0.02

1.4- 1.6

4.23

29

1.46

0.03

1.4-1. 5

3.59

1 1

1984

JONES ET AL .-BLARINA SYSTEMATICS

63

Table 1 . — Continued.

Table 1. — Continued.

Sam-

ID pie

Mean 2 SE Range CV N

Sam-
ID pie

Mean 2 SE Range CV N

B 5

B 12

T 18

T 24

T 13

T 1 1

T 9

C 14

T 6

T 4

T 19

T 8

T 20

T 7

H 17

C 15

C 2

C 3

C 16

C 1

H 21

C 10

H 22

C 23

B 12

B 5

T 18

T 24

T 4

T 13

T 1 1

T 9

T 6

T 7

T 8

T 20

T 19

H 17

C 14

C 1

H 22

H 21

C 10

C 3

C 16

C 23

C 15

C 2

Anterior width of P4 (F = 51.79)
1.61 0.03 1.5-1.

1.59 0.07 1.4-1.

1.42 0.04 1.4-1.

1.41 0.05 1.3-1.

1.41 0.03 1.3-1.

1.41 0.02 1.2-1.

1.38 0.02 1.3-1.

1.37 0.07 1.3-1.

1.36 0.02 1.2-1.

1.36 0.03 1.2-1.

1.31 0.07 1.2-1.

1.29 0.03 1.1-1.

1.25 0.03 1.2-1.

1.24 0.02 1.1-1.

1.24 0.02 1.1-1.

1.23 0.03 1.2-1.

1.23 0.03 1.1-1.

1.22 0.04 1.1-1.

1.20 0.00 1.2

1.20 0.03 1.1-1.

1.20 0.04 1.1-1.

1.20 0.03 1.1-1.

1.18 0.03 1.1-1

1.13 0.03 1.0-1.

Posterior width of P4 (F = 29.82)

2.49 0.05 2.4-2

2.46 0.05 2.3-2

2.40 0.09 2.3-2

2.30 0.11 2.1-2

2.26 0.05 2.0-2

2.25 0.03 2.0-2

2.21 0.03 2.0-2

2.20 0.03 2.0-2

2.18 0.05 1.9-2

2.18 0.03 2.0-2

2.18 0.03 1.9-2

2.17 0.04 2.1-2

2.17 0.04 2.1-2

2.17 0.03 2.0-2

2.17 0.13 2.1-2

2.09 0.06 1.8-2

2.07 0.10 1.9-2

2.06 0.04 2.0-2

2.03 0.05 1.8-2

2.02 0.07 1.9-2

1.98 0.08 1.9-2

1.97 0.03 1.8-2

1.91 0.08 1.7-2

1.91 0.04 1.7-2

7

3.52

18

B

5

7

6.92

10

B

12

5

3.15

5

T

18

5

4.88

7

T

4

7

5.59

38

T

13

6

5.81

78

T

1 1

6

4.87

56

T

9

4

4.23

3

T

19

5

5.06

34

T

24

5

5.53

26

C

14

4

7.12

7

T

6

5

8.07

52

T

8

3

4.19

1 1

T

20

5

6.06

46

T

7

3

4.31

44

C

1

3

4.05

9

H

17

4

5.49

27

C

10

3

4.95

1 1

C

3

0.00

5

H

22

3

4.30

16

C

1 6

3

4.81

7

H

21

4

6.29

24

C

15

2

3.45

6

C

23

3

6.28

29

C

2

6

2.96

10

B

12

6

4.36

18

B

5

5

4.17

5

T

18

5

6.15

7

T

24

5

5.63

26

T

13

4

4.66

38

T

4

6

6.14

78

T

1 1

4

5.17

56

T

9

6

6.34

34

T

19

4

5.06

46

T

6

3

4.25

52

C

14

3

2.98

1 1

T

8

2

2.25

7

T

20

3

3.79

44

H

17

3

5.33

3

T

7

2

5.77

16

H

21

2

5.86

6

H

22

1

2.60

7

C

1

3

5.98

24

C

3

2

5.35

1 1

C

10

1

4.23

5

C

15

2

4.52

29

C

16

1

6.1 1

9

C

23

1

5.49

27

C

2

Labia! length of Ml (F =
2.47 0.03

2.45 0.07

2.28 0.08
2.27 0.04

2.26 0.02
2.25 0.02

2.22 0.02
2.21 0.03

2.21 0.08
2.20 0.00
2.20 0.04

2.14 0.03

2.09 0.04

2.09 0.02

2.08 0.05

2.07 0.02

2.02 0.04

2.02 0.04

2.02 0.03

2.00 0.00
2.00 0.04

1.99 0.05

1.97 0.02

1.93 0.03

58.09)

2.4-2. 6

2.78

18

2. 2-2. 6

4.41

10

2. 2-2. 4

3.67

5

2. 1-2.5

4.79

26

2. 2-2. 4

2.84

38

2. 1-2.4

3.70

78

2. 0-2. 4

3.63

56

2.2-2. 3

1.60

7

2. 1-2.3

4.83

7

2.2

0.00

3

2. 0-2. 4

4.75

34

2. 0-2. 4

4.34

52

2. 0-2. 2

3.35

1 1

2. 0-2. 2

3.15

46

1. 9-2.2

4.38

16

2. 0-2. 2

3.45

44

1. 9-2.2

4.39

24

1. 9-2.1

2.99

1 1

2.0-2. 1

2.02

6

2.0

0.00

5

1. 9-2.1

2.89

7

1. 9-2.1

3.93

9

1 .9-2.1

2.88

29

1. 8-2.1

4.49

27

Lingual length of Ml (F = 50.59)
2.24 0.09 2. 1-2.6

2.17 0.03 2. 0-2. 2

2.12 0.04 2. 1-2.2

2.11 0.08 2.0-2. 2

2.08 0.02 2.0-2. 2

2.05 0.03 1. 9-2.2

2.04 0.02 1. 9-2.2

2.04 0.02 1. 9-2.2

2.03 0.05 1.9-2. 1

2.00 0.03 1. 8-2.2

2.00 0.00 2.2

1.99 0.02 1. 8-2.1

1.98 0.05 1. 8-2.1

1.96 0.02 1. 8-2.1

1.94 0.02 1. 8-2.0

1.93 0.06 1. 8-2.0

1.92 0.06 1. 8-2.0

1.91 0.04 1. 8-2.0

1.86 0.05 1. 7-2.0

1.85 0.03 1. 7-2.0

1.84 0.04 1. 8-1.9

1.84 0.05 1. 8-1.9

1.80 0.03 1. 6-2.0

1.76 0.03 1. 6-1.9

6.38

10

2.74

18

2.1 1

5

5.06

7

2.59

38

4.20

26

3.45

78

3.17

56

3.49

7

3.81

34

0.00

3

3.39

52

3.79

1 1

2.93

44

3.07

46

3.92

7

3.93

6

4.22

16

4.42

1 1

3.80

24

2.86

9

2.98

5

3.98

29

4.19

27

Measurements of the three samples of B. hylophaga
and seven samples of B. carol 'inensis overlapped for
all characters, although samples of B. caro/inensis
averaged smaller. The sample of B. c. shermani ( 1 4,

N = 3) from Lee County, Florida, was indistinguish-
able from eastern populations of B. brevicauda in
all characters.

Statistics for groups of fossils (where N > 4) are

64

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 1 . — Continued.

Table 1. — Continued.

Sam-

ID pie

B

B

T

T

T

T

H

T

T

T

T

T

T

C

T

H

H

C

C

C

C

C

C

C

B

B

T

T

T

T

T

T

T

T

T

T

T

H

C

II

C

C

C

H

C

C

C

C

12

5

18
24
1 1
20

17

6

4
8
9

13

19

14
7

22

21

23

16

1

3

15
2

10

12

5

18

24
1 1

4

13

7

8

6
9

20

19

17

14
22

!

3

10

21

23

16

15
2

Mean

2 SE

Range

cv

N

ID

Sam-

ple

th of Ml (F =

38. 18)

2.51

0.06

2.4-2. 7

3.49

10

B

5

2.45

0.05

2. 2-2. 6

4.16

18

B

12

2.30

0.09

2. 2-2. 4

4.35

5

T

18

2.24

0.10

2. 1-2.5

5.67

7

T

24

2.23

0.03

1. 9-2.6

5.92

78

T

13

2.20

0.05

2. 1-2.3

4.07

1 1

T

1 1

2.18

0.03

2.0-2. 6

4.56

44

T

4

2.18

0.04

2. 0-2. 4

5.61

34

T

19

2.17

0.05

2. 0-2. 5

5.34

26

T

6

2.16

0.04

1. 9-2.4

5.91

52

T

9

2.14

0.03

1. 9-2.4

4.61

56

T

8

2.14

0.03

2. 0-2. 4

4.73

38

T

20

2.1 1

0.07

2. 0-2. 2

4.41

7

C

14

2.10

0.12

2. 0-2. 2

4.76

3

H

17

2.06

0.02

1.9-2. 2

3.55

46

H

22

2.02

0.06

1.9-2. 1

3.73

6

T

7

2.00

0.04

1.9-2. 1

2.89

7

C

16

1.95

0.03

1.8-2. 1

4.21

29

C

1

1.94

0.05

1. 9-2.0

2.82

5

c

3

1.93

0.04

1.8-2. 1

4.52

16

H

21

1.93

0.07

1. 8-2.2

6.18

1 1

C

10

1.92

0.06

1. 8-2.1

4.34

9

C

15

1.92

0.04

1. 6-2.1

5.96

27

C

23

1.92

0.04

1. 8-2.2

5.08

24

C

2

1th of Ml (F =

41.50)

2.66

0.07

2. 5-2.9

4.41

10

B

5

2.64

0.07

2. 2-2. 8

5.55

18

B

12

2.52

0.08

2. 4-2. 6

3.32

5

T

18

2.50

0.1 1

2. 3-2. 7

5.66

7

T

24

2.46

0.03

2. 1-2.8

4.79

78

T

13

2.43

0.05

2.2-2. 7

5.02

26

T

19

2.42

0.03

2.2-2. 6

3.98

38

T

1 1

2.41

0.04

2. 1-2.6

5.32

46

T

4

2.41

0.03

2. 2-2. 6

3.89

52

T

6

2.40

0.04

2. 1-2.6

5.20

34

T

9

2.40

0.03

22-2.6

4.57

56

C

14

2.36

0.04

2.2-2A

2.92

1 1

T

8

2.34

0.06

2.2-2A

3.81

7

T

20

2.31

0.02

2. 1-2.4

3.03

44

H

17

2.30

0.12

2. 2-2. 4

4.35

3

T

7

2.22

0.06

2.1-2. 3

3.40

6

H

21

2.19

0.05

2. 0-2. 4

5.88

16

H

22

2.17

0.06

2. 0-2. 4

4.64

1 1

C

1

2.16

0.04

2. 0-2. 4

4.75

24

C

3

2.16

0.04

2. 1-2.2

2.48

7

C

10

2.15

0.03

2.0-2. 3

4.19

29

C

15

2.14

0.08

2. 0-2. 2

4.18

5

C

16

2.10

0.08

2.0-2. 3

5.32

9

C

2

2.06

0.04

1.8-2. 3

5.17

27

C

23

Mean

2 SE

Range

CV

N

f ml (F= 85.25)

2.48

0.04

2. 3-2. 6

3.62

19

2.46

0.05

2. 3-2. 6

3.47

9

2.32

0.04

2. 3-2. 4

1.93

5

2.29

0.06

2. 2-2. 4

3.65

8

2.27

0.02

2. 2-2.4

3.09

37

2.25

0.02

2. 0-2. 4

3.87

80

2.25

0.03

2. 1-2.4

3.83

24

2.23

0.03

2. 2-2. 3

2.08

7

2.22

0.03

2. 0-2. 4

3.81

33

2.19

0.01

2. 1-2.3

2.49

56

2.12

0.03

1. 9-2.3

4.92

49

2.1 1

0.04

2. 0-2. 2

2.69

10

2.10

0.00

2.1

0.00

3

2.07

0.02

2. 0-2. 2

2.91

43

2.03

0.04

2.0-2. 1

2.54

6

2.02

0.03

1. 7-2.2

5.39

39

2.00

0.00

2.0

0.00

4

1.99

0.03

1. 9-2.1

2.52

15

1.98

0.04

1.9-2. 1

3.52

10

1.97

0.04

1. 9-2.0

2.48

7

1.96

0.03

1. 8-2.2

4.01

24

1.95

0.05

1. 8-2.0

3.72

8

1.91

0.03

1. 8-2.0

3.62

28

1.86

0.03

1. 7-2.0

3.92

27

on id of ml (F =

= 56.08)

1.70

0.04

1.6-1. 9

5.55

19

1.66

0.05

1.6-1. 8

4.39

9

1.62

0.04

1.6-1. 7

2.76

5

1.60

0.05

1.5-1. 7

4.73

8

1.55

0.02

1.5-1. 6

3.26

37

1.54

0.04

1.5-1. 6

3.37

7

1.54

0.02

1.4-1. 9

5.32

80

1.53

0.03

1.4-1. 7

5.51

24

1.51

0.03

1.4-1. 7

5.01

33

1.51

0.02

1.4-1. 6

3.85

56

1.47

0.06

1.4-1. 5

3.94

3

1.44

0.02

1.2-1. 6

5.70

49

1.41

0.04

1.3-1. 5

4.03

10

1.40

0.01

1.3-1. 5

2.43

43

1.37

0.03

1.2-1. 6

6.63

39

1.37

0.04

1. 3-1.4

3.56

7

1.37

0.04

1. 3-1.4

3.78

6

1.37

0.03

1. 3-1.4

3.57

15

1.34

0.03

1. 3-1.4

3.88

10

1.33

0.03

1.2-1. 5

5.28

24

1.31

0.04

1 .2-1.4

5.04

8

1.30

0.00

1.3

0.00

4

1.29

0.02

1. 2-1.4

4.88

27

1.26

0.03

1. 1-1.4

6.43

28

shown in Table 2. Results of Duncan’s multiple range
tests show variability within and among paleofaunas
similar to or greater than that of Holocene popu-

lations. The overlapping measurements result in dif-
ficulty in identifying fossil taxa, particularly when
specimens are incomplete (thus limiting the number

1984

JONES ET AL .-BLARINA SYSTEM ATICS

65

Table 1. — Continued.

Table 1 . — Continued.

Sam-

ple

Range

CV

Sam-

ple

Range

B

B

T

T

T

T

T

T

T

T

C

T

T

T

H

H

C

C

C

H

C

C

C

C

B

B

T

T

T

T

T

T

T

C

T

T

T

T

H

C

H

H

C

C

C

C

C

C

12

5
24
18

13
1 1

4

6

19
9

14
8

20
7

17

22

1

16

10

21

3

15
23

2

12

5
24

13
1 1

4

6

19

9

14
18

8

7

20

17

1

21

22

15
3

16

10
2

23

Length of m2 (F = 48.46)

2.06 0.03 1. 9-2.2

2.03 0.03 1. 9-2.2

1.94 0.04 1. 9-2.0

1.92 0.04 1. 9-2.0

1.92 0.02 1. 8-2.0

1.88 0.02 1. 7-2.0

1.88 0.03 1.7-2. 1

1.86 0.03 1. 7-2.2

1.83 0.04 1. 8-1.9

1.83 0.02 1. 7-1.9

1.80 0.00 1.8

1.78 0.03 1. 5-2.0

1.77 0.04 1.7-1. 9

1.76 0.03 1. 5-2.0

1.75 0.02 1.7-1. 8

1.73 0.04 1.7-1. 8

1.73 0.03 1.6-1. 9

1.70 0.00 1.7

1.70 0.03 1.6-1. 9

1.67 0.06 1.6-1. 8

1.67 0.04 1.6-1. 8

1.66 0.03 1.6-1. 7

1.64 0.02 1.5-1. 7

1.61 0.02 1.5-1. 8

Length of trigonid of m2 (F =

1.36

1.34

1.30

1.28

1.25

1.22

1.22

1.21

1.20

1.20

1.20

1.18

1.18

1.18

1.16

1.16

1.16

1.15

1.14

1.13

1.13

1.12

1.08

1.07

0.07

0.03

0.05

0.03

0.02

0.03

0.03

0.04

0.01

0.00

0.06

0.02

0.02

0.03

0.02

0.03

0.04

0.05

0.04

0.05

0.04

0.02

0.02

0.02

24.70)
1.2-1. 6
1. 2-1.4
1. 2-1.4
1. 1-1.5

1. 1- 1.5

1.1- 1. 3

1. 1- 1.4

1. 1- 1.3
1.1-1. 3

1.2

1. 1-1.3
1.0-1. 3

1.0- 1. 3

1. 1- 1.2

1.1-1. 2

1. 1- 1.3

1.1- 1. 2
1.1-1. 2
1. 1-1.2
1.0-1. 2
1.1-1. 2
1.0-1. 2
1.0-1. 2
1.0-1. 2

4.74
3.23
2.67
2.33

3.14

3.82

4.43

5.43

2.82

3.15
0.00
5.58

3.81
6.22
2.88

2.98
3.96
0.00

3.75
4.52

3.87
3.00
3.80
3.90

8.63

4.52

5.82
6.26
6.47
5.39
6.22
4.45

2.44
0.00
5.89

5.87
6.19
3.57
4.18

5.15
4.62

4.76
4.61
7.12

3.99
5.31
5.27
5.07

9

19

8

5
37
80
24
33

7

56

3
49
10
39
43

6
15

4
24

7

10

8

28

27

9
19

8

37

80

24

33

7

56

3

5
49
39
10
43
15

7

6

8

10

4
24

27

28

B

B

T

T

T

T

T

T

T

T

H

T

T

T

C

H

H

C

C

C

C

C

C

C

B

B

T

T

T

T

T

T

T

C

T

H

H

T

T

C

C

C

H

C

T

C

C

c

5

12

18

24

4

13
1 1

6

19
9

17

7

20

8

14
22

21

3

16

10

1

15

23

2

12

5
24

6
1 1

13
4

19
18

14
9

17

22

8

20
3

15
10
21

1

7

2

16
23

Length of m3 (F = 42.42)

1.67

0.03

1.6-

1.8

3.49

19

1.67

0.07

1.5-

1.8

6.48

9

1.58

0.08

1.5-

1.7

5.30

5

1.55

0.05

1.4-

1.6

4.88

8

1.54

0.03

1.4-

1.7

4.57

24

1.54

0.02

1.3-

1.6

4.78

37

1.53

0.02

1.4-

1.7

4.70

80

1.51

0.02

1.4-

1.6

4.75

33

1.50

0.06

1.3-

1.6

6.01

7

1.47

0.02

1.2-

1.6

4.93

56

1.47

0.02

1.4-

1.6

4.40

43

1.44

0.03

1.2-

1.6

7.56

39

1.43

0.03

1.4-

1.5

3.38

10

1.41

0.03

1.2-

1.6

7.08

49

1.37

0.06

1.3-

1.4

4.23

3

1.37

0.04

1.3-

1.4

3.78

6

1.36

0.04

1.3-

1.4

3.94

7

1.33

0.04

1.2-

1.4

5.08

10

1.33

0.05

1.3-

1.4

3.77

4

1.32

0.03

1.2-

1.5

5.96

24

1.32

0.03

1.2-

1.4

4.97

15

1.31

0.03

1.3-

1.4

2.69

8

1.29

0.03

1.1-

1.4

5.13

28

1.27

0.03

1.0-

1.4

7.21

27

mid of m3 (F

= 20.40)

1.20

0.06

1.0-

1.3

8.78

9

1.14

0.04

1.0-

1.2

6.69

19

1.1 1

0.05

1.0-

1.2

5.76

8

1.08

0.02

1.0-

1.2

6.49

33

1.07

0.02

0.9-

1.2

6.75

80

1.07

0.03

1.0-

1.2

7.17

37

1.05

0.03

0.9-

1.2

6.84

24

1.04

0.04

1.0-

1.1

5.13

7

1.04

0.05

1.0-

1.1

5.27

5

1.03

0.06

1.0-

1.1

5.59

3

1.03

0.01

1.0-

1.1

4.48

56

1.01

0.01

0.9-

1.1

3.82

43

1.00

0.00

1.0

0.00

6

1.00

0.02

0.8-

1.2

6.93

49

0.99

0.02

0.9-

1.0

3.19

10

0.99

0.04

0.9-

1.1

5.73

10

0.99

0.03

0.9-

1.0

3.58

8

0.99

0.02

0.9-

1.1

4.97

24

0.99

0.05

0.9-

1.1

7.00

7

0.98

0.02

0.9-

1.0

4.23

15

0.97

0.03

0.8-

l.l

9.02

39

0.93

0.03

0.8-

IT

7.79

27

0.93

0.05

0.9-

1.0

5.41

4

0.89

0.03

0.8-

1.0

7.96

28

of characters that can be examined) or when sample
size is small. Whether this variability reflects the
true variation within a contemporaneous popula-

tion or an accident in preservation is difficult to
determine. We have tried to indicate (for example,
the Ladds Quarry fauna) where conditions of pres-

66

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 2.— Univariate statistics for samples of fossils (where N > 4) for measurements 3-10 of the upper teeth and 19-24 of the lower
teeth. Vertical lines represent nonsignificant subsets as indicated by Duncan’s multiple range tests. F value for each character is indicated
in parentheses. Asterisks indicate local faunas in which more than one phenon is present. Species are identified by ID as in Table I.
Samples are indicated by 6-character acronym (ARREDO — Arredondo, BACKCR = Back Creek Cave No. 2, BAKERB = Baker Bluff,
BRYNJU = Brynjulfson Cave No. 2, CAVEWI = Cave Without a Name, CHIMNE = Natural Chimneys, CLARKS = Clark’s Cave,
CONARD = Conard Fissure, CRANKS = Crankshaft Cave, CUMBER = Cumberland Cave, FRIESE = Friesenhahn Cave, GUYW1L
= Guy Wilson Cave, HAILEX = Haile XI B, INDIAN = Vero, JASPER = Jasper Saltpeter Cave, KLEINC = Klein Cave, LADDSQ =
Ladds Quarry, MEYERC = Meyer Cave, MOSCOW = Moscow Fissure, NEWPAR = New Paris No. 4, PECCAR = Peccary Cave,
REDDIC — Reddick IA, RIVERS = Riverside Cave, ROBINS = Robinson Cave, SCHMIT = Schmitt Cave, SCHULZ = Schulze Cave.
STELZE = Caverne de St.-Elzear, STRAIT = Strait Canyon Fissure, TROUTC = Trout Cave, WILLAR = Willard Cave, and ZOO-

CA V = Zoo Cave).

ID

Sample

*B

BACKCR

*B

NEWPAR

B

CLARKS

B

MOSCOW

*B

CRANKS

B

ROBINS

B

CHIMNE

*B

BAKERB

T

STELZE

*T

NEWPAR

T

CUMBER

T

RIVERS

*T

BAKERB

T

LADDSQ

T

CONARD

T

PECCAR

T

JASPER

T

WILLAR

T

BRYNJU

*T

MEYERC

H

SCHULZ

*T

BACKCR

*T

CRANKS

H or C

KLEINC

C

CAVEWI

*c

MEYERC

C

ARREDO

c

HAILEX

c

INDIAN

Mean 2 SE Range CV N

Labial length of P4 (F = 55.37)

2.80 0.12

2.80 0.05

2.79 0.09

2.79 0.07

2.75 0.04

2.70 0.08

2.70 0.11

2.61 0.10

2.61 0.04

2.60 0.18

2.57 0.08

2.55 0.06

2.53 0.05

2.50 0.09

2.45 0.06

2.43 0.07

2.40 0.16

2.39 0.05

2.39 0.05

2.37 0.05

2.29 0.06

2.25 0.30

2.25 0.10

2.20 0.09

2.10 0.06

2.10 0.03

1.88 0.05

1.85 0.03

1.84 0.10

2. 7-2.9

3.57

3

2.1-23

5.05

2

2. 6-3.0

4.89

9

2. 7-2. 9

3.23

7

2. 6-2. 9

3.44

20

2. 6-2. 8

3.02

4

2. 5-3.0

5.94

8

2.5-2. 8

5.15

7

2.3-2. 9

4.94

34

2.4-2. 8

7.02

4

2.4-2. 7

4.02

6

2.5-2. 6

2.26

4

2. 4-2. 6

3.25

10

2.4-2. 6

4.00

5

2.3-2. 7

4.33

15

2. 2-2. 8

6.13

20

2. 2-2. 6

6.80

4

2. 2-2. 7

4.97

27

2. 3-2. 5

2.68

8

2.2-2. 6

4.47

17

2. 1-2.4

4.34

10

2. 1-2.4

9.43

2

2. 2-2. 3

3.14

2

2. 1-2.3

4.55

5

2. 0-2. 2

3.89

7

2.0-2. 2

2.92

17

1.8-1. 9

2.67

4

1. 8-2.0

3.37

17

1. 7-2.0

6.20

5

ervation or collection imply the presence of allo-
patric taxa.

In general, fossils were separated by Duncan’s
multiple range tests as in the multivariate canonical
analyses discussed in Paleontology. Results of the
Duncan’s tests appear to support our suspicions that
more than one taxon was present in several samples

(BACKCR, BAKERB, CRANKS, CUMBER,
LADDSQ, MEYERC, NEWPAR, and ROBINS).
There was a high degree of overlap among almost
all samples; only the four groups from Florida (AR-
REDO, HAILEX, INDIAN, and REDDIC) were
significantly different from other samples for most
characters.

1984

JONES ET AL ,-BLARINA SYSTEMATICS

67

Table 2. — Continued.

ID

Sample

B

MOSCOW

*B

NEWPAR

*B

BACKCR

B

CLARKS

*B

CRANKS

B

ROBINS

B

CHIMNE

*T

BAKER B

*B

BAKERB

*T

BACKCR

*T

NEWPAR

T

STELZE

T

RIVERS

T

LADDSQ

T

PECCAR

T

CONARD

T

CUMBER

T

BRYNJU

T

WILLAR

T

JASPER

*x

MEYERC

H

SCHULZ

*T

CRANKS

H orC

KLEINC

C

CAVEWI

*C

MEYERC

C

HAILEX

C

INDIAN

C

ARREDO

B

MOSCOW

T

RIVERS

B

CLARKS

B

CHIMNE

*B

CRANKS

*B

BACKCR

*B

BAKERB

B

ROBINS

*T

BACKCR

*B

NEWPAR

T

LADDSQ

*T

BAKERB

*T

NEWPAR

T

CUMBER

T

WILLAR

T

CONARD

T

PECCAR

T

STELZE

H orC

KLEINC

T

BRYNJU

T

JASPER

H

SCHULZ

*y

MEYERC

*T

CRANKS

C

CAVEWI

*c

MEYERC

c

HAILEX

c

ARREDO

c

INDIAN

Mean 2 SE

Lingual length of P4 (F = 58.46)
2.17 0.08

2.15 0.10

2.13 0.07

2.12 0.08

2.10 0.03

2.08 0.05

2.03 0.03

2.01 0.05

1.99 0.10

1.95 0.06

1.95 0.10

1.94 0.03

1.93 0.10

1.92 0.12

1.92 0.04

1.91 0.05

1.90 0.07

1.89 0.05

1.88 0.04

1.85 0.19

1.81 0.03

1.72 0.06

1.65 0.10

1.62 0.08

1.60 0.00

1.57 0.03

1.42 0.03

1.38 0.04

1.38 0.10

Anterior width of P4 (F = 27.40)
1.70 0.11

1.63 0.13

1.61 0.05

1.56 0.07

1.54 0.04

1.53 0.13

1.51 0.11

1.50 0.08

1.50 0.20

1.50 0.20

1.48 0.08

1.48 0.07

1.48 0.10

1.47 0.07

1.45 0.03

1.45 0.06

1.44 0.04

1.43 0.03

1.38 0.04

1.38 0.03

1.35 0.13

1.33 0.09

1.30 0.02

1.20 0.00

1.19 0.05

1.16 0.03

1.15 0.03

1.13 0.10

1.06 0.05

Range

cv

N

2.0-2. 3

5.12

7

2. 1-2.2

3.29

2

2. 1-2.2

2.71

3

2. 0-2. 3

5.66

9

2.0-2. 2

3.46

20

<N

1

q

<N

2.41

4

2.0-2. 1

2.29

8

1. 9-2.2

3.67

10

1. 8-2.2

6.77

7

1. 9-2.0

2.96

2

1. 9-2.0

3.63

4

1.7-2. 1

4.40

34

1. 8-2.0

4.97

4

1. 7-2.0

6.79

5

1.8-2. 1

4.57

20

1. 8-2.0

4.79

15

1. 8-2.0

4.71

6

1. 8-2.0

4.42

8

1. 7-2.1

5.1 1

27

1. 7-2.1

10.35

4

1. 7-1.9

3.31

17

1.6-1. 8

5.34

10

1.6-1. 7

4.29

2

1.5-1. 7

5.17

5

1.6

0.00

7

1.5-1. 7

4.49

17

1.3-1. 5

3.73

17

1.3-1. 4

3.24

5

1.3-1. 5

6.96

4

1. 5-1.9

8.32

7

1.5-1. 8

7.74

4

1.5-1. 7

4.85

9

1.5-1. 7

5.86

8

1.4-1. 7

5.33

20

1 .4-1.6

7.53

3

1.4-1. 8

9.67

7

1 .4-1.6

5.44

4

1 .4-1.6

9.43

2

1.4-1. 6

9.43

2

1.4-1. 6

5.65

5

1.3-1. 6

6.98

10

1.4-1. 6

6.49

4

1.4-1. 6

5.57

6

1.3-1. 7

5.86

27

1. 3-1.6

7.34

15

1.3-1. 6

5.66

20

1.3-1. 6

6.10

34

1. 3-1.4

3.24

5

1. 3-1.4

3.37

8

1.2-1. 5

9.56

4

1. 1-1.5

10.06

10

1.2-1. 4

3.85

17

1.2

0.00

2

1.1-1. 3

5.82

7

1.1-1. 2

4.38

17

1.1-1. 3

5.44

17

1.0-1. 2

8.51

4

1.0-1. 1

5.17

5

68

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 2. — Continued.

ID

Sample

*B

NEWPAR

B

CLARKS

B

MOSCOW

*B

BACKCR

B

ROBINS

*B

CRANKS

B

CHIMNE

T

STELZE

NEWPAR

T

LADDSQ

*T

BACKCR

*B

BAKERB

T

CONARD

T

CUMBER

T

BRYNJU

T

PECCAR

*T

BAKERB

T

WILLAR

*T

MEYERC

T

RIVERS

T

JASPER

H

SCHULZ

*T

CRANKS

H or C

KLEINC

C

CAVEWI

*C

MEYERC

C

INDIAN

C

HAILEX

C

ARREDO

*B

NEWPAR

B

CLARKS

*B

BACKCR

B

ROBINS

B

CHIMNE

*B

CRANKS

B

MOSCOW

T

STELZE

*T

NEWPAR

T

CUMBER

*B

BAKERB

T

RIVERS

*T

BAKERB

T

CONARD

T

PECCAR

T

LADDSQ

*T

BACKCR

T

BRYNJU

T

WILLAR

*T

MEYERC

H

SCHULZ

T

JASPER

*T

CRANKS

H or C

KLEINC

C

CAVEWI

*C

MEYERC

C

ARREDO

C

HAILEX

C

INDIAN

Mean 2 SE Range CV N

Posterior width of P4 (F = 34.20)

2.70 0.20

2.67 0.08

2.64 0.10

2.63 0.07

2.55 0.06

2.50 0.04

2.48 0.03

2.43 0.05

2.43 0.15

2.40 0.06

2.40 0.20

2.40 0.04

2.34 0.06

2.33 0.10

2.33 0.05

2.33 0.07

2.30 0.08

2.27 0.06

2.24 0.06

2.23 0.10

2.15 0.19

2.14 0.04

2.10 0.20

2.08 0.15

2.00 0.00

1.95 0.05

1.82 0.10

1.79 0.06

1.78 0.10

Labial length of Ml (F = 37.21)
2.55 0.10

2.51 0.13

2.43 0.18

2.43 0.10

2.39 0.07

2.38 0.05

2.37 0.08

2.34 0.04

2.33 0.10

2.32 0.06

2.31 0.09

2.30 0.08

2.27 0.07

2.23 0.04

2.23 0.04

2.22 0.12

2.20 0.20

2.19 0.05

2.18 0.04

2.15 0.06

2.14 0.04

2.13 0.15

2.10 0.20

2.06 0.05

2.03 0.04

1.99 0.04

1.73 0.05

1.71 0.07

1.70 0.06

2. 6-2. 8

5.24

2

2. 5-2.9

2.67

9

2. 5-2.9

4.82

7

2.6-2.1

2.19

3

2. 5-2.6

2.26

4

2. 3-2.6

3.56

20

2.4-2. 5

1.87

8

2. 1-2.7

5.96

34

2. 3-2. 6

6.19

4

2. 3-2. 5

2.95

5

2. 3-2. 5

5.89

2

2. 3-2. 5

2.41

7

2. 1-2.6

5.06

15

2. 2-2. 5

5.19

6

2. 2-2. 4

3.04

8

2. 0-2. 6

6.96

20

2. 1-2.5

5.78

10

2. 0-2. 6

6.96

27

2.0-2. 5

5.69

17

2. 1-2.3

4.30

4

1. 9-2.3

8.91

4

2. 0-2. 2

3.27

10

2. 0-2. 2

6.73

2

1. 9-2.3

7.90

5

2.0

0.00

7

1.7-2. 1

5.16

17

1. 7-2.0

6.02

5

1. 6-2.0

6.96

17

1. 7-1.9

5.39

4

2. 5-2.6

2.77

2

2.2-2. 8

7.57

9

2.3-2. 6

6.28

3

2.3-2. 5

3.95

4

2.3-2. 6

4.15

8

2. 2-2. 6

4.71

20

2. 3-2.6

4.69

7

2.0-2. 5

4.93

34

2.2-2A

4.12

4

2.2-2A

3.25

6

2.2-2. 5

5.25

7

2.2-2A

3.55

4

2. 1-2.5

4.67

10

2. 1-2.4

3.66

15

2. 1-2.4

4.62

20

2. 1-2.4

5.87

5

2. 1-2.3

6.43

2

2. 1-2.3

2.93

8

2. 0-2. 4

4.82

27

2. 0-2. 4

5.22

17

2.0-2. 2

3.27

10

2. 0-2. 3

7.06

4

2. 0-2. 3

6.73

2

2.0-2. 1

2.66

5

2.0-2. 1

2.41

7

1. 9-2.2

3.93

17

1.7-1. 8

2.90

4

1.6-1. 8

2.83

17

1.6-1. 8

4.16

5

1984

JONES ET AL. — BLARINA SYSTEMATICS

69

Table 2. — Continued.

ID

Sample

*B

NEWPAR

B

CLARKS

B

MOSCOW

B

ROBINS

*B

CRANKS

*B

BACKCR

B

CHIMNE

T

RIVERS

BAKER B

T

STELZE

*B

BAKER B

T

CUMBER

*T

BACKCR

NEWPAR

T

LADDSQ

T

CONARD

T

PECCAR

T

BRYNJU

T

WILLAR

H

SCHULZ

*T

MEYERC

T

JASPER

*T

CRANKS

H or C

KLEINC

C

CAVEWI

*C

MEYERC

c

ARREDO

c

INDIAN

c

HAILEX

B

CLARKS

*B

NEWPAR

B

MOSCOW

*B

BACKCR

*T

NEWPAR

B

ROBINS

B

CHIMNE

*B

CRANKS

T

PECCAR

T

STELZE

T

CUMBER

T

RIVERS

T

LADDSQ

T

CONARD

*B

BAKERB

*T

BAKERB

T

WILLAR

*T

BACKCR

T

BRYNJU

MEYERC

H

SCHULZ

*T

CRANKS

H or C

KLEINC

T

JASPER

C

CAVEWI

*C

MEYERC

C

ARREDO

C

HAILEX

C

INDIAN

Mean 2 SE Range CV N

Lingual length of Ml (F = 40.53)

2.45 0.10

2.33 0.09

2.27 0.07

2.23 0.10

2.22 0.03

2.20 0.20

2.19 0.05

2.15 0.13

2.15 0.08

2.15 0.04

2.14 0.07

2.12 0.11

2.10 0.20

2.10 0.08

2.08 0.10

2.08 0.05

2.08 0.05

2.04 0.05

2.01 0.03

1.99 0.04

1.99 0.03

1.98 0.10

1.95 0.10

1.94 0.08

1.90 0.08

1.85 0.03

1.60 0.08

1.60 0.06

1.58 0.03

Anterior width of Ml (F — 26.83)
2.67 0.11

2.65 0.30

2.57 0.13

2.50 0.12

2.48 0.13

2.45 0.06

2.45 0.09

2.41 0.05

2.35 0.07

2.34 0.04

2.33 0.13

2.33 0.13

2.32 0.10

2.32 0.06

2.29 0.14

2.28 0.07

2.26 0.05

2.25 0.50

2.23 0.09

2.17 0.06

2.17 0.06

2.15 0.30

2.12 0.04

2.10 0.22

2.01 0.05

2.01 0.05

1.80 0.12

1.79 0.04

1.74 0.10

2.4-2. 5

2.89

2

2. 2-2. 6

6.06

9

2. 2-2. 4

4.19

7

2. 1-2.3

4.30

4

2. 1-2.4

3.46

20

2.0-2. 3

7.87

3

2.1-2. 3

2.93

8

2.0-2. 3

6.01

4

1. 9-2.4

5.90

10

1.9-2. 5

5.28

34

2.0-2. 3

4.55

7

1.9-2. 3

6.28

6

2.0-2. 2

6.73

2

2.0-2. 2

3.89

4

2. 0-2. 2

5.27

5

2. 0-2. 2

4.14

15

2. 0-2. 3

4.83

20

2. 0-2. 2

3.65

8

1. 9-2.2

3.89

27

1. 9-2.1

2.85

10

1.9-2. 1

3.02

17

1. 9-2.1

4.85

4

1. 9-2.0

3.63

2

1. 9-2.1

4.61

5

1. 8-2.0

5.26

7

1. 8-2.0

3.38

17

1.5-1. 7

5.10

4

1.5-1. 7

4.42

5

1.5-1. 7

3.57

17

2. 4-2. 9

5.93

9

2.5-2. 8

8.01

2

2.4-2. 9

6.63

7

2. 4-2. 6

4.00

3

2. 3-2. 6

5.08

4

2. 4-2. 5

2.36

4

2. 2-2.6

4.88

8

2. 2-2.6

4.24

20

2. 1-2.7

6.69

20

2. 1-2.6

5.37

34

2. 1-2.5

7.00

6

2. 2-2. 5

5.41

4

2. 2-2. 4

4.72

5

2. 1-2.5

4.94

15

2. 0-2. 6

8.16

7

2. 1-2.5

4.98

10

2. 0-2. 6

5.68

27

2.0-2. 5

15.71

2

2. 1-2.5

5.76

8

2. 0-2. 4

5.82

17

2. 0-2. 3

4.37

10

2.0-2. 3

9.87

2

2. 1-2.2

2.1 1

5

1. 9-2.4

10.29

4

1. 9-2.1

3.43

7

1.9-2. 2

4.48

17

1.7-1. 9

6.42

4

1.7-1. 9

4.61

17

1.6-1. 9

6.55

5

Table 2. — Continued.

ID

Sample

*B

NEWPAR

B

MOSCOW

B

CLARKS

*B

BACKCR

B

ROBINS

NEWPAR

B

CHIMNE

*B

CRANKS

T

STELZE

T

BRYNJU

*B

BAKERB

T

LADDSQ

T

PECCAR

T

CUMBER

T

RIVERS

*T

BAKERB

T

WILLAR

T

CONARD

MEYERC

T

JASPER

BACKCR

CRANKS

H

SCHULZ

H or C

KLEINC

*c

MEYERC

C

CAVEWI

C

INDIAN

c

HAILEX

c

ARREDO

*B

ROBINS

*B

BAKERB

*B

CRANKS

B

CLARKS

B

CHIMNE

B

BACKCR

T

ZOOCAV

B or T

TROUTC

LADDSQ

*T

ROBINS

T

STELZE

*T

CUMBER

T

RIVERS

T

WILLAR

T

STRAIT

T

BRYNJU

*T

CRANKS

T

SCHMIT

*T

BAKERB

T

GUYWIL

T

PECCAR

T

CONARD

MEYERC

*c

CUMBER

H

SCHULZ

*H

CRANKS

H

KLEINC

H

FRIESE

*C

LADDSQ

*C

MEYERC

C

REDDIC

c

HAILEX

c

INDIAN

c

ARREDO

Mean 2 SE Range CV N

Posterior width of Ml (F = 32.33)

2.85 0.10

2.81 0.11

2.81 0.12

2.73 0.07

2.70 0.12

2.60 0.14

2.60 0.05

2.60 0.06

2.57 0.05

2.54 0.07

2.51 0.13

2.50 0.06

2.49 0.04

2.48 0.11

2.45 0.13

2.45 0.08

2.44 0.06

2.44 0.08

2.37 0.07

2.35 0.24

2.35 0.30

2.35 0.10

2.28 0.08

2.26 0.10

2.11 0.04

2.10 0.04

1.90 0.09

1.90 0.05

1.85 0.06

Length of ml (F— 103.66)

2.49 0.03

2.45 0.04

2.44 0.02

2.44 0.02

2.42 0.04

2.42 0.10

2.35 0.09

2.34 0.08

2.30 0.06

2.30 0.00

2.28 0.02

2.28 0.06

2.26 0.04

2.25 0.02

2.25 0.06

2.25 0.05

2.24 0.02

2.22 0.08

2.21 0.02

2.20 0.06

2.19 0.03

2.19 0.03

2.17 0.02

2.06 0.05

2.04 0.03

2.01 0.02

1.99 0.05

1.94 0.05

1.93 0.07

1.92 0.02

1.78 0.05

1.73 0.03

1.72 0.04

1.64 0.12

2. 8-2.9

2.48

2

2. 6-3.0

5.59

7

2. 5-3.0

6.52

9

2. 7-2. 8

2.1 1

3

2.6-2. 8

4.28

4

2.5-2. 8

5.44

4

2.5-2. 7

2.91

8

2. 4-2. 9

4.92

20

2. 3-2. 9

5.25

34

2. 3-2. 6

4.18

8

2. 2-2. 7

6.67

7

2. 4-2. 6

2.83

5

2. 3-2.7

4.19

20

2. 3-2. 7

5.35

6

2.3-2. 6

5.27

4

2. 2-2. 6

4.81

10

2. 2-2. 8

6.65

27

2. 2-2. 8

6.35

15

2. 1-2.7

6.1 1

17

2. 1-2.6

10.13

4

2.2-2. 5

9.03

2

2. 3-2.4

3.01

2

2. 1-2.5

5.77

10

2. 1-2.4

5.05

5

2. 0-2. 2

4.06

17

2. 0-2. 2

2.75

7

1. 8-2.0

5.26

5

1. 7-2.1

5.26

17

1.8-1. 9

3.12

4

2.4-2. 7

3.1 1

22

2. 3-2. 7

3.50

18

2.4-2. 5

2.03

19

2. 3-2. 6

3.08

44

2.3-2. 6

3.34

18

2.3-2. 6

4.53

5

2. 1-2.6

6.10

10

2. 2-2. 4

3.82

5

2. 2-2. 4

3.07

5

2.3

0.00

2

2. 0-2. 5

4.03

72

2. 2-2. 4

3.90

8

2. 1-2.3

3.09

10

2. 1-2.5

4.16

59

2. 0-2. 4

5.02

13

2. 2-2. 4

3.65

1 1

2. 2-2. 3

2.25

26

2. 1-2.3

3.77

5

2. 0-2. 4

4.30

61

2. 1-2.3

3.21

5

2. 0-2. 4

5.42

37

2. 1-2.4

3.33

21

2. 1-2.3

3.12

59

1.9-2. 1

3.53

9

1.8-2. 1

3.29

29

1.9-2. 1

2.22

20

1.8-2. 1

3.93

9

1 .9-2.0

2.82

5

1. 9-2.0

2.99

3

1. 8-2.0

3.38

55

1 .6-1.9

5.70

15

1.6-1. 8

3.40

17

1.7-1. 8

2.60

5

1.5-1. 8

8.18

5

1984

JONES ET AL .-BLARINA SYSTEMATICS

71

Table 2. — Continued.

ID

Sample

*B

ROBINS

*B

CRANKS

*B

BAKERB

B

CLARKS

B

CHIMNE

B

BACKCR

*T

ROBINS

T

ZOOCAV

*T

LADDSQ

*T

CUMBER

*T

CRANKS

B or T

TROUTC

T

STELZE

T

RIVERS

T

STRAIT

*x

BAKERB

T

WILLAR

T

SCHMIT

T

BRYNJU

T

PECCAR

T

GUYWIL

*T

MEYERC

T

CONARD

*C

CUMBER

*C

CRANKS

H

SCHULZ

H

KLE1NC

H

FRIESE

*C

LADDSQ

*C

MEYERC

C

REDDIC

C

HAILEX

C

INDIAN

C

ARREDO

Mean

2 SE

Range

cv

N

1 of ml (F =

75.87)

1.74

0.03

1.6-1. 9

3.83

22

1.70

0.03

1.6-1. 8

3.40

19

1.68

0.04

1.5-1. 8

5.23

18

1.66

0.02

oo

3.75

44

1.65

0.03

1.6-1. 8

3.75

18

1.62

0.04

1.6-1. 7

2.76

5

1.60

0.00

1.6

0.00

2

1.59

0.06

1.4-1. 7

5.51

10

1.58

0.04

1. 5-1.6

2.83

5

1.58

0.05

1.5-1. 7

4.49

8

1.57

0.02

1.4-1. 6

3.39

26

1.56

0.05

1. 5-1.6

3.51

5

1.55

0.02

1.4-1. 7

4.71

72

1.54

0.04

1.4-1. 6

4.54

10

1.54

0.04

1.4-1. 6

4.23

13

1.53

0.02

1.4-1. 7

5.38

61

1.52

0.02

1.3-1. 7

5.22

59

1.52

0.08

1.4-1. 6

5.50

5

1.50

0.05

1.4-1. 6

5.16

1 1

1.50

0.03

1.3-1. 6

5.67

37

1.48

0.08

1.4- 1.6

5.65

5

1.47

0.02

1.4-1. 6

4.77

59

1.47

0.03

1.4-1. 6

3.94

21

1.42

0.04

1.3-1. 5

4.69

9

1.38

0.02

1. 2-1.4

3.79

20

1.38

0.03

1.2-1. 5

5.37

29

1.34

0.05

1. 2-1.4

5.40

9

1.32

0.04

1. 3-1.4

3.39

5

1.27

0.07

1.2-1. 3

4.56

3

1.26

0.02

1. 2-1.4

5.36

55

1.23

0.03

1. 1-1.3

5.00

15

1.19

0.02

1. 1-1.3

3.59

17

1.18

0.04

1.1-1. 2

3.79

5

1.12

0.07

1.0-1. 2

7.47

5

TAXONOMY

The three Holocene species recognized in the ge-
nus Blarina currently consist of a total of 18 sub-
species. Blarina brevicauda is represented by 1 3
nominal subspecies arranged herein as representa-
tives of two semispecies, the brevicauda semispecies
and the ta/poides semispecies. Blarina carolinensis
tentatively is represented by four subspecies, where-
as B. hylophaga is represented by only two subspe-
cies. Synonymies of these taxa precede a discussion
of their taxonomic status.

The first citation in synonymies is to the original
description. The second is to the first use of the name
combination presently employed if it differs from
the name originally proposed. Next is a chronolog-
ical list of synonyms and name combinations ap-

plied to the taxon as presently recognized. Extinct
taxa are identified (f). References to relevant taxo-
nomic studies and Hall’s (1981) “The mammals of
North America” are provided when names em-
ployed in those publications differed from names
presently recognized.

Blarina Gray, 1838

Blarina Gray, Proc. Zool. Soc. London for 1837, p. 124, 1838.
Type species Corsira ( Blarina ) talpoides Gray [=Sorex tal-
poides Gapper], by original designation; elevated to generic
rank by Lesson, Nouv. Tableau Mammif., p. 89, 1842 (see

also Baird, Mammals, in Repts. Expl. Surv 8(1 ):36, 1858).

Brachysorex Duvernoy, Mag. de Zool., 2nd ser., 4:37-41, 1842.
Type species Brachysorex brevicaudus Duvernoy [=Sorex
brevicaudus Say], by original designation.

72

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 2. — Continued.

ID

Sample

*B

ROBINS

B

CLARKS

B

BACKCR

*B

BAKERB

B

CHIMNE

*B

CRANKS

B or T

TROUTC

T

ZOOCAV

LADDSQ

T

STELZE

T

RIVERS

*T

CRANKS

T

SCHMIT

T

STRAIT

CUMBER

*T

ROBINS

T

-CONARD

T

BRYNJU

T

WILLAR

T

PECCAR

♦ 'J'

BAKERB

*T

MEYERC

T

GUYWIL

*c

CUMBER

H

SCHULZ

H

KLEINC

*H

CRANKS

H

FR1ESE

*C

MEYERC

*C

LADDSQ

C

REDDIC

c

INDIAN

c

HAILEX

c

ARREDO

Mean

2 SE

Range

cv

N

i2 (F =

69.57)

2.06

0.04

1. 9-2.2

4.18

22

2.05

0.02

1.8-2. 1

3.07

44

2.04

0.08

2. 0-2. 2

4.39

5

2.03

0.02

2.0-2. 1

2.39

18

2.02

0.04

1. 9-2.2

4.66

18

2.02

0.03

1.9-2. 1

2.65

19

1.98

0.08

1.9-2. 1

4.23

5

1.97

0.07

1. 7-2.1

5.89

10

1.96

0.08

1. 9-2.1

4.56

5

1.94

0.02

1. 8-2.1

4.04

72

1.94

0.03

1. 9-2.0

2.66

10

1.94

0.03

1. 8-2.0

3.56

26

1.92

0.07

1. 8-2.0

4.36

5

1.91

0.04

1. 8-2.0

3.36

13

1.90

0.06

1. 7-2.0

4.87

8

1.90

0.20

1. 8-2.0

7.44

2

1.90

0.03

1. 8-2.0

3.33

21

1.88

0.04

1. 8-2.0

3.21

1 1

1.88

0.02

1. 8-2.0

3.78

59

1.86

0.03

1. 7-2.0

5.61

37

1.85

0.02

1. 7-2.0

4.59

61

1.84

0.02

1. 7-2.0

3.47

59

1.82

0.04

OO

1

SO

2.46

5

1.81

0.05

1. 7-1.9

4.32

9

1.81

0.02

1. 7-1.9

3.29

29

1.77

0.05

1.7-1. 9

4.00

9

1.75

0.03

1.6-1. 9

3.93

20

1.68

0.04

1.6-1. 7

2.66

5

1.67

0.02

1.6-1. 8

3.72

55

1.67

0.07

1.6-1. 7

3.46

3

1.53

0.04

1.4-1. 7

7.26

15

1.52

0.04

1.5-1. 6

2.94

5

1.48

0.02

1.4-1. 5

2.96

17

1.44

0.08

1.3-1. 5

6.21

5

Talposorex Pomel, Arch. Sci. Phy. Nat. (Geneva), 9:248, 1848.
Type species Talposorex platyurus Pomel [=Sorex brevicaudus
Say], by original designation.

Anotus Wagner, Suppl. Schreber’s Saugthiere, 5:550-551, 1855.
Type species Sorex (Anotus) carolinensis Bachman, by original
designation.

Blarina brevicauda (Say, 1823)
brevicauda semispecies

Blarina brevicauda brevicauda (Say, 1823)

Sorex brevicaudus Say, in Long, Account of an exped. ... to
the Rocky Mts., 1:164, 1823.

Blarina brevicauda: Baird, Mammals, in Repts. Expl. Surv.
. . . , 8(1):42, 1858.

Blarina costaricensis J. A. Allen, Bull. Amer. Mus. Nat. Hist.,
3:205, 1891; regarded as synonym of B. brevicauda by Mer-
riam, N. Amer. Fauna, 10:12-13, 1895.
f Blarina fossilis Hibbard, Univ. Kansas Sci. Bull., 29 (pt. 2):
238, 1943; regarded as subspecies of B. brevicauda by Hib-

bard. Trans. Kansas Acad. Sci., 60:333, 1957; regarded as
synonym of B. b. brevicauda by Graham and Semken, J.
Mamm., 57:437, 1976.

Blarina brevicauda manitobensis Anderson, 1947

Blarina brevicauda manitobensis Anderson, Bull. Natl. Mus.
Canada, 102:23, 1947.

talpoides semispecies

Blarina brevicauda aloga Bangs, 1902
Blarina brevicauda aloga Bangs, Proc. New England Zool. Club,
3:76, 1902.

Blarina brevicauda angusta Anderson, 1943

Blarina brevicauda angusta Anderson, Ann. Rept. Provancher
Soc. Nat. Hist., Quebec, p. 52, 1943.

Blarina brevicauda churchi Bole and Moulthrop, 1942
Blarina brevicauda churchi Bole and Moulthrop, Sci. Publ.,
Cleveland Mus. Nat. Hist., 5:109, 1942.

Blarina brevicauda compacta Bangs, 1 902

1984

JONES ET AL. — BLARINA SYSTEM ATICS

73

Table 2. — Continued.

!D

Sample

*B

ROBINS

*B

CRANKS

B

BACKCR

B

CLARKS

B

CHIMNE

*B

BAKERB

*T

ROBINS

*T

CRANKS

B or T

TROUTC

*T

LADDSQ

T

STELZE

T

ZOOCAV

*T

CUMBER

T

RIVERS

T

STRAIT

T

CONARD

BAKERB

T

GUYWIL

T

SCHMIT

T

PECCAR

T

WILLAR

*C

CUMBER

T

BRYNJU

*T

MEYERC

H

SCHULZ

*H

CRANKS

H

KLEINC

*c

LADDSQ

*C

MEYERC

H

FRIESE

C

REDDIC

C

INDIAN

C

HAILEX

C

ARREDO

Mean

2 SE

Range

cv

N

5

Kj

II

53.82)

1.39

0.02

1.3-1. 5

3.78

22

1.39

0.02

1.3-1. 5

3.30

19

1.38

0.08

1.3-1. 5

6.06

5

1.38

0.02

1.3-1. 5

3.88

44

1.37

0.03

1.3-1. 5

4.35

18

1.36

0.03

1. 2-1.4

4.47

18

1.35

0.10

1. 3-1.4

5.24

2

1.32

0.03

1.2-1. 4

5.46

26

1.32

0.04

1.3-1. 4

3.39

5

1.32

0.04

1. 3-1.4

3.39

5

1.31

0.02

1.2-1. 5

5.84

72

1.30

0.07

1.1-1. 5

8.1 1

10

1.29

0.05

1. 2-1.4

4.98

8

1.28

0.03

1.2-1. 3

3.29

10

1.27

0.04

1.2-1. 4

4.97

13

1.24

0.03

1. 2-1.4

4.81

21

1.24

0.02

1. 1-1.4

4.98

61

1.24

0.05

1. 2-1. 3

4.42

5

1.24

0.05

1.2-1. 3

4.14

5

1.23

0.03

1. 1-1.4

6.60

37

1.23

0.01

1. 1-1.3

4.44

59

1.22

0.03

1.2-1. 3

3.61

9

1.21

0.03

1.1-1. 3

4.46

11

1.20

0.01

1.1-1. 3

2.67

59

1.18

0.02

1.1-1. 3

4.56

29

1.17

0.02

1. 1-1.2

4.02

20

1.16

0.04

1.1-1. 2

4.56

9

1.13

0.07

1. 1-1.2

5.09

3

1.10

0.02

1.0-1. 2

5.25

55

1.08

0.04

1.0-1. 1

4.14

5

1.07

0.05

1.0-1. 2

8.44

15

1.00

0.00

1.0

0.00

5

1.00

0.02

0
SO

1

3.54

17

0.98

0.04

O

SO

o

4.56

5

Blarina brevicauda compacta Bangs, Proc. New England Zool.
Club, 3:77, 1902.

Blarina brevicauda hooperi Bole and Moulthrop, 1942
Blarina brevicauda hooperi Bole and Moulthrop, Sci. Publ.,
Cleveland Mus. Nat. Hist., 5:110, 1942.

Blarina brevicauda kirtlandi Bole and Moulthrop, 1 942

Blarina brevicauda kirtlandi Bole and Moulthrop, Sci. Publ..
Cleveland Mus. Nat. Hist., 5:99, 1942; tentatively regarded
as distinct species by Graham and Semken, J. Mamm., 57:
445, 1976.

t Blarina brevicauda ozarkensis Brown, 1908

Blarina brevicauda ozarkensis Brown, Mem. Amer. Mus. Nat.
Hist., 9:170, 1908.

Blarina ozarkensis: Graham, M.S. thesis, Univ. Iowa, p. 12,
1972; Graham and Semken, J. Mamm., 57:434, 1976.
Blarina brevicauda pallida R. W. Smith, 1 940
Blarina brevicauda pallida R. W. Smith, Amer. Midland Nat.,
24:223, 1942.

t Blarina brevicauda simplicidens Cope, 1 899

Blarina simplicidens Cope, J. Acad. Nat. Sci. Philadelphia, 1 1:
219, 1899.

Blarina brevicauda simplicidens: Hibbard, Trans. Kansas Acad.
Sci., 60:333, 1957.

Blarina brevicauda talpoides (Gapper, 1830)

Sorex talpoides Gapper, Zool. J., 5:202, 1 830; regarded as syn-
onym of B. brevicauda by Merriam, N. Amer. Fauna, 10:
12, 1895.

Blarina brevicauda talpoides: Bangs, Proc. New England Zool.
Club, 3:75, 1902.

Sorex dekayi Bachman, J. Acad. Nat. Sci. Philadelphia, 7:362-
402, 1837; nomenclatorial history of this name and “ Sorex
dekayi De Kay” were reviewed by Handley and Choate,
Proc. Biol. Soc. Washington, 83:195-202, 1970.

Galemys micrurus Pomel, Arch. Sci. Phy. Nat. (Geneva), 9:
249, 1 849; nomenclatorial history of this name was reviewed
by Handley and Choate, Proc. Biol. Soc. Washington, 83:
195-202, 1970.

Blarina angusticeps Baird, Mammals, in Repts. Expl. Surv.

74

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 2. — Continued.

ID

Sample

*B

CRANKS

B

CLARKS

*B

ROBINS

B

BACKCR

T

SCHMIT

LADDSQ

B

CHIMNE

*B

BAKERB

B or T

TROUTC

T

CONARD

*y

CRANKS

T

ZOOCAV

T

WILLAR

T

BRYNJU

*C

CUMBER

T

STELZE

*T

MEYERC

T

PECCAR

*T

CUMBER

T

GUYWIL

*T

BAKERB

*T

ROBINS

T

STRAIT

T

RIVERS

H

SCHULZ

*H

CRANKS

H

KLEINC

*C

MEYERC

H

FRIESE

LADDSQ

C

REDDIC

c

INDIAN

c

HAILEX

c

ARREDO

Mean 2 SE

Length of m3 (F = 43.85)

1.67

0.03

1.67

0.02

1.67

0.04

1.64

0.10

1.64

0.05

1.64

0.08

1.62

0.04

1.62

0.02

1.60

0.06

1.58

0.04

1.58

0.04

1.58

0.05

1.58

0.02

1.56

0.04

1.56

0.05

1.55

0.02

1.55

0.02

1.53

0.03

1.53

0.05

1.52

0.04

1.51

0.02

1.50

0.00

1.50

0.06

1.48

0.04

1.44

0.03

1.40

0.03

1.38

0.04

1.37

0.02

1.34

0.08

1.33

0.07

1.22

0.03

1.20

0.06

1.18

0.03

1.16

0.05

Range

cv

N

1.6-1. 8

3.90

19

1. 5-1.9

4.89

44

1.5-1. 8

5.03

22

1.5-1. 8

6.95

5

1.6-1. 7

3.34

5

1.6-1. 8

5.45

5

1.5-1. 7

4.51

18

1.5-1. 8

4.86

18

1.5-1. 7

4.42

5

1.4-1. 7

5.52

21

1. 4-1.7

6.46

26

1.4-1. 7

4.99

10

1.4-1. 8

5.18

59

1.5-1. 7

4.31

1 1

1.5-1. 7

4.67

9

1.3-1. 7

5.19

72

1.4-1. 7

4.39

59

1.4-1. 7

6.87

37

1.4-1. 6

4.64

8

1. 5-1.6

2.94

5

1

oo

5.68

61

1.5

0.00

2

1. 4-1.7

7.20

13

1.4-1. 6

4.27

10

1. 3-1.6

5.36

29

1.3-1. 5

5.44

20

1.3-1. 5

4.84

9

1.2-1. 7

5.14

55

1.2-1. 4

6.68

5

1.3-1. 4

4.33

3

1. 1-1.3

5.54

15

1. 1-1.3

5.89

5

1. 1-1.3

5.65

17

1. 1-1.2

4.72

5

. . . , 8( 1 ): 34, 1858; based on “deformed skull” according to
Merriam, N. Amer. Fauna, 10:10, 1895 (see also Bole and
Moulthrop, Sci. Publ., Cleveland Mus. Nat. Hist., 5:111,
1942).

Blarina brevicauda telmalestes Merriam, 1895
Blarina telmalestes Merriam, N. Amer. Fauna, 10:15, 1895;

Hall, The mammals of North America, 1:57, 1981.

Blarina brevicauda telmalestes: Handley, Mammals of the Dis-
mal Swamp: a historical account, in The Great Dismal
Swamp, p. 308, 1979.

Blarina carolinensis (Bachman, 1837)

Blarina carolinensis carolinensis (Bachman, 1837)

Sorex carolinensis Bachman, J. Acad. Nat. Sci. Philadelphia,
7:366, 1837.

[Blarina carolinensis ] carolinensis: Genoways and Choate, Syst.
Zool., 21:114, 1972.

Blarina brevicauda carolinensis: Merriam, N. Amer. Fauna,

10:13, 1895; Hall, The mammals of North America, 1:54,
1981.

Blarina carolinensis minima Lowery, 1943

Blarina brevicauda minima Lowery, Occas. Papers Mus. Zool.,
Louisiana State Univ., 13:218, 1943; Hall, The mammals
of North America, 1:56, 1981.

Blarina carolinensis peninsulae Merriam, 1895

Blarina carolinensis peninsulae Merriam, N. Amer. Fauna, 10:
14, 1895; George et al., J. Mamm., 63:641, 1982.

[ Blarina brevicauda ] peninsulae: Trouessart, Catalogus Mam-
malium . . . , fasc. 1, p. 188, 1897.

Blarina brevicauda peninsulae: Hall, The mammals of North
America, 1:56, 1981.

Blarina carolinensis shermani Hamilton, 1955

Blarina brevicauda shermani Hamilton, Proc. Biol. Soc. Wash-
ington, 68:37, 1955; Hall, The mammals of North America,
1:56, 1981.

B[larina]. carolinensis shermani: George et al., J. Mamm., 63:
643, 1982.

1984

JONES ET AL .-BLARINA SYSTEM ATICS

75

Table 2. — Continued.

!D

Sample

B

CHIMNE

*B

ROBINS

B

BACKCR

B

CLARKS

*B

CRANKS

*B

BAKERB

B or T

TROUTC

*T

ROBINS

*T

CUMBER

T

SCHMIT

*T

LADDSQ

T

STELZE

T

STRAIT

*C

CUMBER

T

CONARD

*T

CRANKS

T

GUYWIL

T

ZOOCAV

T

WILLAR

*T

BAKERB

H

SCHULZ

T

PECCAR

*T

MEYERC

T

RIVERS

T

BRYNJU

H

KLEINC

*C

LADDSQ

*H

CRANKS

H

FRIESE

*C

MEYERC

C

REDDIC

C

INDIAN

C

HAILEX

C

ARREDO

Mean

2 SE

Range

cv

N

i of m3 (F =

31.54)

1.16

0.04

1.0-1. 3

7.40

18

1.15

0.03

1. 1-1.4

6.44

22

1.14

0.05

1. 1-1.2

4.81

5

1.14

0.02

1.0-1. 2

5.40

44

1.1 1

0.03

1.0-1. 2

5.12

19

1.11

0.03

1.0-1. 2

6.56

18

1.10

0.06

1.0-1. 2

6.43

5

1.10

0.00

1.1

0.00

2

1.09

0.05

1.0-1. 2

5.89

8

1.08

0.04

1.0-1. 1

4.14

5

1.08

0.04

1.0-1. 1

4.14

5

1.08

0.01

1.0-1. 2

5.5!

72

1.07

0.06

1.0-1. 3

9.65

13

1.06

0.04

1. 0-1.1

4.99

9

1.04

0.02

1.0-1. 1

4.86

21

1.04

0.02

0
SO

1

5.54

26

1.04

0.05

1. 0-1.1

5.27

5

1.04

0.04

p

so

1

6.72

10

1.03

0.01

1.0-1. 2

5.53

59

1.03

0.01

0
so

1

5.00

61

1.01

0.02

O

SO

\_

4.05

29

1.01

0.02

0.9-1. 2

7.16

37

1.00

0.01

0.9-1. 1

4.52

59

1.00

0.00

1.0

0.00

10

1.00

0.03

0.9-1. 1

4.47

1 1

0.99

0.02

0
SO

1

b

3.37

9

0.97

0.07

p

SO

1

b

5.97

3

0.97

0.02

p

b

L

b

5.07

20

0.94

0.05

p

b

l

o

5.83

5

0.93

0.01

o

00

1

b

5.42

55

0.89

0.03

p

bo

l

b

7.22

15

0.86

0.05

0.8-0. 9

6.37

5

0.84

0.02

0.8-0. 9

6.03

17

0.84

0.05

0.8-0. 9

6.52

5

Blarina hylophaga Elliot, 1899

Blarina hylophaga hylophaga Elliot, 1899

Blarina brevicauda hulophaga [szc] Elliot, Field Columbian
Mus., Zool. Ser., 1:287, 1899.

Blarina brevicauda hylophaga Elliot, Field Columbian Mus.,
Zool. Ser., 6:46 1 , 1905 (correction of previous error in spell-
ing or transliteration).

Blarina hylophaga hylophaga: George et al., Ann. Carnegie
Mus., 50:504, 1981.

Blarina brevicauda carolinensis: Jones and Glass, Southwest-
ern Nat., 5:138, 1 960; Hall, The mammals of North Amer-
ica, 1:54-55, 1981.

[Blarina carolinensis ] carolinensis: Genoways and Choate, Syst.
Zool., 21:114, 1972.

Blarina carolinensis carolinensis: Schmidly and Brown, South-
western Nat., 24:45, 1979.

Blarina hylophaga plumbea Davis, 1941

Blarina brevicauda plumbea Davis, J. Mamm., 22:317, 1941;
Hall, The mammals of North America, 1:56, 1981.

Blarina hylophaga plumbea: George et al., Ann. Carnegie Mus.,
50:510, 1981.

Blarina carolinensis plumbea: Schmidly and Brown, South-
western Nat., 24:45, 1979.

Two additional extinct shrews, Cryptotis adamsi
and Pamcryptotis gidleyi, also are mentioned in the
accounts that follow because they originally were
referred to the genus Blarina. Synonymies of these
taxa are given below.

t Cryptotis adamsi (Hibbard, 1953)

Blarina adamsi Hibbard, J. Paleont., 27:29, 1953.

Cryptotis adamsi: Repenning, U.S. Geol. Surv. Prof. Paper,
565:39, 1967; Choate, Univ. Kansas Publ. Mus. Nat. Hist.,
19:289, 1970.

t Paracryptotis gidleyi (Gazin, 1933)

Blarina gidleyi Gazin, J. Mamm., 14:142, 1933; Repenmng,
U.S. Geol. Surv. Prof. Paper, 565:43, 1967.

76

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Paracryptotis gidleyi: Choate, Univ. Kansas Publ., Mus. Nat.
Hist., 19:294, 1970; Hibbard and Bjork, Contrib. Mus. Pa-
leont., Univ. Michigan, 23:175, 1971.

Ongoing research on geographic variation in Bla-
rina brevicauda almost certainly will reduce the
number of subspecies recognized in that species.
Justification for recognition of two semispecies
within B. brevicauda follows.

Mayr (1969) defined semispecies as “populations
that have acquired some, but not yet all, attributes
of species rank; borderline cases between species and
subspecies.” Semispecies of a species are allopatric
over most of their range, but may exist sympatrically
where their ranges abut (Dobzhansky et al., 1977).
They often hybridize in such areas, but gene flow
between them may be limited (Grant, 1971). Even-
tual elimination of gene flow between semispecies
would complete the process of speciation. We think
the relationship between what are now eastern and
western subspecies of B. brevicauda is explained bet-
ter by their recognition as semispecies than by pos-
sible alternative explanations.

Holocene shrews assigned to the western subspe-
cies B. b. brevicauda and B. b. manitobensis average
appreciably larger than Holocene shrews assigned
to more easterly subspecies although the extremes
of measurements overlap. They do not differ mor-
phologically except in size. Where the ranges of the
larger and smaller phena approach, there is no in-
dication of intergradation; where they occur sym-
patrically in southern Iowa and northern Missouri,
hybridization is suggested by the presence within
samples of both large and small shrews and shrews
variously intermediate between these extremes
(Moncrief et al., 1 982). The larger and smaller phena
have identical standard karyotypes, and in fact share
a balanced polymorphism (Genoways et al., 1977),
so standard karyotypes cannot be employed to as-
sess the extent of hybridization. Preliminary anal-
ysis of unpublished electrophoretic data (J. C. Pat-
ton, personal communication) suggests that allelic
polymorphisms are not evenly distributed between
the larger and smaller phena, which supports the
notion of at least a partial barrier to gene flow. Fi-
nally, the larger and smaller phena are recognizable
in the fossil record as long ago as the early Pleis-
tocene. The distributions of the phena have changed
since that time in response to climatic fluctuations,
but the mensural relationship between them has not.

We interpret these observations as follows. Two
populations of B. brevicauda began to diverge gen-
ically and morphometrically soon after the species

evolved, but gene flow between them prevented spe-
ciation. All known speciation events in Blarina ap-
parently have involved karyotypic rearrangements
(George et al., 1982), and no such rearrangements
have become fixed in the larger and smaller phena
of B. brevicauda. Nevertheless, the distinctive pop-
ulations have behaved essentially as separate species,
except when sympatric, throughout the Quaternary
Period. The mensural variability they exhibit in and
near areas of sympatry, apparently as a result of
hybridization, has been responsible for difficulty in
identifying fossils of Blarina in certain local faunas,
and led Graham and Semken (1976) to suggest the
taxa B. b. kirtlandi and “5. ozarkensis ” are distinct
species. Our data show the phena clearly are not
distinct species, but indicate that recognition as
semispecies is warranted.

Another possible interpretation of this phenom-
enon, which we elected not to accept, is that the
larger and smaller phena of B. brevicauda are sep-
arate “evolutionary species.” Wiley (1978) defined
an evolutionary species as “a single lineage of ances-
tor-descendant populations which maintains its
identity from other such lineages and which has its
own evolutionary tendencies and historical fate.” It
is evident that the ta/poides and brevicauda semi-
species have retained their identities since the early
Pleistocene; in this sense, they would qualify as dis-
tinct evolutionary species. However, the lack of
morphological and morphometric divergence of the
phena over a long period of time and the identical
standard karyotypes in the phena indicate evolu-
tionary stasis (in the sense of Gould and Eldredge,
1977), with the phena continuing to behave like
populations of the same species when they occur
together. We judge that the ta/poides and brevicauda
semispecies do not express separate evolutionary
tendencies and thus do not qualify as separate evo-
lutionary species.

Ongoing research on the status of two nominal
subspecies from Florida, herein referred to the
species Blarina carolinensis on geographic grounds,
might result in taxonomic readjustments. One sub-
species, B. c. peninsulae, has a karyotype more like
that of B. brevicauda than that of B. carolinensis
(George et al., 1982), and might be referred to the
former species or elevated to separate species rank
when electrophoretic and morphometric analyses
are completed. Another subspecies, B. c. shermani,
is known from few specimens and until recently was
thought to be extinct (J. N. Layne, personal com-
munication). Morphometrically it resembles shrews

1984

JONES ET AL. — B LARINA SYSTEMATICS

77

of the talpoides semispecies of B. brevicauda as much
as it does B. carolinensis, and it might be referred
to the former species if additional morphometric,
karyo logic, or electrophoretic data become available
for analysis. These taxonomic changes would entail
reinterpretation of some of the data presented herein,
but doubtfully would alter any major phylogenetic
or paleobiogeographic conclusions.

The taxonomic status of subspecies of Blarina
hylophaga has been reviewed (George et al., 1981),
and no further taxonomic changes are expected. At
the present time, the range of this species is narrowly
sympatric with those of shrews of the brevicauda
semispecies of B. brevicauda in southern Nebraska

and with shrews of both semispecies of B. brevicau-
da in their zone of sympatry in southern Iowa,
northern Missouri, and possibly northeastern Kan-
sas (Moncrief et al., 1982). The range of B. hyloph-
aga approaches that of B. carolinensis in northern
Louisiana (George et al., 1981). The present range
of B. carolinensis apparently is sympatric with that
of the talpoides semispecies of B. brevicauda on the
piedmont in northern Georgia and on the coastal
plain in Virginia and North Carolina (Tate et a!.,
1980). At the present time the three species recog-
nized herein do not occur together at any one lo-
cation, but they might have done so in the past
during a more equable climatic regime.

PALEONTOLOGY

Blancan
Hager man

The Hagerman local fauna (l.f.) was recovered from the Glenns
Ferry Formation west of Hagerman in Twin Falls County, Idaho.
Fossils were collected from approximately 300 sites in fluvial
deposits along the Snake River. The Hagerman is considered one
of the classic faunas of the late Pliocene of North America (Za-
krzewski, 1969). A potassium-argon date of 3.48 ± 0.27 million
years BP was obtained by Evemden et al. (1964). Although once
considered to be younger than the Rexroad l.f. of southwestern
Kansas, recent work suggests that the section at Hagerman rep-
resents a period of time equivalent to that beginning with the
Fox Canyon and ending with the Rexroad (Zakrzewski, personal
communication).

The Hagerman l.f. consists of the remains of molluscs, fishes,
herptiles, birds, and mammals. Skinner et al. (1972) published
the most recent list of mammals and compared the mammals
from Hagerman with those of Rexroad and other Blancan faunas
of Kansas, Texas, and Nebraska.

Evidence suggests that the local habitat was warmer and more
humid and had more vegetation than at present. Remains of
fishes, frogs, and aquatic birds indicate the presence of fresh water
(Brodkorb, 1958; Zakrzewski, 1969; Chantell, 1970). Rodents
represent marsh-meadow and valley slope communities; repre-
sentatives of drier upland habitats are rare but increase toward
the top of the section. Warmer winters are suggested by the pres-
ence of southern elements such as Baiomys (Zakrzewski, 1969).
Bjork (1970) discussed the zoogeography and paleoecology of
carnivores from the deposit. He described the area as “a broad
flood plain with trees and grassland adjacent to the tributary
streams of Lake Idaho.”

Today the area is included in the Artemisian biotic province
of Dice (1943) and is characterized by sagebrush steppe domi-
nated by sagebrush (Artemisia) and wheatgrass ( Agropyron ). Mean
annual precipitation is approximately 25 cm (NOAA, 1974).

Six species of insectivores have been identified
from the Hagerman material, including a “ Blarina ”

(Gazin, 1933). Gazin (1933) named Blarina gidleyi
on the basis of a fragmentary left ramus (USNM
12650 from the locality at T7S, R13E). The speci-
men was described as “closer in size and genera!
proportions to species of Blarina than to any of the
other North American shrews .... The lower mo-
lars . . . show differences which appear to be beyond
the range of variation seen in both Blarina brevi-
cauda and Blarina telmalestes. The trigonid portion
of these teeth is elongate giving the crescentic crest
a more obtuse appearance than in the living species
and the posterior portion of this crest is not so nearly
perpendicular to the inner margin of the tooth, being
directed slightly forward externally. The talonid
portion of the anterior molars is narrower trans-
versely, and the crest meets the posterior wall of the
trigonid at a more inward position somewhat as in
Sorex. Also, the postero-external wall of the talonid
crown in ml and m2 does not project bucally be-
yond the cingulum so noticeably as in Blarina brev-
icauda .... The reduction of the heel of the last
lower molar which is often so marked in Cryptotis
. . . distinguishes Blarina simplicidens Cope . . . from
the living species of Blarina and from Blarina gid-
leyi .... It is possible that were more complete re-
mains known the fossil form would be found to
represent an undescribed genus, presumably related
closely to Blarina ” (Gazin, 1933). Measurements
were given to compare the specimen of B. gidleyi
with one of B. brevicauda (USNM 83073).

Hibbard (1957) identified additional specimens
from the Hagerman l.f. as B. gidleyi. He noted that
this species was distinguished from B. simplicidens

78

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 3.— Univariate statistics for specimens o/'Paracryptotis gid-
leyi from the Hagerman l.f, Twin Falls County, Idaho, for mea-
surements 3-10 of the upper teeth and 19-24 of the lower teeth.

Variable

N

Mean

2 SE

Range

cv

3

6

2.42

0.10

2. 2-2.5

4.84

4

6

1.65

0.10

1.5-1. 8

7.42

5

6

1.27

0.08

1 . 1-1.4

8.15

6

6

2.25

0.04

2. 2-2. 3

2.43

7

6

2.13

0.04

2. 1-2.2

2.42

8

6

2.03

0.04

2.0-2. 1

2.54

9

6

2.17

0.07

2. 1-2.3

3.77

10

6

2.27

0.07

2. 2-2.4

3.60

19

13

2. 12

0.04

2. 0-2. 2

3.79

20

13

1.48

0.03

1 .4-1.6

4.06

21

13

1.77

0.05

1. 6-2.0

5.36

22

13

1.23

0.04

1.1-1. 3

5.12

23

13

1.34

0.04

1.2-1. 4

4.86

24

13

0.93

0.04

o

7

00

o

6.77

by its larger size and more developed cingula on the
lower molars and p4. He also noted certain simi-
larities between B. gid/eyi and Cryptotis kansensis
from the early Pleistocene of Meade County, Kan-
sas. Repenning (1967) examined a well preserved
topotype (UMMP 3304) and listed characteristics
of the specimen which he considered “clearly prim-
itive.” He suggested the possibility that gid/eyi was
intermediate between extant Blarina and Ade/ob-
larina (described from late Miocene deposits in Mal-
heur County, Oregon) and that gid/eyi should not
be assigned to Blarina.

Hibbard and Bjork (1971) reassigned the holotype
of B. gid/eyi, plus additional material from Hager-
man, to the genus Par aery ptotis. Paracryptotis gid-
leyi shares many of the characteristics of P. rex, as
described by Hibbard (1950), but differs in the pos-
terior emargination of P4 and Ml and the more
anterior position of the lower articular condyle.
Paracryptotis was described by Repenning ( 1 967) as
less advanced than Blarina in certain dental features
and in mandibular articulation. Probably P. gid/eyi
represents a lineage that evolved near the origin of
the Tribe Blarinini (Choate, 1970).

We examined eight maxillaries (UMMP 49899,
53270-71, 53346, 55058, and 59935) and 17 den-
taries (UMMP 33904, 34440, 45258, 49900, 49988,
50163-66, 55059; USNM 1 2650, 2 1 385, and 2 1 386)
from Hagerman. Univariate statistics for the most
complete specimens are shown in Table 3.

Dixon

As presently defined, the genus Blarina is first known from the
late Blancan Dixon and White Rock faunas of Kansas. The Dixon

l.f. was recovered from two sites on the Dixon farm southwest
of Kingman, Kingman County (T29S, R8W). This fauna is one
of several from the Meade formation of Kansas (Hibbard, 1956).

Molluscs, fishes, herptiles, birds, and small mammals compose
the Dixon l.f. The mammalian families present are the Soricidae,
Sciuridae, Geomyidae, Castoridae, Cricetidae, and Mustelidae.
All but six of the 14 mammalian species represented are extinct
(Hibbard, 1956, 1970).

The Dixon is the earliest cool fauna recognized in Kansas
(Skinner et ah, 1972). Deposition occurred during a climatic
period of more equable temperature and more effective precip-
itation (Hibbard, 1958, 1970; Brattstrom, 1967; Graham, 1976).
The insectivores and microtines indicate marshy habitat with
trees and shrubs (Hibbard, 1956). Kingman County now is in-
cluded in the Illinoian biotic province described by Dice (1943).

The age of the fauna is pre-Nebraskan (Taylor, 1966; Kurten
and Anderson, 1980) or Nebraskan (Hibbard, 1956, 1958, 1970;
Skinner et ah, 1972). It is the only fauna of its age known from
the Great Plains (Hibbard, 1970).

Part of a right maxillary with M 1 and M2 (UMMP
31966) and an edentulous left dentary (UMMP
31967) from Dixon were referred to Blarina sp. by
Hibbard (1956). Measurements 7 through 14 (2.1,
2.0, 2.2, 2.3, 1.6, 1.5, 2.3, and 2.1, respectively) of
the Ml and M2 were taken for the present study.
These measurements were compared with those of
reference samples of Holocene specimens (Table 1)
and with means derived from the Holocene mea-
surements. Measurements of the fossil are most sim-
ilar to those of modern B. brevicauda (ta/poides
semispecies), and the specimen from Dixon tenta-
tively is assigned to that species.

Fox Canyon

The Fox Canyon local fauna (University of Michigan locality
UM-K1-47) is located on the XI Ranch, sec. 35, T34S, R30W,
in southwestern Meade County, Kansas. It was one of several
fossil sites from the Rexroad Formation of Meade County in-
vestigated by Claude Hibbard. Initially Fox Canyon was consid-
ered a faunule of the Rexroad fauna; later it was recognized as a
separate fauna (Hibbard, 1950, 1967; Zakrzewski, 1967). The
Fox Canyon l.f. pertains to the Blancan land mammal age. Fox
Canyon is considered older than the Rexroad, Benson, and Hag-
erman faunas (Hibbard, 1967; Zakrzewski, 1967, 1969). It was
one of seven local faunas from Kansas from which sediments
were sampled for magnetic polarity by Lindsay et al. (1975). The
sample showed reversed magnetism and so probably was de-
posited during the late Gilbert epoch, approximately 3.5 million
years ago. Fox Canyon was considered the oldest of the faunas
(Lindsay et al., 1975). Dalquest (1978) suggested that the Beck
Ranch l.f. from Scurry County, Texas, dates from between the
periods represented by Fox Canyon and Rexroad. The Fox Can-
yon and Layer Cake (California) local faunas have been correlated
(Kurten and Anderson, 1980).

The fauna of Fox Canyon consists of remains of all vertebrate
classes taken from fluvial deposits along the sides of the canyon.
Hibbard (1950) provided the initial faunal list. Dalquest (1978)
included an unpublished list (compiled by Hibbard) of 34 mam-

1984

JONES ET AL. — B LARIN A SYSTEM ATICS

79

mals (19 of which are types) from Fox Canyon. Six insectivores,
a bat, a rabbit, 18 rodents, seven carnivores, and one deer were
identified by Hibbard. Approximately 24 extinct species are pres-
ent (Kurten and Anderson, 1980). Dalquest (1978) compared the
mammalian faunas from the Fox Canyon, Beck Ranch, Rexroad,
Blanco, and Benson local faunas.

Meade County lies within the High Plains subprovince of the
Great Plains. The High Plains are characterized as relatively flat
and treeless, with elevation ranging from 610 to 1,220 m above
sea level (Self, 1978). The natural vegetation of southwestern
Meade County consists of mixed (bluestem-grama) and short-
grass (grama-bufTalograss) prairie except for floodplain and sand-
sage vegetation along the Cimarron River (Kuchler, 1974). The
climate is semiarid, with the annual water loss by evaporation
exceeding annual precipitation. Mean annual precipitation is ap-
proximately 51 to 56 cm (Self, 1978).

Notiosorex jacksoni was one of five insectivores
described by Hibbard ( 1950, 1 953a) as new species.
Some of the specimens of N. jacksoni later were
reassigned to another new species, Blarina adamsi.
The type of B. adamsi is a maxillary bearing P2
through M2 (UMMP 27267). Paratypes include a
palate (UMMP 27268), a right maxillary (UMMP
27269), the anterior portion of a palate (UMMP
28403), and a second right maxillary (UMMP
27270); each specimen bears partial dentition (Hib-
bard, 1953a). Mandibles identified as B. adamsi
also were recovered from Keefe Canyon (UMMP
25777) (Hibbard, 1953a). A minimum of 30 rami
thus was identified.

Blarina adamsi was described by Hibbard ( 1 953a)
as smaller than what he termed B. b. carolinensis
[=77. hylophaga\, approximately the size of N. jack-
soni. The holotype has a P2 that is “anteroposte-
riorly flattened and does not possess a small lingual
cusp as in Recent forms. Reduced P3 is just visible
from labial side of maxillary. This tooth is not as
small as P3 of the Recent forms of B. b. carolinensis
and B. b. hulophaga [sic] Elliot. The posterior edge
of P4, Ml, and M2 not excavated.” Length of P2-
M2 was 4.5 mm. In UMMP 27268 the location of
large anterior palatine foramina in a depression not
observed in extant Blarina was noted. The mandi-
bles were separated from those of N. jacksoni on
the basis of the shape of the mental foramen and
the articular surfaces of the condyle. Length of m 1-
m3 of 41 mandibles averaged 3.91 mm; length of
c-m3 of eight specimens averaged 5.14 mm (Hib-
bard, 1953a).

Topotypes (UMMP 24352, 27310, and 28411),
all mandibles, from the Fox Canyon l.f. were ex-
amined by Repenning (1967), who reassigned the
species adamsi to the genus Cryptotis. Although the
dental formula of those specimens is the same as

Table 4. — Univariate statistics for specimens o/Cryptotis adamsi
from the Fox Canyon l.f, Meade County, Kansas, for measure-
ments 3-10 of the upper teeth and 19-24 of the lower teeth.

Variable

N

Mean

2 SE

Range

cv

3

9

1.70

0.05

1.6-1. 8

4.16

4

9

1.28

0.06

1.2-1. 4

6.52

5

9

0.90

0.03

O

OO

o

5.56

6

9

1.66

0.06

1.5-1. 8

5.33

7

9

1.62

0.04

1.5-1. 7

4.11

8

9

1.53

0.03

1. 5-1.6

3.26

9

8

1.53

0.03

1. 5-1.6

3.04

10

9

1.74

0.05

1.6-1. 8

4.17

19

13

1.55

0.05

1.4-1. 7

5.67

20

13

1.10

0.05

1.0-1. 2

7.42

21

13

1.42

0.04

1.3-1. 5

4.87

22

13

1.02

0.02

1.0-1. 1

4.29

23

13

1.24

0.03

1.2-1. 3

4.09

24

13

0.89

0.05

0.6-1 .0

10.69

that of extant Blarina (and therefore unlike that of
Cryptotis ), mandibular structure and lower and up-
per teeth were noted as identical to that of living
Cryptotis. Repenning listed the differences between
the fossils and Blarina as “(1) greater anteroposte-
rior shortening of the talonid of ml, (2) more pos-
terior placement of the metaconid of ml, relative
to the position of the protoconid, (3) greater reduc-
tion of the heel of m3, (4) retention of a primitive
blarinine mandibular articulation and associated jaw
structure, and (5) a rectangular M2.” Another man-
dible (USGS M65 1 1 ) from the Christmas Valley l.f.
(Hemphillian) of Lake County, Oregon, also was
identified as C. adamsi (Repenning, 1967). Choate
(1970) examined the material from Fox Canyon and
supported the assignment to Cryptotis. He suggested
that, during the mid- and late Pliocene, lineages of
Cryptotis and Blarina were at a similar grade of
evolution. We agree with assignment of adamsi to
the genus Cryptotis.

Twenty-five specimens (UMMP 24352, 27268-
74, 27305-12, 28403-07, and 28410) of Cryptotis
adamsi were examined for the present study. Uni-
variate statistics for the most complete specimens
examined (nine maxillaries and 13 dentaries) are
shown in Table 4.

White Rock

The White Rock l.f. was recovered from 12 sites in the Belle-
ville formation of Republic County, Kansas. The primary col-
lecting sites are located along the White Rock Canal southwest
of Republic. Deposition occurred in the paleovalley of the an-
cestral Republican River. The fauna was described by Eshelman
(1975).

80

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

The fauna consists of molluscs, crustaceans, fishes, herptiles,
birds, and mammals. Two species of reptiles are extinct. Forty-
three species of mammals are represented, of which 3 1 are extinct.
Small mammals (shrews and mice) predominate in the mam-
malian fauna (Eshelman, 1975). Of the extant species, all have
occurred in the area in historic times (Hall, 1981).

Taxa of the White Rock l.f. are derived from several com-
munities. Permanent water, stream- or river bank, lowland mead-
ow-valley, valley slope, and upland prairie habitats are indicated.
The taxa suggest deposition in more equable climate (Eshelman,
1975). Presently, Republic County is in the Smoky Hills sub-
province of the Great Plains. It is included in the Illinoian biotic
province (Dice, 1943). Natural vegetation is bluestem-grama
prairie, and mean annual precipitation is 71 cm (Eshelman, 1975;
Self, 1978).

The species composition and paleoecology of the White Rock
l.f. are similar to those of the Dixon l.f. of Kingman County. The
evolutionary grade of the fossils from White Rock suggests an
age younger than that of Sand Draw and older than Borchers
(Eshelman, 1975). Eshelman (1975) suggested a pre-Nebraskan
age. He thought that the fauna “accumulated under the influence
of alpine glaciation in the mountains to the west and the early
phases of climatic deterioration, which ultimately resulted in the
Nebraskan continental glaciation.”

The genus Blarina is represented by one specimen
(UMMP V60594) from locality UM-K1 -66 (NW »/4,
SE */4, NE 'A sec. 3, T2S, R5W). The specimen is a
fragment of right dentary with m 1 . Eshelman (1975)
referred the dentary to the species Blarina aff. car-
olinensis based on similarities to Holocene speci-
mens from southeastern Kansas [=B. hylophaga ].
He noted, however, a stubbier ascending ramus and
larger posterointernal fossa. For the present study,
measurements 19, 20, and 25 (2.3, 1 .6, and 4.0) were
taken and were compared with means derived from
the reference samples of Holocene specimens. The
measurements of the molar from White Rock are
most similar to the means of samples of modern B.
brevicauda ( talpoides semispecies). The fossil ten-
tatively is assigned to that species.

Irvingtonian

Conard Fissure

Fossils of the genus Blarina were recovered from two sites in
Newton County, Arkansas. Conard Fissure, located approxi-
mately 6 km west of Willcockson (SE 'A, NW 'A sec. 34, T17N,
R21W), is the older of the two faunas.

Birds, herptiles, and mammals are present in the Conard Fis-
sure l.f. The first study of the mammalian fauna was that by
Brown (1908). He identified 51 species from the site, including
24 extinct species and subspecies. Later revisions of Brown’s
work revealed that some of his new names were assigned incor-
rectly and approximately 44 species are now recognized from
Conard (Kurten and Anderson, 1980). Gidley and Gazin (1938)
compared the Conard l.f. with those of Cumberland Cave and
Port Kennedy. Oesch (1967) and Saunders (1977) compared the
fauna with others from the Ozark Plateau. Graham (1972) re-
viewed Blarina and other small mammals from Conard.

Brown (1908) regarded the fauna as late Wisconsinan, but
Conard Fissure now is recognized as an Irvingtonian fauna. Hib-
bard ( 1958) considered it Illinoian, but Graham (1972) suggested
that the fauna is Kansan.

The biostratigraphy of the site was described by Brown (1908)
and Graham (1972). The latter considered the fauna one bio-
stratigraphic unit and correlative with Cudahy. Some remains
represent inhabitants of the cave. Deposition of other remains
probably was due to activities of owls and other predators (Brown,
1908). The fauna suggests more equable climate at the time of
deposition; the local topography would have permitted a variety
of mesic and xeric habitats in the immediate area (Graham,
1972).

Newton County is part of the Ozark Mountains region of north-
western Arkansas. Dice (1943) included the area in the Caro-
linian biotic province. The natural vegetation of the Ozarks con-
sists mainly of oak-hickory forest, although red cedar glades,
stands of short-leaf pine, beech-maple forest, and upland prairies
also are present (Sealander, 1979). Mean annual precipitation in
Newton County ranges from 1 12 to 142 cm (NOAA, 1974).

Brown ( 1 908) assigned the Blarina of Conard Fis-
sure to a new subspecies, B. b. ozarkensis. Approx-
imately 550 specimens were referred to this taxon.
Diagnostic characters listed by Brown were antero-
posterior compression of the fourth incisor, reduc-
tion of the last unicuspid, reduction of the last lower
molar, and absent or reduced angle on the lower
border of the mandible. The interorbital region and
the posterior palatine foramina were described as
wider, and the heel of the last lower molar smaller,
than in living B. brevicauda. The subspecies differed
similarly from “i?. simplicidens ” of Port Kennedy
Cave.

Graham (1972) compared material from Conard
Fissure (paratypes and additional specimens) with
fossils from Cave Without a Name, Jinglebob, Mt.
Scott, and Cumberland, Meyer, Peccary, and Wil-
lard caves. Fossils also were compared with Holo-
cene specimens from eight states. Based on these
comparisons, he elevated ozarkensis to specific rank.
Diagnostic characters listed by Graham are “pro-
portionally larger and more bulbous cingula of 14,
P2, and lower molars and unicuspids, reduced P3,
reduced talonid of m3, and the mandibular condyle
shape.” Size of B. ozarkensis was noted as inter-
mediate between that of B. b. brevicauda and B. b.
kirtlandi. Graham (1972) and Graham and Semken
(1976) observed extensive variability in specimens
from Conard Fissure and mensural overlap with
Holocene samples, but the differences between spec-
imens from Conard Fissure and Holocene speci-
mens were thought sufficient to merit specific rank.

For the present study, 20 maxillaries and 3 1 den-
taries (KU 58676-77; SUI 35672, 35674-79, 35681,
35684-85, 35689, and 35691-92) were examined.

1984

JONES ET AL.- B LARINA SYSTEM ATICS

Fig. 5.— Canonical analysis of measurements 3-10 of upper teeth
of specimens of Blarina from Conard Fissure. Ellipses enclose
variation in samples of Flolocene taxa of Blarina (see Fig. 3).

The sample appeared homogeneous. We observed
extensive variability, but the degree of variability
seemed comparable to that seen in other Pleistocene
faunas. No morphological differences were observed
that would distinguish these fossils from modern B.
brevicauda.

Morphometric features of ozarkensis are similar
to those observed in Holocene specimens from Iowa
and Missouri (Moncrief et al., 1982). Those speci-
mens exhibit intergradation between the nominal
subspecies B. b. brevicauda and B. b. kirtlandi , and
it is likely that ozarkensis also represents intergra-
dation between larger shrews of the brevicauda
semispecies and smaller shrews of the talpoides
semispecies. This would explain the observation by
Graham and Semken (1976) that “The size of Bla-
rina ozarkensis is intermediate between that of brev-
icauda and kirtlandi .” Pending completion of a sys-
tematic review of the subspecies of B. brevicauda,
ozarkensis is retained herein as a separate subspecies
within the talpoides semispecies of B. brevicauda.

Results of analyses of 15 maxillaries and 21 den-
taries (of which two observations were hidden) are
shown in Figs. 5 and 6. Canonical analyses and com-
parisons with measurements of Holocene samples
(Table 1) indicate that fossils from Conard Fissure
are most similar to Holocene B. brevicauda ( tal-
poides semispecies).

Coleman II A

Some of the most abundant assemblages of Pleistocene animals
have been recovered from Florida. Fossil remains of the genus
Blarina have been reported from a dozen principal Pleistocene

1.60

V2

Fig. 6.— Canonical analysis of measurements 19-24 of lower teeth
of specimens of Blarina from Conard Fissure. Ellipses enclose
variation in samples of Holocene taxa of Blarina (see Fig. 4).

sites in that state (Appendix 1 ). Two of these sites are considered
Irvingtonian.

Chronological correlation of the Pleistocene faunas of Florida
is subject to several major difficulties. Many of the deposits are
located in caves and sinkholes formed in limestone, which is soft
and soluble. As Auffenberg (1963) pointed out, “Each sinkhole
is a unit in itself, containing a fauna which may or may not be
related to that in the next sinkhole." Other deposits, in coastal
stream beds and marshes, are subject to erosion. Fossils often
are discovered through operations such as mining, which usually
disturb the deposit, and early field records often were inaccurate
(Simpson, 1929ft). Many of the fossil assemblages appear to be
heterochronic. Simpson (1929ft) recognized only four faunas that
he considered useful in correlation. To this list, Bader (1957)
added the localities at Arredondo and Reddick. Recent advances
in stratigraphy, and the discovery of early and middle Pleistocene
sites in Florida, have helped clarify some of the difficulties in
correlation (Webb, 1974).

The Coleman IIA fauna of central Florida was recovered from
a quarry near Coleman in Sumter County (SE '4, NW ‘A sec. 7,
T20S, R23E). The site, which later was destroyed by mining, was
a filled sinkhole. Mammalian remains probably represent bats
and the prey of owls that roosted at the site and larger species
that fell into the open sinkhole (Martin, 1974). Martin (1974)
listed thirty-eight species of mammals from Coleman IIA. The
list includes 10 extinct species and four extant species no longer
present in Florida (Conepatus sp., Felis onca, Lepus alleni, and
Erethizon dorsatum).

Martin (1974) and Webb (1974) considered the Coleman IIA
fauna to date from the late Illinoian or early Kansan. Deposition
probably occurred in less than 1,000 years, but the fossils might
not represent sympatric species (Martin, 1974). The presence of
Cryptotis parva, Peromyscus floridanus, Lepus alleni, and Platy-
gonus sp. suggest that a more xeric savanna habitat was present
during the period of deposition, although other species (such as
Neofiber alleni, Sciurus carolinensis, and Mylohyus sp.) imply
mesic habitat (Martin, 1974; Webb, 1974; Frazier, 1977). Martin
(1974) suggested that the Coleman fauna might represent a period
of transition from xeric savanna to mesic forest.

Presently, the climate of Florida is humid and subtropical.

82

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Mean annual precipitation north of the everglades ranges from
132 to 142 cm (NOAA, 1974). Dice (1943) included the state in
the Austroriparian biotic province. The climax vegetation north
of the Everglades is a hardwood forest of pine, oak, magnolia,
and sweetgum (Bailey, 1978).

One mandible with i and ml -2 (UF 1 1626), re-
ferred to the species B. brevicauda, was reported by
Martin (1974). Measurements 19 through 22 and
25 (1.7, 1 .2, 1 .6, 1 .0, and 3. 1 ) of this specimen were
taken for the present study. These measurements
fall within the range of those reported for modern
B. c. peninsulae (sample 2 in Table 1), and super-
ficially the fossil resembles a Holocene specimen of
that taxon (UF 2532 from Alachua County) with
which it was compared. We tentatively assign the
specimen to B. carolinensis pending completion of
a study of the systematic relationships of peninsulae.

Inglis IA

Inglis IA is a fissure deposit located on the western coast of
Florida in Citrus County. The fauna dates from a dry glacial
interval in the very late Blancan or early Irvingtonian (Webb,
1974).

Webb (1974) listed 37 mammalian species from this site. Sev-
enteen extinct species were present, including Megalonyx jeffer-
soni. Eremotherium rusconii, Glossotherium chapadmalense,
Smilodon gracilis. Chasmaporthetes ossifragus. and Equus sp.
Inglis IA is the earliest record of Hydrochoerus holmesv, the oc-
currence of Hydrochoerus and that of other members of the fauna
suggest the presence of coastal savanna and aquatic habitats and
warm winters (Webb, 1974; Kurten and Anderson, 1980).

Webb (1974) noted the presence of Blarina brev-
icauda in the Inglis IA fauna. For the present study,
hve dentaries (UF, uncatalogued material) were ex-
amined. These specimens are similar in shape and
size to modern B. carolinensis of Florida. All molars
are present in three specimens, and measurements
19 through 26 were taken (1.8, 1.3, 1.6, 1.1, 1.2,
0.8, 3.2, 2.0; 1.8, 1.2, 1.6, 1.0, 1.2, 0.9, 3.2, 2.1; 1.7,

1.2, 1.6, 1.0, 1 .2, 0.9, 3.2, 2. 1). On the two remaining
fossils, measurements 19 through 22 and 26 (1.8,

1.3, 1.6, 1.1, 2.0) and 19, 20, and 25 (1.7, 1.2, 3.2)
were taken. The measurements fall within the range
of variation of groups of Holocene B. carolinensis
(Table 1) and are most similar to B. c. peninsulae
(sample 2). We think the fossils from Inglis pertain
to the taxon peninsulae.

Cudahy

Two specimens of Blarina were identified from two sites from
which the Cudahy l.f. has been recovered. The sites are UM-K3-
71 (sec. 7, T32S, R28W) and the Sunbrite ash mine (KU loc. 17,
sec. 26, T32S, R28W). Both sites are located in Meade County
in southwestern Kansas.

The Cudahy l.f. consists of molluscs, fishes, herptiles, birds,
and mammals. Thirty-seven species of mammals (largely insec-
tivores and rodents) have been identified, of which 1 8 are extinct
(Paulson, 1961; Hibbard, 1944, 1970).

Remains from Cudahy were deposited immediately below the
Pearlette type O ash, suggesting a late Kansan age. The Cudahy
was one of five faunas listed by Hibbard (1970) as evidence that
local climate was cooler during the Kansan than during the Ne-
braskan.

Invertebrates and vertebrates of this fauna, particularly the
large number of shrews and microtines, indicate the presence of
a more boreal fauna at the time of deposition (Leonard, 1950;
Blair, 1958; Brattstrom, 1967; Graham, 1976; Hibbard, 1944,
1970). Accumulation occurred in marshy depressions or slow-
moving streams. Pollen associated with the ash is poorly pre-
served (Kapp, 1970).

Presently, Meade County is included in the High Plains phys-
iographic province. Natural vegetation of the area near the Cu-
dahy sites consists of bluestem-grama prairie, with floodplain
forest and savanna (Populus-Salix) along Crooked Creek. Cli-
mate is semiarid; annual water loss through evaporation exceeds
annual precipitation. Mean annual precipitation is 41-51 cm
(Kuchler, 1974; NOAA, 1974; Self, 1978).

A dentary with m 1-3 (UMMP V6 1 263) from UM-
K3-71 and a broken dentary with ml and m2
(UMMP 36802) pertain to the genus Blarina. Mea-
surements 1 9 through 24 (2.2, 1.4, 1.9, 1.2, 1.6, and
1.0) and 19 through 22 (2.1, 1.4, 2.0, and 1.3), re-
spectively, were taken and compared with those of
reference samples of Holocene specimens (Table 1).
The measurements of the fossils are most similar to
those of modern specimens of B. brevicauda ( tal -
poides semispecies), and the two specimens from
Cudahy are referred to that species.

Kanopolis

The Kanopolis l.f. was recovered from a gravel pit north of
Kanopolis, Ellsworth County, Kansas (SW ‘A, NE ‘A sec. 25,
T1 5S, R8W). Kanopolis is the largest Pleistocene fish fauna known
from the Great Plains (Neff, 1975). In addition to 15 species of
fishes, pelecypods, ostracods, 2 1 species of herptiles, and 34 species
of mammals have been identified (Holman, 1972; Hibbard et al.,
1978). Fourteen mammalian species are extinct; all other ver-
tebrates are extant.

The Kanopolis l.f. is considered Yarmouthian. A warm, moist
climate is suggested by the fauna. Permanent stream, marsh, and
forest communities were present, and savannah and grassland
were nearby. Sympatry of extant vertebrates occurs east of the
site (Holman, 1972; Neff, 1975; Frazier, 1977; Hibbard et al.,
1978).

Ellsworth County presently is part of the Smoky Hills sub-
province of the Great Plains. The natural vegetation of the county
is bluestem-grama prairie and transitional bluestem-grama and
bluestem prairie. Mean annual precipitation is 61 to 71 cm
(Kuchler, 1974; NOAA, 1974; Self, 1978).

Hibbard et al. (1978) referred four specimens to
Blarina sp. These specimens are a left dentary with
complete dentition (UMMP V6 1000); two dentaries

1984

JONES ET AL .-BLARINA SYSTEM ATICS

83

with i and m 1-3 (V6041 5, V606 1 2); and an isolated
Ml (V60412). The possible co-occurrence of more
than one species was suggested.

All specimens were examined for the present study.
Measurements 7 through 10 (2.3, 2.1, 2.4, and 2.6)
of the upper molar were taken and compared with
those of Holocene specimens (Table 1). The mea-
surements of the fossil are most similar to those of
samples 5, 12, and 18 of B. brevicauda-, the fossil
might represent the subspecies B. b. brevicauda.
Measurements 19 through 26 of the three remaining
specimens (2.2, 1.6, 1.8, 1.2, 1.4, 1.0, 4.0, 2.5; 1.9,
1.4, 1.8, 1.2, 1.5, 1.0, 4.0, 2.4; 2.2, 1.5, 1.9, 1.3,
1.6, 1.0, 4.0, and 2.5) were taken. The means of
these measurements are 2.1, 1.5, 1.8, 1.2, 1.5, 1.0,
4.0, and 2.5, respectively. Measurements of the
specimens from Kanopolis are most similar to those
of Holocene samples of the talpoides semispecies of
B. brevicauda (Table 1). These analyses lend cre-
dence to the suggestion by Hibbard et al. (1978) that
more than one taxon might be present in the Kan-
opolis l.f., but all specimens apparently pertain to
the species B. brevicauda.

Kentuck

Fossils were found in outcrops along Kentuck Creek in Mc-
Pherson County, Kansas (sec. 13, T18S, R3W). Accumulation
occurred in a stream deposit that cut through an older deposit
of Pearlette-like ash (Hibbard. 1952).

The Kentuck l.f. consists of fishes, herptiles, birds, and mam-
mals. Birds and snakes were discussed by Galbreath (1955) and
Brattstrom (1967). Mammalian faunal lists were published by
Hibbard (1952) and Semken (1966). Some of the rodents were
examined by Nelson and Semken (1970), Martin (1975), and
Van der Meulen (1978). Twenty species of mammals were iden-
tified, of which eight are extinct (Hibbard, 1952).

Stream-border and grassland communities are indicated by the
fauna. Hibbard (1952) originally designated this group of ver-
tebrates as the Kentuck assemblage because the fossils were thought
to represent different communities of Kansan and Yarmouthian
age. Semken (1966) and Brattstrom (1967) concurred with this
interpretation. Semken and others (Zakrzewski, 1975a) now sug-
gest that the remains are a local fauna deposited during more
equable conditions in the Kansan glacial. Van der Meulen (1978)
considered the fauna Aftonian.

Most of McPherson County now is part of the Great Bend
Prairie physiographic region (Self. 1978). Natural vegetation is
mostly bluestem prairie; mean annual precipitation is 61 to 81
cm (Kuchler, 1974; NOAA, 1974).

The genus Blarina is represented in this fauna by
a lower incisor (UMMP V5 1248). The species can-
not be identified.

Rezabek

Fossils of the Rezabek l.f. were recovered from the Rezabek
gravel pit in southwestern Lincoln County, Kansas (sec. 20, T 1 3S,

R10W). The Rezabek l.f. may pertain to the late Irvingtonian or
early Rancholabrean. Hibbard (1970) placed it among the early
Illinoian faunas of the Great Plains. Hibbard and Dalquest (1973)
considered the Rezabek Yarmouthian based on sympatry of Neo-
fiber and Ondatra. Zakrzewski (1975a) suggested the Rezabek
should be retained in the Illinoian, based on the equability model
of Pleistocene climate (as discussed by Graham, 1972). The pres-
ence of microtines assignable to extant species (except Neofiber)
suggests the Rezabek l.f. is Rancholabrean (Zakrzewski, personal
communication).

Because of the abundant remains of fishes, the fauna was thought
to have accumulated in a stream-laid deposit (Hibbard, 1943).
Climate was considered temperate. Early Illinoian faunas of the
Great Plains suggest a moister climate, with cooler summers,
than is present in the region today; a shift toward slightly warmer
winters apparently occurred during the late Illinoian (Hibbard,
1970).

Lincoln County is located in the Smoky Hills subprovince of
the Great Plains. This subprovince is characterized by rolling
hills underlain by Cretaceous rock (Self, 1978). Natural vegeta-
tion of the county consisted of bluestem-grama prairie, floodplain
vegetation along the Saline River, and transitional vegetation
between mixed (bluestem-grama) prairie and bluestem prairie.
Climate is of the temperate continental type; normal annual pre-
cipitation is 66 to 71 cm (Self, 1978).

The Rezabek l.f. consists of molluscs, fishes, herptiles, birds,
and mammals. Fossils of mammals represent the orders Insec-
tivora, Lagomorpha, Rodentia, Perissodactyla, and Artiodactyla.
Two nominal species, Blarina fossilis and Neofiber leonardi, were
described and named in the original account of the fauna (Hib-
bard, 1943).

The genus Blarina is represented in the Rezabek
l.f. by a fragmentary ramus (KU 6675). The speci-
men, which bears one molar, is the holotype of Bla-
rina fossilis. It was distinguished from B. simplici-
dens Cope and B. ozarkensis Brown by its larger
talonid on m3. It was differentiated from modern
species on the basis of size (Hibbard, 1943). Later,
Hibbard (1957) wrote that B. fossilis, B. simplici-
dens, and B. brevicauda could not be distinguished,
and both fossil taxa were relegated to subspecific
rank. Dalquest ( 1 965) assigned three specimens from
the Howard Ranch l.f. to B. b. fossilis because of
their large size. Graham and Semken (1976) saw no
difference between specimens from Howard Ranch
and Rezabek and modern specimens of B. brevi-
cauda. They placed fossilis in synonomy with brev-
icauda.

Measurements 23 through 25 (1.6, 1.2, and 3.7,
respectively) of the specimen from Rezabek were
compared with means of those measurements from
Holocene samples (Table 1). Measurements 23 and
24 of the fossil were most similar to means of the
sample of B. b. brevicauda from Nebraska (sample
12); they were also within the range of measure-
ments for the sample of brevicauda from Iowa (sam-
ple 5). Measurement 25 seems somewhat small for

84

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

any of the samples of brevicauda. Based on the anal-
ysis of the first two measurements and the earlier
work by Graham and Semken (1976), the Rezabek
specimen herein is considered a representative of
the species brevicauda, and most closely resembles
Holocene B. b. brevicauda.

Wathena

The Wathena l.f. was recovered south of Wathena (sec. 33,
T3S, R22E), Doniphan County, Kansas. Remains were found in
lake deposits below the Kansan Nickerson till. Pleistocene de-
posits in this area were described by Bayne (1969).

The fauna was described in an unpublished thesis (Einsohn,
1971). Molluscs, fishes, herptiles, birds, and mammals were pres-
ent. Einsohn (1971) listed 11 mammalian species representing
the families Soricidae, Sciuridae, Geomyidae, Cricetidae, and
Zapodidae. Van der Meulen (1978) discussed Microtus (Allo-
phaiomys) sp. of this fauna. Additional remains of unidentified
rabbits and rodents are housed at the University of Michigan.
The fauna is similar to that of Kentuck l.f. (Zakrzewski, 1975a).

The Wathena l.f. is considered Irvingtonian. Schultz et al. (1978)
placed the fauna in the early Irvingtonian (but considered it more
recent than the faunas of Dixon and White Rock), based on the
presence of “ Allophaiomys.” Van der Meulen (1978) considered
the Wathena Aftonian, but Einsohn (1971) thought that the Wa-
thena was more recent than Cudahy.

Doniphan County was glaciated at least twice (Nebraskan and
Kansan). Presently the county has a temperate continental cli-
mate and receives 81 to 91 cm of precipitation annually. The
natural vegetation of the area is oak-hickory forest, with flood-
plain forest and savanna (Populus-Salix) along the Missouri Riv-
er (Bayne, 1969; Kuchler, 1974; NOAA, 1974; Self, 1978). Little
can be determined regarding climate during the period of de-
position of fossils (Einsohn, 1971).

Shrews are represented by a left dentary fragment
with m2-3 (UMMP V60522) and by an isolated M 1
(UMMP V61 179). The shape and size of both spec-
imens indicate that they pertain to the genus Bla-
rina. Measurements 7 through 10 of the Ml (2.0,
1.9, 2.0, and 2.2) resemble those of Holocene B.
carolinensis. Measurements 2 1 through 26 ( 1 .5, 1.1,
1.3, 1.0, 3.3, and 2.1) of the lower molars and den-
tary also correspond to the means of Holocene sam-
ples of B. carolinensis (Table 1), although careful
examination of the dentary revealed some similarity
to Holocene specimens of B. brevicauda ( talpoides
semispecies). The fragmented condition of the fos-
sils of the Wathena l.f., and the lack of material of
comparable age in eastern Kansas, pose major prob-
lems in the analysis of the Wathena material. We
tentatively assign both specimens to the species B.
carolinensis.

Cumberland Cave

Cumberland Cave is a fissure located 6.4 km northwest of
Cumberland, Allegany County, Maryland, at 39°41'/2'N and

78°47'/4,W (Gidley and Gazin, 1938). The cave was long known
to local residents (Nicholas, 1954). Excavation began after ex-
posure of fossils by a railway cut in 1912. Originally there were
two entrances. One opening ran 61 m and led into the cave 31
m below the ridgetop. This area was destroyed by quarrying. The
second entrance was a sinkhole that led from the top of the ridge
through several chambers to the main room of the cave, about
31 m below the ridgetop (Nicholas, 1954). Fossils generally were
broken but showed no signs of being worn by water. They were
found in unstratified cave clays and breccias (Gidley and Gazin,
1938). Gidley and Gazin suggested that the mode of accumula-
tion was similar to that which occurred at Conard Fissure, as
described by Brown (1908); that is, that materials were carried
into the cave by predators and that accumulation occurred grad-
ually over a long period of time.

Initial reports describing selected elements of the fauna were
produced by Gidley and Gazin; faunal lists were published later
(Gidley and Gazin, 1933, 1938). All remains were mammalian
except for isolated snake vertebrae and a fragment of the humerus
of a ruffed grouse (Bonasa umbellus). The latter was identical to
the modern form that currently inhabits the eastern United States
(Wetmore, 1927). A tooth initially was misidentified as that of
a crocodylid (Richmond, 1963). Eight of the 41 genera and 28
of the 46 species identified from the cave fauna are extinct (Gidley
and Gazin, 1938).

Gidley and Gazin (1933, 1938) recognized that the fauna of
Cumberland Cave included species that presently are allopatric
plus species still present in the area. A few boreal species (such
as the wolverine, Gulo luscus) were present. Western species,
including the thirteen-lined ground squirrel ( Spennophilus tri-
decemlineatus), also were present. The fauna now is considered
autochronic, with deposition having occurred under ecological
conditions not known in any one area today (Guilday, 1971).

The cave is located at an altitude of 255 m in a limestone ridge
in Wills Creek Valley (Nicholas, 1954). This area was included
near the northern boundary of the Carolinian biotic province by
Dice ( 1 943), and within the Appalachian physiographic province
(Abbe. 1900). Mean annual precipitation in this region is 91 to
101 cm (NOAA, 1974).

The age of the Cumberland Cave fauna is Irvingtonian. A
minimum date of 250,000 BP was obtained from racemization
of amino acids in a horse phalanx (Zakrzewski, 19756). Guilday
(1971) considered the fauna Illinoian. Van der Meulen (1978)
suggested a mid-Irvingtonian age based on his study of the Mi-
crotus and Pitymys of Cumberland Cave. Martin (1974) saw some
similarity with the (Irvingtonian) fauna from the Coleman IIA
deposit of Florida; he also noted that the Cumberland material
had not been reviewed for thirty years and considered additional
speculation hazardous. Zakrzewski (19756) pointed out that “If
the Cumberland Cave local fauna had been located on the Great
Plains, the association of Ondatra annectens, Neofiber, and Ato-
pomys, and the lack of Microtus pennsylvanicus would suggest a
pre-Ilhnoian age. Because of the geographic position of Cum-
berland and the lack of other pre-Wisconsin faunas in the area
with which it can be compared, the determination of the exact
age is moot.”

Gidley and Gazin ( 1 938) identified the short-tailed
shrews from Cumberland Cave as Blarina brevi-
cauda. The first specimens to be recovered consisted
of a rostral fragment, two maxillary fragments, and

1984

JONES ET AL .-BLARINA SYSTEM ATICS

85

Fig. 7. — Canonical analysis of measurements 3-10 of upper teeth
of specimens of Blarina from the Cumberland Cave l.f. Centroids
enclose variation in samples of Holocene taxa of Blarina (see
Fig. 3).

eight lower jaws. It was noted that these specimens
were “somewhat larger than average” in Holocene
specimens and resembled the Conard Fissure spec-
imen described by Brown (1908). Graham and
Semken (1976) believed that “three distinct, non-
overlapping populations” were represented that
could be equated to the modern brevicauda, kirt-
landi, and carolinensis. The coexistence of the three
phena was thought to represent a full or early late
glacial climate.

We examined 40 maxillary and dentary fragments
from Cumberland Cave (CM 8032, 20438, 20441-

2, 20445-7, 20449, 2045 1-3, 20457, 20459, 20462-

3, 24218-9, USNM 7779, 12460-2, 12464-7). Of
the nine maxillaries, six were sufficiently complete
for analysis. These plotted within the range of vari-
ation of Holocene specimens of the smaller subspe-
cies ( talpoides semispecies) of B. brevicauda (Fig. 7).
Sixteen dentaries fell within the ranges of variation
for the smaller subspecies of B. brevicauda and for
B. carolinensis (Fig. 8). Univariate statistics (CUM-
BER) are shown in Table 2. Specimens that could
not be studied by canonical analysis had measure-
ments similar to those for the specimens shown in
the two figures. The Cumberland Cave l.f. appar-
ently contains remains of two species of Blarina.
The largest specimens pertain to B. brevicauda ( tal-
poides semispecies), whereas the smallest individ-
uals pertain to the species B. carolinensis.

Angus

Fossils of the Angus l.f. were recovered from a quarry 2.4 km
southwest of Angus, Nuckolls County, Nebraska (SW ‘A, NE 'A

1.60

V2

Fig. 8.— Canonical analysis of measurements 19-24 of lower teeth
of specimens of Blarina from the Cumberland Cave l.f. Centroids
enclose variation in samples of Holocene taxa of Blarina (see
Fig. 4).

sec. 33, T4N, R6W). The fossils were found below the Illinoian
Loveland Loess and the Yarmouthian paleosol and were consid-
ered late Yarmouthian in age (Schultz and Tanner, 1957; Schultz
and Martin, 1970). In the chronology by Schultz et al. (1978),
the Angus l.f. was considered younger than the local faunas of
Port Kennedy, Cudahy, Conard Fissure, Java, and White Rock,
from which specimens of Blarina have been reported.

Fossils recovered from the Angus l.f. include remains of fishes,
amphibians, reptiles, birds, and mammals (Schultz and Tanner,
1957). Schultz and Martin (1970) published a preliminary list of
mammalian species from the site. Remains of extinct species
include those of an unidentified sloth, Cynomys niobrarius, Cas-
t oroides sp.. Ondatra nebracensis, Canis dims nebrascensis,
Mammuthus ( Archidiskodon ) imperator, an unidentified mas-
todont, Mylohyus browni [ =nasutus ], Platygonus sp., Camelops
kansanus, Capromeryxfurcifer, IStockoceros sp., Equus excelsus,
and E. calobatus. To this list Kurten and Anderson ( 1 980) added
Castor accessor and Mammuthus meridionalis, and referred some
of the canid material to Canis armbrusteri. Extant species re-
ported from the site include Sorex cinereus, Blarina brevicauda,
Sca/opus sp., an unidentified bat, Spermophilus tridecemlineatus.
S. franklinii, S. cf. richardsoni, Geomys bursarius, Dipodomys
sp., Perognathus sp., P. hispidus, Castor sp., Onychomys sp.,
Reithrodontomys sp., Peromyscus sp., Neotoma sp., Synaptomys
sp.. Clethrionomys gapperi, Microtus ochrogaster, M. pennsyl-
vanicus, Zapus hudsonius, Lepus sp., Sylvilagus sp., Canis cf.
latrans, Vulpes velox, Mustela frenata, Taxidea cf. tax us. and
Odocoileus sheridanus [ =virginianus ] (Schultz and Martin, 1970;
Kurten and Anderson, 1980). Schultz and Martin (1970) noted
that most of the smaller mammals of the Angus l.f. have modern
descendants in the area today, which suggested that climatic con-
ditions during the period of deposition were no more severe than
those of today.

Presently the area is included in the Illinoian biotic province
(Dice, 1943). The region receives a mean annual precipitation of
60 to 71 cm (NOAA, 1974).

Two maxillary fragments and three mandibles
(KU 33638, 33639, 33658, 33660, and 33664) of

86

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Blarina were examined. Measurements 3 through 6
of two P4’s (2.5, 2.0, 1.3, 2.4; 2.4, 1.8, 1.3, 2.2) and
measurements 19 through 22 (2.3, 1.5, 2.1, 1.3), 23
through 25 (1.7, 1.1, 4.6), and 21 through 26 (1.9,
1.2, 1.5, 1 .0, 4. 1 , 2.5) of lower molars and dentaries
were taken. Measurements of these specimens were
compared with means derived for those measure-
ments in Holocene samples (Table 1). Measure-
ments of the fossils were slightly smaller than the
means of the modern B. b. brevicauda from Ne-
braska (sample 12), and they were comparable in
size to the smaller talpoides semispecies (samples 4,
6, 11, 18, 13, 19, and 24). Fossils of Blarina from
Angus were similar to those of the Cumberland Cave
l.f, which also was considered an Irvingtonian fau-
na. The specimens pertain to the species B. brevi-
cauda.

Bartek Brothers Quarry

The Bartek Brothers Fossil Quarry is located approximately 4
km west of Weston (NE *4 sec. 12, T14N, R5E) in Saunders
County, Nebraska. Fossils were taken from the Sappa Formation.
The age of the deposit was considered post-Kansan by Schultz
and Tanner (1957). Frankel (1963) thought the deposit probably
was Yarmouthian. Molluscs and plant material from the site
suggest the presence of a pond, probably surrounded by marshy
habitat. “The occurrence of fish, reptile, and rodent remains in-
dicates the presence of vertebrates from two distinct environ-
ments; one aquatic, the other, terrestrial. If the Microtinae are
Synaptomys. a marshy environment would be indicated” (Frank-
el, 1963). The presence offour species of gastropods that no longer
occur in the area might indicate a slightly cooler climate at the
time of deposition (Frankel, 1963). The date of the type Sappa
(1.2 million years BP) and the paleoecological interpretation by
Frankel imply a pre-Kansan age for the fauna (Zakrzewski, per-
sonal communication).

The region now is included in the Illinoian biotic province by
Dice (1943). Mean annual precipitation is approximately 61 to
71 cm (NOAA, 1974).

Seeds, molluscs, and vertebrates were recovered from the quar-
ry. Vertebrates include fishes, snakes, rodents (Microtinae), ground
sloth, mastodont, mammoth, horse, camel, llama, wapiti, and
bison (Schultz and Tanner, 1957; Frankel, 1963). Preservation
of vertebrate remains was poor.

Five specimens (USNM 33094-8) of Blarina from
the Bartek Brothers l.f. were examined. All were
fragments of dentaries with incomplete dentition;
isolated teeth also were recovered from the site.
Measurements 19 and 20 (2.2, 1.4) of an additional
m 1 are much smaller; whether they indicate the
presence of a second taxon of B. brevicauda (rep-
resenting the talpoides semispecies) is uncertain.

Berends

Fossils of the genus Blarina have been reported from two sites
in Oklahoma. Both sites, in adjacent counties in western Okla-
homa, are outside the present range of the genus. The region now

is part of the subhumid shortgrass region of the Great Plains
(Dice, 1943; Stephens, 1960). Modern climate is described by
Stephens (1960) and NOAA (1974).

The Berends l.f. was recovered from deposits in the Berends
Sand Draw on the Coy Berends Ranch in Beaver County (secs.
5 and 6, T5N, R28E). This locality is approximately 7.2 km
north and 1.6 km west of Gate (Rinker and Hibbard, 1952;
Taylor, 1954).

Mengel (1952), Smith (1954), Taylor (1954), Taylor and Hib-
bard (1955), and Brattstrom (1967) reported on the molluscs,
fishes, herptiles, and birds of the Berends l.f. Sixteen species of
mammals were reported by Rinker and Hibbard (1952), Starrett
(1956), and Hibbard (1963). Six extinct mammals (Paradipoides
stovalli, Peromyscus berendsensis n. sp.. Ondatra nebracensis,
Equus sp., Mammuthus cf. M. columbi. and Castoroides ohioen-
sis) were identified (Starrett, 1956; Kurten and Anderson, 1980).

Both the location of the Berends l.f. above the Pearlette Ash
and the composition of the fauna suggest an Illinoian age (Rinker
and Hibbard, 1952; Hibbard, 1953L 1958, 1960, 1970; Taylor,
1954). The Berends l.f. might be older than that of Doby Springs
(Jammot, 1972; Kurten and Anderson, 1980). The fauna and
associated pollen suggest deposition during moist conditions,
possibly with cooler climate. Sympatry of extant mammals occurs
in North Dakota (Starrett, 1956). Lake, marsh, and woodland
communities were present (Smith, 1954; Taylor, 1954; Starrett,
1956; Hibbard, 1960, 1970; Kapp. 1970).

Smith (1954) and Starrett (1956) recorded the re-
covery of Blarina from the Berends deposit. The
genus is represented by a left Ml (UMMP V60549)
and by a fragment of a right ramus with ml -m2
(UMMP 31790). Starrett (1956) tentatively identi-
fied the fossils as B. brevicauda. Hibbard (1963)
initially referred the specimens to the subspecies B.
b. brevicauda ; later (1970) he referred them to B. b.
fossil is.

Measurements 7 through 10 (2.3, 2.2, 2.5, and
2.7) and 19 through 22 (2.2, 1.4, 1.8, and 1.2) were
taken. These measurements were compared with
those of the reference samples (Table 1). Measure-
ments of the Ml (V60549) fall within the range of
measurements of samples of B. brevicauda, and are
closest to the means of measurements 7 through 10
recorded for sample 12 (B. b. brevicauda). This spec-
imen is referred to B. b. brevicauda. Measurements
19 through 22 of the other specimen from Berends
are somewhat ambiguous. They fall within the range
of measurements of samples of all three Holocene
species but are outside the range of measurements
of B. b. brevicauda. The second specimen is referred
tentatively to B. brevicauda. Whether two taxa were
present in the fauna at the time of deposition cannot
be ascertained at this time.

Hanover Quarry Fissure

Remains of Blarina were recovered from the Hanover Quarry
Fissure in Adams County, southeastern Pennsylvania. The fauna

1984

JONES ET AL. — BLARINA SYSTEM ATICS

87

Table 5. — Univariate statistics for specimens of Blarina caroli-
nensis from Hanover Quarry Fissure, Adams County, Pennsyl-
vania, for measurements 19-22 of the lower teeth.

Variable

N

Mean

2 SE

Range

cv

19

6

1.80

0.05

1. 7-1.9

3.33

20

6

1.22

0.07

1. 1-1.3

6.56

21

8

1.55

0.06

1.5-1. 7

5.16

22

8

1.01

0.03

1.0-1. 1

3.96

from Hanover Quarry is considered Yarmouthian in age and
probably older than that of Cumberland Cave (Guilday, in litt.,
1982). The Blarina from this fissure are unique in that they are
smaller than remains from most other Appalachian Pleistocene
faunas. Fossils from Hanover are similar in size to some of the
smallest specimens from Cumberland Cave. Most of the fossils
from Hanover consist of edentulous, fragmented dentaries. Other
specimens consist of isolated teeth and dentaries with incomplete
dentition.

Measurements of nine specimens (CM 41213,
41406) were taken. One Ml was examined; mea-
surements 7 through 10(1.9, 1.8, 1.9, 2.0) fall within
the range of variation of samples of Holocene B.
carolinensis (Table 1). Univariate statistics of mea-
surements 1 9 through 22 of the lower molars of eight
dentaries are shown in Table 5. These measure-
ments compare most favorably with those of Ho-
locene B. carolinensis, particularly the small indi-
viduals from Florida and Louisiana (groups 2 and
23, Table 1). No differences in dental morphology
were observed that would merit assignment of the
Hanover material to a different species. The ap-
pearance of carolinensis- sized Blarina in this fauna
suggests temperate conditions during the period of
deposition.

Port Kennedy Cave

Port Kennedy Cave was located in Upper Merion Township,
Montgomery County, in southeastern Pennsylvania. The cave
was a fissure discovered in 1870 during quarrying of limestone.
The fissure later was destroyed by mining. Port Kennedy Cave
is one of the few Appalachian faunas considered to be pre-Wis-
consinan (Guilday, 1971). Its fauna pertains to the Irvingtonian
land mammal age, and generally has been considered Yarmouth-
ian (Hibbard, 1958; Hibbard and Dalquest, 1973) or Illinoian
(Guilday, 1971). Kurten and Anderson (1980), however, tenta-
tively referred the fauna to the Aftonian or early Kansan.

Preliminary investigations of the Port Kennedy l.f. were con-
ducted by Cope (1899). Preservation was poor and the status of
many of the taxa described by Cope is uncertain. Brown (1908)
listed a total of 56 mammalian species, 40 of which are extinct,
from the site. Among the extinct species are ground sloth, Schlos-
ser’s wolverine, gracile sabertooth, mastodont, and two species
of skunk (Kurten and Anderson, 1980). Similarities between the
local faunas of Port Kennedy, Conard Fissure, and Cumberland

Cave were discussed by Brown ( 1 908), Nicholas (1954), and Kur-
ten and Anderson (1980).

Presently, southeastern Pennsylvania is included in the Car-
olinian biotic province as described by Dice (1943). The area is
part of the Piedmont Plateau, characterized by rolling uplands,
low hills, and fertile valleys; mean annual precipitation is ap-
proximately 1 12 cm (NOAA, 1974).

The genus Blarina is represented in Port Kennedy
Cave by part of a left jaw. The specimen (ANSP
1 50) is the holotype of Blarina simplicidens, de-
scribed by Cope in 1899. “It is about the size of B.
brevicauda, and no especial characters can be found
to distinguish it from that species excepting the forms
of the first premolar and the last true molar ....
The crown of the first premolar is somewhat worn
and has an antero-posteriorly oval section. It does
not have the V form as in all the species figured by
Merriam [1895], and which seems to be common
to all the existing species. The last true molar is also
more simple than in the existing species, consisting
of a trigon only, and lacking the heel. The heel is
present in all the species of this genus, of Cryptotis,
Notiosorex, and Sorex, according to Merriam.”
Length of the molar series, length of m2, and depth
of the ramus at m2 were 5, 2.5, and 2 mm, respec-
tively (Cope, 1899).

Hibbard (1957) illustrated and compared the
specimen of B. simplicidens with B. gidleyi, B. fos-
si/is, and B. brevicauda. He saw no structural dif-
ferences between B. simplicidens, B. fossi/is, and B.
brevicauda ; the special form of the premolar noted
by Cope was attributed to wear. Consequently, sim-
plicidens and fossi/is were relegated to the rank of
subspecies of B. brevicauda (Hibbard, 1957) and
later were placed in synonomy with B. brevicauda.

We examined the specimen but did not measure
the teeth. No differences were found that would dis-
tinguish the specimen from B. brevicauda and we
concur with that assignment, as a representative of
the talpoides semispecies.

Java

The Java local fauna was collected in eastern Walworth Coun-
ty, South Dakota, at 45°26'N, 99°52'W (NE >/•> sec 26, T123N,
R75W). Fossils were deposited in a paleostream channel 35 km
east of the Missouri River trench (Martin, 1973ft).

Martin (1973a, 1973ft) described the Java l.f., which consists
of molluscs, fishes, birds, herptiles, and mammals. Twelve fam-
ilies of mammals are represented. The fauna indicates the pres-
ence of woodland or shrubland, grassland, and aquatic com-
munities. Habitat and climate probably were similar to those of
today except that climate might have been somewhat cooler
(Martin, 1973ft). Presently the area is located near the western
boundary of the Illinoian biotic province of Dice (1943).

88

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

The Java l.f. is one of the northernmost Irvingtonian sites
known in North America. A radiocarbon date (>30,000 BP) was
obtained from a sample of molluscs (Martin, 1973a). The rodents
of the fauna suggest a Kansan age (Martin, 19736, 1975). The
Java l.f. was considered older than those of Conard Fissure, Cu-
dahy, and Cumberland Cave by Martin (19736), Schultz et al.
(1978), and Van der Meulen (1978).

The genus Blarina is represented by a single P4
(R. A. Martin, private collection). Measurements 3
through 6 are 1.9, 1.6, 1.1, and 1.7. These mea-
surements are most similar to those of reference
samples of B. caro/inensis (Table 1). The tooth from
Java is referred tentatively to that species. Addi-
tional speculation is premature because of the small
sample size and the age and location of this fauna.

Vera

The Vera l.f. was recovered from three sites in northern Texas.
The three sites are UM-T1-56 (SW '4 sec. 152, Baylor County),
UM-T1-57 (N ‘4 sec. 101, Knox County), and UM-T1-58 (SW
'4 SE ‘4 sec. 1 10, Knox County) (Getz and Hibbard, 1965). Re-
mains were recovered in the Seymour formation immediately
below the Pearlette Ash. Dalquest (1977) estimated the age of
the fauna as ca. 600,000 BP. The fauna is similar to that of
Cudahy.

Molluscs, fishes, herptiles, birds, and mammals are included
in the Vera l.f. Most of the molluscan species occur in sympatry
in eastern South Dakota and southwestern Minnesota (Getz and
Hibbard, 1965). Thirteen species of mammals were identified,
of which five ( Geomys tobinensis, Peromyscus cragini, Ondatra
annectens, Microtus paroperarius, and M. llanensis) are extinct
(Hibbard and Dalquest, 1966). Of the extant species, Blarina
brevicauda and Onychomys sp. no longer occur in Baylor and
Knox counties (Davis, 1978).

The species composition of the Vera l.f. suggests a climate with
cooler, moister summers. Streams and wooded floodplains were
present (Getz and Hibbard, 1965; Hibbard and Dalquest, 1966;
Hibbard, 1970). Presently, the climate of Baylor and Knox coun-
ties is semiarid, with an annual precipitation of 58 to 66 cm (Getz
and Hibbard, 1965; NOAA, 1974).

Hibbard and Dalquest (1966) identified seven
specimens of Blarina from UM-T1-57 (UMMP
45834, 45835, 45850-3, and 45855) and two from
UM-T1 -58 (UMMP 39803 and 45709). These spec-
imens were referred to Blarina cf. B. brevicauda and
a greater range of individual variation than in Ho-
locene faunas was noted. All specimens were ex-
amined for the present study. Specimen 45834 con-
sists of two edentulous dentary fragments; all other
specimens are isolated teeth. Measurements 3
through 6 (2.3, 1.9, 1.4, 2.3) and 3, 5, and 6 (2.2,
1.3, 2.0) of two P4s, measurements 12 through 14
(1.7, 2.0, 1.9) of a M2, and measurements 19 and
20 (2.0 and 1.4) of a ml were taken. These mea-
surements (of specimens 4585 1-3 and 45709) were
compared with measurements of reference samples

of Holocene specimens (Table 1). Measurements of
the specimens from Vera are most similar to the
means of B. brevicauda ; these fragmentary fossils
tentatively are assigned to that species.

Trout Cave

Trout Cave is a limestone cave located along the terrace of the
South Fork of the Potomac River, 4.8 km southwest of Franklin,
Pendleton County, West Virginia (Kurten and Anderson, 1980).
The upper strata of the deposit appear to be Wisconsinan. The
lower strata, separated from the upper levels by a flowstone ap-
proximately 1.8 m below the surface, contain many taxa consid-
ered equivalent to those found in the Irvingtonian fauna of Cum-
berland Cave, Maryland (Zakrzewski, 19756). We examined both
Wisconsinan and Irvingtonian specimens.

The fauna of Trout Cave (primarily small mammals) is under
study. The presence of Ondatra annectens. Ochotona, Phenaco-
mys, sp.. Neofiber leonardi. and Atopomys salvelinus was reported
by Guilday (197 1), Guilday and Parmalee (1972), Frazier (1977),
and Zakrzewski (19756), respectively. These five genera were
recovered from the lower strata of the deposit.

Fifteen specimens of Blarina were examined.
Fourteen were dentaries, most with incomplete den-
tition. Seven specimens (CM 12718) were recovered
from the stratum four feet below the surface and
three (CM 12752) from five feet below the surface.
Both of these strata are above the flowstone, so these
specimens are considered Wisconsinan. A maxillary
fragment was recovered from the four-foot stratum.
All molars were present in the maxillary and mea-
surements 1 through 16 were taken (6.3, 4.4, 2.8,
1.9, 1.6, 2.4, 2.5, 2.2, 2.3, 2.4, 1.7, 1.8, 2.3, 2.2, 0.8,
and 1.6, respectively). These measurements com-
pare most favorably with those near the lower end
of the range of variation for B. b. brevicauda and
are similar to those of B. brevicauda from Manitoba
(sample 18). Of the nine dentaries from these two
strata, five were sufficiently complete for analysis
(Fig. 9). Four specimens were from the four-foot
stratum and one (to the right in the canonical dia-
gram—Fig. 9) was from the five-foot stratum; all fell
within the range of variation shown for the taipoides
semispecies of B. brevicauda. Comparisons were
made between the measurements of the incomplete
dentaries and the means derived for those mea-
surements for extant samples. The fossils all appear
to represent B. brevicauda. Some are large and cor-
respond most closely to B. b. brevicauda from Ne-
braska and Iowa (samples 12 and 5), whereas others
are more similar to the smaller subspecies churchi
(sample 1 1), taipoides (samples 13, 19, and 24), kirt-
iandi (samples 4 and 6), and the B. brevicauda from
Manitoba (sample 18).

The five remaining specimens from Trout Cave

1984

JONES ET AL .-BLARINA SYSTEMATICS

89

V2

Fig. 9. — Canonical analysis of measurements 19-24 of lower teeth
of specimens of Blarina from the Trout Cave l.f. Centroids en-
close variation in samples of Holocene taxa of Blarina (see Fig.
4).

are from eight feet (CM 12822, one dentary), nine
to ten feet (CM 20111, one dentary), and ten to
eleven feet (CM 12880 and 20109, three dentaries)
below the surface. Materials below the flowstone are
considered to be Irvingtonian; the trend from youn-
ger to older materials is gradual, so the period of
deposition cannot be ascertained. Certainly the
specimens designated 12880 and 20109 are from
the bottom of the excavation and probably are Kan-
san. These older specimens, when measurements are
compared with those of Holocene specimens, are
most similar to the smaller talpoides semispecies of
B. brevicauda ( samples 19, 4, 6, and 1 1). There were
no discernible differences among the specimens re-
covered from these strata. Accordingly, all speci-
mens of Blarina from Trout Cave pertain to the
species B. brevicauda, but they apparently represent
both the brevicauda and talpoides semispecies.

Rancholabrean

Peccary Cave

Peccary Cave is located on Ben’s Branch of Cave Creek in
Newton County, Arkansas (SW '/a, SE '/a sec. 22, T15N, R19W).
The cave consists of more than 300 m of passages. Most exca-
vations have been in a series of numbered trenches near the two
entrances of the cave. Preliminary excavations were described
by Davis (1969).

Quinn (1972) reported a series of radiocarbon dates from Pec-
cary Cave that range from 2,230 ± 120 to 16,700 ± 250 BP.
Generally, the Peccary Cave l.f. is considered Wisconsinan in
age.

Remains represent inhabitants of the cave, or were brought in
by predators and water (Davis, 1969; Quinn, 1972). Inverte-
brates, fishes, birds, herptiles, and mammals were found. The
invertebrates and herptiles are similar to those of the area today

(Quinn, 1972). The mammals are still under study. Fifty-one
species of mammals have been identified, of which eight are
extinct (including Dasvpus bellus, P/atygonus compressus, San-
gamona fugitiva. Mammut, and Mammuthus). Twenty-nine
species of small mammals were present (Davis, 1969; Kurten
and Anderson, 1980; Semken, personal communication, 1982).

The fauna indicates the presence of parkland and permanent
water at the time of deposition. Climate was cooler, moister, and
more equable than that of the present (Quinn, 1972; Graham,
1972, 1976). Semken (personal communication) is conducting a
study of the ecology and climate as indicated by the small mam-
mals of trenches 8, 13, and 18. Modern climate and vegetation
of the area are summarized in the discussion of Conard Fissure.

Graham and Semken (1976) identified three
phena— “5. b. brevicauda, B. b. kirtlandi, and B. b.
carolinensis [=B. h. hvlophagal]" — of Blarina from
Peccary Cave. The three contemporaneous phena
at Peccary Cave (and at Cumberland Cave and New
Paris No. 4) were thought to represent more equable
climate during full or early late glacial time.

For the present study, 13 specimens from Trench
8, 20 specimens from Trench 13, 27 from Trench
15, and 4 from Trench 24 (SUI 38333-36, 38338,
38340, 38342-44, 39341, 49840, and 49843) were
measured; only the most complete specimens in the
collection were examined. Measurements of the fos-
sils are most similar to those of groups of Holocene
B. brevicauda (Table 1). The examined sample ap-
peared homogeneous. Results of canonical analyses
of 20 maxillaries and 37 dentaries (of which two
observations are hidden and one was plotted to the
right of Fig. 1 1) are shown in Figs. 10 and 11. All
specimens are shown clustered within or near the
centroid showing the range of variation of modern
B. brevicauda ( talpoides semispecies). In the ex-
amination of these specimens, we saw no indication
that more than one phenon, B. brevicauda, is present
in the Peccary Cave l.f.

Arredondo

The Arredondo fauna was recovered from two pits near Ar-
redondo, Alachua County, Florida (SE Va sec. 22, T10S, R19E).
Remains were found in fissure fills at an elevation of 25 m (Bader,
1957; Webb, 1974).

Remains of herptiles, birds, and mammals were recovered.
Herptiles and birds were studied by Auffenberg (1963, 1967) and
Brodkorb (1959). The most recent list of mammals from Arre-
dondo is that by Webb (1974), who listed 33 species from lo-
calities IA, IB, and 1 1 A. Eleven species are extinct. The paleofauna
indicates the presence of forests, meadows, and freshwater ponds,
and winters possibly were warmer than at present (Brodkorb,
1959; Auffenberg, 1963; Webb, 1974).

The remains from localities IA, IB, and II A are approximately
the same age. The fauna is considered Sangamonian (Kurten,
1965; Webb, 1974; Kurten and Anderson, 1980).

90

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

V2

Fig. 10.— Canonical analysis of measurements 3-10 of upper
teeth of specimens of Blarina from the Peccary Cave l.f. Ellipses
enclose variation in samples of Holocene taxa of Blarina (see
Fig. 3).

Bader (1957) and Webb (1974) reported B. brev-
icauda from Arredondo IB and IIA. We examined
six maxillaries and 1 2 dentaries from Arredondo II
(UF 1719, 2098, 3296, 3352, 3559, 12269, 12271,
56073-74, 56076, 56078, 56081, 56545-49, and
56551). Results of the canonical analyses of four
maxillaries and five dentaries are shown in Figs. 12
and 13; these specimens are represented within or
below the range of variation of Holocene B. c. pen-
insulae and B. carolinensis. One maxillary and one
dentary are not shown on the figures. The two re-
maining maxillaries and seven dentaries have mea-

V2

Fig. 1 1 . — Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Peccary Cave l.f. Ellipses
enclose variation in samples of Holocene taxa of Blarina (see
Fig. 4).

V2

Fig. 12.— Canonical analysis of measurements 3-10 of upper
teeth of specimens of Blarina from the Arredondo l.f. Centroids
enclose variation in samples of Holocene taxa of Blarina (see
Fig. 3).

surements comparable to those of the specimens
represented in the figures. Univariate statistics for
the fossils from Arredondo (ARREDO) are shown
in Table 2. We tentatively conclude that the fossils
pertain to the species B. carolinensis.

Bradenton 5 1st Street

Fossil vertebrates were recovered from two coastal marsh sites
in Bradenton, Manatee County, Florida. One specimen, from the
5 1 st Street locality, was referred to B. brevicauda by Webb ( 1 974).

Fourteen species of mammals have been reported from the
Bradenton sites. Four extant and six extinct species are known
from the 5 1st Street locality (“f?. brevicauda Sigmodon sp., S.

V2

Fig. 1 3. — Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Arredondo l.f. Centroids
enclose variation in samples of Holocene taxa of Blarina (see
Fig. 4).

1984

JONES ET AL .-BLARINA SYSTEMATICS

91

Fig. 14. — Canonical analysis of measurements 3-10 of upper
teeth of specimens of Blarina from the Haile XIB l.f. Centroids
enclose variation in samples of Holocene taxa of Blarina (see
Fig. 3).

V2

Fig. 1 5.— Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Haile XIB l.f. Centroids
enclose variation in samples of Holocene taxa of Blarina (see
Fig. 4).

bakeri. Neofiber alleni, Synaplomys australis, Mammuthus co-
lumbi, Equus sp., Paleolama mirifica, Odocoileus virginiana, and
Bison latifrons) (Webb, 1974). The Bradenton sites were consid-
ered Sangamonian by Webb (1974) and Robertson (1974).

A dentary (UF 2470) with partial dentition was
examined. Measurements 19 through 22, 25, and
26 (1.7, 1.2, 1.5, 1.0, 3.4, 1.9) of the two molars
and dentary were taken. These measurements are
similar to those of the reference samples of Holocene
B. carolinensis (Table 1), and the fossil apparently
pertains to that species (probably to the nominal
subspecies peninsulae).

Haile XIB

Bones were recovered from a series of quarries near Haile,
Alachua County, Florida. The quarries are sinkhole deposits lo-
cated approximately 24 m above present sea level (Webb, 1974).

Herptiles, birds, and mammals are represented in the Haile
deposits. Webb (1974) listed 26 species of mammals from Haile
XIB. Two species, Dasypus bellus and Hemiauchenia macro-
cephala, are extinct. The other mammals are known from Florida
in modem times (Ligon, 1965; Webb, 1974). The presence of
Dasypus bellus and Desmodus magnus implies warmer winters
(Ligon, 1965; Webb, 1974). Prairie, scrub, woodland, marsh, and
aquatic habitats are suggested (Ligon, 1965). The prevalence of
small vertebrates suggests deposition by owls and hawks. The
fauna is considered Sangamonian (Kurten, 1965; Ligon, 1965;
Webb, 1974).

Blarina brevicauda is one of three species of shrews
reported from Haile XIB by Webb (1974). Seven-
teen of the most complete maxillaries and seventeen
dentaries (UF4559, 4561, 4563, 4565, 4569, 4571,
4573, 4575-77, 4579, 4581, 4589-91, 4593, 4594,

4603-05, 4608, 4609, 4611-14, 4616-23) were ex-
amined for the present study. Results of canonical
analyses of these fossils are shown in Figs. 14 and
15. Eight maxillaries and one dentary were plotted
outside the range of the diagrams and are not shown.
Four dentaries are hidden. The remaining speci-
mens represented in the two figures fell within or
near the centroids representing the range of varia-
tion of Holocene B. carolinensis. The fossils from
Haile XIB are approximately the same size as those
from Arredondo, and apparently both pertain to the
species B. carolinensis.

An additional specimen (UF 13094) from Haile
XI1IA also was examined. Measurements 3 through
16(1.9, 1.4, 1.1, 1.8, 1.7, 1.6, 1.8, 1.9, 1.4, 1.3, 1.9,
1 .6, 0.7, 1 .4) were taken. This fossil is indistinguish-
able from those from Haile XIB, and we think that
it, too, pertains to B. carolinensis.

Reddick IA

Vertebrate remains were recovered from a limerock quarry 1.6
km southeast of Reddick, Marion County (SW 'A, NW 'A sec. 14,
T13S, R21E), Florida. Fossils were found in four major deposits
(designated A through D) in the quarry. The four sites probably
were contemporaneous although there might have been minor
temporal differences. Specimens of the short-tailed shrew were
recovered from locality A, known as the “Rodent Beds.” This
extensive deposit contains the remains of many small vertebrates
and is thought to be the result of owl-pellet droppings (Gut and
Ray, 1963).

A comprehensive list of the vertebrate fauna, which includes
herptiles, birds, and mammals, was published by Gut and Ray
in 1963. Only the birds and snakes have been studied extensively

92

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

(Brodkorb, 1957, 1963; AufFenberg, 1963; Hamon, 1964). Pre-
liminary studies on selected elements of the mammalian fauna
were done by Olsen (1958), Gut (1959), and Ray et al. (1963).
Initially, the Reddick site was assigned tentatively to the Illinoian.
More recently it has been considered Sangamonian (Martin and
Webb, 1974; Webb, 1974).

Hypotheses regarding the climate of the area at the time of
deposition are somewhat contradictory. The numerous remains
of bats, owls, vultures, and swallows suggest that Reddick IA was
the site of a Pleistocene cave. Birds of southwestern and northern
affinities are present in the A and C deposits. Brodkorb (1957)
considered the avifauna “fairly typical of a wet grassland or fresh
water marsh” habitat. The presence of birds with northern affin-
ities does not necessarily indicate a cooler climate, because the
modern descendants of those birds are migratory. If the climate
was slightly cooler, Brodkorb suggested that it would be similar
to that of modern Virginia. Hamon ( 1 964) also suggested a slight-
ly cooler climate based on his study of the avifauna. Auffenberg
(1963) noted that the reptiles indicated a dry, open forest, similar
to that present in the area today. He considered the possibility,
however, that the snakes and birds might not have been contem-
poraries. Olsen (1958) considered the presence of the bog lem-
ming ( Synaptomys australis) to be further indication of cooler
climate; however, Simpson (1928) pointed out that this fossil
species often is found in associations that suggest it might not
have been boreal. Martin and Webb (1974) thought the high
density of mammalian species in Florida during the Ranchola-
brean indicated tropical or subtropical conditions. Martin (1977)
also interpreted the presence of the neotropical bat Eumops glau-
cinus in the Melbourne and Monkey Jungle Hammock faunas of
central and southern Florida to indicate warm temperatures dur-
ing this period.

The short-tailed shrews from Reddick were re-
ferred to the species B. brevicauda (Gut and Ray,

1 963). Fifteen dentaries (KU 1 409 1 ; UF 8802, 8803,
and uncatalogued material; UMMP V60863) were
examined for the present study. Univariate statistics
are shown in Table 2. Results of canonical analyses
of these fossils are shown in Fig. 16. One observa-
tion is hidden; another observation fell outside the
range of the diagram. All specimens are within or
near the range of variation of B. carolinensis, which
is the shrew present in the southeastern United States
today (George et al., 1982), and we refer the fossils
to that species.

Vero

The Vero fossil beds are located in Vero Beach, Indian River
County, Florida (SE xk sec. 35, T32S, R39E). Fossils were dis-
covered in the banks of a canal in three sedimentary beds (Sel-
lards, 1937; Weigel, 1962). The top two layers, beds 3 and 2,
contained vegetable debris and vertebrate remains. The first, low-
ermost bed contained remains of marine organisms only (Weigel,
1962).

The discovery of human remains associated with the Pleisto-
cene fauna stimulated early interest in the Vero site. Sellards
(1937) listed most of the early studies of the site. Other vertebrates
from Vero include fishes, amphibians, reptiles, birds, and addi-
tional mammals. Weigel (1962) listed 45 mammals, including

V2

Fig. 16. — Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Reddick l.f. Centroids
enclose variation in samples of Holocene taxa of Blarina (see
Fig. 4).

15 extinct species. Vertebrate remains generally are broken and
disarticulated. The fauna suggests the presence of a variety of
habitats, including swamps, hammocks, and pine woods (Weigel,
1962).

Simpson (1929a, 1929 b) listed Vero as one of four faunas of
Florida which he considered contemporaries (early-late Pleis-
tocene). Deposition is thought to have occurred in the Wiscon-
sinan and early Holocene (Bader, 1957; Kurten, 1965; Weigel,
1962). A series of radiocarbon dates from bed 2 suggests depo-
sition from approximately 30,000 to 3,500 BP; bed 3 contains
Holocene contaminants and was not dated (Weigel, 1962).

Simpson (1929 b) reported “5. b. peninsulae ” from
the Vero deposit. Weigel (1962) reported B. brevi-
cauda from beds 2 and 3. For the present study, five
maxillaries and 1 0 dentaries (UFV74 13, 7416, 7417,
7423, 7428-29, 7432, 7435-36, 7459, 7464, 7469-
70, 7482, 7484), the most complete specimens in
the collection, were examined. All specimens were
from strata 2 and 3. Five maxillaries and five den-
taries were sufficiently complete to be included in
statistical (INDIAN, Table 2) and canonical (Fig.
17) analyses. All fossils are most similar to speci-
mens from reference samples of Holocene B. car-
olinensis, and we refer the Vero material to that
species.

Waccasassa River

The Waccasassa River sites are fluviatile deposits located in
Levy County near the western coast of Florida. The sites are
located approximately 6 m above present sea level (Webb, 1974).
A detailed description of the fauna has not been published.

Blarina brevicauda was reported from Waccasassa River IIB
and III by Webb ( 1974). Thirty-five additional mammalian species
were listed, of which 1 5 are extinct. The fauna is considered
Rancholabrean, probably Wisconsinan, in age (Martin, 1969;
Webb, 1974).

1984

JONES ET AL. — B LARINA SYSTEM ATICS

93

V2

Fig. 17. — Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Vero l.f. Centroids enclose
variation in samples of Holocene taxa of Blarina (see Fig. 4).

The material tentatively assigned to the genus
Blarina (UF 16341 and 16362) consists of eden-
tulous jaws, isolated teeth, and a dentary fragment
with two molars. Measurements 19 through 22 (1.6,
1.1, 1.4, 1.0) were taken for the present study. These
measurements, somewhat smaller than those of ref-
erence samples of B. carolinensis (Table 1), are sim-
ilar to measurements of fossils from Arredondo and
other Florida sites. We think that these fossils per-
tain to the species B. carolinensis.

Williston IIIA

Fossils were recovered from sink deposits in Williston, Levy
County, Florida. Amphibians, reptiles, birds, and mammals from
this site were discussed by Holman (1959a, 1959ft).

Thirty species of herptiles, six birds, and 19 mammals have
been identified. One reptilian species, two birds, and five mam-
mals are extinct; extant forms are present in the area today (Hol-
man, 1959a, 1959ft). The composition of the fauna suggests that
surrounding habitat consisted of marshy pineland, pine forest,
and open sinks during the period of deposition (Holman, 1959ft).
Present vegetation near the site is mesophytic (Holman, 1959a).
Modern vegetation of Florida is summarized in the discussion
of the Coleman IIA fauna.

Holman (1959a) considered the Williston fauna Illinoian. More
recently, the fauna has been placed in the Sangamonian (Kurten,
1965; Martin, 1974; Martin and Webb, 1974).

One dentary (UF V-5848) was referred to the
species B. brevicauda by Holman (19596). The size
of the fossil was thought comparable to that of extant
Blarina of north-central Florida. For the present
study, measurements 19 through 22 (1.6, 1.0, 1.4,
0.9) were taken. We find these measurements some-
what smaller than those of modern B. c. peninsulae
(Table 1), but no morphological features were ob-

served that would distinguish the fossil from the
modern species. We conclude that the Williston
specimen is a small B. carolinensis.

Ladds Quarry

The first major assemblage of Pleistocene vertebrates reported
from Georgia is located in Bartow County. Fossils were recovered
from a matrix of red cave earth (in part firmly cemented as a
cave breccia) located in small fissures exposed in a limestone
quarry (Ray, 1965). The quarry is in the southeastern end of
Quarry Mountain (also known as Ladds Mountain), approxi-
mately 4 km WSW of Cartersville, 34°09'N, 84°50'W (Lipps and
Ray, 1967).

At least 25 species of molluscs, 4 species of birds, 23 species
of herptiles, and 48 species of mammals have been identified at
Ladds. Preliminary faunal lists were compiled by Holman (1967),
LaRocque (1967), Ray (1967), and Wetmore (1967). Mixing of
materials in the deposit was recognized early in the investigation
of the quarry. “The fossils occur in small fissures exposed in a
remnant pinnacle of limestone . . . now isolated from the major
part of the mountain by extensive quarrying .... The disturbed
(by dynamiting) and open nature of the deposit, together with
the presence of certain apparently ecologically incompatible
species, suggest the possibility of a heterochronic assemblage rather
than a unit fauna” (Ray, 1965). A mixed assemblage, at least in
part, is suggested by the molluscs, herptiles, and mammals (Hol-
man, 1967; LaRocque, 1967; Ray, 1965, 1967).

A radiocarbon date for this fauna has not been obtained. The
presence of some of the extinct forms, particularly Peromyscus
cumberlandensis, might indicate a pre-Wisconsinan age for at
least part of the fauna (Guilday, 1971;Guildayetal., 1978). Watts
(1970) compared the results of the analysis of pollen from Ladds
by Benningholf and Stevenson (1967) with pollen spectra ob-
tained from two ponds in Bartow County. In the pollen spectra
from the two Georgia ponds, a pollen assemblage (designated as
zone Ql) was recognized from which basal dates of 20,100 ±
240 and 22,900 ± 400 years BP were obtained. Watts proposed
referral of the assemblage from zone Ql and a similar assemblage
from North Carolina as a southeastern pine-spruce ( Pinus-Picea )
assemblage zone. He raised the possibility that the spectra from
Ladds could be referred to this pine-spruce assemblage, which
he considered “contemporaneous with the main Wisconsin ice
advance.” This is, however, a tentative correlation, due to the
mixing observed in the deposit at Ladds.

Northern elements in the fauna, including the spruce grouse
Canachites canadensis and the wood turtle Clemmys insculpta
(which presently occur no farther south than 44°N and 39°N.
respectively), might indicate a cooler environment (Holman, 1967;
Ray, 1967; Wetmore, 1967; Guilday et al., 1978). The presence
of southern elements in the fauna, including Neofiber alleni (pres-
ently restricted to Florida and southern Georgia) and Didelphis
virginianus, demonstrates the mixed nature of the fauna at Ladds
(Ray, 1967; Frazier, 1977).

Ray (1967) compared the Blarina from the Ladds
l.f. with modern samples from Alabama (Cullman
County), North Carolina (Transylvania County), and
West Virginia (Greenbriar County). Based on the
10 cranial and dental characters used, the fossils
were more heterogeneous in size than were the mod-

94 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 8

V2

Fig. 18. — Canonical analysis of measurements 3-10 of upper
teeth of specimens of Blarina from the Ladds Quarry l.f. Cen-
troids enclose variation in samples of Holocene taxa of Blarina
(see Fig. 3).

V2

Fig. 19. — Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Ladds Quarry l.f. Cen-
troids enclose variation in samples of Holocene taxa of Blarina
(see Fig. 4).

ern samples used for comparison, but were most
similar to specimens from West Virginia. The fossils
were referred to the species B. brevicauda.

Twenty-one specimens (USNM 23384, 23385,
and uncatalogued material) were examined herein.
Five specimens were analyzed for which measure-
ments 3-10 were complete. These specimens clus-
tered within the centroid for the talpoides semispe-
cies of B. brevicauda (Fig. 1 8). These fossils are larger
than the modern shrews of the southeastern United
States, and they are most similar to the samples of
B. b. talpoides from New York and B. b. churchi
(samples 13 and 1 1, respectively). Analyses of mea-
surements of lower teeth were performed for eight
specimens, which plotted in two groups (Fig. 19).
The larger of the two groups falls within the centroid
for the talpoides semispecies of B. brevicauda. The
smaller group consists of three specimens, all of
which were uncatalogued but designated by the let-
ters BB, X, and Y. These animals fall within the
range of variation for modern B. c. caro/inensis.
Univariate statistics for the Ladds fossils (LADDSQ)
are shown in Table 2. Because of the implication
that the Ladds l.f. is heterochronic, it is suggested
that these are B. c. caro/inensis that were deposited
at a later date (possibly Holocene or sub-Recent).
In any event, the fauna evidently consists of two
species of Blarina. B. brevicauda and B. caro/inen-
sis.

Meyer Cave

Meyer Cave is located 6.4 km SSW of Columbia, Monroe
County, Illinois (NW 'A sec. 6, T2S, R10W, New Hanover Town-

ship). The opening of the cave is a narrow passageway in a lime-
stone bluff over the Mississippi River. The passageway leads to
a small room, the floor of which slopes toward a hole. The bell-
shaped fissure located below this room acts as a natural trap for
animals (Parmalee, 1967).

Fossils were recovered from the fissure. The invertebrate fauna
consists of approximately 27 species of molluscs and insects,
whereas the vertebrate fauna consists of approximately 1 1 5 species
of fishes, herptiles, birds, and mammals. The deposit was not
stratified because of burrowing by some of the animals that en-
tered the cave and fell into the fissure. Remains of fishes and
some other vertebrates probably were carried in by predators.
Six samples of soil subjected to palynological analysis were de-
void of pollen (Parmalee, 1967).

No extinct species were found. Fifteen species (including 13
mammals) previously were unknown from Monroe County. Three
of the mammals ( Microtus xanthognathus. Clethrionomys gap-
peri. and Erethizon dorsatum) were not recorded previously from
Illinois. The presence of the pigmy shrew ( Microsorex hoyi) and
other species that have modem ranges to the north of Illinois
indicates deposition during cool, moist conditions (ca. 9,500-
7,500 BC); the presence of the plains pocket gopher (Geomys
bursarius ) and the eastern spotted skunk (Spilogale putorius) sug-
gests additional deposition during a warmer, drier period (ca.
3,500-1,500 BC) (Parmalee, 1967).

The area presently receives approximately 91-96 cm of pre-
cipitation annually (NOAA, 1974). The fauna and flora are con-
sidered part of the lllinoian biotic province as defined by Dice
in 1943 (Parmalee, 1967).

Remains representing a minimum of 262 speci-
mens of Blarina were removed from Meyer Cave.
Two distinct sizes, identified as B. brevicauda cf.
brevicauda and B. b. caro/inensis. were noted by
Parmalee (1967). Graham and Semken (1976)
equated the two phena from Meyer Cave with B. b.
kirtlandi and B. b. caro/inensis. They regarded the
two as contemporaneous phena, and suggested that

1984

JONES ET Ah. — BLARINA SYSTEM ATICS

95

Fig. 20.— Canonical analysis of measurements 3-10 of upper
teeth of specimens of Blarina from the Meyer Cave l.f. Centroids
enclose variation in samples of Holocene iaxa of Blarina (see
Fig. 3).

sympatry at Meyer Cave (as well as at Crankshaft
Cave and Welsh Cave) was indicative of more
equable climate at the time of deposition.

For the present study, 148 specimens from Meyer
Cave were measured. In the analysis of measure-
ments of upper teeth, 34 specimens were plotted
within the centroid for B. carolinensis and the cen-
troid for the smaller subspecies ( talpoides semispe-
cies) of B. brevicauda {Fig. 20). In the analysis of
the lower tooth measurements, most of 1 14 fossils
again were plotted within the centroids for B. car-
olinensis and the talpoides semispecies of B. brevi-
cauda (Fig. 21). Six additional fossils were plotted
to the right of the range of the diagram shown in
Fig. 2 1 ; three of these were just outside the range of
variation shown for B. carolinensis and three were
near the centroid for B. brevicauda. Forty-five spec-
imens were hidden behind other points within the
two centroids. Univariate statistics for the fossils
from Meyer Cave are shown in Table 2 (MEYERC).
Apparently the species of Blarina in the fauna are
B. carolinensis and B. brevicauda, as previously sug-
gested by Graham and Semken (1976); however, it
is not possible with the present data to discount the
presence of a third species, B. hylophaga, which is
approximately intermediate in size between B. brev-
icauda and B. carolinensis. The three species do not
occur together at any locality at the present time,
but might have in the past.

Cherokee Sewer Site

The Cherokee Sewer Site is a bison kill site located 3.2 km
south of Cherokee, Cherokee County, Iowa (W Vi, SW ‘4, SE ‘4

1.60

V2

Fig. 2 1 . —Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Meyer Cave l.f. Centroids
enclose variation in samples of Holocene taxa of Blarina (see
Fig. 4).

sec. 4, T9 IN, R40W). Deposition occurred in an alluvia! fan on
the west floodplain of the Little Sioux River Valley at an elevation
of 360 m (Shutler et al., 1980).

The stratified fan deposits contained three cultural horizons.
Radiocarbon dates of ca. 6,350, 7,300, and 8,400 BP were ob-
tained for horizons I, II, and III, respectively (Hoyer, 1980).
Seeds, molluscs, and mammals were associated with the arche-
ological remains. LInidentified fish, herptiles, and birds also were
present. The site is of special interest because of its location at
the fluctuating prairie-forest border. An interdisciplinary study
was conducted to reconstruct the paleoecology and cultural his-
tory of the area; results were summarized by Shutler et al. ( 1 980).

Remains of canids, bison, cervids, and 23 small mammals were
found at the Cherokee site (Pyle, 1980; Semken, 1980). All mam-
mals are extant. Four species ( Microsorex hoyi, Tamiasciurus
hudsonicus, Perognathus hispidus, and Synaptomys coopen' ) no
longer occur in the area. Accumulation of the small mammals
probably is due to the presence of bison remains and other gar-
bage; raptorial or human predation also might have contributed
to the accumulation (Semken, 1980).

The Cherokee l.f. indicates the presence of forest, arboreal,
meadow, prairie, and aquatic communities. The mammals, in-
cluding the four species no longer present, suggest a local habitat
of open gallery forest (Semken, 1980). Sympatry of the species
from each horizon show a trend of increasing aridity from ap-
proximately 8,400 to 6,350 BP. Climatic conditions of horizon
II are the most similar to those of the present (Semken, 1980).
Analysis of gastropods from Cherokee and pollen analyses from
Lake West Okoboji (80 km north) also indicate a trend of in-
creasing aridity, accompanied by changes in the associated flora
and fauna (Baerreis, 1 980; Baker and Van Zant, 1 980; Wendland,
1980).

Dice (1943) included Iowa in the Illinoian biotic province.
Three grassland associations comprised the tail-grass prairie that
covered the western and central parts of the state; deciduous
forests occurred in eastern Iowa and along waterways and hill-
sides in the west (Bowles, 1975). The climate, geology, and vege-
tation of Iowa were summarized by Bowles (1975). In Cherokee
County the mean annual precipitation is 72.3 cm (Shutler et al.,
1980).

96

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Semken ( 1 980) identified remains of Blarina from
the Cherokee Sewer Site as B. b. brevicauda. Spec-
imens consist of two isolated teeth from horizon I
(SUI 38067A, B); a mandible with ml-3, a man-
dibular fragment with i-m 1 , a mandibular fragment
with m 1-2, two mandibular condyles, a mandibular
fragment with m2-3, and isolated teeth from II (SUI
38068, 44092A-F); a palate and a mandible from
III (SUI 44039A, B). A minimum number of 1, 2,
and 2 individuals thus are represented in horizons
I, II, and III, respectively.

Six of the most complete specimens were exam-
ined for the present study. Measurements 7 through
10 (2.3, 2.2, 2.4, 2.6) and 3 through 16 (2.5, 2.0,
1.4, 2.4, 2.3, 2.1, 2.3, 2.6, 1.7, 1.7, 2.4, 2.0, 0.8, 1.5)
of upper teeth and measurements 19 through 22 of
two mis and m2s (2.2, 1.5, 2.0, 1.4; 2.3, 1.6, 2.0,
1.4) and 19 through 26 of a dentary with complete
dentition (2.4, 1.7, 2.0, 1.3, 1.6, 1.1, 4.4, 2.8), were
compared with those of reference samples (Table 1 ).
Measurements of the specimens from Cherokee were
most similar to those of modern B. b. brevicauda,
and generally fell within the range of variation of
the modern sample from Iowa. The specimens from
the Cherokee Sewer Site pertain to the subspecies
B. b. brevicauda.

Craigmile

The Craigmile and Waubonsie local faunas were recovered
from the Craigmile farm (SE, NW, NW, NE '4 sec. 13, T71N,
R43W) in the Waubonsie Creek watershed, Mills County, Iowa.
The site and the two paleofaunas were described in a dissertation
by Rhodes (1982), which serves as the basis for this report.

The two l.f.s were recovered from superposed Wisconsinan
alluvial fills. Stratigraphy of the site was described by Rhodes
(1982). A minimum age of 23,240 ± 535 RCYBP was obtained
from bone residue in the unit from which the Craigmile l.f. was
recovered.

The Craigmile l.f. is represented by fishes, herptiles, birds, and
mammals. Rhodes (1982) listed 31 mammalian taxa, of which
four species (Dasypus bellus. Mammut sp., Equus sp., and San-
gamona fugitiva) are extinct. Nine species no longer occur in
Mills County. Remains of microtines were the most common
elements in the deposit. All remains were disarticulated and dam-
aged (Rhodes, 1982).

Twenty-nine mammals indicate grassland, meadow, and brushy
edge habitats at Craigmile during the period of deposition. Cooler
summer temperatures are suggested. Taxa no longer present in
the area are boreal or boreomontane species. Twenty-two species
are sympatric in central Wisconsin in an area of oak or pine
savannah with patches of dense forest and prairie (Rhodes, 1982).

Pre-settlement vegetation of the Waubonsie watershed prob-
ably consisted of tail-grass prairie on the Missouri River flood-
plain and exposed slopes, and upland riparian forests (Rhodes,
1982). Modern climate and vegetation of the state were sum-
marized by Bowles (1975).

V2

Fig. 22. — Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Craigmile l.f. Centroids
enclose variation in samples of Holocene taxa of Blarina (see
Fig. 4).

Blarina brevicauda was reported from both the
Craigmile and Waubonsie faunas. Two mandibles,
a maxillary fragment, and isolated teeth (SUI 46405,
MNI = 3) were recovered from the Craigmile de-
posit. The single measurable mandible was referred
to the subspecies B. b. brevicauda (Rhodes, 1982).

These specimens were examined for the present
study. Measurements 7 through 10 of one Ml (2.4,
2.2, 2.5, 2.8) fall within the range of variation of
modern B. brevicauda ( talpoides semispecies) and
B. b. brevicauda ; measurements 7 through 1 0 of a
second Ml (2.0, 2.0, 2.1, 2.3) are slightly smaller
and are more similar to measurements of the tal-
poides semispecies of B. brevicauda (Table 1). Mea-
surements 19 through 26 of a dentary and lower
molars (2.5, 1.8, 2.0, 1.3, 1.6, 1.1, 4.4, 3.0) are most
similar to measurements of modern B. b. brevicau-
da. This specimen is represented at the top of Fig.
22 within the centroid showing the range of varia-
tion of B. b. brevicauda from Iowa and Nebraska.
These data suggest that Mills County, Iowa, was
within the zone of intergradation of the two semi-
species of B. brevicauda 23,000 years ago just as it
is now (Moncrief et al., 1982), and was inhabited
by a mixture of large, small, and intermediate-sized
shrews. The name B. b. brevicauda is appropriate
for intergrades recovered from the Craigmile l.f.

Garrett Farm

The Garrett Farm fauna was recovered from alluvial sediments
along Waubonsie Creek in Mills County, Iowa (NW '4, NE ‘4,
NW '4, SW '4 sec. 8, T71N, R42W). A radiocarbon date of 3, 400
BP was obtained (Fay, 1980).

1984

JONES ET AL. — B LARINA SYSTEM ATICS

97

Remains of plants, molluscs, and vertebrates were recovered.
Fay (1980) listed 21 mammalian taxa from Garrett Farm, of
which three ( Neotoma Imicropus, Clethrionomys gapperi, and
what he termed B. b. kirtlandi) no longer occur in southwestern
Iowa. Mammals associated with steppe habitat dominate the
fauna, but boreal and deciduous forest taxa also are present (Fay,
1980). Fay suggested that the allopatric mammals of the Garrett
Farm and Pleasant Ridge faunas represent “the last stage of ad-
justment from Hypsithermal to modern climatic conditions.”
The modem vegetation and climate of Iowa are summarized in
the discussion of the Cherokee Sewer l.f.

Fay (1980) identified B. b. kirtlandi (MNI = 2)
and B. carolinensis (MNI = 1) from Garrett Farm.
An isolated P4, M2, and a dentary with Ml-3 (SUI
46886) were examined for the present study. Mea-
surements 3 through 6 (2.4, 1 .9, 1.3, 2.4), 1 1 through
14 (1.7, 1.8, 2.4, 2.0), and 19 through 26 (2.2, 1.5,

1.8, 1.2, 1.5, 1.0, 3.6, and 2.5) of these specimens
were taken and compared with measurements from
the reference samples (Table 1). Measurements of
the three specimens from Garrett Farm are most
similar to those of the talpoides semispecies of B.
brevicauda, and the specimens possibly pertain to
the Holocene subspecies B. b. kirtlandi.

Mud Creek

The Mud Creek fauna was recovered from six sites on the
banks of Mud Creek, north of Durant, Iowa. The sites were
located from the west line of sec. 1 (T79N, R1W) in Cedar County
to the east line of sec. 7 (T79N, R1E) in Scott County. The
description of the fauna by Kramer (1972) serves as the basis for
this report.

The fauna consists of molluscs, fishes, herptiles, and mammals.
The mammalian fauna consists of insectivores, rodents, rabbits,
deer, and bison. Four species, including what Kramer termed B.
b. kirtlandi, no longer occur in the area. Sympatry of extant
mammals occurs north and east of Iowa (Kramer, 1972).

A radiocarbon date of 6,220 ± 110 BP was reported for the
Mud Creek l.f. Members of the fauna indicate the presence of
stream and forest habitats. Climate was somewhat cooler and
moister, similar to that of modem Michigan, Illinois, Indiana,
Ohio, and Pennsylvania (Kramer, 1972). A pronounced mete-
orological gradient evidently occurred between Mud Creek and
the more xeric Cherokee I site (Semken, 1980).

Presently, the mean annual precipitation of Cedar and Scott
counties is 81 to 86 cm (NOAA, 1974). Natural vegetation of
the area is summarized in the discussion of the Cherokee Sewer
Site.

Kramer (1972) referred specimens (MNI = 4)
from Mud Creek to B. brevicauda and B. b. kirtlan-
di. Measurements 3 through 6 of an isolated P4 (2.3,

1.9, 1.4, 2.0), 7 through 10 of three mis (2.2, 2.1,
2.4, 2.5; 2.1, 2.1, 2.2, 2.3; and 2.2, 2.1, 2.2, 2.5), 19
and 20 of a ml (2.2, 1.5), and 21 and 22 of a m2
(1.8, 1.2) were obtained for the present study (SUI
35940A and D). Measurements of these teeth were

compared with means obtained for those measure-
ments from samples of Holocene specimens (Table
1). Measurements of the Mud Creek specimens are
smaller than those of the B. b. brevicauda inhabiting
most of Iowa today. Comparisons indicate that the
specimens from Mud Creek pertain to the talpoides
semispecies of B. brevicauda, possibly to the sub-
species B. b. kirtlandi.

Pleasant Ridge

The Pleasant Ridge local biota was recovered from deposits
100 m from the Garrett Farm site in Mills County, Iowa (SE 'A,
NW >A, NW 'A, SW 'A sec. 8, T71N, R42W). A radiocarbon date
of 1,450 BP was reported by Fay (1980).

The Pleasant Ridge fauna consists of molluscs and vertebrates.
Plant remains also were recovered. Twenty-four mammalian taxa
are present; Neotoma floridana is the only species not present in
the modem fauna. Mammals of steppe, deciduous forest, and
boreal habitats are represented equally in the fossil fauna (Fay,
1980).

Fay (1980) referred remains of Blarina (MNI =
5) from Pleasant Ridge to the subspecies B. b. brev-
icauda. For the present study, measurements 3
through 6 (2.5, 2.2, 1.5, 2.6), 11 through 14 (1.8,
1.7, 2.5, 2.1), and 19 and 20 (2.2, 1.5) ofa P4, M2,
and ml from Pleasant Ridge were taken (SUI 45973).
These measurements were compared with those ob-
tained from modern reference samples (Table 1).
This comparison tends to support Fay’s original
identification of the Pleasant Ridge material as B.
b. brevicauda.

Schmitt Cave

Schmitt Cave is located in western Dubuque County, Iowa
(T88N, R1W, sec. 31), near the town of Farley. The site is a rock
shelter and cave located at the base of a cliff, approximately 40
m north of John Creek. Excavation of artifacts was described by
Reese (1972). Accumulation of artifacts and faunal remains oc-
curred approximately 1,900 to 1,300 BP (Reese, 1972; Semken,
personal communication, 1982). Faunal remains represent in-
habitants of the cave and immediate surroundings as well as
species hunted by man (Eshelman, 1972).

Eshelman (1972) described the Holocene fauna, which consists
of 26 fishes, herptiles, birds, and mammals, which was recovered
from the cave. All occur presently in northeastern Iowa except
Erethizon dorsatum. Sympatry occurs in southwestern Wisconsin
and the eastern half of Iowa, indicating little or no climatic changes
since deposition. Permanent stream (probably John Creek), stream
border, forest, and lowland meadow-prairie communities are rep-
resented (Eshelman, 1972). Modem climate and vegetation of
Iowa are summarized in the discussion of the Cherokee Sewer
Site.

Remains of Blarina (SUI 34989) were recovered
from levels 1-3, which represent the top 18 inches
of the deposit. The nine specimens (MNI = 5) from
the site were referred to the species B. brevicauda

98

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

V2

Fig. 23. — Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from Schmitt Cave. Ellipses enclose
variation in samples of Holocene taxa of Blarina (see Fig. 4).

by Eshelman (1972). Two maxillaries and five den-
taries were examined for the present study. Mea-
surements 3 through 16 of upper teeth of the two
maxillaries (2.5, 2.0, 1.6, 2.6, 2.4, 2.2, 2.4, 2.6, 1.7,
1.7, 2.6, 2.2, 0.8, 1.6; 2.4, 1.9, 1.4, 2.3, 2.2, 2.0, 2.3,

2.4, 1.7, 1.7, 2.3, 2.0, 0.8, 1.6) are most similar to
those of Holocene B. b. kirtlandi, B. b. manitoben-
sis, and B. b. churchi (groups 4, 6, 18, and 11 in
Table 1). Measurements 19 through 24 of lower
teeth of the five dentaries also are comparable to
those of modern B. brevicauda ( talpoides semispe-
cies). In canonical analyses these five specimens (one
was hidden) are plotted within the centroid showing
the range of variation of the talpoides semispecies
of Holocene B. brevicauda (Fig. 23). The seven spec-
imens from Schmitt Cave are assigned to the species
B. brevicauda, and probably pertain to the subspe-
cies B. b. kirtlandi.

Thurman

The Thurman I f. was recovered near Thurman, Iowa, from
the Fremont County Quarry (NW lA, NW 'A sec. 23, T70N,
R43W). The fauna was described in an unpublished thesis (Jen-
kins, 1972).

Seeds and remains of gastropods, pelecypods, fishes, amphib-
ians, reptiles, birds, and mammals were recovered from flood-
plain deposits in the quarry. Jenkins (1972) listed 1 4 mammalian
species from the site, only one of which (Mammuthus columbi )
is extinct. The plant and animal remains suggest a temperate
cli mate, possibly with more abundant moisture and more equable
temperatures than at present (Jenkins, 1 972; Graham and Semken,
1976).

A radiocarbon date of 9,800 ± 1 50 BP was obtained (Graham
and Semken, 1976). The presence of mammoths suggests de-
position prior to 10,000 BP (Jenkins, 1972).

Four dentaries pertaining to the genus Blarina
were recovered from the quarry (SUI 35869 and

35888). Graham and Semken (1976) thought the
remains represent two taxa, B. b. brevicauda and B.
b. carolinensis [=B. hylophaga]. Measurements 19
through 26 of three specimens (2.0, 1.3, 1.7, 1.1,

1.5, 1.0, 3.5, 2.3; 2.0, 1.4, 1.7, 1.2, 1.4, 1.0, 3.4, 2.4;
2.1, 1.4, 1.8, 1.2, 1.4, 1.0, 3.5, 2.2) and 21 through
26 of the remaining specimens (2.1, 1.3, 1.7, 1.1,

4.5, 2.6) were taken for the present study. These
measurements were compared with those of Holo-
cene reference samples (Table 1). These compari-
sons indicate that the three smaller specimens rep-
resent the species B. brevicauda ( talpoides
semispecies). The larger specimen apparently per-
tains to B. b. brevicauda (of the brevicauda semi-
species). We assume co-occurrence of these two
phena in the Thurman l.f. represents sympatry in
southwestern Iowa during the late Wisconsinan/ear-
ly Holocene period. The two semispecies of B. brev-
icauda remain sympatric in Fremont County, Iowa,
today.

Waubonsie

The Waubonsie l.f. was found in a channel fill above the unit
from which the Craigmile l.f. was recovered. A radiocarbon date
of 1 4,830 + 1 ,060, — 1 ,220 RCYBP was obtained from charcoal
fragments associated with the vertebrate remains. The Waubon-
sie l.f. was described by Rhodes (1982).

Remains of fishes, herptiles, birds, and mammals were re-
covered. Twenty-three mammalian taxa were represented, of
which 14 occur presently in Mills County. As at Craigmile, pres-
ervation was poor and microtines were the most numerous in-
dividuals present. Eight species that have been extirpated locally
occur sympatrically in central Alberta. The region of maximum
sympatry (17 of 22 taxa considered) is in central Wisconsin and
central Minnesota in areas of prairie-forest transition. The area
of maximum sympatry in Wisconsin is the same as that of Craig-
mile (Rhodes, 1982).

The Waubonsie l.f. indicates the presence of conifer and hard-
wood forests and boreal grasslands (Hallberg et al„ 1974; Rhodes,
1 982). Temperatures probably were similar to those of Craigmile,
but effective precipitation was greater. Rhodes (1982) suggested
that the “environmental mosaic” was more strongly developed
because of changes in topography and climate.

Rhodes (1982) assigned remains of Blarina (SUI
46361, MNI = 7) to the subspecies B. b. brevicauda.
He noted that four of five measurable mandibles fell
within an area of overlap between B. b. brevicauda
and B. b. kirtlandi when using Graham and Semken’s
(1976) mandibular index.

One complete upper toothrow was examined for
the present study. Measurements 3 through 10 (2.4,
2.0, 1.4, 2.4, 2.2, 2.1, 2.3, 2.5) compare most fa-
vorably with measurements of reference samples of
the talpoides semispecies of B. brevicauda (Table 1 ).
Three of the mandibles seen are represented in the

1984

JONES ET AL.- BLARINA SYSTEMATICS

99

Fig. 24. — Canonical analysis of measurements 3-10 of upper
teeth of specimens of Blarina from the Willard Cave l.f. Ellipses
enclose variation in samples of Holocene taxa of Blarina (see
Fig. 3).

canonical diagram in Fig. 22. One specimen felt
within the centroid representing the range of vari-
ation of the talpoides semispecies of B. brevicauda-,
the other two plotted within the area of overlap
between the talpoides and brevicauda semispecies
of B. brevicauda. Measurements of additional fossils
with incomplete dentition are similar to those of the
specimens included in canonical analyses. We, too,
find that these measurements fall into an area of
overlap between those of Holocene B. b. brevicauda
and the smaller talpoides semispecies. As with the
Craigmile l.f., the Waubonsie l.f. appears to have
been deposited in the zone of intergradation be-
tween the brevicauda and talpoides semispecies of
B. brevicauda. However, fossils of Blarina from
Waubonsie appear slightly smaller than modern
specimens from Iowa, and for that reason we suggest
that they be assigned to the talpoides semispecies of
B. brevicauda, subspecies B. b. kirtlandi.

Willard Cave

The Willard Cave l.f. was recovered from Delaware County,
Iowa (NE >/4, SW ‘/4, SW lh sec. 2, R4W, T90N). The fauna was
described in a thesis (Eshelman, 1971} that currently is being
prepared for publication (Eshelman, in iitt, 1982).

Remains of 37 vertebrate species were recovered from Willard
Cave. All species are extant, but eight no longer occur in north-
eastern Iowa. Thirty-one mammalian species were identified; 74%
of the individuals represented pertain to species associated with
deciduous forest (Eshelman, 1971). Semken (1980) noted that
Willard Cave is the first record of prairie species in eastern Iowa.

The species in the fossil fauna presently are allopatric. Depo-
sition might have occurred during more equable climatic con-
ditions, with a mosaic of habitats in the vicinity (Eshelman, in

1.60

V2

Fig. 25.— Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Willard Cave l.f. Ellipses
enclose variation in samples of Holocene taxa of Blarina (see
Fig. 4).

Iitt., 1982). The modem vegetation of the area is summarized in
the discussion of the Cherokee Sewer Site.

Radiocarbon dates indicate that accumulation occurred be-
tween 3,500 and 1,200 BP (Eshelman, 1971). Eshelman (1971)
noted similarities between the Willard Cave l.f. and the faunas
of Crankshaft Cave, Meyer Cave, New Paris No. 4, and Robinson
Cave. He suggested that the Willard fauna accumulated more
recently than these faunas, as evidenced by the absence of extinct
taxa.

The genus Blarina is represented throughout the
deposit by 327 specimens (MNI = 54). These re-
mains (SUI 35185 and 50041) were referred to the
subspecies B. b. brevicauda by Eshelman (1971) and
Graham and Semken (1976).

Twenty-seven maxillaries and 60 dentaries, rep-
resenting the most complete of all the specimens,
were examined for the present study. Analyses of
dental measurements of these specimens (Figs. 24
and 25) indicate that the specimens from Willard
Cave pertain to the talpoides semispecies of B. brev-
icauda.

Jinglebob

Fossils were found in the Jinglebob Pasture of the XI Ranch
(UM-K2-47, SW >/4 sec. 32, T33S, R29W) in Meade County,
Kansas. The Jinglebob is one of four Rancholabrean sites in
Kansas from which specimens of Blarina have been recovered.

The fauna consists of molluscs, fishes, herptiles, and mammals.
Fragments of avian bones were recovered but have not been
studied (Hibbard, 1955). Twenty-three mammalian species have
been identified, of which 10 are extant (Hibbard, 1955, 1970).
Six vertebrate species, including Blarina cf. brevicauda. presently
occur north or south of Meade County. Taylor and Hibbard
(1955) and Hibbard (1963) compared the fauna with others from
southwestern Kansas and northwestern Oklahoma.

The fauna was aged as Sangamonian, but more recently it was
considered early Wisconsinan (Hibbard, 1970; Kapp, 1970; Za-

100

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

krzewski, 1975a). The molluscs and vertebrates, particularly the
presence of both northern and southern elements in the fauna,
indicate more equable climate during the period of deposition
(Hibbard, 1955, 1970; Martin, 1968). Grassland with heavy cov-
er and permanent streams were present (Hibbard, 1955; Hibbard
and Taylor, 1960). Pine and spruce pollen were abundant in
pollen spectra (Kapp, 1970). The modern climate and vegetation
of the area were described in the discussion of the Cudahy l.f.

A relatively great diversity of voles and shrews
was present in the Jinglebob fauna (Graham, 1976).
Eleven specimens of Blarina were referred to B. cf.
brevicauda (Hibbard, 1955, 1970). The remains
consist of parts of five maxillaries (UMMP 29264,
29265, and 29761 ) and six dentaries (UMMP 29266-
70). All specimens were examined for the present
study. The sample appears homogeneous. Measure-
ments of the fossils were compared with those of
Holocene specimens (Table 1). Measurements of
both upper and lower molars are similar to those of
modern B. brevicauda and B. hylophaga. Measure-
ments 3 through 6 of three P4s (2.3, 1.9, 1.3, 2.2;
2.3, 1.8, 1.4, 2.3; 2.2, 1.8, 1.3, 2.1), 7 through 10 of
three Mis (2.1, 1.9, 2.2, 2.3; 2.0, 1.9, 2.1, 2.2; 2.0,
1.8, 2.1, 2.2), and 1 1 through 14 of two M2s (1.8,

1.6, 2.3, 1.9; 1.7, 1.7, 2.2, 1.8) fall within the range
of measurements of modem B. carolinensis, B. brev-
icauda ( talpoides semispecies), and B. hylophaga.
Measurements 19 and 20 of four m Is (2.2, 1.5; 2.0,
1.4; 2.1, 1.4; 2.0, 1.4), 21 and 22 of five m2s (1.8,
1.2; 1.8, 1.2; 1.8, 1.2; 1.7, 1.1; 1.7, 1.1), and 23 and
24 of three m3s (1 .4, 1.0; 1.3, 1.0; 1.4, 1.0) also are
comparable with those of Holocene B. carolinensis,
B. brevicauda, and B. hylophaga. However, mea-
surements of the remains from Jinglebob are most
similar to those of B. brevicauda, and we think that
the fossils pertain to that species.

Mt. Scott

Fossils of the Mt. Scott l.f. were found at three sites on the Big
Springs Ranch in Meade County, Kansas. The three sites are
UM-K4-53 (SE 'A sec. 14, T32S, R29W), UM-K2-59 (SE 'A, SE
■A sec. 18, T32S, R28W), and UM-K1-60 (SW 'A, SW >A sec. 13,
T32S, R29W). Molluscs only were recovered from a fourth site,
UM-K3-60 (NE 'A, SW ‘A sec. 17, T32S, R28W) (Miller, 1966).

In addition to molluscs, remains of fishes, amphibians, reptiles,
birds, and mammals were found at the Mt. Scott sites. Six orders
of mammals were represented by 25 species (Hibbard, 1963).
Mt. Scott was compared to other Pleistocene faunas of Kansas
and Oklahoma by Hibbard (1963) and Semken (1966). Six of the
extant species, including Blarina carolinensis, do not occur pres-
ently in Meade County (Hibbard, 1963).

The Mt. Scott l.f. is considered late Illinoian. Deposition oc-
curred during a period of more effective moisture and more equa-
ble temperatures than now present (Hibbard, 1963, 1970; Smith,
1963; Miller, 1966). The fossil fauna indicates the presence of a
permanent stream, marsh, and shrubs and trees; grassland was

nearby (Etheridge, 1961; Hibbard, 1963; Smith, 1963; Nelson
and Semken, 1970). A shift from colder to warmer temperatures
occurred after the early Illinoian, accompanied by the appearance
of more southern elements as seen in the Mt. Scott fauna (Hib-
bard, 1970). Pollen from Mt. Scott was poorly preserved, but
pollen spectra imply expansion of open plant communities during
the period of deposition (Kapp, 1970). Modem climate and vege-
tation of Meade County are discussed in the section about Cu-
dahy.

Remains identified as B. b. carolinensis were re-
covered from all three sites (Hibbard, 1963). Eigh-
teen dentaries (UMMP 37751, 43804-12, 41233,
44591, and 44600), seven maxillaries (UMMP
438 13-15, 44591), and an isolated incisor (UMMP
43816) were found, and all were examined for the
present study.

Incomplete measurements were obtained from the
seven maxillaries (measurements 7 through 14; 2.1,
2.0, 2.3, 2.4, 1.7, 1.7, 2.3, 1.8; 3 through 14: 2.1,

1.6, 1.2, 2.1, 1.9, 1.8, 2.1, 2.2, 1.5, 1.5, 2.1, 1.7; 3
through 14; 2.1, 1.7, 1.3, 2.1, 1.9, 1.8, 2.1, 2.2, 1.7,

1.6, 2.2, 1 .8; 3 through 6; 2. 1, 1.7, 1.5, 2. 1; 7 through
16: 2.1, 2.0, 1.9, 2.1, 1.7, 1.6, 2.0, 1.7, 0.6, 1.3; 7
through 14: 2.1, 2.0, 2.0, 2.1, 1.7, 1.6, 2.0, 1.7; 11
through 14: 1.7, 1.7, 2.2, 1.8). Comparison with
measurements of the reference samples (Table 1)
reveals that measurements of these fossils fall within
the overlapping ranges of variation of extant B. car-
olinensis and B. hylophaga. The same conclusion
was reached during analysis of the dentaries from
Mt. Scott. Results of canonical analyses of the ten
most complete specimens (one observation was hid-
den) are shown in Fig. 26. We tentatively retain the
previous assignment of the Mt. Scott material to the
species B. carolinensis.

V2

Fig. 26. — Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Mt. Scott l.f. Ellipses
enclose variation in samples of Holocene taxa of Blarina (see
Fig. 4).

1984

JONES ET AL. — BLARINA SYSTEMATICS

101

Robert

Two specimens of Blarina were identified in the Robert l.f. of
southwestern Kansas. This Wisconsinan fauna was recovered
from SW */», SW ‘A sec. 33, T34S, R29W, Meade County (locality
UM-K1-57).

The Robert l.f. consists of molluscs, fishes, herptiles, birds, and
mammals. No pollen has been found associated with the fauna.
All species are extant. Fifteen species of small mammals have
been identified; six species ( Sorex cinereus. S. palustris, B. brev-
icauda, Thomomys cf. T. talpoides, Microtus pennsylvanicus, and
Zapus cf. Z. hudsonius) no longer occur in Meade County (Schultz,
1967, 1969).

Members of the Robert l.f. suggest the presence of marshes
and meadows, and nearby woods and upland prairies (Schultz,
1969). A more equable climate, moister and cooler than at pres-
ent, is indicated (Schultz, 1967, 1969; Miller, 1975). Sympatry
of 1 1 species of the fossil fauna occurs in northeastern South
Dakota (Schultz, 1969). The number of shrew and vole species
also is correlated with mesic climate (Graham, 1976). A descrip-
tion of the modem climate and vegetation of the area is included
in the discussion of the Cudahy l.f.

The Robert is one of the more recent local faunas of south-
western Kansas (Zakrzewski, 1975a). A radiocarbon date of
11,100 ± 390 BP was reported by Schultz (1967, 1969), who
thought that the Robert l.f. could not be correlated with any of
the other faunas in the vicinity.

The remains of Blarina consist of an isolated in-
cisor (UMMP V46004) and a dentary with p4-ml
(UMMP V46022). This material was referred to B.
b. brevicauda by Schultz (1967). Measurements 19
and 20 (2.3 and 1.6) of the molar indicate that the
fossil pertains to the species B. brevicauda (Table
1). The two measurements are somewhat smaller
than the means of the three samples from Iowa,
Nebraska, and Manitoba (5, 12, and 18 in Table 1),
although they fall within the range of variation of
these three groups. Therefore, the previous assign-
ment of the Robert material to the subspecies B. b.
brevicauda tentatively is retained.

Williams

The Williams l.f. was recovered from Rice County in central
Kansas. A description of the fauna has not yet been published.
One new species, Sorex kansasensis, was described by McMullen
(1975). The stratigraphy and composition of the fauna suggest
that deposition occurred during more xeric conditions before the
Wisconsinan glacial (Semken, personal communication). Za-
krzewski (197 5zz) considered the fauna Illinoian.

Rice County now is part of the Great Bend Prairie physio-
graphic province (Self, 1978). Natural vegetation consists of blue-
stem-grama prairie north of the Arkansas, floodplain vegetation
along the river, and sand prairie to the south (Kuchler, 1974).
Mean annual precipitation is 61 to 71 cm (NOAA, 1974).

The genus Blarina is represented by a fragmented
dentary with m2-3 and an isolated incisor (UMMP
V60338). Measurements 21 through 24 of the mo-
lars are 2.0, 1.3, 1.6, and 1.1. These measurements

are most similar to those of Holocene B. brevicauda
(Table 1 ), and the fossil undoubtedly pertains to that
species. It might pertain to the subspecies B. b. brev-
icauda.

Welsh Cave

Welsh Cave is located west of the Cumberland Plateau ap-
proximately 5.6 km southwest ofTroy, Woodford County, Ken-
tucky (84°44'50"W and 37°52'25"N), at an elevation of about 270
m. A preliminary description of the cave and its fauna was pro-
vided by Guilday et al. (197 1). A subsequent change in ownership
of the cave has prevented further research at the site.

Welsh Cave is a “small and relatively featureless” sinkhole
that was formed by the collapse of a more extensive cave system.
Bones generally were in good condition although there was evi-
dence of washing. The surface of the deposit contained mixed
Holocene and fossil material. Materials deeper in the deposit
appeared contemporaneous although no stratification was ob-
served. A carbon- 14 date of 12,950 ± 550 years BP (obtained
from a sample of Platygonus bone) was considered applicable to
most members of the fauna (Guilday et al., 1971).

Twenty-five mammalian species were identified from the Welsh
Cave deposit. Of these, four species now are extinct and thirteen
no longer occur in the area. The most common mammal was the
thirteen-lined ground squirrel ( Spermophilus tridecemlineatus ),
which was a common species in late Pleistocene faunas in the
Appalachians but currently is restricted to the prairies of central
North America (Guilday et al., 1971). The absence of Tamias
and the presence of Spermophilus and Geomys bursarius indicate
the existence of grassland communities in the area at the time of
deposition (Guilday et al., 1 97 1 ). A northern element, consisting
primarily of species of the Canadian/Hudsonian Zone, also was
present in the fossil fauna. Of the 22 extant species represented,
the modem ranges of 19 overlap in an area in eastern Minnesota
and western Wisconsin. This area of overlap suggests "boreal
semi-prairie or parkland” in the Welsh Cave region at the time
of deposition (Guilday et al., 1971). In the analysis by Graham
(1976), the higher density of vole and shrew species during the
late Wisconsinan also was interpreted as evidence of more equa-
ble climate at the time of deposition. Presently, the area is sur-
rounded by farmland and is considered typical of the eastern
Carolinian biotic province as described by Dice in 1 943 (Guilday
et al., 1971).

Three genera of soricids were recovered from the
deposit. The genus Blarina is represented by a right
mandible with complete dentition (CM 20144), a
mandibular fragment, and an isolated incisor
(MNI = 2). The two jaws are of dissimilar size and
were assigned tentatively to the taxa B. b. kirtlandi
and B. b. brevicauda (Guilday et al., 1971). Graham
and Semken (1976) considered these animals iden-
tical with the B. kirtlandi (which they assumed was
a separate species) and B. b. brevicauda from Meyer
Cave, Illinois.

The right mandible was examined in the present
study. Measurements 17 through 26 (6.0, 4.6, 2.2,
1.5, 2.0, 1.2, 1.6, 1 . 1 , 4.4, and 2.7, respectively) were
compared to the means for those measurements for

102

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

the samples of extant animals. The specimen from
Welsh Cave was within the lower end of the range
of variation for B. b. brevicauda (approximately the
same size as B. b. manitobensis ), and undoubtedly
pertains to the species B. brevicauda.

Cata/pa Creek

One specimen of Blarina was recovered from deposits along
Catalpa Creek, Lowndes County, Mississippi. Extant and extinct
vertebrates have been recovered from the site. The deposits have
been reworked, so that the age of the specimen (Pleistocene or
Holocene) is impossible to determine (M. Frazier, in litt., 1980).
The specimen is housed in the Mississippi Museum of Natural
Science (locality cat 1, 47N, 45E).

The specimen is a right mandible with one molar.
Measurements 2 1 and 22(1.9 and 1.3, respectively)
were taken. Comparisons were made with the mea-
surements of samples of modern Blarina from the
southeastern states (Table 1). Measurements of the
specimen from Catalpa Creek fell within the range
of variation of the Holocene sample of B. caroli-
nensis from Alabama (sample 1) and were slightly
larger than the means for the sample of carolinensis
from Georgia (sample 3). The nearest site that in-
cluded Pleistocene specimens, Ladds Quarry (Geor-
gia), contained specimens similar in size to that from
Catalpa Creek. The specimen from Catalpa Creek
therefore is referred to the species B. carolinensis.

Bat Cave

Bat Cave is located in Pulaski County, Missouri, in the SE ’A,
NW 'A, NE 'A sec. 4, T36N, R12W (Hawksley et al., 1973). This
cave, together with the two subsequent ones, are three of 1 0 caves
in the Ozark Plateau of Missouri that are known to contain re-
mains of Pleistocene faunas (Saunders, 1977). Bat Cave is in Bear
Ridge, overlooking the Gasconade River.

Most of the bones were recovered from a crawlway approxi-
mately 150 m long (north of the main entrance and main upper
passage) known as Bone Passage, and from a room (the Devil’s
Kitchen) adjacent to Bone Passage. Bones of bats and fragments
of Ursus americamis were the only remains found outside of Bone
Passage and the Devil’s Kitchen. Possibly there were at least two
modes of deposition of the fossil material. Some of the remains,
including material from a short-faced bear, Arctodus simus. were
found in what seemed to be the original site of a “bear bed.” The
random distribution of other bones and the nature of the matrix
suggest alluvial deposition. The bones might have been carried
into the passage or moved within the passage after original de-
position by water. The species represented in the fauna and the
physical features of the cave indicate that animals entered the
cave of their own volition or were carried in by predators. The
presence of Cryptobranchus suggests the cave was flooded, in
which case some of the mammals might have drowned in the
cave (Hawksley et al., 1973).

Non-mammalian species represented in the Bat Cave fauna
are limited to a gastropod, three genera of amphibians, two snakes,
and a passerine. Fourteen families of mammals are represented.

of which three species are extinct and nine species currently occur
north of the Ozarks. Attempts to analyze pollen from Bone Pas-
sage were unsuccessful. Comparisons with the faunas of Boney
Spring, Missouri (with radiocarbon dates of 16,500 to 13,500
years BP), Crankshaft Cave, Missouri, and Meyer Cave, Illinois,
indicate that the fauna of Bat Cave accumulated in a relatively
short period of time (possibly 16,500 to 13,500 years BP) during
the late Wisconsinan (Hawksley et al., 1973).

Presently, the bluffs in this area are “covered with second or
third growth deciduous forests dominated by oaks and hicko-
ries.” Average annual precipitation is 91 to 1 12 cm (Hawksley
etal., 1973).

Unlike the faunas of Crankshaft and Zoo caves,
only one phenon of Blarina was identified from Bat
Cave. The three identified remains (MNI = 2) ini-
tially were referred to the subspecies B. b. brevicau-
da. The scarcity of material and the presence of only
one phenon of Blarina were interpreted as indicating
deposition during a limited period of relatively sta-
ble climate. The lack of a natural trap in this cave
probably contributed to the low number of soricids
in this fauna (Hawksley et al., 1973).

In the present study, two dentaries (CMSU 388
and 389) with m 1 -m3 present were examined. Mea-
surements 19-24 for these specimens were 2.3, 2.2;
1.6, 1.5; 2.0, 1.8; 1.4, 1.2; 1.7, 1.6; 1.0, 1.0, respec-
tively. Multivariate analyses were not run on the
two specimens; however, comparisons of these mea-
surements with those of Holocene specimens re-
vealed that the fossils are most similar to Holocene
shrews from Indiana, Ohio, New York, and Ontario
(Table 1, samples 4, 5, 13, and 19). Thus, the fossils
are most similar to the talpoides semispecies of B.
brevicauda and likely pertain to that species.

Boney Spring

Boney Spring is located 16 km south of Warsaw in Benton
County, Missouri (NW 'A, SW 'A, SW >A sec. 29, T39N, R22W)
at an elevation of 215 m. The spring is one of eight sites in the
Ozark region from which Blarina has been recovered (Saunders,
1977).

Boney and Trolinger springs are located in the lower Pomme
de Terre River valley. The region is one of transition between
oak-hickory forest and western plains (Saunders, 1977). Inter-
disciplinary research was performed in the area prior to inun-
dation by the Harry S. Truman Reservoir. Saunders (1977) de-
scribed the historical background, excavation, deposition, and
faunas of the two sites.

Remains of ostracods. insects, amphibians, reptiles, and mam-
mals were recovered from Boney Spring. The ostracod and herp-
tile faunas are modern in aspect. Saunders (1977) listed 22 mam-
malian taxa from the site. Five extinct species (Paramylodon
harlani, Castoroides ohioensis, Mammut americanum. Equus sp.,
and Tapirus sp.) are present, and M. americanum (MNI = 31)
is the most abundant species represented in the fossil fauna.
Microtus pennsylvanicus presently occurs north and northeast of
Boney Spring; it has been reported from northern Missouri but

1984

JONES ET AL. — B LARIN A SYSTEMATICS

103

not from Benton County (Saunders, 1977; Hall, 1981). Napaeo-
zapus insignis is found in forested or brushy areas in southeastern
Canada, the northeastern United States, and the Appalachians
(Saunders, 1977; Hall, 1981). The remaining species of mammals
are part of the modem fauna of the study area and are found in
prairie, forest border, oak-hickory forest, and bottomlands forest
and prairie habitats (Saunders, 1977).

The assemblage from Boney Spring represents at least two
communities, one aquatic and one terrestrial (Saunders, 1977).
The remains of terrestrial vertebrates and associated pollen sug-
gest the presence of spruce forest and grassland and cool, humid
conditions during the period of deposition (Mehringer et a!.,
1968, 1970; Hallbergetal, 1974; Saunders, 1977). Accumulation
of the Boney Spring fauna occurred approximately 1 6,200-1 3,600
BP (Saunders, 1977).

Twenty-eight specimens (MNI = 4) of Blarina
from Boney Spring were referred to the species B.
brevicauda by Saunders (1977). Fifteen specimens,
part of the collection of the Illinois State Museum
(E. H. Lindsay field numbers 309-312, 314-317,
474, and BS71), were examined. With the excep-
tions of one maxillary fragment (EHL 311) and one
dentary fragment (EHL BS7 1 ), all specimens consist
of edentulous fragments and isolated teeth. The
sample appears homogeneous. Measurements 3
through 6 of three P4s (2.3, 2.0, 1.4, 2.2; 2.4, 1.8,
1.3, 2.3; 2.3, 2.1, 1.5, 2.4), 7 through 10 of two Mis
(2.3, 2.1, 2.2, 2.7; 2.2, 2.2, 2.2, 2.5), 1 1 through 14
of two M2s (1.6, 1.6, 2.1, 1.9; 1.7, 1.6, 2.3, 2.0), 19
and 20 of four mis (2.4, 1.6; 2.2, 1.5; 2.2, 1.6; 2.1,
1.5), 21 and 22 of four m2s (2.0, 1.3; 1.8, 1.2; 1.9,
1 .2; 1 .9, 1 .2), and 23 and 24 of two m3s (1.6, 1.1;
1.7, 1.1) were taken. Measurements are similar to
those of specimens from Bat and Zoo caves and to
larger specimens from Crankshaft Cave (Table 2).
Comparisons with the measurements of reference
samples (Table 1) suggest referral of the Boney Spring
material to the talpoides semispecies of the Holo-
cene species B. brevicauda.

Brynjulfson Cave No. 1

The Brynjulfson caves are located 5 m above Bonne Femme
Creek, approximately 10 km southeast of Columbia, in Boone
County, Missouri (SW ‘A, NE 'A, SW 'A sec. 16, T47N, R21W).
The faunas of both caves were described in a report by Parmalee
and Oesch (1972), on which this summary is based.

Brynjulfson Cave No. 1 contained the older of the two faunas.
A radiocarbon date of 9,440 ± 760 BP was obtained from a
sample of bone. Gastropods, fishes, herptiles, birds, and mam-
mals were present. Remains of forty-five species of mammals
have been identified; nine species (Brachyprotoma sp., Canis
dims, cf. Megalonyx, Dasypus be 1 1 us, Tapirus sp., Platygonus
compressus, cf. Sangamona, Alces or Cervalces, and Svmbos sp.)
are extinct (Parmalee and Oesch, 1972; Saunders, 1977). Seven
species no longer occur in Missouri. Some of the remains prob-
ably represent inhabitants of the cave; bones also might have

Fig. 27. —Canonical analysis of measurements 3-10 of upper
teeth of specimens of Blarina from Brynjulfson Cave No. 1 (one
observation, represented by a triangle) and Brynjulfson Cave No.
2. Ellipses enclose variation in samples of Holocene taxa of Bla-
rina (see Fig. 3).

been carried in by water and by predators (Parmalee and Oesch,
1972).

The presence of Clethrionomys, Alces or Cervalces, Sanga-
mona, and Symbos suggest cool, moist boreal conditions during
the period of deposition. Plentiful remains of beaver indicate the
presence of open-water marsh. Grassland habitat also was in-
dicated (Parmalee and Oesch, 1972).

In the vicinity of the caves, forests and meadows are present
and swamps occur in the valley bottoms (Parmalee and Oesch,
1972). Mean annual precipitation at Columbia is 94 cm (NOAA,
1974).

Only two soricids were found in the Brynjulfson
deposits. At Brynjulfson Cave No. 1 , the genus Bla-
rina is represented by seven specimens (MNI = 2).
Parmalee and Oesch (1972) referred this material
to the subspecies B. b. brevicauda.

Fragments of two maxillaries and two dentaries
(housed at ISM) were examined for the present study.
The upper teeth of at least one maxillary (repre-
sented by the triangle in Fig. 27) appear large enough
to merit assignment to the brevicauda semispecies
of the species B. brevicauda. Measurements 3 through
6 (2.4, 2.0, 1.5, 2.3), measurements 19 through 23
of three lower molars (2.1, 1.4, 1.7, 1.1, 1.6), and
measurements 19 through 26 of three lower molars
and dentary (2.2, 1.5, 1.9, 1.2, 1.6, 1 .0, 4.2, 2.6) are
more similar to those of the talpoides semispecies
of B. brevicauda. These three specimens are com-
parable in size to the fossils from Brynjulfson
No. 2.

Brynjulfson Cave No. 2

This site is located approximately 137 m from Brynjulfson
Cave No. 1. The methods of accumulation were the same for
both.

104

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

V2

Fig. 28. — Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from Brynjulfson Cave No. 2.
Centroids enclose variation in samples of Holocene taxa of Bla-
rina (see Fig. 4).

A radiocarbon date of 2,460 ± 230 BP was obtained for a
sample of bone from this fauna, which consists of gastropods,
pelecypods, fishes, herptiles, birds, and mammals. Only three
( Dasypus bellus, Canis dims, and Platygonus compressus) of the
37 species of mammals are extinct (Parmalee and Oesch, 1972;
Saunders, 1977). The species Clethrionomys gapperi, Erethizon
dorsatum, and Antilocapra americana no longer occur in Mis-
souri. Fish remains probably indicate periodic flooding. Small
mammals are more numerous and boreal species less prevalent
than in the fauna of Cave No. 1 . Prairie, forest, and marsh hab-
itats are indicated (Parmalee and Oesch, 1972). The presence of
species such as the pronghorn and the northern pocket gopher
indicate a trend toward increasing aridity after the last glacial
retreat (Parmalee and Oesch, 1972).

The genus Blarina is represented by 1 16 remains
(MNI = 21) which were referred to the subspecies
B. b. brevicauda by Parmalee and Oesch (1972). For
the present study, nine maxillaries and 1 1 dentaries
were examined. Measurements 3 through 10 of eight
of the most complete maxillaries were taken. Results
of canonical analyses of these specimens (Fig. 27)
demonstrate the similarity of these measurements
with those of Holocene samples of the talpoides
semispecies of the species B. brevicauda. Results of
analyses of eleven dentaries are shown in Fig. 28;
these observations also are plotted within the cen-
troid showing the range of variation of modern B.
brevicauda. We consider the remains from the de-
posit at Brynjulfson No. 2 to pertain to that species.

Crankshaft Cave

Crankshaft Cave is located in the SW 'A, SW 'A, SE 'A, NE 'A,
NW 1 A sec. 9, T43N. R5E, in Jefferson County, Missouri (Vine-
yard, 1964, in Oesch, 1967). The entrance to the cave is a small
sinkhole that opens into a shaft approximately 1.5 m wide. The

floor of the cave lies 20 m below the entrance. The cave acts as
a natural trap for animals (Parmalee et al., 1969).

A preliminary list of the fauna of Crankshaft Cave was pub-
lished by Oesch in 1967; Parmalee et al. published a second
account in 1969. Remains of molluscs, herptiles, and birds were
identified as genera that inhabit the Ozark region today. Of the
mammals, seven (possibly eight) are extinct species typical of
late Pleistocene faunas. Remains of 1 6 Holocene species no longer
present in Missouri suggest boreal or short-grass prairie habitats
(Parmalee et al., 1969). Attempts were made to date the fauna
more precisely using carbon- 14 techniques on four samples of
bone. The dates obtained seemed too young when compared with
those from similar materials from other sites. Possibly the
groundwater in Crankshaft Cave altered the composition of the
bones and made them useless for dating (Parmalee et al., 1969).

The site presently is surrounded by steep hillsides covered with
mixed hardwood forest. Average annual precipitation is approx-
imately 80 cm (Parmalee et al., 1969).

The genus Blarina is the most abundant com-
ponent of the mammalian remains from Crankshaft
Cave, with at least 663 individuals represented (Par-
malee et al., 1969). Mandibles, teeth, and cranial
fragments of Blarina initially were referred to what
were assumed to be two Holocene subspecies, B. b.
brevicauda and B. b. carolinensis (Oesch, 1967; Par-
malee et al., 1969). The implication was that one
subspecies had geographically replaced the other as
temperature fluctuated with the passage of time.
Graham and Semken (1976) subsequently referred
these materials (CMSU, ISIV^ ifncatalogued mate-
rial) to the nominal taxa kirtlandi and carolinensis ,
which they assumed were contemporaneous species.

In the present study, univariate statistics for the
fossils from Crankshaft (CRANKS) are shown in
Table 2. Canonical analysis separated 22 maxillae
(Fig. 29) into two clusters. Smaller specimens plot-
ted within the centroid for B. hylophaga, whereas
larger specimens clustered primarily within the cen-
troid for the largest subspecies (B. b. brevicauda) of
the species B. brevicauda. The two species were
clearly discriminated by this analysis. Measure-
ments of lower teeth were analyzed twice (Figs. 30
and 31). Most specimens were plotted within the
centroids for B. b. brevicauda, the smaller subspe-
cies of B. brevicauda, and B. hylophaga. Four spec-
imens were plotted outside of the range of variation
for the Holocene samples of Blarina examined in
this study. Further analysis (Fig. 32), in which mea-
surements of upper teeth from Crankshaft were
compared directly with those of Holocene samples,
showed more clearly the similarities between the
Crankshaft material (two ellipses numbered 36) and
the modern B. brevicauda and B. hylophaga. Based
on these canonical analyses, the modern distribu-

1984

JONES ET Ah.- BLARINA SYSTEMATICS

105

V2

Fig. 29. — Canonical analysis of measurements 3-10 of upper
teeth of specimens of Blarina from the Crankshaft Cave l.f. Cen-
troids enclose variation in samples of Holocene taxa of Blarina
(see Fig. 3).

V2

Fig. 31. — Canonical analysis of measurements 19-22 of lower
teeth of specimens of Blarina from the Crankshaft Cave l.f. Cen-
troids enclose variation in samples of Holocene taxa of Blarina
(see Fig. 30).

tion of Blarina in Missouri, and variation in Ho-
locene samples of Blarina from Missouri (Moncrief
et al., 1 982), the fossils from Crankshaft Cave clearly
pertain to the species B. brevicauda and B. hylo-
phaga.

Trolinger Spring

Trolinger Spring is located 4.5 km southeast of Boney Spring
and 19.5 km southeast of Warsaw in Hickory County, Missouri
(NE '4, NW '/4, NE % sec. 9, T38N, R22W) at an elevation of
223 m. Like Boney Spring, Trolinger Spring is in the Pomme de
Terre River valley (Saunders, 1977).

The fauna of Trolinger Spring consists of 137 specimens rep-
resenting seven species of mammals (B. brevicauda, Peromyscus
spp., Synaptomys sp., Mammut americanum, Mammuthus sp..
Equus sp., and Symbos sp.). The most numerous remains in this

V2

Fig. 30. — Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Crankshaft Cave l.f. Cen-
troids enclose variation in samples of Holocene taxa of Blarina
(see Fig. 4).

deposit are those of the mastodont (MNI =11). The three extant
species occur in the modern fauna of Hickory County (Saunders,
1977).

Accumulation occurred in a peat bog between 29,000 and 34,000
BP (Saunders, 1977). The fossils and associated pollen suggest
deposition during the mid-Wisconsinan interstadial when open
pine parkland was predominant in the region (Mehringer et al.,
1970; Saunders, 1977). The pollen record from Boney and Trol-
inger springs shows a transition between approximately 32,000
and 13,500 BP from nonarboreal pollen and pine dominance to
spruce dominance to spruce with deciduous elements (Mehringer
et al., 1970).

1.41

1.35

1.29

MV1

1.23

1.17

1.0S

0.18 0.20 0.22 0.24 0.26 0.28 0.30

MV2

Fig. 32. —Canonical analysis comparing maxillary specimens from
the Crankshaft Cave l.f. with specimens from the Holocene ref-
erence samples. The Crankshaft specimens are designated by the
two ellipses numbered 36. Reference samples are numbered as
indicated in the text. Ellipses indicate one standard deviation
about the mean.

106

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

The Blarina material, like that of other small
mammals from Trolinger Spring, is poorly pre-
served. Two lower incisors (MNI = 1) were re-
covered and are housed at the Illinois State Museum
(E. H. Lindsay field numbers 150 and 151). The
specimens are referred to the species B. brevicauda,
following Saunders (1977).

Zoo Cave

Zoo Cave is located 1 .6 km northeast of Hilda, Taney County,
Missouri. The entrance to the cave is low and narrow, and opens
into a steeply descending crawlway approximately 6.1 m long.
The passage opens into a wide room known as the Armadillo
Room. A maze area is south of this room, and in the past there
probably were numerous openings from the maze area to the
outside. To the north of the Armadillo Room is an area known
as the Bone Passage, where most of the bone material has been
found. The remains might have been carried into the cave by
predators or by water; the fact that no articulated remains have
been found suggests the animals did not die in the cave (Hood
and Hawksley, 1975).

Fossil material in Zoo Cave was recovered from two deposits.
The older of the two deposits contained remains of large verte-
brates, including three extinct mammals. The presence of Da-
sypus bellus is considered an indication of deposition during a
warm period. Radiocarbon dates of specimens of the three extinct
species from other Pleistocene faunas were used to estimate the
period of deposition as approximately 9,000 to 1 3,000 years BP.
The younger deposit, in which stratification was present, con-
tained remains of small vertebrates and invertebrates. These an-
imals serve as the best indicators of climate, but several conflict-
ing communities are indicated by the array of gastropods and
vertebrates in this deposit. The second deposit probably is more
recent than 9,000 BP. Attempts at pollen analysis were unsuc-
cessful (Hood and Hawksley, 1975).

Remains of 59 species were identified in the fauna ofZoo Cave.
Non-mammalian representatives include gastropods (aquatic and
terrestrial), pelecypods, a crustacean, fishes, and herptiles. All
these animals occur in Taney County at present except the fox
snake, Elaphe vulpina (the modem range of which might include
parts of northern Missouri). The remains of fishes and of a hell-
bender ( Crypt obranchus ) indicate a permanent stream commu-
nity. The other herptiles include members of deciduous forest,
prairie, and streamside communities. Thirty-two species of mam-
mals were present, of which three now are extinct and 10 no
longer present in Taney County. One species for which the present
distribution is primarily western was present. Seven species (in-
cluding four shrews) now occur primarily to the north. The area
of current sympatry of the eight northern species (seven mammals
and the fox snake) was described as an area of mixed conifers,
hardwoods, savanna/parkland, and prairie in western Wisconsin
and eastern Minnesota (Hood and Hawksley, 1975).

Zoo Cave is located in the Mark Twain National Forest. The
site is surrounded by deciduous forest dominated by oaks and
hickories with numerous cedar glades (Hood and Hawksley. 1975).
Average annual precipitation is approximately 101 to 111 cm
(NOAA, 1974).

The soricids of Zoo Cave were neither as diverse
nor as numerous as those of Crankshaft Cave (Par-

V2

Fig. 33. — Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Zoo Cave l.f. Centroids
enclose variation in samples of Holocene taxa of Blarina (see
Fig. 4).

malee et al., 1 969). Remains of the short-tailed shrew
were referred to the nominal taxa B. b. brevicauda
(MNI = 4), B. b. carolinensis (MNI = 1), and B. b.
kirtlandi (MNI = 5). Based on earlier assignment of
Holocene shrews in southern Missouri to the taxon
carolinensis. Hood and Hawksley (1975) surmised
that the presence of the more northerly taxa brevi-
cauda and kirtlandi (as well as Sorex arcticus and
S. cinereus ) indicated deposition during a cooler
climatic period.

Nine dentaries (CMSU, uncatalogued) with three
molars present were analyzed for the present study.
Of these, one specimen plotted within the area of
overlap of the centroids for B. carolinensis, B. hy-
lophaga, and the smaller subspecies ( talpoides semi-
species) of B. brevicauda (Fig. 33). The remainder
were plotted within the centroids for B. b. brevicau-
da and the smaller subspecies of B. brevicauda. Based
on these data, it cannot be determined unequivo-
cally that these specimens represent more than one
species. Tentatively, based on the work by Hood
and Hawksley (1975) and Moncriefet al. (1982), the
Blarina of this fauna is referred to the species B.
brevicauda.

Doby Springs

The Pleistocene fauna of Doby Springs was collected at five
localities approximately 15 km west of Buffalo, Oklahoma (N 'A,
SW XA sec. 10, T27N, R24W, Harper County). Fossils were found
in lake sediments deposited in a collapse basin Illinoian in age.
Stephens (1960) described the geology and local fauna.

The Doby Springs l.f. consists of ostracods, molluscs, fishes,
herptiles, birds, and mammals. Smith (1958), Etheridge (1961),
Brattstrom (1967), and Jammot (1972) discussed the fishes, liz-

1984

JONES ET AL.- BLARINA SYSTEMATICS

107

ards, snakes, and shrews ( Sorex ) of the fauna. Stephens (1960)
and Hibbard (1963) listed 24 species of mammals tentatively
identified from the site. Doby Springs is the type locality for
Peromyscus oklahomensis\ seven additional extinct species (Cas-
toroides cf. C. ohioensis, Peromyscus cf. P. cochrani, Mammu-
thus sp., Camelops sp.. Bison cf. B. latifrons, Equus cf. E. nio-
brarensis, and E. sp.) were reported.

Stephens (1960) considered the Doby Springs l.f. Illinoian in
age on the basis of its stratigraphy and composition of species.
The Doby Springs l.f. was compared with the Berends l.f. (Okla-
homa) and other Pleistocene faunas in southwestern Kansas, and
was found most similar to the Berends l.f. (Stephens, 1960). The
Doby Springs and Berends l.f ’s both were considered early Illi-
noian by Hibbard (1970).

The fauna and associated pollen suggest a cooler climate (with
more effective moisture than at present) and the presence of lake,
marsh, lowland meadow, and shrub and tree communities during
the period of deposition (Smith, 1958; Stephens, 1960; Etheridge,
1961; Brattstrom, 1967; Hibbard, 1963, 1970; Kapp, 1970).
Sympatry of extant species of fishes and mammals occurs in the
southeastern comer of North Dakota (Stephens, 1960).

One specimen, UMMP 34752, was referred to the
species B. brevicauda by Stephens (1960). The spec-
imen is a partial left maxillary with P2 and P4; it
was illustrated in Fig. 5 by Stephens (1960). This
fossil was noted as larger than Holocene material
examined at the University of Michigan. Hibbard
(1970) referred the fossil to B. b.fossilis on the basis
of size. For the present study measurements 3
through 6 (2.6, 2.2, 1.5, and 2.4) were taken and
were compared with measurements of the reference
samples (Table 1). The measurements of the spec-
imen from Doby Springs fall within the range of
measurements of samples of B. brevicauda. They
are most similar to the means of samples 5 and 12
of B. b. brevicauda, and the specimen tentatively is
assigned to that subspecies.

Bootlegger Sink

Bootlegger Sink is located 0.4 km east of Emigsville, York
County, Pennsylvania (40°0T00"N and 76°43'30"W) at an ele-
vation of approximately 107 m. Guilday et al. (1966) described
the sinkhole and its fauna.

Remains were recovered from dolomite breccia which occurred
as a ledge projecting from the walls of the sinkhole. The ledge is
a remnant of surface-derived talus that once filled the sink (Guil-
day et al., 1966). Attempts were made to date the deposit by
fluorine and radiocarbon analyses. Results of the fluorine analysis
suggest continuous deposition as the climate changed from boreal
to temperate. The carbon- 14 date (3,722 ± 200 BP) is the min-
imum date for brecciation. Pollen analyses were negative. Based
on these analyses and the composition of the fauna, accumulation
probably occurred from the late Wisconsinan (10,000 to 15,000
BP) through 3,000 or 4,000 years ago (Guilday et al., 1966).

Invertebrate remains from the deposit at Bootlegger Sink were
limited to two species of millipede and one snail. Amphibian
material consisted of remains of an unidentified salamander and
five species of toads and frogs. One species of box turtle, Ter-

rapene Carolina, was recovered and was considered to imply tem-
perate conditions during the later period of accumulation. Six
species of snakes were represented; all occur in the region today
except the mole snake, Lampropeltis calligaster rhombomacu-
lata, which has a present distribution in the southeastern United
States (Guilday et al., 1966).

Remains of 37 species of mammals, representing six orders
and a minimum of 97 individuals, were identified from Boot-
legger Sink. Eight species no longer occur in York County. The
thirteen-lined ground squirrel, Spermophilus tndecemlineatus,
currently is restricted primarily to the plains of central North
America. The long-eared bat, Plecotus sp., presently occurs in
the southern half of the United States; it and the mole snake were
the only species identified from the sink deposit that do not occur
as far north as the site today (Guilday et al., 1966). Other species
represented that are no longer present in York County — Sorex
arcticus, Glaucomys sabrinus. Synaptomys borealis, Microtus
xanthognathus, Napaeozapus insignis, and Rangifer cf. R. tar-
andus— occur to the north of the site (Hall, 1981). Sympatry of
the two voles, M. xanthognathus and M. chrotorrhinus, suggests
boreal conditions during at least part of the period of deposition.
Bootlegger Sink is one of seven Appalachian sites at which the
two species were collected together (Guilday, 1971; Guilday et
al., 1977). The remaining species occur in the area today. Such
a mixed fauna suggests deposition over a long period of time, as
indicated by the fluorine analysis, during varying climatic con-
ditions (Guilday et a!., 1966).

Bootlegger Sink is in the Limestone Valley Section of the Pied-
mont Province. The site is included in the Carolinian biotic
province as described by Dice (1943), but the original deciduous
forest has been removed and the valley now is farmed. Mean
annual precipitation is 101.8 cm (Guilday et al., 1966).

After the preliminary excavation, the genus Bla-
rina was represented by three mandibles, a humerus,
and isolated teeth (MNI = 2). This material (CM
7924) was referred to the species B. brevicauda by
Guilday et al. (1966). Bootlegger Sink was among
the eleven local faunas examined by Graham (1976)
in which greater species density of shrews and voles
was interpreted as an indication of more equable
climate during the Pleistocene. He also indicated
that the Bootlegger material pertained to B. brevi-
cauda.

Measurements 3 through 10 of two maxillaries
(2.8, 2.0, 1.5, 2.8, 2.4, 2.2, 2.5, 2.9; 2.7, 1.9, 1.7,

2.7, 2.4, 2.3, 2.5, 2.7), 1 7 through 24 of two dentaries
(6.3, 5.1, 2.5, 1.7, 2.0, 1.4, 1.6, 1.1; 6.1, 4.9, 2.5,

1.8, 2.0, 1 .4, 1.6, 1.1), and 1 7 through 23 of a third
dentary (5.9, 4.6, 2.1, 1.5, 2.2, 1.4, 1.5) were taken
for the present study (CM 7924 and uncatalogued
material). Measurements of these specimens were
compared with those of Holocene specimens. Al-
though measurements of the fossils vary, they are
most similar to means derived from the two samples
(5 and 12) of extant B. brevicauda from Nebraska
and Iowa. The fossil sample appears homogeneous.

108

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

New Paris No. 4

New Paris Sinkhole No. 4 is located 2.4 km northeast of New
Paris (at 40°5'N and 78°39'W), Bedford County, Pennsylvania.
Guilday et al. ( 1 964) published a detailed description of the sink-
hole and its fauna, on which this report is based.

Originally the sinkhole was a bell-shaped domepit, formed by
circulation of groundwater in the limestone underlying Chestnut
Ridge, on which the sinkhole is located. A narrow vertical fissure,
10 m in depth, was located alongside the domepit and connected
to it by an underground passage. In the past this fissure was open
to the surface, and the fossils recovered from this site were re-
mains of animals that fell down this fissure and were trapped.
The large domepit also was opened to the surface by the collapse
of its ceiling, but apparently animals were not trapped because
of its large opening and rapid rate of fill. Subsequently both the
fissure and the domepit were filled by surface-derived material;
no stratification was evident. The rate of accumulation of the fill
material evidently was faster in the early stages, when the local
climate was cooler, than in the final stages. Preservation generally
was good; many skeletons were recovered uncrushed and artic-
ulated although evidence of disturbance by rodents was noted.
A radiocarbon date of 1 1,300 ± 1,000 BP was obtained from a
sample of charcoal (Guilday et al., 1964).

The fauna of the New Pans Sinkhole No. 4 consists of inver-
tebrates (millipedes and snails), amphibians, reptiles, birds, and
mammals. The invertebrates all are present in the area today,
and it was difficult to separate Pleistocene and Holocene speci-
mens. More than 2,700 vertebrates were recovered. Of the am-
phibians (salamanders, toads, and frogs) all species occur in the
area today with the exception of the northern leopard frog (with
modem records from northwestern Pennsylvania) and a tenta-
tively identified subspecies of toad (Bufo americanus copei) of
the Hudson Bay area (Guilday et al., 1964; Conant, 1975). The
snakes represented in the deposit are species currently in the area.
The absence of lizards and turtles was due to the boreal climate
of the area during the time of deposition (Guilday et al., 1964).
Avian remains are scarce (MNI = 9), and all but one of the seven
species present search for food on the ground. This material
includes remains of the sharp-tailed grouse (Pedioecetes phasi-
anellus) and the ruffed grouse (Bonasa umbellus). The presence
of these birds suggests the forest canopy was closed during the
later stages of deposition, although it is possible that the two
species were not in the immediate area contemporaneously (Guil-
day et al., 1 964). Remains of these two species also were reported
from the Virginian deposits of Back Creek Cave No. 2, Clark’s
Cave, and Natural Chimneys (Guilday et al., 1977).

The mammalian fauna consists of 40 species of insectivores,
bats, rodents, lagomorphs, carnivores, and artiodactyls. The one
artiodactyl ( Mylohyus nasutus, the long-nosed peccary) was the
only extinct species present, although an extinct subspecies of
squirrel ( Tamiasciurus hudsonicus tenuidens) also was present.
Most mammals still are present in the state but not necessarily
at the site. Seven species (Sorex arcticus. S. tridecemlineatus,
Dicrostonyx hudsonius. Phenacomys cf. ungava, Svnaptomys bo-
realis. Microtus xanthognathus, and Mustela nivalis) no longer
occur in the state. The ground squirrel (S. tridecemlineatus) pres-
ently ranges northward to the edge of the Canadian biotic prov-
ince, and is an indicator of grassland near the site during the
period of deposition. The collared lemming (D. hudsonius) in-
dicates the presence of nearby tundra, possibly on Allegheny
Mountain or on the crest of Chestnut Ridge (Guilday et al., 1 964).
The five remaining species presently occur in a variety of boreal

habitats in the Hudsonian and Canadian zones (Guilday, 1971;
Guilday et al., 1964; Hallberg et al., 1974). The fauna, correlated
with analysis of pollen from the deposit, shows a gradual tran-
sition from boreal parkland to boreal forest; both vertebrate and
pollen remains show more temperate elements above the 6 m
level. All mammals from the bottom meter of the excavation
occur at least in part in the Hudsonian or Canadian biotic prov-
inces (Guilday et al., 1964).

The sinkhole is located at an altitude of 465 m. The area
presently is included in the Ridge and Valley physiographic prov-
ince. In Pennsylvania, this province is characterized by alter-
nating forested ridges and farmed valleys. The region is not rugged
but is sufficiently mountainous that distribution of temperature
and precipitation is not uniform (NOAA, 1974). Chestnut Ridge
lies within the rain shadow of Allegheny Mountain. The site is
relatively arid and jack pine may have been present during the
late Pleistocene (Guilday et al., 1964). The biota generally is
transitional between the Carolinian and Canadian biotic prov-
inces of Dice (1943), but Guilday et al. (1964) pointed out the
local variability in topography which influences biotic distribu-
tion.

Guilday et al. (1964) referred specimens of Bla-
rina from New Paris No. 4 (MNI = 37, 38.9% of
all soricids present) to the taxa B. b. brevicauda and
B. b. kirtlandi. The short-tailed shrews were the most
common insectivores in the upper and middle levels
of the deposit, but were absent from the lower level.
Average size of the remains increased toward the
bottom of the deposit, probably due to increasing
contamination by modern B. b. kirtlandi in the up-
per levels. The larger B/arina, “presumably repre-
senting a large northern stock that advanced to the
south with mounting glacial conditions,” were com-
parable to the remains found at Robinson Cave,
Natural Chimneys, and Clark’s Cave (Guilday et
al., 1969, 1977). The stratigraphy of the B/arina
remains was considered supporting evidence for the
trend from boreal parkland to boreal forest indi-
cated by pollen and remains of other vertebrates
(Guilday et al., 1964). Graham and Semken (1976)
detected what they thought were three phena of Bla-
rina in the material from New Paris No. 4; they
interpreted these as sympatric taxa equivalent to
extant brevicauda, carolinensis , and kirtlandi. The
taxon carolinensis apparently was based on one
specimen, which was not seen in the present study.
Graham (1976) considered the greater species den-
sity of shrews during the Wisconsin as evidence for
“a more equable climate with reduced temperature
and moisture gradients,” which correlates well with
the paleoecological hypotheses of Guilday et al.
(1964).

For the present analysis, nine specimens (CM
5843, 5845, 7437, and 7802) were examined. Uni-
variate statistics are shown in Table 2. Six maxil-

1984

JONES ET AL.- BLARINA SYSTEM ATICS

109

vz

Fig. 34. — Canonical analysis of measurements 3-10 of upper
teeth of specimens of Blarina from the New Paris No. 4 l.f.
Centroids enclose variation in samples of Holocene taxa of Bla-
rina (see Fig. 3).

laries fell within the range of variation for modem
samples of B. brevicauda (Fig. 34). One maxillary
that had incomplete measurements appeared iden-
tical to the specimens illustrated in Fig. 34. The two
dentaries examined had measurements that com-
pared most favorably with modern specimens of B.
brevicauda from Nebraska, Iowa, and Manitoba.
Accordingly, all specimens from New Paris No. 4
examined for this study are referred to the species
B. brevicauda, but we cannot say unequivocally that
only one species of Blarina is represented in the
fauna.

Baker Bluff Cave

Remains of Blarina have been recovered from several cave
sites in Tennessee, three of which are located in Sullivan County
(Appendix I and Fig. 2). Sullivan County is included in the Valley
and Ridge physiographic region, which Miller (1974) described
as “characterized by numerous elongate ridges and intervening
valleys, all trending in a northeast-southwest direction.” The area
is in the Carolinian biotic province of Dice (1943). The region
is covered by timbered ridges and rolling farmland; annual pre-
cipitation at Knoxville is 1 19 cm (Guilday et al, 1978).

Baker Bluff Cave is located 13 km southeast of Kingsport, at
36°27'30"N and 82°28'W, Boone Dam quadrangle U.S.G.S. I'h'
topographic map (Guilday et al., 1978). The cave is a large cham-
ber, approximately 10 m long, located at an elevation of 450 m
in a bluff above the South Fork of the Holston River. A detailed
description of the cave and its fauna was provided by Guilday
et al. (1978).

A 3 m column was excavated, from which remains of 1 80 taxa
of invertebrates and vertebrates were recovered. Stratigraphic
levels were designated arbitrarily. The herptiles of the Baker Bluff
l.f. are present in the region today. Two species of birds and 10
species of mammals do not occur presently in the area. Six species
of mammals are extinct. Thirty-one per cent of the mammalian
species have modem ranges in boreal forest, temperate grassland,

V2

Fig. 35.— Canonical analysis of measurements 3-10 of upper
teeth of specimens of Blarina from the Baker Bluff Cave l.f.
Centroids enclose variation in samples of Holocene taxa of Bla-
rina (see Fig. 3).

or in higher elevations in the southern Appalachians. In the faunal
analysis by Guilday et al. (1978), it was noted that the smaller
mammals found in cooler areas today were more abundant in
the upper levels of the deposit.

Carbon- 14 dates were obtained from bone fragments from
three stratigraphic levels in the deposit. The basal date was 1 9, 1 00
± 850 BP. The dates might be too young due to contamination
by plant roots and burrowing by rodents. Based on analysis of
distribution of the small mammals in the deposit. Guilday et al.
(1978) hypothesized that deposition occurred during a transition
from cool-temperate to boreal conditions.

Guilday et al. (1978) identified seven species of
shrews, including two phena of Blarina , from Baker
Bluff Cave. Specimens referred to B. b. kirtlandi
were found in all stratigraphic levels. The larger B.
b. brevicauda occurred in the upper levels (3 to 7
feet) and was absent from the lower three feet of the
deposit. The latter phenon was considered compa-
rable to B. b. cf. brevicauda from Clark’s Cave, Vir-
ginia. As hypothesized by Graham and Semken
(1976), Guilday et al. (1978) assumed that coexis-
tence of the two phena in the higher levels of the
deposit indicated cooler, more equable climate at
the time of deposition.

For the present study, 140 maxillary and dentary
fragments (CM 29746-53) were examined. Univar-
iate statistics for these specimens are shown in Table
2. The individuals shown in Figures 35 and 36 were
separated stratigraphically: a from the 3'-4' level, b
from the 4-5 level, c from the 5'-6' level, d from
the 6'— 7' level, e from the 7'-8' level, f from the 8'-
9' level, and g from the 9'— 1 O' level. Seventeen max-
illary fragments were sufficiently complete for anal-
ysis of upper tooth measurements (Fig. 35). Two

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

1 10

V2

Fig. 36. — Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Baker Bluff Cave l.f.
Centroids enclose variation in samples of Flolocene taxa of Bla-
rina (see Fig. 4).

observations (from strata c and d) were hidden with-
in the centroid for the talpoides semispecies of B.
brevicauda. Although sample size from each strati-
graphic level is small, the plotted specimens show
the occurrence of large specimens of B. brevicauda
in strata b and d and smaller specimens of B. brev-
icauda in all strata shown except b. Results of anal-
ysis of measurements of lower teeth are shown in
Fig. 36. Twenty-three observations (from strata c,
d, f, and g) are hidden, and one observation from
stratum d fell outside the range of the figure (to the
right of the centroid for the talpoides semispecies of
B. brevicauda). The specimens plotted in the cen-
troid for B. b. brevicauda in Fig. 36 are from strata
a, b, c, and d, which are the uppermost strata of the
deposit (3'— 7'). Smaller specimens of B. brevicauda
occurred in all six strata. These results correspond
well with those reported by Guilday et al. (1978).
There is no doubt that all specimens represented
pertain to the species B. brevicauda.

Guy Wilson Cave

Guy Wilson Cave is located approximately 4 km south of BlufT
City, Sullivan County, Tennessee (Corgan, 1976). The cave is on
the south side of the South Fork of the Holston River.

Corgan (1976) reported that at least 25 species of mammals
from this locality have been identified by J. E. Guilday. Nine
species (of the orders Carnivora, Proboscidea, Perissodactyla,
Artiodactyla, Edentata, and Rodentia) have been mentioned in
publication. Remains from three caves in Sullivan County (Guy
Wilson, Baker BlufT, and Beartown) provide the southernmost
record of the caribou Rangifer tarandus in eastern North America
(Guilday et al., 1975). Guy Wilson, Robinson. Baker Bluff, and
Carrier Quarry caves (all in Sullivan County) are noted by Guil-
day (1971) as sites that indicate the southernmost distribution

1.60

V2

Fig. 37. — Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Guy Wilson Cave l.f.
Centroids enclose variation in samples of Flolocene taxa of Bla-
rina (see Fig. 4).

of boreal coniferous forest. A late Wisconsinan date (19,700 ±
600 BP) for Guy Wilson Cave was obtained from carbon- 14
dating of a sample of peccary bone (Guilday et al.. 1975).

We examined 1 8 specimens ( 1 6 dentaries and two
maxillaries) of Blarina from Guy Wilson Cave. Of
these, only five dentaries were sufficiently complete
for analysis of measurements 19-24 of the lower
teeth. The five specimens fell within the range of B.
brevicauda (Fig. 37). Measurements of eight of the
remaining dentaries fell within the range of varia-
tion for those measurements of the samples of extant
B. b. kirtlandi (samples 4 and 6). Two other den-
taries had two measurements slightly larger than
those of extant animals, and one specimen appeared
smaller than modern kirtlandi in five measure-
ments. Measurements of the two maxillaries com-
pared favorably with the means of those measure-
ments of specimens from samples 4 and 6. The
species represented is B. brevicauda.

Riverside Cave

Riverside Cave is located in Sullivan County, Tennessee. Its
fauna has not been described formally. Remains consist largely
of isolated teeth and fragmented jaws, probably accumulated
from the activities of owls and woodrats. The remains probably
include mixed late Wisconsinan and Holocene material (J. E.
Guilday, in litt., 1982). Two genera in the fauna, Sangamona
and Mylohyus, now are extinct. The material identified as San-
gamona (J. E. Guilday, in litt.) might pertain to the fugitive deer,
S’, fugitiva. which has been identified from eight deposits (prob-
ably all Wisconsinan) in the central and eastern United States,
including the Robinson and Whitesburg local faunas ofTennessee
(Kurten and Anderson, 1980). The long-nosed peccary, M. na-
sutus. has been reported from 35 Irvingtonian and Ranchola-
brean faunas from the southeastern quarter of the United States

1984

JONES ET AL .-BLARINA SYSTEM ATICS

1 i 1

Fig. 38. — Canonical analysis of measurements 3-10 of upper
teeth of specimens of Blarina from the Riverside Cave l.f. Cen-
troids enclose variation in samples of Holocene taxa of Blarina
(see Fig. 3).

(Kurten and Anderson, 1980). In Tennessee this species probably
was associated with moist forests and wooded valleys. Remains
also were identified as Tapirus sp. (J. E. Guilday, in litt., 1982),
two species of which have been recorded from eight sites in
eastern Tennessee and which may have been associated with
deciduous forest (Corgan, 1976). The remaining taxa identified
from the deposit are Sorex cinereus, S. fumeus, Blarina, Paras-
calops breweri, Myotis sp., Eptesicus fuscus, Neotoma floridana,
Peromyscus sp., Clethrionomys gappen, Synaptomys cooperi,
Microtus (Pitymys) cf. pinetorum, Tamias striatus, Sciurus sp.,
Glaucomys cf. volans, Zapus hudsonius, cf. Sylvilagns sp., Ursus
americanus, Procyon lotor, Spilogale putorius. Odocoileus virgin-
ianus, and Cervus elaphus (J. E. Guilday, in litt., 1982). All of
these species have occurred in eastern Tennessee in modern times
(Hall, 1981).

Twenty-seven specimens of Blarina (CM, uncata-
logued material) from the Riverside Cave l.f. were
examined. The results of the analysis of maxillary
fragments are shown in Fig. 38; one incomplete
specimen had measurements similar to those of the
four specimens shown. These specimens were plot-
ted within or near the centroid showing the range
of variation for modern specimens of the smaller
subspecies ( talpoides semispecies) of B. brevicauda.
Twenty-two dentaries were examined, nine of which
also were plotted within the range of variation of
the talpoides semispecies of B. brevicauda and one
near the centroid for B. b. brevicauda (Fig. 39). One
observation was hidden. With the exception of the
one large specimen, the measurements of dentaries
were comparable with the means of those measure-
ments for Holocene samples, and were most similar
to the Holocene subspecies manitobensis and chur-
chi (samples 18 and 11). Given the small sample

V2

Fig. 39. — Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Riverside Cave l.f. Cen-
troids enclose variation in samples of Holocene taxa of Blarina
(see Fig. 4).

size, specimens from this deposit appear more sim-
ilar to fossils from Baker Bluff than to specimens
from the Guy Wilson and Robinson caves, and ap-
parently represent only the species B. brevicauda.

Robinson Cave

Robinson Cave is located approximately 13 km southwest of
Livingston, Overton County, Tennessee (36°17'25"N and
85°22'25"W). It is in the east slope of Maxwell Mountain at an
elevation of 366 m (Guilday et al., 1969).

Although bones were found in three sites in the cave, almost
all were discovered in the locality known as the Armadillo Pit.
The floor of Armadillo Pit consists of talus from an extinct sink-
hole. Guilday et al. (1969) thought deposition in Armadillo Pit
might have occurred in three stages. The first stage was repre-
sented by remains of armadillo and large mammals. A late San-
gamonian or early Wisconsinan age was suggested by fluorine
dating techniques and the presence of Dasypus bellus during this
phase of deposition. The second phase was indicated by the de-
position of bats and other small vertebrates; the sinkhole might
have begun to close so that large animals were not trapped. During
the third phase the ceiling was closed, allowing entrance only
through the main cave. Remains accumulated during the third
period of deposition were uncrushed and articulated. The time
of deposition in the second and third stages is thought to be late
or post-Wisconsinan, although a precise date has not been ob-
tained.

Most of the animals from Armadillo Pit were bats and wood-
rats. The other animals entered the cave by their own volition,
were trapped in the sinkhole, or were deposited by predators.
"Remains of at least 2,615 individual mammals, 14 individual
birds, 10 species of reptiles and amphibians, and 461 individual
gastropods were recovered from approximately 5 square yards
of matrix from the floor of this pit . . .” (Guilday et al., 1969).
Six of the vertebrate species are extinct. Forty-nine of the 54
living species are now present in Minnesota and Wisconsin, sug-
gesting a boreal-temperate parkland in the area at the time of
deposition (Guilday et al., 1969). Graham (1976) also noted the

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

112

V2

Fig. 40. — Canonical analysis of measurements 3-10 of upper
teeth of specimens of Blarina from the Robinson Cave l.f. Cen-
troids enclose variation in samples of Holocene taxa of Blarina
(see Fig. 3).

V2

Fig. 41. — Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Robinson Cave l.f. Cen-
troids enclose variation in samples of Holocene taxa of Blarina
(see Fig. 4).

boreal aspect of the fauna, and suggested a more equable climate
at the time of deposition. Guilday (1971) listed Robinson as one
of four Appalachian sites that might indicate the southernmost
distribution of dense coniferous forests. No mammals that are
presently southern in distribution were found.

The site is within the Eastern Highland Rim physiographic
region, which was characterized by Miller (1974) as nearly level
and underlain by limestone. Guilday et al. (1969) described the
immediate vicinity of the cave as an area of broad, cultivated
valleys surrounded by uplands “covered with second or third
growth deciduous woodland . . . Mean annual precipitation is
approximately 142 cm (NOAA, 1974).

Specimens of Blarina (MNI = 83) from Robinson
Cave were identified as B. brevicauda. It was noted
that they are similar in size to the subspecies B. b.
brevicauda presently in the Minnesota-Wisconsin
area (Guilday et al., 1 969). Graham (1976) also con-
sidered the animals B. brevicauda.

For the present study, 28 specimens of Blarina
(CM 8133, 8134, 8146) were examined. Measure-
ments of the upper teeth of four specimens (Fig. 40)
were analyzed; these specimens fell within the range
of variation for Holocene B. b. brevicauda from Iowa
and Nebraska. In the analysis of lower teeth, two
specimens plotted within the centroid for the small-
er talpoides semispecies of B. brevicauda, whereas
other specimens (including four hidden observa-
tions) plotted near or within the range of variation
of Holocene B. b. brevicauda (Fig. 41). Univariate
statistics for fossils from Robinson Cave (ROBINS)
are shown in Table 2. These results are in keeping
with the general boreal aspect of the fauna of Rob-
inson Cave, and indicate that only one species, B.
brevicauda ( brevicauda semi species), was represent-
ed.

Barton Springs Road Site

The Barton Springs Road Site is located in Travis County,
Texas. It is one of several sites on the Edwards Plateau from
which late Pleistocene faunas have been recovered. All of the
Texan sites from which Blarina have been recovered (Appendix
1), except Ben Franklin, Easley Ranch, Howard Ranch, Rex
Rodgers, and the Vera faunule, are located on the Edwards Pla-
teau.

The Edwards Plateau was included in the Comanchian biotic
province by Dice (1943). He described the vegetation on the
plateau as “oaks and junipers . . . often mixed with . . . grass and
mesquite, and on steep rocky slopes these trees may form closed
stands. It is only rarely that any of them reach a height of over
twenty feet. In many places on the plateau the characteristic
vegetation is grass, and trees and shrubs are absent or occur only
in very open stands.” Ford and Van Auken (1982) further de-
scribed the vegetation on the southeastern part of the plateau.
The canyons along the eastern and southern edges are cooler and
more humid than the rest of the plateau. They provide ecological
diversity today and might have done so in the Wisconsinan (Lun-
delius, 1967; Dalquest et al., 1969).

In his analysis of the late Wisconsinan faunas and the paleo-
ecology of central Texas, Lundelius (1967) assigned the species
of those faunas to one of three groups. The first group consisted
of extinct species considered characteristic of the Pleistocene
(Mammut americanum and P/atygonus, for example), some of
which survived until 7,000 or 6,000 years ago. The second group
consisted of 1 1 extant species that are not present now in central
Texas or which occur there as relict populations, including Sorex
cinereus and what he termed Blarina brevicauda. These species
presently occur northeast or northwest of the plateau. The third
group consists of species present in the area today. Almost all
extant species in Texas are represented, with the major exceptions
of Dasvpus novemcinctus and Tavassu tajacu. The Pleistocene
local faunas of Texas appear “to be a reflection of a temperature
gradient in a north-south direction and an east-west change from
forest to grassland or savanna” (Lundelius, 1967).

The short-tailed shrew was noted as “abundant
in most deposits of Wisconsin age in central Texas”

1984

JONES ET AL. — B LARINA SYSTEM ATICS

113

(Lundelius, 1967). He referred all material from the
deposits of central Texas to the subspecies B. b.
carolinensis, following Hibbard (1963). Lundelius
(1967) noted the gradual disappearance of Blarina
from west to east across the Edwards Plateau; the
last western record was from Felton Cave 7,800
years ago. Hall (1982) discussed additional evidence
of increasing dryness in the post-Wisconsinan of
Texas and Oklahoma.

Lundelius (1967) listed 18 species of mammals
recovered from the noncultural unit (3,450 ± 150
BP) and 17 species from the cultural unit (1,015 ±
105 BP) of the Barton Springs site. All species are
extant although not all presently occur in the area
(Davis, 1978).

The genus Blarina is represented in the Barton
Springs fauna by a single specimen (TMM 40627-
60). Measurements 3 through 16 (2.0, 1.5, 1.3, 1.9,
1.8, 1.7, 1.9, 2.0, 1.5, 1.4, 2.0, 1.6, 0.7, and 1.2,
respectively) of this specimen are most similar to
those of fossils from Cave Without a Name and
Felton Cave. These measurements fall within the
lower end of the range of measurements of the ref-
erence samples of B. carolinensis (Table 1). The
specimen from Barton Springs tentatively is referred
to the species B. carolinensis.

Cave Without a Name

Cave Without a Name is located approximately 18 km north-
east of Boeme, Kendall County, Texas. Commercial develop-
ment of the cave was described by Craun (1948). Lundelius (1967)
listed thirty species that have been recovered from this site. Five
extinct species and subspecies ( Dasypus bellus, Procyon lotor si-
mas, Mammui americanum, an ovibovine, and Equus sp.) were
recognized from this deposit. The extinct armadillo < I). bellus)
first appeared in Blancan deposits in Florida, and was identified
from Wisconsinan faunas in Missouri, Tennessee, West Virginia,
Arkansas, Georgia, and New Mexico, in addition to four other
sites in Texas {Kurten and Anderson, 1980). This species might
have been an ecological equivalent of the extant /). novemcinctus
(Lundelius, 1967). In the Miller’s Cave deposit, D. bellus was
found in the travertine unit and D. novemcinctus was found in
the more recent clay unit (Patton, 1963 b). Remains of a raccoon
were referred to an extinct subspecies ( P. t. simus ) first identified
from the Pleistocene of northern California (Kurten and Ander-
son, 1980). Wisconsinan records of the mastodont (M. ameri-
canum) extend from Alaska to Florida. Mastodonts were most
numerous in eastern forest but also inhabited valleys and low-
lands in Texas (Kurten and Anderson, 1980). Six additional
species, including what was termed B. brevicauda. were identified
that have present distributions north of the site, generally in
cooler, wetter climates (Patton, 1963a, Lundelius, 1967). Gra-
ham (1976) also suggested a more equable climate at the site
during the period of deposition based on higher species density
of shrews and voles at the cave during the Wisconsinan.

Kendall County is located along the southeastern escarpment

Fig. 42. — Canonical analysis of measurements 3-10 of upper
teeth of specimens of Blarina from the Cave Without a Name
l.f. Centroids enclose variation in samples of Holocene taxa of
Blarina (see Fig. 3).

of the Edwards Plateau. This region was included in the Com-
anchian biotic province by Dice (1943); his description of the
vegetation of the plateau is included in the account for Barton
Springs. Ford and Van Auken (1982) provided additional infor-
mation regarding the riparian vegetation of Kendall County.

A radiocarbon date of 10,900 ± 190 BP was obtained from a
sample of bone from Cave Without a Name. The site has the
last record of Sorex cinereus, Microtus pennsylvanicus, and Mus-
tela erminea on the plateau. There was an apparent trend for the
species presently distributed farthest north to disappear the ear-
liest from central Texas (Lundelius, 1967).

The short-tailed shrews from Cave Without a
Name were referred to the subspecies B. b. caroli-
nensis by Lundelius ( 1 967). In his analysis, Graham
(1976) also recognized the relatively small size of
the remains and referred them to B. carolinensis.

Ten specimens of maxillary fragments were ex-
amined in the present study (TMM 40450-1254
through 40450-1256, 40450-1260, 40450-1261, and
40450-1263 through 40450-1267). This sample ap-
peared homogeneous. Six specimens were sufficient-
ly complete for analysis (Fig. 42). They clustered
within the centroid showing the range of variation
for modern B. carolinensis, and are referred to that
species.

Easley Ranch

Easley Ranch is located in western Foard County, Texas. The
faunal remains were recovered from a sinkhole on the Easley
Ranch, the type locality of the Good Creek formation (Dalquest,
1962).

The fauna from Easley Ranch has not been dated. Its com-
position is similar to that of the Jmglebob l.f. (Dalquest, 1962).
The Easley Ranch l.f. has been considered Sangamonian (Dalquest,
1962; Slaughter, 1967) and Wisconsinan (Hibbard, 1970) in age.

The mammalian fauna includes 1 1 extinct species. Of the 19

1 14

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

extant species, 13 presently live in the region (Dalquest, 1962).
The Easley Ranch l.f. suggests moderate summers and winters
at the time of deposition, with woodland, grassland, and stream-
bank habitats (Dalquest, 1962; Slaughter, 1967).

Blarina cf. brevicauda is represented in the fauna
by a single incisor (MWU 3119).

Felton Cave

Felton Cave is located in Sutton County, Texas, on the western
portion of the Edwards Plateau. The fauna is dated at 7,770 ±
130 BP (Lundelius, 1967). Fourteen species of mammals have
been recovered from the site; seven species ( B . brevicauda, Cryp-
totis parva, Geomys bursarius, IPeromyscus nasutus, Baiomys
taylori, Microtus sp., and Bison sp.) no longer occur in Sutton
County (Lundelius, 1967; Davis, 1978).

Felton Cave is the last record of Blarina from the western part
of the Edwards Plateau. Short-tailed shrews were present in most
of the fossil faunas from the western side of the plateau until
approximately 1 ,000 BP, reflecting the trend of increasing warm-
ing and drying west to east across the plateau (Lundelius, 1967).

The Blarina of Felton Cave were referred to the
species B. b. carolinensis by Lundelius (1967) and
B. carolinensis by Graham (1976). In the present
study 10 specimens (TMM 41 1 74- 1 9 through 41 1 74-
28), consisting of two maxillaries, seven dentaries,
and an isolated incisor, were examined. All speci-
mens were fragmented and had incomplete denti-
tion. Measurements were most similar to those of
specimens from Barton Springs, Cave Without a
Name, and Longhorn Cavern. The specimens are
referred to the species B. carolinensis.

Friesenhahn Cave

Friesenhahn Cave is located near the southeastern edge of the
Edwards Plateau, approximately 34 km north of San Antonio in
Bexar County. The cave is a solution cavern developed in the
limestone beds underlying the Edwards Plateau. It consists of an
underground chamber, approximately 18 by 9 m, which is con-
nected to the surface by a vertical opening ranging from 2 to 3
m in diameter and approximately 9 m deep. A second, inclined
entrance was open to the surface during the late Pleistocene, but
has filled with debris since that time (Evans, 1961).

Remains were recovered from four distinct units of fill, which
were described by Evans (1961). The oldest unit (zone 1) con-
sisted of a mixture of limestone blocks, gravels, and partially
cemented red clay; remains of turtle shell and of small mammals
were the only vertebrate fossils recovered. Zone 2 consisted main-
ly of clay that was deposited in an underground pond; vertebrate
remains from this unit included articulated specimens. The third
zone, in the central and southeastern parts of the chamber, was
the most fossiliferous of the four units although preservation was
not as good as that in zone 2. The fourth zone was limited to a
channel cut by water flowing into the cave; small vertebrates were
abundant, and some remains were carried in from the older de-
posits within the cave (Evans, 1961). In addition, Sellards (1919)
and Hay (1920) described the recovery of bones of turtles, pro-
boscideans, saber-toothed cats, and other vertebrates from the
surface of the cave floor.

The freshwater pond that once occurred in the main chamber
of the cave might have attracted many of the animals represented
in the cave deposit. Some animals died within the cave; many
others were dragged in by predators. Eventually the entrance was
closed by debris, and the cave was sealed until the opening of
the sinkhole entrance in modem times (Evans, 1961). Most of
the fauna accumulated during the Wisconsinan (zones 2 and 3)
during relatively moist conditions. The age of the material from
zones 1 and 4 is uncertain. Remains from zone 1 probably date
from the early Wisconsinan. Fossils in zone 4, excluding those
washed in from older deposits, might date from the late Wis-
consinan or early Holocene; they represent a period somewhat
more moist than present but drier than the period represented
in zones 2 and 3 (Evans, 1961).

Amphibians, reptiles, and mammals are represented in the
deposit at Friesenhahn Cave. The herptiles (five species of am-
phibians and five reptiles) pertain to Holocene species with the
exception of a new species of Testudo that was described by
Milstead (1956). The remaining herps now are associated in the
Balcoman biotic province, and no extreme faunal shifts in recent
times are indicated (Mecham, 1959). Evans (1961) noted the
presence of turtle remains throughout the deposit, particularly in
zone 3.

Lundelius (1967) listed thirty-two species of mammals from
Friesenhahn Cave. Included are eight extinct species— Canis
( Aenocyon ) dims, Arciodus pristinus, Smilodon sp., Homother-
ium ( Dinobastis ) serus, Mammut americanum, Mammuthus sp.,
Mylohyus sp., and Camelops sp. Two extinct species of Tapirus
and Equus were present also. Remains of the predators (C. dims,
A. pristinus, Smilodon, and H. serus) were found associated with
those of herbivores, probably indicating that the cave regularly
served as a den (Evans, 1961). Remains of the mammoth and
mastodont were numerous and consisted almost wholly of im-
mature individuals, possibly representing a favorite prey of the
cats (Evans, 1961; Meade, 1961). These fossils were found pri-
marily in zones 2 and 3. Bones of Mylohyus were recovered from
zones 2, 3, and 4, and their association with remains of species
such as the tapir and the mastodont suggests the presence of
forests near the deposit (Lundelius, 1960). Members of the genus
Camelops ranged throughout western North America and were
extinct by the late Wisconsinan or early Holocene, as were other
members of the North American megafauna (Kurten and An-
derson, 1980). Blarina brevicauda, Peromyscus nasutus, Cyno-
mys ludovicianus, and possibly Microtus were the only extant
species listed by Lundelius (196.7) that are not present in the
region today. Subsequent absence of P. nasutus was attributed
to increased aridity since the Wisconsinan (Tamsitt, 1957). Ad-
ditional studies of the small mammals of Friesenhahn Cave were
conducted by Kennedy (1956) and Pettus (1956).

Specimens of Blarina from Friesenhahn Cave were
referred to the subspecies B. b. carolinensis by Lun-
delius (1967). In the present study one maxillary
fragment (TMM 933-4008) and seven dentary frag-
ments (TMM 933-4005 through 933-4009 and 933-
4011) were examined. Measurements 3 through 14
of the maxillary (2.2, 1.5, 1.2, 1.9, 2.0, 1.8, 2.0, 2.2,
1.6, 1.6, 2.1, and 1.9, respectively) compare most
favorably with those of the modern taxa B. caroli-
nensis and B. hylophaga plumbea (samples 1,3, 10,

1984

JONES ET AL.- B LARINA SYSTEM ATICS

1 15

V2

Fig. 43. — Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Friesenhahn Cave l.f.
Centroids enclose variation in samples of Holocene taxa of Bla-
rina (see Fig. 4).

15, 16, and 2 1 ) and are slightly smaller than modern
B. h. hylophaga (samples 1 7 and 22). Five of the
dentaries that were sufficiently complete for analysis
plotted within the ellipse showing the range of vari-
ation of modern B. carolinensis (Fig. 43). The two
specimens with incomplete measurements were
comparable to the specimens shown. The fossils ten-
tatively are assigned to B. carolinensis.

Hall’s Cave

Hall’s Cave is located in Kerr County, Texas. Kerr County is
part of the southern edge of the Edwards Plateau; modem vege-
tation of the area was discussed in the account for Barton Springs.

The fauna of Hall’s Cave has not been studied. No carbon
dates are available, but the fauna is considered Holocene by E.
Lundelius, Jr. Bailing wire and barbed wire were found with some
of the remains. Artifacts have not been found in the older part
of the sequence, but there are no indications of a Pleistocene
fauna (E. Lundelius, Jr., in litt. , 1983).

One specimen (TMM 4 1 229-452) of Blarina from
Hall’s Cave was examined. The specimen consists
of a right dentary with i and ml-m3. Measurements
19 through 26 (1.9, 1.3, 1.6, 1.1, 1.2, 0.8, 3.1, and
2.0) were taken. These measurements fall within the
range of variation of Holocene samples of B. hy-
lophaga and B. carolinensis (Table 1). The mea-
surements are most similar to the means of samples
of B. carolinensis, however, and the specimen from
Hall’s Cave is assigned tentatively to that species.

Howard Ranch

The Howard Ranch l.f. was recovered from several deposits
along Groesbeck Creek, northwest of the town of Quanah, in
Hardeman County, Texas. A list of the mammals now present

in the Groesbeck Creek area was included in the description of
the Howard Ranch l.f. by Dalquest (1965).

Fossils were recovered from the Groesbeck formation, which
consists of sediments deposited during the late Wisconsinan. All
of the deposits from which fossils were collected were from the
basin of an extinct lake. A carbon- 14 date of 16,775 ± 565 BP
was obtained from a sample of shells taken from below the ma-
jority of the vertebrate material. The presence of large extinct
mammals in the fauna also suggests a late Wisconsinan age.

The Howard Ranch l.f. consists of the remains of arthropods
(crayfishes), molluscs, fishes, amphibians, reptiles, birds, and
mammals. Vertebrate remains generally were fragmented and
scattered although the quality of preservation varied. Fishes and
amphibians were the most common vertebrates recovered and
mammals were the least common. Forty species of mammals
were identified, seven of which are extinct — Mammuthus (Ele-
phas ) columbi, Equus cf. scotti, E. cf. conversidens. Equus sp.,
Camelops cf. hesternus, Tanupolama sp., and Bison cf. antiquus.
Remains of the mammoth, horses, and camels were fairly com-
mon. Twelve of the remaining mammalian species are extant but
are not present in the Groesbeck Creek area today. Besides Bla-
rina, these include two species of Sorex, Thomomys talpoides,
Oryzomys palustris. Reithrodontomys megalotis, Microtus penn-
sylvanicus, M. ochrogaster and/or M. pinetorum, Ondatra zi-
bethicus, and Synaptomys cooperi. These mammals presently are
distributed to the north, east, and southeast of the Howard Ranch
area. The coexistence of these small mammals during the late
Pleistocene might have resulted from a more equable climate.
The presence of the Wisconsinan ice sheet would have produced
cooler summers and warmer winters as far south as southern
Texas (Dalquest, 1965). The slightly younger fauna of the C2
layer of Schulze Cave was considered additional support for this
thesis (Dalquest et al., 1969).

The area of the site is located near the northern boundary of
the Comanchian biotic province (Dice, 1943). Presently summers
are hot and winters short and mild; average annual precipitation
is between 50 and 61 cm (NOAA, 1974).

Dalquest (1965) described two maxillaries and a
mandible of the short-tailed shrew from deposits at
Howard Ranch. He thought that the three specimens
were larger than the modern races of B. hrevicauda.
Dalquest referred these specimens to B. b. fossi/is,
following Hibbard (1943).

One maxillary and one dentary (MWU 3083) were
examined. Measurements of the fossils were com-
pared to measurements obtained from Holocene
specimens (Table 1). The fossils were smaller than
the large B. b. hrevicauda collected from Iowa and
Nebraska (samples 5 and 12). They were more sim-
ilar to the smaller subspecies of Blarina brevicau-
da—churchi, kirtlandi, manitobensis, and talpoides
(samples 1 1, 4, 6, 18, 13, 19, and 24). They almost
certainly pertain to the species B. hrevicauda.

Klein Cave

Klein Cave is located approximately 19 km WSW of Mountain
Home, Kerr County, Texas. The site is a shallow limestone cave
located near the southern edge of the Edwards Plateau. The mod-

116

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

ern vegetation of the area was discussed in the account for Barton
Springs. A description of KJein Cave and its mammalian fauna
published by Roth (1972) serves as the basis for this account.

The cave is approximately 64 m long, with an entrance ranging
in diameter from 4 to 1 1 m. Roth (1972) identified four layers
of sediment in the cave. Bones of extant Holocene rodents, bats,
and rabbits were found in the second layer. Fossils were located
in a pocket along the south wall of the cave. Few remains of large
mammals were discovered. The bones generally were fragmented,
and the deposit probably accumulated due to activities of owls
and woodrats in the cave. A minimum date of 7,683 ± 643 was
obtained from a sample of bone (Roth, 1972).

Remains of amphibians, reptiles, birds, and mammals have
been recovered from the cave. Roth (1972) listed 43 mammalian
species identified from the deposit. Eight species of bats were
included, making Klein Cave the richest bat fauna reported from
Texas. No mammals were extinct and all were known from other
cave faunas representing the late Pleistocene. Ten species with
northern and eastern distributions are no longer present on or
near the Edwards Plateau. Four of these (Tamias striatus, Mi-
crotus pennsylvanicus, Synaptomys cooperi, and Mustela ermi-
nea ) no longer occur in Texas. All of these previously were re-
ported from at least one other fauna on the Edwards Plateau: T.
striatus from the Schulze Cave fauna (Dalquest et al., 1969), M.
pennsylvanicus from Cave Without a Name (Lundelius, 1967),
S. cooperi from Cave Without a Name, Longhorn Cavern, and
Miller’s Cave (Patton, 1963a). and M. erminea from Schulze
Cave (Dalquest et al., 1969) and Cave Without a Name (Lun-
delius, 1967). These four species, whose present ranges extend
no closer to the Edwards Plateau than the mountains of New
Mexico, are indicative of cooler, moister climate during the pe-
riod of deposition. More equable climate, with lush grasslands
and deciduous forests extending from eastern Texas, was sug-
gested, similar to that indicated by Dalquest et al. (1969) for the
fauna of Schulze Cave (Roth, 1972).

Roth (1972) listed six skulls, 5 upper jaws, and
1 7 lower jaws of what he termed Blarina brevicauda
from the deposit at Klein Cave. Of this material,
six maxillary fragments and 1 3 mandibles (MWU
9120) were examined. Results of the analysis of the
maxillaries are shown in Fig. 44; one specimen had
incomplete (but comparable) measurements and is
not shown. The five specimens plotted in or near
the area of overlap of the ellipses representing B.
carolinensis, B. h. hylophaga, and the smaller sub-
species of B. brevicauda. Four dentaries were in-
complete but were comparable to the nine speci-
mens included in the analysis and shown in Fig. 45.
With the exception of two smaller individuals, these
specimens also plotted in the area of overlap be-
tween B. carolinensis, B. hylophaga hylophaga, and
B. brevicauda. Comparisons were made between the
measurements of the 1 9 individuals from Klein Cave
and the means derived for those measurements for
Holocene samples. The specimens from Klein Cave
were similar to those of modern groups of B. car-
olinensis (samples 1, 3, 10, 15, and 16) and B. h.

V2

Fig. 44.— Canonical analysis of measurements 3-10 of upper
teeth of specimens of Blarina from the KJein Cave l.f. Centroids
enclose variation in samples of Holocene taxa of Blarina (see
Fig. 3).

hylophaga (17 and 22). We suspect that the speci-
mens from Klein Cave represent both B. carolinen-
sis and B. hylophaga, which possibly are sympatric
at the present time in northwestern Louisiana and
are difficult to distinguish (George et al., 1981), but
the present data are inadequate to be certain.

Longhorn Cavern

Longhorn Cavern is located approximately 14 km southwest
of Burnet, Burnet County, Texas, on the eastern half of the Ed-
wards Plateau. The cavern was formed by phreatic solution. Fos-
sils were discovered, and many were lost, during commercial
development of the cavern (described by Craun, 1948). Geology
of Longhorn Cavern was described by Semken (1961) and Mat-
thews (1963). Semken (1961) described the local fauna and stra-

V2

Fig. 45.— Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Klein Cave l.f. Centroids
enclose variation in samples of Holocene taxa of Blarina (see
Fig. 4).

1984

JONES ET AL. — B LARINA SYSTEM ATICS

117

tigraphy; the modern climate and fauna of the area also were
discussed.

Vertebrates were collected from three stratigraphic units within
Longhorn Cavern (Semken, 1961). The floor of the cavern is
overlain by a basal breccia, which was barren of fossils except
for a fragment of a horse-tooth. Red fill, the most extensive unit
in the cavern, overlies the basal breccia and contained remains
of 1 3 genera of mammals and one extinct species of turtle. Black
fill rests in “old stream cuts and in topographic lows” in the red
fill (Semken, 1961). Remains of 16 genera of extant mammals
and of an unidentified fish were recovered from the black fill.
The presence of Mus musculus indicates deposition of the black
fill within the last 200 years; deposition probably occurred since
the 1860’s (Semken, 1961). A fourth fill, the Longhorn breccia,
was identified in non-commercial areas of the cavern. This brec-
cia, probably derived from the red fill, contained remains of two
extinct species of turtles, three extinct species of mammals, and
sixteen extant genera of mammals (Semken, 1961). Accumula-
tion of the Longhorn l.f. (remains from the red fill and the Long-
horn breccia) and of remains from the black fill probably was
due to the activities of carnivores and raptorial birds. Some an-
imals also may have fallen into sinkholes (Semken, 1961).

The Longhorn l.f. was considered late Wisconsinan in age by
Semken (1961) and Hibbard (1970). Pitymys pinetorum, now
restricted in Texas to a relict population near Kerrville and to
the eastern portion of the state (Bryant, 1941; Davis, 1978), is
the most common member of the l.f. The composition of the
Longhorn l.f. suggests the presence of mixed forests and grass-
lands at the time of deposition (Semken, 1961; Patton, 1963a).

Lundelius (1967) reported the presence of B. b.
carolinensis in the Longhorn l.f. Two specimens
(TMM 40279-163 and 40279-164), both from the
red fill, were examined. Measurements 19 through
25 of the first specimen were 1.8, 1.2, 1.6, 1.0, 1.3,
0.9, and 3.4; measurements 19 through 22 of the
second specimen were 1.8, 1.2, 1.5, and 1.0. These
measurements are somewhat smaller than the range
of measurements for samples 1 7 and 22 of extant
B. hylophaga (Table 1). The measurements of the
fossils from Longhorn Cavern fall within the range
of those of extant B. carolinensis (particularly the
smaller taxa represented in samples 2, 10, 15, and
23) and tentatively are assigned to that species.

Miller’s Cave

Miller’s Cave is located 21 km southeast of Llano in Llano
County, Texas. The site is in the Riley Mountains, at an elevation
of 412 m, overlooking Honey Creek. This area is in the central
part of the Edwards Plateau. Dice’s (1943) description of the
vegetation of the plateau is included in the account for Barton
Springs. Patton ( 1 9636) published the description of the geology
and local fauna of Miller’s Cave, on which this report is based.

Miller’s Cave consists of two chambers approximately 55 m
in combined length. Sediments of the chambers were derived
from different sources, so each has a different stratigraphic se-
quence. The present entrance is a shaft opened by collapse of
part of the roof at the junction of the two chambers. Few fossils
were discovered in the north chamber. Most of the l.f. was re-

covered from two stratigraphic units within the south chamber.
The brown clay unit, consisting of a loose clay-silt conglomerate,
contained fossils of bats, armadillo, rabbits, and rodents. A ra-
diocarbon date of 3,008 ± 410 BP was obtained from a sample
of charcoal from this unit. A travertine unit located directly below
the brown clay unit contained fossils of insectivores, bats, ar-
madillos, rabbits, rodents, carnivores, and deer. Analysis of the
composition of the travertine indicated deposition during a cool,
moist period. Bone from the travertine unit was dated at 7,200 ±
300 BP. Few remains were found below the travertine (Patton,
19636).

Unidentified fishes, amphibians, reptiles, and birds were re-
covered from the brown clay and travertine units. Twenty-two
genera and twenty species of mammals were identified tenta-
tively. Remains from the brown clay unit consisted of Myotis
velifer, Dasypus novemcinctus, Sylvilagus sp., Geomys bursarius,
Perognathus hispidus, Reithrodontomys megalotis, Peromyscus
leucopus, Peromyscus sp., Neotoma Jloridana, and Microtus och-
rogaster. All of these species, except D. novemcinctus and an
unidentified bat. also were recovered from the travertine unit. In
addition, the travertine yielded remains of what was termed Bla-
rina brevicauda, and of Cryptotis parva, Scalopus aquaticus, Ep-
tesicus fuscus, Lepus californicus, Neotoma Jloridana, Synapto-
mys cooperi, Ondatra zibethicus, Canis sp., Ursus americanus.
Mephitis mephitis, Spilogaie putorius, and Odocoileus sp. The
only extinct species in the deposit, Dasypus bellus, also was re-
covered from this layer. Patton (19636) considered this date to
mark the minimum terminal date for D. bellus. Lundelius (1967)
regarded this date (7,300 BP) as probably inaccurate. Some species
listed (B. brevicauda, N. Jloridana, O. zibethicus, S. cooperi, and
M. ochrogaster) have modem distributions north, northeast, and
southeast of the Edwards Plateau. Miller’s Cave is the last known
record of S. cooperi (7,300 BP) and M. ochrogaster (3,000 BP).
They seem to have disappeared from the region after the dis-
appearance of the megafauna and the small mammals that pres-
ently are distributed the farthest north from central Texas (Lun-
delius, 1967). Lundelius considered the order of disappearance
due to the warming and drying of the climate to be from west to
east across the plateau. Patton ( 1 9636) suggested there was a more
equable climate at the time of deposition, with stream and marsh,
lowland meadow, shrub and tree, and upland prairie commu-
nities in the vicinity. A warming trend was indicated by the
mineralogy and the change in number and composition of the
fauna of the two units (Patton, 19636).

The short-tailed shrews of the Miller’s Cave de-
posit were identified as B. brevicauda by Patton
(19637>). They were noted as falling within the size
range of what he termed B. b. carolinensis from
Kansas and Oklahoma. Ten specimens, all mandib-
ular fragments with incomplete dentition, were ex-
amined (TMM 40540-20, 40540-229, 40540-497,
40540-499, 40540-501 through 40540-505, and
40540-507). Measurements were compared with
means derived for those measurements from mod-
ern samples. The fossils were most similar to the
modem samples of B. carolinensis, B. h. hylophaga,
and B. h. plumbea (samples 1,3, 10, 15, 16, 17, 22,
and 21). Probably both B. hylophaga and B. caro-
linensis are represented in Miller’s Cave l.f.

118

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Schulze Cave

Schulze Cave is located approximately 45 km northeast of
Rocksprings, Edwards County, Texas. The cave is a sinkhole near
the southern edge of the Edwards Plateau. The Holocene and
Pleistocene faunas of the area were described by Dalquest et al.
(1969).

Holocene vertebrate material was recovered from underneath
the vertical opening. Early Holocene and Pleistocene remains
were recovered from a travertine ledge within the shaft (layers B
and C) and from debris accumulated at the bottom (E). Radiocar-
bon dates were obtained from samples of bone from layers Cl
(9,680 ± 700 BP) and C2 (9,3 10 ± 3 10 BP), indicating deposition
from approximately 1 1,000 to 8,000 years BP (Dalquest et al.,
1969). Most of the remains of small mammals probably accu-
mulated as the result of activities of barn owls. Bones of larger
mammals probably were from animals that fell into the sinkhole
or were washed in or carried in by woodrats (Dalquest et al.,
1969).

Remains from Schulze Cave represent 82 species of mammals,
2 1 of which have not been recorded from the area in modem
times. All taxa listed from layer B are present today on the Ed-
wards Plateau, although not necessarily in the immediate vicinity
of the sinkhole. Layers C 1 and C2 included taxa no longer present
in central Texas. Generally layers Cl and C2 contained the same
taxa, but species typical of northern, cooler areas were more
common in C2, and species typical of warm or semiarid areas
were more common in the upper, younger layer Cl (Dalquest et
al., 1969). The only extinct species represented in the local fauna
are mammoth, horse, and possibly bison. Extant taxa no longer
present on the Edwards Plateau included what was termed Ma-
rina brevicauda. These taxa presently are distributed in the south-
eastern United States and southern Rocky Mountains. Dalquest
(1965) thought that sympatry of these species in central Texas
during the Pleistocene might have resulted from the Wisconsinan
ice sheet, which produced cooler summers and warmer winters.
More equable climate during the period of deposition also was
suggested by Graham (1976) based on his analysis of species
density of shrews and voles. Dalquest et al. (1969) assumed that
a flora and fauna similar to that of an alpine meadow existed in
the area during the late Pleistocene. Deciduous woodland was
present in the canyons and in some of the open areas of the
plateau, and probably was continuous with the woodland in east-
ern Texas (Dalquest et al., 1969).

Specimens of Blarina from Schulze Cave, iden-
tified as B. brevicauda, were recovered from layer
C. Remains of 79 animals were recovered from layer
Cl and 83 from C2, making Blarina the most com-
mon of the four shrews from the deposit (Dalquest
et al., 1969). Dalquest et al. (1969) noted that the
average size of the specimens was the same in both
layers, but that there was considerable variation in
size among the specimens. “The difference in size
between the largest and smallest jaws from both
layers suggest that two species might be involved,
but the measurements do not show a bimodal dis-
tribution and indicate a single rather variable pop-
ulation. All but the smallest specimens are too large
for the small races found in southeastern Texas to-

Fig. 46. — Canonical analysis of measurements 3-10 of upper
teeth of specimens of Blarina from the Schulze Cave l.f. Centroids
enclose variation in samples of Holocene taxa of Blarina (see
Fig. 3).

day; their average size is similar to the average of
B. b. carolinensis [=B. hylophaga hylophaga ], from
northeastern Texas, Oklahoma, and Kansas”
(Dalquest et al., 1969).

Specimens from the Cl layer (MWU 7258) were
available for examination in the present study. Elev-
en maxillaries and 42 jaws were measured. Ten of
the maxillary specimens were sufficiently complete
for canonical analysis and are shown in Fig. 46; the
remaining incomplete specimen is comparable to
those shown. Most of these specimens are somewhat
larger than the modern samples of B. carolinensis.
They are shown within or near the two centroids
that show the range of variation for modern samples
of B. h. hylophaga and the smaller subspecies of B.
brevicauda. Twenty-nine mandibles were analyzed
and are plotted in Fig. 47; eight observations are
hidden. These specimens fell in or near the centroid
showing the range of variation of modern B. h. hy-
lophaga from Kansas and Oklahoma (samples 17
and 22). We tentatively conclude that only one
species, B. hylophaga, is represented in Schulze Cave.

Back Creek Cave No. 2

Fossil Blarina have been recovered from several caves in Vir-
ginia and West Virginia (Fig. 2). This area is included in the
Carolinian biotic province of Dice (1943). However, Guilday
(1962 b) and Guilday et al. (1977) described how variable the
topography and biota are in this mountainous region.

Back Creek Cave, also known as Sheets Cave, is located 24
km west of Clark’s Cave in Bath County, Virginia. The site was
described as a shallow “rock shelter” by Guilday et al. (1977),
and is within the Ridge and Valley physiographic province of
western Virginia. The deposit is comparable in age (Ranchola-

1984

JONES ET AL .-BLARINA SYSTEMATICS

119

l.hO

V2

Fig. 47. — Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Schulze Cave l.f. Centroids
enclose variation in samples of Holocene taxa of Blarina (see
Fig. 4).

brean) to the deposits of Natural Chimneys and Clark’s Cave
(Guilday et al., 1977).

Fossil remains are from a raptor roost deposit with a fauna
similar to that of Clark’s Cave. Avian remains representing 1 7
species (MNI = 55) were included in the faunal list of Clark’s
Cave by Guilday et al. (1977). In several instances identification
was tentative because of the small sample size and generally poor
preservation of bird bones. Most of the birds, such as bobwhite
and woodpeckers, are present in the state today. The spruce grouse
(Canachites canadensis), ruffed grouse ( Bonasa umbellus), and
sharp-tailed grouse ( Pedioecetes phasianellus) are among the nine
species of birds common to the deposits from Back Creek Cave,
Clark’s Cave, and Natural Chimneys. These three species pres-
ently are distributed in Canada; the grouse are boreal species that
currently reside in forests and open glades, whereas the rock
ptarmigan, Lagopus mutus, indicates open or semi-open tundra
vegetation (Guilday et al., 1977). This type of avian fauna implies
open coniferous parkland, perhaps “open, tundra-like ridge crests
and more heavily wooded intermontane valley parklands or bog-
forests” (Guilday et al., 1977).

The mammalian fauna of Back Creek Cave No. 2 is still under
study. Five species were mentioned by Guilday et al. ( 1 977). The
least chipmunk, Eutamias minimus, was identified from the Back
Creek Cave and Clark’s Cave deposits, which are the first records
of this species in the Appalachians. It currently occurs “in open
to brushy, boreal, coniferous forest situations, and reaches its
greatest abundance in open, sandy, pine, and spruce parklands”
(Guilday et al., 1977). The heather vole, Phenacomys interme-
dius, and the yellow-cheeked vole, Microtus xanthognathus, pres-
ently occur in boreal communities in northern Canada. The
northern bog lemming (Synaptomys borealis) ranges through bo-
real forests and taiga, in a variety of habitats, no further south
than Minnesota and New Hampshire. The least weasel, Mustela
nivalis, has a circumboreal distribution and might occur in the
mountainous areas of Virginia. The voles, lemming, and weasel
also were reported from the deposits of Natural Chimneys and
New Paris No. 4 (Guilday et al., 1977).

Fifteen specimens of Blarina from Back Creek
Cave No. 2 were examined (CM, uncatalogued ma-

V2

Fig. 48. — Canonical analysis of measurements 3-10 of upper
teeth of specimens of Blarina from the Back Creek Cave No. 2
l.f. Centroids enclose variation in samples of Holocene taxa of
Blarina (see Fig. 3).

terial). A canonical plot of the five maxillaries that
were examined is illustrated in Fig. 48; univariate
statistics (BACKCR) are shown in Table 2. Three
specimens are within the range of variation shown
for modern B. b. brevicauda, whereas two specimens
fall within the range of variation for the talpoides
semispecies of B. brevicauda. Of the ten dentaries
examined, five were plotted (Fig. 49) within or near
the centroid encompassing the range of variation for
B. b. brevicauda and certainly represent the species
B. brevicauda. One specimen appears much smaller,
and conceivably represents B. carolinensis.

Clark’s Cave

Clark’s Cave is located 12 km southwest of Williamsville
(U.S.G.S. Williamsville quadrangle 15' series, 38°05'10"N,

V2

Fig. 49.— Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Back Creek Cave No. 2
l.f. Centroids enclose variation in samples of Holocene taxa of
Blarina (see Fig. 4).

120

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

79°39'25"W) in Bath County, Virginia. The cave is situated in a
limestone cliff overlooking the Cowpasture River. In the im-
mediate vicinity of the cave (extending approximately 1.5 km
upstream and 1 km downstream) the Cowpasture River valley
is gorgelike and talus at the base of the cliff is extensive. Mesic
forest covers the slope on which Clark’s Cave is located; the
present climate is temperate (Guilday et al.. 1977).

The cave has six major entrances and more than 2,400 m of
passages. Fossils were excavated at the top of loose talus inside
the passageway of entrance No. 2. This deposit is thought to be
a remnant of nesting and roosting debris accumulated from rap-
torial birds. Fossils appeared chemically unaltered but frag-
mented; no articulated remains were found. A detailed descrip-
tion of the cave and analysis of its fauna were published by
Guilday et al. (1977), and serve as the basis for this report.

The age of the deposit is considered late Wisconsinan. The
large number of boreal birds and mammals indicates deposition
during a cooler climatic period. The absence of introduced ani-
mals and the size characteristics of several species (including B.
brevicauda) indicate little or no Holocene accumulation (Guilday
et al., 1977). The exposed nature of the deposit has made dating
techniques unreliable. All floral remains are believed of Holocene
origin. A carbon- 14 date (2,260 ± 85 BP) obtained from bone
fragments is too young.

Remains of crustaceans (crayfish), insects, molluscs, fishes, am-
phibians, reptiles, birds, and mammals were recovered from the
deposit at Clark’s Cave. The two species of crayfish still occur in
the Cowpasture River, and the specimens were probably raptor
food remains. All insect remains were Holocene. Remains of
snails, clams, fishes, and herptiles represent species still present
in the area, and were incidental inclusions or food items. The
herptiles reflect the selection bias of the raptors. Avian remains,
representing 68 species, largely represent raptors and raptor prey.
Most are modern residents or migrants through western Virginia.
Notable exceptions are the spruce grouse ( Canachites canaden-
sis), sharp-tailed grous e (Pediocetes phasianellus), rock ptarmigan
(Lagopus mutus ), and possibly the gray jay (Periosoreus cana-
densis), all of which presently occur in Canada (Guilday et al.,
1977).

The mammalian fauna virtually is identical with that of New
Paris No. 4. Fifty-three species of mammals were identified. One
species (Canis dims) is extinct, whereas all others are extant. Eight
species, all of boreal or western affinities ( Sorex arcticus, Euta-
mias minimus, Spermophilus tridecemlineatus, Synaptomvs bo-
realis, Phenacomys intermedins, Microtus xanthognathus, Maries
americana, and Mustela erminea ) do not occur in the area today.
Eleven species survive at higher elevations and have not been
reported from the Cowpasture River valley. Five species, in-
cluding Blarina brevicauda, are of different size than their modern
counterparts, and are similar to modern boreal species or to
remains from New Paris No. 4 (Guilday et al., 1977). Mammals
accounted for 73.5% of the estimated biomass represented by the
fossils. Voles (Microtinae) were the most numerous in terms of
individuals (45.2%) and body weight (36.6%).

The fauna of the Clark’s Cave deposit suggests that the area
was covered predominantly by open boreal woodland in which
conifers were dominant. No southern birds or mammals were
found. The presence of several species (including P. phasianellus
and 5. tridecemlineatus) implies semi-prairie or parkland con-
ditions. Remains of ptarmigan at Clark’s Cave and Back Creek
Cave No. 2 suggest the presence of tundra conditions in the area.
The topography of the area probably provided ecological diver-
sity, as it does today (Guilday et al., 1977).

V2

Fig. 50. —Canonical analysis of measurements 3-10 of upper
teeth of specimens of Blarina from the Clark’s Cave If. Centroids
enclose variation in samples of Holocene taxa of Blarina (see
Fig. 3).

The fourth most numerous family represented was
the Soricidae. Remains identified as B. brevicauda
were the most abundant, accounting for 42% of the
shrews recovered (MNI = 97). Guilday et al. (1977)
compared measurements of the specimens from
Clark’s Cave with measurements reported for extant
animals from Minnesota and Pennsylvania (Guil-
day et al., 1 964) and from North Carolina and West
Virginia (Ray, 1967). Comparisons also were made
with extant specimens from two localities in Bath
County and with Pleistocene specimens from the
deposits at New Paris No. 4 (Pennsylvania) and
Natural Chimneys (Virginia). The specimens from
Clark’s Cave were comparable to the largest Holo-
cene subspecies (the sample from Minnesota) and
to the larger of the two groups from New Paris
No. 4.

For the present study, 56 specimens of Blarina
from Clark’s Cave were examined (CM 24541 and
24578). Upper tooth row measurements 3-10 were
complete for the nine maxillary specimens exam-
ined (Fig. 50). Five of these specimens plotted with-
in the range of variation of the samples of B. b.
brevicauda from Nebraska and Iowa, the largest Ho-
locene shrews examined for this study. One speci-
men plotted outside but near this ellipse, in the up-
per size limit for intermediate-sized animals, whereas
the three remaining specimens were larger than the
modern B. b. brevicauda sampled for this study.
Thirty-two of the dentaries are plotted in Fig. 51;
1 1 additional observations are hidden. These spec-
imens also clustered in or near the centroid showing

1984

JONES ET AL. — B LARIN A SYSTEM ATICS

121

1.60

V2

Fig. 5 1.— Canonical analysis of measurements 19-24 of lower
teeth of specimens of Biarina from the Clark’s Cave l.f. Centroids
enclose variation in samples of Holocene taxa of Biarina (see

Fig. 4).

the range of variation for the two samples of modem
B. b. brevicauda. The specimens that were not ana-
lyzed because of missing measurements appear no
different from those shown in the figure. All speci-
mens of Biarina from Clark’s Cave represent the
species B. brevicauda.

Jasper Saltpeter Cave

Jasper Saltpeter Cave is located in the southwestern comer of
Virginia in Lee County. This area is part of the Appalachian
Plateau, a region of sharp ridges and deep valleys, with a mean
annual precipitation of 122 cm (NOAA, 1974).

A formal description of Jasper Saltpeter Cave and its fauna
has not been published. The fauna consists of the remains of
small animals recovered from near the surface in the entrance to
the cave. The age of the bone material is unknown, but it is
noteworthy that remains of pika ( Ochotona ) were included. This
genus was thought to have become extinct in the Appalachians
before the Wisconsinan (J. E. Guilday, in litt., 1982). Previous
Appalachian records of the pika were from the Irvingtonian fau-
nas of Cumberland Cave, Maryland, and Rapps and Trout caves.
West Virginia (Kurten and Anderson, 1980).

Five maxillaries and nine dentaries of Biarina
w'ere examined (CM 30244). Results of canonical
analysis of the most complete dentaries are shown
in Fig. 52. The specimens from the cave are shown
to be within the range of variation of the modern
talpoid.es semispecies of B. brevicauda. The fifth
specimen had measurements most similar to those
of the larger specimens at the right of the diagram.
Comparisons were made between measurements of
the fossils (maxillaries and dentaries) and means
obtained for those measurements for the modem
reference samples. The specimens from the Jasper

V2

Fig. 52. — Canonical analysis of measurements 3-10 of upper
teeth of specimens of Biarina from the Jasper Saltpeter Cave l.f.
Centroids enclose variation in samples of Holocene taxa of Bia-
rina (see Fig. 3).

Saltpeter Cave deposit again appear most similar to
modern samples of B. brevicauda ( talpoides semi-
species), and doubtless pertain to that species.

Meadowview Cave

Meadowview Cave is located in Washington County, Virginia.
A description of the cave and its fauna has not been published,
although the presence of Phenacomys in the deposit was reported
by Guilday and Parmalee (1972). Also present are fragmentary
remains of Biarina brevicauda, Neotoma. Clethrionomys gapperi,
Synaptomys sp., Microtus sp., Microtus (Pity my s) sp., Ursus
americanus, Dasypus bellus, cervid sp., and sciurid sp. (J. E.
Guilday, in litt., 1982). The extinct armadillo (D. bellus) has been
reported from Blancan, Irvingtonian, and Wisconsinan localities
in Florida, Georgia, Arkansas, Tennessee, West Virginia, Mis-
souri, Texas, and New Mexico (Kurten and Anderson, 1980).
The genus Phenacomys currently has a Canadian-Hudsonian dis-
tribution (Guilday and Parmalee, 1972). The remaining genera
have been recorded from Virginia in Holocene times (Hall, 1981).
The age of the deposit is late Wisconsinan or early Holocene
(Guilday and Parmalee, 1972).

One dentary (CM 24300) from this deposit was
examined. All molars were present and measure-
ments 18 through 24 and 26 (4.6, 2.3, 1.5, 1.9, 1.3,

l. 5, 1.0, 2.5) were taken. When compared with the
means of modern samples for these measurements
(Table 1), the dentary is most similar to the smaller
subspecies of B. brevicauda {talpoides semispecies)
and doubtless pertains to this species.

Natural Chimneys

The Natural Chimneys are located 1.6 km north of Mt. Solon,
Augusta County, Virginia at 30°22'N and 70°5'W. The site is on
the east bank of the North River valley at an elevation of 414

m. The geology and the fauna of the Natural Chimneys were
described by Guilday (1962 b).

122

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

The Natural Chimneys are dolomite pillars, which are eroded
remnants of the walls of sinkholes that formerly were open on
the hilltop. Brown's Cave and the Cave of the Wooden Steps are
located at the base of the pillars; these are probably remnants of
subterranean connections between the sinkholes (Guilday, 1 962 ft).
The bulk of the bone deposit was recovered from cave fill inside
the mouth of Brown’s Cave, although some specimens also were
found in the side passages of Brown’s Cave and in the Cave of
the Wooden Steps. No stratification was evident. Most of the
bones accumulated from the debris of raptorial birds. Few re-
mains of larger mammals were present, and these probably re-
sulted from denning activities in the cave or were carried in by
woodrats. Except for the larger bones, the remains were well
preserved (Guilday, 1962ft).

The vertebrate fauna (especially the lack of southern mammals
and introduced species) and the nature of the deposit suggest an
early post- Wisconsin age for this cave fauna (Guilday, 1962ft).
The deposit is considered comparable in age to the Bath County
deposits at Back Creek Cave No. 2 and Clark’s Cave (Guilday
et ah, 1977). The fauna also was correlated with that of New
Paris No. 4 of Pennsylvania (Guilday, 1962ft).

The invertebrates of the Natural Chimneys l.f. consist of a
millipede and snails. Fishes are common, but all remains are
small and probably were deposited by predators or by occasional
flooding of the cave. Of the 22 genera of herptiles, only two species
(the coachwhip, Masticophis flagellum, and the diamond-backed
rattlesnake, Crotalus adamanteus) are not found in Virginia to-
day. These also were the only southern animals found in the
fauna, and it is unknown whether they were contemporaneous
with the remainder of the fauna. Forty species of birds, most of
which are found presently in the area as residents or migrants,
were identified from Natural Chimneys. Remains of nine of the
species also were found at Back Creek Cave No. 2 and Clark’s
Cave, including the spruce grouse (Canachites canadensis) and
sharp-tailed grouse (Pedioecetes phasianellus), which currently
have northern boreal distributions (Guilday et al., 1977). The
gray jay ( Perisoreus canadensis ), recovered from Clark’s Cave
and Natural Chimneys, also is an indicator of boreal conditions.
The record of the magpie ( Pica pica) at Natural Chimneys is the
first for the eastern United States. Its presence implies semi-
prairie or parkland conditions and may be an indicator of in-
creased continentality (Guilday, 1962ft, 1 971; Guilday etal., 1977).

Remains of at least 878 mammals were identified, representing
55 species of insectivores, bats, rodents, lagomorphs, and artio-
dactyls. Four species are extinct. The other species are present
in the area today or retreated to higher elevations or latitudes
during post-glacial warming (Guilday, 1962ft). Several species
with modem distributions in the Canadian zone, such as the
arctic shrew (Sorex arcticus) and the spruce vole ( Phenacomys ),
indicate deposition during a period of glacial advance (Guilday,
1962ft). Coexistence of the presently allopatric voles Microtus
xanthognathus and M. chrotorrhinus implies taiga-like condi-
tions (Guilday, 1971). In short, the avian and mammalian re-
mains of Natural Chimneys suggest that coniferous forest covered
the area during deposition (ca. 10,000-15,000 years ago), and
that streams, swampy grasslands, and parkland were in the vi-
cinity (Guilday, 1962ft, 1971).

Soricids are represented in the deposit by seven
species and a minimum of 133 individuals. A min-
imum of 56 individuals was referred to the species
B. brevicauda (Guilday, 1962ft; Guilday et al., 1969).

V2

Fig. 53. — Canonical analysis of measurements 3-10 of upper
teeth of specimens of Blarina from the Natural Chimneys l.f.
Centroids enclose variation in samples of Holocene taxa of Bla-
rina (see Fig. 3).

Guilday ( 1 962 ft) noted that the average size of spec-
imens of Blarina from Natural Chimneys is larger
than that of modern specimens from Pennsylvania,
and that “some specimens are distinctly larger and
more rugged than any modern B. brevicauda kirt-
landi, and compare favorably with the large late
Pleistocene B. brevicauda from the New Paris No.
4 local fauna, Pennsylvania.” Graham (1976) as-
serted that both brevicauda and kirtlandi were pres-
ent in the cave fauna, and that vole and shrew species
density was higher during the Wisconsinan than in
the Holocene. Both of these interpretations were
considered evidence of more equable climate, with
reduced temperature and moisture gradients, during
the period of deposition.

Thirty of the most complete specimens (CM 7544-
7545) were selected for examination in this study.
Nine maxillary fragments were measured, of which
eight were sufficiently complete for canonical anal-
ysis. These specimens, shown in Fig. 53, were plot-
ted in or near the centroid showing the range of
variation of the two modern samples of B. ft. brev-
icauda from Nebraska and Iowa. The ninth speci-
men had measurements comparable to those of the
specimens shown. Of the dentaries examined, eigh-
teen are shown in Fig. 54. These specimens also are
within or near the centroid for modern B. ft. brev-
icauda. The three remaining dentaries appear no
different than those shown. Taking into account the
differences in sample sizes, the two diagrams are
comparable to those for Clark’s Cave (Figs. 50 and
51). Based on the analysis of these specimens, the
Blarina of the Natural Chimneys fauna appear to

1984

JONES ET AL .-BLARINA SYSTEM ATICS

123

V2

Fig. 54. — Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Natural Chimneys l.f.
Centroids enclose variation in samples of Holocene taxa of Bla-
rina (see Fig. 4).

be a relatively homogeneous population of B. brev-
icauda comparable to the extant subspecies B. b.
brevicauda.

Ripplemead Quarry

Bones were recovered from three fissures in the wall of an
abandoned limestone quarry near Ripplemead, Giles County,
Virginia. The quarry is on a bluff overlooking the New River.
Shells of terrestrial snails and remains of one fish, one bird, 5
species of herptiles, and 26 species of mammals were recovered
(Weems, 1972; Weems and Higgins, 1977). Accumulation prob-
ably occurred as bones were washed into the fissures from sur-
rounding slopes (Weems and Higgins, 1 977). The most numerous
remains are those of snakes and rodents. Larger mammals present
include horse, peccary, deer, and bear (Weems and Higgins, 1977;
J. E. Guilday, in litt., 1982). All of the small mammals still live
in the area (J. E. Guilday, in litt., 1982).

Six specimens (CM, uncatalogued material), con-
sisting of fragments of one maxillary and five den-
taries, were examined. The sample appeared ho-
mogeneous. Measurements of these specimens were
compared with those of Holocene specimens (Table
1). Measurements 3 through 14 of the maxillary (2.7,
2.1, 1.7, 2.5, 2.3, 2.2, 2.6, 2.5, 1.8, 1.8, 2.5, 2.1) are
most similar to the means derived for measure-
ments 3 through 14 of the sample of B. b. brevicauda
from Iowa (sample 5). Measurements 19 through
22 (2.5, 1.7, 1.9, 1.3; 2.4, 1.6, 2.0, 1.3) of two den-
taries and measurements 18 through 24 (5.0, 2.3,
1.5, 1.9, 1.3, 1.6, 1.1; 4.9, 2.4, 1.7, 2.0, 1.4, 1.6, 1.1;
4.8, 2.5, 1.7, 2.0, 1.3, 1.5, 1.0) of three dentaries
also were taken. The dentaries compare favorably
with modern samples of the subspecies brevicauda ,
manitobensis, and churchi (samples 5, 12, 18, and

11) and are referred to the species B. brevicauda.
The specimens from Ripplemead Quarry appear
somewhat larger than those from Jasper Saltpeter
Cave (Lee County). They are similar to some of the
specimens from Back Creek Cave No. 2 and Clark’s
Cave (Bath County), Strait Canyon (Highland
County), and Meadowview Cave (Washington
County).

Remains from Ripplemead Quarry have not been
dated. The presence of the tapir, horse, and peccary,
and the absence of domestic animals and of large
mammals typical of the Wisconsinan imply depo-
sition approximately 7,000 to 9,000 BP (Weems and
Higgins, 1977). Possibly some mixing with more
recent material occurred; if the remains are contem-
poraneous, the presence of hardwood forest and cli-
mate similar to that of today is indicated (Weems
and Higgins, 1977).

Strait Canyon Fissure

Strait Canyon Fissure is located in Highland County, within
the Ridge and Valley physiographic province of western Virginia.
A description of the fissure and its fauna is in preparation (J. E.
Guilday, in litt., 1982). Specimens are housed at the Carnegie
Museum of Natural History. The deposit includes remains of
fishes, amphibians, reptiles, birds, and mammals. Remains are
fragmentary, and no stratification is evident. The mammalian
fauna includes four taxa now extinct: Mammut. Tapirus, My-
lohyus, Equus, and Neofiber cf. N. leonardi. The dominant mi-
crotine is Microtus pennsylvanicus (J. E. Guilday, in litt., 1982).
Ray (1965) reported the presence of Tamias. The age of the
deposit is considered early Wisconsinan, and remains of boreal
species (Phenacomys and Synaptomys borealis) are rare (J. E.
Guilday, in litt., 1982).

Thirty-six specimens (four maxillaries and 32
dentaries) collected from talus and from eight strati-
graphic levels (arbitrarily designated) were exam-
ined. Measurements were compared with the means
derived from modern samples. Although the mea-
surements of the fossils were variable, the specimens
from the cave are most similar to the modern sam-
ples of B. brevicauda. Thirteen of the dentaries were
sufficiently complete for analysis; results are shown
in Fig. 55 (one observation was hidden). The spec-
imens shown were recovered from talus (T) and
from levels two feet below the talus (b), nine feet
above the talus (i), eleven feet above the talus (k),
and sixteen feet above the talus (p). All but one of
the specimens are plotted in the centroid that shows
the range of variation for the talpoides semispecies
of B. brevicauda. Clearcut differences among spec-
imens from the different levels are not evident. All
specimens of Blarina from the Strait Canyon Fissure
l.f. tentatively are referred to B. brevicauda.

124

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

V2

Fig. 55. — Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Strait Canyon I f. Cen-
troids enclose variation in samples of Holocene taxa of Blarina
(see Fig. 4).

Eagle Cave

Eagle (Eagle Rock) Cave is one of four caves in Pendleton
County, West Virginia, from which remains of Blarina have been
reported (Appendix 1 ). Pendleton County is included in the Ridge
and Valley physiographic province. The original vegetation of
the area was a chestnut-oak forest, which has been replaced by
a mixed hardwood forest in which oak predominates. Spruce is
found at elevations above 765 m. Grass-sedge meadows occur
occasionally in mountain valleys, and may have persisted since
the late Pleistocene (Guilday and Hamilton, 1973, 1978). Guilday
and Hamilton (1978) reported the average precipitation as 76 to
101 cm annually.

Eagle Cave is located 2 1 .6 km north of Franklin, West Virginia,
at 38°50'N, 79°17'W, 762 m (Onego Quad., U.S.G.S. 15' series)
(Guilday and Hamilton, 1 978). The deposit is in a limestone cave
located in Cave Mountain, overlooking the South Branch of the
Potomac River. The floor of the cave has been greatly disturbed
by human traffic, and no stratification was observed. Fossils were
fragmentary. A description of Eagle Cave and its fauna was pub-
lished by Guilday and Hamilton (1973). Additional notes re-
garding the fauna were published by Guilday (1971), Guilday
and Parmalee(1972), Graham (1976), and Guilday and Hamilton
(1978).

The fauna of Eagle Cave consists of the remains of fishes, frogs,
snakes, and mammals. Most of the remains accumulated in owl
pellets although a few species, such as bats, probably were resi-
dents of the cave. Boreal, continental, and temperate species were
represented. Species that now have northern distributions are
Microtus xanthognalhus, Synaptomys borealis, Sorex arcticus,
and Phenacomys intermedins. Spermophilus tndecemlineatus.
which currently occupies the prairies of central North America,
also was present. A fossil assemblage with these mixed compo-
nents is thought to be typical of mid-Appalachian late Pleistocene
faunas (Guilday and Hamilton, 1978).

Four species of shrews were identified from the
cave deposit. Remains of Blarina (MNI = 7) were
assigned to the species B. brevicauda (Guilday and
Hamilton, 1973). Graham (1976) also assigned the

fossils to B. brevicauda ; he considered members of
the modern fauna to be B. kirtlandi.

Seven dentaries (CM 24397-24399), most with
incomplete dentition, were examined. Measure-
ments 19 through 26 of one dentary (2.5, 1.7, 2.0,

1.4, 1.6, 1.1, 4.4, 2.7), 19 through 22 of three den-
taries (2.4, 1.6, 2.1, 1.4; 2.4, 1.6, 2.0, 1.3; 2.6, 1.8,
2.0, 1.4), 19 through 24 and 26 of one dentary (2.3,

1.5, 1.9, 1.3, 1.6, 1.1, 2.8), and 21 through 24 of
one dentary (2.1, 1.4, 1.6, 1.1) were taken. Mea-
surements 19 and 20 of one molar (2.5, 1.7) also
were taken. Measurements were compared with
means for those measurements obtained from ex-
tant groups (Table 1). The fossils from Eagle Cave
seem referrable to either B. b. brevicauda or B. b.
churchi although there is more variation among the
measurements for the fossils than among their mod-
ern counterparts. Generally, the fossils were larger
than the two modern samples of B. b. kirtlandi (sam-
ples 4 and 6). The fossils obviously represent the
species B. brevicauda.

Hoffman School Cave

Hoffman School Cave is located 8 km south of Franklin, Pen-
dleton County, West Virginia, at 38°34'38"N, 79°21'49"W (Cir-
cleville Quad., U.S.G.S. 15' series) at an elevation of 660 m
(Guilday and Hamilton, 1978). This limestone cave is located
on Thom Creek, a tributary of the South Branch of the Potomac
River. The cave is approximately 30 km south of Eagle Cave.

The bone deposit in Hoffman School Cave, like that in Eagle
Cave, consists mainly of remains of small vertebrates that ac-
cumulated as a result of roosting birds. Cave rats (Neotoma flor-
idana) also inhabited the cave, and probably were responsible
for accumulating the larger mammalian teeth found in the de-
posit. Three components that presently are allopatric— boreal,
continental, and eastern temperate species— are represented in
the deposit (Guilday and Hamilton, 1978). Species no longer
present in the area include Spermophilus tndecemlineatus. Glau-
comys sabrinus, Phenacomys intermedius, Microtus chrotorrhin-
us, Erethizon dorsatum, and Maries americana. The absence of
M. xanthognalhus. common in other late-Wisconsinan sites in
the Appalachians, might indicate that the deposit was formed in
the period of climatic change after the Wisconsinan glaciation,
during which time other boreal forms persisted in the area (Guil-
day and Hamilton, 1978). Otherwise, the assemblage is similar
to that of Eagle Cave in that both are “typical ‘late-boreal’ wood-
land faunas characteristic of the central Appalachians” (Guilday
and Hamilton, 1978).

The soricids recovered from this deposit were
identified as Sorex sp. (MNI = 2) and Blarina brev-
icauda (MNI = 8) by Guilday and Hamilton (1978).
Seven specimens of Blarina (CM 30020) were ex-
amined for this study. Measurements 19 and 20 of
four mis (2.5, 1.8; 2.4, 1.6; 2.4, 1.7; 2.2, 1.5), 21
and 22 of three m2s (2.0, 1.3; 2.1, 1.4; 2.0, 1.3), and

125

JONES ET AL. — B LARINA SYSTEMATICS

1984

Fig. 56.— Canonical analysis of measurements 3-10 of upper
teeth of specimens of Blarina from the Moscow Fissure l.f. Cen-
troids enclose variation in samples of Holocene taxa of Blarina
(see Fig. 3).

23 and 24 of two m3s (1.6, 1.1; 1.7, 1.1) were taken.
All were dentaries with incomplete dentition; con-
sequently, the measurements were not useful for
canonical analysis. Comparisons were made be-
tween measurements of these fossils and means ob-
tained from Holocene samples for those measure-
ments. The fossils were most similar to the sample
of B. b. brevicauda from Iowa and Nebraska (Table
1, samples 5 and 12). One fossil specimen appeared
slightly smaller than the others but obviously per-
tains to the species B. brevicauda. Generally the
fossils of Hoffman School Cave are similar in size
to those of Eagle Cave, which perhaps is to be ex-
pected given the overall similarity of the two faunas.

Moscow Fissure

A Rancholahrean fauna was recovered in the driftless area of
southwestern Wisconsin. The Moscow Fissure l.f. was recovered
approximately 3 km northeast of Blanchardville, Iowa County,
in a deposit along the Blue Mounds Branch (SW ‘A, NE ‘A, SW
‘A sec. 12, T4N, R5E). The fauna was described in a thesis by
Robert L. Foley of the University of Iowa (Semken, personal
communication, 1982).

A radiocarbon date of 17,050 ± 1,500 BP places the Moscow
Fissure l.f. near the late Wisconsin glacial maximum. The fauna
is disharmonius but appears predominantly boreal. Nine species,
including Thomomys talpoides, Synaptomys borealis, Phenaco-
mys intermedins, Dicrostonyx torquatus, Mierotus xanthogna-
thus, and Zapus princeps, no longer occur in the area. The fauna
indicates the presence of marsh, steppe, and spruce parkland.
The l.f. of Moscow Fissure compares most favorably with that
of New Paris No. 4 (Semken, personal communication, 1982).
All specimens, as yet uncatalogued, are housed at the Department
of Geology, University of Iowa (SUI).

Fig. 57. — Canonical analysis of measurements 3-10 of upper
teeth of specimens of Blarina from the Caverne De Saint-Elzear-
de-Bonaventure l.f. Centroids enclose variation in samples of
Holocene taxa of Blarina (see Fig. 3).

Southwestern Wisconsin was included in the Illinoian biotic
province (Dice, 1943). Mean annual precipitation is 76 to 81 cm
(NOAA, 1974).

Foley considered the Blarina of this fauna equiv-
alent to B. b. brevicauda and B. b. manitobensis
[= the brevicauda semispecies of B. brevicauda ]
(Semken, personal communication, 1982). Sixteen
of the most complete specimens (eight maxillaries
and eight dentaries) were examined for the present
study. Results of the canonical analysis of seven
maxillaries are shown in Fig. 56. The observations
are plotted within and near the centroid showing
the range of variation of Holocene B. b. brevicauda
from Iowa and Nebraska. Measurements 19 through
22 of eight m 1 s and m2s (means = 2.46, 1.69, 2.06,
1.40) also are most similar to those of Holocene B.
b. brevicauda. The fossils from Moscow Fissure un-
doubtedly pertain to that subspecies.

Caverne de St. Elzear

Remains of Blarina were recovered from Caverne de Saint-
Elzear-de-Bonaventure, Bonaventure County, Quebec. Bonaven-
ture County is located on the southern coast of the Gaspe Pen-
insula. The area was included in the Canadian biotic province
by Dice (1943).

The deposit of this cave is of Holocene origin. A date of
5,110 ± 150 BP was reported by LaSalle and Guilday (1980).
Microtines from the deposit include Mierotus chrotorrhinus. M.
Pennsylvania, M. xantkognathus, Dicrostonyx, Phenacomys, and
Synaptomys borealis (Zakrzewski, personal communication).
Phenacomys and Dicrostonyx do not occur in the area today
(Hall, 1981).

One hundred and thirty-five specimens (CM
37906-37910 and 37912-37919) of Blarina from

126

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

1.60

V2

Fig. 58. — Canonical analysis of measurements 19-24 of lower
teeth of specimens of Blarina from the Caverne De Saint-Elzear-
de-Bonaventure l.f. Centroids enclose variation in samples of
Holocene taxa of Blarina (see Fig. 4).

PHYLOGENY AND I

We attempted to depict the paleobiogeography of
Blarina in a series of maps (Figs. 59-64). Infor-
mation was gleaned from studies of fossil pollen, in
addition to the paleontological studies previously
discussed. Because of the difficulties in the inter-
pretation and correlation of past climates and pa-
leofaunas, the maps were designed only to show
general trends in the evolution of Blarina ; the faunas
depicted in each figure are not intended to indicate
strict synchrony.

The fossil record of Blarina implies evolution of
the two species brevicauda and carolinensis in the
early Pleistocene. The earliest specimens of the ge-
nus were recovered from late Blancan (early Pleis-
tocene) faunas in Kansas. These specimens, from
Dixon and White Rock, are the size of the modern
talpoides semispecies of B. brevicauda. Fossils of
both faunas indicate cool, moist climatic conditions.
The presence of Blarina on the Great Plains in the
early Irvingtonian prior to the period of maximum
Nebraskan glaciation is implied. The extent of the
range of the genus at this time is unknown because
few Nebraskan faunas have been found.

We examined fossils from 1 7 Irvingtonian sites
(from the central plains, Florida, Arkansas, West
Virginia, Maryland, and Pennsylvania) which com-
prise the Irvingtonian record of Blarina as we know
it. The relative ages of these faunas is uncertain
(Appendix 1), due to the poor preservation of some
of the fossils, the relative scarcity of Irvingtonian
faunas, and the heterochronic nature of some of the

St. Elzear were examined. Univariate statistics are
shown in Table 2. The specimens were recovered
from 1 3 stratigraphic levels, and were separated by
level for canonical analysis. Thirty-two maxillaries
plotted within the centroids for B. b. brevicauda and
the smaller subspecies of brevicauda ; one observa-
tion was hidden (Fig. 57). Seventy-two mandibular
specimens were examined and are illustrated in Fig.
58 (13 observations are hidden). Again, the speci-
mens were plotted within the two centroids repre-
senting subspecies of B. brevicauda. No differences
were observed among specimens from different
levels. Measurements of the St. Elzear material are
most similar to those of the subspecies talpoides,
angusta, and pallida (samples 13, 19, 24, 20, and
9).

deposits, all of which make comparisons among fau-
nas difficult.

The earliest Irvingtonian fossils are thought to be
those of the Inglis IA fauna (Nebraskan — Aftonian)
of western Florida; these are the first known spec-
imens of B. carolinensis. The fauna indicates coastal
savanna habitat. Thus far, remains of Blarina have
not been discovered in any other Aftonian faunas
(unless the Port Kennedy l.f. of Pennsylvania is found
to date from the Aftonian, as tentatively suggested
by Kurten and Anderson, 1980). Semiarid condi-
tions on the Great Plains during the Aftonian (Hib-
bard, 1960, 1970; Eshelman and Hibbard, 1981)
probably drove Blarina into relatively moister areas
to the south- and northeast (see Dorf, 1959, for maps
of generalized climatic zones of a composite of in-
terglacial and glacial stages).

The genus Blarina reappears in the fossil record
during the Kansan glacial stage in faunas in South
Dakota, Kansas, Texas, West Virginia, and Arkan-
sas (Fig. 59); two of these sites are outside the present
range of the genus. Three additional specimens of
what appears to be B. carolinensis were recovered
from Wathena (Kansas) and Java (South Dakota).
Identification of these fossils is tentative due to small
sample size. Comparative material of similar pro-
venience is lacking and both faunas have undergone
preliminary study only. If the fossils from Wathena
and Java prove to be B. carolinensis of Aftonian-
Kansan age, an early, widespread distribution (pos-
sibly associated with grassland) of the species would

1984

JONES ET AL. — B LARINA SYSTEMATIC'S

127

Fig. 59. — Possible distribution of B. brevicauda (closed circles) and B. carolinensis (open circles) in the middle Irvingtonian, as suggested
by Aftonian-Kansan localities; 1. Java; 5. Wathena; 10. Cudahy; 19. Vera; 45. Conard Fissure; 62. Trout Cave. Specimens of B. b.
brevicauda were found at 62.

be argued. Fossils from other sites of this general
period of time (Cudahy, Vera, and Conard) pertain
to the species B. brevicauda. Specimens from Trout
Cave might represent two phena, the brevicauda and
talpoides semispecies of B. brevicauda. The climate
on the plains at this time continued to be cool but
with more effective moisture (Hibbard and Dalquest,

1966; Semken, 1966; Hibbard, 1970; Martin,
\913b). The Conard Fissure I.f. (Arkansas) also sug-
gests a climate with cooler and moister summers.
Climate at Trout Cave probably was similar to that
of today, and the smaller fossils from that cave are
similar in size to B. b. kirilandi. the subspecies pres-
ently in the northeast.

128

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 60. — Possible distribution of B. brevicauda (closed circles) and B. carolinensis (open circles) in the late Irvingtonian, as suggested
by Yarmouthian and early Illinoian localities: 2. Bartek Brothers Quarry; 3. Angus; 6. Rezabek; 7. Kanopolis; 15. Berends; 56. Hanover
Quarry Fissure; 58. Port Kennedy; 59. Cumberland Cave; 77. Coleman IIA. Specimens of B. b. brevicauda were found at 2, 6, 7, and

15.

Fossils of Blarina are known from five faunas of
the Yarmouth interglacial. Two of these (Port Ken-
nedy and Bartek Brothers) might be older than other
Yarmouthian faunas. The Angus (Nebraska) and
Kanopolis (Kansas) local faunas suggest climatic
conditions similar to or moister and warmer than
those of today (Schultz and Martin, 1970; Hibbard

et al., 1978). The difference in size of specimens
from Angus and Bartek Brothers Quarry suggests
there was a zone of contact between two subspecies
in southeastern Nebraska or the periods of depo-
sition were not the same. Based on current knowl-
edge of intergradation between Holocene B. b. brev-
icauda and B. b. kirtlandi in Iowa and Missouri

1984

JONES ET AL. — BLARINA SYSTEM ATICS

129

105 95 85 75 65

Fig. 6 1 . —Possible distribution of B. brevicauda (dosed circles) and B. carolinensis (open circles) in the early Rancholabrean, as suggested
by late Illinoian and Sangamonian localities: 8. Williams; 12. Mt. Scott; 16. Doby Springs; 18. Easley Ranch; 71. Haile XIB; 72.
Arredondo; 73. Reddick IA; 74. Williston III A; 78. Bradenton 5 1st Street. Dotted line indicates extent of Illinoian glaciation. Specimens
of B. b. brevicauda were found at 16.

(Moncrief et al., 1982), the latter alternative seems
more likely. The age of the Sappa formation, from
which the Bartek Brothers l.f. was recovered, sug-
gests that fauna is older (possibly pre-Kansan) than
the Angus fauna. A trend from cool to warm climate
is implied, perhaps accompanied by decreasing
moisure; this might account for the subsequent rein-

vasion of the smaller subspecies observed at Angus.
The size of specimens from Port Kennedy (if, in-
deed, that fauna is Yarmouthian) indicates climatic
conditions during the period of deposition were sim-
ilar to those of today, and B. brevicauda extended
across the mid western and eastern United States
(Fig. 60). Temperate conditions in the east are sug-

130

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 62. — Possible distribution of B. brevicauda (closed circles), B. carolinensis (open circles), and B. hylophaga (closed triangles) as
suggested by early Wisconsinan localities thought to represent full glacial and late glacial (until approximately 10,500 BP) faunas: 1 1.
Jinglebob; 13. Robert; 17. Howard Ranch; 25. Schulze Cave; 27. Cave Without a Name; 33. Craigmile; 36. Waubonsie; 40. Boney
Spring; 41. Trolinger Spring; 43. Bat Cave; 44. Zoo Cave; 46. Peccary Cave; 47. Moscow Fissure; 50. Welsh Cave; 52. Baker Bluff; 53.
Guy Wilson Cave; 55. New Paris Sinkhole No. 4; 57. Bootlegger Sink; 63. Strait Canyon Fissure; 65. Back Creek Cave No. 2; 66.
Clark’s Cave; 70. Ladds Quarry; 79. Vero. Specimens of B. b. brevicauda were found at 13, 33, 47, 50, 52, 55, 57, 65, and 66. Dotted
line indicates maximum extent of Wisconsinan glaciation.

gested by the fauna of Hanover Quarry Fissure ( B .
carolinensis). The presence of B. carolinensis at
Hanover and Cumberland might indicate a zone of
sympatry of the two species, probably prior to the

maximum Illinoian glaciation. Cumberland Cave is
the only Irvingtonian site at which both species are
clearly present. The scarcity of Blarina from the
Yarmouthian probably reflects the sparse fossil rec-

1984

JONES ET AL .-BLARINA SYSTEM ATICS

131

Fig. 63. — Possible distribution of B. brevicauda (closed circles), B. carolinensis (open circles), and B. hvlophaga (closed triangles) during
the pre-boreal and boreal periods (approximately 10,500 to 5,000 BP) as suggested by late Wisconsinan-early Holocene localities: 20.
Miller’s Cave; 22. Felton Cave; 24. Klein Cave; 29. Cherokee; 32. Mud Creek; 38. Brynjulfson Cave No. 1; 42. Crankshaft Cave; 51.
Robinson Cave; 54. Riverside Cave; 60. Eagle Cave; 61. HofTman School Cave; 64. Natural Chimneys; 67. Ripplemead Quarry; 70.
Ladds Quarry; 80. Caveme de Saint-Elzear. Specimens of B. b. brevicauda were found at 29, 42, 51, 61, and 64.

ord on the plains and elsewhere from that intergla-
cial (Hibbard, 1970) and might not be a true indi-
cation of the range of Blarina at that time.

Early Illinoian (late Irvingtonian) records of Bla-
rina consist of the Rezabek (Kansas), Berends (Okla-

homa), and Coleman IIA (Florida) local faunas. The
Rezabek fauna, and others from the Great Plains,
suggest a climate moister and more equable than
now present (Hibbard, 1970; Zakrzewski, 1975a).
The specimen of Blarina that we examined was sim-

132

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 64. — Present distribution of B. brevicauda (B), B. carolinensis (C), and B. hylophaga (H); the restricted distributions of the taxa B.
h. plumbea (P) and “ B . c. shermani ” (S) also are shown (George et al., 1981, 1982; Moncrief et al., 1982). Shaded areas represent
sympatry of species. Localities are indicated as in Fig. 63. Early Flolocene (<5,000 BP) localities of Blarina are included: 28. Barton
Springs; 30. Willard Cave; 31. Schmitt Cave; 34. Garrett Farm; 35. Pleasant Ridge; 37. Thurman; 39. Brynjulfson Cave No. 2; 48.
Meyer Cave. Specimens of B. b. brevicauda were found at 35 and 37.

ilar in size to B. b. brevicauda. The material from
Berends also indicated moister climate and included
specimens that might represent both semispecies of
B. brevicauda. B. carolinensis was present in the
Coleman IIA fauna. The Cumberland Cave (Mary-

land) l.f. also might pertain to the early Illinoian
(Fig. 60 and Appendix 1).

The earliest Rancholabrean records are probably
those from Doby Springs (Oklahoma), Mt. Scott,
and Williams (Kansas). On the plains a transition

1984

JONES ET \h.- BLARINA SYSTEMATICS

133

from moist, equable climate to warmer conditions
by the late Illinoian was accompanied by shifts in
the flora and fauna (Hibbard, 1970; Kapp, 1970).
Whether the B. carolinensis of Mt. Scott (Fig. 6 1 )
reflects this shift, as we suspect, or marks an area
of sympatry between the two species during this
period cannot be ascertained until additional spec-
imens are found. Sangamonian records of Blarina
are limited to several sites in Florida and one (Easley
Ranch) in Texas. Webb (1974) and Delcourt and
Delcourt (1977) cited evidence of warm temperate
conditions in the southeast during at least part of
the Sangamonian. Dalquest ( 1 962), Hibbard (1970),
and Kapp (1970) suggested that the climate on the
central plains during the Sangamonian was warmer
than during the Illinoian but still moist. However,
Graham (1972) suggested, based on the Cragin
Quarry l.f. (Kansas), that the Sangamonian climate
on the plains was semiarid — sufficiently dry to re-
strict Blarina to the east. The scarcity of Blarina
remains from this period seems to support Gra-
ham’s thesis (Fig. 61).

The Wisconsinan fossil record of Blarina is more
extensive than that of earlier periods. Fossil sites
are scattered throughout the southeastern quarter of
the United States, and Wisconsinan biogeography
of the three species of Blarina is complex (Figs. 62
and 63). The history of changes in vegetation during
this period also is complicated, and the composition
of plant communities was not analogous to that of
extant communities (Davis, 1976; Wright, 1981).
An overview of the ecology of the Pleistocene was
provided by Wright (1976).

During the full glacial period (until approximately
1 3,000 BP), climate in the central and eastern United
States was characterized by cooler summers, milder
winters, and more effective moisture than at present
(Slaughter, 1967; Hoffmann and Jones, 1970; Guil-
day, 1971; Wright, 1981). Floral and faunal analyses
suggest that vegetation during this period consisted
of boreal spruce forest, perhaps with patches of de-
ciduous trees (Mehringer et al., 1970; Ross, 1970;
Watts, 1970; Wells, 1970; Wright, 1971). Spruce
forests extended from southern Canada to Kansas
and Missouri, possibly to the western edge of the
modern plains. In the east, forests of spruce and jack
pine might have reached the Atlantic coast and ex-
tended southward as far as the Coastal Plain of
Georgia (Wright, 1968, 1970; Watts, 1970, 1980;
Davis, 1 976). The extent of tundra and isolated gla-
ciation south of the spruce forest is debatable (Guil-
day et al., 1977; Wright, 1981). Prairie and decid-

uous forest probably were limited to the southwest
and southeast (Hoffmann and Jones, 1970; Ross,
1970; Wright, 1981). In the late glacial period (ap-
proximately 13,000 to 10,500 BP) the formerly
widespread boreal forest deteriorated, although it
might have remained in much of the Great Plains
(Hoffmann and Jones, 1970; Wright, 1968, 1970,
1971).

During the moister conditions of the full glacial
period, the geographic range of the genus Blarina
was extended again. Remains from this period have
been recovered from throughout the central and
eastern United States, including areas in Texas and
southwestern Kansas in which Blarina are not now
present (Fig. 62). The faunas and radiocarbon dates
(when available) of many of the Wisconsinan de-
posits from which Blarina were recovered (those of
Peccary, Back Creek, Clark’s, Baker Bluff, Bootleg-
ger, and New Paris caves, Howard Ranch, part of
Ladds, and possibly Bat, Zoo, and Welsh caves)
imply deposition during the full glacial and late gla-
cial periods (for example, Dalquest, 1965; Guilday
etal., 1977, 1978). The shrews of most of the glacial
faunas were identified as B. brevicauda (B. b. brev-
icauda from northern and Appalachian sites), in-
dicating a widespread distribution for that species
during Wisconsinan boreal conditions. Fossils from
Jinglebob fall into an area of overlap between B.
brevicauda, B. carolinensis, and B. hylophaga. Oth-
erwise, B. carolinensis was restricted to the south
during this period. The presence of B. carolinensis
in these faunas evidently reflects the presence of
warmer prairie or forests south of the boreal forest.
Remains from Schulze Cave (late glacial) apparently
represent the species B. hylophaga. Two phena of
Blarina were detected at Baker Bluff, New Paris,
and Back Creek. Sympatry of more than one phenon
of Blarina and of other presently allopatric species
is interpreted as reflecting the more equable climate
of full glacial and late glacial periods (Dalquest, 1 965;
Graham, 1972, 1976, 1979; Graham and Semken,
1976; Klippel and Parmalee, 1982).

During the pre-boreal and boreal periods of the
late Pleistocene and early Holocene (approximately
10,500 to 8,450 BP), climate became increasingly
continental. Spruce forests were replaced by pine or
oak forests in the north, northeast, and in the Ozarks,
and grassland spread northward from the southwest
(Hoffmann and Jones, 1 970; Mehringer et al., 1970;
Watts, 1980; Webb, 1981; Wright, 1970). Although
post-glacial climates have fluctuated, the overall
trend for the last 7,000 years continues to be one of

BLANC AN I I RVI NGTONI AN I RANCHOL ABRE AN

134

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

jL cen i nsulae

T .

J3 . hylophaaa

B.. brev i cauda
(brevicauda sem i spec i es )

B . carol i nens i s

_B . bre v i cauda
(talpo i des semispecies)

Fig. 65. — Dendrogram showing phylogeny of Blarina as indicated by analyses of fossils.

1984

JONES ET AL .-BLARINA SYSTEMATICS

135

increased continentality (Hoffmann and Jones, 1970;
Webb, 1981).

The fossil record of Blarina suggests that the
species became segregated as a result of increased
continentality (Graham and Semken, 1976), as can
be seen in the faunas of the late Wisconsinan and
early Holocene (Fig. 63). Specimens of B. brevicau-
da from these faunas became increasingly indistin-
guishable from modern specimens. The large B. b.
brevicauda was still present in faunas in the north
and northeast (Cherokee, Robinson, Riverside,
Hoffman School Cave, and Natural Chimneys). Cli-
matic fluctuations since the glacial and late glacial
periods have resulted in fluctuating areas of para-
patry or sympatry, such as the post- Wisconsinan
sympatry of three phena at Cheek Bend Cave (Klip-
pel and Parmalee, 1982) and Crankshaft Cave, and
the intergradation and overlapping observed in Ho-
locene distribution (French, 1981; Moncrief et ah,
1982; and others). Blarina hylophaga appeared in
the fossil record in Missouri and Texas, and was
sympatric with B. carolinensis in some of the Texas
faunas. Its relatively recent separation from caro-
linensis is suggested by the past geographic distri-
butions and present similarities of these two species
(George et ah, 1981). As aridity on the central plains
increased, the range of Blarina retracted eastward
across Texas (Lundelius, 1967; Hall, 1982) and
Kansas (Fig. 64). Holocene B. h. plumbea is con-
sidered a remnant of this formerly widespread pop-
ulation.

Based on the results of this study, we propose that
B. brevicauda, or an ancestral species similar to B.
brevicauda, arose from the blarinine stem in the
middle or late Pliocene (Fig. 65). The earliest re-
mains of B. brevicauda (late Blancan) are the size
of the talpoides semispecies. This semispecies seems
to have had a long association with cool, moist cli-
mate. The large semispecies brevicauda might have
appeared as early as the Kansan glacial (at Trout
Cave). It apparently has always occurred near the
northern boundary of the distribution of the genus.
Its distribution is related presumably to its affinity
to the subhumid microthermal regime (as suggested
by Graham, 1972).

Blarina brevicauda likely became divided into two
populations during the Irvingtonian. Differentiation
and chromosomal rearrangement during isolation
resulted in the smaller southern species, B. caroli-
nensis. The early distribution of B. carolinensis is
unclear. It became at least narrowly sympatric with
B. brevicauda during parts of subsequent glacial and
interglacial periods. B. carolinensis became divided
into eastern and western populations shortly after
the Wisconsinan glaciation, and subsequent specia-
tion (again involving both morphological and chro-
mosomal changes) resulted in the species B. hylo-
phaga in the southwestern part of the range of the
genus. All three species possibly were sympatric at
various times in the past.

SPECIMENS EXAMINED

Massachusetts. Dukes Co.: Martha’s Vineyard, 75 (69
UCONN, 6 USNM); West Tisbury, 7 (MCZ).

Blarina brevicauda angusta

New Brunswick. Bald Peak, 7 (NMC); Lac Baker, 3 (NMC).

Quebec. Bonaventure, 1 (NMC).

Blarina brevicauda brevicauda

Iowa. Hamilton Co.: Jewell, 5 (USNM). Henry Co.: Hillsboro,
2 (USNM). Johnson Co.: Iowa City, 3 (USNM). Marion Co.:
Knoxville, 2 (USNM). Pottawattamie Co.: Council Bluffs, 7
(USNM).

Nebraska. Antelope Co.: Neligh, 1 (USNM). Buffalo Co.:
Kearney, 2 (USNM). Cherry Co.: Valentine, 3 (USNM). Platte
Co.: Columbus, 3 (USNM). Washington Co.: Blair, 1 (USNM).

Blarina brevicauda churchi

North Carolina. Great Smoky Mtns. (county uncertain), 1
(USNM). Magnetic City (county uncertain), 5 (USNM). Bun-

combe Co.: Asheville, 1 (USNM). Cherokee Co.: Murphy, 1
(USNM). Haywood Co.: Waynesville, 1 (USNM). Macon Co.:
Highlands, 2 (USNM). Mitchell Co.: Roan Mountain, 1 6 (USNM).
Transylvania Co.: Pisgah Forest, 1 1 (USNM). Watauga Co.:
Boone, 1 (USNM); Grandfather Mountain, 1 (USNM). Yancey
Co.: Mt. Mitchell, 1 (USNM).

South Carolina. Greenville Co.: Caesars Head, 2 (USNM).
Oconee Co.: Walhalla, 1 (USNM).

Tennessee. Campbell Co.: La Follette, I (USNM). Carter Co.:
Roan Mountain, 12 (USNM). Cocke Co.: AVz mi SE Cosby, 2
(USNM); Mt. Guyot, 1 (USNM). Johnson Co.: Shady Valley, 8
(USNM). Sevier Co.: Great Smoky Mountain National Park, 4
(USNM). Sullivan Co.: 17 mi SE Bristol, 2 (USNM).

Virginia. Grayson Co.: Whitetop Mountain, 5 (USNM).

Blarina brevicauda compacta

Massachusetts. Nantucket Co.: Nantucket Island, 46 (6 MCZ,
35 UCONN, 5 USNM); Siasconset, 4 (UCONN); Quidnet, 29
(UCONN).

136

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Blarina brevicauda kirtlandi

Indiana. Allen Co.: Fort Wayne, 1 (USNM). Newton Co.: Rose
Lawn, 1 (USNM). Porter Co.: no particular locality, 24 (USNM).

Ohio. Ashtabula Co.: 5 mi N Geneva, 7 (CM). Athens Co.:
Athens, 6 (USNM). Erie Co.: Sandusky, 1 (USNM). Fairfield
Co.: no particular locality, 10 (USNM). Geauga Co.: Garretts-
ville, 10 (USNM). Hamilton Co.: California, 1 (CM). Meigs Co.:
Carpenter, 1 (USNM).

Blarina brevicauda manitobensis

Manitoba. Red Rock Lake, 1 (NMC); Rennie, 4 (NMC).

Blarina brevicauda pallida

New Brunswick. Hampton, 5 (USNM).

Nova Scotia. Digby, 4 (USNM); Halifax, 1 (USNM); James
River, 6 (USNM); Kedgemakooge Lake, 13 (USNM).

Maine. Aroostook Co.: Presque Isle, 3 (USNM). Penobscot Co.:
East Branch Penobscot River, 14 (USNM). Piscataquis Co.: 20
mi E Mount Katahdin, 3 (USNM); Basin Pond, Mount Katahdin,
1 (USNM); Chimney Pond, Mount Katahdin, 7 (USNM); Ta-
bleland Mountain, Mount Katahdin, 2 (USNM).

Blarina brevicauda talpoides

Ontario. Chatham, 7 (CM); Komoka, 1 (USNM).

Quebec. Montreal, 1 (USNM); Mt. Tremblant Park, 2 (USNM);
St. Roche de Mekinac, St. Maurice, 1 (USNM); St. Rose, 4
(USNM).

New York. Warren Co.: Lake George, 39 (USNM).

Blarina carolinensis carolinensis

Alabama. Autauga Co.: Autaugaville, 4 (USNM). Cullman
Co.: Ardell, 8 (USNM). Hale Co.: Greensboro, 1 (USNM). Rus-
sell Co.: Seale, 1 (USNM). Sumter Co.: York, 2 (USNM).

Georgia. Crawford Co.: Roberta, 1 (USNM). Early Co.:
Blakeley, 1 (USNM). Grady Co.: Beachton, 4 (USNM). Liberty
Co.: LeConte Plantation, Riceboro, 1 (USNM). Randolph Co.:
Cuthbert, 1 (USNM). Richmond Co.: Augusta, 2 (USNM).
Thomas Co.: Boston, 1 (USNM). Tift Co.: Tifton, 1 (USNM).

North Carolina. Wake Co.: Raleigh, 24 (USNM).

South Carolina. Darlington Co.: Society Hill, 1 (USNM).
Dorchester Co.: Saint George, 1 (USNM). Georgetown Co.:
Georgetown, 1 (USNM); Plantersville, 2 (LISNM). Richland Co.:
Columbia, 3 (USNM). Williamsburg Co.: Lane, 1 (USNM).

Tennessee. “Reelfoot Lake,” (county uncertain), 1 (USNM).
Benton Co.: Big Sandy, 1 (USNM). Fayette Co.: Hickory Withe,
3 (USNM).

Blarina carolinensis minima

Louisiana. East Baton Rouge Par.: 2 mi S Baker, 1 (LSU); 2
mi NNE Baton Rouge, 1 (LSU); 3 mi NSE Baton Rouge, 1 (LSU);

College, Baton Rouge, 1 (LSU); Baton Rouge, 2 (LSU); ‘A mi E
on Greenville Springs Road, Baton Rouge, 1 (LSU); 300 yds.
WNW Plank Rd. and Airline Hwy. Baton Rouge, 1 (LSU); 3 mi
SE Baton Rouge, 1 (LSU); 4 mi SE Baton Rouge, 1 (LSU); 7 mi
SE Baton Rouge, 1 (LSU); 3 mi S, 2 mi E Baton Rouge, 1 (LSU);
5 mi N University, 1 (LSU); 3 mi NE University, 1 (LSU); 3A mi
NE Indian Mound, University, 1 (LSU); Indian Mound, Uni-
versity, 2 (LSU); 5 mi E University, 1 (LSU); 2 mi ESE Uni-
versity, 2 (LSU); 2 mi S LSU, Baton Rouge, 1 (LSU); 3 mi S
University, 1 (LSU); 4.5 mi S University, 1 (LSU); 10 mi S LSU
campus, 1 (LSU); 1 mi SE University, 2 (LSU); 13 mi S Uni-
versity Station, 1 (LSU); 7 mi SW Zachary, I (LSU). Red River
Par.: 5 mi N Coushatta, 1 (LSU).

Blarina carolinensis peninsulae

Florida. Brevard Co.: Georgiana, 1 (USNM). Collier Co.: Deep
Lake, 3 (USNM). Dade Co.: 1 mi W Chekika State Recreation
Area (T.55S, R.37E, Sec. 26), 16 (13 CM, 3 MHP); Miami River,

1 (USNM). Highlands Co.: 6 mi S Lake Placid, 1 (MHP) Osceola
Co.: Kissimmee, 1 (USNM). Palm Beach Co.: Ritta, 1 (USNM).
Polk Co.: Winter Haven, 4 (CM).

Blarina carolinensis shermani

Florida. Lee Co.: 2 mi N Ft. Myers, 1 (USNM); Ft. Myers,

2 (USNM).

Blarina hylophaga hylophaga

Kansas. Ellis Co.: 1 mi S, 6'A mi W Antonino (T. 1 5S, R.20W,
Sec. 2), 14 (MHP); 1 mi S, 6 mi W Antonino, 3 (MHP); 16 mi
N, 1 mi W Hays, 3 (MHP); 9 mi N, 4 mi W Hays (T. 12S, R. 19W,
NE ‘A Sec. 14), 3 (MHP). Rooks Co.: 3 mi S, 3 mi W Stockton,
5 (MHP). Russell Co.: 5 mi N, 1 mi E Dorrance (T. 13S, R.l 1 W,
NE ‘A Sec. 1 7), 2 (MHP); 4 mi N, 4 mi E Dorrance (T. 1 3S, R. 1 1 W,
SW 1/4 Sec. 23), 2 (MHP); 3 mi N, 4 mi E Dorrance (T.13S,
R. 1 1 W, NW >4 Sec. 26), 3 (MHP); 6V2 mi S, Vi mi E Lucas (T. 1 2S,
R.l 1W, NE >/4 Sec. 34), 1 (MHP); 8V2 mi S, V2 mi E Lucas (T.13S,
R. 1 1 W, NE ‘/4 Sec. 10), 1 (MHP). Trego Co.: 'h mi N, 3‘A mi W
Ellis (in Ellis Co.) (T.13S, R.21W, NE 'A Sec. 11), 2 (MHP); 'h
mi N, 3 mi W Ellis (in Ellis Co.)(T.13S, R.21W, NE ‘A Sec. 11),
2 (MHP); 2 'A mi W Ellis (in Ellis Co.), 1 (MHP); 11 mi S, 4 mi
W Ellis (in Ellis Co.) (T.15S, R.21W, NE >A Sec. 2), 1 (MHP);
14'/2 mi S, 5‘/2 mi E Ogallah, 1 (MHP).

Oklahoma. Cleveland Co.: 2 mi S, 1 mi W Norman, 1 (MHP).
Garvin Co.: 3 mi N, 2 mi W Davis (by road), 1 (MHP). Murray
Co.: 1.3 mi S, 2.2 mi W Davis (by road), 2 (MHP); 1.3 mi S, 2
mi W Davis (by road), 2 (MHP); 5.7 mi S, 2.3 mi W Davis (by
road), 1 (MHP).

Blarina hylophaga plumbea

Texas. Aransas Co.: Aransas Wildlife Refuge, 7 (TCWC).

ACKNOWLEDGMENTS

Permission to examine specimens used in this study was grant-
ed by L.G. Vostreys (ANSP), J. E. Guilday and A. D. McCrady
(CM), O. Hawksley (CMSU), M. F. Frazier (FACM), R. W. Gra-
ham and J. J. Saunders (ISM), L. D. Martin (KU), J. A. W. Kirsch
(MCZ), W. W. Dalquest (MWU), C. Van Zyll de Jong (NMC),
H. A. Semken, Jr. (SUI), D. J. Schmidly (TCWC), E. L. Lundelius,
Jr. (TMM), S. D. Webb (UF), R. M. Wetzel (UCONN), P. D.

Gingerich (UMMP), M. R. Voorhies (UNSM), T. Bown (USGS),
and R. W. Purdy and others (USNM). R. E. Eshelman (Calvert
Marine Museum), R. S. Rhodes and H. A. Semken, Jr. (SUI), G.
E. Schultz (West Texas State University), E. L. Lundelius, Jr.
(TMM), and J. E. Guilday generously shared unpublished data.
The senior author also thanks Dr. and Mrs. H. H. Genoways,
Mr. and Mrs. Bruce George, and Mr. and Mrs. J. E. Guilday for

1984

JONES ET AL. — B LARIN A SYSTEM ATICS

137

their hospitality while she was working in Pittsburgh and Wash-
ington. Clerical assistance was provided by B. Lange, M. A.
Schmidt, and I. Unrine. S. B. George, J. Golden, C. J. Jones, D.
C. Lovell, N. D. Moncrief, G. S. Morgan, K. W. Navo, D. K.
Tolliver, and J. D. Wieiand provided assistance during various
stages of the study. Help with the computer was given by D.
Eining and N. D. Moncrief. The first author submitted an earlier
draft of this manuscript as a thesis in partial fulfillment of the
requirements for the M.S. degree in Biology at Fort Hays State
University. Members of Jones’ thesis committee (C. A. Ely, E.

D. Fleharty, G. K. Hulett, and R. J. Zakrzewski) made careful
and constructive recommendations regarding the original manu-
script. Special thanks are due N. J. Hildreth and C. B. Jones for
their generous help with the manuscript; J. E. Guilday and R. J.
Zakrzewski for their invaluable comments regarding the Pleis-
tocene of North America; and C. and C. B. Jones for their as-
sistance throughout the course of this study. This study was par-
tially supported by two grants from the National Science
Foundation (DEB77-13120 to J. R. Choate and DEB77- 12283
to H. H. Genoways).

LITERATURE CITED

Abbe, C., Jr. 1900. The physiography of Allegany County. Pp.
27-55, in Maryland Geological Survey Allegany County,
Johns Hopkins Press, Baltimore, 323 pp.

Auffenberg, W. 1963. The fossil snakes of Florida. Tulane
Studies Zook, 10:131-216.

. 1967. Further notes on fossil box turtles of Florida.

Copeia, 1967:319-325.

Bader, R. S. 1957. Two Pleistocene mammalian faunas from
Alachua County, Florida. Bull. Florida State Mus., 2:53-75.
Baerreis, D. A. 1980. Habitat and climatic interpretation from
terrestrial gastropods at the Cherokee site. Pp. 101-122, in
The Cherokee excavations: Holocene ecology and human
adaptations in northwestern Iowa (D. C. Anderson and H.
A. Semken, Jr., eds.), Academic Press, New York, xvi + 277

pp.

Bailey, R. G. 1978. Description of the ecoregions of the United
States. USDA Forest Service, Intermtn. Reg. Ogden, Utah,
77 pp.

Baker, R. G., and K. L. Van Zant. 1980. Holocene vegeta-
tional reconstruction in northwestern Iowa. Pp. 123-138, in
The Cherokee excavations: Holocene ecology and human
adaptations in northwestern Iowa (D. C. Anderson and H.
A. Semken, Jr., eds.), Academic Press, New York, xvi + 277

pp.

Bayne, C. K. 1969. Evidence of multiple stades in the lower
Pleistocene of northeastern Kansas. Trans. Kansas Acad.
Sci., 71:340-349.

Benninghoff, W. S., and A. L. Stevenson. 1967. Pollen anal-
ysis of cave breccia from Ladds Locality, Bartow County,
Georgia. Bull. Georgia Acad. Sci., 25:188-191.

Bjork, P. R. 1970. The Carnivora of the Hagerman local fauna
(late Pliocene) of southwestern Idaho. Trans. Amer. Phil.
Soc., 60(7): 1-54.

Blair, W. F. 1940. Notes on home ranges and populations of
the short-tailed shrew. Ecology, 21:284-288.

. 1941. Some data on the home ranges and general life

history of the short-tailed shrew, red-backed vole, and wood-
land jumping mouse in northern Michigan. Amer. Midland
Nat., 25:681-685.

. 1958. Distributional patterns of vertebrates in the

southern United States in relation to past and present en-
vironments. Pp. 433-468, in Zoogeography (C. L. Hubbs,
ed.), Amer. Assoc. Sci. Publ., 5!:x + 1-509.

Bowles, J. B. 1975. Distribution and biogeography of mam-
mals oflowa. Spec. Publ. Mus., Texas Tech Univ., 9:1-184.
Brattstrom, B. H. 1967. A succession of Pliocene and Pleis-

tocene snake faunas from the high plains of the United States.
Copeia, 1967:188-202.

Braun, J. K., and M. L. Kennedy. 1983. Systematics of the
genus Blarina in Tennessee and adjacent areas. J. Mamm.,
64:414-425.

Brodkorb, P. 1957. New passerine birds from the Pleistocene
of Reddick, Florida. J. Paleo., 31:129-138.

. 1958. Fossil birds from Idaho. Wilson Bull., 70:237-

242.

. 1959. The Pleistocene avifauna of Arredondo, Florida.

Bull. Florida State Mus., 4:269-291.

. 1963. A giant flightless bird from the Pleistocene of

Florida. Auk, 80:111-115.

Brown, B. 1908. The Conard Fissure, a Pleistocene bone de-
posit in northern Arkansas: with a description of two new
genera and twenty new species of mammals. Mem. Amer.
Mus. Nat. Hist., 9:157-208.

Bryant, M. D. 1941. A far southwestern occurrence of Pitymys
in Texas. J. Mamm., 22:202.

Buckner, C. H. 1966. Populations and ecological relationships
of shrews in tamarack bogs of southeastern Manitoba. J.
Mamm., 47:181-194.

Chantell, C. J. 1970. Upper Pliocene frogs from Idaho. Co-
peia, 1970:654-664.

Choate, J. R. 1969. Taxonomic status of the shrew, Notiosorex
( Xenosorex ) phillipsii Schaldach, 1966 (Mammalia: Insec-
tivora). Proc. Biol. Soc. Washington, 82:469-476.

. 1970. Systematics and zoogeography of middle Amer-
ican shrews of the genus Cryptotis. Univ. Kansas Publ., Mus.
Nat. Hist., 19:195-317.

. 1972. Variation within and among populations of the

short-tailed shrew in Connecticut. J. Mamm., 53:1 16-127.

Choate, J. R., and E. D. Fleharty. 1973. Habitat preference
and spatial relations of shrews in a mixed grassland in Kan-
sas. Southwestern Nat., 18:93-1 14.

. 1975. Synopsis of native. Recent mammals of Ellis

County, Kansas. Occas. Papers Mus., Texas Tech Univ., 37:
1-80.

Churcher, C. S., and M. B. Fenton. 1968. Vertebrate remains
from the Dickson Limestone Quarry, Halton County, On-
tario, Canada. Bull. Natl. Speleol. Soc., 30:1 1-16.

Conant, R. 1975. A field guide to reptiles and amphibians of
eastern and central North America. Houghton Mifflin Co.,
Boston, 429 pp.

Cope, E. D. 1 899. Vertebrate remains from Port Kennedy bone
deposit. J. Acad. Nat. Sci. Philadelphia, 1 1:193-267.

138

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Corgan, J. X. 1976. Vertebrate fossils ofTennessee. Tennessee
Div. Geol. Bull., 77, vi + 100 pp.

Craun, V. S. 1948. Commercial caves of Texas. Bull. Natl.
Speleol. Soc., 10:33-45.

Dalquest, W. W. 1962. The Good Creek formation. Pleisto-
cene of Texas, and its fauna. J. Paleo., 36:568-582.

. 1965. New Pleistocene formation and local fauna from

Hardeman County, Texas. J. Paleo., 39:63-79.

. 1977. Mammals of the Holloman local fauna. Pleis-
tocene of Oklahoma. Southwestern Nat., 22:255-268.

. 1 978. Early Blancan mammals of the Beck Ranch local

fauna of Texas. J. Mamm., 59:269-298.

Dalquest, W. W., E. Roth, and F. Judd. 1969. The mammal
fauna of Schulze Cave, Edwards County, Texas. Bull. Florida
State Mus., 13:205-276.

Davis, L. C. 1969. The biostratigraphy of Peccary Cave, New-
ton County, Arkansas. Proc. Arkansas Acad. Sci., 23:1 92—
196.

Davis, M. B. 1976. Pleistocene biogeography of temperate de-
ciduous forests. Geoscience and Man, 13:13-26.

Davis, W. B. 1978. The mammals of Texas. Texas Parks and
Wildlife Dept., Austin, Bull. 41, 294 pp.

Delcourt, P. A., and H. R. Delcourt. 1977. The Tunica Hills,
Louisiana-Mississippi: late glacial locality for spruce and
deciduous forest species. Quaternary Res., 7:218-237.

Dice, L. R. 1943. The biotic provinces of North America. Univ.
Michigan Press, Ann Arbor, 78 pp.

Dobzhansky, T„ F. J. Ayala, G. L. Stebbins, and J. W. Val-
entine. 1977. Evolution. W. H. Freeman, San Francisco,
xiv + 572 pp.

Dorf, E. 1959. Climatic changes of the past and present. Con-
trib. Mus. Paleo., Univ. Michigan, 13:181-210.

Einsohn, S. D. 1971. The stratigraphy and fauna of a Pleis-
tocene outcrop in Doniphan County, northeastern Kansas.
Unpubl. M.S. thesis, Univ. Kansas, Lawrence, 83 pp.

Elliot, D. G. 1 899. Descriptions of apparently new species and
subspecies of mammals from the Indian Territory. Field
Columbian Mus., Zool. Ser., 1:285-288.

Ellis, L. S., V. E. Diersing, and D. F. Hoffmeister. 1978.
Taxonomic status of short-tailed shrews ( Blarina ) in Illinois.
J. Mamm., 59:305-31 1.

Eshelman, R. E. 1971. The paleoecology of Willard Cave, Del-
aware County, Iowa. Unpubl. M.S. thesis, Univ. Iowa, Iowa
City, 72 pp.

. 1972. Faunal analysis of the Schmitt Site. Proc. Iowa

Acad. Sci., 79:59-61.

. 1975. Geology and paleontology of the early Pleistocene

(late Blancan) White Rock fauna from north-central Kansas.
Mus. Paleo., Univ. Michigan Papers on Paleo. (Claude W.
Hibbard Mem. Vol. 4), 13:iv + 60 pp.

Eshelman, R. E., and C. W. Hibbard. 1981. Nash local fauna
(Pleistocene: Aftonian) of Meade County, Kansas. Contrib.
Mus. Paleo., Univ. Michigan, 25:317-326.

Etheridge, R. 1961. Late Cenozoic glass lizards (Ophisaurus)
from the southern Great Plains. Herpetologica, 17:179-186.

Evans, G. L. 1961. The Friesenhahn Cave. Bull. Texas Mem.
Mus., 2:1-22.

Evernden, J. F., D. E. Savage, G. H. Curtis, and G. T. James.
1 964. Potassium-argon dates and the Cenozoic mammalian
chronology of North America. Amer. J. Sci., 262:145-198.

Fay, L. P. 1980. Mammals of the Garrett Farm and Pleasant
Ridge local biotas (Holocene), Mills County, Iowa. AMQUA
6th Bien. Mtg., Abstracts with Program, pp. 74-75.

Ford, A. L., and O. W. Van Auken. 1982. The distribution
of woody species in the Guadalupe River floodplain forest
in the Edwards Plateau of Texas. Southwestern Nat., 27:
383-392.

Frankel, L. 1 963. The biota of a pre-Illinoian pond in eastern
Nebraska. J. Paleo., 37:249-253.

Frazier, M. K. 1977. New records of Neofiber leonardi (Ro-
dentia: Cricetidae) and the paleoecology of the genus. J.
Mamm., 58:368-373.

French, T. W. 1981. Notes on the distribution and taxonomy
of short-tailed shrews (genus Blarina) in the southeast. Brim-
leyana, 6:101-1 10.

Fullington, R. W. 1978. Supplementary data on the Rex
Rodgers site: Mollusca. Pp. 108-1 13, in Archeology at Mac-
kenzie Reservoir (J. T. Hughes and P. S. Willey, eds.), Texas
Historical Commission, Archeological Survey Report 24,
Austin, xiv + 296 pp.

Galbreath, E. C. 1955. An avifauna from the Pleistocene of
central Kansas. Wilson Bull., 67:62-63.

Gazin, C. L. 1933. A new shrew from the Upper Pliocene of
Idaho. J. Mamm., 14:142-144.

Gf.noways, H. H., and J. R. Choate. 1972. A multivariate
analysis of systematic relationships among populations of
the short-tailed shrew (genus Blarina) in Nebraska. Syst.
Zool., 21:106-1 16.

Genoways, H. H., J. C. Patton III, and J. R. Choate. 1977.
Karyotypes of shrews of the genera Cryptotis and Blarina
(Mammalia: Soricidae). Expenentia, 33:1294-1295.

George, S. B., J. R. Choate, and H. H. Genoways. 1981.
Distribution and taxonomic status of Blarina hvlophaga El-
liot (Insectivora: Soricidae). Ann. Carnegie Mus., 50:493-
513.

George, S. B., H. H. Genoways, J. R. Choate, and R. J. Baker.
1982. Karyotypic relationships within the short-tailed
shrews, genus Blarina. J. Mamm., 63:639-645.

Getz, L. L. 1961. Factors influencing the local distribution of
shrews. Amer. Midland Nat., 65:67-88.

Getz, L. L., and C. W. Hibbard. 1965. A molluscan faunule
from the Seymour Formation, of Baylor and Knox counties,
Texas. Papers Michigan Acad. Sci., Arts, and Letters, 50:
275-297.

Gidley, J. W., and C. L. Gazin. 1933. New Mammalia in the
Pleistocene fauna from Cumberland Cave. J. Mamm., 14:
343-357.

. 1938. The Pleistocene vertebrate fauna from Cumber-
land Cave, Maryland. Bull. U.S. Natl. Mus., 171:1-99.

Gould, S. J., and N. Eldredge. 1977. Punctuated equilibria:
the tempo and mode of evolution reconsidered. Paleobiol-
ogy, 3:1 15-151.

Graham, R. W. 1972. Biostratigraphy and paleoecological sig-
nificance of the Conard Fissure local fauna with emphasis
on the genus Blarina. Unpubl. M.S. thesis, Univ. Iowa, Iowa
City, 90 pp.

. 1976. Late Wisconsin mammalian faunas and envi-
ronmental gradients of the eastern United States. Paleo-
biology, 2:343-350.

. 1979. Paleoclimates and late Pleistocene faunal prov-
inces in North America. Pp. 49-69, in Pre-Llano cultures of
the Americas: paradoxes and possibilities (R. L. Humphrey
and D. Stanford, eds.), The Anthropological Society of
Washington, 1 50 pp.

Graham, R. W., and H. A. Semken. 1976. Paleoecological
significance of the short-tailed shrew (Blarina), with a sys-

1984

JONES ET AL. — B LARINA SYSTEM ATICS

139

tematic discussion of Blarina ozarkensis. J. Mamm., 57:
433-449.

Grant, V. 1971. Plant speciation. Columbia Univ. Press, New
York, x + 435 pp.

Guilday, J. E. 1957. Individual and geographic variation in
Blarina brevicauda from Pennsylvania. Ann. Carnegie Mus.,
35:41-68.

. 1962a. Notes on Pleistocene vertebrates from Wythe

County, Virginia. Ann. Carnegie Mus., 36:77-86.

. 19626. The Pleistocene local fauna of the Natural

Chimneys, Augusta County, Virginia. Ann. Carnegie Mus.,
36:87-122.

. 1971. The Pleistocene history of the Appalachian mam-
mal fauna. Pp. 233-262, in The distributional history of the
biota of the southern Appalachians, Part III: Vertebrates (P.
C. Holt, ed.), Res. Div. Monogr. 4, VPI and SU, Blacksburg.

. 1977. Sabertooth cat, Smilodon floridanus (Leidy), and

associated fauna from a Tennessee cave (40 Dv 40), the First
American Bank site. J. Tennessee Acad. Sci., 52:84-94.

Guilday, J. E., and H. W. Hamilton. 1973. The late Pleis-
tocene small mammals of Eagle Cave, Pendleton County,
West Virginia. Ann. Carnegie Mus., 44:45-58.

. 1978. Ecological significance of displaced boreal mam-
mals in West Virginia caves. J. Mamm., 59:176-181.

Guilday, J. E., H. W. Hamilton, E. Anderson, and P. W.
Parmalee. 1978. The Baker Bluff Cave deposit, Tennessee,
and the late Pleistocene faunal gradient. Bull. Carnegie Mus.
Nat. Hist., 11:1-67.

Guilday, J. E., H. W. Hamilton, and A. D. McCrady. 1966.
The bone breccia of Bootlegger Sink, York County, PA. Ann.
Carnegie Mus., 38:145-163.

. 1969. The Pleistocene vertebrate fauna of Robinson

Cave, Overton County, Tennessee. Palaeovertebrata, 2:25-
75.

. 1971. The Welsh Cave peccaries (Platygonus) and as-
sociated fauna, Kentucky Pleistocene. Ann. Carnegie Mus.,
43:249-320.

Guilday, J. E., H. W. Hamilton, and P. W. Parmalee. 1975.
Caribou ( Rangifer tarandus L.) from the Pleistocene of Ten-
nessee. J. Tennessee Acad. Sci., 50:108-1 1 1.

Guilday, J. E., P. S. Martin, and A. D. McCrady. 1 964. New
Paris No. 4: a Pleistocene cave deposit in Bedford County,
Pennsylvania. Bull. Natl. Speleol. Soc., 26:121-194.

Guilday, J. E., and P. W. Parmalee. 1972. Quaternary peri-
glacial records of voles of the genus Phenacomys Merriam
(Cricetidae: Rodentia). Quaternary Res., 2:170-175.

Guilday, J. E., P. W. Parmalee, and H. W. Hamilton. 1977.
The Clark’s Cave bone deposit and the late Pleistocene pa-
leoecology of the central Appalachian mountains of Virginia.
Bull. Carnegie Mus. Nat. Hist., 2:1-87.

Gut, H. J. 1959. A Pleistocene vampire bat from Florida. J.
Mamm., 40:534-538.

Gut, H. J., and C. E. Ray. 1963. The Pleistocene vertebrate
fauna of Reddick, Florida. Quart. J. Florida Acad. Sci., 26:
315-328.

Hall, E. R. 1981. The mammals of North America. John Wiley
and Sons, New York, 2nd ed., l:xviii + 1-600 + 90 and 2:
vi + 601-1 181 + 90.

Hall, E. R., and K. R. Kelson. 1959. The mammals of North
America. Ronald Press, New York, l:xxx + 1-546 + 79.

Hall, S. A. 1982. Late Holocene paleoecology of the Southern
Plains. Quaternary Res., 17:391-407.

Hallberg, G. R., H. A. Semken. and L. C. Davis. 1974. Qua-

ternary records of Microtus xanthognathus (Leach), the yel-
low-cheeked vole, from northwestern Arkansas and south-
western Iowa. J. Mamm., 55:640-645.

Hamon, J. H. 1964. Osteology and paleontology of the pas-
serine birds of the Reddick, Florida, Pleistocene. Bull. Flor-
ida Geol. Surv., 44:vii + 210.

Handley, C. O., Jr. 1979. Mammals of the Dismal Swamp: a
historical account. Pp. 297-357, in The Great Dismal Swamp
(P. W. Kirk, Jr., ed.), Univ. Press Virginia, Charlottesville,
xii + 427 pp.

Hawksley, O., J. F. Reynolds, and R. L. Foley. 1973. Pleis-
tocene vertebrate fauna of Bat Cave, Pulaski County, Mis-
souri. Bull. Natl. Speleol. Soc., 35:61-87.

Hay, O. P. 1920. Descriptions of some Pleistocene vertebrates
found in the United States. Proc. U.S. Natl. Mus., 58:83—
146.

Helwig, J. T., and K. A. Council (eds.). 1979. SAS user’s
guide. SAS Institute, Raleigh, North Carolina, 494 pp.

Hibbard, C. W. 1943. The Rezabek fauna, a new Pleistocene
fauna from Lincoln County, Kansas. Univ. Kansas Sci. Bull.,
29:235-247.

. 1944. Stratigraphy and vertebrate paleontology of Pleis-
tocene deposits of southwestern Kansas. Bull. Geol. Soc.
Amer., 55:707-754.

. 1950. Mammals of the Rexroad formation from Fox

Canyon, Kansas. Contrib. Mus. Paleo., Univ. Michigan, 8:
113-192.

. 1952. Vertebrate fossils from late Cenozoic deposits of

central Kansas. Univ. Kansas Paleo. Contrib. Vertebrata, 2:
1-14.

. 1953a. The insectivores of the Rexroad fauna, upper

Pliocene of Kansas. J. Paleo., 27:21-32.

. 1953 b. The Saw Rock Canyon fauna and its strati-
graphic significance. Papers Michigan Acad. Sci., Arts, and
Letters, 38:387-41 1.

. 1955. The Jinglebob interglacial (Sangamon?) fauna

from Kansas and its climatic significance. Contrib. Mus.
Paleo., Univ. Michigan, 12:179-228.

. 1956. Vertebrate fossils from the Meade formation of

southwestern Kansas. Papers Michigan Acad. Sci., Arts, and
Letters, 41:145-203.

. 1957. Notes on late Cenozoic shrews. Trans. Kansas

Acad. Sci., 60:327-336.

. 1958. Summary of North American Pleistocene mam-
malian local faunas. Papers Michigan Acad. Sci., Arts, and
Letters, 43:3-32.

. 1960. An interpretation of Pliocene and Pleistocene

climates in North America. Ann. Rep. Michigan Acad. Sci.,
Arts, and Letters, 62:5-30.

. 1963. Paleontology. A late Illinoian fauna from Kansas

and its climatic significance. Papers Michigan Acad. Sci.,
Arts, and Letters, 48:187-221.

. 1967. New rodents from the late Cenozoic of Kansas.

Papers Michigan Acad. Sci., Arts, and Letters, 52:1 15-131.

. 1970. Pleistocene mammalian local faunas from the

Great Plains and Central Lowland provinces of the United
States. Pp. 395-433, in Pleistocene and Recent environ-
ments of the central Great Plains (W. Dort, Jr., and J. K.
Jones, Jr., eds.), Univ. Press Kansas, Lawrence, 433 pp.

Hibbard, C. W., and P. R. Bjork. 1971. The insectivores of
the Hagerman local fauna, upper Pliocene of Idaho. Contrib.
Mus. Paleo., Univ. Michigan, 23:171-180.

Hibbard, C. W., and W. W. Dalquest. 1966. Fossils from the

140

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Seymour formation of Knox and Baylor counties, Texas,
and their bearing on the late Kansan climate of that region.
Contrib. Mus. Paleo., Umv. Michigan, 21:1-66.

. 1973. Proneofiber, a new genus of vole (Cricetidae:

Rodentia) from the Pleistocene Seymour Formation of Tex-
as, and its evolutionary and stratigraphic significance. Qua-
ternary Res., 3:269-274.

Hibbard, C. W., and D. W. Taylor. 1960. Two late Pleistocene
faunas from southwestern Kansas. Contrib. Mus. Paleo.,
Univ. Michigan, 16:1-223.

Hibbard, C. W., D. E. Ray, D. E. Savage, D. W. Taylor, and
J. E. Guilday. 1965. Quaternary mammals of North
America. Pp. 509-525, in The Quaternary of the United
States (H. E. Wright, Jr., and D. G. Frey, eds.), Princeton
Univ. Press, Princeton, 922 pp.

Hibbard, C. W., R. J. Zakrzewski, R. E. Eshelman, G. Edmond,
C. D. Griggs, and C. Griggs. 1978. Mammals from the
Kanopolis local fauna. Pleistocene (Yarmouth) of Ellsworth
County, Kansas. Contrib. Mus. Paleo., Univ. Michigan, 25:
11-44.

Hoffmann, R. S., and J. K. Jones, Jr. 1970. Influence of late-
glacial and post-glacial events on the distribution of Recent
mammals on the northern Great Plains. Pp. 355-394, in
Pleistocene and Recent environments of the central Great
Plains (W. Dort, Jr., and J. K. Jones, Jr., eds.), Univ. Press
Kansas, Lawrence, 433 pp.

Holman, J. A. 1 959a. Amphibians and reptiles from the Pleis-
tocene (Illinoian) of Williston, Florida. Copeia, 1 959:96—
102.

. 1959ft. Birds and mammals from the Pleistocene of

Williston, Florida. Bull. Florida State Mus., 5:1-24.

. 1967. A Pleistocene herpetofauna from Ladds, Georgia.

Bull. Georgia Acad. Sci., 25:154-165.

. 1972. Herpetofauna of the Kanopolis local fauna (Pleis-
tocene: Yarmouth) of Kansas. Michigan Academician, 5:87-
98.

Hood, C. H., and O. Hawksley. 1975. A Pleistocene fauna
from Zoo Cave, Taney County, Missouri. Missouri Speleol.,
15:1-42.

Hoyer, B. E. 1980. The geology of the Cherokee Sewer Site.
Pp. 21 -66, in The Cherokee excavations. Holocene ecology
and human adaptations in northwestern Iowa (D. C. An-
derson and H. A. Semken, Jr., eds.), Academic Press, New
York, xvi + 277 pp.

Jammot, D. 1972. Relationships between the new species Sorex
scottensis and the fossil shrews Sorex cinereus Kerr. Mam-
malia, 36:449-458.

Jenkins, J. T., Jr. 1972. The Pleistocene geology and paleon-
tology of the Fremont County Quarry, Fremont County,
Iowa. Unpubl. M.S. thesis, Univ. Iowa, Iowa City, 54 pp.

Jones, J. K., Jr., and B. P. Glass. 1960. The short-tailed shrew,
Blarina brevicauda, in Oklahoma. Southwestern Nat., 5:136-
142.

Jones, J. K., Jr., and R. B. Loomis. 1954. Records of the short-
tailed shrew and least shrew from Colorado. J. Mamm., 35:
110.

Jones, J. K., Jr., D. C. Carter, H. H. Genoways, R. S.
Hoffmann, and D. W. Rice. 1982. Revised checklist of
North American mammals north of Mexico, 1982. Occas.
Papers Mus., Texas Tech Univ., 80:1-22.

Kapp, R. O. 1970. Pollen analysis of pre-Wisconsin sediments
from the Great Plains. Pp. 143-155, in Pleistocene and Re-
cent environments of the central Great Plains (W. Dort, Jr.,

and J. K. Jones, Jr., eds.), Univ. Press Kansas, Lawrence,
433 pp.

Kennerly, T. E., Jr. 1956. Comparisons between fossil and
Recent species of the genus Perognathus. Texas J. Sci., 8:
74-86.

Kirkland, G. L., Jr. 1978. The short-tailed shrew, Blarina
brevicauda (Say), in the central mountains of West Virginia.
Proc. Pennsylvania Acad. Sci., 52:126-130.

Klippel, W. E., and P. W. Parmalee. 1982. Diachronic vari-
ation in insectivores from Cheek Bend Cave and environ-
mental change in the Midsouth. Paleobiology, 8:447-458.

Kramer, T. L. 1972. The paleoecology of the postglacial Mud
Creek biota. Cedar and Scott counties, Iowa. Unpubl. M.S.
thesis, Univ. Iowa, Iowa City, 69 pp.

Krivda, W. 1957. New Manitoba record for the short-tailed
shrew. Canadian Field-Nat., 71:83.

Kuchler, A. W. 1974. A new vegetation map of Kansas. Ecol-
ogy, 55:586-604.

Kurten, B. 1965. The Pleistocene Felidae of Florida. Bull.
Florida State Mus., 9:215-273.

Kurten, B., and E. Anderson. 1980. Pleistocene mammals of
North America. Columbia Univ. Press, New York, 442 pp.

LaRocque, A. 1967. Pleistocene Molluscaofthe Ladds deposit,
Bartow County, Georgia. Bull. Georgia Acad. Sci., 25:167-
187.

LaSalle, P., and J. E. Guilday. 1980. Cavemede Saint-Elzear-
de-Bonaventure. Rapport preliminaire sur les fouilles de 1977
et 1978. Ministere Ener. et Res., Quebec, 31 pp.

Lee, M. R., and E. G. Zimmerman. 1969. Robertsonian poly-
morphism in the cotton rat, Sigmodonfulviventer. J. Mamm.,
50:333-339.

Leonard, A. B. 1950. A Yarmouthian molluscan fauna in the
midcontinent region of the United States. Univ. Kansas,
Paleo. Contrib. Mollusca, 3:1-48.

Ligon, J. D. 1965. A Pleistocene avifauna from Haile, Florida.
Bull. Florida State Mus., 10:127-158.

Lindsay, E. H., N. M. Johnson, and N. D. Opdyke. 1975.
Preliminary correlation of North American land mammal
ages and geomagnetic chronology. Mus. Paleo., Univ. Mich-
igan Papers Paleo., 12:1 1 1-1 19.

Lipps, L., and C. E. Ray. 1967. The Pleistocene fossiliferous
deposit at Ladds, Bartow County, Georgia. Bull. Georgia
Acad. Sci., 25:113-119.

Lundelius, E. L., Jr. 1960. Mylohyus nasutus, long-nosed pec-
cary from the Texas Pleistocene. Bull. Texas Mem. Mus., 1:
1-40.

. 1967. Late-Pleistocene and Holocene faunal history of

central Texas. Pp. 287-319, in Pleistocene extinctions: the
search for a cause (P. S. Martin and H. E. Wright, Jr., eds.),
Yale Univ. Press, New Haven, 453 pp.

Martin, R. A. 1968. Late Pleistocene distribution of Microtus
pennsylvanicus. J. Mamm., 49:265-271.

. 1969. Taxonomy of the giant Pleistocene beaver Cas-

toroides from Florida. J. Paleo., 43:1033-1041.

. 1973a. Description of a new genus of weasel from the

Pleistocene of South Dakota. J. Mamm., 54:924-929.

. 1973ft. The Java local fauna, Pleistocene of South Da-
kota: a preliminary report. Bull. New Jersey Acad. Sci., 18:
48-56.

. 1974. Fossil mammals of the Coleman IIA fauna, Sum-
ter County. Pp. 35-99, in Pleistocene mammals of Florida
(S. D. Webb, ed.), Univ. Press Florida, Gainesville, x + 270
PP

1984

JONES ET AL.- BLARINA SYSTEM ATICS

141

. 1975. Allophaiomys Kormos from the Pleistocene of

North America. Mus. Paleo., Univ. Michigan Papers Paleo.
(Claude W. Hibbard Mem. Vol. 3), 12:97-100.

. 1977. Late Pleistocene Eumops from Florida. Bull. New

Jersey Acad. Sci., 22:18-19.

Martin, R. A., and S. D. Webb. 1974. Late Pleistocene mam-
mals from the Devil’s Den fauna, Levy County. Pp. 114-
145, in Pleistocene mammals of Florida (S. D. Webb, ed.),
Univ. Press Florida, Gainesville, x + 270 pp.

Matthews, W. H., III. 1963. The geologic story of Longhorn
Cavern. Bureau of Economic Geol., Univ. Texas, Austin,
Guidebook, 4:1-50.

Mayr, E. 1969. Principles of systematic zoology. McGraw-Hill,
New York, xi + 428 pp.

McMullen, T. L. 1975. Shrews from the late Pleistocene of
central Kansas, with the description of a new species of
Sorex. J. Mamm., 56:316-320.

Meade, G. E. 1961. The saber-toothed cat, Dinobastis serus.
Bull. Texas Mem. Mus., 2:23-60.

Mecham, J. S. 1958 (printed 1959). Some Pleistocene am-
phibians and reptiles from Friesenhahn Cave, Texas. South-
western Nat., 3:17-27.

Mehringer, P. J., Jr., C. E. Schweger, W. R. Wood, and R. B.
McMillan. 1968. Late- Pleistocene boreal forest in the
western Ozark highlands? Ecology, 49:567-568.

Mehringer, P. J., Jr., J. E. King, and E. H. Lindsay. 1970. A
record of Wisconsin-age vegetation and fauna from the Ozarks
of western Missouri. Pp. 173-183, in Pleistocene and Recent
environments of the central Great Plains (W. Dort, Jr., and
J. K. Jones, Jr., eds.), Univ. Press Kansas, Lawrence, 433

pp.

Mengel, R. M. 1952. White pelican from the Pleistocene of
Oklahoma. Auk, 69:81-82.

Merriam, C. H. 1895. Revision of the shrews of the American
genera Blarina and Notiosorex. N. Amer. Fauna, 10:1-34.

Meylan, A. 1967. Formules chromosomiques et polymor-
phisme Robertsonien chez Blarina brevicauda (Say) (Mam-
malia: Insectivora). Canadian J. Zool., 45:1 1 19-1 127.

Miller, B. B. 1966. Five Illinoian molluscan faunas from the
southern Great Plains. Malacologia, 4:173-260.

. 1975. A sequence of radiocarbon-dated Wisconsinan

nonmarine molluscan faunas from southwestern Kansas-
northwestern Oklahoma. Mus. Paleo., Univ. Michigan Pa-
pers Paleo. (Claude W. Hibbard Mem. Vol. 3), 12:9-18.

Miller, R. A. 1974. The geologic history of Tennessee. Ten-
nessee Div. Geol. Bull., 74:1-36.

Milstead, W. W. 1956. Fossil turtles of Friesenhahn Cave,
Texas, with the description of a new species of Testudo.
Copeia, 1956:162-171.

Moncrief, N. D., J. R. Choate, and H. H. Genoways. 1982.
Morphometric and geographic relationships of short-tailed
shrews (genus Blarina) in Kansas, Iowa, and Missouri. Ann.
Carnegie Mus., 51:157-180.

National Oceanic and Atmospheric Administration (NOAA).
1974. Climates of the states. Water Information Center, Inc.,
Port Washington, New York, 2 vols., 982 pp.

Neff, N. A. 1975. Fishes of the Kanopolis local fauna (Pleis-
tocene) of Ellsworth County, Kansas. Mus. Paleo., Univ.
Michigan Papers on Paleo. (Claude W. Hibbard Mem. Vol.
3), 12:39-48.

Neff, N. A., and L. F. Marcus. 1980. A survey of multivariate
methods for systematics. New York, privately published,

243 pp.

Nelson, R. S., and H. A. Semken. 1970. Paleoecological and
stratigraphic significance of the muskrat in Pleistocene de-
posits. Bull. Geol. Soc., 81:3733-3738.

Nero, R. W. 1960. Short-tailed shrew north of the North Sas-
katchewan River. Blue Jay, 18:41-42.

Nicholas, G. 1954. Pleistocene ecology of Cumberland Bone
Cave. American Caver, 16:29-39.

Ober, L. D. 1 978. The Monkey Jungle, a late Pleistocene fossil
site in southern Florida. Plaster Jacket, 28:1-13.

Oesch, R. D. 1 967. A preliminary investigation of a Pleistocene
vertebrate fauna from Crankshaft Pit, Jefferson County, Mis-
souri. Bull. Natl. Speleol. Soc., 29:163-185.

Olsen, S. J. 1958. The bog lemming from the Pleistocene of
Florida. J. Mamm., 39:537-540.

Parmalee, P. W. 1967. A Recent cave bone deposit in south-
western Illinois. Bull. Natl. Speleol. Soc., 29:119-147.

Parmalee, P. W., and R. D. Oesch. 1972. Pleistocene and
Recent faunas from the Brynjulfson caves, Missouri. Illinois
State Mus. Rep. Invest., 25:1-52.

Parmalee, P. W., R. D. Oesch, and J. E. Guilday. 1969. Pleis-
tocene and Recent vertebrate faunas from Crankshaft Cave,
Missouri. Illinois State Mus. Rep. Invest., 14:1-37.

Patton, T. H. 1963a. Fossil remains of southern bog lemming
in Pleistocene deposits of Texas. J. Mamm., 44:275-277.

. 19636. Fossil vertebrates from Miller’s Cave, Llano

County, Texas. Bull. Texas Mem. Mus., 7:1-41.

Paulson, G. R. 1961. The mammals of the Cudahy fauna.
Papers Michigan Acad. Sci., Arts, and Letters, 46:127-153.

Peterson, O. A. 1926. The fossils of the Frankstown Cave,
Blair County, Pennsylvania. Ann. Carnegie Mus., 16:249-
315.

Pettus, D. 1956. Fossil rabbits (Lagomorpha) of the Friesen-
hahn Cave deposit, Texas. Southwestern Nat., 1:109-1 15.

Pruitt, W. O., Jr. 1953. An analysis of some physical factors
affecting the local distribution of the shorttail shrew (Blarina
brevicauda) in the northern part of the Lower Peninsula of
Michigan. Misc. Publ. Mus. Zool., Univ. Michigan, 79:1-
39.

. 1959. Microclimates and local distribution of small

mammals on the George Reserve, Michigan. Misc. Publ.
Mus. Zool., Univ. Michigan, 109:1-27.

Pyle, K. B. 1980. The Cherokee large mammal fauna. Pp. 1 7 1-
196, in The Cherokee excavations: Holocene ecology and
human adaptations in northwestern Iowa (D. C. Anderson
and H. A. Semken, Jr., eds.). Academic Press, New York,
xvi + 277 pp.

Quinn, J. H. 1972. Extinct mammals in Arkansas and related
C'4 dates circa 3000 years ago. 24th Inter. Geol. Cong., 12:
89-96.

Ray, C. E. 1965. A new chipmunk, Tamias aristus, from the
Pleistocene of Georgia. J. Paleo., 39:1016-1022.

. 1967. Pleistocene mammals from Ladds, Bartow Coun-
ty, Georgia. Bull. Georgia Acad. Sci., 25:120-150.

Ray, C. E., S. J. Olsen, and H. J. Gut. 1963. Three mammals
new to the Pleistocene fauna of Florida, and a reconsider-
ation of five earlier records. J. Mamm., 44:373-395.

Reese, J. L. 1972. The Schmitt Cave. Proc. Iowa Acad. Sci.,
79:56-58.

Repenning, C. A. 1967. Subfamilies and genera of the Soric-
idae. U.S.G.S. Prof. Paper, 565:iv + 74 pp.

Rhodes, R. S., II. 1 982. Mammalian paleoecology of the Farm-
dalian Craigmile and the Woodfordian Waubonsie local fau-

142

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

nas, southwestern Iowa. Unpubl. Ph.D. dissert., Univ. Iowa,
Iowa City, vi + 132 pp.

Richmond, N. D. 1963. Evidence against the existence of croc-
odiles in Virginia and Maryland during the Pleistocene. Proc.
Biol. Soc. Washington, 76:65-67.

Rinker, G. C., and C. W. Hibbard. 1952. A new beaver, and
associated vertebrates, from the Pleistocene of Oklahoma.
J. Mamm, 33:98-101.

Robertson, J. S., Jr. 1974. Fossil Bison of Florida. Pp. 214-
246, in Pleistocene mammals of Florida (S. D. Webb, ed.),
Univ. Press Florida, Gainesville, x + 270 pp.

Ross, H. H. 1970. The ecological history of the Great Plains:
evidence from grassland insects. Pp. 225-240, in Pleistocene
and Recent environments of the central Great Plains (W.
Dort, Jr., and J. K. Jones, Jr., eds.), Univ. Press Kansas,
Lawrence, 433 pp.

Roth, E. L. 1972. Late Pleistocene mammals from Klein Cave,
Kerr County, Texas. Texas J. Sci., 24:75-84.

Saunders, J. J. 1977. Late Pleistocene vertebrates of the west-
ern Ozark highland, Missouri. Illinois State Mus. Rep. In-
vest., 33:1-1 18.

Schmidly, D. J., and W. A. Brown. 1979. Systematics of short-
tailed shrews (genus Blarina) in Texas. Southwestern Nat.,
24:39-48.

Schultz, C. B., and L. D. Martin. 1970. Quaternary mam-
malian sequence in the central Great Plains. Pp. 341-353,
in Pleistocene and Recent environments of the central Great
Plains (W. Dort, Jr., and J. K. Jones, Jr., eds.), Univ. Press
Kansas, Lawrence, 433 pp.

Schultz, C. B.. and L. G. Tanner. 1957. Medial Pleistocene
fossil vertebrate localities in Nebraska. Bull. Univ. Nebraska
State Mus., 4:59-81.

Schultz, C. B., L. D. Martin, L. G. Tanner, and R. G. Corner.
1978. Provincial land mammal ages for the North American
Quaternary. Trans. Nebraska Acad. Sci., 5:59-64.

Schultz, G. E. 1967. Four superimposed late-Pleistocene fau-
nas from southwest Kansas. Pp. 321-336, in Pleistocene
extinctions: the search for a cause (P. S. Martin and H. E.
Wright, Jr., eds.), Yale Univ. Press, New Haven, 453 pp.

. 1969. Geology and paleontology of a late Pleistocene

basin in southwest Kansas. Geol. Soc. Amer., Spec. Paper
105 : viii + 1-85.

. 1978. Supplementary data on the Rex Rodgers site:

Micromammals. P. 114, in Archeology at Mackenzie Res-
ervoir (J. T. Hughes and P. S. Willey, eds.), Texas Historical
Commission, Archeological Survey Report 24, Austin, xiv
+ 296 pp.

Sealander, J. A. 1979. A guide to Arkansas mammals. River
Road Press, Conway, Arkansas, x + 3 1 3 pp.

Self, H. 1978. Environment and man in Kansas. Regents Press
Kansas, Lawrence, 288 pp.

Sellards, E. H. 1919. The geology and mineral resources of
Bexar County. Univ. Texas Bull., 1932:1-202.

. 1 937. The Vero finds in the light of present knowledge.

Pp. 193-210, in Early man (G. G. MacCurdy, ed.), J. P.
Lippincott Co., Philadelphia, 362 pp.

Semken, H. A., Jr. 1961. Fossil vertebrates from Longhorn
Cavern, Burnet County, Texas. Texas J. Sci., 13:290-310.

. 1966. Stratigraphy and paleontology of the McPherson

Equus beds (Sandahl local fauna) McPherson County, Kan-
sas. Contrib. Mus. Paleo., Univ. Michigan, 20:121-178.

. 1980. Holocene climatic reconstructions derived from

the three micromammal bearing cultural horizons of the
Cherokee Sewer Site, northwestern Iowa. Pp. 67-99, in The
Cherokee excavations: Holocene ecology and human ad-
aptations in northwestern Iowa (D. C. Anderson and H. A.
Semken, Jr., eds.). Academic Press, New York, xvi + 277

pp.

Sheldon, C. 1936. The mammals of Lake Kedgemakooge and
vicinity. Nova Scotia. J. Mamm., 17:207-215.

Shutler, R., Jr., D. C. Anderson, L. S. Tatum, and H. A.
Semken, Jr. 1980. Excavation techniques and synopsis of
results derived from the Cherokee project. Pp. 1-20, in The
Cherokee excavations: Holocene ecology and human ad-
aptations in northwestern Iowa (D. C. Anderson and H. A.
Semken, Jr., eds.), Academic Press, New York, xvi + 277

pp.

Simpson, G. G. 1928. Pleistocene mammals from a cave in
Citrus County, Florida. Amer. Mus. Novitates, 328:1-16.

. 1929a. Pleistocene mammalian fauna of the Seminole

Field, Pinellas County, Florida. Bull. Amer. Mus. Nat. Hist.,
56:561-599.

. 1929A The extinct land mammals of Florida. Florida

State Geol. Survey Annual Report, 20:229-279.

Skinner, M. F., C. W. Hibbard, E. D. Gutentag, G. R. Smith,
J. G. Lundberg, J. A. Holman, J. A. Feduccia, and P. V.
Rich. 1972. Early Pleistocene preglacial and glacial rocks
and faunas of north-central Nebraska. Bull. Amer. Mus. Nat.
Hist., 148:1-148.

Slaughter, B. H. 1967. Animal ranges as a clue to late-Pleis-
tocene extinctions. Pp. 155-167, in Pleistocene extinctions:
the search for a cause (P. S. Martin and H. E. Wright, Jr.,
eds.), Yale Univ. Press, New Haven, 453 pp.

Slaughter, B. H., and B. R. Hoover. 1963. Sulphur River
formation and the Pleistocene mammals of the Ben Franklin
local fauna. J. Grad. Res. Center, 31:132-148.

Smith, C. L. 1954. Pleistocene fishes of the Berends fauna of
Beaver County, Oklahoma. Copeia, 1954:282-289.

. 1958. Additional Pleistocene fishes from Kansas and

Oklahoma. Copeia, 1958:176-180.

Smith, G. R. 1963. A late Illinoian fish fauna from southwestern
Kansas and its climatic significance. Copeia, 1963:278-285.

Smith, R. L. 1959. Abundance of small mammals in conifer
plantations. J. Mamm., 40:253-254.

Starrett, A. 1956. Pleistocene mammals of the Berends fauna
of Oklahoma. J. Paleo., 30:1 187-1 192.

Stephens, J. J. 1960. Stratigraphy and paleontology of a late
Pleistocene basin, Harper County, Oklahoma. Bull. Geol.
Soc. Amer., 71:1675-1702.

Tamsitt, J. R. 1957. Peromyscus from the late Pleistocene of
Texas. Texas J. Sci., 9:355-363.

Tate, C. M., J. F. Pagels, and C. O. Handley, Jr. 1980. Dis-
tribution and systematic relationship of two kinds of short-
tailed shrews (Soricidae: Blarina) in south-central Virginia.
Proc. Biol. Soc. Washington, 93:50-60.

Taylor, D. W. 1954. A new Pleistocene fauna and new species
of fossil snails from the high plains. Occas. Papers Mus.
Zook, Univ. Michigan, 557:1-16.

. 1966. Summary of North American Blancan nonmar-
ine molluscs. Malacologia, 4:1-172.

Taylor, D. W., and C. W. Hibbard. 1955. A new Pleistocene
fauna from Harper County, Oklahoma. Oklahoma Geol.
Survey, 37:1-23.

Van der Meulen, A. J. 1978. Microtus and Pitymys (Arvi-

1984

JONES ET AL . — BLA R IN A SYSTEM ATICS

143

colidae) from Cumberland Cave, Maryland, with a com-
parison of some New and Old World species. Ann. Carnegie
Mus., 47:101-145.

Voorhies, M. R. 1974. Pleistocene vertebrates with boreal af-
finities in the Georgia Piedmont. Quaternary Res., 4:85-93.

Watts, W. A. 1 970. The full-glacial vegetation of northwestern
Georgia. Ecology, 51:17-33.

. 1980. The late Quaternary vegetation history of the

southeastern United States. Ann. Rev. Ecol. Syst., 1 1:387-
409.

Webb, S. D. 1974. Chronology of Florida Pleistocene mammals.
Pp. 5-31, in Pleistocene mammals of Florida (S. D. Webb,
ed.), Univ. Press Florida, Gainesville, x + 270 pp.

Webb, T., III. 1981. The past 1 1 ,000 years of vegetational change
in eastern North America. Bioscience, 31:501-506.

Weems, R. E. 1972. Vertebrate remains from a (?) post-Wis-
consin fissure deposit near Pearisburg, Virginia. Virginia J.
Sci., 23:137 (Abst.).

Weems, R. E., and B. B. Higgins. 1977. Post-Wisconsinan
vertebrate remains from a fissure deposit near Ripplemead,
Virginia. Bull. Natl. Speleol. Soc., 39:106-108.

Weigel, R. D. 1962. Fossil vertebrates of Vero, Florida. Florida
Geol. Surv., Spec. Publ., 10:1-59.

Wells, P. V. 1970. Vegetational history of the Great Plains: a
post-glacial record of coniferous woodland in southeastern
Wyoming. Pp. 185-202, in Pleistocene and Recent environ-
ments of the central Great Plains (W. Dort, Jr., and J. K.
Jones, Jr., eds.), Univ. Press Kansas, Lawrence, 433 pp.

Wendland, W. M. 1980. Holocene climatic reconstructions on
the prairie peninsula. Pp. 139-148, in The Cherokee exca-
vations. Holocene ecology and human adaptations in north-
western Iowa (D. C. Anderson and H. A. Semken, Jr., eds.).
Academic Press, New York, xvi + 277 pp.

Wetmore, A 1927. A record of the ruffed grouse from the
Pleistocene of Maryland. Auk, 44:561.

. 1967. Pleistocene Aves from Ladds, Georgia. Bull.

Georgia Acad. Sci., 25: 151-153.

Wiley, E. O. 1978. The evolutionary species concept recon-
sidered. Syst. Zool., 27:17-26.

Wright, H. E., Jr. 1968. The roles of pine and spruce in the
forest history of Minnesota and adjacent areas. Ecology, 49:
937-955.

. 1970. Vegetational history of the Central Plains. Pp.

157-172, in Pleistocene and Recent environments of the
central Great Plains (W. Dort, Jr., and J. K. Jones, Jr., eds.),
Univ. Press Kansas, Lawrence, 433 pp.

. 1971. Late Quaternary vegetational history of North

America. Pp. 425-464, in The late Cenozoic glacial ages (K.
K. Turekian, ed.), Yale Univ. Press, New Haven, 606 pp.

. 1976. Pleistocene ecology— some current problems.

Geoscience and Man, 13:1-12.

. 1981. Vegetation east of the Rocky Mountains 18,000

years ago. Quaternary Res., 15:1 13-125.

Yates, T. L., and D. J. Schmidly. 1977. Systematics of Seal-
opus aquaticus (Linnaeus) in Texas and adjacent states. Oc-
cas. Papers Mus., Texas Tech Univ., 45:1-36.

Zakrzewski, R. J. 1967. The primitive vole, Ogmodontomys,
from the late Cenozoic of Kansas and Nebraska. Papers
Michigan Acad. Sci., Arts, and Letters, 52:133-150.

. 1969. The rodents from the Hagerman local fauna,

upper Pliocene of Idaho. Contrib. Mus. Paleo., Univ. Mich-
igan, 23: 1-36.

. 1975a. Pleistocene stratigraphy and paleontology in

western Kansas: the state of the art, 1974. Mus. Paleo., Univ.
Michigan Papers Paleo. (Claude W. Hibbard Mem. Vol. 3),
12:121-128.

. 1 975Z>. The late Pleistocene arvicoline rodent Atopo-

mys. Ann. Carnegie Mus., 45:255-261.

Address (Jones and Choate): Museum of the High Plains, Fort Hays State University, Hays, Kansas 6760 1 .

Present address (Jones): Department of Zoology, University of Florida, Gainesville, Florida 3261 1.

Address (Genoways): Section of Mammals, Carnegie Museum of Natural History, 4400 Forbes Avenue,

Pittsburgh, Pennsylvania 15213.

APPENDIX 1

Pliocene, Pleistocene, and Holocene sites from which Blarina have been reported. Carbon- 14 and geomagnetic dates, when provided, are
in years before present. Samples examined in this study are numbered as in Fig. 2. Computer acronyms are indicated in parentheses.

Site

County

Age

Taxon

Reference

Arkansas

Rancholabrean

46 Peccary Cave (PECCAR)
Irvingtonian

Newton

Wisconsinan (16,700 ±
250)

B. brevicauda

Graham, 1976; Gra-
ham and Semken,
1976

45 Conard Fissure (CONARD)

Florida

Rancholabrean

Newton

Kansan

B. brevicauda

Graham, 1972

Devil’s Den

Levy

Holocene (possibly
7,000 to 8,000)

“ B . brevicauda ”

Martin and Webb,
1974

Melbourne

Brevard

Late Wisconsinan

“ B . brevicauda ”

Webb, 1974

Monkey Jungle Hammock

Dade

Late Pleistocene

“ B . cf. brevicauda"

Martin, 1977; Ober,
1 978, and in litt.

144

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

APPENDIX 1

Continued.

Site

County

Age

Taxon

Reference

79 Vero 2 and 3 (INDIAN)

Indian River

Late Wisconsinan

B. carolinensis

Webb, 1974

Kendrick IA

Marion

Wisconsinan

“ B . brevicauda”

Webb, 1974

75 Waccasassa River IIB and III

Levy

Wisconsinan

B. carolinensis

Webb, 1974

72 Arredondo IB, IIA (ARREDO)

Alachua

Late Sangamonian

B. carolinensis

Webb, 1974

78 Bradenton 51st St.

Manatee

Sangomonian

B. carolinensis

Webb, 1974

71 Haile XIB, XIIIA (HAILEX)

Alachua

Sangamonian

B. carolinensis

Webb, 1974; Martin
and Webb, 1974

73 Reddick IA (REDDIC)

Marion

Sangamonian

B. carolinensis

Webb, 1974

Waccasassa River VIA

Levy

Sangamonian

“ B . brevicauda ”

Webb, 1974

74 Williston III A

Levy

Sangamonian

B. carolinensis

Martin, 1974; Webb,
1974

Irvingtonian

77 Coleman IIA

Sumter

Late Kansan to early 11-
linoian

B. carolinensis

Martin, 1974; Webb,
1974

76 Inglis IA

Citrus

Nebraskan — Aftonian

B. carolinensis

Webb, 1974

Georgia

Rancholabrean

70 Ladds Quarry (LADDSQ)

Bartow

Late Wisconsinan

B. brevicauda and
B. carolinensis

Ray, 1967; Kurten and
Anderson, 1980

Illinois

Rancholabrean

48 Meyer Cave (MEYERC)

Monroe

Early Holocene

B. brevicauda and
B. carolinensis

Parmalee, 1967

Iowa

Rancholabrean

37 Thurman

Fremont

Holocene (980 ± 1 50)

B. brevicauda and
B. b. brevicauda

Jenkins, 1972

35 Pleasant Ridge

Mills

Holocene (1,450)

B. b. brevicauda

Fay, 1980

30 Willard Cave (WILLAR)

Delaware

Holocene (1,255 ± 55;
1,605 ± 65; 3,500 ±
60)

B brevicauda

Eshelman, 1971 and in
litt.

31 Schmitt Cave (SCHMIT)

Dubuque

Holocene (ca. 1,300-
1,900)

B. brevicauda

Eshelman, 1972

34 Garrett Farm

Mills

Holocene (3,400)

B. brevicauda

Fay, 1980

32 Mud Creek

Cedar and
Scott

Holocene (6,220 ±110)

B. brevicauda

Kramer, 1972

29 Cherokee Sewer Site

Cherokee

Holocene (ca. 6,350-
8,400)

B. b. brevicauda

Hoyer, 1980; Semken,
1980

36 Waubonsie

Mills

14,830 + 1,060,
-1,220

B. brevicauda

Rhodes, 1982

33 Craigmile

Mills

23,240 ± 535

B. b.' brevicauda

Rhodes, 1982

Kansas

Rancholabrean

1 3 Robert

Meade

Late Wisconsinan
(1 1,100 ± 390)

B. b. brevicauda

Schultz, 1969

1 1 Jinglebob

Meade

Early Wisconsinan

B carolinensis

Kapp, 1970; Zakrzew-
ski, 1975a

8 Williams

Rice

Illinoian

B. brevicauda

Zakrzewski, 1975a

12 Mt. Scott

Meade

Late Illinoian

B. carolinensis

Hibbard, 1963

Irvingtonian

6 Rezabek

Lincoln

Yarmouthian or Illinoi-
an

B. b. brevicauda

Hibbard and Dalquest,
1973; Zakrzewski,
1975a

1984

JONES ET AL .-BLARINA SYSTEM ATICS

145

APPENDIX 1

Continued.

Site

County

Age

Taxon

Reference

7 Kanopolis

Ellsworth

Yarmouthian

B. brevicauda (and
possibly B. b.
brevicauda )

Hibbard et al., 1978

9 Kentuck

McPherson

Kansan

Blarina sp.

Semken, 1966;
Zakrzewski, 1975a

10 Cudahy

Meade

Late Kansan (600,000)

B. brevicauda

Paulson, 1961; Lindsay
et al., 1975

5 Wathena

Doniphan

Aftonian or Kansan

B. carolinensis

Einsohn, 1971; Van
der Meulen, 1978

Blancan

14 Dixon

Kingman

Nebraskan

B. brevicauda

Hibbard, 1956, 1970;
Skinner et al., 1972

4 White Rock

Republic

Pre-Nebraskan

B. brevicauda

Eshelman, 1975

Kentucky

Rancholabrean

50 Welsh Cave

Woodford

Wisconsinan (12,950 ±
550)

B. b. brevicauda

Guilday, Hamilton,
and McCrady, 1971

Maryland

Irvingtonian

59 Cumberland Cave (CUMBER)

Allegany

Possibly late Kansan or
early Illinoian

B. brevicauda and
B. carolinensis

Van der Meulen, 1978;
Guilday, in litt.

Mississippi

Rancholabrean

49 Catalpa Creek

Lowndes

Pleistocene to Holocene

B. carolinensis

Frazier, in litt.

Missouri

Rancholabrean

39 Brynjulfson Cave No. 2 (BRYNJU)

Boone

Holocene (2,460 ± 230)

B. brevicauda

Parmalee and Oesch,
1972

38 Brynjulfson Cave No. 1

Boone

Late Wisconsinan
(9,440 ± 760)

B. brevicauda

Parmalee and Oesch,
1972

42 Crankshaft Cave (CRANKS)

Jefferson

Late Wisconsinan — Ho-
locene

B. b. brevicauda, B.
brevicauda, B.
hylophaga

Parmalee, Oesch, and
Guilday, 1969

44 Zoo Cave (ZOOCAV)

Taney

Wisconsinan — Holocene

B. brevicauda

Hood and Hawksley,
1975

43 Bat Cave

Pulaski

Late Wisconsinan

B. brevicauda

Hawksley, Reynolds,
and Foley, 1973

40 Boney Spring

Benton

Late Wisconsinan
(13,700 ± 600;
16,580 ± 220)

B. brevicauda

Mehringer, King, and
Lindsay, 1970; Saun-
ders, 1977

41 Trolinger Spring

Elickory

Mid- Wisconsinan
(25,650 ± 700;
32,200 ±

1,900, -1,600)

B. brevicauda

Mehringer, King, and
Lindsay, 1970; Saun-
ders, 1977

Nebraska

Irvingtonian

2 Bartek Brothers

Saunders

Yarmouthian

B. b. brevicauda,
possibly B. brevi-
cauda

Frankel, 1963

3 Angus

Nuckolls

Late Yarmouthian or
Early Illinoian

B. brevicauda

Schultz and Martin,
1970; Kurten and
Anderson, 1980

146

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

APPENDIX 1

Continued.

Site

County

Age

Taxon

Reference

Oklahoma

Rancholabrean

16 Doby Springs

Harper

Ulinoian

B. b. brevicauda

Stephens, 1960

Irvingtonian

1 5 Berends

Beaver

Early Illinoian

B. b. brevicauda
and possibly B.
brevicauda

Hibbard, 1970; Star-
red, 1956

Pennsylvania

Rancholabrean

Meadowcroft Rockshelter

Washington

Holocene

“ B . brevicauda ”

Guilday, in litt.

57 Bootlegger Sink

York

Late Wisconsinan — Ho-
locene (1 1,550 ±

100)

B. b. brevicauda

Guilday, Hamilton,
and McCrady, 1966

Frankstown

Blair

Wisconsinan

“ B . brevicauda"

Peterson, 1926; Kurten
and Anderson, 1980

55 New Paris No. 4 (NEWPAR)

Bedford

Wisconsinan (9,540 to
11,300 ± 1,000)

B. b. brevicauda. B.
brevicauda

Guilday, Martin, and
McCrady, 1964

Irvingtonian

56 Hanover Quarry Fissure

Adams

Yarmouthian

B. carolinensis

Guilday, in litt., 1982

58 Port Kennedy Cave

Montgomery

Aftonian, Kansan, or
Yarmouthian

B. brevicauda

Hibbard, 1958; Guil-
day, 1971; Kurten
and Anderson, 1980

South Dakota
Irvingtonian

1 Java

Walworth

Kansan

B. carolinensis

Martin, 19736

Tennessee

Rancholabrean

First American Bank Site

Davidson

Late Pleistocene — Holo-
cene (9,4 10 to
10,034)

“ B . brevicauda"

Guilday, 1977; Za-
krzewski, personal
communication

54 Riverside Cave (RIVERS)

Sullivan

Late Wisconsinan and
Holocene

B. brevicauda

Guilday, personal com-
munication

Cheek Bend Cave

Maury

Late Wisconsinan and
Holocene

“ B . brevicauda ”

Klippel and Parmalee,
1982

52 Baker Bluff Cave (BAKERB)

Sullivan

Late Wisconsinan
(10,560 ± 220 to
19,100 ± 850)

B. b. brevicauda
and B. brevicau-
da (Ikirtlandi)

Guilday et al., 1978

53 Guy Wilson Cave (GUYWIL)

Sullivan

Late Wisconsinan
(19,700 ± 600)

B. brevicauda

Guilday, 1971

51 Robinson Cave (ROBINS)

Overton

Wisconsinan

B. b. brevicauda

Guilday, Hamilton,
and McCrady, 1969

Texas

Rancholabrean

28 Barton Springs Road

Travis

Holocene (1,015 ± 105)

B. carolinensis

Lundelius, 1967

23 Hall’s Cave

Kerr

Holocene

B. carolinensis

Lundelius, in litt.,
1983

21 Longhorn Cavern

Burnet

Holocene

B. carolinensis

Lundelius, 1967

20 Miller’s Cave

Llano

Holocene (7,200 ± 300)

B. carolinensis and
B. hylophaga

Patton, 19636

24 Klein Cave (KLEINC)

Kerr

Holocene (7,683 ± 643)

B. carolinensis and
B. hylophaga

Roth, 1972

1984

JONES ET AL. — B LARINA SYSTEM ATICS

147

APPENDIX 1

Continued.

Site

County

Age

Taxon

Reference

22 Felton Cave

Sutton

Holocene (7,770 ± 130)

B. carolinensis

Lundelius, 1967; Gra-
ham, 1976

Rex Rodgers Site

Briscoe

Mid to late Wisconsi-
nan

“Marina sp.”

Fullington, 1978;
Schultz, 1978

25 Schulze Cave (SCHULZ)

Edwards

Late Wisconsinan— ear-
ly Holocene ( 3 1 ,000-
8,000)

B. hylophaga

Dalquest et al., 1969

27 Cave Without a Name (CAVEWI)

Kendall

Wisconsinan (10,900 ±
190}

B. carolinensis

Lundelius, 1967

Ben Franklin

Delta

Late Wisconsin (9,550
± 375; 11,135 ±
450)

“ Blarina sp.”

Slaughter and Hoover,
1963

17 Howard Ranch

Hardeman

Wisconsinan

B. brevicauda

Dalquest, 1965

26 Friesenhahn Cave (FRIESE)

Bexar

Wisconsinan

B. carolinensis

Evans, 1961; Lunde-
lius, I960

18 Easley Ranch

Foard

Sangamonian

B. cf. brevicauda

Dalquest, 1962

Irvingtonian

19 Vera

Baylor and
Knox

Late Kansan (ca.
600,000)

B. brevicauda

Hibbard and Dalquest,
1 966; Dalquest,

1977

Unnamed locality

Swisher

Kansan

“ Blarina sp.”

G. E. Schultz, in litt,
1983

Virginia

Rancholabrean

69 Meadowview Cave

Washington

Late Wisconsinan or
early Holocene

B. brevicauda

Guilday and Parmalee,
1972

67 Ripplemead Quarry

Giles

Late Wisconsinan or
early Holocene

B. brevicauda

Guilday, in litt.

64 Natural Chimneys (CHIMNE)

Augusta

Early post- Wisconsinan

B. b. brevicauda

Guilday, 19626

65 Back Creek Cave No. 2 (BACKCR)

Bath

Wisconsinan

B. b. brevicauda
and B. brevicau-
da

Guilday, Parmalee, and
Hamilton, 1977

66 Clark’s Cave (CLARKS)

Bath

Late Wisconsinan

B. b. brevicauda

Guilday, Parmalee, and
Hamilton, 1977

63 Strait Canyon (STRAIT)

Highland

Early Wisconsinan
(29,870, +1,800-
1,400)

B. brevicauda

Guilday, in litt.; Za-
krzewski, personal
communication

Unknown

Early’s Cave

Wythe

Mid-Pleistocene

“ B . cf. brevicauda ”

Guilday, 1962a

68 Jasper Saltpeter Cave (JASPER)

Lee

Unknown

B. brevicauda

Guilday, in litt.

West Virginia

Rancholabrean

6 1 Hoffman School Cave

Pendleton

Late Wisconsinan— ear-
ly Holocene

B. b. brevicauda

Guilday and Hamilton,
1978

Mandy Walters Cave

Pendleton

Late Wisconsinan — ear-
ly Holocene

“B. brevicauda ”

Guilday and Hamilton,
1978

60 Eagle Cave

Pendleton

Wisconsinan

B. brevicauda

Guilday and Hamilton,
1973

Irvingtonian

62 Trout Cave (TROUTC)

Pendleton

Irvingtonian (possibly
late Kansan) — Wis-
consinan

B. brevicauda and
B. b. brevicauda

Zakrzewski, 1975&;
Guilday, in litt.

148

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

APPENDIX 1

Continued.

Site

County

Age

Taxon

Reference

Wisconsin

Rancholabrean

47 Moscow Fissure (MOSCOW)

Iowa

17,050 ± 1,500

B. b. brevicauda

Semken, personal com-
munication, 1982

Ontario

Rancholabrean
Dickson Limestone Quarry

Halton

Minimum age = 215
years

“ B . brevicauda ”

Churcher and Fenton,
1968

Quebec

Rancholabrean

80 Caveme de Sainte-Elzear-de-
Bonaventure (STELZE)

Bonaventure

Holocene (5,1 10 ± 150)

B. brevicauda

LaSalle and Guilday,
1980; Zakrzewski,
personal communi-
cation

ARMADILLOS IN NORTH AMERICAN LATE
PLEISTOCENE CONTEXTS

Walter E. Klippel and Paul W. Parmalee

ABSTRACT

Remains of climatically incompatible mammal species fre-
quently occur together in North American late Pleistocene con-
texts. An apparent incongruence of Neotropical Dasypus remains
with those of temperate and boreal mammal species is evidenced
from the faunal assemblages from Cheek Bend Cave, Tennessee,
and other contemporaneous sites. A size comparison of band
scutes from both late Pleistocene ( Dasypus bellus) and modem
( Dasypus novemcinctus) armadillos indicate that many speci-
mens from the Midsouth are intermediate in size between the
large extinct form and the smaller, modem, nine-banded arma-
dillo.

Elements of Dasypus are recorded from 50 late Pleistocene
deposits in 1 2 states and all locations except one in southwestern
Iowa are located south of the 40th parallel. It is suggested that
the distributional limits of the extinct species or form ofarmadillo
was, like the nine-banded armadillo, limited by aridity in the
west and cold extremes in the north. Utilization of natural den-
ning sites such as limestone caves during periods of extreme cold
may have been one means which enabled D. bellus to survive in
northern latitudes.

INTRODUCTION

Animals represented in late Pleistocene faunal as-
semblages whose present-day counterparts have
markedly different habitat requirements (for ex-
ample, Neotropical versus boreal mammals) have
posed serious problems when attempting credible
interpretations of paleoenvironments in portions of
North America. Such associations are frequently at-
tributed to posthumous mixing of faunal-bearing
deposits. However, Ray (1967:123) noted that “al-
though it is difficult to imagine such seemingly eco-
logically disparate animals as Dasypus bellus and
Martes pennanti as members of the same fauna, the
possibility should not be excluded out of hand. A
growing body of evidence indicates the impropriety
of rigidly interpreting Pleistocene communities in
terms of present ones . . Guilday et al. (1978:61)
observed and reported Wisconsinan “. . . contem-
poraneity of both temperate and boreal species . . .”

as far south as Tennessee and suggested that D. bel-
lus remains in Appalachian late Pleistocene sites
indicated milder winter extremes. Over the past two
decades a number of other authors (for example,
Dalquest et al., 1969; Graham and Semken, 1976;
Hibbard, 1960; Holman, 1980, 1981; Lundelius,
1974; Martin and Gilbert, 1978; Martin and Neu-
ner, 1978; Morgan, 1972; Slaughter, 1961, 1967,
1975) have argued that extralimital southern and
northern species that are climatically incompatible
today could represent late Pleistocene unit faunas
rather than heterochronic assemblages resulting from
posthumous mixing.

Dasypus remains, recovered with those of tem-
perate and boreal mammals from a stratified cave
deposit in Middle Tennessee, have prompted the
following consideration of armadillos from late
Pleistocene contexts in North America.

CHEEK BEND CAVE MAMMALS AND LATE PLEISTOCENE
ENVIRONMENTS IN THE MIDSOUTH

During the summer of 1978 several rockshelters
situated along the Duck River were tested for ar-
chaeological deposits. Excavations in one of these,
Cheek Bend Cave located ca 1 3 km ESE of Colum-
bia, Maury County, Tennessee, indicated that strat-
ified deposits extended to a depth of more than 4.5
m. Three 1 by 2 m test units were excavated along
the east wall near the mouth of the cave. Two major,
distinct fill episodes were delineated. These major
divisions were further subdivided into eight strata

based on physical differences in the deposits (Klip-
pel and Parmalee, 1982 b).

The lower strata (I— III) included faunal species
whose present day ranges are far from the cave site.
The five strata closest to the surface (IV-VIII) con-
tained mostly remains of animals extant in the area
today. Strata V through VIII contained evidence of
prehistoric human occupation. Stone, bone, and ce-
ramic artifacts indicate that the cave was occupied
sporadically for more than 7,000 yr. Radiocarbon

149

150

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

dates on charcoal from Strata V, VI, and VII indi-
cate that these units were aggrading at 7,505 ± 440
(GX 7855), 4,655 ± 75 (UGa 2775), and 2,630 ±
255 (UGa 2859), respectively.

The remains of only four mammalian species
whose present distributions are generally restricted
south of the 40th parallel were recovered from the
cave deposits and all of those occurred in Holocene
strata. Dasypus, whose present distribution is also
restricted south of the 40th parallel today, was rep-
resented in late Pleistocene strata along with re-
mains of two mammals whose present-day distri-
butions are entirely restricted north of the 40th
parallel (that is, Sorex arcticus and Microtus xan-
thognathus). Numerous other boreal species repre-
sented in the late Pleistocene strata only occur south
of the 40th parallel at high elevations in the Smoky
and Rocky mountains (for example, Sorex palustris,
Martes americana, Mustela rixosa, Glaucomys sa-
brinus, Clethrionomys gapperi, Phenacomys inter-
medius, and Napaeozapus insignis). The remains of
several temperate mammal species (for example,
Synaptomys cooperi, Tamias striatus, Geomys bur-
sarius) from late Pleistocene strata also occur with
the boreal yellow-cheeked vole (M. xanthognathus).
These species (temperate versus boreal) are also al-
lopatric under present-day climatic conditions and
conform to the observation by Guilday et al. (1978)
that boreal and temperate species were sympatric
during the late Pleistocene as far south as Tennessee.
However, our primary interest here is in the pres-
ence of armadillos with clearly boreal mammal
species whose remains occur in late Pleistocene con-
text at Creek Bend Cave (Table 1).

Boreal mammal species represented at Cheek Bend
Cave are congruent, in many respects, with the evi-
dence that jack pine-spruce-lir forests dominated
Middle Tennessee during the full glacial (approxi-
mately 22,000 to 16,500 Y.B.P.) and were probably
most similar to forests of southern Manitoba today
(Delcourt, 1979:291). Boreal-like conditions were
followed by Late Wisconsinan mixed coniferous-

THE NINE-BANDED ARMA

Historic records indicate that D. novemcinctus did
not exist in North America beyond the southern-
most portion of Texas (Fig. 1) during the 19th cen-
tury (Baird, 1857; Cope, 1880). Numerous subse-
quent considerations of its range expansion (for
example, Bailey, 1 905; Cleveland, 1970; Fitch et al.,

deciduous forests which lasted from 16,500 to 12,500
years ago. Delcourt (1979:270) has suggested that
at 16,500 B.P. vegetation of Middle Tennessee was
similar to modern conditions in northeast Minne-
sota, and by 1 2,500 conditions were similar to those
found in northeast Wisconsin today. During the ear-
ly Holocene (ca. 12,500-8,000 Y.B.P.) cool tem-
perate mixed mesic forest prevailed throughout this
portion of the Midsouth; vegetation was comparable
in many respects to that of the Allegheny Plateau
region of Ohio and West Virginia at 12,000 and
10,000 years ago, respectively.

While boreal mammals from late Pleistocene strata
at Cheek Bend Cave generally fit expectations based
on Delcourt’s ( 1 979) reconstruction of Wisconsinan
environments for Middle Tennessee (Klippel and
Parmalee, 1982a; Parmalee and Klippel, 1981), ar-
madillos seem out of place in this context unless D.
bel/us either had strikingly different tolerance limits
than the modern nine-banded armadillo ( Dasypus
novemcinctus ) or periods of equable climates al-
lowed the Neotropical Dasypus to range into areas
inhabited by boreal mammals.

“It is unclear whether D. novemcinctus evolved
directly from D. bel/us or whether it ecologically
replaced D. bel/us ” (McNab, 1980:624). Slaughter
(1961:31 1) has suggested that D. bel/us and D. no-
vemcinctus are osteologically identical except for the
greater size of D. bellus. Auffenberg (1957) empha-
sized this point in his description of a nearly com-
plete D. bellus from Mefford Cave in Florida. Kur-
ten and Anderson (1980:131) contend that “even
though definite records of Dasypus novemcinctus do
not appear until Holocene times, there is no doubt
that it is the ecological equivalent of Dasypus bellus,
the larger Pleistocene armadillo.” Others have also
implied or argued for the ecological equivalents of
D. bellus and D. novemcinctus (for example, Lun-
delius, 1967:297; Slaughter, 1961:313; Ray, 1967:
123; Guilday et al., 1978:32; Hood and Hawksley,
1975:22).

XO IN NORTH AMERICA

1952; Kalmbach, 1943: Taber, 1939; Strecker, 1926)
attest to the observation by Buchanan and Talmage
(1954:142) that “one of the most interesting phe-
nomena concerning the armadillo ... is the speed
with which it is extending its North American range.”
In a recent summary of the zoogeography of D.

1984

KLIPPEL AND PARMALEE-NORTH AMERICAN ARMADILLOS

151

Table {.—Mammal remains (numbers of elements) recovered from stratified deposits at Cheek Bend Cave (40MU261). Only specimens
identified to genus and species are included: specimens from equivocal contexts (e.g.. Stratum IIl/IV) are excluded (from Klippe! and
Parmalee 1982 b; Tables 19. 21-27; late Pleistocene strata = /, II. Ill; Holocene strata = IV. V, VI, VII, and VIII). Superscript a = five
elements in an early Holocene stratum possibly intrusive; b = one element in a late Pleistocene stratum probably intrusive; * = mammals

with present day ranges south of the 40th parallel.

Species

Late Pleistocene
strata

Late Pleistocene
and Holocene strata

Holocene strata

Sorex palustris, water shrew

5

Sorex arcticus, arctic shrew

27

Microsorex hoyi, pygmy shrew

12

Condylura cristata, star-nosed mole

567

Spermophilus tridecemlineatus, 1 3-lined ground squirrel

139

Glaucomys sabrinus, northern flying squirrel

53

Clethrionomys gapperi, red-backed vole

423

Phenacomys intermedius, heather vole

9

Microtus xanthognathus, yellow-cheeked vole

1,419

Geomys bursarius, plains pocket gopher

1 ,474a

Napaeozapus insignis, woodland jumping mouse

6

Ursus americanus, black bear

1

Martes americana, pine martin

1

Mustela rixosa, least weasel

58

Mustela vison, mink

14

Dasypus bellus, “beautiful” armadillo

69

Megalonyx jeffersonii, Jefferson's ground sloth

1

Blarina brevicauda, short-tailed shrew

4,777

Sorex longirostris/cinereus, southeastem/masked shrew

58

Sorex fiumeus, smoky shrew

16

Scalopus aquaticus, eastern mole

503

Myotis spp., myotis bats

683

Eptesicus fuscus, big brown bat

3,352

Nycticeius humeralis, evening bat

17

Tamias striatus, eastern chipmunk

950

Tamiasciurus hudsonicus, red squirrel

42

Microtus pennsylvanicus, meadow vole

1,125

Microtus spp., vole

16,352

Ondatra zibethica, muskrat

25

Synaptomys cooperi, southern bog lemming

483

Peromyscus spp., deer/white-footed mouse

722

Zapus hudsonius, meadow jumping mouse

310

Erethizon dorsatum, porcupine

5

Sylvilagus cf. floridanus, rabbit

301

Mustela frenata, long-tailed weasel

59

Odocoileus virginianus, white-tailed deer

78

Cryptotis parva, least shrew

4,71 9b

Pipistrellus subflavus, eastern pipistrel

32

Lasiurus borealis, red bat

2

Plecotus sp., big-eared bat*

41

Marmota monax, woodchuck

10

Sciurus carolinensis, gray squirrel

6

Sciurus niger, fox squirrel

6

Sciurus spp., squirrel

140

Glaucomys voians, southern flying squirrel

1,050

Reithrodontomys cf. humulis, harvest mouse*

3

Oryzomys palustris, marsh rice rat*

12

Sigmodon hispidus, hispid cotton rat*

18

Neotoma floridana, eastern woodrat

186

Castor canadensis, beaver

5

Sylvilagus aquaticus, swamp rabbit*

1

(Jrocyon cinereoargenteus, gray fox

1

Canis sp., canid

1

Procyon lotor, raccoon

4

Mylohyus nasutus, long-nosed peccary

3

152

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 1— Distributions of Dasypus novemcinctus showing the rapid Historic Period spread of the nine-banded armadillo in North
America. Distributions for 1880-1940 taken from Kalmbach (1943); ca. 1980 distribution from Hall (1981).

novemcinctus in the United States, Humphrey (1974:
457) describes the armadillo as having a strong pi-
oneering capability. Invasion from Rio Grande City,
Texas, toward the northeast occurred at roughly 10
km per year between 1880 and 1972. Between 1905
and 1914 expansion to the northeast occurred at an
even greater speed of over 25 km per year (Hum-
phrey, 1974:fig. 4). Pioneering by the nine-banded
armadillo is most successful in river valleys (ripar-
ian environments are preferred habitats for the ar-
madillo, Fitch et al., 1952) which serve as dispersal
conducts where invasion can be especially rapid.

Barriers to the armadillo’s continued expansion
in North America appear to be aridity and cold
temperatures. Arid conditions limit the armadillo’s
primary food source of soil and surface macroin-
vertebrates; these are obtained in mesic areas by
probing under rotted logs and in soil rich in humus
and leaf litter. In xeric areas they are restricted to
valley floors where invertebrates are generally more
abundant in riparian habitats. Humphrey (1974) has
suggested a minimum annual precipitation of 380
mm for armadillo survival.

Heat, per se, apparently is not a deterrent to sur-
vival in that armadillos adjust their above ground
foraging activities according to the seasons and time
of day (Fitch et al., 1952). “High temperatures seem
effective only in causing their activities to be con-
fined largely to the cooler hours of darkness; low
temperatures have the opposite effect, bring them
abroad during the hottest part of the day in search
of food” (Taber, 1945:212, 213). Clark (1951) sug-
gests that armadillo burrows and natural dens (crev-
ices in limestone bluffs) serve as limited sources of
insects during drought periods. Cold temperatures
also affect insect availability in that “ . . . foraging
at air temperatures below freezing would be of little
value especially if the soil were frozen or covered
with snow” (McNab, 1980:621).

Additionally, D. novemcinctus has a high thermal
conductance which “results from the replacement
of a fur coat by protective armor” (McNab, 1980:
616). It can only withstand short periods of near
freezing temperature (that is, adults of ca. 5 kg, 9 to
10 days at near 0°C). Cold avoidance for short pe-
riods of time by this non-hibernating animal is usu-
ally accomplished by burrowing into the ground or

1984

KLIPPEL AND PARMALEE-NORTH AMERICAN ARMADILLOS

153

seeking shelter in caves and crevices in limestone
bluffs.

McNab (1980) has demonstrated that armadillos
with greater body mass have a longer survival time
during cold weather and suggested the 20 to 30 kg
D. bellus that occupied portions of North America
during the Pleistocene might be expected to have
had an increased “survival time” of 1 2 days at near
0°C. An even longer period of survival would be
anticipated at 10°C. However, burrowing as a means

of protection from cold by such large animals would
be nearly impossible; natural shelters (caves) might
have served as the only refuge under such condi-
tions. Taber’s (1945) observation of D. novemcinc-
tus from the Edwards Plateau in Texas indicates that
where caves and crevices are available and abundant
in limestone cliffs few burrows are excavated, thus
indicating a preference for such features even when
the alternative of burrowing is possible for the small
nine-banded armadillo.

PLEISTOCENE DASYPUS IN NORTH AMERICA

Dasypus remains in faunal assemblages are es-
pecially apparent because of the large number of
diagnostic elements that are potentially recoverable.
D. novemcinctus, for example, has over 3,300 scutes
per animal (Davis, 1969), including between 500
and 600 band scutes (UT #4489). These are in ad-
diton to elements such as teeth, mandibles, and limb
bones that are usually employed in the identification
of most mammals.

Remains of the “highly visible” Dasypus have
been recovered from Pleistocene sites throughout
much of southern North America from Florida and
the Carolinas to New Mexico (Fig. 2, Table 2). With
one known exception in southwestern Iowa (Rhodes,
1982), all Pleistocene Dasypus remains have been
found south of the 40th parallel. However, a rela-
tively large number have been recovered from out-
side the present range of the nine-banded armadillo
(Fig. 1) as delineated by Hall (1981).

Deposits in North America from which Dasypus
remains have been recovered are predominantly
Rancholabrean in age (Slaughter, 1961; Kurten and
Anderson, 1980) or from equivocal contexts. How-
ever, Dasypus from Florida have been assigned to
both the Blancan (Haile XVA and Santa Fe R.IB)
and Irvingtonian (Inglis I A, Santa Fe R.IIA, and
Coleman IIA) Land Mammal Ages (Webb, 1974).

Some evidence suggests a general increase in size
of North American Dasypus from Blancan through
the Rancholabrean (Kurten and Anderson, 1980).
D. bellus moveable dermal plates from Irvingtonian
contexts are smaller than many of those from Ran-
cholabrean deposits in Florida (Martin, 1974a:43).
However, Guilday and McCrady (1966) reported
specimens from a site in West Virginia and one from
Tennessee that appear intermediate in size between
Rancholabrean D. bellus from Florida and the nine-
banded armadillo that inhabits Florida today. They

noted that “perhaps the relative harshness of the
environment is reflected in the smaller size of the
Tennessee and West Virginia specimens as com-
pared to those from Missouri and Florida Pleisto-
cene sites” (Guilday and McCrady, 1966:183).

Since the mid-1960s, a number of other Dasypus
remains have been recovered from the Midsouth.
We have chosen to assess the Guilday and McCrady
(1966) observation concerning size on a number of
these specimens by measuring band scutes. We have
somewhat modified a procedure described by Mar-
tin (1974a:42) and have recorded maximum thick-
ness at the slight dorsal ridge where band scutes
overlap; maximum scute width was taken at this
point of maximum thickness. We excluded lateral
scutes (the first three from each side of each band
row) in that they introduce significant size variation
when included with interior bands. A number of
these lateral scutes in a small sample could unnec-
essarily make thickness measurements for an assem-
blage appear erroneously small. We have also ex-
cluded scutes (most from the ninth band) that are
not straight or only slightly convex at the distal end.
These modifications to Martin’s (1974a) method
tend to result in reduced sample sizes. However,
they also exclude considerable measurement vari-
ation in animals we compared (that is, lateral versus
interior band scutes of a nine-banded armadillo, UT
#4489, as well as those of the large sample of D.
bellus from Crankshaft Cave). Since few samples
from any given site are large enough to treat with
more than basic statistics (that is, ca. < 10), we have
excluded obvious sources of variation at the expense
of having only slightly reduced samples.

Results of scute measurements of both modern
and Pleistocene Dasypus are presented in Table 3
and Fig. 3. Two Rancholabrean sites (Crankshaft
Cave in Missouri and Melbourne in Florida) pro-

154

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 2. — Locations of Pleistocene sites that contain remains of Dasypus bel/us. Locality numbers correspond to numbers and information
provided in Table 2.

duced relatively large samples of measurable band
scutes; one standard deviation (n — 1 ) for each sam-
ple is illustrated in Fig. 3. Random samples of 51
band scutes from a modern nine-banded armadillo
of unknown weight (UT #4489) and another of 56

band scutes from a 4.3 kg armadillo from Missis-
sippi (UT #6479) are also illustrated (Hall, 1981,
suggests that modem armadillos in Texas may reach
7.7 kg but usually range around 5 kg).

There is no overlap in size (thickness or width)

Table 2 . — Pleistocene sites in North American that contain Dasypus remains. Numbers at the left of this table correspond to locations

in Fig. 2.

Location
(see Fig. 2)

State/site

County

Citation

Florida

1

Vero Beach

Indian River

Weigel, 1962

2

Melbourne

Brevard

C. E. Ray: unpublished

3

Sebastion Canal

Brevard

Webb, 1974

4

Mefford Cave

Marion

Auffenberg, 1957

5

Reddick

Marion

Gut and Ray, 1963

6

Haile XVA

Alachua

Webb, 1974

7

Kendrick IA

Marion

Webb, 1974

8

Haile XIVA

Alachua

Martin, 19746

9

Arredondo IB

Alachua

Webb, 1974

10

Coleman 1 1 A

Sumter

Webb, 1974

1 1

Inglis IA

Citrus

Webb, 1974

12

Seminole Field

Pinellas

Webb, 1974

13

Saber-Tooth Cave

Citrus

Simpson, 1928

1984

KLIPPEL AND PARMALEE-NORTH AMERICAN ARMADILLOS

155

Table 2. — Continued.

Location
(see Fig. 2)

State/site

County

Citation

14

Wilhlacoochee River VIIA

Citrus

Webb, 1974

15

Waccasassa River HR & III

Levy

Webb, 1974

16

Williston III A

Levy

Webb, 1974

17

Sante Fe River IB
South Carolina

Gilchrist

Webb, 1974

18

Edisto Island
North Carolina

Charleston

A. Sanders: unpublished

19

Harriett’s Pit
Georgia

Jones

C. E. Ray: unpublished

20

Ladds Quarry

Bartow

Ray, 1967

21

Kingston Saltpeter Cave

Bartow

Sneed, 1983

22

Yarborough Cave
Tennessee

Bartow

Sneed, 1983

23

Lookout Mountain Cave

Hamilton

R. Wilson: unpublished

24

Cumberland Caverns

Warren

Smyre, 1978

25

Cheek Bend Cave

Maury

Klippel and Parmalee, 19826

26

Robinson Cave

Overton

Guilday and McCrady, 1966

27

Finger Quarry

Blount

Robison, 1981

28

Baker Bluff Cave
Virginia

Sullivan

Guilday et al., 1978

29

Meadowview Cave
West Virginia

Washington

A. Guilday: unpublished

30

Organ-Hedricks Cave
Indiana

Greenbrier

Guilday and McCrady, 1 966

31

Anderson Pit Cave

Monroe

Richards, 1980

32

Prairie Creek
Iowa

Daviess

Tomak, 1982

33

Craigmile

Missouri

Mills

Rhodes, 1982

34

Brynjulfson Caves

Boone

Parmalee and Oesch, 1972

35

Cherokee Cave

St. Louis

Simpson, 1949

36

Crankshaft Cave

Jefferson

Parmalee et al., 1969

37

Zoo Cave
Arkansas

Taney

Hood and Hawksley, 1975

38

Peccary Cave

Newton

Davis, 1969

39

10 Mile Rock Site
Oklahoma

Washington

Sealander, 1979

40

Bar M

Texas

Harper

Slaughter, 1967

41

Slaton Quarry

Lubbock

Dalquest, 1967

42

Easley Ranch

Ford

Slaughter, 1967

43

Ben Franklin

Fannin/Delta

Slaughter and Hoover, 1963

44

Clear Creek

Denton

Slaughter and Ritchie, 1963

45

Hill-Shuler

Dallas/Denton

Slaughter et al., 1 962

46

Moore Pit

Dallas

Slaughter, 1966

47

Miller’s Cave

Llano

Patton, 1963

48

Cave Without a Name

Kendall

Lundelius, 1967

49

Kincaid Shelter
New Mexico

Uvalde

Lundelius, 1967

50

Brown Sand Wedge

Roosevelt

Slaughter, 1964

156

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

0-1 i — r 1 1 1 1 1 r

I 2 3 4 5 6 7 8

MAXIMUM THICKNESS (mm)

Fig. 3. — Mean band scute size of D. novemcinctus and D. bellus : 1) modern D. novemcinctus, UT #4489, Florida; 2) modern D.
novemcinctus, UT #6479, Mississippi; 3) Rancholabrean D. bellus. Robinson Cave, Tennessee; 4) Rancholabrean D. bellus , Cheek Bend
Cave, Tennessee; 5) Rancholabrean D. bellus. Kingston Salt Peter Cave, Georgia; 6) Rancholabrean D. bellus, Organ-Hedrick Cave,
West Virginia; 7) Rancholabrean D. bellus. Lookout Mt., Tennessee; 8) Ranchlobrean D. bellus, Baker Bluff Cave, Tennessee; 9)
Rancholabrean D. bellus. Meadow View Cave, Virginia; 10) Rancholabrean D. bellus, Anderson Pit Cave, Indiana; 1 1) Rancholabrean
D. bellus, Reddick, Florida; 12) Rancholabrean D. bellus, Ladds Quarry, Georgia; 13) Rancholabrean D. bellus, Melbourne, Florida;
14) Rancholabrean D. bellus, Crankshaft Cave, Missouri; 1 5) Blancan D. bellus, Haile XVA, Florida; 16) Irvingtonian D. bellus, Coleman
IIA, Florida. One standard deviation (n — 1) and observed ranges are illustrated for samples with over 50 scutes (that is, 1,2, 13, and
14).

between D. bellus (Crankshaft Cave and Melbourne)
and the nine-banded armadillo of unknown weight
from Florida. Neither do scutes from within the first
standard deviation of the modern 4.3 kg specimen
from Mississippi and D. bellus from Crankshaft Cave
have overlapping ranges. However, their total ob-

served ranges do overlap in maximum width. When
means (or single values) of several smaller samples
from the Midsouth are plotted (including the spec-
imens from Robinson Cave and Organ-Hedricks
Cave reported by Guilday and McCrady, 1966), they
fall intermediate between D. novemcinctus and

1984

KLIPPEL AND PARMALEE- NORTH AMERICAN ARMADILLOS

157

Table 3. — Dasypus band scute measurements (mm) from modern and Pleistocene contexts in eastern North America. Superscript “ =
maximum dimensions for any or all band scutes; h = maximum dimensions of inner band scutes at slight dorsal ridge where scutes

overlap; c — n — 1 standard deviation; d = observed range.

Maximum

width

Maximum

thickness

Species/locality

N

Mean

SD‘

ORd

N

Mean

SDC

ORd

Recent

D. novemcinctus (FL)
Martin (1974:43)

235“

5.4

3.5-8. 1

235“

1.6

0.9-2. 4

D. novemcinctus (FL)

UT Zooarchaeology (#4489)

5 lb

5.7

0.87

3. 9-7.7

51b

1.6

0.22

1. 2-2.0

D. novemcinctus (MS)

UT Zooarchaeology (#6479)

56b

6.9

0.98

4. 3-9.0

56b

2.3

0.21

1. 8-2.6

Rancholabrean
D. bellus (FL)
Martin (1974:43)

163“

11.4

6.6-16.5

163“

3.9

1.9-6. 1

D. bellus (FL)
Reddick IA

4b

11.7

1.97

8.8-13.0

4b

6.2

0.63

5.6-7. 1

D. bellus (FL)
Melbourne

55b

12.8

1.53

9.0-16.8

55b

5.7

0.54

3.9-6. 6

D. bellus (GA)

Ladds Quarry

2b

13.2

12.5-13.9

2b

6.6

6.6

D. bellus (GA)

Kingston Salt Peter Cave

6b

9.1

2.66

6.5-12.5

6b

3.5

0.48

2. 8-3. 9

D. bellus (IN)
Anderson Pit

3b

12.4

0.93

11.3-13.0

4b

5.4

0.66

4. 5-6.0

D. bellus (MO)
Crankshaft Cave

22 1 b

12.4

2.27

7.9-18.4

22 1 b

5.2

0.86

3. 5-7. 3

D. bellus (TN)
Cheek Bend Cave

10b

9.9

1.33

7.6-12.0

2b

2.9

2. 5-3. 2

D. bellus (TN)
Robinson Cave

5b

6.8

0.61

5.9-7. 5

5b

3.0

0.54

2.0-3. 3

D. bellus (TN)
Lookout Mt.

lb

9.1

lb

4.1

D. bellus (TN)
Baker Bluff Cave

!b

8.5

lb

4.5

D. bellus (VA)
Meadow View Cave

lb

11.0

lb

5.1

D. bellus (WV)
Organ-Hendricks

lb

—

—

8.7

lb

—

3.4

Irvingtonian
D. bellus (FL)

Coleman IIA; Martin (1974a:43)

21“

8.2

6.2-10.3

21“

4.0

2. 7-6. 8

Blancan
D. bellus (FL)

Haile XVA; Martin ( 1 974<a:43)

10“

10.5

8.2-12.4

10“

3.5

2.9-4. 2

“typical” large D. bellus which is considered to have
been at least twice as large as the nine-banded ar-
madillo. It is also interesting to note that the means
of measured scutes (measured with slightly more
variance included in samples) of specimens from
the earlier Blancan (Haile XVA) and Irvingtonian
(Coleman IIA sites), reportedly smaller than Ran-

cholebrean D. bellus (Martin, 1974a), fall in this
intermediate area as well.

There is considerable variation in the size of adult
nine-banded armadillos (Russell, 1953; Wetzel and
Mondolfi, 1979). According to Hall (1981) arma-
dillos in Texas attain weights nearly twice that of
the 4.3 kg specimen from Mississippi reported here

158

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

(Fig. 3:2, UT #6479). One cannot help but surmise
that scutes from 7 to 8 kg armadillos might be as
large, if not larger, than those falling between D.
novemcinctus (from Florida and Mississippi) and
the large Pleistocene Dasypus (for example, from
Crankshaft Cave and Melbourne) represented in Fig.
3. In fact, if we were to assume that twice the weight
equated with the twice the scute size, large D. no-
vemcinctus mean scute size would clearly fall within
the range of the first standard deviation of these

from Crankshaft Cave D. bellus. Our intent is not
to attempt to demonstrate that D. bellus from
Crankshaft Cave weighed only slightly over 8 kg,
but rather to point up the considerable variation in
nine-banded armadillo size. Judging from evidence
presented here (Table 3), Rancholabrean Dasypus
must have also had a considerable range in size
which, in some instances, approached the upper size
limits of D. novemcinctus.

SUMMARY

As is the case at Cheek Bend Cave in Middle
Tennessee, Dasypus remains are frequently found
in late Pleistocene contexts with those of other
mammals that are presently restricted to boreal hab-
itats. On the surface, this association seems unlikely,
but when habits, tolerance limits, and the ecology
of the nine-banded armadillo (modern analogue) are
taken into account along with evidence for more
equable climates during the late Pleistocene, the as-
sociation becomes considerably more credible.

Dasypus has been in North America (at least as
far north as central Florida) since Blancan times.
Armadillos are notorious at pioneering and can
spread rapidly under suitable environmental con-
ditions. If modern records for D. novemcinctus can
be used as an indication of the potential rate of
dispersal (that is, ca. 25 km/yr) for the larger Pleis-
tocene form, this mammal could have spread from
the Florida Peninsula to Middle Tennessee in con-
siderably less than a half century. From the Gulf
Coast (for example, Pensacola, Florida, or Mobile,
Alabama) it could have reached Middle Tennessee
in less than three decades. Equable climatic condi-
tions characterized by cool summers and winters
without prolonged periods below freezing are not
entirely unthinkable over the many thousands of
years represented by the Rancholabrean. Such peri-
odic conditions would undoubtedly have been con-
ducive to large populations of insects and other in-
vertebrates upon which D. bellus probably fed.

The stratigraphic position of faunal remains in
Cheek Bend Cave suggests that the armadillo in-
habited Middle Tennessee at least during the late
Wisconsinan. Palynological evidence suggests that
mixed coniferous-deciduous forests were replacing
the jack pine-spruce-fir forests of the full Wiscon-
sinan in Middle Tennessee during this period.

It is probable that during the late glacial (Wis-
consinan) armadillos were able to rapidly extend
their range northward during even relatively short
intervals with above freezing temperatures. The
preference of this mammal for natural den sites (as
suggested by numerous occurrences of its remains)
could account in part for its prevalence in cave con-
texts. It is also probable that armadillos no larger
than the individual represented in Cheek Bend Cave,
as well as those from other sites in the Midsouth
(Fig. 3), could have burrowed for protection for short
periods of cold temperatures as do modern nine-
banded armadillos today. These smaller animals
could represent a late Wisconsinan adaptation in
the Midsouth that was not conducive to the survival
of larger individuals of D. bellus if the latter were
not able to find sufficient protection in caves or pro-
vide their own protection by burrowing. If this is
the case then larger armadillos whose remains have
been recovered from more northern latitudes may
represent Dasypus expansions during more mod-
erate interglacial periods (for example, Sangamoni-
an) of the Late Pleistocene.

ACKNOWLEDGMENTS

To the following individuals we express our appreciation for
the loan of Dasypus bellus specimens and for information relative
to this material housed or curated under their care: Russell W.
Graham, Alice M. Guilday, Oscar Hawksley, Larry D. Martin,
Clayton E. Ray, Ronald L. Richards, Albert E. Sanders, Joel M.

Sneed, and Ronald C. Wilson. We are especially grateful to Terry
Faulkner for the preparation of the figures. Excavations at Cheek
Bend Cave (40MU26 1 ) and identification of the faunal materials
were undertaken under the auspices of the Tennessee Valley Au-
thority (contract No. TVA TV-49244 A and TVA TV-530 13A).

1984

KLIPPEL AND PARMALEE-NORTH AMERICAN ARMADILLOS

159

LITERATURE CITED

Auffenberg, W. 1957. A note on an unusually complete spec-
imen of Dasypus bellus (Simpson) from Florida. Quar. J.
Florida Acad. Sci., 20:233-237.

Bailey, V. 1905. Biological survey of Texas. North Amer. Fau-
na, 25:1-222.

Baird, S. F. 1857. Mammals of North America. Pacific R.R.
Rept., 8: 1-757.

Buchanan, G. D„ and R. V. Talmage. 1954. The geographical
distribution of the armadillo in the United States. Texas
Acad. J. Sci., 6:142-150.

Clark, W. K. 1951. Ecological life history of the armadillo in
the eastern Edwards Plateau region. Amer. Midland Nat.,
46:337-358.

Cleveland, A. G. 1970. The current geographic distribution
of the armadillo in the United States. Texas Acad. Sci., 22:
90-93.

Cope, E. D. 1880. Zoological position of Texas. Bull. U.S. Nat.
Mus., 17:1-51.

Dalquest, W. W. 1967. Mammals of the Pleistocene Slaton
Local Fauna of Texas. Southwestern Nat., 12:1-30.

Dalquest, W. W., E. Roth, and F. Judd. 1969. The mammal
fauna of Schulze Cave, Edwards County, Texas. Bull. Florida
State Mus., Biol. Ser., 13:205-276.

Davis, L. C. 1969. The biostratigraphy of Peccary Cave, New-
ton County, Arkansas. Proc. Arkansas Acad. Sci. 23:192.

Delcourt, H. R. 1979. Late Quaternary vegetation history of
the eastern Highland Rim and adjacent Cumberland Plateau
of Tennessee. Ecol. Monogr., 49:255-280.

Fitch, H. S., P. Goodrum, and C. Newman. 1952. The ar-
madillo in the southeastern United States. J. Mamm., 33:
21-37.

Graham, R. W., and H. A. Semken. 1976. Paleoecological
significance of the short-tailed shrew ( Blarina ), with a sys-
tematic discussion of Blarina ozarkensis. J. Mamm., 57:
433_449.

Guilday, J. E., H. W. Hamilton, E. Anderson, and P. W.
Parmalee. 1978. The Baker Bluff Cave deposit, Tennessee,
and the late Pleistocene faunal gradient. Bull. Carnegie Mus.
Nat. Hist., 1 1:1-67.

Guilday, J. E., and A. D. McCrady. 1 966. Armadillo remains
from Tennessee and West Virginia caves. Bull. Nat. Speleo.
Soc., 28:183-184.

Gut, H. J., and C. E. Ray. 1963. The Pleistocene vertebrate
fauna of Reddick, Florida. Quar. J. Florida Acad. Sci., 26:
315-328.

Hall, E. R. 1981. The mammals of North America. John Wiley
and Sons, New York. 2 vols., 1,181 pp.

Hibbard, C. W. 1960. Pliocene and Pleistocene climates in
North America. Annual Rept. Michigan Acad. Sci., Arts,
and Letters, 62:5-30.

Holman, J. A. 1980. Paleoclimatic implications of Pleistocene
herpetofaunas of eastern and central North America. Trans.
Nebraska Acad. Sci., 8:131-140.

. 1981. A review of North American Pleistocene snakes.

Michigan State Univ. Paleontol. Ser., 1:263-306.

Hood, C. H., and O. Hawksley. 1975. A Pleistocene fauna
from Zoo Cave, Taney County, Missouri. Missouri Speleo.,
15:1-42.

Humphrey, S. 1974. Zoogeography of the nine-banded arma-

dillo (Dasypus novemcinctus) in the United States. Bio-
Science, 24:457-462.

Kalmbach, E. R. 1943. The armadillo: its relation to agriculture
and game. Game, Fish, and Oyster Commission, Austin,
Texas, 61 pp.

Klippel, W. E., and P. W. Parmalee. 1982a. Diachronic vari-
ation in insectivores from Cheek Bend Cave and environ-
mental change in the Midsouth. Paleobiology, 8:447-458.

. 19826. The paleontology of Cheek Bend Cave: Phase

II Report. Report submitted to the Tennessee Valley Au-
thority, Norris, Tennessee, xiii + 249 pp.

Kurten, B., and E. Anderson. 1980. Pleistocene mammals of
North America. Columbia Univ. Press, 442 pp.

Lundelius, E. L., Jr. 1967. Late Pleistocene and Holocene
faunal history of central Texas. Pp. 287-319, in Pleistocene
extinctions: the search for a cause (P. S. Martin and H. E.
Wright, eds.), Yale Univ. Press, New Haven, x + 453 pp.

. 1974. The last fifteen thousand years of faunal change

in North America. Pp. 141-160, in History and prehistory
of the Lubbock Lake Site, The Mus. J., 15:1-160.

Martin, L. D., and B. M. Gilbert. 1978. Excavations at Nat-
ural Trap Cave. Trans. Nebraska Acad. Sci., 6:107-1 16.

Martin, L. D., and A. M. Neuner. 1978. The end of the
Pleistocene in North America. Trans. Nebraska Acad. Sci.,
6:1 17-126.

Martin, R. A. 1974a. Fossil mammals from the Coleman II A
fauna, Sumter County. Pp. 35-99, in Pleistocene mammals
of Florida (S. D. Webb, ed.), Univ. Florida Press, Gaines-
ville, x + 270 pp.

. 19746. Fossil vertebrates from the Haile XI VA Fauna,

Clachua County. Pp. 100-1 13, in Pleistocene mammals of
Florida (S. D. Webb, ed.), Univ. Florida Press, Gainesville,
X + 270 pp.

McNab, B. K. 1980. Energetics and the limits to a temperate
distribution in armadillos. J. Mamm., 61:606-627.

Morgan, J. M. 1972. An analysis of penglacial climatic in-
dicators of late glacial time in North America. Unpublished
Ph.D. dissert., Univ. Wisconsin, Madison, 160 pp.

Parmalee, P. W., and W. E. Klippel. 1981. A late Pleistocene
record of the heather vole ( Phenacomys intermedins ), in the
Nashville Basin, Tennessee. J. Tennessee Acad. Sci., 56: 1 27—
129.

Parmalee, P. W., and R. D. Oesch. 1972. Pleistocene and
recent faunas from the Brynjulson Caves, Missouri. Illinois
State. Mus. Repts. Investigations, 25:1-52.

Parmalee, P. W., R. D. Oesch, and J. E. Guilday. 1969. Pleis-
tocene and Recent vertebrate faunas from Crankshaft Cave,
Missouri. Illinois State Mus. Repts. Investigations, 14:1-37.

Patton, T. H. 1963. Fossil vertebrates from Miller’s Cave,
Llano County, Texas. Bull. Texas Mem. Mus., 7:41.

Ray, C. E. 1967. Pleistocene mammals from Ladds. Bartow
County, Georgia. Bull. Georgia Acad. Sci., 25:120-150.

Rhodes, R. S. 1982. Mammalian paleoecology of the Farm-
dahan Craigmile and the Woodfordian Waubonsie local fau-
nas, southwestern Iowa. Unpublished Ph.D. dissert., Univ.
Iowa, Iowa City, 132 pp.

Richards, R. L. 1980. Rice rat (Oryzomys cf. palustris) remains
from southern Indiana caves. Proc. Indiana Acad. Sci., 89:
425-431.

160

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Robison, N. D. 1981. A description of the Pleistocene faunal
remains recovered from Finger Quarry, Blount County, Ten-
nessee. J. Tennessee Acad. Sci., 56:68-71.

Russell, R. J. 1953. Description of a new armadillo (Dasypus
novemcinctus) from Mexico with remarks on geographic
variation of the species. Proc. Biol. Soc. Washington, 66:21-
26.

Sealander, J. A. 1979. A guide to Arkansas mammals. River
Road Press, Conway, Arkansas, pp. 313.

Simpson, G. G. 1928. Pleistocene mammals from a cave in
Citrus County, Florida. Amer. Mus. Novitates, 328:1-16.

. 1949. A fossil deposit in a cave in St. Louis. Amer.

Mus. Novitates, 1408:1-46.

Slaughter, B. H. 1961. The significance of Dasypus bellus
(Simpson) in Pleistocene local faunas. Texas J. Sci. 13:31 1 —
315.

. 1964. An ecological interpretation of the Brown Sand

Wedge local fauna, Blackwater Draw, New Mexico, and a
hypothesis concerning late Pleistocene extinction. Paleoecol.
Llano Estracado, 2:1-37.

. 1966. The Moore Pit local fauna. Pleistocene of Texas.

J. Paleont., 40:70-91.

. 1967. Animal ranges as a clue to late Pleistocene ex-
tinctions. Pp. 155-167, in Pleistocene extinctions: the search
fora cause (P. S. Martin and H. E. Wright, eds.), Yale Univ.
Press, New Haven, 453 pp.

. 1975. Ecological interpretation of the Brown Sand Wedge

local fauna. Pp. 179-192, in Late Pleistocene environments
of the southern High Plains, Publ. Burgwin Res. Center,
Rancho de Taos, New Mexico, 9.

Slaughter, B. H., and B. R. Hoover. 1963. Sulphur River
formation and the Pleistocene mammals of the Ben Franklin

local fauna. Southern Methodist Univ., Grad. Res. Center
J., 31:132-148.

Slaughter, B. H., W. W. Crook, Jr., R. K. Harris, D. C.
Allen, and M. Seifert. 1962. The Hill-Shuler local faunas
of the Upper Trinity River, Dallas and Denton Counties,
Texas. Rept. Investigations Univ. Texas Bur. Econ. Geol.,
48:1-75.

Slaughter, B. H., and R. Ritchie. 1963. Pleistocene mammals
of the Clear Creek local fauna, Denton County, Texas. South-
ern Methodist Univ., Grad. Res. Center J., 31:1 17-131.

Smyre, J. 1978. Recent bone finds at Cumberland Caverns.
Speleonews (NSS Nashville and Chattanooga Grottoes), 22( 1 ):
3-7.

Sneed, J. M. 1 983. Another Pleistocene site in Georgia. Month-
ly Breakdown (NSS Grotto No. 285), 3:9-12.

Strecker, J. K. 1926. Extension of the nine-banded armadillo.
J. Mamm., 7:206-210.

Taber, F. W. 1939. Extension of the range of the armadillo. J.
Mamm., 20:489-493.

. 1945. Contribution on the life history and ecology of

the nine-banded armadillo., J. Mamm., 26:21 1-226.

Tomak, C. H. 1982. Dasypus bellus and other extinct mammals
from the Prairie Creek Site. J. Mamm., 63:158-160.

Webb, S. D. 1974. Chronology of Florida Pleistocene mammals.
Pp. 5-3 1 , in Pleistocene mammals of Florida (S. D. Webb,
ed.), Univ. Florida Press, Gainesville, x + 270 pp.

Weigel, R. D. 1962. Fossil vertebrates of Vero, Florida. Florida
Geol. Surv., Spec. Publ., 10:1-59.

Wetzel, R. M., and Mondolfi. 1979. The subgenera and species
of long-nosed armadillos, genus Dasypus L. Pp. 43-63, in
Vertebrate ecology in the northern Neotropics (J. F. Eisen-
berg, ed.), Smithsonian Institution Press, Washington, D.C.

Address (Klippel): Department of Anthropology, The University of Tennessee, Knoxville, Tennessee 37916.

Address (Parmalee): Frank H. McClung Museum, The University of Tennessee, Knoxville, Tennessee

37916.

A PLEISTOCENE OCCURRENCE OF GEOMYS
(RODENTIA: GEOMYIDAE) IN
WEST VIRGINIA

Frederick Grady

ABSTRACT

Remains of the pocket gopher, Geomys sp., have been re-
covered from the lower levels of an excavation in New Trout
Cave, Pendleton County, West Virginia. This is the first record
of Geomys from West Virginia and is some 480 km east of the
present range of Geomys bursarius and 500 km north of the
present range of Geomys pinetis. The New Trout Cave Geomys

elements occur 60 or more cm below a level that has been dated
by C14 at 29,400 ± 1,700 B.P. The associated fauna includes
Ochotona sp. but lacks several extinct small mammal taxa char-
acteristics of the late Irvingtonian faunas of Trout and Cumber-
land Caves.

INTRODUCTION

In 1979 a rich Pleistocene bone deposit was dis-
covered in New Trout Cave, Pendleton County, West
Virginia. The cave is located at 30°36'10"N Lati-
tude, 79°22'08"W Longitude and is at 548.6 m al-
titude (Grady and Garton, 1981). The deposit, con-
sisting of bone-rich matrix, was collected in 30 cm
intervals to a depth of 220 cm and wet-screened

through 5 mm and 1.5 mm mesh. A fauna including
fish, amphibians, reptiles, birds, and at least 60
species of mammals was recovered. The specimens
have been deposited in the collections of the De-
partment of Paleobiology of the National Museum
of Natural History.

DISCUSSION

The upper 90 cm of the New Trout Cave deposit
contain a fairly typical late Pleistocene fauna in-
cluding northern species such as Dicrostonyx hud-
sonius, Microtus xanthognathus, and Sorex arcticus
(Grady and Garton, 1981). Also present are two
species now found in the Midwest, Spermophilus
tridecemlineatus and Taxidea taxus.

Among the more interesting remains in the three
levels from 120-21 cm in the New Trout Deposit,
were 1 1 specimens of the pocket gopher, Geomys
sp., numbered USNM 336244-336254. The spec-
imens include six parts of upper incisors, one partial
lower incisor, a distal part of a tibia, and an ungual
phalanx. Indentification to species level was not pos-
sible due to the fragmentary nature of the material.
The New Trout Cave specimens probably are either
the plains pocket gopher, Geomys bursarius , or the
southeastern pocket gopher, Geomys pinetis. These
two species are differentiated osteologically by the
shape of the nasal bones, which are highly variable
(Guilday et al., 1971) and not represented in the
New Trout collection. The upper incisor parts ranged
in width from 2.2 to 2.8 mm, within the ranges of
incisor widths of both G. bursarius and G. pinetis.
G. S. Morgan (personal communication) suggested

that the New Trout tibia was closer to Geomys bur-
sarius in size though Parmalee and Klippel (1981)
demonstrated considerable overlap in the postcra-
nial elements of the two species.

There have been a number of fossil records of
Geomys in the eastern United States outside of the
current ranges of Geomys bursarius and Geomys
pinetis (Fig. 1 ) summarized by Parmalee and Klippel
(1981). The localities in Illinois, Indiana, Kentucky,
and Tennessee are all, except possibly Harrodsburg,
Indiana, late Pleistocene or Holocene in age (Guil-
day et ah, 1971; Guilday, 1977; Parmalee and Klip-
pel, 1981). The New Trout Cave locality is some
480 km east of the current range of Geomys bur-
sarius and 500 km north of Geomys pinetis (Fig. 1)
(Hall, 1981). The Geomys specimens from New
Trout Cave are the first known from West Virginia.
They occur in levels 60 or more centimeters below
a C14 date of 29,400 ± 1,700 BP. The associated
fauna in the levels of New Trout Cave containing
Geomys lacks several of the northern species present
in the upper levels and includes Ochotona sp., pre-
viously associated with late Irvingtonian faunas of
the Central Appalachians (Guilday, 1979). A num-
ber of extinct small mammal taxa present in the late

161

162

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 1. -Geographic ranges of Geomys bursarius and G. pinetis with fossil localities outside the present range (modified from Parmalee
and KJippel, 1981).

Irvingtonian faunas of Trout and Cumberland Caves
are not present in New Trout Cave. It seems likely
that the lower levels of New Trout Cave are early
to middle Rancholabrean in age.

The climatic implications of Geomys in West Vir-
ginia are unclear because the material is not specif-
ically identifiable. Historically the vegetation in West
Virginia consisted of mixed coniferous and decid-
uous forest (McKeever, manuscript). Geomys bur-
sarius inhabits prairie and semi-prairie from Can-
ada to Texas, whereas G. pinetis inhabits open pine
woods with sandy soil in the Southeast (Guilday et
al., 1971). Based on specimens from Kentucky,
Guilday et al. (1971) suggested there was a contin-
uous distribution of Geomys from the Midwest to
Florida. It seems likely that Geomys expanded its
range into West Virginia during a dry, probably gla-

cial interval prior to the latest Wisconsin Glaciation.
The absence of Geomys in the rodent-rich upper
levels of the New Trout Cave deposit and in other
late Rancholabrean faunas of the Central Appala-
chians indicates that Geomys did not survive to this
interval in West Virginia. A thomomine pocket go-
pher, Thomomvs potomacensis, is present in the
Irvingtonian fauna of nearby Trout Cave and be-
came extinct prior to Rancholabrean time (Kurten
and Anderson, 1980).

Much work remains to be done on the New Trout
Cave Fauna. Despite removal of some four and a
half tons of matrix, much more remains for future
investigators. New localities in nearby Trout and
Hamilton Caves give promise of further additions
to the Pleistocene faunal record of West Virginia.

1984

GRADY — GEOMYS FROM WEST VIRGINIA

163

ACKNOWLEDGMENTS

John Guilday supported the New Trout Cave project in many
ways from 1 979 to 1 982. Ray Garton was co-director of the New
Trout Cave project and assisted in ali phases of the collection
and processing of the materials. The National Speleological So-
ciety, current owner of New Trout Cave has provided continuing
support for paleontological research in the cave. Members of two
cave clubs, D. C. Grotto and Monongahela Grotto, provided the

volunteers that made possible the collection of some four and a
half tons of matrix. Allen McCrady and Harold Hamilton pro-
vided assistance in the processing of matrix from New Trout
Cave at the New Paris field laboratory. Gary Morgan assisted in
the identification of the tibia. Robert Emry and Clayton Ray
critically read this paper.

LITERATURE CITED

Grady, F., and E. R. Garton. 1981. The collared lemming
Dicrostonyx hudsonius (Pallas) from a Pleistocene cave de-
posit in West Virginia. Proc. Intemat. Congress of Speleol.,
8:279-281.

Guilday, J. E. 1977. Sabertooth cat, Smilodon floridanus (Lei-
dy) and associated fauna from a Tennessee Cave (40 Dv 40),
the First American Bank Site. J. Tennessee Acad. Sci., 52:
84-94.

. 1979. Eastern North American Pleistocene Ochotona

(Lagomorpha: Mammalia). Ann. Carnegie Mus., 48:435-
444.

Guilday, J. E., H. W. Hamilton, and A. D. McCrady. 1971.
The Welsh Cave peccaries (Platygonus compressus) and as-

sociated fauna, Kentucky Pleistocene. Ann. Carnegie Mus.,
43:249-320.

Hall, E. R. 1981. The mammals of North America. John Wiley
and Sons, New York, 2 vols., 1,180 pp.

Kurten, B., and E. Anderson. 1980. Pleistocene mammals of
North America. Columbia Univ. Press, New York, 442 pp.

McKeever, S. (MS). Ecology and distribution of the mammals
of West Virginia. Unpublished PhD thesis, Univ. Micro-
films, Ann Arbor, 335 pp.

Parmalee, P. W., and W. E. Klippel. 1981. A late Pleistocene
population of the pocket gopher, Geomys cf. bursarius, in
the Nashville Basin, Tennessee. J. Mamm., 62:831-835.

Address: Department of Paleobiology, National Museum of Natural History, Smithsonian Institution,

Washington, D.C. 20560.

NEOTOM A IN THE LATE PLEISTOCENE OF NEW MEXICO

AND CHIHUAHUA

Arthur H. Harris

ABSTRACT

Selected characters apt to be present in fossil specimens were
studied in modem woodrats ( Neoloma ) in an effort to determine
discriminatory features. Emphasis was on the lower first molar.
Most modern specimens of the eight species studied (N. albigiila,
N. cinerea, N. floridana, N. goldmani, N. lepida, N. mexicana,
N. micropus, and N. stephensi) can be identified correctly to species
by the use of standard statistical methods and discriminant anal-
ysis. Particularly important in preliminary separation into major
groups is the absence or near-absence (<2 mm in height) of the
lateral dentine tract of the first lower molar in one group (N.
albigida, N. floridana, some N. lepida, and N. micropus) versus
presence (>2 mm in height) in a second group (N. cinerea, N.
goldmani, most N. lepida, N. mexicana, and N. stephensi). With-
in the group possessing the developed tract, N. goldmani and N.
lepida tend to have the tract lower in height than do the other
members of the group.

Application of the discriminatory data to over 500 fossil spec-
imens from 24 late Pleistocene and early Holocene sites located
in New Mexico and southern Chihuahua reveals profound dif-
ferences between interstadial, stadial, and early Holocene wood-
rat faunas. Interstadial faunas (ca. 25,000 to 33,000 B.P.) were
characterized by presence of two undescribed species (apparently
related to N. cinerea and N. goldmani) along with N. albigida
and N. micropus. Most stadial sites were dominated by N.
cinerea, with other species represented being N. albigula, N.
? goldmani, N. floridana, and N. micropus. Of these, N. floridana
and possibly N. micropus appeared only toward the end of stadial
times. Jimenez Cave, in southern Chihuahua, is not certainly
stadial in age; N. lepida was common. Early Holocene sites had
woodrat faunas similar to those of today except for the additional
presence of N. mexicana.

INTRODUCTION

Virtually every area of the western United States
and northern Mexico is inhabited by one or more
species of woodrats, genus Neotoma. Their remains
often are abundant in late Pleistocene cave faunas
and their contribution to our knowledge of Pleis-
tocene ecology by means of their preserved middens
is well known. Their own skeletal remains have con-
tributed relatively little to our knowledge of the
Pleistocene, however, because identification to
species often is difficult and, to the skeptical, sus-
pect. Reasonably sure identifications would add sig-

nificantly to our knowledge of Pleistocene biogeog-
raphy and, to a lesser degree, ecology. The aims of
this study were to produce discriminating criteria
usable with commonly preserved fossil elements for
those species apt to be found in late Pleistocene of
the Southwest and northern Mexico, to apply these
criteria to late Pleistocene/early Holocene speci-
mens available to me, and to interpret the finding
in terms of systematics, biogeography, and paleo-
ecology.

MATERIALS AND METHODS

Modern comparative material was assembled for species cur-
rently living in the southwestern region and for species judged
to have possibly occurred in the area during the late Pleistocene.
These species are Neotoma albigula Hartley (white-throated
woodrat), N. cinerea (Ord) (bushy-tailed woodrat), N. floridana
(Ord) (eastern woodrat), N. goldmani Merriam (Goldman’s
woodrat), N. lepida Thomas (desert woodrat), N. mexicana Baird
(Mexican woodrat), N. micropus Baird (southern plains woodrat),
and N. stephensi Goldman (Stephens’ woodrat). Limitations of
time and scarcity of specimens in collections ( N . goldmani) have
resulted in several samples of less than ideal size for statistical
treatment.

Most effort has been directed toward the dentary with its teeth,
particularly m 1 . This in part reflects commonness of recovery in
fossil faunas and in part the potential for identification.

Measurements on both modem and fossil material were taken
with an ocular micrometer to the nearest 0. 1 mm except for the

greatest width of loph 2 of m 1 and depth of incisor, which were
taken to 0.01 mm with dial calipers. Measurements consistently
taken on lower jaw elements were: 1) length of alveolar cheek-
tooth row (LG-ALV); 2) mid-length of ml (LG-M1); 3) greatest
width of loph 2 of ml (WD-M1); 4) height of antero-lateral
dentine tract of ml (TRACT); 5) distance from base of lingual
fold 1 to base of fold 2 of ml (F1-F2); 6) distance from base of
lingual fold 2 to anterior face of ml (ANT-F2); 7) development
of the antero-intemal reentrant fold of loph 1, ml (FOLD); and
8) a ratio comparing depth of the ml anterointemal reentrant
fold to the width of ml (RATIO). These measurements are in-
dicated in Fig. 1.

LG-M1 was taken from the lingual side at a level estimated to
lie between one-third and one-half the height of an unworn tooth
and perpendicular to the vertical axis of the tooth (not parallel
to the wear surface). Occlusal length proved to be too variable
with wear and angle of wear to be as useful a measurement.

164

1984

HARRIS- LATE PLEISTOCENE NEOTOM A

165

b

Fig. 1.— Qualitative characters and methods of taking measurements. A) Lingual view of left ml of Neotorna cinerea, showing method
of taking LG-M1, ANT-F2, and F1-F2. B) Occlusal view of left ml of Neotoma albigula, showing measurements for calculation of
RATIO (a/b). C) Labial view of left ml of N. cinerea, showing method of measuring dentine tract (TRACT) and presence of accessory
cusp (ac). D) Occlusal view of left m3 of Neotoma goldmani, showing accessory fold (af).

WD-M 1 was taken with caliper blades approximately vertical.
The bulge present in some teeth at the base of F2 was avoided.
A series of measurements on the same specimen occasionally will
vary by several hundreths millimeter; heavily worn teeth tend to
give an underestimate. Very young teeth (roots broadly open)
also often give smaller measurements than comparable adult teeth
of the same species even when the enamel appears complete.

The anterolateral dentine tract (Fig. 1) was, to the best of my
knowledge, first noted by Lundelius (1979) in N. cinerea and N.
mexicana. Differing in presence and development within the ge-
nus, it is an extension of the enamel-less area ventral to the crown
onto the lateral surface of loph 1 (lesser developed tracts may
occur on the other lophs and on those of m2 and m3). Height
was measured from a line parallel to the wear surface extended
anteriorly from the base of fold 1 (Fig. 1). In most cases, a sharp
demarcation between the normal enamel surface and the dentine
was present and used for the upper limit. In some specimens,
the transition from dentine to normal enamel is attenuate, with-
out sharp limits; the uppermost appearance of dentine was used
in these cases. With wear, the uppermost portions of the tract
may be worn away and a conservative estimate of tract height

was made. Tract measurements of the high-dentine forms thus
are biased slightly to the low side.

Measurements involving lingual folds were taken from the most
ventral point of the fold, which may not be the median point.
For ANT-F2, the measurement was made to the anterior wall of
the tooth above any abrupt basal constriction.

Development of the anterointemai reentrant fold was coded
as shown in Fig. 2. Except for the 0. 1 coding, depth may vary
greatly (though fairly typical examples are shown)— closeness of
the fold base to the tooth base is the criterion used. This is
somewhat subjective and wear may obliterate categories 0.2 and
0.3.

RATIO was used in an attempt to quantify the depth of the
anterointemai fold in relation to the width of loph 1 (Fig. 1).
Corrections for wear were made. Wear categories are 1 ) very light
(depressions on occlusal surface not yet obliterated or enamel
walls not fully worn to a flat surface); 2) light (base of labial folds
not erupted to level of lateral alveolar wall, or judged to have
not reached that stage in the case of isolated teeth; tooth usually
notably tapered toward the top); 3) moderate (fold bases erupted
but wear not sufficiently close to base as to distort enamel pat-

166

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 2. — Lingual and occlusal views of ml’s showing character of antero-internal grooves used for coding. 1) Neotoma albigula, coding
0.1. 2) Neotoma micropus, coding 0.2. 3) Neotoma goldmani, coding 0.3. 4) Neotoma mexicana, coding 0.4.

terns); 4) heavy (enamel patterns obviously distorted in propor-
tions); and 5) very heavy (portions of enamel wall other than at
the dentine tract missing, often fold bases isolated as islands of
enamel surrounded by dentine). Very heavy wear specimens were
not used in the analysis. Correctives for wear are given in Table
1 and were obtained by calculating the factors necessary to correct
to the average moderate-wear category in the modern samples.

In general, more subjective weight was given to LG-M1, WD-
M 1 , TRACT, and FOLD than to the other characters.

Additional data taken include samples of lower incisor depth
(greatest distance from dorsal to ventral surface perpendicular to
the tooth axis) and several qualitative characters. These included
the presence of accessory cusps at the fold bases in ml and m2
(Fig. 1), presence of an accessory fold in m3 (Fig. 1), the devel-
opment of the capsule at the base of the incisor (Fig. 3), the
character of the mandibular foramen (Fig. 3), and the relative
depths of the external and internal reentrant folds of m3.

The quantitative data were subjected to standard statistical

Table 1 —Correction for wear for measurement R.AT10. The raw measurement is multiplied by the wear corrective.

Taxon

Wear category

Very light

Light

Moderate

Heavy

Neotoma cinerea. N. mexicana, N. goldmani

1.89

1.31

1.00

0.96

Neotoma lepida

1.00

1.31

1.00

0.96

Neotoma albigula, N. floridana, N. micropus

-

1.22

1.00

0.96

1984

HARRIS- LATE PLEISTOCENE NEOTOMA

167

Fig. 3. — Mandibular foramen types (left) and incisor capsule types
foramen. B) Neotoma albigula, showing enlarged incisor capsule,
foramen. D) Neotoma goldmani, showing reduced incisor capsule

treatment, but also multivariate techniques were utilized, in-
cluding discriminant analysis (BMD07M, SPSS) and clustering
techniques (NT-SYS).

Although the sample statistics and scattergrams can be used
for identification, a somewhat more objective method is the use
of discriminant analysis. This procedure is most powerful when
used between pairs of taxa, in that the weighted characters that
best separate three or more taxa are not necessarily the best for
separation of any single pair of taxa. In practice, unknown spec-
imens that might belong to any of several taxa are processed with
samples of the several known taxa, the results being used to
segregate the unknowns to fewer probable choices. The data given
in the text and in Table 2 and Figs. 5 to 7 generally should be
sufficient to make a preliminary reduction to two or three prob-
able taxa. The data given in Tables 3 and 4 may then be used to

(right). A) Neotoma albigula, showing ventrally oriented mandibular
C) Neotoma goldmani, showing more laterally directed mandibular

assign discriminant scores to individual specimens and make
identifications based on these scores.

Several cautions should be observed: 1 ) if the specimen belongs
to a taxon not represented in the discriminant analysis, a spurious
identification will result; 2) some specimens of one taxon overlap
with members of another taxon to the point that they are indis-
tinguishable on the basis of the characters used, again resulting
in an incorrect or equivocal identification; 3) Neotoma has a nasty
habit of occasionally having an otherwise rather stable character
state shift to the mode seen in a different taxon in rare individuals;
if this is a heavily weighted character, assignment may be in-
correct. For this reason, measurements should be scanned for
reasonableness. As an example, several unknown specimens were
assigned to N. micropus, but WD-M 1 was considerably less than
minus two standard deviations from the N. micropus mean.

168

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 2. — Basic statistics for samples of modern Neotoma spec-
imens.

Taxon

Mean

SD

Observed range

N

Neotoma albigula

LG-M1

3.09

0.136

2. 7-3. 3

36

WD-M1

1.815

0.073

1.63-1.97

72

TRACT

0.01

0.028

0.0-0. 1

72

ANT-F2

2.29

0.126

2.0-2. 6

36

F1-F2

0.85

0.091

0

1

b

36

FOLD

0.18

0.059

0. 1-0.3

36

RATIO

0.671

0.062

0.55-0.83

36

Neotoma cinerea

LG-M1

3.52

0.122

3. 3-3. 7

16

WD-M1

1.954

0.102

1.79-2.14

16

TRACT

1.14

0.242

O

be

ON

16

ANT-F2

2.68

0.181

2. 4-3.0

16

F1-F2

1.03

0.101

0.9-1. 2

16

FOLD

0.32

0.098

0. 1-0.4

16

RATIO

0.481

0.144

0.26-0.71

16

Neotoma floridana
LG-M1 3.46

0.130

3. 3-3. 8

22

WD-M1

2.070

0.084

1.93-2.26

22

TRACT

0.02

0.053

0.0-0. 2

22

ANT-F2

2.57

0.109

2. 4-2. 8

22

F1-F2

0.98

0.102

0.9-1. 2

22

FOLD

0.27

0.089

0. 1-0.4

22

RATIO

0.479

0.076

0.32-0.59

22

Neotoma goldmani
LG-M1 2.78

0.128

2. 6-3.0

8

WD-M1

1.639

0.071

1.54-1.79

8

TRACT

0.35

0.120

0.2-0. 5

8

ANT-F2

1.93

0.158

1. 7-2.1

8

F1-F2

0.80

0.053

0.7-0. 9

8

FOLD

0.20

0.093

0. 1-0.3

8

RATIO

0.645

0.083

0.56-0.75

8

Neotoma lepida

LG-M1

3.04

0.137

2. 8-3. 3

28

WD-M1

1.708

0.081

1.54-1.83

28

TRACT

0.28

0.188

0.0-0. 7

28

ANT-F2

2.30

0.129

2. 1-2.5

28

F1-F2

0.84

0.057

0.7-0. 9

28

FOLD

0.25

0.079

0. 1-0.3

28

RATIO

0.569

0.087

0.42-0.70

28

Neotoma mexicana

LG-M1 3.17

0.145

2. 9-3. 5

42

WD-M1

1.723

0.063

1.61-1.85

42

TRACT

1.32

0.345

0.7-2. 2

42

ANT-F2

2.39

0.144

2.1-2. 7

42

F1-F2

0.87

0.077

0.7-1. 1

42

FOLD

0.39

0.034

0.2-0. 4

42

RATIO

0.546

0.081

0.39-0.77

42

Neotoma micropus

LG-M1 3.17

0.106

2. 9-3. 4

26

WD-M1

1.998

0.090

1.84-2.21

26

TRACT

0.004

0.020

0.0-0. 1

26

ANT-F2

2.29

0.141

2. 0-2. 6

26

F1-F2

0.94

0.110

0.6-1. 1

26

Table 2. — Continued.

Taxon

Mean

SD

Observed range

N

FOLD

0.069

0.014

0. 1-0.3

26

RATIO

0.609

0.072

0.47-0.77

26

Neotoma stephensi

LG-M1

2.79

0.178

2.5-3. 1

12

WD-M1

1.713

0.087

1.62-1.93

12

TRACT

1.17

0.328

0.6-1. 7

12

ANT-F2

2.14

0.144

1. 9-2.4

12

F1-F2

0.84

0.051

0. 8-0.9

12

FOLD

0.29

0.051

0. 2-0.4

12

RATIO

0.564

0.110

0.43-0.77

12

Some idea of the trustworthiness of discriminant analysis for
identifying members of two taxa can be gained by using discrim-
inant criteria to assign identifications to individual members of
the modem samples. Results are given in Tables 3 and 4. The
assignment error of non-sample specimens is expected to be great-
er, particularly when the discriminant analysis has available as
“knowns” only samples of small size.

Final identifications of the fossil material were by an amal-
gamation of all multivariate and statistical techniques as tem-
pered by characteristics of the specimens and samples themselves
(wear, preservation, identifications of other Neotoma in the site,
etc.).

Table 3. — Pair-wise unstandardized canonical discriminant func-
tion coefficients for computing discriminant scores of non-dentine
tract Neotoma ml’s apt to be confused. To determine score for
an unidentified individual, each measurement is multiplied by the
respective coefficient and these are summed and added to the
constant. The specimen is assigned to species according to its
position along the discriminant function (greater or less than the
division point). Some measurements are not of value for some
species and are omitted. Taxa are identified by initial of species
name.

Coefficients for species pairs

Variable

A/F

A/L

A/M

F/M

LG-Ml

—

—

—

5.483

WD-M1

8.892

-7.236

-1 1.757

—

TRACT

—

6.146

—

9.495

ANT-F2

—

2.134

6.830

7.058

F1-F2

—

—

-4.515

-8.473

FOLD

—

—

5.189

—

RATIO

-9.851

-5.748

1 1.771

-2.622

Constant

-11.044

10.674

1.941

-25.741

Division point

-0.514

0.182

-0.267

0.156

Species above
point

Nf

N.l.

N.a.

Nf.

Modern sample
discriminated

100%

91%

100%

96%

1984

HARRIS-LATE PLEISTOCENE NEOTOMA

169

Table 4.— Pair-wise discriminant function coefficients for dentine tract Neotoma. See caption of Table 3 for explanation.

Variable

Coefficients for

species pairs

C/M

c/s

G/L

G/S

L/S

M/S

LG-Ml

-2.270

6.431

—

_

-3.120

5.685

WD-M1

13.109

3.642

—

7.779

3.408

-4.880

TRACT

—

—

—

3.908

3.913

—

ANT-F2

5.825

—

7.378

—

—

—

F1-F2

—

—

—

—

—

-4.038

FOLD

-16.030

-4.952

—

—

—

15.803

RATIO

-

-

-

-

-

-

Constant

-24.390

-25.844

-16.333

-16.375

1.296

-1 1.519

Division point

1.148

-0.387

-0.761

-0.377

0.853

-0.998

Species above point

N.c.

N.c.

N.l.

N.s.

N.s.

N.m.

Modem sample discriminated

100%

100%

89%

100%

100%

94%

More than 500 specimens of Neotoma from the late Pleistocene
and early Holocene of Chihuahua and New Mexico were avail-
able for study. These specimens are preserved in the Resource
Collections, Laboratory for Environmental Biology, University
of Texas at El Paso (UTEP). Sites and site data are given in Table
5; site localities are shown in Fig. 4.

RESULTS AND DISCUSSION

The basic statistics of the modern species samples
are given in Table 2 and in part displayed in Figs.
5 to 7.

Specimens of Neotoma can be divided into two
major groups on the basis of the development of the
lateral dentine tract on m 1 . Most specimens of N.
albigula, N. floridana, and N. micropus basically
lack the tract, though a few individuals have the
tract developed to a height of ca. 0. 1 mm (0.2 mm
in N. floridana). The remaining species have a tract
height of >0.2 mm except for some N. lepida (ca.
36% of the modern sample).

Qualitative characters of the dentary differing be-
tween the two groups include generally greater de-
velopment of the incisor capsule and a somewhat
more ventrally oriented mandibular foramen in the
non-dentine tract group (Fig. 3). Neither character
is entirely clearcut in all cases, and N. lepida is more
similar to the non-dentine tract species.

Within the group possessing a dentine tract, there
is a dichotomy between those that possess a rela-
tively low tract ( N . goldmani and N. lepida) and
those with a high tract ( N . cinerea, N. mexicana,
and N. stephensi) (Fig. 7).

Table 6 shows the identifications of the Neotoma
found in each site treated.

Fig. 4. — Sketch of New Mexico and Chihuahua, showing late
Pleistocene sites. 1) Isleta Caves; 2) Howell’s Ridge Cave; 3)
Baldy Peak Cave; 4) Khulo Site; 5) Conkling's Cavern and Shelter
Cave; 6) Anthony Cave; 7) Dry Cave; 8) Dark Canyon Cave; 9)
Jimenez Cave.

170

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 5.

—Sites from which fossil Neotoma were examined. Those marked with an asterisk are in the Dry Cave system.

UTEP no.

Name

Age (BP)

Reference

1, 17

*Lost Valley

29,290 ± 1,060

Harris, 1977

4

*Bison Chamber

<14,470, >10,730

Harris, 1977

5

*Sabertooth Camel Maze

25,160 ± 1,730

Harris, 1977

6

*Harris’ Pocket

14,470 ± 250

Harris, 1977

21

Khulo Site

<10,000

Harris, 1977

22

*Animal Fair

15,030 ± 210

Harris, 1977

23

*Stalag 17

1 1,880 ± 250

Harris, 1977

24

*Entrance Chamber

<1 1,880

Harris, 1980

25

*Camel Room

Est. > 1 2,000

Harris, 1977

26, 27

*Rm Vanishing Floor

33,590 ± 1.500

Harris, 1977

28

*Rick’s Cenote

Est. 11,000

Unpublished

29

Anthony Cave

stadial

Harris, 1977

30

Shelter Cave

stadial, Holocene

Harris, 1977

32

Howell’s Ridge Cave

stadial, Holocene

Harris, 1977; Van Devender and
Wiseman, 1977

41

Isleta Cave No. 1

stadial, Holocene

Harris and Findley, 1964

46

Isleta Cave No. 2

stadial, Holocene

Harris and Findley, 1964

54

*TT II

10,730 ± 150

Harris, 1977

75

Dark Canyon Cave

stadial

Harris, 1977

90

Conkling Cavern

stadial

Harris, 1977

91

Jimenez Cave

?stadial and ?Holocene

Unpublished

94

Baldy Peak Cave

stadial and ?Holocene

Unpublished

122

*Pit N & W Animal Fair

?early stadial

Harris, 1977

Table 6.

— Taxa identified from each site.

Taxon

Neotoma

Neotoma

Neotoma

Neotoma Neotoma Neotoma

Neotoma

UTEP Loc.

albigula

cinerea

jloridana

goldmani lepida mexicana

micropus

A

B

Interstadial

1

X

X

X

X

5

X

17

X

X

X

26

X

X

27

X

X

Stadial

4

9

6

X

X

cf.

22

X

X

?

23

X

X

25

X

28

cf.

29

X

X

9

30

X

32

X

X

41

X

X

9

46

X

X

9

54

cf.

cf.

X

75

X

X

90

X

X

91

X

?

cf.

X

X

94

X

X

122

X

Holocene

21

X

X

9

24

X

X

X

1984

HARRIS- LATE PLEISTOCENE NEOTOM A

171

Fig. 5. — Scattergram of LG-M1 (ordinate) against WD-M1 (abscissa). Circles, modern Neotoma albignia: dots, modem Neotoma
micwpus\ X, modem Neotoma floridana\ triangles, selected fossil specimens from sites shown by numbers. Line shows best separation
between N. albigula and N. micropus.

Species Accounts
Neotoma albigula

The white-throated woodrat now occurs at or near
all sites. It apparently was as ubiquitous in the past,
occurring in interstadial, stadial, and early Holocene
sites.

This rat is easily separable from other Neotoma
except for N. micropus and some individuals of N.
lepida on the basis of ml . Dalquest et al. (1969) and
Lundelius (1979) have discussed separation criteria
between N. albigula and N. micropus. Dalquest et
al. found no overlap in width of loph 2 of ml, A.
albigula having a width of <1.94 mm and N. mi-
cropus a width of > 1 .94 mm. Lundelius, in a sample
of 32 N. albigula and 30 N. micropus , found 9.4%

of the former with breadths > 1.94 mm and 16.7%
of the latter with widths < 1.94 mm. In the present
study, a sample of 72 A. albigula showed three (4.2%)
exceeding 1.94 mm, and five of 26 (19.2%) A. mi-
cropus with measurements of < 1.94 mm. A bivar-
iate scattergram of LG-M1 and WD-M1 (Fig. 5)
does a somewhat better job of separation, but some
misidentifications on the basis of these measure-
ments are inevitable.

Confusion between A. albigula and A. lepida re-
sults from the inability to separate those A. lepida
with dentine tracts <0.2 mm from A. albigula.

There is some variability within the species (Fig.
8), with some tendency for smaller size in the stadial
sites and with the length of ml from Jimenez Cave
(UTEP 91) and the Khulo Site (UTEP 21) being

NO. 8

1 72 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

Fig. 6. — Dispersal of plotted points of LG-M1 (ordinate) against WD-M1 (abscissa) for modern specimens of five high dentine tract
species, fossil specimens of extinct Species A, and two fossil specimens from UTEP 91 (species identified by initials). Polygraphs were
constructed by connecting marginal points.

somewhat large. Jimenez Cave lies within the cur-
rent range of N. albigula durangae, which Anderson
(1972) notes as being somewhat intermediate be-
tween N. a. albigula and N. micropus ; this may be
showing up here. No such explanation is available
for the Khulo Site sample, nor do width measure-
ments indicate much chance of N. micropus biasing
the sample by wrongful inclusion.

Today, N. albigula occurs in a variety of habitats
from desert to pinyon-juniper woodlands and pon-
derosa pine forests, though seldom found in pure
grassland. In various parts of its range it may be
associated with all of the species considered here,
though only marginally with some.

Neotoma cinerea

The bushy-tailed woodrat occurs in every stadial
fauna for which there is at least a moderate-sized
sample, and often makes up the majority of the
recovered specimens.

There is remarkably little variation from site to
site and from Pleistocene populations to modern
populations, no significant differences being found
among the three characters looked at closely (LG-
Ml, WD-M1, TRACT). The two specimens (que-
ried) from Jimenez Cave, however, are exception-
ally short, though of normal width and dentine tract
height (Fig. 6); these suggest intraspecific variation
at this far southern extension of the geographic range.

1984

HARRIS- LATE PLEISTOCENE NEOTOMA

173

Fig. 7. — Dispersal of plotted points of TRACT (ordinate) against WD-M1 (abscissa). Specimens as in Fig. 6, without the two fossil
specimens from UTEP 91.

but may represent an unknown taxon. Table 7 gives
the basic statistics for three measurements from the
three largest samples.

Lundelius (1979) identified N. cinerea from Pratt
Cave (located in the Texas portion of the Guadalupe
Mountains just south of the New Mexico border;
age is considered Holocene by Lundelius, but some
material may be Pleistocene) and from Dark Can-
yon Cave (Texas Memorial Museum specimens). He
suggested that these specimens represented popu-
lations with larger ml’s than those of modern N.
cinerea from New Mexico; that they were of a size
more like modern specimens from Wyoming. How-
ever, there is no significant difference between the
present small sample of Recent Neotoma from New
Mexico (n = 9) and the UTEP sample from Dark
Canyon Cave (n = 9 adults). The Dark Canyon sam-
ple of Lundelius averaged 0.41 mm larger than his
small sample (n - 4) of modern New Mexican spec-
imens, whereas the difference in the present study

is 0.1 mm (the Lundelius measurement, however,
was occlusal length of ml rather than mid-length).
Neotoma cinerea apparently was absent from the

Table 7 ,—LG-Ml , WD-M1, and TRACT statistics Tor Neotoma
cinerea from UTEP 22, 29, and 75.

Locality

Mean

SD

Observed range

N

UTEP 22

LG-M1

3.54

0.214

3.2-4. 1

29

WD-M1

1.962

0.115

1.78-2.21

29

TRACT

1.10

0.219

0.7-1. 6

29

UTEP 29

LG-M1

3.49

0.151

3. 2-3. 7

1 1

WD-M1

1.986

0.071

1.88-2.12

1 1

TRACT

1.16

0.401

0

4-

1

V£>

1 1

UTEP 75

LG-M1

3.56

0.167

3. 3-3. 8

9

WD-M1

1.970

0.105

1.85-2.14

9

TRACT

1.04

0.255

0.6-1. 4

9

174

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

J I 1 1 I I I L

Loc. I
Loc. 21
Loc. 22
Loc. 24
Loc. 29
Loc. 90
Loc. 91
Holocene

— i 1 1 -t 1 1 1 1 —

2.7 2.8 2.9 3.0 3.2 3.3 3.4 3.5

Fig. 8. — ml lengths of fossil and modem Neotoma albigitla. Vertical lines, sample means; horizontal lines, observed range; box, 95%
confidence interval of the mean; numbers to right of figures, sample size.

interstadial deposits of Dry Cave (but see comments
below, in the “Species A, undescribed” account) and
from deposits that are surely Holocene, but as noted
above, was virtually ubiquitous during stadial times
with little morphological variation. Likely there was
an essentially continuous population throughout the
region. Holocene occurrence is possible at several
sites, including the Isleta Caves, Howell’s Ridge Cave
(UTEP 32, possibly representing TV. mexicana ), and
Pratt Cave, but its absence and replacement by TV.
mexicana at the presumably uncontaminated En-
trance Chamber deposit (UTEP 24) at Dry Cave
and the Khulo Site (UTEP 21) suggest otherwise.

Some earlier workers have interpreted presence
of TV. cinerea at low elevations south of their present
range as indicative of the forests where they are
found now at their point of nearest occurrence in
the Sangre de Cristo Mountains of northern New
Mexico (for example, Murray, 1957). This, how-
ever, ignores their occurrence under far different
conditions and at lower elevations in northwestern
New Mexico (extending into pinyon-juniper wood-
land and occasionally lower) and their widespread
habitation of sagebrush and the like farther north
(Cary, 1917). Likely TV. cinerea inhabited all vege-
tational zones from timberline down to, and in-

1984

HARRIS- LATE PLEISTOCENE NEOTOM A

175

eluding, steppe-woodland (terminology after Harris,
in press), an open woodland with well developed
growth of grasses, herbs, and shrubs. Pure grassland,
however, probably was not inhabited. Within these
vegetationai types, they likely were limited as today
(Armstrong, 1972) to areas of cliffs, jumbled rocks,
and caves.

Neotoma floridana

The most convincing evidence of the eastern
woodrat in the southwestern Pleistocene is the cen-
tral portion of a skull (UTEP 23-79) from the Stalag
1 7 site in Dry Cave. Qualitative features, including
the presence of a well-forked palatine spine, fit N.
floridana rather than N. micropits. Dalquest et al.
(1969) suggest that the two species are separable on
the basis of breadth of the maxillary molar rows (N.
micropus. 8.0 to 8.7 mm; N. floridana. 8.7 to 9.2
mm). The method of taking molar row breadth is
not described. In the present study, measurement
was done by applying caliper blades to the outer
sides of the toothrows at the alveolar line (mea-
surements to the most lateral tooth surfaces are much
larger). In 26 N. micropus, breadth measurements
ranged from 7.67 to 8.64 mm; in 26 N. floridana,
8.06 to 9.48 mm, with 10 of the specimens <8.7
mm. Molar row breadth in 23-79 is 8.85 mm, well
beyond any N. micropus measured by myself,
Dalquest et al. (1969), or Lundelius (1979). Two
dentaries with m 1 ’s from the same site are identified
as N. floridana by discriminant analysis (see also
Fig. 5). Both ml’s have the deep anterointernal fold
that occurs with fair frequency in N. floridana but
is rare in TV. micropus.

Two specimens from the early Holocene Khulo
Site (UTEP 21) could be TV. floridana, but are as-
signed here to TV. ? micropus. If TV. floridana, they
would be at the small end of their size range. A
similar situation occurs in the Holocene Entrance
Chamber deposits (UTEP 24) of Dry Cave, but here
TV. micropus also seems present and the possible N.
floridana again would have to be at the lower limits
of its size distribution. Other equivocal specimens
come from UTEP 46; they are identified here as TV.
? micropus.

A much better candidate for TV. floridana is spec-
imen 28-1, an ml from Rick’s Cenote (UTEP 28)
within Dry Cave. Identified by discriminant anal-
ysis as this species, its length (3.5 mm) is beyond
any in the modern sample of TV. micropus (longest,
3.4 mm). This site should be very late Pleistocene,
as is another site in Dry Cave, UTEP 54 (approxi-

Table 8 . — Basic statistics for combined sample of Neotoma
?go!dmani from UTEP 22 and 29.

Measure-

ment

Mean

SE

Observed range

N

LG- Ml

2.86

0.052

2. 8-2. 9

8

WD-M1

1.698

0.055

1.60-1.77

8

TRACT

0.23

0.1 16

0.0-0. 4

8

ANT-F2

2.04

0.074

1.9-2. 1

8

F1-F2

0.79

0.064

0.7-0. 9

8

FOLD

0.18

0.071

0. 1-0.3

8

RATIO

0.673

0.059

0.57-0.76

8

mately 10,730 radiocarbon years, but with extinct
fauna). Two of the three available specimens seem
to be TV. micropus, but the third clearly falls into the
TV. floridana area (Fig. 5). Four ml specimens from
Jimenez Cave (UTEP 91) in Mexico are identified
as TV. floridana by discriminant analysis, and another
as TV. micropus. Four of the specimens, including
the latter, fall into the intermediate length-width
area, but the fifth specimen is unequivocally within
the TV. floridana area (Fig. 5).

Thus, in summary, TV. floridana seems defendably
identified from near the Pleistocene-Holocene
boundary in UTEP 23, and the evidence is rather
strongly suggestive of presence in two other Dry
Cave sites of latest Pleistocene age. Evidence also
suggests presence at the undated Jimenez Cave, in
southern Chihuahua, Mexico. Occurrence at two
other sites (UTEP 21 and 46), both sites with early
Holocene materials, is possible but judged unlikely.

Neotoma Igoldmani

Five ml’s from the Animal Fair site (UTEP 22)
of Dry Cave and three from Anthony Cave (UTEP
29) are closer in most measurements (Table 8) to
TV. goldmani than to TV. lepida. Width and dentine
height are slightly more similar to TV. lepida than to
TV. goldmani, nearly reaching a significant difference
from TV. goldmani on dentine height in the com-
bined sample (n = 8). In all other traits, however,
TV. lepida is more dissimilar (significantly so for M 1 -
LG, P < 0.001; ANT-F2, P < 0.001; F1-F2, P =
0.033; GROOVE, P = 0.028; and RATIO, P =
0.003). In addition, one of two m3’s associated with
the ml’s has an accessory fold and the other a sug-
gestion of such. Six of eight modern TV. goldmani
possess this fold and it was probably developed on
a seventh (the foldlet becomes unclear and then dis-
appears with moderate to heavy wear). In 30 TV.
lepida from east of California, only one definitely

176

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

has such a fold. Of 15 specimens of TV. lepida from
California and Baja California, however, six show
folds (this population, however, is much larger in
size than interior TV. lepida).

Today, TV. goldmani occurs in eastern Mexico al-
most north to the Big Bend of Texas. A continuous
distribution northward through the mountain ranges
of Trans-Pecos Texas and into southeastern New
Mexico would not be startling, and this rare rat may
well occur today north of its recognized range. It
probably is an inhabitant today of rocky areas in
arid mountains.

Neotoma lepida

The ml of this animal seems to differ from that
of TV. albigula primarily in the possession in about
two-thirds of the individuals of a dentine tract >0.2
mm in height; length averages significantly less, but
with much overlap. Six of 72 ml’s of TV. albigula
had a dentine tract height of 0. 1 mm; none had a
higher tract. Thus possession of a tract >0.2 mm
rules out TV. albigula beyond reasonable doubt. Ba-
sic statistics for UTEP 91 are given in Table 9. A
comparison of dentine tract heights between the
UTEP 9 1 sample and that of modem TV. lepida shows
a significantly lower tract in modern TV. lepida ( P =
0.017); this probably is biased, however, because
fossil TV. lepida m 1 's with a dentine tract that is <0.2
mm cannot be separated from those of TV. albigula
except when the teeth are notably smaller, and thus
most such low-height tract teeth would not appear
in the sample. The UTEP 91 population also differs
significantly from modern TV. lepida in slightly
shorter length of ml (P = 0.023) and the measure-
ment ANT-F2 ( P < 0.001), and in larger RATIO
measurement (P = 0.001). These differences are in-
terpreted as geographic/chronologic variation, Ji-
menez Cave being far from any modern population
of TV. lepida. A modern sample identified as TV. lep-
ida from California and Baja California shows far
more difference in ml length, for example (mean =
3.25, n= 13).

Neotoma mexicana

Demonstrable TV. mexicana are surprisingly rare
in the sites considered, and those that do occur do
so in generally unexpected circumstances. Unless a
few individuals are mixed in with the large number
of TV. cinerea identified from stadial sites (and this
is a possibility), TV. mexicana is absent from all sta-
dial sites with the exception of Baldy Peak Cave
(UTEP 94). This site had little fill, and post-Pleis-
tocene remains would be inseparable from older fos-

Table 9 .—Basic statistics for Neotoma lepida from UTEP 91.

Measure-

ment

Mean

SD

Observed range

N

LG-Ml

2.95

0.1 12

2.8-3. 2

23

WD-M1

1.733

0.061

1.61-1.82

23

TRACT

0.39

0.122

0. 1-0.5

23

ANT-F2

2.07

0.118

1. 9-2.3

23

F1-F2

0.83

0.076

0.7-0. 9

23

FOLD

0.25

0.095

o

T

o

23

RATIO

0.665

0.117

0.48-0.90

23

sils. This is the only site where both TV. cinerea and
TV. mexicana are solidly identified.

Other sites with TV. mexicana are Holocene. The
largest sample is from UTEP 2 1 , the Khulo Site (n =
7), and it seems comparable to modern New Mex-
ican TV. mexicana. From the Entrance Chamber of
Dry Cave (UTEP 24), however, two of the three
specimens are notably small for TV. mexicana, but
resemble no other known taxon.

Neotoma micropus

This large woodrat occurs from interstadial time
to the present, though not found in large numbers.
It may have been absent during full stadial condi-
tions. Difficulties in discrimination from TV. albigula
and TV. floridana have been described in those ac-
counts.

Species A, undescribed

An interstadial population of relatively large
woodrats with moderate development of the dentine
tract is represented by 12 ml’s as well as by addi-
tional material. The specimens clearly are separable
from all other interstadial populations (Figs. 6 and
7). They do show a relationship in morphological
characters to TV. mexicana and TV. cinerea. They
differ significantly from these species as a population
(Table 10), though some individual specimens of TV.
cinerea and TV. mexicana cannot be discriminated.
Although intermediate in several characters be-
tween those two species, the height of the dentine
tract averages conspicuously lower than in either. A
qualitative character, the presence of small acces-
sory cusps at the base of fold 2 of m2 on three
specimens, suggests closer relationship to TV. cinerea
than to TV. mexicana.

Species B, undescribed

A woodrat from the interstadial deposits of Dry
Cave (UTEP 1, 17, 26, 27) currently is being de-
scribed.

1984

HARRIS-LATE PLEISTOCENE NEOTOMA

177

Table 10. — Comparison of UTEP 1 "cinerea -type" specimens
(=Species A) (n = 10) with modern N. cinerea (n = 16) and N.
mexicana (n = 42). A = mean UTEP 1(1 SD). B = mean N.
cinerea (1 SD). C = t between UTEP 1 and N. cinerea. D = mean
N. mexicana (1 SD). E = t between UTEP 1 and N. mexicana.

*, significant at P

< 0.05;

**, P < 0.01;

*** p <

0.001.

Character

A

B

c

D

E

LG-Ml

3.29
(0.1 10)

3.52

(0.122)

4 8 1 ***

3.17

(0.145)

2.37*

WD-M1

1.814

(0.056)

1.954

(0.102)

3.95***

1.723

(0.063)

4 93***

TRACT

0.65

(0.158)

1.14

(0.242)

5.71***

1.32

(0.345)

5 97***

ANT-F2

2.50

(0.149)

2.68

(0.181)

2.56**

2.39

(0.144)

2.24*

F1-F2

0.92

(0.9)

1.03

(0.101)

2.82**

0.87

(0.077)

1.65

FOLD

0.40

(0.0)

0.32

(0.098)

2.60*

0.39

(0.034)

0.66

RATIO

0.487

(0.082)

0.481

(0.144)

0.11

0.546

(0.081)

2.07*

This taxon is significantly different from N. lepida
(P < 0.001) in all but GROOVE and RATIO, and
from N. goldmani in TRACT ( P < 0.001), ANT-
F2 (P = 0.042), and RATIO (P = 0.003). The most
meaningful difference from N. goldmani probably
is in the dentine tract height, which averages almost
0.2 mm higher in Species B (0.54 mm versus 0.35
mm). Three of four m3’s have accessory folds pre-
served.

Species B presumably is most closely related to
N. goldmani among all extant Neotoma. It is more
than twice as common in the interstadial deposits
than the next most common species, N. albigu/a.

General Discussion

Interstadial, stadial, and early Holocene Neotoma
faunas all differ radically from the modern condi-
tion, emphasizing the uniqueness of modern climate
and biology.

The recognized interstadial faunas from the re-
gion are all from Dry Cave, in southeastern New
Mexico. These faunas indicate conditions with
greater effective moisture than today, though prob-
ably less than during stadial times (or differently
distributed), an absence of cold winter temperatures,
and rather warm summer temperatures (Harris,
1 977; Van Devender et al., 1 976; Harris and Crews,
1983).

The interstadial woodrat faunas include the mod-

ern forms N. albigu/a and N. micropus. These two
species occur today in grassland and desert situa-
tions, usually with brush, cacti, or rocky areas avail-
able. They show local differences in habitat pref-
erence, but occur in close physical proximity. N.
micropus becomes rare to the west now, apparently
not reaching west as far as the New Mexico-Arizona
border.

The other two species in the interstadial faunas
appear to represent undescribed, extinct species. One,
the Species A of this study, seems to be allied with
N. mexicana or N. cinerea , more probably the latter.
A reasonable scenario would be isolation in the
southeastern New Mexican moderate to high ele-
vations following an earlier stadial expansion of N.
cinerea. This would allow a period of differentiation
during isolation, perhaps even from early Wisconsin
stadial conditions. Late Pleistocene re-expansion of
N. cinerea could then cause extinction by compe-
tition or genetic swamping.

The other species (Species B) seems to have its
affinity with the N. goldmani-lepida group. The rar-
ity of N. goldmani in collections and the large geo-
graphic variation seen in modern nominal N. lepida
make assessment of exact relationships difficult.
Species B might well be ancestral to N. goldmani or
both N. goldmani and N. lepida , or of course be a
distinct taxon becoming extinct without leaving is-
sue.

By mid-stadial times, the woodrat fauna had
changed considerably. An N. goldmani- like form
occurs at Dry Cave, a possible descendant of Species
B (the occurrence of the goldmani- type woodrat at
Anthony Cave, north of El Paso on the Texas-New
Mexico border, is undated, but apparently full-sta-
dial, at least in part). N. albigula remains are wide-
spread, but N. cinerea of modern character now is
the most commonly represented species in the cave
faunas. As the complete faunas make clear (Harris,
in press), some of the sites considered here were
below the coniferous forest zone, lying in sagebrush-
grasslands or steppe-woodlands. Notably cooler
summers, cold winters (but probably not to the ex-
tremes of today), and more effective moisture al-
lowed invasion of a vast area by the bushy-tailed
woodrat, even possibly to southern Chihuahua.

Changing conditions near the end of the Pleis-
tocene allowed invasion by N . floridana into the Dry
Cave area, probably in the context of better grass-
land habitat. Neotoma cinerea still is present, how-
ever, as are other species of mammals now extir-
pated from the lower elevations of southern New
Mexico (Harris, 1977). Jimenez Cave, in southern

178

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Chihuahua, may show presence of N. floridana, but
the age is not definitely known to be late stadial.

Neotoma micropus may have been present during
full stadial times (UTEP 6, cf.; UTEP 4, ?), but is
not reasonably surely present until latest Pleistocene
(UTEP 54). It also occurred at Jimenez Cave. As
with N. floridana, it may denote increasing summer
temperatures and possibly increased emphasis on
summer precipitation.

At the close of the Pleistocene, a major change
occurs— A. cinerea abruptly disappears, replaced by
N. mexicana (this assumes presence of N. cinerea
at Pratt Cave is Pleistocene in age). With the ex-
ception of N. mexicana, the Neotoma fauna is mod-
ern. Midden evidence indicates the early Holocene
maintained woodland in the lowlands of the South-
west until at least 8,000 B.P. (Van Devender and

Spaulding, 1979). It seems evident that A. mexicana
could replace A. cinerea under those conditions,
either by out-competing it or by moving in as A.
cinerea succumbed to other factors. A. mexicana
appears equipped to survive under such conditions,
hanging on even today in a few jumbled-rock areas
far below its more common elevational and vege-
tational range (Findley et al., 1975).

Of the original list of species considered, only A.
stephensi, now occurring from western New Mexico
across central and northern Arizona, has not been
identified. With the present study as a base, it can
be hoped that examination of Neotoma from other
sites in the western United States and Mexico can
clarify the geographic and chronologic distribution
of late Pleistocene Neotoma.

ACKNOWLEDGMENTS

I wish to thank T. Yates and W. Barber of the Museum of
Southwestern Biology, The University of New Mexico; D. Hoff-
meister of the Illinois Natural History Museum, The University
of Illinois at Urbana-Champaign; and R. Fisher of the Fish and
Wildlife Service Museum Section, National Museum of Natural
History, for the loan of modem comparative material. Other

comparative material was collected under permits from the New
Mexico Department of Game and Fish. Many of the fossil spec-
imens reported on here were collected under a grant from the
National Geographic Society with the permission of the Bureau
of Land Management. H. Messing allowed study of Neotoma
from the Jimenez Cave fauna, currently under study by him.

LITERATURE CITED

Anderson, S. 1972. Mammals of Chihuahua, taxonomy and
distribution. Bull. Amer. Mus. Nat. Hist., 148:149-410.

Armstrong, D. M. 1972. Distribution of mammals in Colo-
rado. Mus. Nat. Hist., Univ. Kansas Monogr., 3:1-415.

Cary, M. 1917. Life zone investigations in Wyoming. N. Amer.
Fauna, 42:1-95.

Dalquest, W. W., E. Roth, and F. Judd. 1969. The mammal
fauna of Schulze Cave, Edwards County, Texas. Bull. Florida
State Mus., Biol. Sci., 13:205-276.

Findley, J. S., A. H. Harris, D. E. Wilson, and C. Jones. 1975.
Mammals of New Mexico. Univ. New Mexico Press, Al-
buquerque, 360 pp.

Harris, A. H. 1977. Wisconsin age environments in the north-
ern Chihuahuan Desert: Evidence from the higher verte-
brates. Pp. 23-52, in Transactions of the symposium on the
biological resources of the Chihuahuan Desert Region, United
States and Mexico (R. H. Wauer and D. H. Riskind, eds.),
Natl. Park Serv. Trans. Proc. Ser., 3:1-658.

. 1980. The paleoecology of Dry Cave, New Mexico.

Natl. Geogr. Soc., Res. Rep., 12:331-338.

. In press. Late Pleistocene paleoecology of the West.

Univ. Texas Press, Austin.

Harris, A. H., and C. R. Crews. 1983. Conkling’s roadrun-
ner— a subspecies of the California roadrunner? Southwest-
ern Nat., 28:407^112.

Harris, A. H., and J. S. Findley. 1964. Pleistocene-Recent
fauna of the Isleta Caves, Bernalillo County, New Mexico.
Amer. J. Sci., 262: 1 14-120.

Lundelius, E. L., Jr. 1979. Post-Pleistocene mammals from
Pratt Cave and their environmental significance. Pp. 239-
258, in Biological investigations in the Guadalupe Moun-
tains National Park, Texas (H. H. Genoways and R. J. Baker,
eds.), Natl. Park Serv. Proc. Trans. Ser., 4:1-442.

Murray, K. F. 1957. Pleistocene climate and the faunal record
of Burnet Cave, New Mexico. Ecology, 38:129-132.

Van Devender, T. R., and W. G. Spaulding. 1979. Devel-
opment of vegetation and climate in the southwestern United
States. Science, 204:701-710.

Van Devender, T. R., and F. M. Wiseman. 1977. A prelimi-
nary chronology of bioenvironmental changes during the
paleoindian period in the monsoonal Southwest. Pp. 13-27,
in Paleoindian lifeways (E. Johnson, ed.), West Texas Mus.
Assoc., The Mus. J., 17:1-197.

Van Devender, T. R., K. B. Moodie, and A. H. Harris. 1976.
The desert tortoise ( Gopherus agassizi) in the Pleistocene of
the northern Chihuahuan Desert. Herpetologica, 32:298-
304.

Address: Laboratory for Environmental Biology, University of Texas at El Paso, El Paso, Texas 79968.

THE EVOLUTION OF COTTON RAT BODY MASS

Robert A. Martin

ABSTRACT

Body mass of fossil Sigmodon is estimated by an equation
generated from a relationship between length of the first lower
molar and body mass in extant cricetines. Replacement sequences

of fossil cotton rats in four limited geographical regions of the
continental United States manifest the Cope-Deperet Rule. Ex-
planatory hypotheses are briefly examined.

INTRODUCTION

Body mass (or weight) of mammals is the key to
estimating other meaningful biological parameters
in extinct species. Mass is highly correlated in living
species with metabolic rate (Kleiber, 1961; Martin,
1980), home range (McNab, 1963; Harestad and
Bunnell, 1979), population density (Martin, 1981;
Damuth, 1981), and other variables. As I showed
in an earlier paper (Martin, 1980), certain morpho-
metric variables are tightly correlated with body mass
in mammals and therefore can be mathematically
manipulated to estimate any other variables cor-
related with mass. This simple procedure will allow
paleobiologists to test new hypotheses dealing with
mammalian extinction, morphological evolution,
and the evolution of faunal energetics.

The purpose of this essay is to provide some base-
line information for cotton rats, genus Sigmodon,

and to speculate on the evolution of body size in
this genus. This paper is part of a larger project
which will examine the overall evolution of cotton
rat energetics. In an earlier treatment (Martin, 1980)
I had used two morphometric variables, the greatest
width across the occipital condyles and the greatest
width of the femoral head, to estimate body mass.
As skulls are rare paleontologic finds, and femora
not always available or readily identifiable, another
measurement, length of the first lower molar (M,)
was used in this study. First lower molars are com-
mon in deposits containing Sigmodon, and are also
taxonomically useful (Martin, 1979). As will be
demonstrated, the length of this tooth is highly cor-
related with body mass in Sigmodon and other cri-
cetine rodents.

MATERIALS AND METHODS

Length of the M, (first lower molar) was measured with an
American Optical (AO) filar micrometer eyepiece coupled to a
Spencer binocular dissecting microscope. The eyepiece was cal-
ibrated with an AO two millimeter micrometer slide. Measure-
ments were taken from 33 individuals representing six cricetine
species. Mass (W) data were taken from skins associated with
each mandible. Species, number of individuals, range of M , length,
and range of observed W are as follows: Reithrodontomys hu-
mulis (1; 1.27 mm; 7.1 g), Peromyscus floridanus (2; 1.87-2.02
mm; 40.3-43.6 g), P. gossypinus (3; 1.58-1.72 mm; 17.4-19.5

g), P. polionotus (6; 1.30-1.39 mm; 9.1-18.3 g), P. leucopus (10;
1 .38-1 .70 mm; 1 1 .6-24.0 g), Sigmodon hispidus from the Florida
peninsula (11; 2.44-2.72 mm; 79.8-120.6 g). A least squares
straight line was fitted to the logarithm of W plotted against the
log of M, length (Fig. 1). Body mass data were then calculated
for extinct and extant cotton rat species based on the equation
that was generated. With the exception of the S. hispidus sample
noted above and a sample of Sigmodon medius from the Blanco
fauna, these calculations are made from measurements published
by Martin (1979).

RESULTS AND DISCUSSION

The relationship between M , length and body mass
in cricetines is illustrated by the least squares line
in Fig. 1. The equation for this line, converted to
its exponential form, is

W = 4.05L3-33 (1)

where L = length of the first lower molar in milli-

meters and mass is in grams. The two variables are
highly correlated (r = 0.98) and the 95% confidence
limits of /3 are calculated as 3.33 ± 0.35. Ninety-
five percent confidence limits of antilog y are 9.0 ±
1.1 (at minimum x), 30.0 ± 1 . 1 (at x), and 111.4 zt
1.1 (at maximum x). The confidence interval con-
tracts from 12.5 percent of y at minimum x to one

179

180

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Length Mi
(mm)

Fig. 1— Least squares regression line running through scatter
diagram of body mass plotted against length of the first lower
molar in extant cricetines.

percent when W > 100 g. Predicted values of body
mass for Sigmodon, in the upper levels of mass,
average less than 10% (usually closer to 1%) different
from values observed on specimen tags. The average
mass of S. hispidus predicted by equation (1) from
the first lower molars measured in this study was
101.4 g. The observed mean value was 101.6 g.

Body mass data for extinct and extant Sigmodon
are presented in Table 1 and Fig. 2. Within all of
the Sigmodon replacement chronologies ( not phy-
togenies) illustrated in Fig. 2, there is a general trend
towards increased body size. Reversals are seen in
the S. medius to Sigmodon minor lineage in Kansas
and Arizona, and from the Inglis IA Sigmodon cur-
tisi to Haile XVIA Sigmodon libitinus in Florida,
but these smaller animals are subsequently replaced
by larger species.

Based upon mass data generated by equation ( 1 ),
living Sigmodon species average around 78 g (range
of averages; 58-108 g). This level of average body
size was attained during the early Irvingtonian land
mammal age (approximately 1.8 million years ago)
in the southwest and Great Plains regions by Sig-
modon hudspethensis and S. curtisi. Nevertheless,

Table 1 . — Body mass of extinct and extant species q/'Sigmodon.
Mass data estimated by equation (1) in text.

Taxon

N

Mean
length
of first
lower
molar
(mm)

Mean

body

mass

(g)

Observed
range of
body mass (g)

Extinct

Sigmodon medius 1
Sigmodon minor

90

2.06

45.0

28.7-63.9

Curtis Ranch,
Borchers)

47

1.90

34.4

24.7-55.1

Sigmodon curtisi 2
Sigmodon libitinus

19

2.38

72.8

52.6-84.5

(Haile XVIA)

26

2.00

51.0

28.7-77.9

Sigmodon hudspethensis
(Red Light, Hudspeth)
Sigmodon bakeri

4

1.83

68.7

55.1-95.1

(Coleman IIA,
Williston IIIA)

23

1.84

69.0

43.5-98.8

Extant

Sigmodon hispidus
(Florida Recent)
Sigmodon hispidus

1 1

2.63

101.4

78.8-113.4

(Reddick I A)

18

2.49

84.5

59.4-1 13.4

Sigmodon hispidus 3

8

2.35

69.7

54.3-84.5

Sigmodon ochrognathus

5

2.26

61.2

47.9-84.5

Sigmodon fulviventer

8

2.27

62.1

55.9-73.7

Sigmodon alleni

5

2.41

75.8

62.1-93.9

Sigmodon mascotensis

3

2.22

57.7

54.3-60.3

Sigmodon arizonae

4

2.45

80.1

76.8-83.4

Sigmodon leucotis

9

2.54

90.3

79.0-1 17.6

Sigmodon peruanus

9

2.68

107.9

65.8-162.4

1 Mean of means from the following localities: Benson, Tusker, Rexroad Loc. 3,
Sanders, Sand Draw, Haile XVA, Wendell Fox Pasture, Blanco.

2 Curtis Ranch, Kentuck, Inglis IA combined.

3 S. h. berlandierf a small extant subspecies from Texas and Mexico.

for those areas from which we have a replacement
chronology, the trend towards increased size con-
tinued; albeit with the reversals noted above. I have
previously suggested (Martin, 1 979) that the smaller
average size of S. minor relative to S. medius is the
result of ecological separation (character displace-
ment) away from the larger, more progressive S.
curtisi with which it was contemporaneous.

Why is it that the fossil history of Sigmodon, like
so many other mammalian phylogenies, demon-
strates evolution towards increased size? This ten-
dency is often referred to as Cope’s Rule, although
Stanley (1973) indicates that it was first explicitly
stated by Deperet (1909). Two hypotheses have been
presented. The first, which we may label as Xhe fun-
damental advantage hypothesis, suggests that, all
other things being equal, it is more beneficial to be
large than small. Some of these advantages are listed

1984

MARTIN — SIGMODON BODY MASS

81

Fig. 2. — Replacement chronologies of cotton rats, comparing the evolution of body mass in Sigmodon with the distribution of body
mass seen in modem species. The horizontal lines within each chronology separate land mammal ages. B = Blancan, I = Irvingtonian.
Ra = Rancholabrean, Re = Recent. Solid circles represent fossil deposits; open circles are areas from which samples of extant species
only were captured. The solid black line on the United States map depicts the present limits of cotton rat distribution. Numbers within
illustrations are average mass values for the samples. At each locality, replacement sequences are listed from oldest to youngest sample.
1 = northcentral Florida: S. medius (Haile XVA ) — S. curtisi (Inglis IA)— S. libitinus (Haile XVIA)— S', bakeri (2 samples; Coleman 1IA
and Williston IIIA) — S. hispidus (2 samples; Reddick IA and Recent). 2 = Meade Basin: S', medius (Rexroad Loc. 3) — S. medius
(Sanders) — S. minor (Borchers)— cf. S. curtisi (Kentuck)— S. hispidus (Recent). 3 = Vallecito-Fish Creek Beds: S. medius (Layer Cake)—
S. medius (Arroyo Seco)— S. medius (transition zone between Arroyo Seco and Vallecito Creek faunas) — cf. S. curtisi (Vallecito Creek).
4 = San Pedro Valley: S. medius (Benson and Tusker combined)— S. minor (36 g; Curtis Ranch) and S. curtisi (76 g; Curtis Ranch) —
S. arizonae (Recent). 5 = S. fulviventer. 6 = S. leucotis. 7 = S. alleni. 8 = S. hispidus berlandieri. 9 — S. mascotensis. 10 = S. peruanus,
a large South American species. Not included on this illustration is S. (Sigmomys) alstoni, an average-sized South American Sigmodon
with grooved upper incisors.

182

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

by Stanley (1973), and include . . improved abil-
ity to capture prey or ward off predators, greater
reproductive success, increased intelligence (with
increased brain size), better stamina, expanded range
of acceptable food, decreased annual mortality, ex-
tended individual longevity, and increased heat re-
tention per unit volume.” The second hypothesis,
proposed by Stanley (1973), can be for convenience
referred to as the structural limitation hypothesis.
It is a probabilistic model which can be summarized
as follows: (1) large species of a given body plan are
more specialized than are small ones, (2) large scale
adaptive breakthroughs occur at small body size,
and (3) the tendency of taxa in a lineage to evolve
larger size represents movement towards an opti-
mum size for the adaptive zone being filled, not
evolution away from small body size.

According to this hypothesis, size increase is not
innately favored as a particular adaptive strategy,
but simply represents the normal inertial sequence
during diversification. By this principle, the dimin-
utive Sigmodon minor and S. libitinus are no more
special than the larger species which appeared before
and after them (unless, as small species, they were
ancestral to larger, more advanced ones!). Van Valen
(1975, 1978) implies that the structural limitation
hypothesis should pertain only to the start of a trend
towards large size, not to its maintenance. That is,
why does a trend towards large size continue in a
phylogeny after the initial stages of diversification?
If body size is not being regulated, should not there
be an equal tendency towards dwarfing or gigantism
in a given lineage following an initial explosion of
taxa? 1 am not certain that this follows. It would
seem to depend on how close a species is to an
optimum size for its new adaptive zone. Stanley
(1973) does not deny that selection may continue
to favor large size or that large size is not adaptive
(for one or more of the reasons listed as fundamental
advantages); merely that dwarfing is not nearly as
likely as size increase because of the structural con-
straints associated with largeness.

The late Pliocene to Recent evolution of cotton
rats, spanning a time period of about four million
years, is clear in one regard — each local replacement
chronology moves inexorably towards animals of
increased average body mass (Fig. 3). Further, dwarf-
ing appears to occur only just prior to and during
invasion and replacement by larger species occu-
pying similar adaptive zones. I do not know if this
is a general trend for other mammalian groups, but
it should be examined. The results of this study

Time

Fig. 3.— Change of Sigmodon body mass through late Pliocene
and Pleistocene time. W = mass, B = Blancan, I = Irvingtonian,
Ra = Rancholabrean, Re = Recent. Solid circles = Meade Basin,
Open circles = Vallecito-Fish Creek Beds, solid stars = north-
central Florida, open squares = San Pedro Valley. See Fig. 2 and
text for further detail.

generally support Stanley’s (1973) structural limi-
tation hypothesis. A single species, Sigmodon medi-
us, exists unchanged for almost two million years
in pastoral habitats on the central Great Plains and
southern United States from coast to coast. Towards
the end of Blancan time it undergoes a relatively
rapid dwarfing throughout most of its range. By the
end of the Blancan it is sufficiently diminished in
size to be recognized as other species, S. minor. At
least in Arizona, it is seen briefly cohabiting with
the advanced Sigmodon curtisi. From late Blancan
through Irvingtonian time two or three Sigmodon
species are recognizable during a given temporal
interval. Four species— V. hispidus, S. fulviventer,
S. ochrognathus, and S. ari zonae— are recognized
by neomammalogists as extant in these areas today.
It would be very difficult to separate dentitions of
these species if interred in the same accumulation,
and I suspect that only two or three would be ac-
cepted as valid paleospecies. One could argue that,
having made a behavioral-anatomical breakthrough
in changing from a granivorous, browsing mode of
foraging to grazing, a trend of evolution towards
larger size then began, continuing into present time.
This increase in body size was complemented by an

1984

MARTIN -SIGMODON BODY MASS

183

increase in dental hypsodonty and the nature by
which the tooth crown is supported (see Martin,
1979, for details). It is likely that an optimum size
for Sigmodon has yet to be reached. Some of the
dental specializations seen as variants in S. hispidus
(such as lamination and root capture) suggest to me

that this group is moving structurally in a direction
reminiscent of the large Central and South Ameri-
can hystricomorphs. While Stanley’s structural lim-
itation hypothesis may seem to be a tautology (that
is, today’s specialization may be tomorrow’s gen-
eralization), it can be supported by available data.

LITERATURE CITED

Damuth, J. 1981. Population density and body size in mam-
mals. Nature, 290:699-700.

Deperet, C. 1909. The transformations of the animal world.
D. Appleton and Co., New York, 360 pp.

Harestad, A. S., and F. L. Bunnell. 1979. Home range and
body weight — a reevaluation. Ecology, 60:389-402.

Kleiber, M. 1961. The fire of life. Wiley, New York, 454 pp.

Martin, R. A. 1979. Fossil history of the rodent genus Sig-
modon. Evol. Monogr. 2:1-36.

— . 1980. Body mass and basal metabolism of extinct mam-

mals. Comp. Biochem. Physiol., 66A:307-314.

1981. On extinct homimd population densities. J. Hu-
man Evol., 10:427-428.

McNab, B. K. 1963. Bioenergetics and the determination of
home range size. Amer. Nat., 97:133-140.

Stanley, S. M. 1973. An explanation for Cope’s Rule, Evo-
lution, 27: 1-26.

Van Valen, L. 1975. Group selection, sex, and fossils. Evo-
lution, 29:87-94.

. 1978. [Review of] Patterns of evolution, illustrated by

the fossil record, edited by A. Hallam. Paleobiology, 4:21 0—
216.

Address: Department of Biological and Allied Health Sciences, Fairleigh Dickinson University, Madison,
New Jersey 07940.

BIOSTRATIGRAPHY AND BIOGEOGRAPHY OF QUATERNARY
MICROTINE RODENTS FROM NORTHERN YUKON
TERRITORY, EASTERN BERINGIA

Richard E. Morlan

ABSTRACT

Recent work in northern Yukon Territory has produced sam-
ples of microtine rodent fossils from well documented strati-
graphic contexts spanning much of late Pleistocene and Holocene
time. Clethrionomys, Microtus, Lemmus, and forms of the Di-
crostonychini are present throughout the sequence which is also
punctuated by two appearances of Phenacomys. Evolutionary
changes can be recognized within the Dicrostonyx lineage, the
fossils of which are analyzed in terms of morphotypes defined
by Agadjaman and von Koenigswald (1977). Vole fossils bearing
similarities to Microtus paroperarius are labelled “ Microtus sp.

X” and are separated from both the former and the morpholog-
ically similar tundra vole, M. oeconomus, by means of measure-
ments proposed by van der Meulen (1973, 1978).

The Yukon sequence is arranged in terms of intervals defined
by Hopkins (1982), and possible correlations with Alaskan and
Siberian sediments and faunas are discussed. The Beringian mi-
crotine record is seen to be woefully incomplete, poorly dated,
and comprised of samples that are often very small. Nonetheless,
some tentative proposals are put forth concerning biostratigraphy
and biogeography, and gaps in the record are identified.

INTRODUCTION

The pivotal role of Beringia in the formation of
Holarctic biota is well known (see Hopkins et al.,
1982, for a recent overview). Not only has this re-
gion been an important pathway between the Pale-
arctic and Nearctic, but also it may have been a
center of evolution for many taxa. Of practical im-
portance to paleontologists is the fact that much of
Beringia, from the Kolyma Lowland of Siberia to
the Mackenzie valley of northwestern Canada (Fig.
1), was not glaciated and therefore offers opportu-
nities to study the evolution of faunal and floral
communities over long periods. The non-glacial and
periglacial environments of Beringia served as re-
fugia for both plants and animals during glacials and
were sources for the dispersal of life forms following
deglaciation (Hughes et al., 1981:330).

For several reasons, however, primary vertebrate
paleontological data have accumulated slowly and
have seldom been adequate for defining long se-
quences of faunas in datable stratigraphic contexts.
The permafrost that confers excellent preservational
environments in much of Beringia also inhibits ac-
cess to Pleistocene sediments. Many areas are re-
mote, accessible only by boat or helicopter, and are
logistically difficult or very expensive to study. In
central Alaska and Yukon, most vertebrate remains
have been recovered from perennially frozen col-
luvial silts (“mucks”) exposed in placer mines where
good stratigraphic controls are difficult to maintain
and where concentrations of microfaunal remains
are often low (for example, Guthrie, 1968:Fig. 5).

Only recently have fossil remains been recovered
from other kinds of commercial excavations such
as roadcuts (for example, Weber et al., 1981).

In northern interior areas of Siberia, Alaska, and
Yukon, many streams have exposed fossiliferous
sediments along their banks, and these have begun
to produce small samples of vertebrate remains in
primary stratigraphic contexts. One of the most pro-
ductive streams is Old Crow River, a northern Yu-
kon tributary of Porcupine River, where large con-
centrations of vertebrate fossils on the modern sand
and gravel bars reflect the presence of abundant fos-
sils yet to be eroded or excavated from the river
banks (Harington, 1977, 1978:62-63; Morlan,
1980#, 1 9806; Jopling et al., 1981:table 2). Work in
this area has only begun to reveal the potential for
such studies. The samples are too small, too poorly
dated, and too few in number to support a thorough
paleoecological analysis, but they are adequate to
define taxonomic, biostratigraphic and biogeo-
graphic problems and to suggest a few solutions to
some of the problems.

In addition to research in the lowlands, excava-
tions at an upland site, the Bluefish Caves, have
produced large samples of late Pleistocene and Ho-
locene vertebrate remains (Cinq-Mars, 1979, 1982;
Morlan and Cinq-Mars, 1982; Morlan, 1983#).
These samples will be mentioned briefly in order to
represent the critical millennia at the close of the
Pleistocene, and their paleoecological implications
will be explored elsewhere.

184

1984

MORLAN- EASTERN BERINGIAN MICROTINES

185

Fig. 1. — Map of northern Yukon Territory showing localities mentioned in the text. See inset for other Beringian localities.

186

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 1 .—Summary of taphonomic variables and references to associated fossils for microtine rodent samples from the northern Yukon

Territory. Letters at left margin refer to Fig. 2.

Samples

Enclosing sediment

Redeposition

Recovery techniques

References to associated fossils

Recent

K.

Klo-kut

Overbank allu-

Nil

Trowelling

Morlan, 1973

vium

J.

Bluefish Caves, Holo-

Organic-rich rub-

Nil

Trowelling, dry

Cinq-Mars, 1979, 1982; Morlan

cene

ble

and wet screen-

and Cinq-Mars, 1982; Ritchie

ing

et al., 1982

Duvann

y Yar Interval

I.

Bluefish Caves, Pleisto-

Loess

Nil

Trowelling, dry

Cinq-Mars, 1979, 1982; Morlan

cene

and wet screen-

and Cinq-Mars, 1982; Ritchie

ing

et al., 1982

Boutellier Interval

H.

HH62-228, “Unit 2b“

Lacustrine

Unknown

Unknown

Delorme, 1968; Harington,

1977, 1978; Lichti-Federov-
ich, 1974; Matthews, 1975;
McAllister and Harington,
1969

G.

HH68-10, Unit 2b

Loess{?)

Nil

Bulk removal,

(None)

handpicked

F.

Loc. 12, Disconform-

Overbank allu-

Significant

Trowelling, wet

Jopling et al., 1981 (Fauna 7);

lty A

vium

screen (1.6
mm, 50 liters)

Morlan, 1981; this report

E.

Loc. 15, Disconform-

Overbank allu-

Significant

Trowelling, wet

Bobrowsky, 1982; Cumbaa et

ity A

vium

screen (1.6

al., 1981; Janssens, 1981;

mm, 560 liters)

Lichti-Federovich, 1973;
Morlan, 1980a, 1980ft; Mor-
lan and Matthews, 1983

pre- Happv Interval

D.

Loc. 1 5, Unit 2a

Point bar(?) allu-

Significant

Trowelling

Cumbaa et a!., 1981; Morlan,

vium

1980a, 1980ft; this report

C.

Loc. 1 1, Unit 2a

Point bar alluvium

Significant

Trowelling, wet

Jopling et al., 1981 (Fauna 6);

screen (1.6
mm, 300 liters)

this report

B.

Loc. 12, Unit 2a

Point barf?) allu-

Significant

Trowelling, wet

Jopling et al., 1981 (Fauna 2)

vium

screening

A.

Locs. 44, 45, 64,

Channel-bottom

Significant

Trowelling, hos-

Harington, 1977, 1978; Lichti-

Unit 2a

alluvium

ing, wet screen-

Federovich, 1973; Matthews,

mg

1975

This paper will focus primarily on the Old Crow
Basin of northern Yukon Territory and will present
new data on microtine rodents from several strati-
graphic and taphonomic contexts. Many of the
northern Yukon localities have produced wide va-
rieties of fossils in addition to the microtine rodents

(see Table 1 for references). The multi-disciplinary
research needed to integrate the northern Yukon
fossil record has already produced a substantial vol-
ume of published material (summaries in Hughes
etal., 1981:331; Joplinget al., 1981), and additional
reports are in preparation.

METHODS

Recovery techniques are listed in Table 1 with mesh size of
screen and volume of sample shown in parentheses when known.

At the Bluefish Caves, Cinq-Mars (1979:5) recovered microtine
rodents by trowelling and dry screening all matrix on a 3 mm

1984

MORLAN-EASTERN BERINGIAN MICROTINES

187

mesh (ca. 4 mm diagonal sieve openings) and by wet screening
(0. 1 8 mm sieve openings) ca. 5 kg of residue that passed through
the coarse screen from each excavation level; many other spec-
imens will be recovered in the laboratory from bulk samples that
have not yet been analyzed.

Identifications were made by means of direct comparisons with
reference specimens of extant taxa and by means of literature
searches for extinct taxa. Drawings of many molars were made
at approximately 15 x with a drawing tube on a Wild-Leitz M5
microscope. The drawing tube facilitated reconstruction of some
fragmentary teeth.

Locality designations on Old Crow River were assigned by
Hanngton (1977) except for HH68-10, studied by O. L. Hughes

who also designated HH62-228 on Porcupine River. Specimens
from Locs. 44, 45, 64 and HH62-228 are kept in the Paleobiol-
ogy Division, National Museum of Natural Sciences, Ottawa,
and are catalogued with five-digit numbers following the prefix
“NMC-.” Specimens reported by Jopling et al. (1981) are cur-
rently housed in the Department of Anthropology, University of
Toronto, and have not been examined for the purposes of this
report. All others are kept in the Archaeological Survey of Can-
ada, National Museum of Man, Ottawa, under catalogue desig-
nations based on the Borden system. The catalogue numbers
appear in parentheses in the following list: Loc. 1 1 (MkVI-9);
Loc. 12 (MkVl-10); Loc. 15 (M1V1-2); HH68-10 (M1V1-8); Blue-
fish Caves (MgVo-1, MgVo-2); Klo-kut (MjVI-1).

GROSS STRATIGRAPHY AND CHRONOLOGY

Microtine rodent samples have been recovered
from seven localities in Old Crow Basin, two lo-
calities in Bluefish Basin, and two caves in the up-
lands immediately south of Bluefish Basin (Fig. 1).
None of these areas was glaciated during the Pleis-
tocene, but all were indirectly affected by glaciers
that reached the eastern flanks of Richardson Moun-
tains. During late Wisconsinan time, meltwater from
Laurentide ice that occupied Bonnet Plume Basin
and McDougall Pass created large lakes in Old Crow,
Bluefish, and Bell basins (Hughes et al., 1981). As
the lakes drained, silt from the newly exposed lake
bottom in Bluefish Basin supplied loess that accu-
mulated in the Bluefish Caves to the south (Cinq-
Mars, 1979, 1982). The drainage of the lakes took
place through a canyon newly cut by Porcupine Riv-
er through bedrock near the Alaska/Yukon border
(Hughes, 1972; Thorson and Dixon, 1983), and the
lower base level created by downcutting enabled
Porcupine River and its tributary, Old Crow River,
to dissect thick sequences of Quaternary sediment
in Bluefish and Old Crow basins, respectively.

Exposures along Old Crow River can be charac-
terized stratigraphically in terms of four gross units
(see Hughes, 1972; Morlan and Matthews, 1978;
Morlan, 1979, 1980a, 19836): Unit 1, a lacustrine
silty clay at the base, largely concealed by the river;
Unit 2, approximately 18-20 m of sand, silt, and
clay representing an interlacustrine sequence con-
sisting primarily of alluvium, subdivided into Units
2a and 2b at some sections; Unit 3, approximately
5 m of glaciolacustrine clay; and Unit 4, silt and
peat, usually about 1 -3 m thick, that forms the mod-
ern surface (Fig. 2). Some of these units can be re-
lated to a series of intervals defined by Hopkins
(1982) as a framework for integrating Beringian stra-

tigraphy, chronology, and paleogeography. For ex-
ample, the glaciolacustrine clay of Unit 3 represents
the Duvanny Yar Interval and can be correlated
with various regional late Wisconsinan glacial ad-
vances in Beringia as well as with oxygen isotope
stage 2 (Hopkins, 1982:fig. 2; Shackleton and Op-
dyke, 1973). A radiocarbon date of 25,170 ± 630
B.P. (NMC-1232) on a proboscidean tusk only 50
cm below the base of the clay in Bluefish Basin
provides a maximum age for glacial lake formation
in both basins. A minimum age for drainage of the
lakes is provided by radiocarbon dates of approxi-
mately 12,000 B.P. on Bison crassicornis bones re-
covered from a channel incised into the top of the
lake clay in Old Crow Basin (Harington, 1977:827-
834, table 5, 1978:55; Morlan, 1980a:fig. 9.3).

As the glacial lakes drained, the newly exposed
lake bottom in Bluefish Basin supplied loess that
buried bones and artifacts in the Bluefish Caves, but
dates as early as 18,000 years ago at the latter site
(Cinq-Mars, personal communication, 1982) sug-
gest that the level of the glacial meltwater lake was
lowered considerably or episodically during the last
millennia of its existence.

Where Unit 2 can be subdivided, the boundary
between units 2a and 2b is marked by an erosional
contact informally known as Disconformity A. Ra-
diocarbon dates on Unit 2b range from >51,000
B.P. (GSC-2559-2) on an autochthonous peat near
the base of the unit in Old Crow Basin to 25,170
B.P. (on the proboscidean tusk mentioned above).
The erosional contact was created after an episode
of climatic warming that thawed ice wedges and
replaced them with sediment-filled pseudomorphs
that are truncated by Disconformity A. The discon-
formity and Unit 2b together represent middle Wis-

188

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 2. — Schematic stratigraphic profile of Quaternary sediments in Old Crow Basin, northern Yukon Territory (after Hughes, 1972;
Morlan and Matthews, 1978; Morlan, 19837>). Circled letters (A-G) mark approximate positions of microtine rodent samples from Old
Crow Basin. Letters outside the column (H-K) refer to samples from Bluefish Basin and Bluefish Caves that are correlated with Old
Crow Basin on the basis of radiocarbon dates. Roman numerals at left edge of column refer to pollen assemblage types defined by
Lichti-Federovich (1973). Lines in Unit 2a represent bedding planes, and “xxx” marks the position of Old Crow tephra. V-shaped
marks at the upper contacts of Units 2a and 2b represent ice wedge pseudomorphs. Named intervals are after Hopkins (1982).

consinan time, the Boutellier Interval of Hopkins
(1982), approximately equivalent to oxygen isotope
stage 3.

Presumably the cold Happy Interval, to which

many regional Beringian glacial advances are as-
signed (Hopkins 1982:fig. 2), is represented in Old
Crow Basin by the growth of ice wedges mentioned
above. The Happy Interval is the oldest named in-

1984

MORLAN- EASTERN BERING1AN MICROTINES

189

terval in Hopkins’ framework and is approximately
equivalent to oxygen isotope stage 4. The lower
boundary of the interval is not yet defined, but it
includes Old Crow tephra (Hopkins, 1 982:6). At five
sections in Old Crow Basin, Old Crow tephra has
been found just below Disconformity A, and fission
track analysis indicates that this tephra is
< 120,000 years old (Naeser et ah, 1982; Westgate
et ah, 1983). Old Crow tephra has also been found
in southern Yukon and at several Alaskan localities
(Westgate, 1 982), and Schweger and Matthews (n.d.)
have collated paleoenvironmental and chronomet-
ric evidence from several of these sites to narrow
the possible time of deposition of the tephra to be-
tween 87,000 and 105,000 years B.P. Although this
range does not identify the “instant” of deposition,
it suggests that Old Crow tephra was deposited
sometime during oxygen isotope stage 5 (Shackleton
and Opdyke, 1973:table 3).

Ten or more meters of Unit 2a alluvium underlie
Old Crow tephra at each of the sections where the
tephra has been observed, but variable amounts of
time may be represented by such sediment thick-
ness. Some sections expose the dipping beds of point
bars that could have accumulated rapidly, but in
other exposures all the bedding is finely laminated
and flat-lying and could represent a long time pe-
riod. Some of the fossils, such as spotted skunk,
from the lower part of Unit 2a suggest an age within
the Sangamon Interglaciation (Harington, 1978:62-
63) equivalent to some part of isotope stage 5e and
therefore older by definition than the Happy Inter-
val. Even greater age is implied by four uranium
series dates (160,000 years ago or more) on bones
from approximately the middle of Unit 2a (Morlan,
1983/>:table 4), and some of the microtine rodent
remains to be described below are suggestive of a
longer time scale than has usually been contem-
plated for the sequence exposed along Old Crow
River.

Obviously, no precise age can be given for the

Unit 1 clay at the base of the Old Crow sections. A
thick lacustrine unit in Bluefish Basin is reversely
magnetized and “probably lies within the Matuya-
ma Epoch or an earlier epoch” (Pearce et al., 1982:
926). Hughes (1972) correlated this unit with Unit
1 in Old Crow Basin, believing both to be glacio-
lacustrine in origin. Subsequent discovery of high
percentages of pine pollen in the Bluefish Basin unit
(Lichti-Federovich, 1974) cast doubt on that mode
of origin, but no additional data are available either
to deny or to confirm the correlation with Old Crow
Basin Unit 1. Either a paleomagnetic study of the
Unit 1 clay or a date on the recently discovered
Little Timber tephra (just upsection from the clay)
could provide a critical limit on the chronological
framework for Old Crow Basin (Schweger and Mat-
thews, n.d.).

Detailed analysis of stratigraphy in the northern
Yukon basins is now in preparation, but the fore-
going outline is adequate for the purposes of this
paper. The gross stratigraphic context of each mi-
crotine rodent sample from Old Crow Basin is shown
by a letter (A-G) on the composite sketch in Fig. 2,
and samples from Bluefish Basin and Bluefish Caves
(H-K) are positioned outside the stratigraphic col-
umn on the basis of radiocarbon dating. Tapho-
nomic variables and references to associated fossils
are summarized in Table 1. It is noteworthy that
there are no samples to represent the Happy Inter-
val, because the fossils associated with Disconfor-
mity A are actually contained in the alluvium that
covers the erosional contact and therefore belong to
the Boutellier Interval. Because the glaciolacustrine
clay of Unit 3 is devoid of fossils, only upland sites,
such as Bluefish Caves, can provide a record for the
Duvanny Yar Interval. The Klo-kut archaeological
site is included here for two reasons: ( 1 ) as a lowland
counterpart to the Holocene sample from Bluefish
Caves; and (2) to correct misidentifications pub-
lished 10 years ago (for example, Dicrostonyx did
not occur there).

SYSTEMATICS AND MORPHOLOGY

Of the arvicoline (“microtine”) rodents known
from northern Yukon, the voles and lemmings are
described here, and muskrat ( Ondatra zibethicus)
specimens will be presented elsewhere. Systematic
study of the northern Yukon fossils is essentially
complete for most taxa but is in a preliminary stage
for some. Those requiring further study are fossils

of Microtus and Dicrostonyx that resemble extinct
species described from other areas. For these, the
analysis has been based entirely on published lit-
erature rather than on comparisons with original
fossil materials. Hence this is an interim report in
which current knowledge is summarized, and a more
detailed study is planned.

190

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 2. — Summary of microtine rodent remains from the northern Yukon Territory.

Sample

Cleth-

riono-

mys

cf.

rutilus

Phe-

nac-

omys

cf.

inter-

medins

Micro-
tus
sp. X

M.

xan-

tho-

gna-

thus

M.

miurus

M.

miurus

or

penn-

syl-

vani-

cus

M.

penn-

syl-

vani-

cus

M.

oeco-

no-

mus

Micro-

tus

spp.

Lem-

mus

sibir-

icus

cf.

Dicros-

tonyx

D.

D cf. guliel- D.
simpli- mitor- tor-
cior quatus quatus

Total

NISP

Recent

K. Klo-kut

18

\*

5

10

4

i

39

J. Bluefish Caves, Holocene

164

478

325

119

1,086

Duvanny Yar Interval

I. Bluefish Caves, Pleistocene

160

322

223

12

513

346

329

1,905

Boutellier Interval

H. HH62-228, “Unit 2b”'
G. HH68-10, Unit 2b

+

41

41

F. Loc. 12, Disconformity A2

+

+

+

+

+

?

F. Loc. 12, Disconformity A

3

2

3

15

4

27

E. Loc. 15, Disconformity A

7

i

6

7

6

5

20

102

33

187

pre Happy Interval

(Unit 2a, middle)

C. Loc. 1 1 , Unit 2a

16

22

39

5*

7

1 1*

85

194

42

421

C. Loc. 11, Unit 2a2

+

+

+ *

+

?

+

+

D. Loc. 15, Unit 2a

4

2

6

(Unit 2a, lower)

B. Loc. 12, Unit 2a2

+

+

?

+

+

A. Loc. 44, Unit 2a

1

6

2

4*

1

2*

29

17

62

A. Loc. 45, Unit 2a

u

1

1

3

A. Loc. 64, Unit 2a

8

8

9

8

33

1 Hanngton (1977:12 1-126); 2 Jopling et al. (1981 :Table 2); ? Identification questioned in this report; + Present; * Tentative species identification (cf.); letters at left margin refer
to Fig. 2. NISP. Number of Identified Specimens.

The distribution of northern Yukon microtine ro-
dent taxa is summarized in Table 2. It is believed
that Clethrionomys and Phenacomys are monotypic
in the northern Yukon, but no dental characters
have been found to distinguish the eastern Beringian
species from others that live in the temperate zone
(Rausch and Rausch, 1 975a; Guilday and Parmalee,
1972). Clethrionomys , the red-backed vole, is pres-
ent throughout the sequence but is rare in Pleisto-
cene deposits in Old Crow Basin. Its absence from
the Klo-kut record is probably a sampling error in
view of its modern abundance at the site (Savage,
1977). Phenacomys, the heather vole, is extremely
rare in the fossil assemblages and does not live in
the northern Yukon today (Youngman, 1975:map
26).

Lemmus sibiricus, the brown lemming, is con-
specific with the Siberian L. obensis but cannot be
separated on the basis of tooth characters from the
more westerly L. lemmus (Rausch and Rausch,
1975/)). It is abundant throughout the northern Yu-
kon fossil record (Table 2).

Of the five species of Microtus that now live in

the Yukon and on the Alaskan mainland, only M.
longicaudus, the long-tailed vole, has not been rec-
ognized among the fossils reported here. The taiga
vole, M. xanthognathus, is distinctive for its large
size as well as its occlusal patterns on lower first and
upper third molars, and it is prominent in most of
the northern Yukon fossil assemblages.

Other species of voles are often difficult to distin-
guish from one another. In referring the Microtus
fossils to species, I have made extensive use of van
der Meulen’s (1973, 1978) measurements and ratios
and have enlarged his techniques to encompass the
highly complex teeth of M. miurus and M. penn-
sylvanicus, the singing and meadow voles, respec-
tively. Despite such careful study, I have been able
to refer only some of the fossils to species (see Guil-
day et al., 1964:166, fig. 22, for a similar problem)
and have left others under the label “A/, miurus or
Pennsylvania^ .” Future studies of Beringian mi-
crotine fossils should focus on methods of distin-
guishing these species from one another as well as
from their Palearctic counterparts ( M. grega/is, M.
agrestis). The diagnostic upper second molar of M.

MORLAN-EASTERN BERINGIAN MICROTINES

191

1984

Fig. 3. — Molars of “Microtus sp. X” from Old Crow Basin. Black lines represent enamel, white areas represent dentine, and stippled
areas represent crown cementum. Dashed lines are reconstructions based on complete teeth. A, Loc. 64 (NMC-251 1 1); B-S, Loc. 1 1
(MkVl-9:44). A, Lml-Lm3; B-C, Lml; D-I, Rml; J-K, Lm3; L-N, Rm3; O-R, LM3; S, RM3.

pennsy/vanicus is present in samples from Unit 2a
(Loc. 1 1), Disconformity A (Locs. 12 and 15), and
Klo-kut. It is not present in the large Biuefish Caves
sample where all lower first molars with 5-6 tri-
angles that Sack taiga vole characteristics have been

referred to M. miurus (Morlan, 1983a). M. miurus
disappears abruptly in the latest Pleistocene levels
at Biuefish Caves and is now confined to moun-
tainous areas of the Yukon (Youngman, 1975:map
31).

192

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 3 . — Morphotypes of maxillary molars of varying lemmings from Old Crow Basin (after Agadjanian and von Koenigswald, 1977,
except for "</" as explained in text). "?" means not recorded. See Fig. 4.

Sample

Upper M 1

Upper M2

Upper M3

<1

I

11

II/III

III

I

11

ii/iii

in

I

II

III

Modem, n. Yukon

2

2

26

i

29

?

?

?

BFC, Pleistocene

3

3

29

2

2

39

3

5

8

HH68-10, Unit 2b

10

2

5

2

6

Loc. 1 5, Disc. A

2

1

1

3

1

4

2

Loc. 1 1, Unit 2a

6

7

1

1

Loc. 44, Unit 2a

2

4

1

1

1

Loc. 64, Unit 2a

2

The tundra vole, M. oeconomus, is unique among
living Beringian voles in having only four closed
triangles on the lower first molar. Good examples
of this species have been found in the Pleistocene
levels of the Bluefish Caves and at Klo-kut, but no
specimens are known for the Boutellier Interval.
Both the middle and lower portions of Unit 2a have
yielded four-triangled lower first molars that are list-
ed under “ Microtus sp. X” in Table 2. These teeth
frequently exhibit a fourth buccal salient angle that
is extremely rare in M. oeconomus (Fig. 3A-I), and
they have lower A/L ratios and higher B/W ratios
(see van der Meulen, 1978) than the modern tundra
vole. One mandible from Loc. 64 contains all its
teeth (Fig. 3A) and shows that the distinctive first
molar is associated with an unusual third molar on
which the apex of the first buccal re-entrant angle
bisects the second lingual salient angle (rather than
meeting the apex of the first lingual re-entrant angle
as in most species of this genus); six isolated lower
third molars of this type are present in the sample
from Loc. 1 1 (Fig. 3J-N). Also from Loc. 1 1, there
are six isolated upper third molars that have more
primitive occlusal patterns than any living species
(Fig. 30-S).

The Beringian taxon most similar to “ Microtus
sp. X” is the ancient M. deceitensis from the Cape
Deceit Formation in western Alaska (Guthrie and
Matthews, 1971), but the Cape Deceit vole is much
larger than the specimens from Old Crow Unit 2a.
The most similar taxa outside Beringia are M. par-
operarius ( Hibbard, 1944:fig. 11; Guthrie, 1965:fig.
2; van der Meulen, 1978:fig. 1 1A-G) and Microtus
sp. C as described by van der Meulen (1973:87-88)
from Layer 10 at the southern Hungarian site of
Villany-8. Van der Meulen (1978:124) argues that
M. paroperarius is an immigrant from Eurasia, di-
rectly descending from Microtus sp. C, and I suggest
that the “ Microtus sp. X” fossils could be Beringian

representatives of that immigration. I have labelled
the fossils “ Microtus sp. X” without intending to
imply that they represent a formal taxon; small sam-
ple sizes and considerations of significant redepo-
sition history (Table 1) preclude the formal defini-
tion of a species at this time, and future study must
entail direct comparisons with other fossils in ad-
dition to the literature review mentioned here. Even
in this preliminary stage of the study, however, it is
reasonable to question the report of M. oeconomus
fossils from Unit 2a (Jopling et ah, 1981:table 2,
Faunas 2 and 6) and to suggest that they might rep-
resent “ Microtus sp. X.” If so, the modern tundra
vole would appear to have arrived in eastern Ber-
ingia in late Wisconsinan time.

Study of varying lemming fossils from Old Crow
Basin has benefited from the concepts advanced by
Agadjanian (1972, 1976; Agadjanian and von Koe-
nigswald, 1977) and Zazhigin (1976), but the re-
sulting sequence of taxa should be regarded as ten-
tative (Table 2). The morphotypes defined for the
maxillary dentition of Dicrostonyx (Agadjanian and
von Koenigswald, 1977) are adequate to character-
ize the Old Crow sequence from the middle of Unit
2a onward through time, but some of the fossils
from the lower part of Unit 2a are even more “prim-
itive” than Morphotype I and have been labelled
“<I” in Table 3 (see Fig. 4H-K). The latter match
published drawings of Predicrostonyx maxillary teeth
(Guthrie and Matthews, 1971:fig. 12; Zazhigin, 1976:
fig. 1 ), but there are too few specimens in the samples
to lend much confidence to an identification.

If larger samples were to support the patterns seen
in Table 3, we could conclude that lower Unit 2a
(at Locs. 44 and 64) contains fossils of Predicros-
tonyx and/or a primitive species of Dicrostonyx (Fig.
4H-M), but Loc. 44 has also yielded a single upper
first molar (Fig. 4N) that represents a more ad-
vanced form. The Loc. 1 1 sample, from the middle

1984 MORLAN- EASTERN BERINGIAN MICROTINES 193

Fig. 4. — Maxillary molars of Dicrostonychini from Old Crow Basin, Unit 2a. Black lines represent enamel and white areas represent
dentine. Dashed lines are reconstructions based on features preserved below the plane of the occlusal surfaces. A-G, Loc. 1 1 (MkVI-
9:54); H-I, Loc. 64 (NMC-25101, -25135); J-N, Loc. 44 (NMC-25640, -25637, -25638, -25641, -25639). A-C, LM1, Morphotype I;
D-F, LM2, Morphotype I; G, RM3, Morphotype II; H-I, RM1, Morphotype <1; J, LM1, Morphotype <1; K, RM1, Morphotype <1;
L-M, RM1, Morphotype I; N, RM1, Morphotype III.

of Unit 2a, appears to represent D. cf. simplicior (or
else an early manifestation of D. hudsonius\ Fig. 4A-
G). All teeth from younger samples are more ad-
vanced than Morphotype L Those from Discon-
formity A exhibit a range of variation seen in D.
gulielmi (Agadjanian and von Koenigswald, 1977),
and I have borrowed a semantic device from Rus-
sian colleagues (Sher et al., 1979:55) in referring
them to D. gulielmi-torquatus. Specimens from
HH68-10 (upper Unit 2b) and the Bluefish Caves
do not differ significantly from the modern northern
Yukon population and are referred to D. torquatus.

All these identifications are based on maxillary
molar morphology. The mandibular dentition in the
samples from Unit 2a exhibits more variation than
I had expected on the basis of published statements
and illustrations. It would appear that the presence
or absence of anterior spurs on the lower second
molar permits discrimination between Predicros-
tonyx and Dicrostonyx (compare Guthrie and Mat-
thews, 1971 :figs. 5-6 and Zazhigin, 1976:Fig. l),but
the specimens from Loc. 1 1 exhibit the entire range
of variation. Similar variations are seen in the sam-
ples from Locs. 44 and 64 (Morlan, 1983c), but all
samples younger than Unit 2a contain lower molars

like those of D. torquatus (that is, both buccal and
lingual spurs on the lower second molar and on most
lower third molars).

Jopling et al. (1981:21) concluded that “the ear-
liest Dicrostonyx in Old Crow Basin were of the
Misothermus line which was replaced by D. torqua-
tus in early Wisconsinan time.” This conclusion may
be premature in that all the available samples are
small and were derived from fluvial sediments that
could contain mixtures of different species repre-
senting various time periods. Furthermore, we must
account for the very primitive tooth patterns that
resemble Predicrostonyx , bearing in mind that this
genus supposedly evolved into Dicrostonyx by at
least early Irvingtonian time (Repenning, 1980:43).
There are several possible explanations for the mix-
ture of morphotypes seen at Loc. 44 (Fig. 4J-N): (1)
the “primitive” teeth belong there, and the single
“advanced” specimen (Fig. 4N) is intrusive because
the sample was derived from sediments below the
seasonal high water mark of Old Crow River; (2)
the advanced specimen belongs there, and the more
primitive teeth were redeposited from older sedi-
ments; or (3) the sequence of morphotypes and
named taxa that has been inferred from Eurasian

194

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

evidence has limited utility in eastern Beringia. I
favor the first alternative, because we have docu-
mented the problem of intrusive specimens below
the high water mark elsewhere in the basin; all max-
illary teeth in the Loc. 1 1 sample from the middle
of Unit 2a represent Morphotype I; and the weight
of evidence from Old Crow Basin supports the Eur-
asian-based sequence for this lineage.

The presence of D. torquatus in early Wisconsinan

time cannot yet be demonstrated, but this species
is definitely present near the top of Unit 2b in sed-
iments that immediately underlie the glaciolacus-
trine clay of the late Wisconsinan (=Duvanny Yar
Interval). Varying lemmings disappear abruptly at
the end of the Pleistocene at the Bluefish Caves, and
they are now confined to the Arctic Coastal Plain
and to mountainous areas of northern and central
Yukon (Youngman, 1975:map 35).

BIOGEOGRAPHY

Youngman (1975:table 1) summarized the refug-
ial origins of microtine rodents now living in Yukon
Territory. The Beringian refugium was home to
Clethrionomys rutilus, Microtus miurus, M. oecon-
omus, Lemmus sibiricus trimucronatus, and Di-
crostonyx torquatus. A Rocky Mountain refugium
held populations of Lemmus sibiricus helvolus, and
all others are seen as postglacial immigrants from
south of Wisconsinan ice limits— Phenacomys in-
termedins, Microtus pennsylvanicus, M. longicau-
dus, M. xanthognathus, and Synaptomys borealis.
Of those he elected to discuss, Macpherson (1965)
agreed with all these assessments. Guthrie (1968:
239) saw the distribution of M. xanthognathus as
indicating “post-glacial expansions from Alaska
rather than to it,” and Rausch and Rausch (1975&:
27) considered Lemmus sibiricus helvolus to be “the
result of post-glacial extension from the north” (see
Burns, 1980).

The northern Yukon fossil record sheds new light
on the histories of some taxa but remains mute on
others. Microtus longicaudus has not been recog-
nized, and no specimens of Synaptomys, the bog
lemming, have been found in eastern Beringia (given
that the specimen from Tofty, Alaska, Repenning
et al., 1964, may have been incorrectly identified
and has since been lost; Harington, 1978:76). Of
extant taxa represented by northern Yukon fossils,
all except Microtus pennsylvanicus and Phenacomys
intermedins were present at the Bluefish Caves dur-
ing the Duvanny Yar Interval which embraces the
late Wisconsinan glaciation (Table 2). The two ex-
ceptions are interesting, because their remains have
been found in sediments representing earlier inter-
vals, and they may have been extirpated from east-
ern Beringia during the last glaciation. Unfortu-
nately we do not yet have samples representing the
relatively cold Happy Interval in the Yukon, but

present evidence suggests that the meadow and
heather voles were present there during the warm
Boutellier Interval and during pre-Happy Interval
episodes that may represent relatively warm periods
of isotope stage 5.

The modern range of M. xanthognathus forms a
large trapezoid with apices near Churchill, Mani-
toba, the Mackenzie Delta, west-central Alaska, and
west-central Alberta (Banfield, 1974:map 92). Elev-
en Rancholabrean faunal assemblages in the eastern
and central United States include M. xanthognathus
(Hallberg et al., 1974), and it is the dominant form
in some of them (for example, Guilday et al., 1964:
164, 1977:64). Because some of these occurrences
have been known for 20 years or more, whereas the
late Wisconsinan record from the Bluefish Caves
was only recently discovered (Cinq-Mars, 1979), the
taiga vole has usually been interpreted as a postgla-
cial southern immigrant that followed retreating gla-
ciers northward (for example, Youngman, 1975:99;
Guilday et al., 1977:65-67). The Bluefish Caves
samples now show that the taiga vole was also res-
ident in Beringia during late Wisconsinan glaciation,
and it seems remarkable that two large demes of the
modern population could have remained isolated
from one another during the late Wisconsinan with-
out differentiation to at least the subspecific level.
The northernmost of the U.S. fossil occurrences is
1,800 km south of Churchill, and none of the U.S.
fossils has been dated to less than 1 1,000 years ago
(cf. Hallberg et al., 1974:640, fig. 1; Graham, 1979:
fig. 6). Unless more northerly and younger mid-
continental fossils are found, it seems likely that the
southern deme of M. xanthognathus became extinct
and that the modern population was derived from
the northern deme that, on the basis of the Bluefish
Caves evidence, persisted in the northern Yukon
throughout the late Wisconsinan and Holocene.

1984

MORLAN-EASTERN BERINGIAN MICROTINES

195

Varying lemming fossils have also been found
south of the Wisconsinan ice terminus in both east-
ern and western North America. The eastern form,
as represented at New Paris No. 4, Pennsylvania
(Guilday et al., 1964:160), has been identified as
Dicrostonyx hudsonius, and western fossils have been
referred to D. torquatus (Guilday, 1968; Martin et
al., 1979; Burns, 1980). Guilday (1968) noted that
the late Pleistocene occurrence of separate western
and eastern forms indicated either that the perigla-
cial tundra belt was discontinuous or that a contin-
uous tundra belt was occupied by varying lemmings
that exhibited clinal variations in tooth form. The
recent discovery of midwestern Dicrostonyx fossils
(for example, Voorhies, 1981; Rhoades, 1982) may
hold promise for choosing between these alterna-
tives. The northern Yukon fossils show that pre-

Wisconsinan Dicrostonyx populations contained all
the variations needed to produce a dine, but they
shed no light on the speciation of the two modern
varying lemmings. Given the complete glacial cover
of Ungava Peninsula and Hudson Bay (Prest, 1 969),
it seems logically impossible to derive the modern
D. hudsonius population from any other source than
northward postglacial migration, but, as Bums ( 1 980:
1510) has noted, the southern population of D. tor-
quatus did not necessarily move northward follow-
ing deglaciation. It could have become extinct, as
suggested here for the taiga vole, with the modern
northern population descending from Beringian res-
idents that are known to have lived in the Yukon,
and in Alaska (Guthrie, 1968; Guthrie and Mat-
thews, 1971), during the Duvanny Yar Interval (Ta-
ble 2).

BIOSTRATIGRAPHY

Repenning (1980, n.d.) has shown that microtine
rodents can provide chronometric evidence because
of their rapid evolutionary changes punctuated by
episodes of widespread dispersal. The data pre-
sented here from Old Crow Basin (Table 2) indicate
some promise in this regard but do not yet permit
unequivocal conclusions. For example, the very
primitive varying lemming molars in the lower part
of Unit 2a (Table 3) suggest that Quaternary ex-
posures in Old Crow Basin may span the period of
time during which Dicrostonyx evolved from Pre-
dicrostonyx. This evolutionary change is docu-
mented in the Olyor Suite of the Kolyma Lowland
in Siberia, but the dating of the Olyor Suite is open
to more than one interpretation. A paleomagnetic
reversal at the Unit Illa/IIIb contact within the Olyor
Suite at the Krestovka section is interpreted by Sher
et al. (1979) and Giterman et al. (1982:52) to rep-
resent the Matuyama-Bruhnes boundary of approx-
imately 700,000 years ago, and the earliest speci-
mens of Dicrostonyx appear just below that contact.
Repenning (n.d.) calibrates the paleomagnetic rec-
ord of the Kolyma Lowland on a much longer time
scale, placing the reversely magnetized Olyor Ilia
prior to the Reunion Subchron, more than 2 m.y.
ago, and the normal Olyor Illb in the Olduvai Sub-
chron (1.7-1. 8 m.y. ago). As mentioned above, a
paleomagnetic reversal of late Matuyama or older
age has been identified in Bluefish Basin (Pearce et
al., 1982), but it is not known whether the Unit 1

clay in Old Crow Basin should be correlated with
the reversed lacustrine unit in Bluefish Basin. Hence
the implications of the varying lemmings from Old
Crow Basin cannot be understood until three prob-
lems are solved: (1) confirmation of the presence or
absence of Predicrostonyx in the lower part of Unit
2a; (2) correlation of Unit 1 with Bluefish Basin and
in turn with the Kolyma Lowland sequence; and (3)
consensus on calibration of the paleomagnetic rec-
ord from Kolyma Lowland.

Subsequent changes within the Dicrostonyx lin-
eage in Old Crow Basin parallel those recorded in
Siberia and Europe (Zazhigin, 1 976; Agadjanian and
von Koenigswald, 1977). The stages also match
meager evidence from Alaska. The “ henseli zone”
in Unit 1 of the Deering Formation at Cape Deceit
can now be understood to have yielded a D. gu-
lielmi-torquatus type of lemming that represents the
Happy Interval. Superjacent Unit 2 in the Deering
Formation represents the Duvanny Yar Interval and
contains D. torquatus (Guthrie and Matthews, 1971;
Matthews, 1974; Hopkins, 1982; Giterman et al.,
1982). The report of D. torquatus from an Alaskan
deposit representing the pre-Illinoian Kotzebuan
transgression cannot be taken at face value, because
the single specimen is a mandible containing ml-
m3 that was not seen in situ (Pewe and Hopkins,
1967:268-269).

The possible chronometric significance of “Mi-
crotus sp. X” in Old Crow Unit 2a is also uncertain.

196

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Even if we assume that the fossils placed in this
category represent a valid taxon related to the im-
migration of Microtus paroperarius, the picture re-
mains unclear. Van der Meulen (1978:143) dates
the arrival of M. paroperarius at “just prior to 0.7
million years” in the mid-continent of North Amer-
ica, but apparently it reached the Wellsch Valley of
Saskatchewan at about the time of the Olduvai Event
(1.7-1. 8 m.y. ago; Repenning, n.d.; Stalker et al.,
1982).

The primitive Dicrostonyx and “ Microtus sp. X”
fossils might be taken to imply that Unit 2a contains
an Irvingtonian fauna, but the associated fossils of
the meadow vole comprise a classic marker of the
beginning of the Rancholabrean when “M. penn-
sylvanicus appears as a flood in almost every ap-
propriate fossil locality east of the Rocky Moun-
tains . . .” (Repenning, 1980:42). Obviously, it
should not be assumed that the land mammal ages
of mid-continental North America have the same
temporal limits and species composition as those
that might be defined in eastern Beringia.

The data reported here represent the first step
toward defining land mammal ages for eastern Ber-
ingia, and even this preliminary framework may aid
in solving an interesting problem. The fauna from
Loc. 12 reported by Jopling et al. (1981; letter “B”
in Fig. 2 and Tables 1-2 of this report) was obtained
from dipping beds for which two markedly different
interpretations have been offered. To Jopling and

his colleagues, a "pingo origin is the most rational
explanation for this particular occurrence” (Jopling
et al., 198 1: 16-1 7). According to this interpretation,
some of the vertebrate fossils were deposited in al-
luvium while others occurred in colluvial deposits,
and the resulting assemblage is little younger than
the lake clay of Unit 1 (said to be Illinoian; Jopling
et al., 1981:18, 29-30). An alternate interpretation
of the dipping beds is that they represent the foreset
beds of point bars laid down by a meandering pre-
cursor of Old Crow River (O. L. Hughes, personal
communication, 1978, 1981). The entire suite of
beds might be attributed to a single fluvial cycle with
vertical relief of approximately 15 m between chan-
nel bottom and floodplain (similar in scale to the
modern river). This interpretation would derive the
vertebrate assemblage from a stream channel much
younger in age than the Unit 1 clay and possibly
similar in age to Disconformity A. Interest in the
age of the assemblage is heightened by the recovery
of altered bones that are interpreted as ancient ar-
tifacts (Jopling et al., 1981:26-29). The microtine
rodent remains could aid in determining the ap-
proximate age of the assemblage. If “ Microtus sp.
X” and primitive species of Dicrostonyx are asso-
ciated, a pre-Happy Interval age would be suggested.
If all the Microtus specimens resemble modern taxa
and the varying lemmings exhibit maxillary mor-
photypes II and III, a Happy Interval or early Bou-
tellier Interval age would be indicated.

CONCLUSION

This summary has shown that substantial prog-
ress has been made in the recovery and interpre-
tation of microtine rodent fossils in eastern Beringia,
but it has shown even more clearly that research
should focus more specifically and more deliberately
on the problems and prospects posed by small mam-
mals. The need for larger samples and better chrono-
metric controls restricts not only our views of bio-
stratigraphy but also the use of the record for

paleoecological reconstruction. This should now be-
come a primary goal, because the relatively restrict-
ed and well defined habitat requirements, dietary
preferences, and intra- and interspecific interactions
of microtine rodents (for example, Batzli and Jung,
1 980) will provide a powerful tool in paleoenviron-
mental analysis, especially if their fossil record can
be integrated with that of other animals and plant
remains.

ACKNOWLEDGMENTS

Laboratory work on the microtme rodent samples from Locs.
11, 12, and 1 5 was begun in 1978. 1 sent a number of troublesome
specimens and a draft manuscript on the samples from Discon-
formity A to Dr. John Guilday who kindly replied at length and
encouraged me to continue the study. An early draft of this paper
was nearly ready for mailing to Dr. Guilday when I learned of

his death, and I am grateful to Drs. Dawson and Genoways for
the opportunity to publish the report in this memorial volume.

Catherine Craig-Bullen undertook the laborious task of picking
microtine teeth and other fossils from samples that were collected
during field work of the Yukon Refugium Project, sponsored by
the National Museum of Man, the Geological Survey of Canada,

1984

MORLAN- EASTERN BERINGIAN MICROTINES

197

and the University of Alberta. I wish to thank Jacques Cinq-
Mars, Archaeological Survey of Canada, National Museum of
Man, for the opportunity to study the Bluefish Caves samples. I
am grateful to C. R. Harington, Paleobiology Division, National
Museum of Natural Sciences, for making his Old Crow Basin
specimens available for this analysis. C. van Zyll de Jong and
David Campbell, Vertebrate Zoology Division, National Mu-
seum of Natural Sciences, kindly loaned me a series of reference
specimens and assisted during several work sessions in the Col-
lection of Mammals.

Earlier drafts of this paper were read by C. R. Harington, J.
V. Matthews, Jr., and C. A. Repenning, all of whom made con-
structive comments that improved the presentation. I benefited
especially from long discussions with John Matthews regarding
pan-Beringian correlations, chronology and paleoenvironment.
Obviously none of these colleagues is responsible for errors that
may remain in this report.

This is Contribution No. 74 of the Yukon Refugium Project.

LITERATURE CITED

Agadjanian, A. K. 1972. Lemmingovye fauny srednego i
pozdnego Pleistotsena (Hoofed lemming fauna of the middle
and late Pleistocene). Biulletin Komissii po Izuchenyiu Chet-
vertichnogo Perioda, 39:67-81.

. 1976. Istoriia stanovleniia kopytnykh lemmingov v

Pleistotsene (Formation history of hoofed lemmings in the
Pleistocene). Pp. 289-295, in Beringiia v Kainozoe (V. L.
Kontrimavichus, ed.), Khabarovsk: Nauka.

Agadjanian, A. K., and von Koenigswald, W. 1977. Merk-
malsvershiebung an den oberen Molaren von Dicrostonyx
(Rodentia, Mammalia) im Jungquartar. Neues Jahrbuch Geol.
Palaont. Abhandlungen, 153:33-49.

Banfield, A. W. F. 1974. The mammals of Canada. Univ.
Toronto Press, Toronto, 438 pp.

Batzli, G. O., and H.-J. G. Jung. 1980. Nutritional ecology
of microtine rodents: resource utilization near Atkasook,
Alaska. Arctic and Alpine Res., 12:483-499.

Bobrowsky, P. T. 1982. The quantitative and qualitative sig-
nificance of fossil and subfossil gastropod remains in ar-
chaeology. Unpublished M.A. thesis, Simon Fraser Univ.,
294 pp.

Burns, J. A. 1980. The brown lemming, Lemmus sibiricus
(Rodentia, Arvicolidae), in the late Pleistocene of Alberta
and its postglacial dispersal. Canadian J. ZooL, 58:1507-
1511.

Cinq-Mars, J. 1979. Bluefish Cave I: a late Pleistocene eastern
Beringian cave deposit in the northern Yukon. Canadian J.
Arch., 3:1-32.

. 1982. Les grottes du Poisson-Bleu. GEOS, 1 1:19-21.

Cumbaa, S. L., D. E. McAllister, and R. E. Morlan. 1981.
Late Pleistocene fish fossils of Coregonus, Stenodus, Thy-
mallus, Catostomus, Lola, and Cottus from the Old Crow
Basin, northern Yukon, Canada. Canadian J. Earth Sci., 18:
1740-1754.

Delorme, L. D. 1968. Pleistocene freshwater Ostracoda from
Yukon, Canada. Canadian J. Zook, 46:859-876.

Giterman, R. E., A. V. Sher, and J. V. Matthews, Jr. 1982.
Comparison of the development of steppe-tundra environ-
ments in west and east Beringia: pollen and macrofossil evi-
dence from key sections. Pp. 43-73, in Paleoecology of Ber-
ingia (D. M. Hopkins, J. V. Matthews, Jr., C. E. Schweger,
and S. B. Young, eds.), Academic Press, New York, 489 pp.

Graham, R. W. 1979. Paleoclimates and late Pleistocene faunal
provinces in North America. Pp. 49-69, in Pre-Llano cul-
tures of the Americas: paradoxes and possibilities (R. L.
Humphrey and D. Stanford, eds.), Anthro. Soc. Washington,
Washington, D.C., 150 pp.

Guilday, J. E. 1968. Pleistocene zoogeography of the lemming
Dicrostonyx (Cricetidae: Rodentia), a reevaluation. Univ.
Colorado Studies, Ser. Earth Set., 6:63-71.

Guilday, J. E., and P. W. Parmalee. 1972. Quaternary peri-
glacial records of voles of the genus Phenacomyx Merriam
(Cricetidae: Rodentia). Quaternary Res., 2:170-175.

Guilday, J. E., P. S. Martin, and A. D. McCrady. 1964. New
Paris No. 4: a late Pleistocene cave deposit in Bedford Coun-
ty, Pennsylvania. Bull. Nat. Speleo. Soc., 26:121-194.

Guilday, J. E., P. W. Parmalee, and H. W. Hamilton. 1977.
The Clark's Cave bone deposit and the late Pleistocene pa-
leoecology of the central Appalachian Mountains of Virginia.
Bull. Carnegie Mus. Nat. Hist., 2:1-87.

Guthrie, R. D. 1965. Variability in characters undergoing rap-
id evolution, an analysis of Microtus molars. Evolution, 19:
214-233.

. 1968. Paleoecology of the late Pleistocene small mam-
mal community from interior Alaska. Arctic, 21:223-244.

Guthrie, R. D., and J. V. Matthews, Jr. 1971. The Cape
Deceit fauna— Early Pleistocene mammalian assemblage from
the Alaskan Arctic. Quaternary Res., 1:474-510.

Hallberg, G. R., H. A. Semken, and L. C. Davis. 1974. Qua-
ternary records of Microtus xanthognathus (Leach), the yel-
low-cheeked vole, from northwestern Arkansas and south-
western Iowa. J. Mamm., 55:640-645.

Harington, C. R. 1977. Pleistocene mammals of the Yukon
Territory. Unpublished Ph.D. dissert., Univ. Alberta, Ed-
monton, 1060 pp.

. 1978. Quaternary vertebrate faunas of Canada and

Alaska and their suggested chronological sequence. Nat. Mus.
Nat. Sci., Syllogeus, 15:1-105.

Hibbard, C. W. 1944. Stratigraphy and vertebrate paleontology
of Pleistocene deposits of southwestern Kansas. Bull. Geol.
Soc. Amer., 55:707-754.

Hopkins, D. M. 1982. Aspects of the paleogeography of Ber-
ingia during the late Pleistocene. Pp. 3-23, in Paleoecology
of Beringia (D. M. Hopkins, J. V. Matthews, Jr., C. E.
Schweger, and S. B. Young, eds.), Academic Press, New
York, 489 pp.

Hopkins, D. M., J. V. Matthews, Jr., C. E. Schweger, and S.
B. Young, (eds.). 1982. Paleoecology of Beringia. Aca-
demic Press, New York, 489 pp.

Hughes, O. L. 1972. Surficial geology of northern Yukon Ter-
ritory and northwestern District of Mackenzie, Northwest
Territories. Geol. Surv. Canada Paper, 69-36:1-1 1.

Hughes, O. L., C. R. Harington, J. A. Janssens, J. V. Matthews,
Jr., R. E. Morlan, N. W. Rutter, and C. E. Schweger.

198

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

1981. Upper Pleistocene stratigraphy, paleoecology, and ar-
chaeology of the northern Yukon interior, eastern Beringia,
1: Bonnet Plume Basin. Arctic, 34:329-365.

Janssens, J. A. P. 1981. Subfossil bryophytes in eastern Ber-
ingia: their paleoenvironmental and phytogeographical sig-
nificance. Unpublished Ph.D. dissert., Univ. Alberta, Ed-
monton, 163 pp.

Jopling, A. V., W. N. Irving, and B. F. Beebe. 1981. Strati-
graphic, sedimentological and faunal evidence for the oc-
currence of Pre-Sangamonian artefacts in northern Yukon.
Arctic, 34:3-33.

Lichti-Federovich, S. 1973. Palynology of six sections of Late
Quaternary sediments from the Old Crow River, Yukon
Territory. Canadian J. Botany, 51:553-564.

. 1974. Palynology of two sections of Late Quaternary

sediments from the Porcupine River, Yukon Territory. Geol.
Surv. Canada Paper, 74-23:1-6.

Macpherson, A. H. 1965. The origin of diversity in mammals
of the Canadian Arctic tundra. Syst. Zool., 14:153-173.

Martin, L. D., B. M. Gilbert, and S. A. Chomko. 1979. Di-
crostonyx (Rodentia) from the late Pleistocene of northern
Wyoming. J. Mamm., 60:193-195.

Matthews, J. V., Jr. 1974. Quaternary environments at Cape
Deceit (Seward Peninsula, Alaska): evolution of a tundra
ecosystem. Bull. Geol. Soc. Amer., 85:1353-1385.

. 1975. Insects and plant macrofossils from two Qua-
ternary exposures in the Old Crow-Porcupine region, Y ukon
Territory, Canada. Arctic and Alpine Res., 7:249-259.

McAllister, D. E., and C. R. Harington. 1969. Pleistocene
grayling, Thymallus, from Yukon, Canada. Canadian J. Earth
Sci„ 6:1185-1190.

Morlan, R. E. 1973. The later prehistory of the middle Por-
cupine drainage, northern Yukon Territory. Nat. Mus. Man,
Arch. Surv. Canada, Mercury Series, 1 1:1-583.

. 1979. A stratigraphic framework for Pleistocene arti-
facts from Old Crow River, northern Yukon Territory. Pp.
125-145, in Pre-Llano cultures in the Americas: paradoxes
and possibilities (R. L. Humphrey and D. Stanford, eds.),
Anthro. Soc. Washington, Washington, D.C., 150 pp.

. 1980a. Taphonomy and archaeology in the Upper

Pleistocene of the northern Yukon Territory: a glimpse of
the peopling of the New World. Nat. Mus. Man, Arch. Surv.
Canada, Mercury Ser., 94:1-380.

. 19806. Yukon Refugium Project. Pp. 65-68, in Ar-
chaeological survey of Canada annual reviews 1977-1979
(R. J. M. Marois, ed.), Nat. Mus. Man, Arch. Surv. Canada,
Mercury Ser., 95:1-106.

. 1981. Quaternary exposures in Old Crow Basin, north-
ern Yukon Territory, 1. Loc. 12(MkVI-10). Manuscript No.
1880, Scientific Records, Archaeological Survey of Canada,
National Museum of Man, Ottawa, 71 pp.

. 1983a. Arvicoline rodents from the Bluefish Caves,

northern Yukon Territory. Manuscript No. 2202, Scientific
Records, Archaeological Survey of Canada, National Mu-
seum of Man, Ottawa, 68 pp.

. 19836. Pleistocene archaeology in Old Crow Basin: a

critical reappraisal. In New evidence for the Pleistocene peo-
pling of the Americas (A. L. Bryan, ed). Center for the Study
of Early Man, Univ. Maine, Orono, in press.

. 1983c. Varying lemming fossils from Old Crow Basin,

northern Yukon Territory. Manuscript No. 2209, Scientific

Records, Archaeological Survey of Canada, National Mu-
seum of Man, Ottawa, 23 pp.

Morlan, R. E., and J. Cinq-Mars. 1982. Ancient Benngians:
human occupations in the late Pleistocene of Alaska and the
Yukon Territory. Pp. 353-381, in Paleoecology of Beringia
(D. M. Hopkins, J. V. Matthews, Jr., C. E. Schweger, and S.
B. Young, eds.). Academic Press, New York, 489 pp.

Morlan, R. E., and J. V. Matthews, Jr. 1978. New dates for
early man. GEOS Winter, 1978:2-5.

. 1983. Taphonomy and paleoecology of fossil insect

assemblages from Old Crow River Loc. 1 5, northern Yukon
Territory. Geographic Physique Quatemaire, 37:147-158.

Naeser, N. D., J. A. Westgate, O. L. Hughes, and T. L. Pewe.
1982. Fission-track ages of late Cenozoic distal tephra beds
in the Yukon Territory and Alaska. Canadian J. Earth Sci.,
19:2167-2178.

Pearce, G. W., J. A. Westgate, and S. Robertson. 1982. Mag-
netic reversal history of Pleistocene sediments at Old Crow,
northwestern Yukon Territory. Canadian J. Earth Sci., 19:
919-929.

Pewe, T. L., and D. M. Hopkins. 1967. Mammal remains of
pre-Wisconsin age in Alaska. Pp. 266-270, in The Bering
Land Bridge (D. M. Hopkins, ed), Stanford Univ. Press,
Stanford, 495 pp.

Prest, V. K. 1969. Retreat of Wisconsin and Recent ice in
North America. Geol. Surv. Canada Map, 1257A.

Rausch, R. L., and V. R. Rausch. 1975a. Relationships of the
redbacked vole, Clethrionomys rutilus (Pallas), in North
America: karyotypes of the subspecies dawsoni and albiven-
ter. Syst. Zool., 24:163-170.

. 19756. Taxonomy and zoogeography of Lemmus spp.

(Rodentia: Arvicolinae), with notes on laboratory-reared
lemmings. Zeit. Saugetierk., 40:8-34.

Repenning, C. A. 1 980. Faunal exchanges between Siberia and
North America. Canadian Journal of Anthropology, 1(1):
37-49.

. (n.d.) Quaternary rodent biochronology and its corre-
lation with climatic and magnetic stratigraphies. Unpub-
lished manuscript.

Repenning, C. A., D. M. Hopkins, and M. Rubin. 1964. Tun-
dra rodents in a late Pleistocene fauna from the Tofty placer
district, central Alaska. Arctic, 17:176-197.

Rhoades, R. S., II. 1982. Mammalian paleoecology of the
Farmdalian Craigmile and the Woodfordian Waubonsie lo-
cal faunas, southwestern Iowa. Unpublished Ph.D. dissert.,
Univ. Iowa, Iowa City, 132 pp.

Ritchie, J. C., J. Cinq-Mars, and L. C. Cwynar. 1982. L'en-
vironnement tardiglaciaire du Y ukon septentrional, Canada.
Geographic Physique Quatemaire, 34:241-250.

Savage, H. 1977. Post- Pleistocene zoology aspects of the
Northern Yukon Research Programme, June 1976 to April
1977. Interim Report: 1977, Northern Yukon Research Pro-
gramme, University of Toronto. Unpublished manuscript.

Schweger, C. E., and J. V. Matthews, Jr. n.d. OldCrowtephra
and its implications for the synoptic paleoecology/paleocli-
matology of Alaska and the Yukon during mid-Wisconsinan
time. Unpublished manuscript.

Shackleton, N. J., and N. D. Opdyke. 1973. Oxygen isotope
and palaeomagnetic stratigraphy of equatorial Pacific core
V28-238: oxygen isotope temperatures and ice volumes on
a 105 year and 106 year scale. Quaternary Res., 3:39-55.

1984

MORLAN- EASTERN BERINGIAN MICROTINES

199

Sher, A. V., T. N. Kaplina, R. E. Giterman, A. V. Lozhkin,
A. A. Arkh angelov, S. V. Kiselyov, Yu. V. Kouznetsov,
E. I. Virina, and V. S. Zazhigin. 1979. Late Cenozoic of
the Kolyma Lowland. XIV Pacific Science Congress, Guide
to Tour XI, 1 16 pp.

Stalker, A. MacS., C. S. Churcher, and R. S. Hill. 1982. Ice
age deposits and animals from the southwestern part of the
Great Plains of Canada. Geol. Surv. Canada Miscel. Report
3 1 (wall chart).

Thorson, R. M., and E. J. Dixon, Jr. 1983. Alluvial history
of the Porcupine River, Alaska: role of glacial-lake overflow
from northwest Canada. Bull. Geol. Soc. Amer., 94:576-
589.

van der Meulen, A. J. 1973. Middle Pleistocene smaller mam-
mals from the Monte Peglia, (Orvieto, Italy) with special
reference to the phylogeny of Microtus (Arvicolidae. Ro-
dentia). Quatemaria, 17:1-144.

. 1978. Microtus and Pitymys (Arvicolidae) from Cum-
berland Cave, Maryland, with a comparison of some New
and Old World species. Ann. Carnegie Mus., 47:101-145.

Voorhies, M. R. 1981. A fossil record of the porcupine (Er-
ethizon dorsatum) from the Great Plains. J. Mamm., 62:
835-837.

Weber, F. R., T. D. Hamilton, D. M. Hopkins, C. A. Re-
penning, and H. Haas. 1981. Canyon Creek: a late Pleis-
tocene vertebrate locality in interior Alaska. Quaternary Res.,
16:167-180.

Westgate, J. A. 1982. Discovery of a large-magnitude, late
Pleistocene volcanic eruption in Alaska. Science, 218:789-
790.

Westgate, J. A., T. D. Hamilton, and M. P. Gorton. 1983.
Old Crow tephra: a new late Pleistocene stratigraphic marker
across north-central Alaska and western Yukon Territory.
Quaternary Res., 19:38-54.

Youngman, P. M. 1975. Mammals of the Yukon Territory.
Publ. Zook, Nat. Mus. Nat. Sci., 10:1-192.

Zazhigan, V. S. 1976. Rannie etapy evoliutsii kopytnykh lem-
mingov (Dicrostonychini, Microtinae, Rodentia)— kharak-
temykh predstavitelei sybarkticheckoi fauny Beringu (Early
evolutionary stages of collared lemmings (Dicrostonychini,
Microtinae, Rodentia) — as characteristic representatives of
Beringian subarctic fauna). Pp. 280-288, in Beringiia v Kai-
nozoe (V. L. Kontrimavichus, ed.), Khabarovsk, Nauka.

Address: Archaeological Survey of Canada, National Museum of Man, Ottawa, Ontario K1A 0M8.

NEW ARVICOLINES (MAMMALIA: RODENTIA) FROM THE
BLANCAN OF KANSAS AND NEBRASKA

Richard J. Zakrzewski

ABSTRACT

Study of arvicolmes from five local faunas of Blancan age in
Kansas (Dixon, White Rock) and Nebraska (Sand Draw, Mullen,
Big Springs) permits the redefinition of the genus Pliophenaco-
mys, the naming of two new genera, Guildayomys and Hibbar-
domys, and six new species — P. dixonensis, G. hibbardi, H. mar-

thae, H. skinned, H. fayae, and PI. voorhiesi. The genera are
distinguished on the basis of the occlusal pattern of ml and M3
and the species on the relative development of dentine tracts on
the posterior loop and anteroconid complex of ml.

INTRODUCTION

While studying the small mammals from the Big
Springs local fauna (l.f.) from the Blancan of north-
eastern Nebraska (Voorhies and Zakrzewski, manu-
script) a problem arose in attempting to describe the
specimens that had been tentatively assigned to
Pliophenacomys osborni. Based on the range of vari-
ation established for the ml by previous studies
(Martin, 1972; Eshelman, 1975) one highly- variable
hypsodont seven-triangled arvicoline was present,
but two distinct morphotypes of the ml could be
observed. One in which the sixth alternating triangle
was normally developed and the dentine tract found
on the side of the triangle interrupted the occlusal
pattern when a millimeter or more of wear had oc-
curred; the second had a sixth triangle that was gen-
erally reduced and rounded in appearance and the
dentine tract found on the side of the triangle in-
terrupted the occlusal pattern very early in life.

Eshelman (1975) working with a larger sample
(N = 69) than Martin (N = 5) observed that the
specimens from the White Rock l.f. (Blancan of
north-central Kansas) resembled those from the
Mullen l.f. (Blancan of central Nebraska) but may
have been aware of the differences I stated when he

suggested that the specimens might belong in a taxon
other than Pliophenacomys as the mis from the
White Rock looked more like specimens that Guth-
rie and Matthews (1971) had described as Pliomys
from the Pleistocene of Alaska. Eshelman question-
ably assigned his specimens to Pliophenacomys os-
borni along with specimens of an arvicoline from
the Dixon l.f. (Blancan of south-central Kansas) that
Hibbard had cataloged as new but never described.

Continued examination of the Big Springs ma-
terial along with material from the local faunas men-
tioned above showed that two morphotypes could
be discerned among each of the other molars, al-
though it was not obvious at first to which m 1 mor-
photype the other molars belonged. With additional
study it became apparent that not one but three
different taxa were represented among the Big
Springs, Mullen, and White Rock specimens con-
sidered to be Pliophenacomys osborni. One of these
is P. osborni. The other two are new genera. A re-
definition of P. osborni and a description of the new
genera and their contained species form the basis
for this report.

METHODS

The specimens consist mainly of isolated teeth. The kind and
amount are listed under the appropriate taxon. Standard mea-
surements (Zakrzewski, 1 967, Fig. 1 ) were made using a Gaertner
Measuring Microscope. The terminology used in the description

of the ml (Fig. 1A) follows van der Meulen (1973); terminology
for the M3 (Fig. IB) follows Zakrzewski (1967). Abbreviations
used with this terminology are given below. These abbreviations
are used throughout the text.

ABBREVIATIONS

ACC = anteroconid complex
AC = anterior cap
AL = anterior loop
BRA = buccal reentrant angle

200

LRA = lingual reentrant angle
PL = posterior loop
T = triangle

1984

ZAKRZEWSKI — BLANCAN ARVICOLINES

201

Fig. 1.— Occlusal patterns of arvicoline teeth (A, Lml ; B, RM3) listing terminology used in text. See list of abbreviations in text for
explanation.

SYSTEMATIC PALEONTOLOGY

Guildayomys, new genus

Diagnosis. — Arvicoline with evergrowing teeth,
lacking cement in the reentrant angles. Dentine tracts
extending the entire height of the crown are found
on both sides of the PL and the anterior face of the
lower teeth, and on both sides of the AL and the
posterior face of the upper teeth. Dentine tracts may
also be present on T6 and 7 of the ml and T1 of
the Ml.

Type species.-— Guildayomys hibbardi, new species.

Etymology’. —Named for the late John E. Guilday, Carnegie
Museum of Natural History.

Guildayomys hibbardi. new species
(Fig. 2)

Pliophenacomys osborni Martin, 1972 (in part). Bull. Univ. Ne-
braska St. Mus., 9:174-176.

Pliophenacomys ? osborni Martin; Eshelman, 1975 (in part), Univ.
Michigan Papers Paleo., 13:41-43.

Holotype. — UNSM 51839, partial left dentary with

I, m 1 -m 2 .

Horizon and type locality. — Long Pine formation.
Big Springs l.f., Antelope Co., Nebraska.

Paratypes. — Big Springs l.f., UNSM 52700, 90800, partial right
dentary with ml-m2; 52794, partial right dentary with ml; 90801,
partial right dentary with I; 52695, 52701, 52705, 527 1 2-527 1 4,
52721, 52726-52727, 52729-52730, 52733, 52735-52739,
52741, 52744-52745, 52748, 52751, 52754, 52761-52762,
52764-52766, 52769, 52774, 52776-52778, 52781, 52783-

52784, 52786, 90802-90807, 43 mis; 90808-90815, 8 m2s;
90816-90817, 2 m3s; 90818-90823, 6 Mis; 90824-90827, 4
M2s; 52893-52897, 52899, 52901, 52903, 90828-90831, 12 M3s.
Mullen l.f., UNSM 39569, m 1 . White Rock l.f., UM-K1 -66: UM
60593, ml.

Diagnosis. — Same as for genus.

Etymology.— Named for the late Claude W. Hibbard.

Description of holotype.— The holotype (Fig. 2A)
represents an individual in the adult stage of de-
velopment. The ml is composed of a PL, seven
alternating triangles, and a simple AC3. A thin strip
of dentine connects the alternating triangles to each
other and in turn to the PL and AC. Triangles 6 and
7 are nearly confluent and open broadly into the
AC. Dentine tracts are present on both sides of the
PL, T6 and 7, and the anterior face of the AC. They
are high enough so that the enamel on the occlusal
surface is interrupted at each of these points. In
addition to being interrupted, the enamel on the
occlusal surface is differentiated into thick and thin
segments. Except for where it is interrupted the
enamel is thicker on the AC and the anterior faces
of the triangles and PL. The enamel remains thick
around the apex of the alternating triangles onto the
posterior face to a point not quite halfway to the
midline of the tooth where it begins to thin until it
reaches the apex of the reentrant angle. The enamel
on the posterior face of the PL is thin. The reentrant

202

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

C VD

Fig. 2 . — Guildayomys hibbardi. A) UNSM 51839, partial left dentary with ml-m2, type, occlusal view of teeth, buccal view of dentary.
B) UNSM 52897, LM3, occlusal and buccal view. C) UNSM 90818, LM1, occlusal and lingual view. D) UNSM 90825, LM2, occlusal
and lingual view. Short bar equals 1 mm on dentary; long bar equals 1 mm on other specimens.

angles are broad and rounded at their apices. They
do not constrict near the midline and the direction
defined by a line which bisects the apex is variable.

The m2 is composed of a PL and four alternating
triangles. The fourth triangle is oval shaped. Dentine
tracts are present on both sides of the PL and the
anterior face of T4. Characteristics of the enamel,
triangle closure, and reentrant angles are as in m 1 .

The occlusal length of the ml-m2 is 4.52 mm.

The length of ml is 3.01 mm, width is 1.21 mm.
For m2 the same measurements are 1.68 and 1.12
mm. Crown height could not be measured.

The dentary is nearly complete. The masseteric
ridges and crests are well developed. The mental
foramen is lateral. There is no capsular process for
the reception of the incisor. There is no temporal
fossa between the m3 and ascending ramus and the
base of the alveoli for m3 is not greatly exaggerated.

1984

ZAKRZEWSKI - BLANCAN ARVICOLINES

203

Description of paralyses. — The remaining dentaries are similar
to the holotype. The characters mentioned above and size make
the dentaries of Guildayomys almost identical to those of Phenac-
ornys intermedins Memriam. Dentaries of Pliophenacomys are
also similar except that the mental foramen is more ventrally
located. This difference can be seen easily when examining the
dentary in a dorsal view. Dentaries of Pitymys and Microtus have
a well developed temporal fossa and the base of the alveoli of
m3 is greatly enlarged to accomodate the tooth.

The m 1 s are similar in occlusal pattern and most characteristics
to that of the holotype. The occlusal pattern in terms of number,
closure, and relative size of alternating triangles is Microtus- like.
The shape of the reentrant angles are similar to Pliophenacomys.
Dentine tracts are present on both sides of the PL and on the
anterior face of the AC in all the teeth. The enamel on the occlusal
surface is also interrupted at all these sites with the exception of
one tooth (UNSM 52774) from an immature individual, where
only the buccal tract on the PL is interrupted. The majority of
mis (41/48) also have dentine tracts on T6 and 7 and in most
cases (36/41) the tracts cause an interruption of the enamel on
the occlusal surface. Five specimens (including the m 1 from White
Rock) have tracts on both triangles that do not extend the entire
length of the crown; three have tracts only on T6; two only on
T7; and two (including the Mullen specimen) lacks tracts on both
T6 and 7.

The paratypic m2s are similar to that of the holotype. The m3
is composed of a posterior loop and three alternating triangles.
Except for size and one less triangle the m3 is like the m2 in the
position and length of dentine tracts, differentiation of enamel,
shape of reentrants, and closure and size of triangles.

The Ml is composed of an AL, three closed alternating tri-
angles, and a fourth triangle that expands confluently into a PL
(Fig. 2C). The buccal triangles are slightly larger than the lingual
triangles. The enamel of the occlusal surface tends to be thicker
on the posterior faces of the triangles and thinner on the anterior
faces. The enamel of the anterior face of the PL is thick. This
enamel is interrupted on the posterior face of the PL, the apex
of T 1 , and both sides of the AL as dentine tracts are present
along the sides of the crown, throughout its entire length, at these
points. The apices of the BRAs are directed posteriorly, whereas
those of the LRAs tend to be variable.

The M2 is composed of an AL, two closed alternating triangles,
and a third triangle that expands confluently into a PL (Fig. 2D).
Smaller size, one less triangle, and the development of dentine
tracts only on the loops distinguishes the M2 of Guildayomys
from the M 1 .

The M3 (Fig. 2B) is composed of an AL that opens broadly
into a small buccal triangle. This small triangle is generally dosed
off from a large lingual triangle, which in turn is generally dosed
off from a PL. This pattern most closely resembles that in Plio-
phenacomys and Pliolemmus. Like the other molars the enamel
of the occlusal surface is differentiated into thick and thin seg-
ments. The enamel is thinner on the anterior face of the AL and
T2 and the posterior face of LRA2. The enamel pattern of the
occlusal surface is interrupted by dentine tracts on both the buccal
and lingual side of the AL and on the posterior face of the PL.
In three of the seven specimens the tract on the PL is wide enough
to cover the buccal side of the loop as well. The direction that
the apices of the reentrant angles take with respect to the midline
of the tooth is variable, but they tend to be perpendicular to it
for the most part. In two specimens, a small BRA3 is present.
Measurements of various parameters of the teeth of Guildayomys
are given in Table i.

Table 1 . —Measurements of teeth in Guildayomys from the Big
Springs.

Tooth

Character

N

Mean

SD

Observed range

ml

GL

23

2.99

0.19

2.65-3.35

GW

23

1.99

0.07

1.00-1.35

GH

19

3.71

0.50

2.45-4.99

m2

GL

7

1.74

0.04

1.68-1.79

GW

7

1.10

0.05

1.03-1.20

GH

4

3.16

0.27

2.86-3.42

m3

GL

2

1.36

0.05

1.32-1.39

GW

2

0.80

0.09

0.73-0.86

GH

2

2.21

0.28

2.01-2.41

MI

GL

4

2.50

0.11

2.38-2.61

GW

4

1.33

0.11

1.24-1.48

GH

4

3.83

0.09

3.70-3.90

M2

GL

4

1.93

0.04

1.89-1.96

GW

4

1.19

0.06

1.12-1.27

GH

4

3.53

0.43

2.92-3.83

M3

GL

7

1.74

0.08

1.63-1.81

GW

7

1.02

0.05

0.96-1.08

GH

7

2.88

0.37

2.42-3.50

LBRAI

7

0.28

0.04

0.23-0.33

LBRA3

7

0.04

0.07

0.00-0.16

See Table 3 for abbreviations not listed in text.

Discussion. — The occlusal pattern of the teeth in
Guildayomys most closely resemble those of Plio-
phenacomys. Guildayomys can be easily distin-
guished from Pliophenacomys and all other pre-Ir-
vingtonian voles except for Pliolemmus, by the fact
that the molars are evergrowing. Pliolemmus, the
only other Blancan vole with evergrowing molars,
can be distinguished from Guildayomys by the fact
that in many specimens of Pliolemmus all of the
alternating triangles and loops of the teeth possess
dentine tracts. Because of this feature the tips of the
triangles and loops in Pliolemmus tend to be blunt.
In addition, the M3 of Pliolemmus has an additional
triangle and it and Pliophenacomys have a shorter
BRA I on the M3. The PL of the upper molars tend
to be more elongate in Guildayomys than in either
Pliophenacomys or Pliolemmus.

Guildayomys differs from all the other voles with
evergrowing teeth except for the extant steppe lem-
ming Lagurus, and the extinct Eolagurus and Pro-
lagurus, by the lack of cement in the reentrant an-
gles. These latter three taxa known from Asia, have
a M3 with an additional triangle, a deeper BRA1,
and more elongate PL. The sage brush vole, extant
in western North America and generally assigned to
Lagurus, has a similar M3 but has cement in the

204

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

reentrants. It also has a dentine tract on T7 of the
m 1 .

The similarities of occlusal pattern between Guil-
dayomys and Pliophenacomys suggests that perhaps
the latter may stand in an ancestral position to the
former. Guildayomys relationship to Pliolemmus,
advanced voles with cement in their reentrants, and
Lagurus is not as obvious. Except for the two iso-
lated mis from the Mullen and White Rock, the
taxon is known only from the material in the Big
Springs l.f. Additional material from North America
and a detailed study of the Asian material is nec-
essary before any definitive statement on relation-
ships can be made.

Hibbardomys , new genus
(Figs. 3-5)

Diagnosis.— Hibbardomys is a primitive vole with
rooted teeth, lacking cement in the reentrant angles.
The lingual reentrants of the lower molars are nar-
rower than in Pliophenacomys. The reentrants tend
to constrict and turn anteriorly as they approach the
midline of the tooth. The enamel on the occlusal
surface of the lower molars is differentiated into
thick and thin segments. The anterior faces of the
alternating triangles, the ACC, and both faces of the
PL exhibit thick enamel, whereas the enamel is thin
on the posterior faces of the alternating triangles.

The ml is composed of a PL, three to seven al-
ternating triangles, and an AC. The complexity of
the AC varies. It is more complex in specimens with
fewer triangles and becomes simpler as triangles are
added. A well developed sixth triangle is seldom, if
ever, present. In primitive species what would be
counted as the sixth triangle is rectangular in shape
and appears to be an outgrowth of the AC; a con-
dition generally referred to as trilobate when seen
in other Blancan voles (Hibbard and Zakrzewski,
1967). In advanced species T6 tends to be blunt or
rounded. In a few specimens T6 may even appear
to be lacking, but no matter what the configuration
a well developed dentine tract is always present where
T6 would form. The buccal side of the AC generally
lacks a dentine tract but a very thin tract may occur
on the anterior face of the AC.

The M3 is composed of an AL, three alternating
triangles, and a PL. Though smaller than T2, T1
and 3 are better developed than in Pliophenacomys.
The BRA1 is well developed (>0.35 mm). The pos-
terior lingual reentrants of Ml (LRA3) and M2
(LRA2) tend to be shallower than in Pliophenaco-

mys. Reentrant pits (Zakrzewski, 1969, fig. 7c) are
variously developed on both sides of the last triangle
on the Ml and M2. The height of the dentine tract
on the lingual side of the AL on the M 1 is signifi-
cantly smaller than the tract on the buccal side.

Type species. — Hibbardomys voorhiesi, new
species.

Etymology. — Named for the late Claude W. Hibbard.

Recognized species and occurrences.— H. marthae, new species.
White Rock and Big Springs; H. skinneri. new species, Sand
Draw; H. fayae, new species, Dixon; H. voorhiesi, new species,
White Rock, Mullen, and Big Springs.

Hibbardomys marthae, new species
(Fig. 3A-B)

Ogmodontomys sp .; Eshelman, 1975, Papers Paleo., Univ. Mich-
igan, 13:38-39.

Pliophenacomys cf. P. primaevus Hibbard; Eshelman, 1975, Pa-
pers Paleo. Univ. Michigan, 13:43-44.

Holotype. — UNSM 51279, partial right dentary
with i, and ml.

Horizon and locality.— Long Pine formation, Big
Springs l.f., Antelope Co., Nebraska.

Paratypes. — Big Springs, UNSM 52770, 52905, 52906, 52908,
52910, 52912, 90832-90836, 1 1 mis; 90837, Ml; 90838-90844,
7 M2s; 52898, 52900, 2 M3s. White Rock-UM-Kl-66: UM
61657, ml; 81750, M3; UM-K3-69: 61858, 3 mis; UM-K4-72:
61743, Ml, 81751, 2 M2s; UM-K5-72: 61744, 5 Mis, 81752, 4
M2s.

Diagnosis.— Hibbardomys marthae is distin-
guished from other members of the genus by its
shorter dentine tracts and the complexity of the ACC
on the m 1 .

Etymology.— Named for my wife, Martha.

Description of holotype.— The. ml (Fig. 3 A) con-
sists of a PL, three alternating triangles and a com-
plicated ACC. The ACC is composed of a fourth
triangle that is modified by a Mimomys- ridge and
prism fold and a fifth triangle that is confluent with
a very small third lobe at the position of the sixth
triangle. These in turn open into a small AC.

The enamel of the occlusal surface is interrupted
at the third lobe = T6 by a well developed dentine
tract on the buccal side of the ACC. Dentine tracts
are slightly developed on the buccal side of the tooth
on PL and T2. The actual height of the tracts and
crown height of the tooth could not be measured
because the base of the tooth is covered by the den-
tary. The occlusal length of the tooth is 2.90 mm,
occlusal width 1.41 mm.

1984

ZAKRZEWSKI — BLANCAN ARVICOLINES

205

The ml is two rooted and lacks cement in the
reentrant angles. The majority of the reentrant an-
gles are narrow and their apices are oriented ante-
riorly. However, BRA3 is round, shallow and its
“apex” is directed posteriorly. The apex of LRA4
is perpendicular to the midline of the tooth.

With the exception of the PL the enamel of the
posterior faces of the remaining triangles is thinner
than that of the anterior faces. The enamel of the
anterobuccal face of the AC appears to be thinner
as well.

The dentary is broken posterior to the alveoli for
the m3. Muscle attachments for the masseter com-
plex on the buccal side of the jaw are well developed.
There is no evidence of a temporal fossa between
the tooth row and ascending ramus.

Description of paratypes.— The other m 1 s are generally similar
in occlusal pattern and dentine tract development. The major
differences between the teeth are in the relative complexity of
the ACC. The complexity is determined by the presence or ab-
sence of a prism fold, enamel ridge, or a trilobed pattern. All
m 1 s have at least one of these characters developed, some have
two. A summary of the variation seen in the ACC is listed in
Table 2. Twelve of the 16 mis have the enamel interrupted at
the position of T6. The tooth that shows enamel interruption at
the earliest age is one with a crown height of 2.74 mm. The
shortest tract where the crown is uninterrupted is at 1.70 mm.
Measurements of other parameters of the m 1 are listed in T able 3 .

The Ml and M2 are assigned to this taxon because of the
development of reentrant pits (Fig. 3B). The M3s are assigned
because of occlusal pattern.

Hibbardomys skinneri, new species
(Fig. 3F-H)

Pliophenacomys primaevus Hibbard; Hibbard, 1972 (in part).

Bull. Amer. Mus. Nat. Hist., 148:102-105.

Table 2.— Variation of ACC of ml in Hibbardomys marthae.

Character

Pnsm

Enamel

Third

PF

ER

Locality

fold

ridge

lobe

3L

3L

Big Springs

1

1

4

5

1

White Rock

0

0

0

4

0

Ho/otype. — UM 57192, Rml.

Horizon and type locality. — Keim formation, UM-
Nb3-67 Sand Draw l.f.. Brown Co., Nebraska.

Paratype. — Magill: UM 57017, ml; 81753, 2 m2s; 81754, m3;
81755, 3 Mis; 81756, 2 M2s; 81757, M3; Owl Pellet: 81758,
m3: 81759, M2.

Diagnosis. — Hibbardomys skinneri is distin-
guished from H. marthae by its slightly larger size,
higher dentine tracts, and less complicated ACC on
m 1 . It differs from other members of the genus by
its lower dentine tracts and slightly complex ACC
on ml.

Etymology. — Named for Morris F. Skinner, American Mu-
seum of Natural History.

Description of holotvpe. — The holotype repre-
sents an individual in an early adult stage of devel-
opment. The tooth (Fig. 3F-H) is composed of a
PL, three alternating triangles, and a slightly com-
plex ACC. The ACC is composed of a fourth and
fifth alternating triangle; T5 is confluent with what
could be considered a sixth triangle but the latter is
rectangular in shape. These two triangles open
slightly into a simple AC. The occlusal length of the
tooth is 2.87 mm, the width is 1.32 mm. Crown

Table 3 .—Measurements of ml in Hibbardomys marthae.

Locality

N

Mean

SD

Observed range

N

Mean

SD

Observed range

GL

GW

Big Springs

12

2.74

0.14

2.46-3.00

12

1.36

0.07

1.23-1.47

White Rock

2

2.61

0.06

2.56-2.65

2

1.31

0.01

1.30-1.31

GH

DTH T6

Big Springs

1 1

1.99

0.55

1.15-2.74

3

2.18

0.52

1.70-2.74*

White Rock

2

2.07

0.85

1.47-2.67

1

2.16

HBDT

HLDT

Big Springs

11

0.68

0.07

0.60-0.82

11

0.37

0.04

0.32-0.45

White Rock

2

0.69

0.08

0.63-0.74

2

0.32

0.01

0.31-0.33

N = number of specimens; SD = standard deviation; GL = greatest length; GW = greatest width; GH = greatest height; DTH T6 = dentine tract height of triangle 6; HBDT =
height of buccal tract on PL; HLDT = height of lingual tract on PL; * = occlusal pattern interrupted, therefore, may not be representative of maximum height; if on lower value,
may not be minimum height.

206

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 3 .—Hibbardomys and Pliophenacomys. A-B) H. marthae; A) UNSM 51279, partial right dentary with i and ml, type, occlusal
view of tooth, buccal view of dentary; B) UNSM 90837, LM1, occlusal and posterior views. C-E) P. primaevus, UM 57183, Rml, C)
occlusal view, D) buccal view, E) lingual view. F-H) H. skinneri, UM 57192, Rml, type, F) occlusal view, G) buccal view, H) lingual
view. Short bar equals 1 mm on dentary and side views; long bar equals 1 mm on occlusal and posterior views.

height is 2.82 mm. Dentine tract height on the PL
is 0.83 mm on the buccal side and 0.30 mm on the
lingual. The height of the dentine tract on T6 is
sufficient to interrupt the enamel of the occlusal sur-
face at this point.

Description of paratypes. — The other ml (UM 57017) differs
from the type in that T6 though not as well developed is highly
crenulated. In addition a shallow LRA5 is developed on the ACC

that delimits a small seventh triangle. Measurements of this tooth
are listed in Table 4. The other teeth are assigned to Hibbardomys
because they have the generic characters. Their measurements
are listed in Tables 5-9.

Hibbardomys fayae , new species
(Fig. 4A-F)

Pliophenacomys meadensis Hibbard; Hibbard, 1956 (in part).
Papers, Michigan Acad. Sci., Arts & Letters, 41:164-167.

1984

ZAKRZEWSKI — BLANCAN ARVICOLINES

207

Table 4.— Measurements of ml in Hibbardomys and Pliophenacomys.

Taxon

N

Mean

SD

Observed range

N

Mean

SD

Observed range

GL

GW

H. skinneri

2

2.93

0.08

2.87-2.99

2

1.40

0.1 1

1.32-1.47

H. fayae

8

2.92

0.25

2.45-3.16

7

1.26

0.08

1.1 1-1.37

H. voorhiesi'

15

2.89

0.15

2.58-3.03

15

1.27

0.07

1.15-1.37

H. voorhiesi 2

12

2.96

0.19

2.66-3.55

12

1.30

0.07

1.19-1.38

H. voorhiesi 3

1

3.00

1

1.36

P. primaevus

1

2.86

1

1.22

P. dixonensis

5

2.91

0.12

2.78-3.03

5

1.22

0.08

1.10-1.30

P. osborni'

10

2.93

0.12

2.75-3.13

10

1.19

0.05

1.11-1.26

P. osborni 2

7

3.12

0.18

3.00-3.40

9

1.22

0.09

1.11-1.32

P. osbornP

2

3.08

0.35

2.83-3.32

2

1.25

0.23

1.09-1.41

GH

HLDT

H. skinneri

2

2.72

0.13

2.62-2.81

2

0.30

0.01

0.29-0.30

H. fayae

8

2.52

0.95

1.24-4.1 1

8

0.63

0.15

0.37-0.88

H. voorhiesi'

15

3.58

0.79

1.60-4.43

I

H. voorhiesi 2

10

2.86

0.86

1.34-4.02

I

H. voorhiesi 3

1

3.00

I

P. primaevus

1

3.63

1

0.74

P. dixonensis

5

2.67

0.32

2.39-3.20

5

1.96

0.30

1.68-2.39*

P. osborni'

10

3.50

0.57

2.44-4.30

I

P. osborni 2

9

2.67

0.84

1.34-3.54

I

P. osbornP

2

3.35

0.08

3.29-3.41

I

1 Big Springs.

2 White Rock.

3 Mullen.

I specimens have interrupted enamel, except for immature, see text for measurements. See Table 3 for other abbreviations.

Gen. et sp. indet.; Hibbard and Zakrzewski, 1967, Contr. Mus.

Paleo. Univ. Michigan, 21:262.

Dliophenacomys ? osborni Martin; Eshelman 1975, Univ. Mich-
igan Papers Paleo., 13:41-43.

Holotype. — UM 54948, Lml.

Horizon and type locality. — ''" Belleville” forma-
tion (Hibbard, 1972), Dixon l.f., Kingman Co.; Kan-
sas.

Paratypes. -UM 54945, 54947, 55445, 12 mis; 54950, 55446,
55449, 1 1 m2s; 54954, 55448, 55453, 4 m3s; 54951, 54953,
55454, 55455, 14 Mis; 55450, 55451, 55456, 10 M2s; 54952,
55458, 5 M3s.

Diagnosis.— Hibbardomys fayae has dentine tracts
and anatomical roots that are intermediate in de-
velopment between H. skinneri and H. voorhiesi.

Etymology.— Named for Faye Ganfield Hibbard, wife of the
late Claude W. Hibbard.

Description of the holotype.— The holotype rep-
resents an individual that is in an early adult stage
of development. The tooth (Fig. 4A-C) consists of
a PL, three closed, alternating triangles, and a com-
plex ACC. The ACC consists of five additional tri-
angles and a small AC. Triangles 4 and 5 form a
pair in continuity with Tl-3. Triangle 6 has a rect-

angular rather than a triangluar shape. The anterior
edge of the enamel on the occlusal surface of T6 is
interrupted by a dentine tract along its side. Tri-
angles 7 and 8 are confluent and open broadly into
the AC. Apices of the reentrant angles are directed
anteriorly except for the most anterior three, which
are directed posteriorly. BRA4 is shallower than it
is in H. voorhiesi.

The tooth has an occlusal length of 2.98 mm,
width of 1.22 mm and a crown height of 3.26 mm.
In addition to the dentine tract on T6, tracts are
present on both sides of the PL. The buccal tract is
2.80 mm, the lingual 0.88 mm.

Description of paratypes. — The other mis are similar to the
holotype, except that their ACCs are not as complex. Six of the
seven paratypic mis, wherein T6 could be observed, have the
enamel of the occlusal surface interrupted at that position. The
other specimen represents an immature individual with a crown
height of4.1 1 mm. The height of the tract on T6 in this individual
is 3.59 mm. Five of the mis also have the enamel interrupted
on the buccal side of the PL. It seems that the enamel will always
be interrupted on the buccal side when wear on the crown reduces
its height to approximately 2.35 mm. The highest crowned tooth
in which the enamel is interrupted measures 2.37 mm, and the
shortest tract that can be measured on a tooth that does not
possess interrupted enamel is also 2.37 mm. The highest buccal
tract that can be measured is on the immature individual men-

208

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 4.—Hibbardomys and Pliophenacomys. A-F) H. fayae\ A-C) UM 54948, Lml, type; A) occlusal view, B) buccal view, C) lingual
view. D-F) UM 55458, RM3; D) occlusal view, E) buccal view, F) lingual view. G-L) P. dixonensis; G-I) UM 55444, Lml, type, G)
occlusal view, H) buccal view, I) lingual view. J-L) UM 55452, RM3, J) occlusal view, K) buccal view, L) lingual view. Short bar equals
1 mm on side views of mis; long bar equals 1 mm on occlusal view of all teeth and side views of M3s.

tioned above, 2.85 mm. The enamel on the lingual side of the
PL will not be interrupted until crown height wears down to 0.88
mm, which is the height of this tract on the type. Measurements
on the mis are listed in Table 4. The remaining teeth were as-
signed on the basis of generic characters. Six of the 1 1 m2s have
the enamel interrupted on the buccal side of the PL, whereas
only two teeth have the enamel interrupted on the lingual side.
Measurements of the m2 are listed in Table 5, those of m3 in
Table 6.

The Mis of Hibbardomys can be easily distinguished from
those of Pliophenacomys in the Dixon l.f. by possessing a shal-

lower LRA3, shorter dentine tracts (although there is no appre-
ciable difference between the crown heights of the two taxa no
Hibbardomys specimens exhibits interrupted enamel on the oc-
clusal surface), and anatomical roots are better developed (all 14
Hibbardomys Mis have three roots, whereas two of six Plio-
phenacomys M 1 s have two roots). Measurements of M 1 are listed
in Table 7.

The M2s of Hibbardomys are distinguished from Pliophenac-
omys by a shallower LRA2 and less developed reentrant pits.
Unlike the Mis the dentine tracts on the AL of the M2s in H.
fayae are high enough so that two specimens have an interrupted

1984

ZAKRZEWSKI — BLANCAN ARVICOLINES

209

Table 5 . — Measurements of m2 o/ Hibbardomys and Pliophenacomys.

Taxon

N

Mean

SD

Observed range

N

Mean

SD

Observed range

GL

GW

H. skinneri

1

1.82

2

1.22

0.10

1.15-1.29

H. fayae

1 1

1.69

0.12

1.55-1.91

1 1

1.20

0.09

0.98-1.29

H. voorhiesi '

5

1.72

0.08

1.61-1.81

5

1.20

0.08

1.1 1-1.32

H. voorhiesi2

15

1.68

0.08

1.60-1.79

16

1.17

0.08

1.05-1.32

H. primaevus

3

1.74

0.03

1.71-1.77

3

1.13

0.06

1.07-1.17

P. dixonensis

5

1.72

0.05

i. 67-1. 79

5

1.09

0.05

1.02-1.15

P. osborni'

2

1.76

0.19

1.62-1.89

2

1.12

0.11

1.04-1.20

P. osborni 2

5

1.66

0.09

1.56-1.75

5

1.06

0.09

0.92-1.14

GH

H. skinneri

2

2.39

0.51

2.03-2.75

H. fayae

1 1

2.18

0.80

1.08-3.12

H. voorhiesi'

5

2.76

0.56

2.35-3.68

H. voorhiesi2

15

2.41

0.64

1.65-3.59

P. primaevus

3

2.54

0.43

2.07-2.92

P. dixonensis

5

2.09

0.57

1.13-2.57

P. osborni'

2

3.01

0.31

2.79-3.23

P. osborni 2

5

2.55

0.28

2.30-2.89

HBDT

HLDT

H. skinneri

2

LOO

0.21

1.15-1.29

2

0.52

0.28

0.32-0.71

H. fayae

5

2.23

0.22

1.90*-2.44

10

0.93

0.28

0.59-1.34*

H. voorhiesi'

2

2.96

1.02

2.24-3.68*

2

2.64

1.48

1.59-3.68*

H. voorhiesi 2

3

3.14

0.40

2.78*-3.57

6

2.40

0.80

1.54-3.35

P. primaevus

3

1.47

0.21

1.28-1.70

2

0.99

0.15

0.88-1.09

P. dixonensis

2

2.35

0.32

2.12*-2.57*

4

1.44

0.17

1 .23*— 1 .65

P. osborni'

2

2.00

1.90

0.66-3.34*

2

1.88

2.07

0.41-3.34*

P. osborni 2

2

2.25

0.35

2.00-2.49

3

2.23

0.35

1.89-2.58

See Table 3 for abbreviations.

enamel pattern on the buccal side of the AL and two specimens
have both sides interrupted. Measurements of M2 are listed in
Table 8.

The M3s of Hibbardomys fayae are distinguished by their oc-
clusal pattern (Fig. 4D-F). Two of the M3s have interrupted
enamel on both sides of the AL. One of these specimens has a
crown height of 2.04 mm. This is the highest tract present on the
lingual side. Measurements of the M3 are listed in Table 9.

Hibbardomys voorhiesi, new species
(Fig. 5)

Pliophenacomys osborni Martin, 1972 (in part), Bull. Nebraska
St. Mus., 9:174-176.

Pliophenacomys ? osborni Martin; Eshelman, 1975 (in part), Univ.
Michigan Mus. Paleo., Papers Paleo., 13:41-43.

Hololype. — UM 55523, partial left dentary with

m l-m2.

Horizon and type locality. — Belleville fm., White
Rock l.f., Loc. UM-K1-66, Republic Co., Kansas.

Paratypes.— White Rock l.f.— UM-K 1-66: 81760, 3 mis; 8 1 761,
4 m2s; 81762, m3; 60590, 6 Mis; 81763, 2 M2s; 81764, M3.
UM-K3-69: 6 1 736, 24 m 1 s; 8 1 765, 20 m2s; 8 1 766, 4 m3s; 81767,
29 Mis; 81768, 17 M2s; 81769, 8 M3s. UM-K4-72: 61737, 2

mis; 81770, 2 m2s; 81771, 2 m3s; 81772, 3 Mis; 81773, 2 M2s.
UM-K4-72 + 12' SE: 81774, 2 mis; 81775, m2; 81776, 2 Mis.
UM-K4-72 + 200' NW: 81777, ml; 61739, 2 Mis. UM-K5-72:
61740, 1 1 mis; 81788, 10 m2s; 81779, 2 m3s; 81780, 10 Mis;
81781, 3 M2s; 81782, 7 M3s. UM-K6-72: 8 1 783, 2 mis; 81784,
2 m2s; 81785, 2 M3s. UM-K7-72: 61742, Ml. Mullen l.f.-
UNSM Coll Loc. Cr. 10, Pit 3: 39570-39571, 2 mis. Big Springs
l.f.: 52697-52699, 52703-52704, 52706-52711, 52715-52718,
52722-52725, 52728, 52732, 52742, 52750, 52753, 52755-
52759, 52768, 52771, 52773, 52775, 52779-52780, 52782, 52785,
52788-52789, 52791-52793, 52795, 90845, 44 mis; 90846-
90850, 5 m2s; 90851, m3; 90852-90902, 5 1 Mis; 90903-90910,
8 M2s; 52874-52875, 52877-52880, 52883-52886, 52888-
52889, 52891-52892, 9091 1, 15 M3s.

Diagnosis.— Hibbardomys voorhiesi is distin-
guished from the other members of the genus by its
greater hyposodonty and higher dentine tracts.

Etymology. — Named for Michael R. Voorhies, University of
Nebraska-Lincoln.

Description of holotype. — The holotype (Fig. 5 A)
represents an individual in an adult stage of devel-
opment. The back of the jaw posterior to the alveoli

210

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 6.— Measurements of m3 in Hibbardomys and Pliophenacomys.

Taxon

N

Mean

SD

Observed range

N

Mean

SD

Observed range

GL

GW

H. skinneri

2

1.66

0.18

1.53-1.79

2

mi

0.04

1.08-1.13

H. fayae

4

1.63

0.20

1.34-1.79

4

1.00

0.10

0.85-1.08

H. voorhiesi'

1

1.56

1

0.92

H. voorhiesi 2

4

1.55

0.14

1.37-1.71

4

0.96

0.08

0.87-1.03

P. primaevus

2

1.56

0.02

1.54-1.57

2

0.97

0.07

0.92-1.02

P. dixonensis

1

1.33

1

0.86

P. osborni'

1

1.46

1

0.90

P. osborni 2

5

1.46

0.10

1.36-1.61

5

0.93

0.05

0.85-0.98

GH

H. skinneri

2

1.90

0.20

1.34-1.79

H. fayae

4

1.68

0.36

1.15-1.92

H voorhiesi 1

1

3.40

H. voorhiesi 2

4

2.04

0.88

1.41-3.21

P. primaevus

2

1.82

0.33

1.58-2.05

P. dixonensis

1

2.40

P. osborni 1

1

1.87

P. osborni 2

5

2.26

0.10

1.57-2.24

HBDT

HLDT

H. skinneri

2

0.65

0.11

0.57-0.73

2

0.44

0.16

0.32-0.55

H. fayae

3

1.25

0.48

0.77-1.73

2

0.97

0.18

0.84-1.10

H. voorhiesi 1

1

2.42

1

2.42

H. voorhiesi 2

2

1.98

0.33

1.75-2.21*

3

1.59

0.28

1.27-1.80

P. primaevus

2

0.64

0.31

0.42-0.86

2

0.61

0.06

0.57-0.65

P. dixonensis

1

1.90

1

0.88

P. osborni'

1

1.12

1

0.83

P. osborni 2

5

1.80

0.26

1.57-2.24*

4

1.48

0.23

1.27-1.88

See Table 3 for abbreviations.

for m3 is missing. The m 1 consists of a PL, seven
alternating triangles, and a simple AC. The PL is
closed off from Tl.A thin strip of dentine succes-
sively joins each of the first six alternating triangles.
Triangle 6 opens broadly into T7. Seven, in turn,
opens into the AC. Dentine tracts are present on
both sides of the PL, T6, and the anterobuccal edge
of the AC. The tracts are high enough so that the
enamel of the occlusal surface is interrupted at these
points. The tract on T6 is so developed that the
triangle has a rounded appearance.

The apices of the reentrant angles are directed
anteriorly except for LRA 5, which is perpendicular.

The m2 consists of a PL, three alternating trian-
gles, and a small oval-shaped anterior loop or fourth
triangle. Details of enamel closure and thickness,
reentrant direction, and dentine tract development
are similar to that of the ml. The major difference
is that the entire anterior face of the fourth triangle
lacks enamel.

The occlusal length of ml-m2 is 4.68 mm. The
length of ml is 3.10 mm and m2, 1.58 mm. The

width of the teeth is 1.30 and 1.20 mm, respectively,
crown and dentine tract height could not be mea-
sured.

The dentary is too fragmentary to provide for an
adequate description. The diastemal region is short
and the crests for muscle attachment are poorly de-
veloped. But whether these are primary or second-
ary conditions is not determinable. The dental fo-
ramen is in a more ventral position than it is in
Pliophenacomys. It appears that a very slight mas-
seteric fossa may have been present between the m3
and the ascending ramus, but a dentary in which
this area is better preserved is needed before a more
positive statement can be made.

Description of paratypes.— Although variation is present, the
remaining mis are well characterized in the description of the
holotype. The pattern of the ACC with its rounded T6 on which
the enamel is interrupted because of the dentine tract height
appears to be so consistent that m 1 s of this taxon can be separated
from all others so far known with relative ease.

Only one (UNSM 52755) of the 27 mis that could be measured
does not have an interrupted occlusal surface at the site of T6.

1984

ZAKRZEWSKI — BLANCAN AR VICO LINES

211

Fig. 5 .—Hibbardomys voorhiesi. A) UM 55523, partial left dentary with m l-m2, type, occlusal view of teeth and buccal view of dentary.
B-D) UNSM 9091 1, LM3; B) occlusal view, C) lingual view, D) buccal view. E-G) UNSM 90852, LM1, E) occlusal view, F) lingual
view, G) buccal view. H-J) UNSM 90903, LM2, H) occlusal view, I) lingual view, J) buccal view. Short bar equals 1 mm on side views,
long bar equals 1 mm on occlusal views.

This specimen represents an immature individual that has a crown
height of 4.32 mm. The tract height of T6 is 4.08 mm. This
measurement may represent a minimum height for this tract in
the taxon as all mis with a crown height of less than 4.00 mm
have interrupted enamel, and three m 1 s that have a crown height
equal to or greater than 4.32 mm have the enamel interrupted
at the site ofT6. These three teeth represent immature individuals
as well and are from the Big Springs l.f. Three of the four m 1 s
mentioned above and an immature individual from the White
Rock l.f. lack interrupted enamel on the PL as well. The crown
height and the tract height of the buccal tract and lingual tract,
respectively are 4.06, 3.71, 3.21; 4.32, 3.90, 3.78; 4.32, I, 4.25;

4.43, 3.85, 3.70 mm. The immature specimen from the Big Springs
that exhibits interrupted enamel on the PL has a crown height
of 4.37. These measurements may represent the potential range
in tract height at these sites. Other measurements of the m 1 s are
listed in Table 4.

Isolated m2s and m3s were considered to pertain to this taxon
if they met the generic criteria of constricted and anteriorly di-
rected reentrant angles, and thick enamel on the posterior face
of the PL. Four of 20 m2s do not exhibit interrupted enamel on
the buccal side of the PL and seven of 20 do not exhibit this
condition on the lingual side. Measurements of tract height and
other parameters of the m2s are listed in Table 5. Ranges of tract

212

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 1 . — Measurements of Ml in Hibbardomys and Pliophenacomys.

Taxon

N

Mean

SD

Observed range

N

Mean

SD

Observed range

GL

GW

H. skinneri

2

2.49

0.18

2.36-2.62

2

1.57

0.12

1.48-1.65

H. fayae

14

2.40

0.13

2.17-2.64

14

1.50

0.09

1.32-1.66

H. voorhiesi'

10

2.40

0.09

2.27-2.52

10

1.48

0.08

1.36-1.61

H. voorhiesi 2

22

2.33

0.13

2.07-2.55

22

1.44

0.11

1.14-1.58

P. dixonensis

6

2.39

0.05

2.33-2.46

6

1.40

0.07

1.31-1.50

P. osborni 1

1

2.42

1

1.43

P. osborni2

9

2.37

0.08

2.25-2.50

9

1.41

0.05

1.30-1.48

GH

LLRA3

H. skinneri

2

3.25

0.07

3.20-3.30

2

0.25

0.02

0.23-0.26

H. fayae

14

3.07

0.78

2.1 1-4.73

14

0.21

0.03

0.17-0.27

H. voorhiesi'

10

4.71

0.55

3.38-5.22

10

0.24

0.03

0.20-0.29

H. voorhiesP

22

4.00

0.56

2.50-4.85

22

0.23

0.03

0.18-0.29

P. dixonensis

6

2.79

0.82

1.65-3.90

5

0.35

0.03

0.32-0.39

P. osborni'

1

4.37

1

0.39

P. osborni 2

9

3.73

0.98

1.68-4.94

8

0.35

0.03

0.27-0.37

HBDT

HLDT

H. skinneri

2

0.65

0.13

0.55-0.74

2

0.29

0.12

0.20-0.37

H. fayae

14

1.24

0.33

0.63-1.67

14

0.54

0.14

0.37-0.84

H. voorhiesi'

5

4.57

0.59

3.73-5.19*

9

3.34

0.89

1.44-4.26*

H. voorhiesi 2

15

3.38

0.88

1.42-4.59*

22

1.21

0.80

0.32-2.81

P. dixonensis

4

2.60

0.31

2.37*-3.06

6

1.77

0.25

1.33-2.03*

P. osborni'

1

3.79

1

2.52

P. osborni 2

4

3.74

0.25

3.45-4.04

7

2.61

0.94

1.23-3.66

See Table 3 for abbreviations.

height on the PL of m3 cannot be established with certainty
because of the small sample (N = 5). Two teeth do not exhibit
interrupted enamel on the buccal side of the PL; whereas only
one has the enamel interrupted on the lingual side of the PL.
This tooth has a crown height of 1.41 mm. The potential range
of tract heights and other measurements of m3 are listed in Table 6.

The Ml is composed of an AL and four alternating triangles
(Fig. 5E-G). T nangle 1 tends to be rounded rather than triangular.
The AL and triangles are either closed off from each other or a
very thin strip of dentine may connect successive triangles. The
enamel of the anterior faces of the triangles is thinner than on
the posterior faces of the AL. Dentine tracts are present on both
sides of the AL, Tl, and the posterior end of T4. The tracts are
shorter on the lingual side of the AL so that the enamel pattern
is interrupted later in the wear of the tooth than on the buccal
side. None of the Mis from White Rock exhibit an interrupted
enamel pattern on the lingual side of the AL and only two of ten
from the Big Springs show interruption at this site. Potential range
of tract heights as well as other measurements are given in Table
7. Apices of the reentrant angles tend to be directed posteriorly
except for LRA3, which, if present, is very shallow and directed
anteriorly. Anatomical roots do not develop until relatively late
in the life of an individual. The majority of teeth that have
developed anatomical roots have two, a few have three. The third
root, if present, lies under Tl and is greatly reduced. More gen-
erally it fuses with the anterior root, which underlies the AL, to
form a single large root.

The M2 is similar in occlusal pattern to the M 1 except that it
has one less triangle and the differences between the dentine tract
height on the sides of the AL do not appear to be as great (Fig.

5H-J). All seven M2s from the Big Springs exhibit an interrupted
enamel pattern on the buccal side of the AL. The maximum
crown height from this sample is 4.91 mm. Three of the seven
teeth do not exhibit an interrupted pattern on the lingual side.
But the heights of these tracts fall within the range exhibited by
those teeth that are interrupted. Nine of the 13 M2s exhibit
interrupted enamel on the buccal side of the AL and eight of 1 3
are interrupted on the lingual side. The range of tract heights are
listed in Table 8 along with other measurements of the M2s. The
tooth has two roots.

The M3 consists of an AL, three alternating triangles and a
PL. Infrequently, a shallow LRA 3 is present on the PL so that
a very small fourth triangle is formed (Fig. 5B-D). Dentine tracts
are present on the sides of the AL and the posterior face of the
PL. The enamel of the triangles is thinner on the anterior faces
and thicker on the posterior, whereas the enamel of the loops is
of equivalent thickness on both faces. Measurements of M3 are
found in Table 9. The tooth has two roots.

Discussion. — That the specimens assigned to Hib-
bardomys might represent a distinct taxon has been
known for some time (Hibbard and Zakrzewski,
1 967). As mentioned above, Eshelman (1975) ques-
tionably placed the specimens from the White Rock
into Pliophenacomys and stated that the White Rock
and Mullen specimens appeared to be more similar
to Pliomys deeringensis Guthrie and Matthews from
the Cape Deceit l.f. (Irvingtonian of Alaska) than to

1984

ZAKRZEWSKI - BLANCAN ARVICOLINES

213

Table 8.— Measurements of M2 in Hibbardomys and Pliophenacomys.

Taxon

N

Mean

SD

Observed range

N

Mean

SD

Observed range

GL

GW

H. skinneri

3

2.1 1

0.18

1.91-2.22

3

1.40

0.14

1.24-1.49

H. fayae

10

1.98

0.05

1.91-2.08

10

1.31

0.06

1.22-1.45

H. voorhiesi'

7

2.00

0.07

1.88-2.07

7

1.34

0.06

1.28-1.42

H. voorhiesi2

13

1.99

0.06

1.86-2.07

13

1.33

0.08

1.23-1.47

P. primaevus

4

2.11

0.10

2.00-2.23

4

1.37

0.12

1.22-1.49

P. dixonensis

1

1.95

1

1.16

P. osborni'

1

2.05

1

1.28

P. osborni 2

7

1.98

0.06

1.88-2.07

7

1.27

0.05

1.21-1.34

GH

LLRA2

H. skinneri

3

2.74

0.33

2.36-2.93

3

0.22

0.05

0.17-0.25

H. fayae

10

3.17

0.66

1.73-3.77

9

0.21

0.03

0.17-0.27

H. voorhiesi'

7

3.94

0.54

3.34-4.91

7

0.24

0.03

0.22-0.28

H. voorhiesi 2

13

3.02

0.82

1.44-4.08

13

0.23

0.03

0.20-0.28

P. primaevus

4

2.82

0.96

1.52-3.81

4

0.30

0.02

0.28-0.33

P. dixonensis

1

3.26

1

0.30

P. osborni'

1

4.54

1

0.39

P. osborni 2

7

2.66

0.45

2.16-3.48

7

0.34

0.03

0.29-0.38

HBDT

HLDT

H. skinneri

3

0.77

0.06

0.72-0.82

3

0.80

0.12

0.69-0.93

H. fayae

6

2.70

0.45

1 ,86*-3. 10

7

2.28

0.23

1.97*-2.65

H. voorhiesi'

7

I

4

3.75

0.16

3.52-3.90*

H. voorhiesi2

5

2.77

1.16

0.76-3.77*

6

3.04

0.40

2.60-3.77

P. primaevus

4

0.83

0.22

0.58-1.08

4

0.87

0.12

0.74-1.02

P. dixonensis

1

2.40

1

2.13

P. osborni'

I

I

1

I

P. osborni 2

1

2.87

2

2.49

0.44

2.18-2.80*

See Table 3 for abbreviations.

Table 9.

—Measurements of M3 in Hibbardomys and Pliophenacomys.

Taxon

N

Mean

SD

Observed range

N

Mean

SD

Observed range

GL

GW

H. skinneri

1

1.63

i

1.05

H. fayae

5

1.94

0.23

1.65-2.23

5

1.17

0.08

1.08-1.30

H. voorhiesi'

10

1.89

0.20

1.53-2.10

10

1.07

0.09

0.93-1.21

H. voorhiesi 2

9

1.99

0.10

1.82-2.10

9

1.18

0.06

1.08-1.22

P. primaevus

4

1.90

0.08

1.79-1.99

4

1.08

0.14

0.96-1.27

P. dixonensis

4

1.73

0.11

1.62-1.87

4

1.08

0.07

1.01-1.09

P. osborni 2

7

1.76

0.10

1.63-1.87

7

1.05

0.10

0.90-1.15

GH

LLRA1

H. skinneri

1

2.74

1

0.25

H. fayae

5

2.32

0.62

1.49-2.94

5

0.48

0.11

0.34-0.63

H. voorhiesi'

10

2.85

0.67

1.72-3.95

10

0.47

0.05

0.40-0.55

H. voorhiesi 2

9

2.23

0.62

1.26-2.98

9

0.51

0.04

0.45-0.58

P. primaevus

4

1.99

0.59

1.12-2.39

4

0.27

0.04

0.22-0.30

P. dixonensis

4

2.16

0.02

2.14-2.19

4

0.26

0.09

0.18-0.35

P. osborni 2

7

2.33

0.50

1.35-2.75

7

0.26

0.04

0.20-0.31

HBDT

HLDT

H. skinneri

1

0.72

1

0.76

H. fayae

3

1.58

0.55

1.13-2.19

4

1.45

0.40

1.20-2.04*

H. voorhiesi'

6

2.20

0.77

0.80-3.06

6

2.31

0.64

1.48-3.07

H. voorhiesi 2

3

2.11

0.49

1.56-2.48*

2

2.57

0.55

2.18-2.96*

P. primaevus

4

0.82

0.13

0.68-1.00

4

0.91

0.12

0.82-1.09

P. dixonensis

4

1.61

0.15

1.45-1.74

4

1.55

0.26

1.35-1.93

P. osborni 2

5

2.31

0.15

2.11-2.49

3

2.37

0.23

2.16-2.62

See Table 3 for abbreviations.

214

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

the Pliophenacomys from the Fox Canyon l.f. based
on the height of the dentine tracts and thickness of
enamel. I was of a similar opinion until I had the
opportunity to examine the specimens in detail.

As can be seen from their diagnoses Hibbardomys
is distinct from Pliophenacomys on the basis of the
occlusal pattern of the teeth, especially m 1 and M3.
Hibbardomys is distinct from Pliomys for the same
reason. Pliophenacomys and Pliomys are similar in
the occlusal pattern of the ml and M3. Pliophen-
acomys and Pliomys may be congeneric but I did
not have a large enough sample of Pliomys on which
to base a conclusion.

Based on the small sample of Pliomys deeringen-
sis I examined it appears that the species belongs to
neither Hibbardomys nor Pliomys but, again, a larg-
er sample is needed before a judgement can be made.

Characteristics of Hibbardomys such as the oc-
clusal pattern of M3, reentrant pits on the posterior
triangles of Ml and M2, prism folds on the ml of
the primitive species, and constricted, anteriorly-
directed apices on the reentrants of the lower molars
suggest that it may have evolved from some species
of Cosomys. Only in Cosomys are all of the above
characters found, but no Cosomys has a well de-
veloped dentine tract at the position of T6. Og-
modontomys and Ophiomys are other possibilities,
but I know of no Ophiomys with reentrant pits on
the upper teeth and although primitive species have
a complex ACC, prism folds and pits seem to be
lost fairly early in phylogeny of the genus. A similar
statement can be made regarding Ogmodontomys
and in addition it retains a three rooted M3 and the
lingual reentrants of the m 1 in the most advanced
species tend to be broad and perpendicular to the
midline of the tooth. At present Hibbardomys is
known from only four sites in the northern and east-
ern parts of the central Plains. Though not the most
common of the arvicolines in the faunas where it is
found it appears to be replacing Pliophenacomys,
which it resembles in gross dental morphology (ml
with five to seven triangles). Hibbardomys becomes
more abundant with respect to Pliophenacomys in
the younger White Rock and Big Springs local fau-
nas.

Pliophenacomys Hibbard, 1938
(Fig. 3-4, 6)

Phenacomys ( Pliophenacomys ) Hibbard, 1938; Trans. Kansas

Acad. Sci., 40:248-249.

Pliophenacomys Hibbard, 1950; Contr. Mus. Paleo. Univ. Mich-
igan, 8:150-157.

Emended diagnosis.— Pliophenacomys is a prim-
itive vole with rooted teeth, lacking cement in the
reentrant angles. The lingual reentrants of the lower
molars tend to be broader than in Hibbardomys with
their apical angles directed nearly perpendicular to
the midline of the tooth. The apices of the reentrants
do not constrict as they approach the midline.

The enamel of the occlusal surface tends to be
differentiated into thick and thin segments. The an-
terior faces of the alternating triangles and posterior
loop in the lower molars exhibit thick enamel,
whereas the posterior faces of the same elements
exhibit thin enamel.

The ml is composed of a PL, five to seven alter-
nating triangles, and an AC. Dentine tracts on the
ACC are confined generally to the buccal side (AC
and T6).

The M3 is composed of an AL, two alternating
triangles, and a PL. Triangle 1 is poorly developed,
confluent with the AL, and generally closed off- from
T2. The BRA1 is very shallow (<0.35 mm). The
M3 has two roots. The posterior lingual reentrants
of Ml (LRA3) and M2 (LRA2) tend to be deeper
(>0.27 mm) than in Ogmodontomys, Ophiomys, or
Hibbardomys.

Recognized species and occurrence. — P. finneyi Hibbard and
Zakrzewski, 1972; Fox Canyon; P. primaevus Hibbard, 1938;
Rexroad Loc. 2 and Sand Draw; P. dixonensis, new species; Dix-
on; P. osborni Martin 1972, Mullen, Big Springs, and White
Rock.

Pliophenacomys primaevus Hibbard, 1938
(Fig. 3C-E)

Phenacomys (Pliophenacomys) primaevus Hibbard, 1938; Trans.

Kansas Acad. Sci., 40:248-249.

P. primaevus Hibbard, Hibbard 1972 (in part); Bull. Amer. Mus.

Nat. Hist., 148:102-105.

Holotype. — KU 3905, right dentary with ml-m2.

Type locality. — Rexroad formation, Rexroad Loc.
2, Meade County, Kansas.

Referred material. — Rexroad Loc. 2, KU 5976 ml, m3. Sand
Draw l.f., Magill: UM 57183, ml; UM 81814, m2; UM 81816,
4 M2s; UM 8 1 8 1 7, 2 M3s; Owl Pellet: UM 8 1 8 1 3, 2 partial m 1 s;
UM 59815, 2 m2s; UM 81815, 2 m3s; UM 81818, 2 M3s; UM
57181, M1-M3.

Emended diagnosis.— Pliophenacomys primaevus
is slightly more hypsodont and has higher dentine
tracts than P. finneyi. It has shorter dentine tracts
than the remaining taxa in the genus.

Description. — Remains of Pliophenacomys pri-
maevus (Fig. 3C-E) possess the characteristics listed

1984

ZAKRZEWSKI — BLANCAN ARVICOLINES

215

in the generic diagnosis, therefore, additional de-
tailed description is not warranted. Measurements
of various parameters of the teeth are listed in Ta-
bles 4-9. These measurements suggest that the four
species of Pliophenacomys are similar in size and
separated on the development of the dentine tract
on the PL of m 1 . Although the dentine tract on the
buccal side of the posterior loop is much higher in
P. primaevus than it is in P. finneyi, in no specimen
is the tract high enough to interrupt the enamel on
the occlusal surface until substantial wear has oc-
curred. The dentine tract on the lingual side (Fig.
3E) of the PL is much shorter than that on the buccal
(Fig. 3D), a characteristic that can be used to sep-
arate P. primaevus from P. osborni.

Pliophenacomys dixonensis, new species
(Fig. 4G-L)

Holotype.- UM 55444, Lml.

Paratypes. — UM 54946, 54949, 5 mis; 55447, 5 m2s; 54946,
m3; 81787, 6 Mis; 81788, M2; 55452, 55457, 4 M3s.

Horizon and type locality. — “Belleville” forma-
tion (Hibbard, 1972), Dixon l.f., Kingman County,
Kansas.

Diagnosis.— Pliophenacomys dixonensis is distin-
guished by the relative height of its dentine tracts,
which are higher than those of P. primaevus but
shorter than those of P. osborni.

Etymology’.— This species is named for its occurrence in the
Dixon l.f.

Description of holotype.— The holotype (Fig. 4G-
I) possesses the characters of the genus. It represents
an individual in the adult stage of development. The
tooth measures 2.99 mm in length, 1.20 mm in
width, and 2.69 mm in crown height. The maximum
height of the buccal dentine tract on the PL cannot
be determined, as the crown has been reduced enough
so that the enamel pattern of the occlusal surface is
interrupted at this point. The height of this tract was
at least 2.69 mm, whereas the tract on the lingual
side of the PL is 2.13 mm high. The tract on the
anterior face of the AC is also high enough to in-
terrupt the enamel of the occlusal surface at this
point.

Description of paratypes.— The remaining mis are similar to
the holotype. Only one of the mis has a dentine tract on the
buccal side of the PL that does not interrupt the enamel of the
occlusal surface. This specimen is in an immature stage of de-
velopment (no anatomical roots have developed). It has a crown
height of 3.20 mm and the height of the buccal tract is 2.80 mm.

The remaining mis have a crown height below 2.70 mm. Mea-
surements of ml are listed in Table 4.

Only one of the m2s has a dentine tract on the buccal side of
the PL that does not interrupt the enamel of the occlusal surface.
The height of the tract is 2.12 mm, whereas the crown height is
2.44 mm. A height of 2.12 mm is not the maximum height as
m2s with crown heights of 2.25 and 2.57 mm have interrupted
enamel patterns on the buccal side. Interrupted enamel is not
established on the lingual side of the PL until substantial wear
takes place. A m2 with the crown height of 1.13 mm exhibits an
interrupted enamel pattern on the lingual side of the PL. The
highest tract of 1 .65 mm does not exhibit an interrupted pattern.
Pliophenacomys dixonensis is similar to P. primaevus in that the
buccal tract is better developed than the lingual tract on the PL
in the lower molars.

In addition to having the LRA3 deeper, Mis of P. dixonensis
can be distinguished from those of Hibbardomys by the height
of the dentine tracts on the sides of the AL. Buccal tracts are
probably greater than 2.00 mm and lingual tracts greater than
1.30 mm. Only one of six Mis had three distinct roots. Three
of the six had the small median root fused to the large anterior
one, and the remaining two had only two roots.

The only M2 that is known has two roots like P. primaevus.
Other characters of the M2 and those of the M3 (Fig. 4J-L) are
typical of the genus; except for the tract height which is typical
for the species. Measurements for the teeth can be found in Tables
5-9.

Pliophenacomys osborni Martin
(Fig. 6)

Pliophenacomys osborni Martin, 1972 (in part). Bull. Nebraska
St. Mus., 9:174-176.

Pliophenacomys ? osborni Martin; Eshelman, 1975 (in part). Pa-
pers Paleo. Univ. Michigan, 13:41-43.

Holotype. — UNSM 392 1 6, partial left dentary with
m l-m3.

Type locality. — Unnamed formation, UNSM Coll.
Loc. Ho- 103, Hooker County, Nebraska.

Referred material. — UNSM Coll. Loc. Cr. 10, Pit 3, 39568, ml;
Big Springs l.f., UNSM 52707, 52731, 52734, 52740, 52747,
52749, 52752, 52763, 52772, 52787, 90912, 90913, 12 mis;
90914, 90915, 2 m2s; 90916, m3; 90917, 90918, 2 Mis; 90919,
90920, 2 M2s; White Rock l.f., UM-K1-66: UM 81789, 3 mis;
81790, 2 m2s; 81791, M2; 81792, M3. UM-K3-69: 81793, 11
mis; 81794, 8 m2s; 81795, 4 m3s; 81796, 11 Mis; 81797, 5
M2s; 81798, 2 M3s. UM-K4-72: 81799, 2 mis; 81800, m3;
81801, 5 Mis; 80802, 3 M2s; 81803, 2 M3s. UM-K4-72 + 12'
SE: 61738, ml; 81804, m2; 81805, Ml; 81806, 2 M2s. UM-K5-
72: 81807, 2 m2s; 81808, 4 m3s; 81809, 9 Mis; 81810, M2;
81786, 8181 1, 3 M3s. UM-K6-72: 61741, 2 mis; 81812, Ml.

Emended diagnosis. — Pliophenacomys osborni is
distinguished from the other species in the genus by
having the highest dentine tracts on the sides of the
PL in the lower molars and by the fact that the
lingual tract is generally as high as the buccal tract.

Description. — With the exception of the charac-
ters mentioned in the emended diagnosis P. osborni

216

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 6 . — Pliophenacomys osborni. A-C) UNSM 52772, Rml; A) occlusal view, B) buccal view, C) lingual view. D-F) UM 81812, LM1,
D) occlusal view, E) lingual view, F) buccal view. G-I) UM 81810, LM2, G) occlusal view, H) lingual view, I) buccal view. J-L) UM
8 1 786, LM3, J) occlusal view, K) lingual view, L) buccal view. Short bar equals 1 mm on side views, long bar equals 1 mm on occlusal
views.

does not differ significantly from P. primaevus or P.
dixonensis. Although specimens of P. osborni ap-
pear to be more hypsodont than those of P. pri-
maevus or P. dixonensis , measurements of crown
height (Tables 4-9) do not readily support this view.

Only two of the m 1 s (Fig. 6A-C) of P. osborni do
not have an interrupted enamel pattern on the PL.
These teeth (UNSM 52752 and 52731) are both
from the Big Springs l.f. and in a young adult stage
of development. The crown height, buccal tract
height, and lingual tract height of these specimens
are 3.48, 3.38, 2.60 and 3.26, 3.09, 3.13 respec-

tively. Ten specimens with crown heights greater
than these have interrupted patterns.

The two m2s from the Big Springs l.f. are prob-
lematical. UNSM 90915 has a buccal tract height
of 0.66 mm and a lingual tract height of 0.41 mm,
whereas UNSM 90914 with a crown height of 3.34
mm has an interrupted occlusal pattern. These teeth
were assigned to Pliophenacomys on the shape of
the reentrant angles and the presence of thin enamel
on the posterior face of the PL. Though variable the
five m2s from the White Rock appear to represent
a more coherent sample. The two specimens from

1984

ZAKRZEWSKI — BLANCAN ARVICOLINES

217

the Big Springs either represent the extremes in pop-
ulation variation or pertain to different taxa.

The m3 from Big Springs is substantially different
in tract height. It could represent a different taxon
or the Big Springs shows a clinal relationship in tract
height with the White Rock.

No Mis (Fig. 6D-F) of P/iophenacomys osborni
were found that exhibit three distinct roots. Of those
specimens that exhibited anatomical roots, the ma-
jority had a slight groove or depression on the me-
dian lingual side of the anterior root suggesting some
fusion had taken place. The remaining teeth were
at a stage of development in which no anatomical
roots had formed. Crown height and wear on the
occlusal surface suggests that anatomical roots de-
velop later in life in P. osborni than they do in the
other species of the genus. Except for a higher den-
tine tract height and a slight increase in hypsodonty
with the correlative retardation of the development
of anatomical roots the M2s (Fig. 6G-I) and M3s
(Fig. 6J-K) of P. osborni do not differ from those
of other species of Pliophenacomys described in this
paper.

Discussion. — After its occurrence in the Fox Can-
yon l.f. where Pliophenacomys is represented by at
least two and perhaps as many as three thousand
individuals it becomes a rare element in subsequent

Blancan faunas. The reason for this rarity is not
intuitively obvious, because when it first appears
Pliophenacomys is relatively advanced in terms of
dentition when compared to coeval arvicolines. The
advancement continues as shown by the increase in
hypsodonty, increase in dentine tract height, de-
crease in number of roots on Ml and 2, and the
formation of roots later in life. Perhaps the rarity is
a function of sampling, with one exception, Loc.
2A, all non-finneyi Pliophenacomys are found well
to the north and east of the well-sampled Meade
County area. In addition to geography perhaps Plio-
phenacomys was an upland dweller and, therefore,
not well represented in fossiliferous deposits that
are primarily lowland in origin.

Also of interest is the decrease in number of Plio-
phenacomys with respect to Hibbardomys through
time. When the two genera occur together for the
first time in the Sand Draw l.f., though both taxa
are rare, they are represented by an approximately
equal number of individuals. However, in the later
faunas Hibbardomys outnumbers Pliophenacomys
approximately 2 to 1 . This relationship could be a
function of sampling, reflect habitat preference, or
replacement of one by the other. Additional mate-
rial at other localities will be necessary to determine
which of these alternatives is correct.

ACKNOWLEDGMENTS

I thank the following individuals for their aid in this study; for
the loan of specimens in their care Drs. Michael R. Voorhies,
Vertebrate Paleontology Section, University of Nebraska State
Museum, Lincoln (UNSM), Philip D. Gingerich, Museum of
Paleontology, University of Michigan, Ann Arbor (UM), Charles

A. Repenning, United States Geological Survey, Menlo Park,
California, Jerry R. Choate, Museum of the High Plains, Fort
Hays State University; for critical reading of the manuscript Drs.
Choate, Repenning, and Voorhies; for typing of the manuscript
and preparation of the line drawings Gwenne Cash.

LITERATURE CITED

Eshelman, R. E. 1975. Geology and paleontology of the early
Pleistocene (late Blancan) White Rock fauna from the north-
central Kansas. Univ. Michigan Papers Paleo., 13 (C. W.
Hibbard Mem. Vol. 4): 1-60.

Guthrie, R. D., and J. V. Matthews. 1971. The Cape Deceit
fauna— early Pleistocene mammalian assemblage from the
Alaskan Arctic. Quaternary Res., 1:474-510.

Hibbard, C. W. 1972. Class Mammalia. In Early Pleistocene
pre-glacial and glacial rocks and faunas of north-central Ne-
braska (M. F. Skinner and C. W. Hibbard et al.) Bull. Amer.
Mus. Nat. Hist., 148:1-148.

Hibbard, C. W., and R. J. Zakrzewski. 1967. Phyletic trends
in the late Cenozoic microtine Ophiomys gen. nov., from
Idaho. Contr. Mus. Paleo., Univ. Michigan, 21:255-271.

Martin, L. D. 1972. The microtine rodents of the Mullen
Assemblage from the Pleistocene of north central Nebraska.
Bull. Univ. Nebraska St. Mus., 9:173-182.

Meulen, A. J. van der. 1973. Middle Pleistocene smaller
mammals from the Monte Peglia, (Orvieto, Italy) with spe-
cial reference to the phylogeny of Microtus (Arvicolidae,
Rodentia). Quatemaria, 17:1-144.

Zakrzewski, R. J. 1967. The primitive vole, Ogmodontomys,
from the late Cenozoic of Kansas and Nebraska. Papers
Michigan Acad. Sci., Arts, & Letters, 52:133-150.

. 1969. The rodents from the Hagerman local fauna,

upper Pliocene of Idaho. Contr. Mus. Paleo., Univ. Michi-
gan, 23:1-36.

Address: Sternberg Memorial Museum and Department of Earth Sciences, Fort Hays State University,
Hays, Kansas 67601.

GEOGRAPHIC DIFFERENTIATION IN THE RANCHOLABREAN DIRE
WOLF ( CANIS DIRUS LEIDY) IN NORTH AMERICA

Bjorn Kurten

ABSTRACT

The late Rancholabrean dire wolf of California and Mexico species. Within the nominate subspecies, which inhabited North

differs from the nominate subspecies in details of dental pro- America east of the Rocky Mountains, microevolution on the

portions and especially in the limb bones which average more deme level may be observed in the course of the Rancholabrean.

than 10% shorter, and is referred to C. dims gidldavi, new sub-

INTRODUCTION

The discovery of the extinct dire wolf, Canis dims
Leidy, dates back to 1854, but it was only with the
excavations at Rancho La Brea that adequate ma-
terial of this species became known (Merriam, 1912).
The species has now been found at about 100 sites
in North America (Nowak, 1979), ranging from
southern Canada to Mexico. It is also known from
South America (Churcher, 1959), but has not been
recorded from Alaska or northern Canada.

With such a range in space, and also an appre-
ciable stratigraphic range, some degree of local and/
or temporal differentiation may well be expected.
In fact a number of species of the dire wolf group
have been proposed from time to time. However,
study of such infraspecific differentiation has been
hampered by lack of adequate statistical data. Ob-
viously, the great collection from Rancho La Brea
would be an ideal sample for population studies. To
date, only a number of limb elements have been
studied in this manner (Nigra and Lance, 1947; Stock
and Lance, 1 948). Comparison with limb bones from
other sites indicates that geographic differentiation

existed (Hawksley et al., 1963; Galbreath, 1964;
Lundelius, 1972; Kurten and Anderson, 1980).
However, limb bones are not particularly common
except at the major fossil-producing sites. Still, when
limb bones from a large number of sites were pooled,
a very marked difference between the dire wolves
of Rancho La Brea on one hand, and those from
North America east of the Rocky Mountains on the
other, became apparent.

A metric study of dire wolf dentitions revealed
that concomitant differences between western and
eastern-central dire wolves existed, as well as geo-
graphic or temporal differentiation on the deme level.
Unfortunately, lack of time prevented a study of the
main Rancho La Brea collections in Los Angeles,
but a respectable sample was obtained in other col-
lections, the largest being that of the Field Museum
of Natural History, Chicago. A second important
dire wolf mass occurrence is at San Josecito Cave,
Mexico; this material was studied in the Natural
History Museum of Los Angeles County.

METHODS AND MATERIALS

Material from a total of about 40 sites was studied in the course
of 1970-1972, and additional material seen in 1981. The work
was partially assisted by National Science Foundation Grant No.
GB 31297 to Bryan Patterson, Harvard University. Support was
also given by the American Educational Foundation and by the
Societas Scientiarum Fenmca.

Many persons have made material and/or information avail-
able, among them E. Anderson, W. Auffenberg, C. S. Churcher,
W. W. Dalquest, T. Downs, D. Fortsch, J. E. Guilday, O. Hawks-
ley, C. W. Hibbard, R. Hoffman, E. L. Lindsay, E. L. Lundelius,
R. M. Nowak, P. Parmalee, B. Patterson, C. E. Ray, K. Richey,
H. A. Richmond, D. E. Savage, C. B. Schultz, H. Semken, G. G.
Simpson, B. Slaughter, J. Soiset, R. Tedford, W. B. Turnbull, S.
D. Webb, D. Whistler, and J. A. White.

Institutions and collections are identified by the following ac-
ronyms:

AMNH American Museum of Natural History
ANSP Academy of Natural Sciences at Philadelphia
BM(NH) British Museum (Natural History)

CIT California Institute of Technology, collection now in

LACM

CM Carnegie Museum of Natural History

CMS Central Missouri State University

FM Field Museum of Natural History

FSM Florida State Museum, University of Florida

ISUM Idaho State University Museum

KUM University of Kansas Museum

218

KU RTEN — CAN IS DIRUS TAXONOMY

219

1984

LACM Los Angeles County Museum of Natural History
MCZ Museum of Comparative Zoology, Harvard Univer-
sity

SMU Southern Methodist University Museum of Paleon-

tology

TMM Texas Memorial Museum

UCMP University of California Museum of Paleontology
UMMP University of Michigan Museum of Paleontology

UNSM University of Nebraska State Museum
USNM National Museum of Natural History

All measurements are in millimeters. Most are self-explana-
tory. Palatal length was measured along midline from prosthion
to internal narial opening. Width of upper molars is maximum
width from anteroextemal base of crown to most distant lingual
point; M' length was measured parallel to buccal face of crown.
Limb bone lengths are maximum lengths.

TAXONOMY

Genus Canis Linaeus, 1758
Subgenus Aenocyon Merriam, 1918

Type species. — Canis dims Leidy, 1858.
Diagnosis (mainly adapted from Merriam,
1 9 1 2). — Large Canis with very big head; jaws elon-
gate; camassial teeth larger and more massive than
in other Canis ; palate, frontal region, and zygomatic
arches relatively broad; sagittal crest high; inion
overhanging; nasal bones extending relatively far
back; nasal processes of frontals relatively short;
postpalatine foramina opposite to posterior ends of
P4; optic foramen and foramen lacerus anterior in
common pit; M1 with reduced hypocone; P4 with
somewhat reduced protocone.

Canis dirus Leidy, 1858

Synonyms under subspecies.

Diagnosis. — Sole known species of the subgenus
Aenocyon.

Stratigraphic range. — Rancholabrean Land-
Mammal Age.

Geographic range.— North America from south-
ern Canada to the south; western South America.

Canis dirus dirus Leidy, 1858

Canis primaevus Leidy, 1 854 (not C. primaevus Hodgson, 1833).
Canis dirus Leidy, 1858.

Canis indianensis Leidy, 1869.

Canis mississippiensis Allen, 1876.

Cams ayersi Seliards, 1916.

Type. — ANSP 11614, left maxilla.

Type locality. — Evansville, Indiana; age late Wis-
consin (9,400 ± 250 B.P., Hester, 1967).

Diagnosis. — Limb bones longer, especially dis-
tally, than in C. d. guildayr, P2 relatively short.
Stratigraphic range. — Rancholabrean.

Geographic range. — Range of species east of North
American continental divide; possibly Idaho, Utah.

Material referred to C. d. dirus'. — Arredondo, Alachua Co.,
Florida. FSM 2571 tibia, 2977 P4. Age Sangamon.

Aucilia River, Jefferson Co., Florida. Maxilla, teeth. Age Wis-
consin.

Bat Cave, Pulaski Co., Missouri. CMS, remains of at least four
individuals as listed in Hawksley et al. (1963, 1973). Age late
Wisconsin, probably about 16,000-10,000 B.P.

Blue Mounds, Dane Co., Wisconsin. MCZ 10988 humerus, tibia.
Type material of C. mississippiensis Allen (1876). Age presum-
ably Rancholabrean.

Brynjulfson Caves, Boone Co., Missouri. CMS, radius, radius
fragment, MC 2, MC 3-5, MC 4, 2 MT 4, MT 5. Age late Wis-
consin to Early Holocene, 9,440 ± 760 B.P.

Carroll Cave, Camden Co., Missouri. Partial skeleton, data from
Hawksley et al. (1963, 1973). Age Wisconsin.

Cherokee Cave, St. Louis Co., Missouri. Metapodials, data from
Simpson (1949). Age late Wisconsin.

Cragin Quarry, Meade Co., Kansas. KUM 4613 mandible, M1.
Age Sangamon.

Eichelberger Cave, Marion Co., Florida. FSM 1622 mandible,
1624 M|, 2177 M1. Age probably Sangamon.

*Flamingo Waterway, Florida. FM 2367 1 M,. Age probably Wis-
consin.

Frankstown Cave, Blair Co., Pennsylvania. CM, remains of 4
individuals as listed by Peterson (1926). Age Wisconsin.

Friesenhahn Cave, Bexar Co., Texas. TMM 933, 2 mandibles, 2
C1, P3, M1, 2 C,, 3 humeri, 3 radii, 3 femora, tibia, 2 MC 3, MT
4, fragments. Age Wisconsin.

Herculaneum, Jefferson Co., Missouri. FM WC 1736, P4, M2.
Age Wisconsin.

Hermit’s Cave, Eddy Co., New Mexico. UNSM 19212 maxilla.
Age Wisconsin, 12,900 ± 350 and 11,850 ± 350 B.P.

Hornsby Sink, Alachua Co., Florida. FSM 3986 P4, 3987 man-
dible, 3988 maxilla. Age Wisconsin.

1 This and subsequent lists record material seen or used on the basis of published

material; in the latter case, the source is indicated. For most sites, further references

may be found in Nowak (1979) and Kurten and Anderson (1980). Records not in
Nowak (1979) or substantially augmenting his data are denoted with an asterisk.

220

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Ichetucknee River, Columbia Co., Florida. FSM 8005 mandible,
8006 maxilla, 8208 mandible, 8209 maxilla, 8210 MC 2, 821 1
MT 4, 12899 mandible, 17717 mandible. Age Wisconsin.

Ingleside, San Patricio Co., Texas. TMM, remains of at least two
individuals as listed by Lundelius (1972). Age early Wisconsin.

Laubach Cave, Williamson Co., Texas. SMU 61172 femur, 61173
humerus. Age may be Wisconsin interstadial, 45,000-25,000 B.P.

♦Merritt Island, Brevard Co., Florida (Kurten, 1965). FSM 8217
P4. Age Wisconsin.

Moore Pit, Dallas Co., Texas. SMU 60979 MT 4. Age Wisconsin,
50,000-25,000 B.P.

Orr Cave, Beaverhead Co., Montana. CM 1 2088 mandible. Age
Wisconsin.

Powder Mill Creek Cave, Shannon Co., Missouri. Skeleton, data
from Galbreath (1964). Age late Wisconsin (13,170 ± 600 B.P).

Reddick, Marion Co., Florida. FSM 2903 humerus, radius, fe-
mur, MC 2-5, MT 2-5, 2923 skull with mandible, 3081 skull,
6595 M,, unnumbered palate, mandible, teeth. Age Sangamon.

Renick Cave, Greenbrier Co., West Virginia. CM 24327 man-
dible. Age Wisconsin.

Sabertooth (Lecanto) Cave, Citrus Co., Florida. AMNH 23404
MC 5, MT 3, MT 5. Age Sangamon or Wisconsin.

Vero, Indian River Co., Florida. MCZ 17945 M,. Type skull of
C. ayersi Sellards (not seen) also from this site. Age late Wis-
consin.

Welsh Cave, Woodford Co., Kentucky. CM 12625 casts of skull,
mandible, various bones (2 individuals). Age late Wisconsin,
12,950 ± 550 B.P.

Zoo Cave, Taney Co., Missouri. Remains of one individual, data
from Flood and Hawksley (1975). Age late Wisconsin, 13,000-
9,000 B.P.

In addition, numerous sites have yielded material probably
referable to C. dirus dims (lists in Nowak, 1979:1 12-1 15, Ar-
kansas, Florida, Illinois, Indiana, Kansas, Louisiana, Missouri,
Nebraska, New Mexico, Oklahoma, Tennessee, Texas, Virginia,
Wisconsin).

Canis dirus guildayi, new subspecies

Type.— CIT 10834, skull, figured Merriam (1912,
Figs. 1-4), and associated mandible (Merriam,
1912, Figs. 1, 5-6).

Type locality. — Rancho La Brea, Los Angeles Co.,
California; age late Wisconsin.

Diagnosis. — Limb bones shorter, especially dis-
tally, than in nominate subspecies; P2 relatively long.
Stratigraphic range. — Sangamon?— Wisconsin.
Geographic range. — California, Mexico.

Material referred to C. d. guildayi.— Coal Quarry, Eldorado
Co., California (Univ. California Loc. No. V-4805). UCMP man-
dible. Age Wisconsin.

*E1 Tajo Quarry, Mexico (Univ. California Loc. No. V-2501).
UCMP 26650, skull and mandible. Age Rancholabrean.

McKittrick, Kern Co., California. UCMP, 2 maxillae, 2 man-
dibles, teeth as listed by Schultz (1938). Age Wisconsin.

♦Potrecito near Cienega, Mexico. AMNFI 67302 mandible. Age
Pleistocene.

Rancho La Brea, Los Angeles Co., California. FM 5 skulls, 2
palates, 24 maxillae, 12 mandibles, teeth, long bones, metapo-
dials; UMMP skull, 6 maxillae, 3 mandibles, teeth; USNM 6
mandibles; AMNH, BM(NH), CM, MCZ, UCMP, 10 skulls, 9
mandibles, teeth, limb bones. Age mostly late Wisconsin.

Samwel Cave, Shasta Co., California. UCMP 9566 mandible,
teeth. Age Wisconsin.

San Josecito Cave, Nuevo Leon, Mexico. LACM 192, 5 skulls,
4 palates, 1 1 maxillae, 46 mandibles, teeth. Age Wisconsin.

Material referred to Canis dirus, subspecies unknown.—
♦American Falls, Power Co., Idaho. ISUM 4800/000052 skull,
47001/17004, 48001/1563, 49001/2623 mandibles. While con-
firming the tentative identification by Gazin (1935, as Canis cf.
dirus), the allocation to subspecies of this and other Idaho ma-
terial remains uncertain. Age Sangamon or early Wisconsin.

♦Dam, Power Co., Idaho. ISUM 26798 mandible with P2_^, 27776
C1. Robust proportions indicate C. dirus. Age Wisconsin, 26,500 ±
3,500 B.P.

♦Fossil Lake, Lake Co., Oregon. UCMP 26910 mandible. It ver-
ifies the tentative recognition of C. cf. dirus based by Elftman
(1931) on a camassial fragment. The dentition is extensively
damaged and the specimen is not subspecifically determinable.
Geographically, it is closer to C. dirus guildayi than to the nom-
inate subspecies. Age early or middle Wisconsin.

Rainbow Beach, Power Co., Idaho. ISUM 2623 mandible. Age
Wisconsin, between 31,300 ± 2,300 and 21,500 ± 700 B.P.

DISCUSSION

Limb proportions

Galbreath (1964), Hawksley et al. (1963, 1973),
and Hood and Hawksley (1975) have commented
upon the large size of dire wolf limb bones, espe-
cially metapodials, from Missouri (see also Nowak,
1979:114). Such large bones were found in Bat,
Brynjulfson, Carroll, Cherokee and Powder Mill
Creek caves. Similar characters are found in other

samples from east of the Rocky Mountains, that is,
from Ichetucknee River and Reddick (Florida), Blue
Mounds (Iowa), Welsh Cave (Kentucky), Franks-
town Cave (Pennsylvania), Friesenhahn Cave, In-
gleside, Laubach Cave, Moore Pit (Texas), as noted
by Kurten and Anderson (1980:171). Occasional
smaller specimens, that is, from Arredondo and Le-
canto Cave (Florida), Cragin Quarry (Kansas), Zoo

1984

KURTEN — CANIS DIR US TAXONOMY

221

Table 1 .—Limb bone lengths in Canis d. dirus compared with
means for C. d. guildayi from Rancho La Brea (RLB).

Bone

N

Observed

range

Mean

SD

RLB1

Humerus

1 1

224-254

243.7

± 3.0

9.9

217.0 ± 0.4

Radius

9

222-259

243.0 ± 4.0

12.0

209.4 ± 0.3

Femur

6

256-278

266.3

± 3.2

7.7

241.8 ± 0.5

Tibia

9

230-267

250.0

± 4.9

14.7

231.6 ± 0.3

MC 2

15

80-98

91.1

± 1.3

4.9

77.2 ± 0.09

MC 3

14

94-111

102.4

± 1.5

5.2

88.1 ± 0.10

MC 4

13

95-110

103.6 ± 1.3

4.6

97.1 ± 0.10

MC 5

15

76-100

89.5

± 1.6

6.2

73.8 ± 0.09

MT 2

9

88-112

99.8

± 2.2

6.7

83.3 ± 0.10

MT 3

9

96-116

106.3

± 2.8

8.4

94.0 ± 0.1 1

MT 4

15

102-127

112.9

± 1.7

6.7

96.2 ± 0.1 1

MT 5

10

88-111

100.6 ± 2.4

7.5

88.1 ± 0.10

1 Data from Stock and Lance (1948) and from Nigra and Lance (1947, for left
metapodials; standard error emended).

Cave (Missouri), and Slaton (Texas) may be regard-
ed as distal variants of this population.

In Table 1 all of this material has been pooled
and compared with mean values for Rancho La Brea
as published by Nigra and Lance (1947) and Stock
and Lance ( 1 948). As might be expected in a sample
so scattered geographically and temporally (ranging
from the Sangamon to the late Wisconsin), the C.
d. dims material shows a somewhat greater vari-
ability than that from the Californian site. Whereas
the coefficients of variation for the very large Ran-
cho La Brea sample (in which the number of spec-
imens ranges from 313 femora to more than 2,500
MC 5) average 4.05, that for C. d. dirus averages
5.6. The latter figure, although only slightly higher,
may suggest moderate (deme-level) local and/or
temporal differentiation.

The limb bones of C. d. dirus average significantly
longer than those of C. d. guildayi throughout. The
joint overlap, which may be determined by the
method outlined by Mayr et al. (1953), is minimal.
For radius length, for instance, it is only about 4%,
whereas the corresponding figure for length of MC
4 is only about 1%. In effect, then, 99% of a mixed
sample of MC 4 can be allocated to the correct taxon
on this basis, which greatly exceeds the subspecific
differentiation limit (90%) suggested by Mayr et al.

Relative proportions of limb bones also differ
markedly in the two subspecies. Data on the three
main segments of the front and hind limb are sum-
marized in Fig. 1 , which is a ratio diagram (Simpson,
1941) in which Rancho La Brea means serve as the
standard of comparison. Means for other samples

Table 2 .—Lengths of front limb (humerus + radius + MC 3) and
hind limb (femur + tibia + MT 3) in samples of Canis lupus and
Canis dirus.

Front

Sample

Front

Hind

in percent
of hind

C. lupus. MCZ 267

498

540

92.2

C. lupus, sample CL

567

617

91.9

C. dirus guildayi. RLB

515.4

567 .4

90.8

C. d. dirus

589.1

622.6

94.6

are expressed as log and percentage differences from
the standard. Apart from the greater length of all
the limb segments, the nominate subspecies is char-
acterized by relatively longer radii and metapodials
(more than 15% longer than in C. d. guildayi ),
whereas the excess of the length in the humeri, fem-
ora, and tibiae averages only 10%.

Also, it is evident that the front limb as a whole
is longer, in relation to the hind limb, in C. d. dirus.
The sum humerus + radius + MC 3 lengths in C.
d. dirus is 14.3% greater than in C. d. guildayi ; for
femur + tibia + MT 3, the difference is only 9.7%.
A further difference is seen in the fact that the radius
of C. d. dirus is longer relative to the humerus,

Table 3. — Cranial and mandibular dimensions in samples of

Canis d. guildayi and C. d. dirus.

Observed

Sample

N

range

Mean

SD

Condylobasal length

Rancho La Brea

14

252-282

262.0 ± 2.0

7.9

San Josecito Cave
and El Tajo

3

258-270

264.7 ± 2.9

C. d. dirus

6

265-306

284.3 ± 5.7

15.3

Palatal length

Rancho La Brea

15

136-155

143.9 ± 1.2

4.8

San Josecito Cave
and El Tajo

4

134-152

144.0 ± 3.2

C. d. dirus

5

147-168

154.6 ± 3.4

7.6

Width over P^-M1, inclusive
Rancho La Brea 14

88-102

93.8 ± 0.9

3.5

Rancho La Brea1

62

88-104

96.2 ± 0.5

3.9

San Josecito Cave
and El Tajo

5

95-103

99.2 ± 1.6

3.5

C. d. dirus

4

98-101

100.0 ± 0.6

-

Mandibular length

Rancho La Brea

13

197-224

209.1 ± 2.0

7.3

San Josecito Cave
and El Tajo

9

206-218

21 1.0 ± 1.2

3.8

C. d. dirus

6

205-245

221.2 ± 5.6

15.0

1 Data from Nowak (1979:149).

222

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

LOG DIFFERENCE SCALE

- 0.04 -0.02 0 + 0.02 +0.04 +0.00

PERCENTAGE SCALE

Fig. 1. — Ratio diagram showing relative lengths of limb bones in samples of Canis. Standard of comparison: RLB, C. dims guildayi.
Rancho La Brea (means). Other samples are: CDD, C. d. dims (means); CL, C. lupus. Recent (means); MCZ 267, C. lupus, Recent
(small single specimen).

whereas the tibia is not elongated relative to the
femur.

I have included data on Recent Canis lupus in
the ratio diagram. MCZ 267 is a relatively small
specimen, whereas CL denotes mean values for a
sample of eight Recent wolves published by Stock
and Lance ( 1 948). The pattern in both cases deviates
from C. d. guildayi of Rancho La Brea in the same
manner as that of C. d. dims, except that the tibia
is more elongated in C. lupus. In absolute size, the
Stock and Lance sample (CL) is close to C. d. dims.

The relationships between the lengths of the front
limb (humerus + radius + MC 3) and hind limb
(femur + tibia + MT 3) are set down in Table 2. In
Canis lupus and in C. d. guildayi, the front limb
averages 91-92% of the length of the hind. In C. d.
dims it is markedly longer, almost 95%. Such an
elongation of the front limb is commonly seen in
large mammals of the open plains, which typically
have a sloping back.

Skull and Mandible

Measurements of skulls and mandibles are sum-
marized in Table 3. Samples are Rancho La Brea
(RLB), Mexico (SJC + ET), and Canis d. dims
(CDD). Length measurements for Californian and
Mexican samples (C. d. guildayi) coincide closely,
whereas those for the nominate subspecies average
significantly longer. Muzzle width (across P4-M') is
relatively great in the Mexican sample and the dif-
ference from my RLB sample is significant. How-
ever, the larger sample (62 specimens) measured by
Nowak (1979) gave a somewhat higher mean for
Rancho La Brea. Age differences may affect the rel-
ative width of the skull.

Dentition

Table 4 shows measurements of the teeth of Canis
dims. The material was split into four samples —
Rancho La Brea (RLB), San Josecito Cave (SJC),

1984

KURTEN — CAN IS DIRUS TAXONOMY

223

Table 4 . — Tooth dimensions in

samples of Canis d.

. guildayi and C.

d. dirus.

Measure-

ment

Sample

N

Observed range

Mean

SD

Length C1

Rancho La Brea

i i

13.7-16.8

14.86 ± 0.29

0.97

San Josecito Cave

10

14.0-18.0

15.65 ± 0.36

1.13

C. d. dirus (Sangamon-Early Wisconsin)

3

16.5-18.3

17.17 ± 0.57

—

C. d. dims (Late Wisconsin)

5

14.5-16.0

15.02 ± 0.28

0.63

Length P2

Rancho La Brea

12

14.3-17.1

15.73 ± 0.21

0.74

San Josecito Cave

7

15.8-17.2

16.40 ± 0.23

0.60

C. d. dims (Sangamon-Early Wisconsin)

5

15.0-16.6

15.98 ± 0.29

0.64

C. d. dims (Late Wisconsin)

7

14.1-15.5

14.71 ± 0.19

0.51

Width P2

Rancho La Brea

12

6.5-8. 1

7.40 ±0.15

0.52

San Josecito Cave

7

7. 3-7. 9

7.53 ± 0.09

0.25

C. d. dims (Sangamon-Early Wisconsin)

5

6. 7-8. 3

7.62 ± 0.27

0.61

C. d. dims (Late Wisconsin)

6

6. 2-7.8

6.87 ± 0.25

0.62

Length P3

Rancho La Brea

15

16.4-19.7

17.95 ± 0.25

0.98

San Josecito Cave

10

16.4-21.1

18.52 ± 0.39

1 .22

C. d. dims (Sangamon-Early Wisconsin)

6

17.9-21.2

19.10 ± 0.47

1.16

C. d. dims (Late Wisconsin)

1 1

16.6-18.9

17.85 ± 0.28

0.92

Width P3

Rancho La Brea

15

6.9-9. 7

7.95 ± 0.12

0.45

San Josecito Cave

10

7. 4-9.0

8.36 ± 0.15

0.46

C. d. dims (Sangamon-Early Wisconsin)

6

8. 0-8. 9

8.48 ± 0.16

0.38

C. d. dims (Late Wisconsin)

1 1

7. 0-9.0

7.91 ±0.17

0.58

Length P4

Rancho La Brea

46

28.5-35.4

31.32 ± 0.23

1.54

San Josecito Cave

17

29.4-34.1

31.88 ± 0.31

1.28

C. d. dims (Sangamon-Early Wisconsin)

8

30.7-34.8

31.99 ± 0.45

1.28

C. d. dims (Late Wisconsin)

13

29.6-32.5

31.52 ± 0.27

0.96

'Width BP4

Rancho La Brea

48

10.9-14.5

12.25 ± 0.12

0.82

San Josecito Cave

17

11.5-14.2

12.65 ± 0.18

0.76

C. d. dims (Sangamon-Early Wisconsin)

8

12.1-14.2

13.14 ± 0.22

0.63

C. d. dims (Late Wisconsin)

13

1 1.8-13.6

12.73 ± 0.16

0.58

Width M'

Rancho La Brea

56

22.8-28.3

25.40 ± 0.18

1.32

San Josecito Cave

22

22.7-26.7

25.11 ± 0.19

0.90

C. d. dims (Sangamon-Early Wisconsin)

10

24.5-28.5

26.08 ± 0.32

1.02

C. d. dims (Late Wisconsin)

10

24.4-27.9

26.25 ± 0.37

1.17

Length M1

Rancho La Brea

56

17.3-22.0

19.48 ± 0.16

1.18

San Josecito Cave

23

17.6-20.9

19.38 ± 0.19

0.91

C. d. dims (Sangamon-Early Wisconsin)

9

18.5-21.3

19.99 ± 0.29

0.87

C. d. dims (Late Wisconsin)

10

18.7-20.5

19.99 ± 0.17

0.54

Width M2

Rancho La Brea

33

13.2-17.0

14.98 ± 0.17

0.98

San Josecito Cave

15

13.6-16.2

14.85 ± 0.17

0.67

C. d. dims (Sangamon-Early Wisconsin)

9

14.3-17.8

15.94 ± 0.38

1.15

C. d. dims (Late Wisconsin)

8

14.4-16.7

15.60 ± 0.28

0.79

Width C,

Rancho La Brea

19

10.1-13.7

11.51 ± 0.18

0.79

San Josecito Cave

16

10.1-12.4

11.51 ± 0.17

0.68

C. d. dims (Sangamon-Early Wisconsin)

3

12.0-12.3

12.17 ± 0.09

—

C. d. dims (Late Wisconsin)

7

10.5-11.9

11.00 ± 0.25

0.66

Length P2

Rancho La Brea

20

14.4-17.9

15.72 ± 0.20

0.89

San Josecito Cave

30

13.2-16.4

15.22 ± 0.13

0.69

C. d. dims (Sangamon-Early Wisconsin)

4

15.2-16.1

15.70 ± 0.23

0.47

C. d. dims (Late Wisconsin)

9

12.2-15.8

14.78 ± 0.40

1.19

Width P2

Rancho La Brea

20

6. 3-8. 2

7.43 ± 0.09

0.39

San Josecito Cave

30

6.5-8. 1

7.21 ± 0.07

0.39

C. d. dims (Sangamon-Early Wisconsin)

4

6. 5-7. 6

7.15 ± 0.25

0.51

C. d. dims (Late Wisconsin)

10

5. 2-7.5

6.70 ± 0.21

0.67

224

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table A. — Continued

Measure-

ment

Sample

N

Observed range

Mean

SD

Length P3

Rancho La Brea

20

14.1-18.7

16.27 ± 0.22

1.00

San Josecito Cave

32

15.0-17.3

16.04 ± 0.11

0.62

C. d. dirus (Sangamon-Early Wisconsin)

6

15.8-17.5

16.70 ± 0.31

0.75

C. d. dirus (Late Wisconsin)

13

13.5-17.4

16.06 ± 0.30

1.08

Width P3

Rancho La Brea

20

7. 3-9. 2

8.15 ± 0.12

0.52

San Josecito Cave

34

7. 0-8. 9

7.97 ± 0.08

0.44

C. d. dirus (Sangamon-Early Wisconsin)

6

7. 1-8.9

8.15 ± 0.27

0.67

C. d. dirus (Late Wisconsin)

14

6. 7-9.0

7.69 ± 0.23

0.86

Length P4

Rancho La Brea

26

17.6-21.8

19.95 ± 0.19

0.97

San Josecito Cave

32

17.5-20.7

19.60 ± 0.13

0.75

C. d. dirus (Sangamon-Early Wisconsin)

7

18.5-22.0

20.27 ± 0.46

1.23

C. d. dirus (Late Wisconsin)

16

17.8-21.2

19.36 ± 0.32

1.28

Width P4

Rancho La Brea

25

9.4-11.1

10.28 ± 0.10

0.52

San Josecito Cave

33

9.3-11.9

10.20 ± 0.10

0.60

C. d. dirus (Sangamon-Early Wisconsin)

8

9.0-10.8

10.15 ± 0.23

0.64

C. d. dirus (Late Wisconsin)

16

8.3-10.8

9.61 ± 0.18

0.71

Length M,

Rancho La Brea

33

32.9-40.2

35.54 ± 0.27

1.57

San Josecito Cave

44

31.5-39.3

35.50 ± 0.23

1.52

C. d. dirus (Sangamon-Early Wisconsin)

10

32.5-39.1

36.09 ± 0.61

1.93

C. d. dirus (Late Wisconsin)

14

32.1-37.4

35.16 ± 0.52

1.93

Width M,

Rancho La Brea

29

12.3-14.9

13.31 ± 0.13

0.71

San Josecito Cave

40

12.4-15.0

13.62 ± 0.10

0.61

C. d. dirus (Sangamon-Early Wisconsin)

1 1

12.8-15.5

14.10 ± 0.30

0.98

C. d. dirus (Late Wisconsin)

22

11.8-16.3

13.40 ± 0.24

1.13

1 Length TM,

Rancho La Brea

27

22.3-27.0

24.36 ± 0.20

1.02

San Josecito Cave

33

22.0-27.9

24.64 ± 0.19

1.09

C. d. dirus (Sangamon-Early Wisconsin)

6

23.5-26.3

24.53 ± 0.44

1.08

C. d. dirus (Late Wisconsin)

15

21.0-26.4

24.01 ± 0.44

1.72

Length M2

Rancho La Brea

25

11.7-15.5

13.33 ± 0.16

0.79

San Josecito Cave

33

11.2-14.2

12.76 ± 0.14

0.80

C. d. dirus (Sangamon-Early Wisconsin)

7

11.3-15.8

13.74 ± 0.55

1.46

C. d. dirus (Late Wisconsin)

13

12.4-14.8

13.49 ± 0.22

0.81

Width M2

Rancho La Brea

23

9.2-10.5

9.86 ± 0.10

0.47

San Josecito Cave

34

8.9-10.4

9.60 ± 0.06

0.34

C. d. dirus (Sangamon-Early Wisconsin)

5

9.7-11.7

10.48 ± 0.34

0.75

C. d. dirus (Late Wisconsin)

12

8.9-11.5

10.05 ± 0.21

0.73

1 B, blade; T, trigonid.

consinan sample is widely scattered geographically.
(Individual CV values for the two CDD samples are
variable but this appears to be due to small-sample
effects.)

Table 5.— Averages and observed ranges for coefficient of varia-
tion in 22 dental variates of Canis dirus.

Observed Mean

Taxon Sample range CV

C. d. guildayi Rancho La Brea 4.2-6. 5 5.6

C. d. guildayi San Josecito Cave 3. 3-7. 2 4.9

C. d. dirus Sangamon-Early Wisconsin 1.2-10.6 5.5

C. d. dirus Late Wisconsin 3.0-11.2 6.4

Canis d. dirus of probably Sangamonian and early
Wisconsinan age (CDD(l)), and Canis d. dirus of
late Wisconsinan age (CDD(2)). In the ratio diagram
(Fig. 2) the principal dimensions of the teeth are
compared.

To test for possible heterogeneity within samples,
the coefficients of variation for the dental variates
was averaged, with results as given in Table 5. In
an analogous study of four samples of Canis lepo-
phagus, C. arnensis, and C. latrans, the mean CV
(coefficient of variation) varied between 5. 6-5. 9
(Kurten, 1974). In the present case, only the sample
CDD(2) gave a slightly higher average; this late Wis-

1984

KURTEN — CAN IS DIRUS TAXONOMY

225

LOG DIFFERENCE SCALE

-0.12 - 0.10 - 0.08 - 0.06 -0.04 - 0.02 0 + 0.02 - 0.04

PERCENTAGE SCALE

Fig. 2. — Ratio diagram showing relative size of teeth (widths of M'-2, C,, lengths of other teeth) in samples of Canis (means). Standard
of comparison: SJC, C. dims giiildayi, San Josecito Cave. Other samples are: RLB, C. d. guildayi, Rancho La Brea; CDD( 1 ) and CDD(2),
C. d. dims, early and late samples respectively (see text); CL, C. lupus, late Wisconsinan to early Holocene.

The Californian and Mexican samples differ only
moderately from each other and none of the differ-
ences reaches the significance level of P — 0.01 (see
Table 6). However, there are probably slight differ-
ences on the deme level. The two samples of the
nominate subspecies show more consistent differ-
ences from C. d. guildayi, for instance in the rela-
tively small size of P2 and the large size of M1 and
M2. Most measurements for late Wisconsinan C. d.
dims (CDD(2)) are smaller than those for the earlier
(CDD(l)) sample which suggests a slight phyletic
dwarfing within the nominate subspecies. Other de-

tails may be gleaned from the ratio diagram and
from Table 6 in which P-values for significant or
probably significant differences have been collected.

It is of interest to compare the means for C. dims
with those for C. lupus. For this purpose, a late
Wisconsinan/early Holocene sample of C. lupus
(mainly from Moonshiner Cave, Idaho) was studied
and the means recorded in the ratio diagram (CL).
The pattern for C. lupus differs in several respects
from those for the various dire wolf samples, no-
tably in the small size of the canines and carnassials,
and the large size of the postcamassial teeth.

226

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 6. — Selected comparisons between dental dimensions in
samples of Canis dirus. For data and samples see Table 4. Body
of table gives values for P (two-sided).

Rancho La Brea/

San Josecito Cave/

C. d. dirus (Sangamon-

San Josecito Cave

Early Wisconsin)

LP2

<0.05

WM1

<0.02

LM,

<0.02

WM2

<0.02

Rancho La Brea/C. d. dirus

WC, <0.01

San Josecito Cave/

(Sangamon-Early Wisconsin)

C. d. dirus (Late Wisconsin)

LC'

<0.01

LP2

<0.01

LP3

<0.05

lm2

<0.01

WBP4 <0.01

WC, <0.01

Rancho La Brea/C. d. dirus

C. d. dirus (Sangamon-
Early Wisconsin)/

(Late Wisconsin)

C. d. dirus (Late Wisconsin)

LP2

<0.01

LC1

<0.05

WP,

<0.01

LP2

<0.01

WP4

<0.01

WC,

<0.01

Conclusions

The taxa Cams d. dirus and C. d. guildayi differ
in body size, as reflected in the size of the head and
limbs, in limb proportions, and to some extent in
dental proportions. The adaptive significance of the
differences in limb proportions is evident; presum-
ably the nominate subspecies was somewhat fleeter
of foot than C. d. guildayi and this may reflect cer-
tain differences in the mode of life.

Although the two taxa differ in many respects, the
similarity in many characters, especially dental, is
such that, in my opinion, they should be regarded
as subspecies of a single species.

The history of the two subspecies is incompletely
known. Large C. d. dirus with long metapodials date
back to the Sangamon, whereas the first appearance
of the shorter-limbed C. d. guildayi may be slightly
more recent. This, and the resemblance between C.
d. dirus and C. lupus in limb proportions, might
suggest that the condition in C. d. guildayi is the
derived one. As long as the origin of the dire wolf
is unknown such considerations must remain spec-
ulative.

LITERATURE CITED

Allen, J. A. 1876. Description of some remains of an extinct
species of wolf and an exinct species of deer from the Lead
Region of the Upper Mississippi. Amer. J. Sci., 40:47-51.

Churcher, C. S. 1959. Fossil Canis from the Tar Pits of La
Brea, Peru. Science, 130:564-565.

Elftman, H. O. 1931. Pleistocene mammals of Fossil Lake,
Oregon. Amer. Mus. Nov., 481:1-21.

Galbreath, E. C. 1 964. A dire wolf skeleton and Powder Mill
Creek Cave, Missouri. Trans. Illinois State Acad. Sci., 57:
224-242.

Gazin, C. L. 1935. Annotated list of Pleistocene Mammalia
from American Falls, Idaho. J. Washington Acad. Sci., 25:
297-307.

Hawksley, O., J. F. Reynolds, and R. L. Foley. 1973. Pleis-
tocene vertebrate fauna of Bat Cave, Pulaski County, Mis-
souri. Bull. Nat. Spel. Soc., 35:61-87.

Hawksley, O., J. F. Reynolds, and J. McGowan. 1963. The
dire wolf in Missouri. Missouri Speleol., 5:63-72.

Hester, J. J. 1967. The agency of man in animal extinctions.
Pp. 169-192, in Pleistocene extinctions: the search for a
cause (P. S. Martin and H. E. Wright, eds.), Yale Univ. Press,
New Haven and London, x + 453 pp.

Hood, C. H., and O. Hawksley. 1975. The Pleistocene fauna
from Zoo Cave, Taney County, Missouri. Missouri Speleol.,
15:1-42.

Kurten, B. 1965. The Pleistocene Felidae of Florida. Bull.
Florida State Mus., 9:215-273.

. 1974. A history of coyote-like dogs (Canidae, Mam-
malia). Acta Zool. Fennica, 140:1-38.

Kurten, B., and E. Anderson. 1980. Pleistocene mammals of

North America. Columbia Univ. Press, New York, xvii + 443

pp.

Leidy, J. 1854. Notice of some fossil bones discovered by Mr.
Francis A. Lincke, in the banks of the Ohio River, Indiana.
Proc. Acad. Nat. Sci. Philadelphia, 7:199-201.

. 1858. Notice of remains of extinct Vertebrata, from the

Valley of the Niobrara River, collected during the exploring
expedition of 1857, in Nebraska, under the command of
Lieut. G. K. Warren, U.S. Top. Eng., by Dr. F. V. Hayden.
Proc. Acad. Nat. Sci. Philadelphia, 1858:20-29.

. 1869. The extinct mammalian fauna of Dakota and

Nebraska, including an account of some allied forms from
other localities, together with a synopsis of the mammalian
remains of North America. J. Acad. Nat. Sci. Philadelphia,
ser. 2, 7:1-472.

Lundelius, E. L., Jr. 1972. Fossil vertebrates from the late
Pleistocene Ingleside fauna, San Patricio County, Texas. Rep.
Invest. Bur. Econ. Geol., Austin, 77:1-74.

Mayr, E., E. G. Linsley, and R. L. Usinger. 1953. Methods
and principles of systematic zoology. McGraw-Hill New York,
328 pp.

Merriam, J. C. 1912. The fauna of Rancho La Brea. Part II.

Canidae. Mem. Univ. California, 1:215-272.

Nigra, J. O., and J. F. Lance. 1947. A statistical study of the
metapodials of the dire wolf group from the Pleistocene of
Rancho La Brea. Bull. So. California Acad. Sci., 46:26-34.
Nowak, R. M. 1979. North American Quaternary Canis.

Monogr. Mus. Nat. Hist., Univ. Kansas, 6:1-154.
Peterson, O. A. 1926. The fossils of Frankstown Cave, Blair
County, Pennsylvania. Ann. Carnegie Mus., 16:249-315.

1984

KURTEN — CAN IS DIRUS TAXONOMY

227

Schultz, J. R. 1938. A late Quaternary mammal fauna from
the tar seeps of McKittrick, California. Publ. Carnegie Inst.
Washington, 487: 1 1 1-215.

Sellards, E. H. 1916. Human remains and associated fossils
from the Pleistocene of Florida. Rep. Florida Geol. Surv.,
8:121-160.

Simpson, G. G. 1941. Large Pleistocene felines of North Amer-
ica. Amer. Mus. Nov., 1 136:1-27.

. 1949. A fossil deposit in a cave in St. Louis. Amer.

Mus. Nov., 1408:1-46.

Stock, C, and I. F. Lance. 1948. The relative length of limb
elements in Cams dims. Bull. So. California Acad. Sci., 47:
79-84.

Address: Department of Geology, University of Helsinki, 00170 Helsinki 17, Finland.

USE AND ABUSE OF DOGS

Elizabeth S. Wing
ABSTRACT

Dog remains are abundant in most archeological sites in Mex-
ico. Their remains have been recovered from human burials and
from contexts which are interpreted as food refuse. Early Spanish
chronicles describe the use of dogs for food and ritual. Whether
these dogs were hairless, as is frequently stated, is the question.
Genetically hairless dogs can be identified among osteological
collections by the extreme abnormalities in the dentition which

accompanies the hairless condition. No such abnormalities have
been reported in faunal samples. Artificially produced hairless
dogs may account for some of the early reports of hairless dogs
but these can not be detected archeologically. Evidence of another
example of dogs which were altered is the sample of dog burials
from West Mexico. A majority of these dogs had intentionally
broken canines and incisors.

INTRODUCTION

Extraordinary interdependence has developed be-
tween man and dog during the 12,000 years of this
close association. Throughout the millenia people
have modified dogs through selection and when the
results of selection did not satisfy, dogs were tailored
to fit a need or ideal. They can with some justifi-
cation be called man-made (Clutton-Brock, 1977)
though the relationship between man and dog is a
mutual one.

Dog remains occur commonly in North American
archeological sites and are a hallmark of Mexican
sites. Dogs are also frequently mentioned in the early
Spanish chronicles describing Aztec and Maya ways
of life (Maudslay trans. Diaz del Castillo, 1 956; Dur-
an, 1967; Anderson and Dibble trans. Sahagun, 1950;
Tozzer, 1941). Duran (1967) describes the market-
place at Acolman in 1539, where over 400 dogs were
for sale at higher prices than other meat. The dog

meat was consumed for special feasts. Diaz del Cas-
tillo (Maudslay trans., 1956) also mentions that small
dogs were fattened for consumption in the home
and for sale at the market at Tlatelolco. Landa (Toz-
zer, 1941) describes the place held by dogs in Maya
sacrificial ceremonies and after their ceremonial of-
fering they were eaten. Dogs also played a part in
burial rites of bom Aztecs and Maya (Tozzer, 1941:
143). Sahagun (Anderson and Dibble, 1950) has il-
lustrated a dog called a chichi with long yellow hair,
erect ears, and curly tail which is just the kind of
dog used to carry a dead one “across the place of
the nine rivers in the land of the dead.” Sahagun
applies the name xoloitzcuintli to the hairless dogs
which he claims are made permanently hairless by
rubbing turpentine on their skin while they are pup-
pies. In addition, he notes that some say dogs are
bom without hair.

RESULTS AND DISCUSSION

Remains from archeological sites must be relied
upon to verify these observations. One type of ar-
cheological information which is often described in
the context of prehistoric dog use are the famous
Colima dog statues. These have been found in great
numbers associated with human burials in the pres-
ent state of Colima on the West Coast of Mexico.
The Colima dogs are uniformly fat with proportions
similar to that of a puppy and with what appears to
be a full complement of teeth though they are too
uniform in size and shape to be accurate represen-
tations of dog dentition. This context has suggested
the interpretation that these statues represent the
dogs described by Sahagun as the guides for the dead

(Wright, 1960). They have also been interpreted as
hairless dogs intentionally fattened for a food animal
(Coe, 1962; Wright, 1960).

The other type of archeological material that can
provide information about dog use is the remains
of dogs themselves. Three of the questions which
may be answered by these remains are: whether dogs
are associated with human burials, in which case
they may represent the guide dog to the land of the
dead; whether dogs were used for food; and whether
either of these roles were played by hairless dogs.
Dog remains have been found in clear association
with human burials (Wing, manuscript). This as-
sociation is particularly frequent in the sites of the

228

1984

WING-ARCHEOLOGICAL SITE DOGS

229

Table {. — Recorded crosses of hairless dogs.

T T

9 f ?

m x 2

Ti ■

«fx • X ?

— I I —

1~ l 1 1 S 1

9 9 9 cf o

O*" - male or female normal hair, normal dentition

- male or female hairless with abnormal dentition
¥ - mafe or female normal hair, abnormal dentition

- animal born dead

- condition of hair and dentition unknown

- gender unknown, normal hair and dentition

" litter too young to determine dental abnormalities

Data collected in Peru by Cappeiia Hays,

Zoology Department, University of Florida

T

?

O

o

230

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 1. — Muzzle of dog (78-100) associated with human bunal at the site designated N12/1, Marismas Nacionales. A) Oblique view of
palate showing broken incisors and canines; B) Lateral view of left dentary showing broken canine. Drawings by Wendy Zomlefer.

Marismas Nacionales in the states of Sinaloa and
Nayarit on the West Coast of Mexico just north of
Colima. Dog remains are also very frequently as-
sociated with other food remains and share with
them butchering marks and burned skeletal ele-
ments characteristic of food remains (Flannery , 1976;
Hamblin, 1980; Pohl, 1976; Pohl and Feldman,
1982; Wing, 1978). Dogs constitute as much as 25%
of the animals used for food (Wing, 1981). Whether
any of these dogs was hairless is open to question
and warrants some discussion of the characteristics
of the congenital hairless condition.

Hairless dogs are known from many places in-
cluding Mexico, the Greater Antilles, Peru, Para-
guay as well as in the Old World (Allen, 1920;
Braund, 1975; Weiss, 1976; Wright, 1960). The con-
dition of hairlessness, when not induced by rubbing
the skin with turpentine, is a genetic disorder passed
to following generations by an autosomal dominant
pattern with incomplete penetrance and variable ex-
pressivity (Graber, 1978; Hutt, 1979). Both the de-
gree of hairlessness and the degree of abnormality
and congenital absence of the teeth are affected by

this condition. The results of experimental matings
of hairless dogs indicates that the individuals which
are homozygous for this trait are usually born dead
with extreme abnormalities, and the heterozygous
individuals are hairless (Hutt, 1979; Letard, 1930).
The ratio of normal to hairless viable to hairless
dead from experimental crosses is 9:12:4 which fits
the expected 1 :2: 1 from the segregation of dominant
genes from two heterozygous parents (Letard, 1930,
reported by Hutt, 1979). The crosses presented on
Table 1 are incomplete but indicate the penetrance
of dental abnormalities even in animals with normal
coats. The extent of congenital tooth absence is pro-
found (Table 2). The American Kennel Club in writ-
ing standards for the breed recognize the absence of
the premolars associated with the hairlessness and
do not penalize an animal for the absence of the
incisors as well.

If these hairless dogs represent the same genetic
condition as the xoloitzcuintli then the remains of
hairless dogs can be recognized with confidence by
their abnormal dentition even though the degree of
abnormality varies. As far as I know the degree of

1984

WING-ARCHEOLOGICAL SITE DOGS

231

Table 2. — Documented absence of teeth in hairless dogs or normally haired dogs from hairless ancestors.

Name

Gender

Teeth

in upper and lower jaw

Age in years

I,1

h2

h3

c

P, •

IV

p3j

p44

M,1

m22

m3°

Flor

F

2 '/2— 3

+

+

+

+

+

+

+

+

+

+

+

+

+

Chispi

F

Vh

+

+

+

+

+

+

Chaski

F

4

+

+

+

+

+

+

+

+

+

+

+

+

+

Nanau

F

7

+

+

a

a

+

+

+

+

+

+

+

+

+

Mitsuko

M

7

+

+

+

+

+

+

+

+

+

+

Nanpu

M

4

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

No name

M

1 1

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

Lalau

M

8

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

Juanacho

M

4-5

+

+

+

+

+

+

Tali normal hair

M

3/4

+

+

Sumac normal hair

F

4

+

-j-

+

+

blank = tooth present.

+ = tooth absent,
a = tooth present but abnormal.

Data collected in Peru by Coppelia Hays, Zoology Department, University of Florida.

tooth absence permitted in the breed by the Amer-
ican Kennel Club standards or seen by Coppelia
Hays in the Peruvian examples has never been re-
ported in dog remains from Mexico. The Colima
dog statues clearly have a full complement of teeth
so their interpretation as hairless dogs is open to
question. The large sample of dog burials from the
post classic sites of the Marismas Nacionales north
of Colima also have normal though modified den-
tition and would by this criterion have normal hair.
None of the many dog remains described as asso-
ciated with food remains are reported to have ab-
normal dentition of the extent seen in known hair-
less dogs. This lack of archeological evidence for the
presence of hairless dogs does not mean that hairless
dogs did not exist. It does, however, suggest that
genetically hairless dogs were not as abundant as
normal dogs.

Artificially hairless dogs, those whose hair was
intentionally removed with turpentine, can not be
detected archeologically. Thus, no way exists of de-

termining how wide spread this practice reported
by Sahagun (Anderson and Dibble, 1950) was.

Though probably not related to artificial simu-
lation of the hairless dog, many of the dog burials
from the five sites in the Marismas Nacionales have
broken incisors and canines (Fig. 1 and Table 3).
These broken teeth are darkened by the blood pig-
ments typically seen in dead teeth that are retained
in the mouths of live animals. Two of the broken
canines show slight wear subsequent to the damage
to the crown. No evidence of fractured bone is ap-
parent. Fifty-three of the total sample of 96 canines
from the dog burials are broken. The most frequent-
ly damaged incisor is the third which is the largest
and the closest to the canine. In addition to the dog
burials, four raccoons were included with the human
burials and only one of these raccoons escaped the
tooth breaking treatment. These data suggest that
the front teeth particularly the canines were inten-
tionally broken during the animal’s life prior to their
burial with a human. Furthermore, this was a fre-

232

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 3. — Broken and unbroken teeth from dogs and raccoons
(from Chalpa only) associated with human burial sites in the
Maris mas Nacionales, West Mexico.

Sites

Incisors

Broken

Canines

Unbroken

canines

N12/1

5

3

3

Tecualillo

2

6

7

Cristo Rey

9

11

6

Rincon

3

4

18

Chalpa

Canis

6

24

6

Procyon

0

5

3

Total

25

53

43

quent practice and extended to raccoons as well as
dogs. Raccoons apparently had a role similar to dogs
in the minds of the people inhabiting the estuary.
What motivated these people to mutilate the ani-
mals, which may have been intended as guides to
the after life, can only be a matter of speculation.

Though it may not be possible to reconstruct mo-
tivations, the results of prehistoric action may be
revealed. Clearly, dogs were very important in the
lives and perhaps after lives of the prehistoric people
of Mexico. There can be no doubt that dogs were
used for food and for ceremonial purposes related
to burial rites. The role of hairless dogs, either ge-
netically or artificially hairless, remains a tantalizing
question.

ACKNOWLEDGMENTS

This paper is gratefully dedicated to John Guilday for his pi-
oneering work in Zooarcheology and setting goals and standards
for the growing numbers of zooarcheologists to strive towards.
Thanks are also due Coppelia Hays for collecting data on the

dental condition of hairless dogs and to Wendy Zomlefer for her
drawing of the brutalized muzzle. I am also most grateful to Sylvia
Scudder for help in searching the literature.

LITERATURE CITED

Allen, G. M. 1920. Dogs of the American aboriginies. Bull.
Mus. Comp. Zool., 63:43 1 -5 1 7;478-48 1 .

Braund, K. 1975. The uncommon dog breeds. Arco Publ. Co.
Inc., New York, pp. 194-229.

Clutton-Brock, J. 1977. Man-made dogs. Science, 197:1340-
1342.

Coe, M. D. 1962. Mexico. Ediciones Lara, Mexico City, Mex-
ico, 245 pp.

Diaz del Castillo, B. 1517-1521. The Discovery and Con-
quest of New Mexico. Translated by A. P. Maudslay London
1908-16 (American ed., Farrar, Straus and Cudahy, New
York, 1956).

Duran, F. D. 1967. Historia de Las Indias de Nueva Espana
y Islas de Tierra Firma. Vol. II, pp. 218-219. Introduction
by J. F. Ramirez. Editora Nacional, Mexico.

Flannery, K. V. 1976. The early Mesoamerican village. Ac-
ademic Press, New York, 377 pp.

Graber, L. W. 1978. Congenital absence of teeth: a review with
emphasis on inheritance patterns. Jada, 96:266-275.

Hamblin, N. L. 1980. Animal Utilization by the Cozumel Maya:
Interpretation through faunal analysis. Unpublished Ph.D.
dissert. Univ. Arizona, Tucson, 349 pp.

Hutt, F. B. 1979. Genetics for dog breeders. W. H. Freeman,
San Francisco, 245 pp.

Letard, E. 1930. Le Mendelisme experimental Experiences sur
heredite mendelienne du caractere “peau nue” dans l’espece
chien. Rev. Vet. et J. Med. Vet., Toulouse, 82:553-570.

Pohl, M. E. D. 1976. Ethnozoology of the Maya: an analysis
of fauna from five sites in the Peten, Guatemala. Unpub-

lished Ph.D. dissert., Harvard Univ., Cambridge, Massa-
chusetts, 317 pp.

Pohl, M., and L. H. Feldman. 1982. The traditional role of
women and animals in lowland Maya economy in Maya
subsistence. Pp. 295-31 1, in Studies in memory of Dennis
E. Puleston (K. V. Flannery, ed.), Academic Press, New York,
368 pp.

Sahagun, F. B. de. 1950. General history of the things of New
Spain translated from the Nahuatl by Arthur Jo Anderson
and Charles E. Dibble. Sante Fe, Nos. 1-5; 7-9; 12.

Tozzer.A. M. 1941. Landa’s relacionde Las Cosasde Yucatan.
Papers Peabody Mus. Amer. Arch. Ethnol., Harvard Univ.,
18:1-394.

Weiss, P. 1976. El Perro Peruano Sin Pelo. Museo Nac. An-
tropologia y Arqueologia ser.: Paleobiologia No 1 reprodu-
cido de la revista 1 970 Acta Herediana. Univ. Peruana Cay-
etano, Heredia, 3:33-54.

Wing, E. S. 1978. Use of dogs for food: an adaptation to the
coastal environment in prehistoric coastal adaptations edited
by B. Stark and B. Voorhies. Academic Press, New York,
pp. 29-41.

. 1981. A comparison of Olmec and Maya food ways in

the Olmec and their neighbors edited by Elizabeth P. Benson.
Dumbarton Oaks Research Library and Collections, Har-
vard Univ., Washington D.C., pp. 20-28.

. ms. Vertebrate remains from the Archeological sites in

the Marismas Nacionales.

Wright, N. P. 1960. El enigma del xoloitzcuntli. Inst. Nac.
Antropologia e Historia, Mexico, 102 pp.

Address: Florida State Museum, University of Florida, Gainesville, Florida 3261 1.

TIME OF EXTINCTION AND NATURE OF ADAPTATION OF THE
NOBLE MARTEN, MARTES NOB I LIS

Donald K. Grayson

ABSTRACT

The noble marten, Maries nobilis, is known from at least 12
sites in the western United States and the Yukon. This extinct
marten is generally thought to have become extinct at the end of
the Pleistocene, and to have been adapted to cool, if not boreal,
conditions. Two sites, however, have yielded specimens that date
to between 3,000 and 3,500 B.P., one of which is reported here
for the first time. It is argued that the noble marten may not have

become extinct until late in the Holocene. In addition, the ver-
tebrate taxa with which Maries nobilis has now been found in
the western United States suggest that it was adapted to a broad
variety of environmental settings, and that it could not have
suffered from competition with the American marten (Martes
americana) throughout its range.

INTRODUCTION

In 1926, E. R. Hall described a new subspecies
of marten, Martes caurina ( =americana ) nobilis on
the basis of eight maxillae and mandibles from the
Pleistocene deposits of Samwel and Potter Creek
caves, northern California (Hall, 1926). Although
Hall (1936) later concluded that the material he had
described did not differ sufficiently from the local
M. americana sierrae to merit subspecific recogni-
tion, Anderson (1970) was able to demonstrate con-
sistent differences between the skeletons of modern
M. americana and a sizeable series of marten cranial
and postcranial remains from four western United
States cave faunas. Accordingly, she assigned the
paleontological material to Martes nobilis, the noble
marten (see taxonomic review in Anderson, 1970).
Since Anderson’s revision, M. nobilis has been iden-
tified from a number of additional paleontological
and archaeological faunas. This large, extinct mar-
ten is now known from at least 1 2 sites in the western
United States (Colorado, Wyoming, Idaho, Nevada,

and California) and the Yukon (Webster, 1978; Kur-
ten and Anderson, 1980, Grayson, 1982a).

The noble marten is generally thought to have
become extinct at the end of the Pleistocene, and to
have been adapted to cool, if not boreal, conditions
(for example, Anderson, 1970; Hager, 1972; Zie-
mens and Walker, 1974; Miller, 1979; Kurten and
Anderson, 1980). Given this apparent adaptation
and apparent date of extinction, terminal Pleisto-
cene climatic change has often been assigned a role
in explaining the extinction of the noble marten (for
example, Anderson, 1970; Kurten and Anderson,
1 980). The recent discovery of Martes nobilis in the
prehistoric fauna of Hidden Cave, Nevada, coupled
with a previously published but generally over-
looked record from Dry Creek Rockshelter, Idaho
(Webster, 1978), suggests that current hypotheses
concerning the nature of adaptation and timing of
extinction of M. nobilis may be incorrect.

HIDDEN CAVE AND THE NOBLE MARTEN

Hidden Cave (Nv-Ch-16; elevation 1,251 m) is
located on the northern face of Eetza Mountain in
the southern Carson Desert of western Nevada, ap-
proximately 27 km southeast of the town of Fallon
(see Figs. 1-2). The site has been the scene of three
professional archaeological excavations, the first
conducted by the Nevada State Park Commission
in 1940, the second by the University of California,
Berkeley in 1951, and the third by the American
Museum of Natural History (AMNH) in 1979 and
1980 under the direction of David H. Thomas

(Thomas, 1982). Although the Berkeley excavations
provided a great deal of information on the archae-
ology and stratigraphy of the Hidden Cave deposits
(Morrison, 1964; Roust and Clewlow, 1968), the
AMNH excavations were much more extensive, and
provided, among other things, a stratified sequence
of vertebrate faunal remains that span roughly the
past 21,000 years (see Fig. 3). To date, 6,603 bones
and teeth belonging to 45 taxa (genera and species)
of mammals have been identified from the 14 de-
fined strata within the site.

233

234

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Today, the area surrounding Hidden Cave is arid.
Vegetation on the slopes adjacent to the site is dom-
inated by little greasewood ( Sarcobatus baileyi) and
shadscale ( Atriplex confertifolia), while the Carson
Desert lowlands to the immediate north of Eetza
Mountain are dominated by black greasewood ( Sar-
cobatus vermiculatus). Although now quite arid.

however, the desertification of the Eetza Mountain
area did not occur long ago. All strata within Hidden
Cave with sizeable faunal samples, including de-
posits above a layer of Mono tephra dated to ca.
1,500 years before the present (B.P.), contain yel-
low-bellied marmots ( Marmota flaviventris ) and
bushy-tailed wood rats ( Neotoma cinerea ), species

1984

GRAYSON -EXTINCTION OF NOBLE MARTEN

235

that require relatively moist environments and that
are no longer found on or near Eetza Mountain. In
addition, cat-tail ( Typha latifolia and T. angusti-
folia) pollen is abundant in the site, accounting for
approximately 20% of the identified pollen above
Mono tephra (Wigand and Mehringer, 1982). The
precise timing of the acidification of the Eetza
Mountain area is not known, but is currently under
detailed investigation.

Among the identified mammalian bones and teeth
from Hidden Cave is a single specimen of Martes
nobilis, a left M1 (AMNH HC-184; Fig. 4). This
tooth is much larger than the corresponding tooth
of M. americana: compared to M. pennanti, the
inner lobe of the tooth is expanded, producing a
deeper constriction between inner and outer lobes.
These are characters in which the tooth agrees with
M. nobilis (Anderson, 1970). Plotting the width of

236

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 3.— An internal view of Hidden Cave, late in the 1979 excavations. Photograph by Albert A. Alcorn, Fallon, Nevada.

M1 against the length of the inner lobe of the tooth
on the scattergram provided by Anderson (1970)
shows the Hidden Cave specimen to lie well above
the distribution of M. americana, at the lower end
of the distribution for M. pennanti, and at the upper
end of that for M. nobilis. The measurements of the
tooth (width M1 = 10.0 mm, length M1 inner = 6.3
mm, length M1 mid = 4.4 mm, length M1 outer
4.4 mm; see Anderson, 1970 for measurement def-
initions) fall within known ranges for M. nobilis
(Anderson, 1970; Hager, 1972). Identification of the
specimen as M. nobilis has been confirmed by An-
derson (personal communication).

This single tooth was recovered from Hidden Cave
stratum II. Although the stratum II fauna is small
(279 specimens identified to date), the composition
of this fauna is similar to that of other Hidden Cave
strata in that it contains mesic species, including
Marmot a flaviventris and Neotoma cinerea, that do
not exist on Eetza Mountain today (Table 1). The
precise dating of stratum II is somewhat problem-
atic, though the general age of this stratum is not.

Two radiocarbon dates are available for this stra-
tum-810 ± 80 B.P. (WSU-2457) and 3,850 ± 110
B.P. (WSU-2458). The younger of these determi-
nations is clearly too young, as it underlies Mono

0 mm. 10.0

Fig. 4.— The Hidden Cave Martes nobilis M1.

1984

GRAYSON — EXTINCTION OF NOBLE MARTEN

237

Table l.— Numbers of identified specimens (NISP) per taxon —
Hidden Cave stratum II mammals.

Taxon

NISP

Sylvilagus sp.

72

Sylvilagus nuttallii

1

Lepus sp.

119

Marmota flaviventris

9

Spermophilus sp.

1

Spermophilus cf. townsendii

3

Spermophilus townsendii

8

Thomomys sp.

2

Thomomys cf. bottae

3

Thomomys bottae

1

Perognathus sp.

1

Perognathus longimembris

i

Dipodomys sp.

1 1

Dipodomys microps

3

Peromyscus maniculat'us

1

Neotoma sp.

11

Neotoma cf. lepida

7

Neotoma lepida

1

Neotoma cf. cinerea

2

Neotoma cinerea

6

Microtus sp.

9

Cams latrans

2

Cams lupus

1

Maries nobiiis

1

Mustela vison

1

Taxidea taxus

1

Spilogale gracilis

1

Total

279

tephra dated to ca. 1,500 B.P. The older date, in
turn, may be too old, because it overlies younger

dates. Eight radiocarbon determinations from the
deeper stratum IV fall between ca. 2,000 and 5,400
B.P., with four of these dates lying between ca. 3,500
and 3,800 B.P. Basing his interpretation on the in-
ternal fit of radiocarbon determinations, on geolog-
ical considerations, and on the archaeological evi-
dence, Jonathan O. Davis (personal communication;
see also Davis, 1982) estimates that stratum II was
deposited between about 3,700 and 3,600 B.P. Stra-
tum II seems to have formed the surface of the site
between ca. 3,600 and 1,500 B.P. , during which time
very little deposition occurred (Davis, persona!
communication). Stratum I dates to between ca.
1,500 B.P. and historic times.

Unfortunately, some of the Hidden Cave sedi-
ments have been disturbed by rodent activity, and
by the activities of prehistoric peoples. As a result,
the stratigraphic integrity of the deposits from which
the M. nobiiis tooth came cannot be given unques-
tioning faith. Indeed, part of the unit from which
the tooth came (Lambda 10, level 3; see Fig. 5 for
the location of the AMNH Hidden Cave excava-
tions units) was disturbed by rodents. Because the
tooth was discovered during the screening of de-
posits from that unit, it is not possible to know
whether the specimen was originally located in the
disturbed area. All that can be concluded from the
Hidden Cave M. nobiiis tooth is that either the noble
marten survived much longer than has been thought,
or that this tooth has been displaced upwards from
Pleistocene strata. Both options are open.

THE NOBLE MARTEN: TIME OF EXTINCTION AND ADAPTATION

Although the stratigraphic placement of the Hid-
den Cave Maries nobiiis specimen cannot be re-
garded as secure, it is also true that this does not
represent the first late Holocene record for the noble
marten. Kurten and Anderson (1972), for instance,
identified 45 specimens of M. nobiiis from Jaguar
Cave, southern Idaho, of which two came from a
stratum that provided a single radiocarbon date of
ca. 4,000 B.P. (Sadek-Kooros, 1972). These speci-
mens, however, may also have reached this position
through disturbance, whereas Sadek-Kooros (1972)
believes the 4,000 B.P. date may be too young. More
important, then, is the fact that S. J. Miller has
identified a single tooth of M. nobiiis from Dry' Creek
Rockshelter in the Boise River Valley of south-
western Idaho (in Webster, 1978), an identification

that has also been verified by Anderson (personal
communication). Bone from the stratum that pro-
vided the M. nobiiis tooth was dated to 3,270 ±
110 B.P. (WSU-1574), whereas a date of 3,530 ±
85 B.P. (WSU-1486) is available for the stratum
beneath, and of 2,090 ± 80 B.P. (WSU-1503) for
the next dated stratum above. These dates are fully
in line with a rti factual stylistic markers from the
same strata. Significantly, the stratum that provided
WSU-1486 sits directly on the bedrock floor of the
cave, so that the M. nobiiis specimen in Dry Creek
Rockshelter could not have been derived from deep-
er Pleistocene strata within the site (see the strati-
graphic profile in Webster, 1978). The Dry Creek
Rockshelter M. nobiiis, then, seems to provide firm
support for the survival of this animal into the late

238

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 5.— The grid system utilized by the American Museum of Natural History during the 1979-1980 excavations. The Martes nobilis
specimen came from unit Lambda 10.

Holocene. Given that this is the case, the Hidden
Cave M. nobilis may, in fact, indicate the late sur-
vival of the noble marten in western Nevada as well.

The likelihood that M. nobilis survived into late
Holocene times in the arid west suggests that subfos-
sil specimens of Martes from this area should be
examined extremely closely. For instance, Spiess
(1974) reported two specimens of Martes americana
from Bronco Charlie Cave, Ruby Range, eastern

Nevada. One of these specimens was a lower right
canine for which Spiess did not provide prove-
nience. The second specimen, however, was a lower
right mandible retaining M,. Although precise in-
formation on the provenience of this specimen was
not provided by Spiess, it did come from deposits
that also contained archaeological materials. The
oldest projectile points in Bronco Charlie Cave are
Elko points (Casjens, 1974), a style that dates to

1984

GRAYSON -EXTINCTION OF NOBLE MARTEN

239

between ca. 3,500 and 1,200 B.P. in the central and
western Great Basin (O’Connell, 1967; Heizer and
Hester, 1973; Thomas, 1981), although earlier dates
are available to the east (Aikens, 1970). Thus, as-
suming no disturbance, the Bronco Charlie Martes
mandible came from sediments that would seem to
be late Holocene in age. Although Martes ameri-
eana is no longer found on the Ruby Range (the
closest known modem population are some 325 km
to the east: Durrant, 1952), the extinction of boreal
mammals in the Great Basin during Holocene times
is now extremely well-documented (Grayson, 1977,
1981, 1982 b\ Thompson and Mead, 1982), and the
Bronco Charlie Martes may represent another such
instance. However, it seems equally possible that
the specimen was misidentified and actually per-
tains to Martes nobilis. It should be reexamined to
see whether this is the case.

I have noted that Martes nobilis is generally
thought to have been adapted to cool conditions.
This adaptation has been inferred from the fact that
noble marten remains have been found associated
with the remains of boreal taxa at a number of sites.
At both Little Box Elder Cave, Wyoming, and Jag-
uar Cave, Idaho, for instance, late Pleistocene sed-
iments incorporated specimens of both M. nobilis
and the collared lemming, Dicrostonyx cf. torquatus
(Anderson, 1968, 1970; Kurten and Anderson,
1 972). In contrast, however, and as Anderson (1970)
has discussed, no boreal taxa were associated with
M. nobilis at Samwel and Potter Creek caves in
Shasta County, California; with the possible excep-
tion of the extinct large mammals, the species iden-
tified from those caves do not require an environ-
ment much different from that which characterizes
the area today (Sinclair, 1903; Furlong, 1906).

Whether or not the Hidden Cave M. nobilis came

from an undisturbed stratigraphic setting, it was not
associated with a boreal fauna — there are no boreal
species of mammals or birds in the Hidden Cave
fauna. In fact, excluding the remains of extinct horse
(Equus sp.). the set of mammalian and avian species
represented in the Hidden Cave fauna identified to
date (and the identifications are nearly complete)
can be duplicated by extant faunas in many well-
watered parts of the Great Basin today. Similarly,
Dry Creek Rockshelter lacked boreal taxa. When
arrayed alongside such sites as Samwel, Potter Creek,
and Little Box Elder caves, Hidden Cave and Dry
Creek Rockshelter suggest that the noble marten was
adapted to a broad variety of environmental set-
tings, and certainly cannot be taken as indicative of
boreal conditions. In addition, it would seem to
have occupied a wider variety of environments than
M. americana. If so, the suggestion that competition
with the American marten was in part responsible
for the extinction of the noble marten throughout
its range (Anderson, 1970; Kurten and Anderson,
1972, 1980) loses much of its force.

In short, accumulating information on the noble
marten suggests that it was adapted to a broad va-
riety of environmental settings, that it would not
have suffered from competition with the American
marten throughout its range, and that it did not
become extinct until late in the Holocene, perhaps
after 3,000 B.P. If these hypotheses are correct, it
must be asked why the noble marten survived the
major environmental changes associated with the
end of the Pleistocene at about 10,000 B.P. and
those associated with the establishment of approx-
imately modem environmental conditions in the
arid west between about 8,000 and 7,000 B.P. (Meh-
ringer, 1977; Van Devender and Spaulding, 1979),
only to become extinct during late Holocene times.

ACKNOWLE DGM ENTS

I thank Elaine Anderson for verifying my identification of the
Hidden Cave Martes nobilis specimen, and for so freely sharing
her thoughts on the adaptation and extinction of the noble mar-
ten. I thank E. Anderson, J. O. Davis, C. Maser, P. J. Mehringer,
Jr., and D. H. Thomas for providing critical comments on this
manuscript, and Brian Hatoff for making the reexcavation of
Hidden Cave possible. Figures 1, 2, 4, and 5 were drawn and
provided by Dennis O’Brien of the American Museum of Natural
History; his help is once again gratefully acknowledged. Casts of
the Hidden Cave Martes nobilis specimen have been deposited
in the collections of the Burke Memorial Museum, University

of Washington, the Frank H. McClung Museum, University of
Tennessee, and the Department of Geosciences, University of
Arizona. The specimen itself is housed in the collections of the
Department of Anthropology, American Museum of Natural
History. The research reported herein was supported by the
American Museum of Natural History and the Bureau of Land
Management (Carson City, Nevada), with additional support from
the Richard Lounsbury and Speidel Foundations. The final report
on the results of the Hidden Cave excavations will be issued as
an Anthropological Paper of the American Museum of Natural
History.

240

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

LITERATURE CITED

Aikens, C. M. 1970. Hogup Cave. Univ. Utah Anthro. Papers.,
93:1-286.

Anderson, E. 1968. Fauna of the Little Box Elder Cave, Con-
verse County, Wyoming. The Carnivora. Univ. Colorado
Studies, Ser. Earth Sci., 6:1-59.

. 1970. Quaternary evolution of the genus Maries (Car-
nivora, Mustelidae). Acta Zool. Fennicae, 130:1-132.

Casjens, L. 1974. The prehistoric human ecology of the south-
ern Ruby Valley, Nevada. Unpublished Ph.D. dissert., Har-
vard Univ., Cambridge, Massachusetts 593 pp.

Davis, J. O. 1982. Sediments and geological setting of Hidden
Cave. Pp. 183-242, in The archaeology of Hidden Cave (D.
H. Thomas, ed.), Report submitted to the Bureau of Land
Management, Carson City, Nevada, 886 pp.

Durrant, S. D. 1952. The mammals of Utah: taxonomy and
distribution. Univ. Kansas Publ., Mus. Nat. Hist., 6:1-549.

Furlong, E. L. 1906. The exploration of Samwel Cave. Amer.
J. Sci., 172:235-237.

Grayson, D. K. 1977. On the Holocene history of some north-
ern Great Basin lagomorphs. J. Mamm., 58:507-513.

. 1981. A mid-Holocene record for the heather vole,

Phenacomys cf. intermedius, in the central Great Basin and
its biogeographic significance. J. Mamm., 62:1 15-121.

. 1982a. The paleontology of Hidden Cave: birds and

mammals. Pp. 286-363, in The archaeology of Hidden Cave
(D. H. Thomas, ed.). Report submitted to the Bureau of
Land Management, Carson City, Nevada.

. 1982 b. Toward a history of Great Basin mammals dur-
ing the past 15,000 years. Pp. 82-101, in Man and environ-
ment in the Great Basin (D. B. Madsen and J. F. O’Connell,
eds.), Soc. Amer. Archaeol. Papers, 2:1-242.

Hager, M. W. 1972. A late Wisconsin-Recent vertebrate fauna
from the Chimney Rock Animal Trap, Laramie County,
Wyoming. Univ. Wyoming Contrib. Geol., 2:63-71.

Hall, E. R. 1926. A new marten from the Pleistocene cave
deposits of California. J. Mamm., 7:127-130.

. 1936. Mustelid mammals from the Pleistocene of North

America with systematic notes on some recent members of
the genera Mustela, Taxidea, and Mephitis. Carnegie Inst.
Washington Publ., 473:41-1 19.

Heizer, R. F., and T. R. Hester. 1973. Review and discussion
of Great Basin projectile points: forms and chronology. Ar-
chaeological Research Facility, Dept. Anthro., Univ. Cali-
fornia, Berkeley, ii + 39 pp.

Kurten, B., and E. Anderson. 1972. The sediments and fauna
of Jaguar Cave. II. The fauna. Tebiwa, 1 5( 1):2 1—45.

. 1 980. Pleistocene mammals of North America. Colum-
bia Univ. Press, New York, 442 pp.

Mehringer, P. J., Jr. 1977. Great Basin late Quaternary en-
vironments. Pp. 113-167, in Models and Great Basin pre-
history (D. D. Fowler, ed.). Desert Res. Inst. Publ. Soc. Sci.,
12:1-213.

Miller, S. J. 1979. The archaeological fauna of four sites in
Smith Creek Canyon. Pp. 272-329, in The archaeology of
Smith Creek Canyon, eastern Nevada (D. R. Tuohy and D.
L. Rendall, eds.), Nevada State Mus. Anthro. Papers, 17:1—
394.

Morrison, R. B. 1964. Lake Lahontan: geology of southern
Carson Desert, Nevada. U.S. Geol. Surv. Prof. Papers, 401:
1-156.

O’Connell, J. 1 967. Elko eared/Elko comer-notched points as
time markers in the Great Basin. Univ. California Archaeol.
Surv. Report, 70:129-140.

Roust, N. L., and C. W. Clewlow, Jr. 1968. Projectile points
from Hidden Cave (Nv-Ch-16), Churchill County, Nevada.
Univ. California Archaeol. Surv. Report, 71:103-1 16.

Sadek-Kooros, H. 1972. The sediments and fauna of Jaguar
Cave. I. The sediments. Tebiwa, 15(1): 1-20.

Sinclair, W. J. 1903. A preliminary account of the exploration
of Potter Creek Cave, Shasta County, California. Science,
17:708-712.

Spiess, A. 1974. Faunal remains from Bronco Charlie Cave
(26EK801), Elko County, Nevada. Pp. 452-486, in The pre-
historic human ecology of the southern Ruby Valley, Ne-
vada, by L. Casjens. Unpublished Ph.D. dissert., Harvard
Univ., Cambridge, 593 pp.

Thomas, D. H. 1981. How to classify the projectile points from
Monitor Valley, Nevada. J. California and Great Basin An-
thro., 3:7-43.

. 1982. Previous research at Hidden Cave. Pp. 80-109

in The archaeology of Hidden Cave (D. H. Thomas, ed.),
Report submitted to the Bureau of Land Management, Car-
son City, Nevada, 886 pp.

Thompson, R. S., and J. I. Mead. 1982. Late Quaternary en-
vironments and biogeography in the Great Basin. Quater-
nary Res., 17:39-55.

Van Devender, T., and W. G. Spaulding. 1979. Development
of vegetation and climate in the southwestern United States.
Science, 204:701-710.

Webster, G. S. 1978. Dry Creek Rockshelter: cultural chro-
nology in the western Snake River region of Idaho ca. 4, 1 50
B.P. -1,300 B.P. Tebiwa, Misc. Papers Idaho State Univ.
Mus. Nat. Hist., 15:1-35.

Wigand, P. E., and P. J. Mehringer, Jr. 1982. Pollen and
seed analysis for Hidden Cave. Pp. 252-285, in The ar-
chaeology of Hidden Cave (D. H. Thomas, ed.). Report sub-
mitted to the Bureau of Land Management, Carson City,
Nevada, 886 pp.

Ziemens, G., and D. N. Walker. 1974. Bell Cave, Wyoming:
preliminary archaeological and paleontological investiga-
tions. Pp. 88-90, in Applied archaeology and geology: the
Holocene history of Wyoming (M. Wilson, ed.). Geol. Surv.
Wyoming Rept. Inv., 10:1-127.

Address: Department of Anthropology and Burke Memorial Museum, University of Washington, Seattle,
Washington 98195.

LAVA BLISTERS AS CARNIVORE TRAPS

John A. White, H. Gregory McDonald, Elaine Anderson, and

James M. Soiset

ABSTRACT

Analysis of the bones of 58 kinds of vertebrate animals found
in two Idaho lava blister caves indicates that in one cave, Moon-
shiner, 30% (42% of biomass) of the animal remains are of carniv-
orous mammals, such as wolves, coyotes, foxes, badgers, wol-
verines, martens, and weasels. Carnivores make up 42% (40% of
biomass) in the other cave, Middle Butte. In contrast, the per-

centage of carnivores in the Recent regional fauna or in a “nor-
mal” fossil cave fauna, normally ranges from 2 to 5%. This strongly
suggests that both caves functioned as selective carnivore traps.
Evidently the bone deposit in Moonshiner Cave began accu-
mulating at an earlier time than the deposit in Middle Butte Cave.

INTRODUCTION

One of the major problems in interpreting the
paleontology of a fossil fauna is the recognition of
bias. An accurate and feasible interpretation of the
paleoecology of a fossil fauna requires that all pro-
cesses, biological, climatological, or geological, which
alter the representation of animals in a fossil deposit
from that which occurred when those animals were
living, should be identified. An ideal fauna for paleo-
ecological analysis is one in which the preservation
of individuals is proportional to their abundance in
the living community so that the relative abundance
of different species is preserved. Unfortunately, this
is generally not the case and studies have identified
various types of bias in the fossil record (Voorhies,
1969). Even though a random unbiased sample is
ideal, studies of biased samples are also informative
and may highlight interesting events that took place
in the past and shed light on the paleoethology of a
species. Large accumulations of bones of single
species are presumably nonrandom. Such deposits
include— the Menoceras and Stenomylus quarries
of Miocene age at Agate Springs, Nebraska; the Plio-
cene Hagerman horse quarry in Idaho; the accu-

mulation of bones of cave bear, Ursus spe/aeus, in
Pleistocene caves in Europe; and the asphalt de-
posits at Rancho La Brea, California. The over-rep-
resentation of Menoceras bones at Agate Springs is
attributed to stream transport of the bones (Peter-
son, 1923), whereas that of Stenomylus is inter-
preted as a catastrophic event which overwhelmed
a herd while on its bedding ground (Brown, 1 929).
The wealth of cave bear bones is attributed to the
gradual accumulation in dens over many genera-
tions in bear defended caves (Kurten, 1 976; Soergel,
1940), whereas the over- representation of large car-
nivores such as the sabertooth, Smilodon fatalis,
and the dire wolf, Canis dims, in tar pits, is inter-
preted as selective trapping as a result of scavenging
food habits by individuals as they attempted to feed
on entrapped herbivores (Stock, 1972). We propose
that the large accumulation of carnivore bones in
the lava caves of Idaho, discussed in this paper, is
due to carnivores entering the caves to feed on en-
trapped herbivores and becoming entrapped them-
selves.

METHODS

The procedures used in excavating and collecting the bones
from the two lava blister caves were as follows.

Moonshiner Cave

A test trench was excavated across the cave floor near the
confluence of the northeast and southeast tunnels (Fig. 1), re-
vealing the fine-grained silt to be deposited in pockets in the floor
of the cave. The deepest of the latter pockets was less than 0.5
m deep. Thus the sediments ranged in thickness from 0 to 0.5
m. Almost no bones were found to be in articulation.

All medium to large-sized bones were collected from the sur-
face of the floor of the cave, and almost none of the large bones

were completely buried in the sediments. An estimated 10% of
the bone-bearing sediments was removed at random and screened
to produce approximately 1.5 metric tons of concentrate.

Middle Butte Cave

Two test trenches were excavated through the sediments down
to the basalt floor. One trench reached a depth of 3.7 m and the
other 2.6 m. The deeper trench was located 2 m west of the cave
opening and the shallower one, 2 m north of the rock pile (Fig.
2). A shallow trench directly below the cave entrance revealed
the rock fall, which is related to the formation of the cave en-
trance, to be confined to the upper meter of the sediments. This

241

242

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

/

/ /

suggests that the sediments entered via another opening and that
most of the sedimentation had occurred prior to the formation
of the present opening.

Most of the bones (99%) were encountered in the upper 0.3 m
of the sediments in the deeper trenches mentioned above. The
sediments are in the clay to silt-size range, are very cohesive, and
can be removed in well-defined blocks. A fluviatile origin is in-
dicated by the fine laminations present throughout the entire
deposit and by the presence of cross-bedding characteristic of
stream deposits. The surface bones were intermixed with dung
from Neotoma and plant material presumably brought into the
cave by packrats.

While excavating the deeper trenches two layers of ash were
found. The upper ash was at a depth of 25 cm and the lower ash
was at a depth of 120 cm. The upper ash is Mazama and the

lower Glacier Peak (Owen Davis, personal communication, 1982).
Mazama ash is dated at 6,600 years B.P. and Glacier Peak 1 2,000
years B.P. This gives a rate of deposition of 95 cm in 5,400 years
or one cm every 57 years. Using this average rate of sedimen-
tation the 25 cm above the Mazama ash represents approximately
1 ,425 years. Because the majority of the bone was collected above
the top of the consolidated sediments a maximum age for most
of the bones would be approximately 5,175 years B.P.

All medium to large-sized bones were collected from the sur-
face of the cave floor, and almost none of the large bones were
completely buried in the sediments. Almost no bones were in
articulation. An estimated 10% of the bone-bearing sediments
were removed at random and screened to produce approximately
one metric ton of concentrate.

GEOLOGIC SETTING

The lava fields of the Snake River Plain of Idaho
contain numerous lava tubes (tunnels) and lava blis-
ters (tumuli of Daly, 1914:133) as well as deep fis-

sures such as the Great Rift near Aberdeen, Idaho.
Tumuli are formed by the expansion of entrapped
gases within the molten rock, and pressure ridges

1984

WHITE ET AL. -CARNIVORE TRAPS

243

MIDDLE BUTTE CAVE

Fig. 2. — Map of Middle Butte Cave with cross sections.

and cracks appear on the ground surface over the
cavities. These surface features can be traced to de-
termine the underground extent of a tumulus. Lava
tubes or tunnels are formed by the cooling of the
surface of the molten lava prior to the deeper por-
tions of the flow. The underlying still molten lava
then flows out from beneath the hardened crust leav-
ing a cylindrical cavity. Cavities in the lava or basalt,
whether formed by gas or by flowing lava, range
from a few centimeters to tens of meters in diameter
and/or length. The collapse of a portion of the roof
of these structures provides the entrance through
which debris can accumulate in the caves. Debris

accumulation can be attributed to climatic events
such as eolian or fluviatile sedimentation, to rap-
torial birds, carnivore denning, human habitation,
or some other source. The type and amount of debris
accumulation ranges from many meters to none at
all and depends on the number, size, and position
of openings, external and internal drainage patterns,
time of exposure to the outside, and the nature of
the prevailing winds.

Although lava blisters may have but a single open-
ing formed by collapse of part of the roof, the cracks
formed on the surface of the ground above the roof
enables water to trickle into the cave below.

THE CAVES

Moonshiner Cave

This cave, which was inhabited by a moonshiner
during Prohibition, is a lava blister formed in un-
differentiated basalts of Pleistocene age (Nace et al.,
1975). It is located 48 km west of Idaho Falls, Bing-
ham County, Idaho, and 0.8 km east of East Butte,
an extinct volcano.

The cave entrance is approximately 1.5 m wide
and 2 m long. A pile of breakdown is present 2.5
m below the entrance; it was formed by the collapse
of the portion of the roof which formed the entrance.

A careful search of the surface along pressure ridges
and of the inside of the cave revealed only one en-
trance, although the numerous cracks in the pressure
ridges would permit water to trickle into the cave.
The cave is Y-shaped (Fig. 1 ) with the tail projecting
90 m to the west, one arm projecting to the northeast
and the other to the southeast. All portions of the
floor lie essentially in the same plane with a slight
sloping to the east. The highest portion of the cave
ceiling is at the entrance. From this point, the ceiling
slopes domelike in all directions. Ten meters west

244

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

of the entrance the ceiling is 1 m high and it main-
tains this height to the end of the tunnel. The ceiling
of the northeast arm decreases to 0. 1 m only 1 5 m
from the entrance. The same reduction in ceiling
height is reached in the southeast arm 20 m from
the entrance. Immediately under the cave entrance
there is, from spring to late summer, a luxuriant
growth of a fern, Pteridium aquilinium (Fig. 1). In
winter and early spring, a snow cone forms which
has been observed to reach to within 0.5 m of the
cave entrance (Fig. IB). The largest concentration
of bone is situated near the entrance of the southeast
arm.

The small amount of sediments in the cave sug-
gests an eolian origin for the following reasons: 1)
the cave opening evidently has been present for
thousands of years; 2) water entered the cave through
the cave opening and seeped through the cracks in
the pressure ridges; 3) the pressure ridges are almost
denuded of sediments and water tends to drain away
to the sides. Thus it can be inferred that few sedi-
ments were washed into the cave. Wakefield Dort
(personal communication) suggested there may have
been some washing of sediments down slope into
the northeast and southeast arms (each of which was
traced over 100 m on the surface by following the
pressure ridges), thus removing the sediments from
the main chamber.

Middle Butte Cave

Middle Butte Cave is a lava blister which is con-
nected to a lava tube. Evidently the lava blister was

THE ]

The bones from both caves are essentially unal-
tered. Some of the specimens have been encrusted
with a carbonate deposit. This mineral encrustation
is present also on the ceilings, marking places where
water seeped through cracks in the basalt. A few
mummified remains of animals that had recently
fallen into the caves were found, but the majority
of the bones were disarticulated and scattered on
the cave floor. Although many bones had been
gnawed by rodents, none of the limb bones exhibited
unhealed fractures. This would suggest that the fall
into the cave was not sufficient to cause breakage of
limb bones and that not much scavenging occurred.
A few pathological specimens are present in the
samples but most are old and healed injuries, not
related to entrapment.

The outstanding feature common to the faunas of

formed before the underlying lava tube was formed
making the lava blister an outpocketing of the lava
tube. The surface features on the ground above the
lava blister and the connected lava tube can be traced
70 m northeast of the entrance to the cave to a large
collapsed feature. A portion of the lava tube that
has not collapsed extends south and west of the
collapsed feature. This preserved portion has filled
to within 1.5 m of the ceiling with fine laminated
sediments. This seems to be the entrance by which
water entered to deposit sediments in Middle Butte
Cave. Sedimentation had essentially closed off this
opening farther back in the lava tube prior to the
formation of the entrance into the cave.

Middle Butte Cave is located 8 km east of Atomic
City, Bingham County, Idaho, and 1.6 km south of
Middle Butte, another extinct volcano, from which
it gets its name. The distance between Middle Butte
Cave and Moonshiner Cave is 8 km.

This cave is roughly triangular in outline when
viewed from above (Fig. 2). The ceiling is domed,
the highest point being approximately 2.5 m west
of the entrance. The cave entrance is 2 m above the
floor. Unlike the vertical entrance of Moonshiner
Cave, the opening of Middle Butte Cave is sloped
at about a 45 degree angle. The passage from the
surface into the cave is about 2.5 m long. The sides
of the cave are smooth except in areas where rock
fall from the ceiling has occurred.

The temperature inside both caves remains con-
sistently around 8° C.

the two caves is the abnormally high number of
carnivores (Figs. 3 and 4), a condition that is com-
parable to that described for the fauna from the
Rancho La Brea tar pits (Marcus, 1960; Stock, 1972).

The majority of the bones from both caves are of
pygmy rabbits, cottontails, jack rabbits, sagebrush
voles, red foxes, coyotes, badgers, and long-tailed
weasels, all of which occur in the sagebrush-grass
community found in the region today. Martes amer-
icana is is present in Middle Butte Cave, whereas
M. nobilis is in the Moonshiner Cave fauna. The
absence of M. nobilis from Middle Butte Cave sug-
gests that entrapment in this cave occurred later
than in Moonshiner Cave. Grizzly bear, wolf, kit
fox, marten, black-footed ferret, wolverine, wapiti,
and bison, although present in the cave faunas, are
not found in the area today. The current absence of

1984

WHITE ET AL. -CARNIVORE TRAPS

245

30% CARNIVORES 2/o

70% HERBIVORES 98%

BARRO COLORADO
ISLAND

42%

Carnivores

5%

58%

Herbivores

95%

MIDDLE BUTTE CAVE

18 NON-CARNIVORE
TRAP CAVES

Fig. 3. — Pie diagrams, based on minimum numbers of individuals, comparing the ratios of carnivores to herbivores in the two selective
trap caves, a Recent fauna, and 18 combined non-selective trap caves (See Table 2).

bison, wapiti, and wolf is probably related to the
agency of man, whereas the disappearance of mar-
ten, black-footed ferret, and kit fox may be related

to environmental changes. A detailed study of all
faunal elements is in progress and will be published
at a future date. A faunal list is given in Table 1.

CAVES AS CARNIVORE TRAPS

Observing the caves as they are today, one can
hypothesize how the caves may have functioned as
selective carnivore traps. The preponderance of car-
nivores in the faunas of both caves suggests that

there must have been a bias of some type occurring
during the accumulation of the bones. The use of
caves as dens comes to mind. The difficulty to exit
after entry into the caves precludes this possibility.

246

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

42% Carnivores

40%

58% Herbivores

60%

MOONSHINER CAVE MIDDLE BUTTE CAVE

Fig. 4. — Pie diagrams, based on biomass, comparing the ratios of carnivores to herbivores in the two selective trap caves (See Table 1).

If the accumulation of animal bones had taken place
in a cave with an entrance 7 m or more above the
floor and creating a deadfall situation, or a cave with
an entrance that provided easy ingress and egress,
the resulting fauna probably would have a normal
carnivore-herbivore distribution.

In Moonshiner Cave, from late spring to early
autumn, the luxuriant growth of ferns immediately
under the entrance gives the illusion that the dis-
tance to the floor is 1.5 m, an easy jump for car-
nivores such as coyotes and bobcats. An herbivore
caught in the cave would have had little chance to
escape, and a hungry carnivore, enticed by the sight,
smell, or noise of an entrapped animal, could have
jumped into the cave after it. Although an easy jump
into the cave, the 2.5 m height of the entrance would
require running in a straight line to obtain the nec-
essary impetus to jump out. The rubble resulting
from rock fall from the ceiling would have made
running in a straight line difficult, thus trapping most

carnivores. Weakened or dead from starvation, a
carnivore could become additional bait for the trap.
Carnivores are opportunistic, and as Ricklefs (1973:
531) puts it, “One would expect extreme selectivity
by predators that spend much time searching for
and pursuing prey; the predators should choose the
easiest catch, the individual with a relatively small
chance of escape

In winter, at Moonshiner Cave, a soft and pre-
cipitous snow cone develops which reaches to within
a half meter of the ground level (Fig. 1 B). This cone
is formed by falling and drifting snow and is easily
broken down by even a small object falling on it.
Small herbivores or carnivores could easily jump
onto the cone which would collapse under them and
thus trap them in the cave. This kind of entrapment
could occur by voluntary action of both herbivores
and carnivores or as the result of pursuit.

In contrast, the entrance into Middle Butte Cave
slants through approximately 2 m of basalt. The

1984

WHITE ET AL. — CARNIVORE TRAPS

247

Table 1 . — Faunal list with minimum number of individuals for Moonshiner Cave and Middle Butte Cave, Bingham County, Idaho. "P"
indicates a group to be represented in the fauna, for which an accurate identification and census was not made. "IVt” indicates the

average weight for each taxon, which was used to calculate the biomass.

Taxon

Moonshiner

Cave

Middle Butte
Cave

Wt

Class Reptilia— reptiles
Order Lacertilia — lizards

p

p

Order Ophidia— snakes

Family Colubridae— racers, king snakes, and others

p

p

Family Crotalidae— rattlesnakes

p

p

-

Class Aves — birds
Order Falconiformes— raptors
Family Cathartidae — New World vultures
Cathartes aura— Turkey Vulture

i

Family Accipitridae— eagles and hawks
Aquila chrysaetos— Golden Eagle

l

_

Order Galliformes— grouse, ptarmigan, and others
Family Tetraonidae— pheasants, quail, and others
Centrocercus urophasianus— Sage Grouse

4

6

Order Strigiformes— owls
Family Strigidae— owls
Asia flammeus— Short-eared Owl

1

Order Piciformes— woodpeckers
Family Picidae — woodpeckers
Colaptes awra/Ms-Common Flicker

1

Order Passeriformes— perching birds
Family Corvidae— crows and jays

1

Family Stumidae — starlings

Sturnus vulgaris— Starling

1

—

—

Class Mammalia — mammals
Order Insectivora— moles, shrews, and others
Family Soricidae — shrews
Sorex cf. S. cinereus

163

7 g

Order Chrioptera— bats
Family Vespertilionidae — vespertilionid bats
unidentified bats

21

1

Order Lagomorpha — rabbits, hares, pikas
Family Ochotonidae— pikas
Ochotona cf. O. princeps— pika

1

142 g

Family Leporidae — rabbits and hares

Brachvlagus idahoensis— pygmy rabbit

230

139

440 g

Sylvilagus nuttallii— Nuttall’s cottontail

211

82

755 g

Lepus americanus— snowshoe hare

9

17

1.4 kg

Lepus cf. L. californicus— black-tailed jack rabbit

2

3

2.3 kg

Lepus cf. L. townsendii— white-tailed jack rabbit

15

12

3.4 kg

Order Rodentia — rodents
Family Sciuridae— squirrels
Eutamias minimus— least chipmunk

27

35 g

Spermophilus townsendii— Townsend’s ground squirrel

27

—

120 g

Spermophilus richardsoni— Richardson's ground squirrel

4

—

410 g

Spermophilus lateralis— golden-mantled ground squirrel

6

—

226 g

Spermophilus sp.— ground squirrel

57

69

200 g

Marmota flaviventris— yellow-bellied marmot

56

4

6 kg

Family Geomyidae— pocket gophers
Thomomys cf. T. talpoides— northern pocket gopher

123

17

150 g

Family Heteromyidae — pocket mice, and others
Perognathus cf. P. parvus— silky pocket mouse

23

24 g

Family Castoridae— beavers
Castor canadensis— beaver

1

_

22 kg

248

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 1. — Continued

Taxon

Moonshiner

Cave

Middle Butte
Cave

Wt

Family Cricetidae— New World rats and mice

Peromyscus maniculatus— deer mouse

561

3

35 g

Neotoma cinerea— bushy-tailed woodrat

48

75

395 g

Microtus montanus— montane vole

1

—

60 g

Microtus longicaudus-\ong-tai\ed vole

1

—

60 g

Lagurus cf. L. curtatus— sage vole

1

-

30 g

Clethrionomys cf. C. gapperi— red-backed vole

1

—

30 g

Phenacomys cf. P. intermedius— heather vole

1

—

35 g

Unidentified voles

229

5

35 g

Family Erethizontidae — porcupines

Erethizon dorsatum— porcupine

1

1

5 kg

Order Artiodactyla— even-toed hoofed mammals
Family Cervidae— deer, wapiti, and others

C'ervus cf C. elephas— wapiti

2

2

275 kg

Family Antilocapridae — pronghorns

Antilocapra americana— pronghorn

4

1

50 kg

Family Bovidae — bison, cattle, sheep

Bos laurus— cattle

P

P

—

Bison bison — American bison

P

P

—

Bos or Bison

7

2

900 kg

Ovis aries— domestic sheep

1

1

75 kg

Order Carnivora — carnivores
Family Canidae — dogs

Canis lupus— timber wolf

11

4

51 kg

Canis latrans— coyote

32

23

18 kg

Vulpes vulpes— red fox

120

46

4.8 kg

Vulpes macrotis— kit fox

2

1 1

2 kg

Family Ursidae— bears

Ursus a ret os— grizzly bear

13

2

205 kg

Family Mustelidae— mustelids

Mustela frenata — long- tailed weasel

293

153

300 g

Mustela erminea— ermine

35

13

190 g

Mustela sp. (M. frenata or M. erminea)

109

25

245 g

Mustela nivalis (=M. rixosa)— least weasel

8

—

40 g

Mustela nigripes— black-footed ferret

2

—

700 g

Mustela vison— mink

1

—

1 kg

Martes nobilis— extinct pine marten

25

—

1.5 kg

Martes cf. M. americana— pine marten

—

2

850 g

Gulo gulo — wolverine

25

5

21 kg

Taxidea taxus— badger

55

15

8.2 kg

Spilogale gracilis— spotted skunk

24

6

600 g

Mephitis mephitis— striped skunk

1

—

1.7 kg

Family Felidae— cats

Lynx canadensis— C anada lynx

12

—

10.1 kg

Lynx rufus— bobcat

12

—

11.2 kg

Lynx sp. (L. canadensis or L. rufus)

-

14

10.2 kg

sides of this opening are smooth and almost without
crevices for feet or claws to obtain footholds. The
opening, which is 1 m2 and passes through 2.5 m
of basalt, permits but a few minutes of direct sun-
light to enter the cave each day and provides for an
abrupt change from light to darkness. Dim light near
the entrance may have affected the depth perception

of a carnivore making ready to jump into the cave,
or, in the case of pursuit, there would be insufficient
time to adjust to the difference of light intensity.
Any attempt to leap out of the cave would be made
considerably more difficult without “footholds” on
the sides of the entrance, and by the presence of
rubble on the floor.

1984

WHITE ET AL. — CARNIVORE TRAPS

249

B

i i i i i i i i i i

30 35mm 40

Fig. 5. — Modified Dice-Leraas diagram comparing the basilar length of skull (of Hensel) of male, extant Mustela erminea invicta from
Idaho (A), with M. erminea from Moonshiner Cave (B). Data for “A” from E. R. Hall (personal communication).

AGE OF THE FAUNAS IN MOONSHINER AND MIDDLE BUTTE CAVES

The presence of Maries nobilis , the only extinct
species found in Moonshiner Cave, suggests that it
may have been acting as a trap longer than Middle
Butte Cave. Jaguar Cave in Lemhi County, Idaho,
has M. nobilis as a member of the fauna with a date
of 10,370 ± 530 years B.P. (Kurten and Anderson,
1972). Webster (1978) reported M. nobilis from a
level dated between 3,300 and 2,550 years B.P. in
Dry Creek Rock Shelter, Ada County, Idaho. There
are no extinct taxa from Middle Butte Cave. As
discussed above, the fauna from Middle Butte Cave
cannot be any older than 5,175 years B.P. as esti-
mated from the rate of sedimentation and is prob-
ably much younger. Despite the evidence that M.
nobilis may have survived until relatively recent
time in the state, we think the fauna from Moon-
shiner Cave is older than that from Middle Butte
Cave and that it started functioning as a trap at an
earlier date.

The presence of pine marten, wolverine, ermine,
least weasel, and Canada lynx indicates a cooler

climate. Specimens of Mustela erminea from
Moonshiner Cave are significantly larger than spec-
imens of this species occurring north of the area
today, M. e. invicta, and are about the size of M. e.
richardsoni from near Fort Franklin, Mackenzie,
some 2,400 km to the north (Fig. 5) (Hall, 1951).
This situation is akin to that in the Conard Fissure
in Arkansas. Regarding the latter. Hall (1951:167)
writes, “It may be significant that the cranial char-
acters of the female ermine from there are most
nearly approximated among Recent weasels by those
which live along the southern edge of the frozen
tundra.”

The bone accumulation in both caves must have
taken place over a long period of time, because it
seems unlikely that 30 wolverines and 446 weasels
could exist in the same vicinity at the same time.
The caves are separated by 8 km and Moonshiner
Cave was open at the time Middle Butte Cave began
functioning as a trap.

COMPARISONS WITH OTHER CAVES

Comparisons with other cave deposits provide
information pertinent to our carnivore trap hypoth-
esis. The fauna from Meyer Cave, Illinois (Parma-
lee, 1967), includes a large number of carnivores;
however, the percentage (5.9%) is not so dispro-
portionate as in our two Idaho cave faunas. The
Meyer Cave bone deposit is at the bottom of a ver-
tical fissure with a drop of 4.2 m. The bone deposit
terminated at a depth of 6.4 m below the top of the
fissure. The opening into the fissure is connected to
the outside of the bluff by a 4.5 m horizontal crawl-
way. Easy access to the crawlway is made possible

by a talus slope. Remains of mountain lion kittens,
raccoon pups, numerous bobcat kittens, and fox pups
in Meyer Cave suggest that it was used as a den,
and that the accumulation of carnivore bones in the
bottom of the fissure was due to occasional acci-
dental falls. The relatively small number of adult
carnivores from the bottom of the fissure suggests
that the opening to the pitfall was generally avoided.

There is a striking similarity in the composition
of the fauna and distance from the cave opening to
the floor, between our two Idaho caves and the
Chimney Rock Animal Trap Cave, Larimer County,

250

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Colorado (Hager, 1972). In the latter cave the dis-
tance from the cave entrance to the floor is approx-
imately 2.5 m and the fauna has a . . preponder-
ance of carnivores . . .” (Hager, 1972:68).

The bone deposit at New Paris No. 4, Pennsyl-
vania (Guilday et al., 1964), accumulated in a dome
pit formed in a narrow fissure connected to the sur-
face and adjacent to a sterile sinkhole, 9 m in di-
ameter. The large size of the sinkhole makes it an
obvious obstacle to avoid, and the rubble formed
by the collapse of the roof formed a slope too gentle
to function as an efficient trap. The small adjacent
dome pit with its restricted opening at the surface
functioned as an efficient pitfall trap which resulted
in the collection of a random sample of the fauna
in the area; it included only five species of carni-
vores, all mustelids. The vertical distance of 9 m
from the top of the fissure to the bottom seems too
large to induce carnivores to enter, despite the wealth
of carcasses accumulated on the bottom. The depth
and restricted dimensions of the fissure may have
prevented carnivores from knowing of the trapped
animals.

A final comparison can be made with an open
sink hole accumulation, Nichol’s Hammock, Flor-
ida (Hirschfeld, 1968). The sinkhole is 3.7 m to 4.6
m deep and 3 m wide. Unlike the large sinkhole at
New Paris, the walls are vertical with a slight inward
curvature at the top. This configuration insures that
any animal trapped in the pit would not be able to
get out. The mammalian fauna from the sinkhole
was composed of a minimum of 44 individuals of
13 species, 28% (12 individuals) of which were car-
nivores. Perhaps the trapped herbivores in the sink-
hole served to attract carnivores to the edge, and in
an attempt to find a way down they might have
accidentally fallen into the sinkhole. Unlike Moon-
shiner and Middle Butte Caves, the drop into the

plainly visible sinkhole would be far too great to
assume that any would jump in voluntarily.

Rather than extending these comparisons to in-
clude all known Pleistocene cave faunas, a synopsis
is presented in Table 2. Caves for which MNI’s have
been published are listed along with the percentage
of carnivores in each one. The high representation
of carnivores in Moonshiner and Middle Butte Caves
can easily be seen. At the bottom of the table, the
percentage of carnivores for all faunas is only 4.7 in
contrast to 32.8 for the two Idaho caves. The per-
centage of carnivores in the fauna generally ranges
from zero to 1 1%. Three of these faunas have higher
than expected numbers of carnivores, although not
so high as in the Idaho caves. Guilday and Adam
(1967) attribute the high number of carnivores at
Jaguar Cave to its use as a denning site. The high
percentage of carnivores in the other two faunas,
Nichol’s Hammock and the First American Bank
Site, are a result of small samples, 44 and 3 1 MNI,
respectively. In smaller samples, the representation
of rarer individuals (in this case the carnivores) will
be exaggerated (Grayson, 1978; Payne, 1972). The
samples from Moonshiner and Middle Butte Caves
are large enough to overcome this problem. It must
be admitted that continued counting of small ro-
dents in the Idaho cave samples would mathemat-
ically decrease the percentage of medium and large-
sized carnivores; but, at the same time, it would not
greatly affect the percent biomass they represent (see
Fig. 5). An increased number of mice in the faunal
count would not alter the absolute number of car-
nivores from the caves. To explain the large number
of carnivores in the Idaho caves, we offer the hy-
pothesis of selective entrapment based on the pe-
culiar configuration of the two caves and the be-
havior of the carnivores.

FACTORS RELATING TO CAVES AS CARNIVORE TRAPS

Age Distribution

The age distribution of medium to large-sized car-
nivores, especially in Moonshiner Cave, suggests
that juveniles and very old adults were more often
entrapped than prime adults. This may be related
to the smaller size of younger animals and the over-
all weakened condition of the older adults. Thus,
prime adult animals may have been able to escape.

Diversity Index

The calculation of the diversity index (H') and
equilibrity (E) for each of the faunas in Table 2 also
reflects the degree of bias toward carnivores in the
fauna. Species diversity (H') of each fauna was com-
puted with the formula:

S

H' = 2p ,ln P.

i= 1

1984

WHITE ET AL. -CARNIVORE TRAPS

251

Table 2. — The numbers of mammalian carnivores and herbivores recorded from several North American caves. Bats are excluded. See
text for explanations of diversity index, maximum diveristy, and equiiikriiy.

Locality

Number

of

species

MNI

MNI

carnivores

Percent

carnivores

Diversity

index

(H')

Maximum

diversity

(H^

Equili-

brity

(E)

Diversity

index

carnivores

Percent

diveristy

carnivores

Arizona:

Papago Springs cave (Skinner, 1 942)

27

217

14

6.5

2.297

3.296

0.697

0.306

13.3

Florida:

Nichol’s Hammock (Hirschfeld, 1968)

14

44

10

22.7

2.227

2.639

0.844

0.646

29.0

Idaho:

Jaguar Cave (Guilday and Adam, 1967;
Kurten and Anderson, 1972)

39

549

88

16.0

2.679

3.664

0.371

0.679

25.3

Middle Butte Cave (this paper)

29

756

319

42.2

2.458

3.367

0.730

1.031

42.0

Moonshiner Cave (this paper)

43

2,591

780

30.1

2.666

3.761

0.709

0.893

33.5

Illinois:

Meyer Cave (Parmalee, 1967)

40

8,870

523

5.9

1.763

3.689

0.478

0.224

12.7

Kentucky:

Welsh Cave (Guilday et al., 1971)

23

127

6

4.7

2.149

3.135

0.685

0.207

9.6

Missouri:

Brynjulfson Cave 1 (Parmalee and
Oescfa, 1972)

45

247

27

10.9

3.130

3.807

0.822

0.511

16.3

Brynjulfson Cave 2 (Parmalee and
Oesch, 1972)

37

781

63

8.1

2.424

3.611

0.671

0.351

14.5

Crankshaft Cave (Parmalee et al.,
1969)

55

1,926

47

2.4

2.454

4.007

0.612

0.148

6.0

Pennsylvania:

Bootlegger Sink (Guilday et al.,
1966)

32

63

9

12.0

3.145

3.466

0.908

0.464

14.8

Hosterman’s Pit (Guilday, 1967)

8

12

0

0.0

1.814

2.079

0.873

0

0

New Paris No. 4 (Guilday et al.,
1964)

36

1,363

5

0.3

2.469

3.584

0.689

0.024

1.0

Sheep Rock Shelter (Guilday and
Parmalee, 1965)

32

287

29

10.3

2.429

3.466

0.701

0.349

14.4

Tennessee:

First American Bank Site (Guilday,
1977)

20

31

7

22.6

2.891

2.996

0.965

0.669

23.1

Robinson Cave (Guilday et al., 1969)

41

725

21

2.9

2.653

3.714

0.714

0.160

6.0

Virginia:

Clark’s Cave (Guilday et al., 1977)

45

2,754

18

0.7

2.537

3.807

0.666

0.044

5.1

Natural Chimneys (Guilday, 1962)

50

523

14

1.9

3.130

3.912

0.800

0.159

5.1

West Virginia:

Eagle Cave (Guilday and Hamilton,
1973)

26

53

2

3.3

2.928

3.258

0.899

0.150

5.1

WVoming:

Little Box Elder Cave (Anderson,
1968)

48

2,917

120

4.1

2.327

3.871

0.601

0.204

8.7

Subtotal: (excluding Moonshiner and
Middle Butte Caves)

21,489

1,003

4.7*

11.9*

Subtotal (Moonshiner and Middle Butte
Caves)

3,347

1,099

32.8*

37.8*

Grand total:

24,836

2,102

= means.

252

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

where p, represents the proportion of the total num-
ber contributed by the ith species, s represents the
number of species in the fauna (Pielou, 1966), and

E = H/Hmax

where Hmax is the natural logarithm of the number
of observed species. The carnivores in the faunas
from Moonshiner and Middle Butte Caves contrib-
ute 33.5 and 42.0, respectively, to the faunal diver-
sity. None of the carnivores in the other cave faunas
come close to these numbers with respect to their
contribution toward faunal diversity. Nichol’s
Hammock, Jaguar Cave, and the First American
Bank Site do stand out in that the carnivores con-
tribute greatly to the total faunal diversity but they
still do not approach those of the carnivores in the
Idaho caves. Because of energy flow and their tro-
phic position, carnivores do not form a major por-
tion of a modem community in relation to other
animals, neither in number of individuals nor in
number of species (Odum, 1968, 1971).

Few complete studies of modern vertebrate fau-
nas are available. The study of nonvolant mammals
on Barro Colorado Island by Eisenberg and Thor-
ington (1973) lists the carnivores as comprising 2.1%
of the individuals and 2.7% of the biomass. They
also cite a study of Suriname in which the carnivores
contribute 2.06% of the individuals and 1.16% of
the biomass. Many of the caves listed in Table 2,
such as Meyer, Welsh, Crankshaft, Robinson, Eagle,
and Little Box Elder caves, compare with these per-
centages for modern faunas.

That the carnivores in Moonshiner and Middle
Butte Caves contribute such a large percentage to-
ward total numbers of individuals in the faunas and
to overall faunal diversity, attests to the uniqueness
and effectiveness of the caves as traps.

Closed Cave System

In addition to the trapping of animals, the con-
figuration of the Idaho caves also prevents post-
mortem removal of bones. In stream deposits there
may be a selective winnowing of bones (Voorhies,
1969) and on the surface there may be a scattering
of the bones by scavengers resulting in the loss of
skeletal elements (Hill, 1980). In cave systems such
as we have described for the selective traps, theo-
retically, there is no loss of bone. Although carcasses
may be tom apart and scattered by the entrapped
carnivores, they still remain in the cave. Specimens
can only be lost by breakage due to chewing or gnaw-
ing, being eaten and completely digested, being sub-

jected to leaching and decomposition resulting from
cave dampness, or removal by humans prior to the
time we made the collections.

Population Density and Body Size

A major consideration is the population density
of a particular species. The absolute number of in-
dividuals within a given geographical area is related
to body size. It is a standard axiom in ecology that
the larger the body size of an animal, the fewer the
numbers of individuals within a given area. For
example, the large number of long-tailed weasels,
Mustela frenata (293 in Moonshiner Cave and 178
in Middle Butte Cave), versus the number of grizzly
bears, Ursus arctos (13 and 2, respectively) fits this
axiom.

Territoriality

Territoriality is also a factor. The wolverine, Gulo
gulo, for example, is a solitary animal; a single male
may range over an area of 3,000 km2 which it will
share with two or three females. Covering a large
territory requires a great deal of time and the chances
are slim that a wolverine would discover an her-
bivore shortly after it had fallen into the cave. There
would be a time lag before a new individual would
move into the newly vacated territory. Even after
the invasion of the new territory, it is still possible
that the new animal would not come into the vi-
cinity of the cave. Because wolverines are solitary,
antisocial animals, the presence of 25 individuals
in Moonshiner Cave would argue not only for the
effectiveness of the trap but also for the considerable
length of time over which it functioned.

Rare Species

The smallest carnivore in the cave faunas and in
North America today is the least weasel, Mustela
rixosa ( M . nivalis of some authors). It is also one of
the rarest mammals, both in fossil and modem fau-
nas. This pattern is duplicated in Moonshiner Cave
where only eight individuals were recovered. The
presence of eight individuals of a rare carnivore pre-
served in the cave strengthens the idea that the cave
was acting as a baited trap over a long period of
time.

Body Size and Height of Ceiling

One of the main criteria for the selectivity of the
Idaho caves as traps is the floor to ceiling height.
This criterion applies especially to large carnivores
because the cave opening to floor distance is small

1984

WHITE ET AL. -CARNIVORE TRAPS

253

in proportion to body size, but it does not account
for the abundance of Mustela frenata. The average
size of the individuals of this species is less than one
tenth of the distance from the cave floor to the en-
trance. Therefore it is doubtful that weasels would
jump down into the cave from late spring to early
autumn, although some may have fallen in while
trying to find a way down into the cave.

The long-tailed weasel is known for its voracious
feeding behavior. It has a well-developed sense of
smell and follows the trails of its prey by scent,
loping along until it overtakes its victim, then kills
it by biting the base of the skull. It may kill more
than it can eat but often caches the excess for future
meals, especially in late summer and autumn. Stud-
ies on the food habits of M. frenata (Hall, 1951;
Quick, 1951) show that mice, especially Microtus
and Peromyscus, are favored prey. When mice are
scarce, long-tailed weasels take pocket gophers,
chipmunks, ground squirrels, rabbits, shrews, and
other animals. Thus, almost any mammal, ranging
in size from a rabbit down to a shrew, could serve
to bait the caves as traps for weasels. Weasels eat
about one third of their weight every 24 hrs. Pop-
ulation density ranges from a low of one per 1 .3 km2
(Quick, 195 1) to as many as one per 0.25 km2 (Burt
and Grossenheider, 1976). They have a small home
range and a short cruising radius.

As noted above, from late autumn to early spring,
snow drifts through the entrance of Moonshiner Cave
and forms a large cone that almost reaches up to
ground level. This snow cone reduces the floor to
ceiling height to less than a meter and it seems prob-
able that smaller-sized carnivores could make the
relatively easy jump onto the snow cone. Since the
cone is formed by drifting snow, it probably would
break down to some extent by the weight of the
jumping animal, making it difficult to jump out again.
Entrapment of large carnivores in Moonshiner Cave,
therefore, probably occurred in summer and winter,
whereas smaller carnivores were probably only
trapped in winter.

The age distribution of the numerous individuals
of Mustela frenata in the two caves support our
contention of its winter entrapment. Employing the
criteria used by Hall (1951: 25), the crania of M.
frenata from the caves were segregated into age
groups, as follows: juveniles— individuals with milk
teeth, birth to 3 months; young— individuals with
open sutures between the maxilla and premaxilla
and between premaxilla and nasals, 3 to 7.5 months;
subadults— individuals with sutures between max-

illa and premaxilla visible but indistinct, 7.5 to 10
months; adults— individuals with bones of rostrum
coalesced and no traces of sutures visible to the
naked eye, over 10 months.

Of 293 individuals from Moonshiner Cave, 66.6%
were adults, 26.6% were subadults, and 6.8% fell
into the young age group. No juveniles were present.
The sample of 178 crania from Middle Butte Cave
produced 71.3% adults, 23.4% subadults, 5.3%
young, and no juveniles.

Generally, weasels are bom in April (Wright,
1948). It would be highly improbable for a juvenile
to become entrapped, because by the time the young
weasels moved away from the den, the snow cone
would have melted away from underneath the cave
entrance (Fig. 6). By the time they had matured
enough to hunt for themselves, it would be late au-
tumn or early winter, at which time they would have
achieved the young or early subadult age class. This
coincides with the first snow fall in the area of the
caves between late October and early December and
the first appearance of a snow cone under the en-
trance (Fig. 1). Although one might assume that
more of the younger, inexperienced animals would
be trapped, this is not the case. Most of the indi-
viduals are in the adult age class.

At Middle Butte Cave, the slope of the entrance
precludes the formation of a snow cone and a growth
of ferns, and it reduces the amount of light in the
cave, thus insuring an abrupt change in light inten-
sity. In daytime, weasels or any other predator in
pusuit of prey entering the cave would have insuf-
ficient time to adjust to the abrupt change in light
intensity and, consequently, would fall into the cave.
At night, a pursuing predator may not have been
able to stop in time on the smooth, inclined surface
of the entrance. The relatively large number oflepo-
rid bones in this cave (Fig. 3), coupled with the
knowledge that rabbits often go into any available
shelter to escape a predator, suggests that many car-
nivores were trapped while in pursuit of rabbits. In
this cave, we cannot account for the absence of ju-
venile weasels.

Carnivore Habits

The biased representation of carnivores in the
Idaho selective-trap caves is related essentially to
two factors: 1) the configuration of the caves, which
has already been discussed and can be considered a
constant; 2) the behavior and habits of carnivores
leading to entrapment in the caves. The latter factor

254

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

cr

CD

Fig. 6. — Diagram with the age-distribution of extant weasels by month, and the average depth per month of snow on the ground at the
Idaho National Engineering Laboratory (data from O. Doyle Markham, personal communication).

is extremely variable and depends on the behavior
and habits of the species in question.

A brief description of the habits of some carni-
vores (Burt and Grossenheider, 1976; Walker, 1975)
will round out this discussion. The spotted skunk,
Spilogale gracilis, is carnivorous, but its food varies
seasonally. In summer, it feeds mainly on vegetable
matter and insects, whereas in winter it feeds on
rodents and other small mammals. They are noc-
turnal and have a population of more than one per
ha and a home range of 64 ha. The small size of
these skunks suggests that they probably were trapped
in Moonshiner in winter.

Although the pine marten, Maries nobilis, from
Moonshiner Cave is extinct, its habits probably were
similar to those of the living M. americana which
is present in Middle Butte Cave. This small, active,
mainly nocturnal carnivore feeds on rodents and
rabbits but takes carrion in the winter. A population
density of fewer than one per km2 is probably high,
and the presence of 25 MNI in Moonshiner Cave
contributes to our interpretation of selective en-
trapment.

The badger, Taxidea taxus, is fossorial, preys on

rodents and rabbits, and uses carrion as an impor-
tant part of its diet. As are wolverines, badgers are
solitary animals but have a smaller home range of
about 850 ha (Long, 1973). The relatively larger
number of badgers in Moonshiner Cave (55 badgers
and 25 wolverines) correlates with the smaller home
range, thus increasing the chances for an individual
badger to find the cave, certainly better than for
wolverines.

The canids, Vulpes vulpes, Canis latrans , and
C. lupus, have keen senses of smell and hearing,
hunt at night, and often feed on carrion. They pro-
vide another example of smaller carnivores with
small home ranges being more numerous than larger
ones with larger home ranges. In Moonshiner Cave,
there were 120 red foxes, 32 coyotes, and 1 1 wolves,
and in Middle Butte Cave there were 46 red foxes,
23 coyotes, and 4 wolves.

A final example is the genus Lynx (L. rufus and
L. canadensis ), solitary and nocturnal cats that feed
on rabbits, hares, rodents, and birds. A minimum
number of 22 Lynx were identified at Moonshiner
and 14 at Middle Butte Cave.

Several points are evident from the above:

1984

WHITE ET AL. — CARNIVORE TRAPS

255

1 . Most carnivores hunt by scent and feed on car-
rion.

2. Most carnivores are nocturnal.

3. Small carnivores are generally more abundant
and have smaller home ranges than larger ones.

4. Based on what is known of home range, popu-
lation density, and food habits of modern car-

nivores, the proportion of carnivores to carni-
vores in the caves is the same as in a modern
regional fauna.

All these factors lead us to conclude that the caves
operated over a long period of time as baited traps
which were easy to enter but difficult to escape.

ACKNOWLEDGMENTS

We greatly acknowledge the financial support of Mr. and Mrs.
E. Hadley Stuart, Jr., in the excavation and field study especially
of Middle Butte Cave. For permission to collect materials from
Moonshiner Cave, we thank the officials of the Idaho National
Engineering Laboratory. Tom Amberson, Penn Gildersleeve,
Mark Johnson, Scott McDonald, and Jim White aided in the
collection of the cave materials. Wakefield Dort and William H.
Burt provided many invaluable comments on the geology and

zoology of the caves, respectively. We thank O. Doyle Markham,
Department of Energy, Idaho Operations Office, for providing
data on snowfall during 14 years at the Idaho National Engi-
neering Laboratory. E. Raymond Hall kindly provided mea-
surements of extant Mustela erminea invicta from Idaho. John
E. Guilday, Barry L. Keller, Ernest L. Lundelius, Jr., and Michael
R. Voorhies critically read the manuscript and made many valu-
able suggestions. John Moeller made the illustrations.

LITERATURE CITED

Anderson, E. 1968. Fauna of the Little Box Elder Cave, Con-
verse County, Wyoming: The Carnivora. Univ. Colorado
Stud. Ser., Earth Sci., 6:1-59.

Brown, B. 1929. A Miocene camel bed-ground. Nat. Hist., 29:
658-662.

Burt, W. H., and R. P. Grossenheider. 1976. A field guide
to the mammals. Houghton Mifflin Co., Boston, 3rd ed., 289

pp.

Daly, R. A. S914. Igneous rocks and their origin. McGraw-
Hill, New York, 563 pp.

Eisenberg, J. F., and R. W. Thorington, Jr. 1973. A prelim-
inary analysis of a Neotropical fauna. Biotropica, 5: 1 50-16 1 .

Grayson, D. K. 1978. Minimum numbers and sample size in
vertebrate faunal analyses. Amer. Antiq., 43:53-65.

Guilday, J. E. 1962. The Pleistocene local fauna of the Natural
Chimneys, Augusta County, Virgina. Ann. Carnegie Mus.,
36:87-122.

. 1967. The climatic significance of the Hosterman’s Pit

local fauna. Centre County, Pennsylvania. Amer. Antiq., 32:
231-232.

. 1977. Sabertooth cat, Smilodon floridanus (Leidy), an

associated fauna from a Tennessee Cave (40 DV 40), the
First American Bank Site. J. Tennessee Acad. Sci., 52:84-
94.

Guilday, j. E., and E. K. Adam. 1967. Small mammal remains
from Jaguar Cave, Lemhi County, Idaho. Tebiwa, 10:26-
36.

Guilday, J. E., and H. W. Hamilton. 1973. The Late Pleis-
tocene small mammals of Eagle Cave, Pendleton County,
West Virginia. Ann. Carnegie Mus., 44:45-58.

Guilday, J. E., and A. D. McCkady. 1966. The bone breccia
of Bootlegger Sink, York County, Pennsylvania. Ann. Car-
negie Mus., 38:145-163.

. 1969. The Pleistocene vertebrate fauna of Robinson

Cave, Overton County, Tennessee. Palaeovertebrata, 2:25-
75.

. 1971. The Welsh Cave peccaries (Platygonus) and as-

sociated fauna, Kentucky, Pleistocene. Ann. Carnegie Mus.,
43:249-320.

Guilday, J. E., and P. W. Parmalee. 1965. Animal remains
from the Sheep Rock Shelter (36 Hu 1), Huntingdon County,
Pennsylvania. Pennsylvania ArchaeL, 35:34-49.

Guilday, J. E., P. S. Martin, and A. D. McCrady. 1964. New
Paris No. 4: a Late Pleistocene cave deposit in Bedford Coun-
ty, Pennsylvania. Bull. Natl. Speleol. Soc., 26:121-194.

Guilday, J. E., P. W. Parmalee, and H. W. Hamilton. 1977.
The Clark’s Cave bone deposit and the Late Pleistocene
paleoecology of the central Appalachian Mountains of Vir-
ginia. Bull. Carnegie Mus. Nat. Hist., 2:1-87.

Hager, M. W. 1972. A Late Wisconsin-Recent vertebrate fauna
from the Chimney Rock Animal Trap, Larimer County, Col-
orado. Univ. Wyoming Contrib. Geol., 11:63-71.

Hall, E. R. 1951. American weasels. Univ. Kansas Publ., Mus.
Nat. Hist., 4:1-466.

Hill, A. P. 1980. Early postmortem damage to the remains of
some contemporary East African mammals. Pp. 131-152,
in Fossils in the making, vertebrate taphonomy and paleo-
ecology, Univ. Chicago Press, Chicago, Illinois, 338 pp.

Hirschfeld, S. E. 1968. Vertebrate fauna of Nichol’s Ham-
mock, a natural trap. Quart. J. Florida Acad. Sci., 31:1 77—
189.

Kurten, B. 1976. The cave bear story. Columbia Univ. Press,
New York, 163 pp.

Kurten, B., and E. Anderson. 1972. The sediments and fauna
of Jaguar Cave: II the fauna. Tebiwa, 15:21-45.

Long, C. A. 1973. Taxidea taxus. Mammalian Species, 26:

1-4.

Marcus, L. 1960. A census of the abundant large Pleistocene
mammals from Rancho La Brea. Los Angeles County Mus.,
Contrib. Sci., 38: 1-11.

Nace, R. L., P. T. Voegeli, J. R. Jones, and M. Deutsch. 1975.
Generalized geologic framework of the National Reactor
Testing Station, Idaho. U.S.G.S. Prof. Paper, 725B: 119-

147.

256

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Odum, E. P. 1968. Energy flow in ecosystems: a historical re-
view. Amer. Zool., 8: 1 1-18.

. 1971. Fundamentals of ecology. Saunders, Philadel-
phia, 3rd ed., 574 pp.

Parmalee, P. W. 1967. A Recent cave bone deposit in south-
western Illinois. Bull. Natl. Speleol. Soc., 29:1 19-147.

Parmalee, P. W„ and R. O. Oesch. 1972. Pleistocene and
Recent faunas from the Brynjulfson Caves, Missouri. Illinois
St. Mus. Rep. Invest., 25:1-52.

Parmalee, P. W., R. O. Oesch, and J. E. Guilday. 1969. Pleis-
tocene and Recent vertebrate faunas from Crankshaft Cave,
Missouri. Illinois St. Mus. Rep. Invest., 14:1-37.

Payne, S. 1972. On the interpretation of bone samples from
archaeological sites. Pp. 65-82, in Papers in economic pre-
history (E. S. Higgs, ed.), Cambridge Univ. Press, Cam-
bridge, 219 pp.

Peterson, O. A. 1923. A fossilbearing slab of sandstone from
the Agate Springs Quarries of western Nebraska exhibited
in the Carnegie Museum. Ann. Carnegie Mus., 15:91-93.

Pielou, E. C. 1966. The measurement of diversity in different
types of biological collections. J. Theor. Biol., 13:131-144.

Quick, H. F. 1951. Notes on the ecology of weasels in Gunnison
County, Colorado. J. Mamm., 32:281-290.

Ricklefs, R. E. 1973. Ecology. Chiron Press, Newton, Mas-
sachusetts, 861 pp.

Skinner, M. F. 1942. The fauna of Papago Springs Cave, Ar-
izona. Bull. Amer. Mus. Nat. Hist., 80:143-220.

Soergel, W. 1940. Die Massenvorkommen des Hohlenbaren.
G. Fisher. Jena, 1 12 pp.

Stock, C. 1972. Rancho La Brea, a record of Pleistocene life
in California. Los Angeles Co. Mus. Sci., Ser. 20 (Paleo. no.
1 1, 7th printing): 1-81.

Voorhies, M. R. 1969. Taphonomy and population dynamics
of an Early Pliocene vertebrate fauna, Knox County, Ne-
braska. Univ. Wyoming Contrib. Geol., Spec. Paper, 1:1-
69.

Walker, E. P. 1975. Mammals of the World. Johns Hopkins
Press, Baltimore, 3rd ed., 1500 pp.

Webster, G. S. 1978. Dry Creek Rockshelter: Cultural Chro-
nology in the Western Snake River Region of Idaho. Tebiwa,
15:1-35.

Wright, P. L. 1948. Breeding habits of captive long-tailed
weasels (Mustela frenata). Amer. Midland Nat., 39:338-344.

Address (White): Idaho Museum of Natural History, Idaho State University, Pocatello, Idaho 83209.

Address (McDonald): Vertebrate Paleontology, Royal Ontario Museum, 100 Queen’s Park, Toronto, On-

tario M5C 2C6, Canada.

Address (Anderson): 730 Magnolia Street, Denver, Colorado 80220.

Address (Soiset): 1284 South Fisher Avenue, Blackfoot, Idaho 83221.

REVIEW OF THE SMALL CARNIVORES OF NORTH AMERICA
DURING THE LAST 3.5 MILLION YEARS

Elaine Anderson

ABSTRACT

During the last 3 . 5 million years 30 genera and about 6 5 species
of small carnivores (mustelids, small canids, procyonids, and
small felids) have inhabited North America. The number of gen-
era has remained constant (18-19) during each land mammal

age, but the number of species has fluctuated between 25 and 36.
At the present time 18 genera and 30 mainland species are rec-
ognized. Extinction patterns between small and large carnivores
are compared.

INTRODUCTION

Smilodon, Arctodus, Canis dims— to most people
those animals are the Pleistocene carnivores. The
smaller species are usually neglected or overlooked,
yet the weasels, martens, badgers, otters, skunks,
foxes, coyotes, raccoons, lynxes, and small cats have
played and continue to play an important role in
the ecosystem. In this paper I will discuss briefly the
small carnivores, those ranging in size from the least
weasel, Mustela rixosa (called nivalis by some work-
ers), the smallest known carnivore weighing only 38
to 63 g to the sea otter, Enhydra lutris that may
reach a weight of 37 kg. The time span covered is
approximately 3.5 million years, the Blancan, Ir-
vingtonian and Rancholabrean land mammal ages
and the Holocene period in North America.

A population explosion of microtine rodents oc-
curred in the early Blancan, and this event can be
correlated with a radiation of small carnivores. They
filled all available niches and many were adapted to
feed on the hordes of rodents. Helping control ro-
dent populations continues to be a major role for
the small carnivores.

John Guilday often kidded me about my carni-
vores eating his mice. He frequently sent me isolated
teeth and fragmentary bones of carnivores from his
Appalachian faunas to identify. In the Baker Bluff
fauna (Guilday et al., 1978) I was delighted to find
fragmentary remains of four pine martens, a fisher,
10 weasels, a badger, and four skunks; he patiently
tabulated hundreds of microtines. At Strait Canyon

Fissure, Virginia (manuscript), I recorded 61 mus-
telids and 30 procyonids; he counted 656 cricetids
and 44 sciurids. This brings out a significant point,
namely carnivores are rare in most Blancan and
Pleistocene faunas. Only at carnivore trap sites
(Rancho La Brea, Moonshiner Cave) where they are
attracted to the site by the scent and struggles of
entrapped animals, are the remains numerous, and
these localities represent deposition over a long pe-
riod of time. At such sites the bones of young and
old animals predominate. At raptor sites (Clark’s
Cave) the relative scarcity of carnivore remains re-
flects the collecting bias of the birds of prey (Guilday
et ah, 1 977). Due to the alertness and agility of most
small carnivores, they seldom fall in sinkholes (New
Paris No. 4). Caves used as dens may contain more
complete material of animals that died there or were
brought in as prey (Little Box Elder Cave). At some
sites, carnivore remains are almost non-existent, for
example, at Hansen Bluff, a new Irvingtonian lo-
cality in southern Colorado, only a canid phalanx
has been found.

Tables 1-3 list Blancan, Irvingtonian, and Ran-
cholabrean species of small carnivores and give their
geologic and geographic ranges, number of reported
faunas, size, habitat, and food. Additional data such
as new faunal records and taxonomic changes are
given in the brief species accounts that follow. For
more detailed information see Kurten and Ander-
son (1980).

SPECIES ACCOUNTS

Family Mustelidae

*Martes diluviana

The Irvingtonian fisher, the earliest known Amer-
ican species of Maries, has been found at Port Ken-
nedy, Cumberland, and Conard Fissure.

Martes pennanti

The extant fisher had a wider range in the late
Rancholabrean than it does today and was found
throughout the Appalachians as well as in the South
and Midwest. It has recently been identified at Strait
Canyon Fissure, Virginia.

257

258

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 1 . — Small carnivores of the Blancan Land Mammal Age in North America, f = extinct genus. * — extinct species. Superscript' =

first appearance of species.

Species

Geologic

range

Geographic range
Blancan

No. B1
faunas

Size

Habitat

Food

Mustelidae

* Mustela rexroadensis

M BP

w u.s.

4

med. weasel

savanna

small rodents

Mustela frenata

L Bl'-R

Kansas

1

85-340 g

savanna

small rodents

*MusteIa meltoni

M BP

SW Kansas

1

mink size

swampy areas

microtines

f Ferinestrix vorax

BP

Idaho

1

>Gulo

savanna

carrion

t Trigonictis macrodon

BP

U.S.

12

ca. 2-5.5 kg

various

small animals

t Trigonictis cookii

BP

W U.S.

5

2/3 size of

various

small rodents

f Sminthosinus howleri

M BP

Idaho, Nebraska

2

T. macrodon
< T. cookii

marshy areas

small microtines

fCammarles cumminsi

L BP

N Texas

1

mod. large

open plains

small animals

Taxidea taxus

BP

W U.S.

7

5.8-11.2 kg

prairies

rodents, rabbits

t Satherium piscinarium

BP

U.S.

5

>Lutra

aquatic

aquatic animals

t Buisnictis breviramus

E'-M B1

W U.S.

5

ca. 85-340 g

swampy areas

omnivorous

f Buisnictis burrowsi

M BP

Nebraska

1

ca. 85-340 g

savanna

omnivorous

*Spilogale rexroadi

B1

W U.S.

3

size of V

savanna

omnivorous

*Spi!ogale{ ?) microdens

E BP

Texas

1

S', pygmaea
454-999 g

savanna

mice, insects

Spilogale putorius

L Bl'-R

Kansas

1

363-567 g

savanna

mice, insects

Mephitis mephitis

M Bl'-R

Nebraska

1

2. 7-6.3 kg

brushy areas

omnivorous

f Brachyopsigale dubius

E BP

Kansas, Texas

2

Mephitis- size

near water

omnivorous

Canidae

*Canis lepophagus

BP-E Irv

U.S.

15

ca. 9-22 kg

various

small-med.

*Urocyon progressus

BP

Kansas, Texas

2

ca. 6 kg

brushy areas

animals
rodents, rabbits

Procyonidae

*Nasua pronarica

E BP

Texas

1

>N. narica

open forests

omnivorous

*Procyon rexroadensis

M BP

Kansas, Texas

2

>P. lot or

swampy areas

omnivorous

*Bassariscus casei

E'-M B1

W U.S.

3

900-1,300 g

savanna

small animals

\Parailurus angelicus

M B1

Washington, Europe

1

>Ailurus

gallery forest

omnivorous

Felidae

*Felis lacustris

BP-E Irv

W U.S.

8

large Lynx size

near water

small-med.

*Felis rexroadensis

E Bl-E Irv

Kansas

1

lynx size

savanna

animals

rodents, rabbits

Lynx rufus

L BP-R

Texas

1

6.7-15.7 kg

wooded areas

rabbits, rodents

*Martes nobilis

Remains of the noble marten have recently been
identified at Little Canyon Creek Cave, Washakie
County, Wyoming 14C dated 10,170 ± 250 B.P.
(Walker, D. P., personal communication) and at two
Holocene archeological sites, Hidden Cave, Chur-
chill County, Nevada, dated 3,700 years B.P. and
Dry Cave, Ada County, Idaho 14C dated 3,270 ±
1 10 B.P. (Grayson, this volume). At Little Canyon
Creek Cave, it was associated with boreal and plains
species, but there are no boreal species at the other
two sites. A more equable climate probably explains
these occurrences.

Martes americana

Pine marten remains are rare in Pleistocene fau-
nas. A late immigrant from Eurasia, it first appears
in North America in the late Rancholabrean. It has
recently been identified at Strait Canyon Fissure.

*Mustela rexroadensis

This medium sized Blancan weasel has been found
at Hagerman, Rexroad, Fox Canyon, and White
Rock. It may have been ancestral to Mustela frenata.

Mustela frenata

First recognized in the late Blancan Borchers fau-

Table 2. — Small carnivores of the In'ingtonian Land Mammal Age in North America, f = extinct genus. * = extinct species. Superscript1 = first appearance

of species.

1984

ANDERSON -SMALL NORTH AMERICAN CARNIVORES

259

ZJ=

S 2
u c
- o

D.

U

<D

o

cd

3
O

I *

£ >

-

cd u

T) cS
*> £

A ° ° -

O -t r- — <

(N — I

2 p
cd i«

<2

'S C

2 |

a .5

0 g

1 ^

> on d

|p

,o >

Cd

Q. "O
O V

E §
8 is

S)

00

M (N ' J

00 00 00
M -x -x

c

JS i I
*-* oo

r-

'O ^ N

V 33 Jv SO fN

rJ, I I

V5 cj fN O

n~)

fN

fN

rn

fN

fN

fN

fN

so

<St

T3

2

G

G

cd

<D

C/3

cd

u.

<

£■

cd

s

<S>

s

2

o

M

cd

a

o

Wh

3

UJ

cd

b

UJ

cd

c

cd

>

(A

t/T

G

cd

Q

6

cd

G

c/5

C/3

2

2

£

C/3

C/3

C/3

<2

O

C/3

(S>

G

b

b

e

c

2

<

3

o

b

b

b

*cd

N

c

b

C

<u

UJ

w

£

X

C/3

w

UJ

UJ

u

<

UJ

Oh

T3 T3

•c -C
o o

E E

Oi ftS

qS j j
- ~ t l

i ffl H M

M J J J

S u

m oq

0 i X & X ~ _i X
Qi J J — r m J

oaii-iiujSi:

>5 3 A

1 I ■§ 1 | I

1 ^ 3 I | §
^ - cq o ^ U

G • G is <st
cd & rz r. rz

_T\ cd S) Cd

■s s 1 1 1

£ 2 § ° §

~ C — .5 —

cs ^

E 8

■8 8
O O
£ £

<zf c/T
<D <L>

•c -c

2 'O

g *8
> §
2 £

<N

fN

I OO
Os fN —

. <N

cd I cd ^
O Os O

| OO Os
10 >ri (N
- I I
(N OO

m >Ti N f9 fS

cd

T3

■C

o

E

b

£

2 c/3
c 5

<L>

Ou, UJ

cd

£

o

sz

cd

2

O

cd"

M

</i

cd

-D

>• i Oh Ot

~ > i OJS

w i & -' C

5 2 w ^ j

cd

T3

D

8

O

£

T3

■c

O

E

Oi

1

_J

*Felis lacustris Bl-E Irv Arizona 1 Lynx-size near water rodents, rabbits

*Felis rexroadensis E Bl-E Irv Nebraska 1 Lynx- size savanna rodents, rabbits

Felis yagouroundi ?Irv-R Pennsylvania 1 6.7-8 kg brushy areas small animals

Lynx issiodorensis E Irv1 Nebraska 1 > L. canadensis savanna med. animals

Lynx rufus L Bl-R U.S. 6 6.7-15.7 kg forests, brush rodents, rabbits

260

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 3.— Small carnivores of the Rancholabrean Land Mammal Age in North America, f = extinct genus. *= extinct species. Superscript^ first
appearance of species. Superscript 2 = survives outside of North American today.

Species

Geologic

range

Geographic range
Rancholabrean

No.

RLB

faunas

Size

Habitat

Food

Mustelidae

Martes pennanti

M RLB'-R

Yukon Terr., E U.S.

i i

2.0-5. 5 kg

mixed hardwoods

small-med.

animals

*Martes nobilis

LRLB'-EH

W U.S.

i i

ca. 0.9-3 kg

forests, brush

small animals

Martes americana

L RLB'-R

Yukon Terr., W U.S.,
Appalachians

l i

681-1,248 g

coniferous

forests

small animals,
fruit

Mustela frenata

LB1-R

S Canada, U.S.

35

85-340 g

various

small animals

Mustela erminea

LIrv-R

Alaska, Yukon Terr.,
U.S. S to Texas

14

28-170 g

tundra, forests

mice

Mustela rixosa

L RLB'-R

U.S.

9

38-63 g

meadows, brush

small mice

Mustela vison

L Irv-R

Alaska, U.S.

20

567-1,362 g

near water

aquatic animals

Mustela eversmanni 2

L RLB1

* ssp. Alaska

1

>M. nigripes

steppes

rodents

Mustela nigripes

RLB'-R

Canada, W U.S.

17

0.681-1.6 kg

prairies

rodents, Cynomys

Gulo gido

RLB'-R

Alaska, Canada, W U.S.

10

16-27 kg

tundra, taiga

carrion,

mammals

Taxidea taxus

Bl-R

W U.S., Alaska, prairie
corridor E U.S.

39

5.8-11.2 kg

prairies

rodents

Lutra canadensis

Irv-R

E U.S.

15

4.5-11.2 kg

inland waterways

aquatic animals

*Enhydra macrodonta

Irv-RLB

California

1

>E. lutris

marine

marine inverte-
brates

En hydra lutris

M RLB'-R

N Pacific coast

3

13.5-37 kg

marine kelpbeds

marine inverte-
brates

Spilogale putorius

L Bl-R

U.S., S Canada

46

363-567 g

prairies, brush

mice, insects, eggs

Mir achy proloma
obtusata

Irv-RLB

E Cent. U.S.

3

small Spilogale

forests

hard-shelled

insects

Mephitis mephitis

M Bl-R

U.S.

48

2. 7-6. 3 kg

semiopen areas

omnivorous

Conepatus leuconotus

Irv-R

U.S., Mexico

10

0.9-2. 7 kg

semiwooded areas

insects

Canidae

Canis latrans

M Irv-R

S Canada, U.S.

60+

9-22 kg

prairies, woods

rabbits, rodents

Canis rufus

E Irv-R

S U.S., Mexico

8

18-31.5 kg

brushy areas

small animals,
crabs

*Canis cedazoensis

M RLB'

Mexico

1

<C. latrans

prairies

small animals,
carrion

Canis familiaris

L RLB'-R

Canada, U.S.

14

2 size classes

various

small animals,
carrion

Cuon a/pinus2

LRLB'

Beringia, Mexico

3

10-20 kg

forest, steppes

med. -large
animals

Urocyon cinereoar-

Irv-R

U.S.

30

3. 2-5. 8 kg

brushy areas

omnivorous

genteus

Vulpes velox

LIrv-R

Great Plains

7

1.8-2. 9 kg

open plains

small animals

Vulpes macrotis

L RLB'-R

California, SW U.S.

1 1

1. 4-2.7 kg

desert scrub

rodents, rabbits

Vulpes vulpes

M RLB'-R

North America

28

4. 5-6. 7 kg

mixed forests

rodents, rabbits

Alopex lagopus

L RLB'-R

Yukon Terr.

1

3.2-6. 7 kg

tundra

carrion

Procyonidae

Procyon lotor

L Irv-R

U.S.

58

5.4-15.8 kg

wooded areas

omnivorous

*Bassariscus

L RLB'

Arizona, Mexico

2

>B. astutus

rocky areas

small animals

sonoitensis

Bassariscus astutus

L RLB'-R

SW U.S., California

22

900-1,130 g

rocky areas

small animals

Felidae

Fehs pardalis

M RLB'-R

Florida

1

9-18 kg

forests

mammals

*Felis amnicola

RLB'

Florida

7

ca. 6.5-8 kg

brushy areas

small animals

Felis yagouroundi

?Irv-R

Texas, Mexico

2

6.7-8. 1 kg

thorn thickets

small animals

Lynx canadensis

M RLB'-R

Alaska, S Canada,
Idaho, Utah

5

6.7-13.5 kg

boreal forests

small animals

Lynx rufus

L Bl-R

U.S.

57

6.7-15.7 kg

forests, brush

rodents, rabbits

1984

ANDERSON -SMALL NORTH AMERICAN CARNIVORES

261

na, the long-tailed weasel has the longest stratigraph-
ic record and the greatest geographic range of any
American weasel. Found in almost all terrestrial
habitats except deserts, Mustela frenata is often the
most abundant small carnivore in a fauna (see White
et al., this volume). New records include Strait Can-
yon Fissure, Upper Sloth Cave, Texas, and Muskox
Cave, New Mexico.

Mustela erminea

The circumboreal ermine reached America in the
late Blancan or early Irvingtonian. Geographical
variation in size and sexual dimorphism are pro-
nounced, and there is size overlap between the larger
M. frenata and the smaller M. rixosa. The ermine
had a wider range in the late Pleistocene and it has
recently been identified at Strait Canyon Fissure and
Smith Creek Cave, Nevada.

Mustela rixosa

Smallest of the Pleistocene and extant carnivores,
the least weasel is probably an evolutionary offshoot
of M. nivalis and reached America in the late Ran-
cholabrean. It has recently been recognized in the
Strait Canyon Fissure fauna.

*Mustela meltoni

This poorly known Blancan species has only been
found at Rexroad.

Mustela vison

Youngman (1982) showed that M. vison and M.
lutreola, the European mink, are not closely related
and placed the American species in a separate
Nearctic, endemic subgenus (Vison) that also in-
cludes the extinct sea mink, M. macrodon. Am-
phibious in habits, the mink is a good indicator of
nearby permanent water. New records of the species
include Smith Creek Cave, Strait Canyon Fissure,
and Christensen Bog, Indiana.

* Mustela macrodon

The largest known American mink, this species
has only been found along the northeast coast from
the Bay of Fundy to Massachusetts where its re-
mains have been recovered from Indian middens.
It probably survived until a few hundred years ago.

Mustela nigripes

Additional late Rancholabrean records of the
black-footed ferret include Little Canyon Creek
Cave, Wyoming, and January Cave, Alberta; both

sites are within the historic range of the species. A
population of this endangered mammal has recently
been discovered in northern Wyoming.

Mustela eversmanni

An extinct subspecies, M. e. beringiae, of the Eur-
asian steppe ferret has been recognized in the Fair-
banks fauna. Apparently it did not extend its range
beyond the unglaciated steppes of Beringia.

t Tisisthenes parvus

This small, poorly known weasel-like mustelid is
known only from the Irvingtonian Java fauna. South
Dakota.

*Gulo schlosseri

Schlosser’s wolverine, the earliest recognized
species of Gulo, has been found in contemporane-
ous-age deposits in eastern North America and cen-
tral Europe. It was ancestral to Gulo gulo.

Gulo gulo

This rare circumboreal species first appears in late
Rancholabrean faunas. Late Pleistocene wolverines
were larger than early Holocene and Recent speci-
mens.

t Ferinestrix vorax

This poorly known Blancan species is known only
from two specimens found at Hagerman. Bjork
(1970) placed it in the subfamily Mellivorinae.

f Trigonictis macrodon

A recent revision of Trigonictis (Ray et al., 1981)
showed that a large galictine mustelid, now called
T. macrodon was widespread in the Blancan. It is
known from Charles County, Maryland (type lo-
cality), Hagerman, Rexroad, White Bluffs, Grand
View, Broadwater-Lisco, Sand Draw, Deer Park,
near Channing, Texas, Vallecito Creek, near Safford
Arizona, Smith Mill River in Wayne County, North
Carolina, and Santa Fe River VIII A. Trigonictis
probably reached North America from Eurasia in
the early Blancan, perhaps somewhat earlier.

f T rigon ictis cooki i

Differing from T. macrodon by only slightly
smaller size, T. cookii has been found at Hagerman,
Grand View, Sand Draw, Broadwater, Red Corral,
and Haile XVIA. In addition, Trigonictis sp. is re-
corded from Cita Canyon, Arroyo Seco, and Inglis
IA. To have two closely related galictine mustelids

262

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

occurring together in at least four faunas has puzzled
all who have studied them. Until more material
becomes available to clarify their relationship, they
are regarded as distinct species occupying slightly
different niches.

t Sminthosinus bowleri

This small galictine mustelid has only been found
at Hagerman and Broadwater. It may only be sub-
specifically distinct from Trigonictis.

fCanimartes cumminsi

This species is only known from one specimen
found in the Blanco fauna. It may not even be a
mustelid, because M2 is apparently present.

Taxidea taxus

The fossorial badger first appears in the Blancan,
and by the late Rancholabrean they were common
in western faunas. Prairie corridors provided an
eastern extension of their range (Baker Bluff, Welsh,
Peccary, Bootlegger Sink), and they also spread
northward to unglaciated eastern Beringia (Fair-
banks, Gold Run Creek, Dominion Creek) where
steppe conditions prevailed in the late Pleistocene.
Badger remains have recently been found at Dutton,
Little Canyon Creek Cave, and Conkling Cavern,
New Mexico.

f Satherium piscinarium

The Blancan otter has recently been recognized
in the Vallecito Creek fauna. It was probably an-
cestral to Pteroneura brasiliensis, the flat-tailed otter
of South America.

Lutra canadensis

The aquatic river otter has recently been recog-
nized at Warm Mineral Springs, Florida, and Strait
Canyon Fissure, Virginia. River otters probably
reached North America from China in the early
Irvingtonian.

*En hydra macrodonta

Specimens from three sites in California and Or-
egon have been referred to this species. It is not
known if it was ancestral to the extant species, En-
hvdra lutris.

Enhydra lutris

In the Pleistocene the range of the sea otter ex-
tended from Point Barrow, Alaska, to southern Cal-
ifornia. Although strictly protected today, they oc-

cupy only about one-fifth of their former range and
are regarded as a threatened species.

t Buisnictis breviramus

This small, short-faced skunk has been found in
early and middle Blancan faunas in Kansas, Texas,
and Idaho. Its relationship to other small skunks is
uncertain for its P4 is Mustela- like but the M1 re-
sembles Mephitis and the M, is four-rooted as in
typical mephitines (Bjork, 1974).

f Buisnictis burrowsi

More advanced than B. breviramus, this small
skunk has only been found at Sand Draw.

*Spilogale rexroadi

Ancestral to later species of Spilogale, the prim-
itive Rexroad skunk has been found at Rexroad,
Blanco, and Beck Ranch.

* Spilogale(‘l) microdens

This skunk is known only from a single well pre-
served mandible found in the early Blancan Beck
Ranch, Texas fauna. As large as a male S. putorius,
it has relatively small teeth. Until upper dentitions
are found, its taxonomic position remains uncer-
tain.

Spilogale putorius

Spotted skunks are common in Rancholabrean
faunas throughout the country. New records include
Monkey Jungle, Florida; Strait Canyon Fissure, Vir-
ginia; Riverside Cave, Tennessee; Howell’s Ridge
and Muskox caves, New Mexico. The eastern and
western populations are now generally regarded as
distinct species— V. putorius in the eastern half of
the county does not have delayed implantation; S.
gracilis the western form has a long (210-260 days)
delayed implantation (Mead, 1 968, 1981). Of course,
these physiological differences are not reflected in
the fossil material, and size and morphological dif-
ferences between the two populations are small. Un-
known is when or why delayed implantation de-
veloped in these skunks.

| Brachyprotoma obtusata

The short-faced skunk has been found at several
sites in east-central United States. Brachyprotoma
is the only genus of small carnivores to have become
extinct in the late Rancholabrean/early Holocene;
why, no one knows.

1984

ANDERSON -SMALL NORTH AMERICAN CARNIVORES

263

f Osmotherium spelaeum

This skunk is only known from the early Irving-
tonian Port Kennedy fauna. It may not be generi-
cally distinct from Mephitis.

Mephitis mephitis

Striped skunks make their first appearance in the
late Blancan Broadwater fauna, and their remains
are common in Rancholabrean faunas. This is partly
explained by their use of caves as dens and by their
propensity for falling down fissures. M. mephitis has
recently been identified in the Strait Canyon Fissure
fauna.

Mephitis macoura

The hooded skunk, primarily a Mexican species
that barely ranges into the United States today, has
not been found in any U.S. Pleistocene fauna.

Conepatus leuconotus

In the Pleistocene the range of the hog-nosed skunk
extended to Florida and Georgia; today its eastern
limit is Texas. It has recently been identified at
Muskox Cave, New Mexico.

f Brachyopsigale dubius

This poorly known skunklike mustelid is known
only from two mandibular fragments recovered from
the early Blancan Fox Canyon and Beck Ranch fau-
nas. Characterized by an extremely short jaw and
crowded teeth, it was about the size of Mephitis.

Family Canidae

* Can is lepophagus

This ancestral coyote was widespread in the Blan-
can, and it survived into the early Irvingtonian when
it was replaced by C. latrans.

Can is latrans

Remains of coyotes have been found at more than
100 Pleistocene localities making it one of the most
common species. It is especially numerous at trap
sites (Rancho La Brea, Moonshiner Cave) because
of its scavenging habits.

*Canis cedozoensis

Smaller than the coyote, this Mexican canid has
counterparts at several sites in the U.S. Their re-
lationships are uncertain.

Canis rufus

Nowak (1979) in his study of Quaternary Canis
referred specimens from the following Pleistocene
sites to C. rufus: Port Kennedy, Inglis IA, Vero,
Melbourne, Crystal River Power Plant, Haile VIIA,
Devil’s Den, Miller’s Cave, Eddy Bluff, Arkansas,
and the Upper Becerra Formation in Mexico, and
from archeological sites in Florida, Alabama, Ar-
kansas, Virginia, West Virginia, Pennsylvania, and
Ohio. Today the endangered red wolf is only found
in parts of southeastern United States.

Canis familiaris

Domestic dog remains are rare in Rancholabrean
faunas and may be difficult to distinguish from those
of coyote or wolf. The specimen from Old Crow
River 1 1A has not been dated yet but is probably
as old or older than the ones from Jaguar Cave 14C
dated 10,370 ± 350 B.P. (Beebe, 1980).

Cuon alpinus

Remains of the Asian dhole with its distinctive
trenchant dentition have been found in Ranchola-
brean faunas in Beringia (Fairbanks, Old Crow Riv-
er 14N) and Mexico (San Josecito). Its New World
extirpation is unexplained.

*Urocyon progress us

The earliest known gray fox is this Blancan species
that has been found at Rexroad, Red Light, and
probably Broadwater. It was ancestral to U. cinere-
oargenteus.

Urocyon cinereoargenteus

This extant species is first recognized in the Ir-
vingtonian, and its remains are quite common in
Rancholabrean faunas. New records include Dut-
ton, Monkey Jungle, Warm Mineral Springs, and
Harrodsburg Crevice, Indiana.

Vulpes velox

Swift and kit foxes have usually been united in a
single species, V. velox. Recent studies (Egoscue,
1979; McGrew, 1979) have demonstrated that these
two small foxes have distinct, non-overlapping (ex-
cept possibly in eastern New Mexico) ranges and
can be separated by several external and cranial
characters. This geographic separation is apparent
in Pleistocene faunas containing swift fox, V. velox
(prairies east of the Rocky Mountains) or kit fox,
V. macrotis (deserts and semideserts of western

264

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

North America). I am tentatively referring speci-
mens from the following sites to V. velox : Cragin
Quarry, Angus, Chimney Rock, Berends, Lubbock
Lake, and Klein. Dalquest (1978) tentatively re-
ferred some small canid remains from Beck Ranch
(early Blancan) to V. velox', if this identification is
correct, it is the earliest known occurrence of the
species.

Vulpes macrotis

The desert-dwelling kit fox differs from V. velox
in having a longer tail, larger ears, and a narrower
skull with a more slender rostrum (McGrew, 1979).
Specimens from the following faunas are tentatively
referred to V. macrotis : Ventana, Burnet, Gray Sand
fauna of Blackwater Draw, Isleta, Dry, Smith Creek,
Schulze, Moonshiner, and Middle Butte caves.

Vulpes vulpes

The Holarctic red fox reached North America from
Eurasia in the Rancholabrean. It has recently been
identified at Natural Trap, Smith Creek Cave, and
Monkey Jungle.

A l ope: c lagopus

Remains of the arctic fox have only been found
at Old Crow River. It apparently did not reach the
Nearctic until the late Rancholabrean.

Family Procyonidae
*Nasua pronarica

A single lower fourth premolar found in the early
Blancan Beck Ranch fauna is the only known spec-
imen of N. pronarica. Because coatimundi remains
have not been identified at other Blancan sites. Beck
Ranch may represent the northern limits of this
South American representative.

Nasua narica

Although the coatimundi is now found in south-
ern Arizona, the extant species is not known from
any Pleistocene locality in the United States. It is
probably descended from N. pronarica.

*Procyon rexroadensis

The Blancan Rexroad raccoon has been found at
Rexroad, Fox Canyon, Cita Canyon, and the Anza
Borrego faunas.

Procyon lotor

Raccoons are common in Pleistocene faunas from
the late Irvingtonian onward. Its adaptability to a

wide variety of habitats and its omnivorous habits
contribute to its success. It has recently been iden-
tified at Monkey Jungle, Warm Mineral Springs,
Rancho La Brea, Strait Canyon Fissure, and Chris-
tensen Bog.

*Bassariscus casei

This Blancan species has been found at Rexroad,
Beck Ranch, and Vallecito Creek. It was ancestral
to B. astutus.

*Bassariscus sonoitensis

The Sonoita ringtail is known only from skull and
postcranial material found in the late Ranchola-
brean faunas of Papago Springs and San Josecito.

Bassariscus astutus

Remains of the extant ringtail have recently been
identified in the following late Rancholabrean
southwestern cave faunas: Vulture, Stanton’s, Pratt,
Upper Sloth, and Muskox.

t Parailurus angelicus

The only record of the English panda in North
America is an isolated M1 found in the early Blancan
Taunton fauna in southwestern Washington.

Family Felidae

*Felis lacustris

This lynx-sized cat was widely distributed in the
Blancan and survived until the early Irvingtonian.

*Felis rexroadensis

About the same size as F. lacustris, this distinct
Blancan species shows affinities to Lynx.

Felis pardalis

The ocelot has only been identified in the Reddick
fauna. Hunted for its fur, it is now classified as an
endangered species.

*Felis amnicola

Remains of several small cats in the margay-jag-
uarundi size range have been referred to this extinct
species found in Rancholabrean faunas in Florida.

Felis yagouroundi

Fossils of the jaguarundi have been found at
Schulze and San Josecito caves, and an early form
may be represented at Port Kennedy.

1984

ANDERSON-SMALL NORTH AMERICAN CARNIVORES

265

Felis wiedii

The Neotropical margay has only been reported
from a middle Holocene fauna in coastal Texas. Its
range extends into southern Texas today.

*Lynx issiodorensis

This species, the earliest known Lynx, was wide-
spread in Old World faunas ranging in age from early
Pliocene through the Villafranchian; it has tenta-
tively been identified from Mullen I, an early Ir-
vingtonian fauna. In appearance it resembles Felis
rather than recent lynxes with its shorter legs and
relatively larger skull. Werdelin (1981) believes it
was ancestral to later species.

Lynx canadensis

An immigrant from Eurasia, this boreal species
is rare in Rancholabrean faunas. Its population fluc-
tuates with that of Lepus americanus.

Lynx rufus

The bobcat is common in Pleistocene faunas
throughout the country. It has recently been iden-
tified at Monkey Jungle, Harrodsburg Crevice,
Conkling Cavern, Muskox Cave, and Strait Canyon
Fissure.

DISCUSSION

The number of genera of small carnivores has
remained constant during the last 3.5 million years
while the number of species increased from 25 in
the Irvingtonian to 36 in the Rancholabrean, mainly
due to new immigrants. The most successful species
have a long stratigraphic range, a wide geographic
range, live in a variety of habitats, utilize a number
of food sources, and are adaptable to changing con-
ditions. These species include Mustela frenaia , Tax-
idea taxus, Spilogale putorius, Mephitis mephitis,
Cams latrans, Procyon lotor, and Lynx rufus. All of
them are autochthonous, native American species.
In general, species having a restrictive range, narrow
habitat requirements, and limited food sources be-
came extinct or are threatened with extinction to-
day.

In contrast to these common species, 25 species
are poorly known— a few fragmentary specimens
found in three or fewer faunas. Whether their scar-
city is due to chance collecting or actual rarity of
the animal is unknown. The rarity of some of the
Blancan and Irvingtonian species makes their taxo-
nomic relationships uncertain— this is especially true
for the Blancan skunks. Hopefully more material
wall be found so that comparisons can be made and
the range of variation and sexual dimorphism as-
certained. This may result in the number of genera
and species being reduced or perhaps increased.

Table 4 shows differences in extinction patterns
between small and large carnivores from the Blan-
can to the present. During that period of time, there
were 30 genera and 65 species of small carnivores
and 16 genera and 27 species of large carnivores

(>40 kg, a mustelid, large canids, ursids, large felids,
and a hyaenid). Of these, 1 1 genera (36%) and 34
species (52%) of the small carnivores became ex-
tinct, whereas 10 genera (62%) and 21 species (74%)
of the large carnivores disappeared. At the present
time 18 genera and 33 species of small carnivores
inhabit North America, whereas only four genera
and six species of large carnivores ( Canis lupus, Ur-
sus americanus, Ursus arctos, Ursus maritimus, Felis
coneolor, and Panthera onca) survive here.

More small carnivores became extinct in the
Blancan than in the following mammal ages. In con-
trast more large carnivores succumbed to Ranchola-
brean extinctions. Brachyprotoma was the only ge-

Table 4.— Comparison of extinction patterns between small and
large carnivores during the last 3.5 million years.

Land mammal
age and size

No.

genera

No.

species

No.

extinct

genera

No.

extinct

species

Blancan

Small carnivores

19

26

8

21

Large carnivores

9

10

3

5

Irvingtonian

Small carnivores

18

26

3

1 1

Large carnivores

1 1

16

3

7

Rancholabrean

Small carnivores

19

36

1

6

Large carnivores

9

13

5

8

Holocene

Small carnivores

18

30

0

2

Large carnivores

4

6

0

0

266

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

nus of small carnivore to vanish at the end of the
Rancholabrean. Disappearance of prey, competi-
tion, climatic change, and man’s activities in the
late Rancholabrean have usually been cited as caus-

es of carnivore extinction, but the small carnivores
were less affected by these factors than the large
ones, so the cause(s) of their extinction remains an
enigma.

ACKNOWLEDGMENTS

My thanks go to the late John Guilday for many stimulating McDonald, Horace Quick, Sam Erlinge, Mikael Sandell, and Car-

discussions on Pleistocene mammals and to Bjorn Kurten for olyn King provided data for this paper,

steering me to the small carnivores. Donald Grayson, Greg

LITERATURE CITED

Beebe, B. 1980. A domestic dog (Canis familiaris L.) of prob-
able Pleistocene age from Old Crow, Yukon Territory, Can-
ada. Canadian J. Archaeol., 4:161-168.

Bjork, P. R. 1970. The Carnivora of the Hagerman local fauna
(late Pliocene) of southwestern Idaho. Trans. Amer. Phil.
Soc., ns, 60(7): 1-54.

. 1974. Additional carnivores from the Rexroad For-
mation (Upper Pliocene) of southwestern Kansas. Trans.
Kansas Acad. Sci., 76:24-38.

Dalquest, W. W. 1978. Early Blancan mammals of the Beck
Ranch local fauna of Texas. J. Mamm., 59:269-298.
Egoscue, H. J. 1979. Vulpes velox. Mammalian Species, 122:
1-5.

Guilday, J. E., H. W. Hamilton, E. Anderson, and P. W.
Parmalee. 1978. The Baker BluffCave deposit, Tennessee,
and the late Pleistocene faunal gradient. Bull. Carnegie Mus.
Nat. Hist., 11:1-67.

Guilday, J. E., P. W. Parmalee, and H. W. Hamilton. 1977.
The Clark’s Cave bone deposit and the late Pleistocene pa-
leoecology of the central Appalachian Mountains of Virginia.
Bull. Carnegie Mus. Nat. Hist., 2:1-88.

Kurten, B., and E. Anderson. 1980. Pleistocene mammals of
North America. Columbia Univ. Press, New York, 442 pp.

McGrew, J. C. 1979. Vulpes macrotis. Mammalian Species,
123:1-6.

Mead, R. A. 1968. Reproduction in western forms of the spot-
ted skunk (genus Spilogale). J. Mamm., 49:373-390.

. 1981. Delayed implantation in mustelids, with special

emphasis on the spotted skunk. J. Reprod. Fert., Suppl., 29:
11-24.

Nowak, R. M. 1979. North American Quaternary Canis.
Monogr. Mus. Nat. Hist., Univ. Kansas, 6:1-154.

Ray, C. E., E. Anderson, and S. D. Webb. 1981. The Blancan
carnivore Trigonictis (Mammalia, Mustelidae) in the eastern
United States. Brimleyana, 5:1-36.

Werdelin, L. 1981. The evolution of lynxes. Ann. Zool. Fen-
nici, 18:37-71.

Youngman, P. M. 1982. Distribution and systematics of the
European mink ( Mustela lutreola Linnaeus, 1761). Acta Zool.
Fennica, 166:1-48.

Address: 730 Magnolia Street, Denver, Colorado 80220.

THE PLEISTOCENE DUNG BLANKET OF BECHAN CAVE, UTAH

Owen K. Davis, Larry Agenbroad, Paul S. Martin, and

Jim I. Mead

ABSTRACT

Boluses of dung rich in graminoid stems dropped by a large
herbivore and comparable in size and content to African elephant
(Loxodonta) dung were discovered recently in Bechan Cave,
southeastern Utah. Two boluses were radiocarbon dated at 1 1 ,670
and 12,900 yr B.P., respectively. They are embedded in a 255
m3 dung blanket along with fecal remains and hair of ground
sloths, artiodactyls, and small mammals. While no diagnostic
bones of mammoth were recovered, the deposit yielded long
coarse hair attributed to mammoth.

This unusual discovery supports the previous report by Hansen
(in Jennings, 1980) of mammoth dung at Cowboy Cave, Utah.

Both deposits constitute major finds for paleoecologists and are
the most impressive discovery of their kind since the discovery
in the 1930s of dry caves yielding ground sloth dung in Nevada,
Arizona, New Mexico, and west Texas. The Bechan Cave deposit
accumulated during an interval of vegetation change when blue
spruce and water birch declined, oak became more abundant,
and the regional vegetation was sagebrush steppe. Blue spruce
and water birch currently occupy higher elevations in southern
Utah. Megaherbivore occupation of Bechan Cave evidently end-
ed hundreds of years before mammoth and ground sloth (Noth-
rotheriops ) extinction in the region, ca. 1 1,000 yr B.P.

INTRODUCTION

In November of 1982 a National Park Service
team undertaking a grazing survey in Glen Canyon
National Recreation Area entered a large dry cavern
in a narrow valley north of the Colorado River. The
west facing cave (Fig. 1 ), unusually large for caverns
in the Navajo Sandstone, lies above a relatively steep
slope. Unlike certain other shelters in the vicinity
it contained no surface deposit of cow manure or
other evidence of feral cattle.

Attention of the exploring party, which included
Larry Belli and Charles Berg of the National Park
Service and NPS consultants Steve Carruthers and
Lauren Haury of Flagstaff, was drawn to dry dung
fragments exposed in the walls of shallow holes dug
by pot hunters in the sandy floor of the cave. Aware
of the interest in fossil ground sloth coprolites found
in other caves in the region, Carruthers forwarded
a sample to Tucson via Dr. Art Phillips of the Mu-
seum of Northern Arizona. Inspection revealed that
the fragment was rich in graminoid culms, rather
than in browse fragments as is typical of Shasta
ground sloth dung. In shape and texture the sample
was unlike the numerous samples of Shasta ground
sloth dung in the University of Arizona Paleoen-
vironmental Laboratory collection.

Finely chewed ruminant dung could be eliminat-
ed from comparison. While of a coarser texture and
dominated by fragments 0. 3-2.0 mm in length
(Spaulding and Martin, 1979), equid droppings lack
large grass culms. The unknown sample resembled

droppings of Loxodonta (African elephant). A visit
to the cave was planned.

In February 1983 guided by Belli and other NPS
personnel, accompanied by Utah State Archaeolo-
gist Dave Madsen and Utah Paleontologist Jim
Madsen, Jim and Emilee Mead and Larry Agen-
broad made a brief reconnaissance. The cave, des-
ignated “Bechan Cave” from the Navajo word for
excrement, is roughly 173 ft (52.8 m) deep, 103 ft
(3 1 .4 m) wide, up to 30 ft (9. 1 m) in height and well
lit during the daytime through its cavernous mouth.
The dung blanket proved extensive. To determine
its thickness a small pit was dug close to the north
wall of the cave (see Fig. 2). The contents of the pit
(“pit M”) included two large boluses of unbroken
dung. The larger, designated M-l, measures 230 by
170 by 85 mm (Fig. 3); the smaller, M-2, is sub-
spherical and 225 m in diameter (Fig. 4). Both closely
resemble boluses of African elephant dung (Fig. 5).
They lack twigs typical of dung balls of the Shasta
ground sloth (Fig. 6). The M-l and M-2 dung balls
were adjoining in a small pit and their pollen con-
tents (Table 2) are very similar. While their carbon
fourteen ages (Table 1) are significantly different
(<0.01 ), their carbon thirteen values are identical
(Table 1) and we infer that they were dropped by
the same animal.

These remarkable specimens may not be the first
record of fossil proboscidean dung found in Utah.
A similar dung blanket was found in Cowboy Cave

267

268

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 1. — Mouth of Bechan Cave within Navajo Sandstone. Utah juniper, Indian rice grass, and various shrubs grow on sandy soils
within the canyon.

immediately west of Canyonlands National Park
(Jennings, 1980). Fecal remains from a variety of
herbivores, living and extinct and including ele-
phant dung fragments were reported by Hansen
(1980) from Cowboy Cave.

Subsequent excavations and screening of test pits
at Bechan Cave have yielded fragments but no ad-
ditional intact boluses of the quality of M-l and M-
2. Mead and Agenbroad recovered fecal pellets com-
parable to elk or Harrington’s mountain goat, in-
cluding some pellets 26 by 19 by 18 mm in size
which resemble unknown herbivore droppings (“elk-
camel”) reported in Hansen (1980:1 82). In addition
they found twigs and other plant macrofossils ap-
parently transported by Neotoma and a strand of

coarse hair 1 50 mm long, and 1.15 mm in diameter,
too coarse to be attributed to living members of the
Utah megafauna (that is, elk, pronghorn, mountain
sheep, mule deer, black bear, wolf, and mountain
lion).

In view of the rich finds made in sample pit “M”
more extensive work in the cave was proposed and
approved by National Park Service managers. A
protective policy was adopted which we seek to fol-
low in not identifying the site by its map coordi-
nates. Such information can be requested from the
Glen Canyon National Recreation Area, Page, Ar-
izona. Despite these efforts unauthorized digging
within the cave continues.

In March 1983, Agenbroad and Mead returned

1984

DAVIS ET AL.-BECHAN CAVE DUNG BLANKET

269

Fig. 2. — Excavation of sample pit “M” by Agenbroad and Mead (kneeling), February 1983.

for five days with two Northern Arizona University
students, Richard Ryan and Deborah Meier to be
joined by P. S. Martin and O. K. Davis. The cave
floor was mapped (Fig. 7), “Test Pit 1“ was dug
adjacent to sample pit “M,” contents of the test pit
were screened with window screen (2 mm mesh),
and samples were removed for radiocarbon dating
and paleobotanical study. At 35 cm Test Pit 1 yield-
ed a ground sloth dung ball packed with twigs, in-
cluding an acorn of Quercus gambelii. Davis and
Martin began a plant survey outside the cave while
Mead and Agenbroad undertook geological recon-
naissance of the alluvial fill exposed in deep cuts
within the canyon.

In May 1983 Agenbroad, Davis, and Martin ac-
companied by Brett and Finn Agenbroad and Mary

Kay O’Rourke, returned to Bechan Cave for four
days of field work. Geological and botanical recon-
naissance continued. Thickness of overburden and
thickness of the dung blanket itself was measured
in a series of 49 auger holes positioned on a 5-m
grid, yielding an isopatch map (Fig. 8). Test Pit 2
(position shown on Fig. 7) was dug to sterile sand
near the back of the cave and its contents screened.
The trampled edge of the dung blanket at the mouth
of the cave was faced, photographed, and back-filled.
Future visitors will be able to examine details of
dung blanket texture and stratigraphy by removing
the backfill rather than digging within the cave. An
auger hole to 5 m near the cave mouth (“N” on Fig.
7) indicated dry sterile sand becoming wet at 1.5 m
beneath the dung blanket.

270

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 3. — Bolus M-l from sample pit “M”; plant stems are of graminoids.

CAVE CONTENTS

Based on auger probes the dung blanket of Bechan
Cave is estimated to contain 370 m3 of organic ma-
terials. Overburden near the cave walls constitutes
1 0-20 cm of loose sand thickening to 90 cm or more
toward the center of the cave and exceeding that
near the mouth where dislodged sandstone blocks
prevented augering. In 2.1 m3 of screened test pit
material (less than one percent of the deposit) and

in auger probes no diagnostic bones of large mam-
mals were found. Large bone fragments within the
dung blanket near the mouth of the cave were not
suitable for taxonomic purposes. Occasional jaws
and other small mammal bones were collected.
Packrats use and evidently have used the cave for
12,000 years at least. A packrat midden near the
back of the cave yielded archeological corn cobs.

1984

DAVIS ET AL.-BECHAN CAVE DUNG BLANKET

271

Fig. 4. — Bolus M-2 from sample pit “M”; plant stems are of graminoids.

Evidently Paiute or Navajo placed a brush shelter
of juniper branches against the north wall. Corn cobs
lay near rock cysts at the mouth of the cave; other
contents, if any, had been plundered before our vis-
its. No pottery and only a few stone artifacts in-
cluding metates were found; layers of charcoal 10
cm thick and plant debris in sand above the dung
blanket indicate Elolocene occupation. It is possible
that a rich Archaic record comparable to that at
Cowboy Cave lies beneath the ceiling collapse near

the mouth of the cave. Filled voids (krotovinas) dug
by rat-size rodents were examined in the cleared
exposure at the mouth of the cave. Charcoal was
abundant in the krotovinas but we encountered no
in situ charcoal and no artifacts in the dung blanket
itself.

In the absence of diagnostic bones we sought to
identify the ancient occupants by two means — com-
parison with fecal samples of living and extinct
megaherbivores, and by identification of hair. Sev-

272

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 5 .—Loxodonta (African elephant) dung from Amboseli, Kenya.

Table 1 .—Bechan Cave radiocarbon dates on megaherbivore dung from dung blanket.

Sample

no.

Age

513C

Comment

A-3212

1 1,670 ± 300

-23.2%

Sample pit “M,” bolus M-l of mammoth(?) dung, coll. Mead and
Agenbroad, February 1983

A-3213

12,900 ± 160

-23.2%

Sample pit "M,” bolus M-2 of mammoth(?) dung, coll. Mead and
Agenbroad, February 1983

A-3298

12,620 ± 220

-18.4%

Test Pit 1, 10N 2E; fragment of bolus of mammothf?) dung from
base of unit; M-4 coll. Mead, March 1983

A-3297

12,400 ± 250

-21.9%

Test Pit 1, 1 IN IE, top of dung unit, loose material, coll. Mead,
March 1983

A-3296

1 1,850 ± 160

-25.7%

Test Pit 1, 1 IN IE, middle to base of dung unit, fragment of mam-
moth(?) dung, M-3 coll. Mead, March 1983

GX-9371

13,505 ± 580

Sample pit “M,” mixed dung blanket material, coll. Agenbroad,
February 1983

1984

DAVIS ET AL.-BECHAN CAVE DUNG BLANKET

273

Fig. 6. —Nothrotheriops (Shasta ground sloth) dung from Rampart Cave, Arizona; plant stems are of various shrubs, not graminoids.

era! dozen individual hairs were removed during
screening of the dung blanket for macrofossils, and
occasional hairs or hair masses were found during
facing of the dung blanket at the mouth of the cave.
Preliminary study of the collection by Charles Bo-
len, a graduate student in Anthropology at Northern
Arizona University, has revealed the presence of
hair comparable in thickness to mammoth, hair of
ground sloths (cf. Glossotherium and Nothrother-
iops), several artiodactyls, Equus, and various small
mammals.

Four radiocarbon dates, in addition to samples

from M-i and M-2, were obtained: one (GX-9371)
from the sample pit “M” and the other three from
Test Pit 1 (see Table 1). A-3297 was collected from
the top of the dung blanket stratigraphically above
A-3296. A-3298, a fragment of cf. mammoth dung,
was collected at the base of the dung blanket on
another comer of Test Pit 1. All three dates fall
within the range established by M-l and M-2 (A-
32 1 2 and A-32 1 3). The sample of dung blanket ma-
terial collected by Agenbroad from sample pit “M”
overlaps within one sigma the age of A-32 13.

Thus the radiocarbon samples can be viewed as

274

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 2.

— Concentration of identifiable

remains in Bechan dung layer (1 IN, IE, NE corner).

Taxa of plants

Plant part
identified

Depth in dung

0-10

10-20

20-30

30-40

Betulaceae (361)

Betula occidentalis

bracts

9

6

12

8

nutlets

7

2

3

3

twigs

55

122

83

51

Cactaceae (519)

Opuntia sp.

spines

162

148

109

96

epidermis

1

O. polyacantha

seed

1

Sclerocactus sp.

seeds

2

Caprifoliaceae (76)

Sambucus sp.

twigs

30

7

3

24

Syrnphoricarpos sp.

seeds

2

3

2

twigs

1

3

1

Chenopodiaceae ( 1 54)

Atriplex sp.

twigs

38

55

34

7

Atriplex cf. canescens

fruits

5

2

leaf

1

A. cf. confertifolia

fruits

2

Chenopodium sp.

seeds

1

2

1

Corispermum sp.

seeds

2

1

3

Cleomaceae (5)

Cleome sp.

seeds

1

2

2

Compositae (97)

Artemisia cf. tridentata

leaves

3

2

1

3

twigs

30

25

16

10

Cirsium sp.

achene

1

1

2

2

cf Heleanthus

achene

1

Comaceae ( 1 3)

Cornus stolonifera

stones

4

4

2

3

Cupressaceae (4)

Juniperus sp.

male bract

1

Juniperus communis

seeds

3

Cyperaceae (707)

Carex cf. lenticularis

nutlets

75

16

28

14

Carex cf interior

nutlets

210

52

44

38

Carex lasiocarpa

nutlets

99

33

41

54

Scirpus sp.

nutlets

3

Fagaceae ( 1 2)

Quercus sp.

wood

1

8

3

Gramineae (204)

cf Agropyron

florets

3

8

12

5

Oryzopsis hymenoides

caryopses

53

37

49

28

Panicum

caryopses

2

1

Spartina gracilis

spikelet

1

1

1

2

Sporobolis

spikelet

1

Gramineae + Cyperaceae (708)

culms

127

406

109

66

Labiatae ( 1 )

cf Marrubium

seed

1

Leguminosae (8)

cf Dalea

seeds

4

2

2

1984

DAVIS ET AL.-BECHAN CAVE DUNG BLANKET

275

Table 2. — Continued.

Taxa of plants

Plant part
identified

Depth in dung

0-10

10-20

20-30

30-40

Liliaceae ( 1 )

cf. Smilacina

seed

1

Malvaceae (2)

Sphaeralcea sp.

seeds

2

Pinaceae (59)

Picea pungens

needles

1

16

22

14

twigs

2

1

2

1

Potamogetonaceae ( 1 )

Potamogeton diversifoiium

nutlet

1

Rosaceae (342)

Amelanchier sp.

achenes

1

2

1

Purshia sp.

leaf

1

Rosa sp.

thorns

33

20

31

25

twigs

45

62

53

26

achenes

16

7

8

10

Rubus sp.

achene

1

Saxifragaceae (3)

Ribes sp. seeds 3

Total (3,277)

Total Graminoids (1,619) = 49.4%

ranging from 1 1,700 to 12,900 years in age and the
dung blanket appears to have been deposited, per-
haps intermittently, over roughly 1,000 years. Var-
ious authors (Long and Martin, 1974; Thompson et
al., 1980) have stressed the 9th millennium B.C. as
the critical time for extinction of Shasta ground sloths
and other animals in this region. On the basis of the
six radiocarbon dates in Table 1 it would appear
that Bechan Cave was not occupied by Pleistocene
herbivores after 1 1,700 years ago, which is slightly

but significantly older than the time of extinction of
Shasta ground sloths in western Arizona as deter-
mined by Long and Martin (1974).

Abandonment can be explained in at least two
ways, either temporary access to the cave ended with
erosion of a dune deposit, left as a stabilized rem-
nant 100 m southwest of the cave, or the vegetation
adjacent to the cave may have been attractive to the
megafauna for a brief interval only.

GEOLOGIC SETTING

Bechan Cave occupies a canyon cut into the Na-
vajo Sandstone and, in places, the upper units of
the Kayenta Formation. The cave itself appears to
have been formed by the spalling of the cross bedded
sandstone, perhaps aided by moisture infiltrating
from joint controlled surface drainage, spray from
an adjacent plunge pool, and mechanical forces
caused by crystal growth of calcite, due to evapo-
rating ground water.

Entrance to the cave is steep and hazardous. There

is geomorphic suggestion of a ramp in the past, with
access provided by a climbing dune. A remnant of
this dune still exists, although much of it has been
removed by the plunge pool and surface drainage.
The stabilized climbing and falling dunes within the
canyons testify to abundant eolian deposition in the
past. Terrace remnants indicate a former valley fill
of wind blown sand and silt.

A side valley below Bechan Cave contains up to
40 m of alluvium. One measured section indicates

276

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

black cienega soils and a water table fluctuating dur-
ing the accumulation of the alluvium. The valley fill
has been dissected subsequently and the main stem
head cut has migrated up gradient to the terminus

of the box canyons. For large quadrupeds seeking
access to the valley harboring Bechan Cave there is
only one entry point— across steeply sloping bare
sandstone.

MODERN VEGETATION

The vegetation near Bechan Cave occupies a va-
riety of habitat types ranging from bare walls and
ridge tops of Navajo sandstone vegetated only along
cracks or in pockets to rich riparian growth along a
semipermanent stream. The region is also noted for
its hanging “gardens” (Welsh and Taft, 1981), face
and foot wall gardens and plunge basins fed by seep-
age that may harbor primose ( Primula specula ),
maiden-hair fern ( Adiantum ), rock spirea ( Petro -
phytum), cardinal flower ( Lobelia ), death camas
( Zigadenus ), orchids ( Epipactis ), monkey flowers
( Mimulus ), columbine ( Aquilegia ), and other aquat-
ic or mesic species unusual in this arid region (mean
annual ppt. at Page, Arizona = 14.6 cm). Above the

canyon bottoms waterpockets (deep plunge pools)
supporting willows ( Salix exigua) maintain a per-
manent supply during the summer season. The ri-
parian habitat near Bechan Cave includes Populus
fremontii (Fremont cottonwood), Rhamnus betuli-
folia (birch leaf buckthorn), Carex spp. (sedge), poi-
son ivy (Toxicodendron rydbergii), and other aquat-
ic or mesic species. Fraxinus anomala (single leaf
ash) occupies cracks in the bedrock.

Scattered trees along the canyon bottom include
Juniperus osteosperma (Utah juniper), Quercus
gambelii (Gambel oak), and Amelanchier utahensis
(Utah service berry). Shrubs include Coleogyne ra-
mosissima (Black brush), Ephedra spp. (Mormon

1984

DAVIS ET AL.-BECHAN CAVE DUNG BLANKET

277

CONTOURS OF EQUAL THICKNESS
OF DUNG LAYER C.I = IOcm

tea), Atriplex canescens (four-winged salt bush),
Grayia spinosa (spiny hop sage), and Opuntia
phaeacantha major (prickly pear). Common grasses
include Oryzopsis hymenoides (Indian rice grass),
Stipa comata (needle grass), and Sporobolus sp.

(sacaton). Above the canyon and away from bare
cliffs the vegetation is dominated by Coleogyne ra-
mosissima, Juniperus osteosperma, and Oryzopsis
hymenoides.

PALEOECOLOGICAL ANALYSIS

Pollen was extracted from samples collected from
sample pit “M” by Mead and Agenbroad in Feb-
ruary (Fig. 1) and from samples collected from the
east wall of pit “1” excavated on 17 March 1983.
The first set of samples was from within the dung
layer and included samples from the two large dung
boluses (M-l and M-2). The second set included
samples taken at 2 cm intervals from the sand above

and below the dung layer, which was at 23 to 39
cm.

Pollen samples were soaked in dilute HC1 and
Lycopodium tracer tablets were added to allow cal-
culation of pollen concentrations. The samples were
then screened to remove coarse debris and trans-
ferred to test tubes. After treatment with concen-
trated HC1 and HF, the samples were acetylized

278

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

(Faegri and Iversen, 1975), treated with 10% KOH,
and transferred to silicone oil. Over 500 pollen grains
were counted in each sample at 400 x and 1 ,000 x .
Identification of the pollen types is based on the
reference collection and library of the Paleoenvi-
ronmental Laboratory at the University of Arizona.
Preservation was excellent in all samples.

Bulk samples of dung for macrofossil analysis were
taken from a 10 by 10 cm column in the northeast
comer (11N-1E) of the test pit excavated on 17
March 1983. The dung layer was 40 cm thick at that
locality and the four samples, each 1 0 cm thick, were
approximately 1,000 cc in volume.

Macrofossil samples were soaked in water for 24
hrs, then screened through #6 (3.35 mm) and #20
(0.85 mm) ASTM screens. After the residue had
dried, the identifiable remains were removed for
examination. Identifications are based on the mac-
rofossil collections at the Paleoenvironmental Lab-
oratory and the University of Arizona Herbarium.
All spruce ( Picea ) needles were either thin-sectioned
or examined in cross-section to determine the size
and placement of the resin canals. Twigs that lacked
external identifying characters (thorns, buds, leaf-
scars, distinctive bark) were examined in cross-sec-
tion.

RESULTS AND INTERPRETATIONS

The dung samples are composed primarily of what
appear to be crushed culms and leaves of graminoids
(Gramineae and Cyperaceae), and most identifiable
macrofossils represent these two families (Table 2).
Apart from cactus spines, nutlets of sedges are the
most abundant type of fossil in the dung layer. The
concentration of graminoids is greatest in the up-
permost samples, which is true for the concentra-
tions of most other macrofossil types.

The relationship of the fossils in the column to
the diet of any one organism is difficult to establish
because the samples are probably from the dung of
more than one animal. Four different types of dung
were found in the pit. Neotoma (packrat) fecal pel-
lets are present in the dung layer and this animal
favors cactus pads in house construction. Abundant
spines of Opuntia (prickly pear) in the dung layer
may result primarily from the activities of Neotoma.
However, packrats do not gather grass culms in
quantity. As these are evident in the two large dung
balls (Fig. 2, 3), the organism that deposited them,
cf. Mammuthus, may be responsible for the prev-
alence of graminoids in the identifiable remains.

With a few notable exceptions, the macrofossil
composition in samples from different depths is very
similar. The lack of substantial differences among
samples may result from the short period of accu-
mulation of the dung layer, which is suggested by
the narrow range of radiocarbon dates (Table 1).
Lack of a clearcut stratigraphic sequence within the
millennium embraced by the radiocarbon dates
could reflect trampling and disturbance of the dung
blanket by those mammals that deposited it.

Although specific dietary information cannot be
gained from the column macrofossils, the data pro-

vide evidence of the nature of local vegetation. The
fossil assemblage includes plants from several types
of habitats. As is true today, the riparian community
was then an important vegetation type in the narrow
canyon. However, its members were different from
those of today. Three species of sedges were iden-
tified based on their perigynia and nutlets (Fig. 9).
The southern limits of two of these, Carex lasio-
carpa and C. cf. lenticularis, are currently in Idaho
and Montana. The ancient community contained
water birch ( Betu/a occidentalis ) which at the lati-
tude of the cave occurs today only at higher eleva-
tions.

Several of the shrub species found in the dung
layer were probably more abundant in the riparian
community than in the uplands— Sambucus (elder-
berry), Symphoricarpos (snowberry), Rosa (wild
rose), Rubus (raspberry), and Ribes (currant). Given
the small amounts of Picea pollen in the dung sam-
ples (Table 3) and spruce’s tendency to reach its
lowest elevational limits along stream sides, we sus-
pect that spruce too may have favored the riparian
community. We have collected blue spruce ( Picea
pungens) in stream bottoms in the adjacent Henry
Mountains about 1,890 m where it associates with
Betula and is parasitized by nymphs of the gall aphid,
Chernies coo/eyi Gill. The distinctive cone galls
caused by this parasite were found in the dung blan-
ket (Fig. 9). Juniperus communis, present in the up-
permost fossil sample, is also presently restricted to
higher elevations.

Dune vegetation represents another type of hab-
itat reflected in the fossil assemblage. Several species
characteristic of arenaceous habitats today were re-
covered in the ancient dung layer. Fossil members

1984

DAVIS ET AL.-BECHAN CAVE DUNG BLANKET

279

Fig. 9. — Selected plant macrofossils from Bechan Cave. (A) Carex cf. lenticularis modem achene and perigynium (left), fossil achene
and perigynium (right). (B) Carex cf. interior fossil perigynium front and back (left), fossil achene (right). (C) Carex lasiocarpa modem
perigynium (left), fossil perigynium (right). (D) Potamogeton cf. diversifolium fossil. (E) Panicum fossil. (F) Cleome fossil. (G) Sclerocactus
fossil. (H) Betula occidentalis modem bract (left), achene (right). (I) Betula occidentalis fossil bract (left), achene (right). (J) Spartina
fossil. (K) cf. Agropyron fossil. (L) Oryzopsis fossil. (M) Corispermum (fossil). (N) Opuntia polyacantha fossil. (O) Cornus stolonifera
fossils. (P) cf. Smilacina fossil. (Q) cf. Dalea fossils. (R) Amelanchier fossil. (S) Svmphoricarpos fossil. (T) cf. Atriplex fossil wood. (U)
Rosa fossil achenes and thorns. (V) Picea pungens fossil needles and twig. (W) Artemisia modem A. bigelovii (top), fossil A. tridentata
(middle two pieces), modem A. tridentata (bottom). (X) Opuntia fossil spines. (Y) Spruce cone galls modem from Picea pungens (left),
fossil (right). Photos are shown at three different magnifications, note length bar in three different groupings (A-G, FI-X, and Y) separated
by solid lines.

280

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 3.— Pollen percentages for more common pollen types in Bechan Test Pit 1, 11N-1E.

Pollen types

Sample numbers

0-2

10-12

20-22

Dung layer
top bottom

40-42

78-80

Dung balls
Ml M2

Deteriorated

1.4

5.4

8.3

10.4

12.9

13.7

26.5

7.8

8.5

Unknown

2

.8

.4

.4

1.2

0

.3

.2

0

Juniperus

41.6

12.1

13

11.8

8.5

4.2

.9

.4

1

Picea

.4

0

.2

.2

2.2

8.9

.6

0

0

Pinus

17.3

22.9

33.7

6.1

14.1

27.6

29.8

4

2.2

Ostrya

0

0

0

.8

.4

10.3

1.8

.2

.2

Quercus

5.2

7

6.1

6.3

1.8

0

0

0

.2

Ephedra

2.6

2.6

0

0

0

Chenopodiaceae-Amaran.

7.2

8.2

4.5

5.7

15.4

9.3

8.1

33.2

39.4

Artemisia

1 1

18.5

21.5

12.4

30.9

19.3

18.7

20.8

21.4

Ambrosia

3.8

8.7

9.5

7.9

1.4

1

2.1

11.2

7.5

Cirsium

0

.4

.2

.2

.4

1

1.5

0

0

Other compositae

3.4

10.3

2

.4

1.8

1.8

6.9

6

6

Gramineae

1.2

.8

.4

34.1

7.5

2.4

1.5

12.6

9.2

Sphaeralcea

0

0

0

2.2

.2

0

0

1.2

1.5

Umbelliferae

0

0

0

.2

0

0

0

2.2

2.5

Acer negundo

0

0

0

.4

0

0

0

0

0

Betula

0

0

0

0

8.3

3

1.5

.8

.2

Cyperaceae

0

0

0

.4

2.8

1.2

0

.6

.7

Typha

0

0

0

6.7

0

0

0

11.8

11

Sporormiella

.2

0

0

0

0

0

0

16

16.2

Other fungal

1

1

.2

38.4

8.1

6

.6

5.8

6.2

Tracers added ( x l ,000)

56.6

56.6

56.6

90.6

90.6

56.6

56.6

90.6

90.6

Tracers recovered

20

9

7

47.5

4.5

134

621

21.6

23.7

Vol. sedim. used

15

15

15

.5

2

15

15

1.4

1

Cone. ( x 1,000)

94.029

211.051

273.606

484.9

20

14.175

2.019

3

1.6

Pollen sum

498

503

507

508

495

503

332

500

401

of this community include Opuntia polyacantha (a
different species occurs in the modem vegetation),
Sclerocactus sp., Atrip/ex canescens, and Oryzopsis
hymenoides.

The natural upland vegetation growing on Navajo
Sandstone during the time of dung accumulation is
difficult to reconstruct from the macrofossil re-
mains. Today the regional vegetation at the eleva-
tion of the cave is black brush steppe. The relatively
high percentages (30.9%) of Artemisia pollen in the
dung layer (Table 3) may indicate that sagebrush
steppe was the regional vegetation type while the
dung accumulated. Artemisia cf. tridentata macro-
fossils are present in the dung layer; the species is
not now found near Bechan Cave.

Although the dung layer may have experienced
some mixing by trampling, some trends in the mac-
rofossils (Table 2) deserve mention because they
may indicate vegetation change during the period
of dung accumulation. The abundance of aquatic
plants at the top of the dung is greater than in lower
samples. The three-fold increase in sedge achenes

in the 0-10 cm sample is far greater than the increase
in concentration of any other type, and this sample
contains two other aquatic taxa— Scirpus sp. and
Potamogeton diversifolium that do not occur in low-
er samples. Other taxa ( Symphoricarpos , Quercus,
and Picea pungens) show the opposite trend— being
more abundant in the three lower samples. Betula
occidenta/is (Fig. 9) is most abundant in the two
middle samples.

Preliminary analysis of pollen from the dung unit
(‘Top’ and ‘Bottom’ samples in Table 3) only par-
tially confirms this trend. The percentage of Picea
pollen is greater in the sample from the bottom of
the dung, as is the abundance of Betula pollen. How-
ever, Quercus pollen is more abundant in the top
sample, and Symphoricarpos pollen was not re-
covered.

Although macrofossils and pollen from the dung
layer provide limited evidence for vegetation change,
preliminary analysis of pollen in cave fill above and
below the dung layer (Table 3) suggests strong tem-
poral differences. The sand surrounding the dung

1984

DAVIS ET AL.-BECHAN CAVE DUNG BLANKET

281

should be free from the effects of trampling by large
mammals. Differences in Picea pollen percentages
above and below the dung layer are in the same
direction as the macrofossil concentrations and dung
pollen percentages. Immediately below the dung
layer Picea pollen reaches 8.9% and above the layer
the percentages drop to 0.2% (Table 3).

Several other taxa show changes with hop horn-
beam (Osirya), box elder ( Acer negundo), Gramine-
ae, and Cyperaceae all more abundant below the
layer than above. Juniperus, Quercus, and Ambrosia
showed the opposite trend and are more abundant
above the dung layer than below (Table 3). Although
the pollen data are preliminary they are sufficient
to indicate that the dung layer accumulated during
an interval of major vegetation change when spruce

and birch became less abundant and oak became
more abundant.

One last fossil is worthy of note: Sporormiella
spores were common only in the two dung balls,
M-l and M-2 (Table 3). Sporormiella is a dung fun-
gus whose abundance in certain lake sediments has
been correlated with disturbance by grazing animals
(Davis et al., 1977). Its restricted distribution in the
dung layer reflects its poor dispersal, and its ap-
pearance in the surface sample probably results from
the introduction of livestock by European immi-
grants. The abundance of Sporormiella spores in the
two dung balls is of considerable interest. It suggests
that an abundance of the spore type in ancient lake
sediments may be used to indicate the presence of
an abundant megafauna in the surrounding range.

DISCUSSION AND SUMMARY

Around the 10th millennium B.C. and within a
few hundred years of the last records of Shasta ground
sloths and Columbian mammoths in the region, sev-
eral species of large extinct herbivores occupied Be-
chan Cave in southern Utah. They left a distinct
stratigraphic unit, a dung blanket containing abun-
dant plant material and animal droppings up to 40
cm in thickness and estimated at 255 m3 in volume.
No diagnostic bones of the large mammals have
been found to date; however, based on the prelim-
inary examination of dry boluses of herbivore dung
and collections of hair we believe the cave was vis-
ited by mammoth. In addition at least one and prob-
ably two kinds of ground sloth, and various un-
known species of ungulates ranging in size from that
of mountain goats to elk were present. Subsequent
use of Bechan Cave by large mammals was largely
restricted to Homo sapiens ; charcoal from hearths
commonly occurs above the dung blanket. The dung
blanket at Bechan Cave is larger than a similar de-
posit of comparable age found beneath abundant
archeological material at Cowboy Cave in eastern
Utah. Neither site has yielded evidence of human
occupation during the time of dung blanket depo-
sition around 12,000 years ago.

As determined by pollen and plant macrofossil
analysis the environment outside Bechan Cave
roughly 12,000 years ago constituted a riparian gal-
lery of water birch, blue spruce, elderberry, snow-
berry, currant, sedges, and cattail. Other species
present in the sandy valley or along the cliff-rimmed

canyon beyond the cave include (in order of abun-
dance of fossil remains), big sagebrush ( Artemisia
cf. tridentata), prostrate juniper ( Juniperus com-
munis), Utah cactus ( Sclerocactus sp.), prickly pear
( Opuntia sp.), salt bush ( Atriplex sp.), thistle ( Cir -
sium), oak ( Quercus sp.), Indian rice grass ( Oryzop -
sis hymenoides) and other grasses, and hop horn-
beam (Osirya sp.). Many of these species do not
occur in the region today; presently the extra-local
species can be found at higher elevations or higher
latitudes such as at 1 ,800 m in canyons of the Henry
Mountains or Aquarius Plateau. The regional vege-
tation around 12,000 years ago may have been pri-
marily sagebrush rather than blackbrush ( Coleo -
gyne) steppe as it is now.

The canyon outside the cave continues to support
large herbivores as shown by the small herds of feral
cattle and horses dating from the establishment of
the Glen Canyon National Recreation Area. The
potential of the region to support the variety of
species present 11,600 to 12,900 years ago is less
certain. Did their habitat fail? All the plant taxa
present in the dung layer can be found today at
higher elevations in the Colorado Plateau. As in the
case of the Shasta ground sloth (Long and Martin,
1974; Hansen, 1980), dietary information on the
extinct fauna indicates that plant taxa ingested by
the vanished large herbivores continue to grow at
no great distance from the fossil site.

The most remarkable items found to date in Be-
chan Cave are the two dung boluses, M-l and M-

282

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

2. They are too large to represent most of the 39
genera of megaherbivores known from the late Pleis-
tocene (Rancholabrean) of North America. While
we cannot be certain that they were not voided by
a large grazing ground sloth such as Glossotherium,
they do not resemble fragments of Glossotherium
dung from Mylodon Cave, Chile. Within the pro-
boscidea the American mastodont, genus Manumit,
is known in Utah from the late Pleistocene. How-
ever, late Pleistocene mastodonts are rare in this
part of North America and they are thought to have
been browsers, not grazers. The dung boluses in
question are packed with large graminoid stems.
They most closely resemble African elephant ( Lox -
odonta) dung in size and texture and we attribute

them to mammoth ( Mammuthus ), a common Pleis-
tocene fossil in western North America.

Finally, it is of interest that the dung balls harbor
spores of the coprophilous fungus Sporormiella. Da-
vis et al. (1977) reported an increase in Sporormiella
spores at Wildcat Lake, Washington, coincident with
the introduction and heavy grazing impact of sheep.
With the extinction of two-thirds of the large mam-
mal species in western North America 1 1,000 years
ago a decline in Sporormiella may be expected. The
decline in Sporormiella and possibly other cop-
rophilous fungi should be sought in suitable spore-
rich sediments or cave deposits. Independently of
fossil animal remains, the change could reflect the
late Pleistocene extinction of megafauna.

ACKNOWLEDGMENTS

This study began during John Guilday’s last month; we think
it would have attracted his wide ranging paleontological interest.
Mr. John Lancaster, Superintendent of Glen Canyon National
Park, Dr. Adrienne Anderson, Larry Belli and Vic Vierra, all of
the National Park Service, provided outstanding help and valu-
able logistic aid. With their encouragement and that of Emil
Haury of the University of Arizona, we sought and received
National Geographic research support to continue our work on
Bechan Cave. Early phases of the work were funded by NSF grant

#BSR82- 14939 to P. S. Martin and T. R. Van Devender. Cave
maps and geological data were drafted by Agenbroad, pollen and
plant macrofossil analyses were contributed by Davis, with grass
identification by Larry Toolin; Mead analyzed fossil dung sam-
ples; Charles Bolen identified hair samples. Bonnie Fine Jacobs
responded to an appeal for African elephant dung from Kenya
and Cynthia Moss of the World Wildlife Fund, Nairobi, verified
our opinion about the similarity between the Bechan Cave sam-
ples and Loxodonta dung balls.

LITERATURE CITED

Davis, O. K., D. A. Kolva, and P. J. Mehringer, Jr. 1977.
Pollen analysis of Wildcat Lake, Whitman County, Wash-
ington: the last 1,000 years. Northwest Science, 51:13-30.

Faegri, K., and J. Iversen. 1975. Textbook of modem pollen
analysis. Hafner Press, New York, 295 pp.

Hansen, R. M. 1980. Late Pleistocene plant fragments in the
dungs of herbivores at Cowboy Cave. Pp. 179-189, in Cow-
boy Cave (J. D. Jennings), Univ. Utah Anthro. Papers, Salt
Lake City, 104:1-224.

Jennings, J. D. 1980. Cowboy Cave. Univ. Utah Anthro. Pa-
pers, Salt Lake City, 104:1-224.

Long, A., and P. S. Martin. 1974. Death of American ground
sloths. Science, 186:636-640.

Spaulding, W. G., and P. S. Martin. 1979. Ground sloth dung
of the Guadalupe Mountains. Pp. 259-269, in Biological
Investigations in the Guadalupe Mountains National Park,
Texas (H. H. Genoways and R. J. Baker, eds.). National Park
Service Proc. and Trans. Ser., 4:xvii + 1-442.

Thompson, R. S., T. R. Van Devender, P. S. Martin, A. Long,
and T. Foppe. 1980. Shasta ground sloth (Nothrotheriops
shastensis Hoffstetter) at Shelter Cave, New Mexico: envi-
ronment, diet, and extinction. Quaternary Res., 14:360-376.

Welsh, S. L., and C. A. Taft. 1981. Biotic communities of
hanging gardens in southeastern Utah. National Geographic
Soc. Res. Rept., 1972 Projects, pp. 663-681.

Address (Davis and Martin): Department of Geosciences, University of Arizona, Tucson, Arizona 85721.

Address (Agenbroad): Department of Geology, Northern Arizona University, Flagstaff, Arizona 8601 1.

Address (Mead): Center for the Study of Early Man, University of Maine, Orono, Maine 04473.

PLEISTOCENE TAPIRS IN THE EASTERN UNITED STATES

Clayton E. Ray and Albert E. Sanders
ABSTRACT

The name Tapirus haysii is reinstated for the larger Pleistocene nearly complete skull of T. veroensis from South Carolina is

tapirs, its type locality in North Carolina reestablished, topotypic described, and other material recorded, especially as it supple-

and other material reported, and its distribution reviewed. A ments previously known distribution.

INTRODUCTION

Simpson (1945) laid the foundation upon which
has been built all subsequent research on North
American Pleistocene tapirs. Although Leidy (1855:
200; 1 859: 106, for example), with his usual acumen,
understood that the fossils generally fell into two
size groups, one similar in size to the living T. ter-
restris ( =T . americanus) and another, larger form,
his T. haysii, specific assignments through the years
frequently seemed subjective, if not capricious, prior
to Simpson’s careful analysis. The notorious mor-
phological conservatism of tapirs has hampered spe-
cific identification of the most common specimens
(partial dentitions and isolated teeth) and contrib-
utes to a continuing suspicion, already clearly ex-
pressed by Leidy (1869:391), that phantom taxa may
be concealed within the ostensibly homogeneous,
stereotyped dentitions. The suspicion is reinforced
by the co-occurrence today in northern Colombia
of three species ( T. terrestris and T. bairdii sym-
patrically, and T. pinchaque nearby at higher ele-
vations; Hershkovitz, 1954:490-491, 494; fig. 62)
that are well demarcated in life and in cranial char-
acters, but broadly overlapping in most dental char-
acters, including size. There would seem to be no a
priori reason why as many taxa could not have left
their remains within a small area of eastern North
America given its altitudinal range and the time and
changing environments of the Pleistocene. The
widespread and fairly abundant fossil material is
generally inadequate for rigorous specific identifi-

cation, and it must be understood that our concept
of the number of species in the Pleistocene of North
America and of their geographic and temporal dis-
tribution remains very insecurely founded.

So few skulls are known that each one discovered
(Sellards, 1918; Simpson, 1945; Lundelius and
Slaughter, 1976; Bogan et al., 1980) has facilitated
significant advancement in understanding. Thus, a
primary purpose of this communication is to make
known an essentially complete skull of T. veroensis
from South Carolina, rivaled in quality only by the
holotype.

Until well preserved cranial material of a larger
North American Pleistocene tapir is available, its
relationship to T. veroensis, and more broadly with-
in the Tapiridae, will remain equivocal. Meanwhile,
T. haysii is resurrected for the large eastern tapir on
the basis of identification of its type locality, avail-
ability of topotypes, and Simpson’s (1945:65) rec-
ognition of the holotype as a P4 rather than M2.

We also discuss new or obscure specimens of both
taxa. As it has not been feasible to reexamine by
any means all recorded material, we have concen-
trated on that which extends or fills in the geographic
distribution and/or aids in understanding temporal
distribution.

In addition to Tapirus haysii and T. veroensis, all
nominal taxa from the Pleistocene of North Amer-
ica are reviewed.

METHODS AND MATERIALS

The following abbreviations for institutions are used through-
out:

AAS Arkansas Archeological Survey, State University
AMNH American Museum of Natural History, New York
ANSP Academy of Natural Sciences of Philadelphia
BM(NH) British Museum (Natural History), London

CAS California Academy of Sciences, San Francisco
CM Carnegie Museum of Natural History, Pittsburgh
ChM Charleston Museum, Charleston, South Carolina
MCZ Museum of Comparative Zoology, Harvard Univer-
sity, Cambridge, Massachusetts

NCSM North Carolina State Museum of Natural History,
Raleigh

283

284

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

NYSM New York State Museum, Albany
SCMC South Carolina Museum Commission, Columbia
UF Florida State Museum, University of Florida, Gaines-

ville

UF/FGS Florida State Museum/Florida Geological Survey
Collections, Gainesville
USGS United States Geological Survey
USNM [United States] National Museum of Natural History,
Smithsonian Institution, Washington, D.C.

All measurements are in millimeters; abbreviations for mea-
surements of cheek teeth are: L, anteroposterior diameter of crown;
W, maximum transverse diameter of crown; AW, width of crown
across anterior lobe; PW, width of crown across posterior lobe.

For comparative material of the living species we have relied
primarily on the collections of ANSP and USNM, which together
have provided 50 skulls of T. terrestris, 6 of T. pinchaque, 6 1 of
T. bairdii, and 12 of 7'. indicus. Adequate representation was
available for all except T. pinchaque, for which additional spec-
imens of all ages are needed. Three of the most important studies
of tapir systematics (Hatcher, 1896; Simpson, 1945; Radinsky,
1965) relied on a single skull of T. pinchaque for critical com-
parisons. This situation unfortunately is not unusual for large
mammals, museum collections of which frequently are inade-
quate in quantity and quality for study of individual, sexual,
ontogenetic, and geographic variation.

NOMINAL TAXA

All nominal taxa of North American Pleistocene
tapirs are discussed here in chronological sequence.
In much of the nineteenth century literature the fos-
sils were not clearly distinguished from living tapirs,
and in many instances were recorded under the name
Tapirus americanus, a junior synonym then current
for T. terrestris.

Tapirus mastodontoides was described by Harlan
(1825:223-225) on the basis of an isolated tooth
from Big Bone Lick, Kentucky. Because deinotheres
were then considered to be giant tapirs, he regarded
his new form as very small, only a little larger than
living tapirs. The tooth was recognized almost im-
mediately as an anterior milk molar of American
mastodon (Cooper, 1831:163; Hays, 1834:324). In
rebuttal to Cooper, Harlan (1835:265-267), having
made further comparisons with Cuvier’s collections
in Paris, insisted that the tooth was that of a tapir,
in fact the P1, and that its size and the structure of
its roots distinguished it from teeth of mastodon.
After Harlan’s death, when his collection ultimately
came to ANSP, Hays identified the holotype and
Leidy (1858:12) confirmed the conclusion of Cooper
and Hays that “the specimen was a first milk molar
of the Mastodon.” Many years later however, Simp-
son (1942:162) again raised the possibility “that it
is really a tapir, in which case the name is valid and
antedates any other for our fossil tapirs,” a suppo-
sition presumably quickly abandoned as the name
is not mentioned by Simpson (1945).

Until or unless some totally unexpected evidence
comes to light. Hays’ recognition of the holotype
must be accepted at this late date. The specimen,
ANSP 1 326 1 , illustrated here for the first time (Fig.
1A, B, I, M), appears to be a left DP2 of Mammut
americanum (DP2 according to Gillette and Colbert,
1976:33). The anteroposterior length of the crown

is approximately 33.4 mm and the maximum height
of the tooth 48 mm. It is deeply worn, and much
mutilated, lacking the posterolingual comer, and
most of the enamel of the anterolabial comer and
posterior margin. This damage must have occurred
after the original description, when Harlan (1825:
224) characterized the crown as “nearly quadran-
gular,” which would have made it more similar in
appearance to DP2 in Mammut than it now is. This
tooth is not common in collections, and seemingly
is rather variable.

An unworn right DP2, CM 2332F, from the
Frankstown Cave, Pennsylvania, with roots broken
away, is rather smaller (crown length 28. 1 mm) and
the crown more quadrangular in plan and the cross
lophs more transversely oriented (Peterson, 1926:
275, pi. 23). An unworn presumed left DP2, USNM
12062, from Cumberland Cave, Maryland (Gidley
and Gazin, 1938:70), is similar in crown shape and
pattern to that from Frankstown, but is larger (length
32.6 mm) and has essentially complete roots that
match those of T. mastodontoides almost exactly
(Fig. 1 E, F, K, O). A left DP2, MCZ 11106, described
and figured (but unfortunately subsequently dam-
aged) in the classic monograph by Warren (1855:
67; pi. 2, fig. 1; pi. 8, fig. 1) is very similar to Harlan’s
holotype (Fig. 1G, H, L, P). Harlan’s specimen is
matched best, however, by USNM 4986, an isolated
tooth thought to be a right DP2, from Kimmswick,
Missouri, with crown length of 33.9 mm, and max-
imum height as preserved of 51.1 mm. It was de-
scribed and illustrated by Hay (1912:679-680; pi.
17, figs. 6 and 7; 1914:346; pi. 47, figs. 9 and 10;
text-fig. 1 15), and is refigured here (Fig. 1C, D, J,
N).

The detailed similarity between the holotype of
Tapirus mastodontoides and teeth of known mas-

1984

RAY AND SANDERS- PLEISTOCENE TAPIRUS

285

Fig. 1. — Holotype of Tapirus mastodontoides Harlan 1825 and some examples of left DP2 of Mammut americanum, in occlusal (A-
H; stereo pairs), labial (I-L), and lingual (M-P) aspects, xQ.8. T. mastodontoides'. ANSP 13261, A-B, I, M; M. americanum : USNM
4986, C-D, J, N (all figures reversed); USNM 12062, E-F, K, O; MCZ 11106, G-H, L, P.

286

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

todonts seems sufficient to lay T. mastodontoides to
rest forever as a junior synonym of Mammut ameri-
canum. It seems probable however that T. masto-
dontoides is not without significance for fossil tapirs,
as it may help to explain the otherwise groundless
linkage between Big Bone Lick (the source of T.
mastodontoides) and Tapirus haysii, beginning with
Leidy’s lapsus ( 1859: 1 06) and carried forward to the
present. Although Leidy (1859:107, footnote) had
the opportunity to examine the type of T. masto-
dontoides, and recognized it as a mastodon milk
tooth, prior to publication of his paper in 1859, at
the time of writing the body of that paper he said
of Tapirus haysii that “it may be questioned if it
had not already been noticed by Dr. Harlan, under
the name of Tapirus mastodontoides. The specimen
described by Dr. Harlan, on which the latter was
founded, is also stated to have been a lower molar,
from Big-bone-lick.”

Tapirus americanus fossilis was introduced into
the North American paleontological literature by
Leidy (1849) with reference to: a left P4 from the
vicinity of Opelousas, Louisiana, reported and fig-
ured by Carpenter ( 1 842); to an immature left man-
dibular ramus and a mature left maxilla with P4-
M\ both from the Brazos River near San Felipe,
Texas, also reported and figured by Carpenter ( 1 846:
245, 247-249; figs. 3 and 4); and to an isolated,
deeply worn tooth regarded as a left DP3, from near
Natchez, Mississippi. Leidy ( 1 859: pi. 17, figs. 1 and
6) also refigured the specimens from Texas. This
name has been generally neglected and would best
remain so had it not been suggested (in unpublished
notes) that it qualifies as a valid senior synonym of
T. veroensis, a suggestion with which we do not
concur.

In the first place, it is questionable whether Leidy
intended that appellation as a formal trinomial, for,
in both the original (1849) and subsequent (1859:
106; 1869:391) publications, he characterized the
specimens as nearly or entirely indistinguishable
from their counterparts in living T. americanus (= T.
terrestris). Furthermore, although “ fossilis ” has been
used as a formal species-group epithet through the
years, it was also used at times in the nineteenth
century in quite another manner, simply to label as
fossils specimens otherwise indistinguishable from
their modern counterparts. Leidy (1853:9-10) him-
self employed this system in an annotated list of
North American fossil mammals, where this Pleis-
tocene tapir and at least five other taxa listed under
the names of living species were listed in the format,

Tapirus americanus (fossilis), with the word fossilis
not only parenthetical, but printed in a different and
smaller type-face than the binomial.

If it is insisted that the more rigid modern pro-
cedural standards be applied retroactively, and that
Leidy therefore established a species-group taxon
perhaps unintentionally, we would argue that Tap-
irus fossilis (Leidy 1 849) can be ignored as a for-
gotten name, because we can find no instance of its
use in publications of the past 50 years, including
Gillette and Colbert’s (1976) Catalogue of Types in
ANSP. (Simpson’s, 1945:65, mention does not con-
stitute usage.)

In the event that others wish to consider the sub-
ject, we hereby designate the immature left man-
dibular ramus from the Brazos River, Texas, ANSP
1 1 500, as the lectotype. Further, we note that Lund
earlier (1841:264) employed a similar notation
“Tapirus fossT) for remains from Brazilian caves,
at least in a species list. However, we have found
nothing to qualify as a description, and Winge ( 1 906)
did not perpetuate Lund’s usage. In any case, more
than one effective suppressant seems to be available
against resurrection of the name for North Ameri-
can Pleistocene tapirs.

Tapirus haysii, introduced as a nomen nudum in
1852 and validated in 1859, both by Leidy, but set
aside as indeterminate by Simpson (1945), is the
subject of detailed attention elsewhere in this paper,
where we propose its reinstatement.

Tapirus calif or nicus was described by Merriam
(1913) as T. haysii californicus on the basis of an
isolated tooth crown, probably a left M2, although
possibly a P4, from near Sonora, Tuolumne County,
California. In reviewing its status and literature,
Simpson (1945:67-69) gave the dimensions of the
holotype as L 25.3, AW 17.8, PW 17.5, noted that
the tooth is relatively narrow but not clearly sepa-
rable in size or structure from corresponding teeth
of T. veroensis, T. terrestris, or T. bairdii, and point-
ed out that it is significantly smaller than M2 of T.
haysii (his T. copei). Although the weakness of Sel-
lards’ (1918:65-66) effort to distinguish T. veroensis
from californicus was confirmed by Simpson’s ob-
servations, he wisely chose not to pursue the taxo-
nomic implications and instead to recognize this
smaller west coast tapir tentatively as T. californi-
cus, pending discovery of more diagnostic material.
This remains the most satisfactory course, although
new material from the vicinity of Los Angeles, listed
by genus but not described by Miller (1971:36, 52-
54), should aid in resolving the problem. We have

1984

RAY AND SANDERS- PLEISTOCENE TAPIRUS

287

had available the right maxilla with M'-M3 from
the Elk River fauna of Oregon, CAS 101, generally
referred to T. californicus, most recently by Leffler
(1964). Its dental dimensions are L M1, 21.9; AW
M1, 27.5; PW M1, 24.9; L M2, 25.4; AW M2, 30.8;
PW M2, 26.8; L M3, 26.3; AW M\ 29.5; PW M3,
24. 1 . On the basis of size and structure, the specimen
would be assigned routinely to T. veroensis had it
been found in the eastern United States.

It should be noted that T. californicus, whatever
its ultimate taxonomic fate, falls within the smaller,
T. veroensis size group, not the larger, T. havsii-T.
copei size group, contrary to Lundelius and Slaugh-
ter (1 976:228, 239) and Kurten and Anderson ( 1 980:
292).

Tapirus veroensis, described by Sellards ( 1 9 1 8) on
the basis of a fine skull from Florida, was the first
species of North American Pleistocene tapir to be
founded on fully satisfactory, confidently diagnostic
material, and it retains this distinction. The species
is discussed at length elsewhere in this paper, where
we record new material, most notably an additional
fine skull.

Tapirus merriami was established by Frick (1921:
311-314) on the basis of the crowns of two right
lower molars, each broken through the transverse
valley between anterior and posterior lophids, and
the fragments disarranged but embedded in matrix
with an associated mandibular fragment (Frick,
1921: fig. 26). The specimen came from beds of the
Bautista Creek badlands in southern California, lat-
er assigned an Irvingtonian age. The tooth fragments
were reassociated to form a slightly worn, nearly
complete crown of an anterior tooth thought to be
M,, and an unworn, presumably unerupted, incom-
plete crown of a posterior tooth thought to be M2.
No reason was given for identifying the teeth as
and M2 rather than M2 and M3. The proportions of
the teeth are at least as compatible with the latter
assignment, if not more so, and the sizes of the teeth
would then be more compatible with Simpson’s
(1945) sample of 7 . copei. The measurements of the
teeth (from Frick, 1921:313) are L M,(M2?), (32.5);
AW Mj(M2?), 23.7; PW M,(M2?), 22; L M2(M3?),
(32.5); AW M2(M3?), 23.6; PW M2(M3?), 20.5.

The lengths of the crowns were regarded as im-
precise by Frick, but, judging from his illustrations,
it seems also that the widths, especially PW of the
unerupted posterior tooth, might well be suspect.
Simpson (1945:69) accepted Frick’s placement of
the teeth and thought that the species would prove
to be valid because M; was significantly larger than

any other from the North American Pleistocene. He
did, however, point out that differences between
anterior and posterior widths of the teeth were with-
in the range for T. copei, and that the two forms
were closer to one another than to any other on the
basis of the meager comparisons possible, though
he refrained from suggesting special relationship.

A small quantity of scattered, mostly fragmentary,
disparate elements of large Pleistocene tapirs has
accumulated since 1921 and mostly since 1945. This
material is reviewed under Distribution of Tapirus
haysii below. Meager as it is, it goes far toward clos-
ing the gap in size, morphology, and distribution
between Port Kennedy and Bautista Creek, and sug-
gests to us that the twain indeed shall meet within
the foreseeable future.

Tapirus tennesseae was named by Hay (1920:88-
90) on the basis of 10 isolated teeth, USNM 8949,
probably representing at least two individuals, from
a cave near Whitesburg, Tennessee. Simpson ( 1 945:

65) considered it “a possible synonym [of T. ver-
oensis], essentially indeterminate at present, and
properly ignored.” Simpson’s conclusions are rein-
forced by additional material from the region, no-
tably the skull from Claiborne Cave, Tennessee, re-
ported by Bogan et al. (1980) and described further
herein. We feel that T. tennesseae should not be
ignored but properly regarded as a junior synonym
of T. veroensis.

Tapirus copei was proposed by Simpson (1945:

66) for the large tapir in the Irvingtonian fauna of
Port Kennedy, Pennsylvania. This material had pre-
viously been referred to T. haysii, which Simpson
(1945) regarded as indeterminate. Elsewhere in this
paper we argue for reinstatement of T. haysii, and
the relegation of T. copei to its junior synonymy.

Tapirus excelsus was named by Simpson (1945:
70) on the basis of a juvenile skull with partial skel-
eton and another partial younger skeleton, from a
sinkhole near Enon, Missouri. He recognized its
general similarity to T. veroensis, and Lundelius and
Slaughter (1976) have shown that the supposed cra-
nial distinctions are attributable to ontogenetic
change, a conclusion reinforced by our additional
material. Lundelius and Slaughter (1976) left open
the possibility of recognizing excelsus as a large,
midcontinental subspecies of T. veroensis. We feel
that subspecies are not usefully employed in the
present state of knowledge of fossil tapirs. Never-
theless, as additional material accumulates, espe-
cially from the midcontinent, the possibility should
be kept in mind that T. excelsus may yet prove to

288

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

represent a distinguishable geographic entity. Re- broad teeth continue to be found in the region (Par-

mains with rather large (for T. veroensis), rather malee et al., 1969:26-30; Lundelius, 1972:60-63).

SYSTEMATIC ACCOUNTS

Tapirus haysii Leidy 1859

Type description.— The specific name was intro-
duced by Leidy (1852a) as a nomen nudum, and
additional material was referred to it in the same
year (Leidy, 1852 b), still without description. Later,
in referring yet another specimen to the species, he
alluded (Leidy, 1 855:200) indirectly to the large size
of T. haysii. but not until 1859 did he provide mea-
surements and illustrations of the holotype, in what
is generally regarded as the valid type description
(Leidy, 1859:106; pi. 17, figs. 7 and 8). This last
publication is usually cited as 1 860. It was published
in Holmes’ “Post-Pleiocene Fossils of South-Car-
olina,” which appeared in parts from 1858 to 1860.
Blackwelder and Ward (in press) have presented
persuasive evidence that the parts including Tapirus
haysii were published in 1859.

Type specimen.— There has never been any ques-
tion about the holotype being the isolated tooth
crown with remnants of roots catalogued as ANSP
1 1 504, although Hay (1912:591; figs. 38 and 39) in
his figure captions mistakenly identified his repro-
duction of Leidy’s (1859, pi. 17, figs. 9 and 10)
illustrations of the Evansville, Indiana, tooth (ANSP
11505) as representing the type, an error that he
later corrected (Hay, 1927:76).

Leidy (1859:106) regarded the holotype as a sec-
ond lower molar, but Simpson (1945:65) pointed
out that the proportions of the crown correspond
much better to those of the fourth lower premolar.
We regard it as a right P4 (Fig. 2D).

Type locality. — Simpson ( 1 945:65) stated that “the
locality is not now known and can never be estab-
lished with any degree of certainty.” The uncertainty
arose from Leidy’s (1 859: 106) assertion that the type
specimen was “supposed to have been obtained from
Big-bone-lick, Kentucky.” As indicated previously
in the discussion of Tapirus mastodontoides, the
association between the two in Leidy’s mind at the
time of writing his most substantial account of T.
haysii may help to explain the spurious connection
of the latter with Big Bone Lick. There is in fact no
evidence for occurrence of any tapir at Big Bone
Lick in any of the massive collections made there
(Schultz et al., 1963). Although Leidy was arguably

the best vertebrate paleontologist that this country
has produced, this would be by no means a unique
instance of his garbling data (cf. Emry and Purdy,
in press).

In the first published reference to the type (Hays,
1852), “Dr. Hays stated that the tooth of the fossil
Tapir presented by him this evening [16 March
1852], was found in the bed of a canal in North
Carolina. It had been in his possession for several
years, and was the first fossil Tapir tooth found in
North America.” Also, in the “Donations to Mu-
seum” (of ANSP) for that date is listed “Molar of
a fossil Tapir, from North Carolina; and mould in
plaster, with several casts of the same. Presented by
Dr. Isaac Hays.” (The mould and casts are still pre-
served with the specimen.)

Fortunately, evidence does exist not only to sup-
port the derivation of the tooth from North Caro-
lina, but also to localize its source to the north bank
of the Neuse River, 16 mi below New Bern (Fig. 3,
locality 1). During the first quarter of 1832 Thomas
Nuttall made a southern field trip funded ($260) by
Harvard University, and in his report of 21 June to
the Corporation he indicated that “the principal part
of my time was spent in the vicinity of Newbem in
North Carolina” and that “on the banks of the river
Neuse, while in the neighbourhood of Newbern, I
carefully examined for fossils of wh. an immense
stratum is here exposed to view by the washing of
the river” (Graustein, 1967:260-261). Nuttall went
on to discuss the fossils and their implications at
some length and promised that “an arranged col-
lection of these fossils will be placed in the museum
of the University.” Recent efforts to locate such a
collection in the MCZ have failed.

Graustein (1967:261) went on to state that “He
was fortunate in some of his fossil finds. At a meet-
ing of the Academy on June 19, Dr. Hays exhibited
the anterior molar of a mastodon that Nuttall got
and more than a year later announced from the same
source a tooth of ‘A Tapir,’ the first evidence found
in North America of this animal.” The hand-writ-
ten, unpublished minutes of meetings of the ANSP
preserved in their archives record the following ver-
bal communications: for 19 June 1832, “Dr. Hays

1984

RAY AND SANDERS -PLEISTOCENE TAPIRUS

289

Fig. 2. — Lower dentition of Tapirus haysii in occlusal aspect, x I. A) USNM 347270, left M2; B) USNM 243692, right M2 (cast); C)
USNM 215061, right M3; D) ANSP 11504, right P4 (Holotype); E) AAS 69-658, left mandibular ramus with M.-M, and posterior

lophid of P4; F) USNM 244439, left mandibular ramus with P2,
Hancock Ditch, Arkansas; F) Hanover Quarry, Pennsylvania.

exhibited the anterior molar of mastodon found by
Mr. Nuttal (sic) near Newbem N. Ca.”; for 6 August
1833, “Dr. Hays announced that he possesses the
fossil tooth of a Tapir, from Newbem N. Ca. found

3, DP4) M„ and M- (cast). A-D) Neuse River, North Carolina; E)

by Mr. Nuttall, which the Dr. states to be the first
instance of the detection of the remains of this an-
imal on the North American continent.” Obviously
Hays (no friend to Harlan; Simpson, 1 942: 1 64-165)

290

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 3. — Distribution map for Tapirus haysii. Localities correspond to numbered localities in the text. Locality 1 is the type locality;
locality 1 8 is the type locality for T. merriami.

had by that time already rejected T. mastodontoides
Harlan 1825 as a tapir. One can only surmise the
satisfaction with which Hays recognized the Nuttall
specimen as tapir, especially plucked from under
Harlan’s nose, as Harlan had noted as early as 1835
(p. 267) that “Mr. T. Conrad possesses specimens
[of Equus] from the Pliocine on the river Neuse, 1 6
miles below Newbem, N. C.” and later wrote the
following note: “April 19th, 1837. Examined a col-
lection of fossil bones collected by Mr. Nuttall near
Newbern, North Carolina,” and listed the animals
identified, which did not include tapir (in Conrad,
1838:1 1 [ 1 3], footnote).

Unfortunately, Nuttall’s field journals seem to
have been destroyed after having survived for more
than a century (Graustein, 1967:397), else we surely
would have had much more explicit data direct from
the collector, who never carried out his plan to pub-
lish extensively on the collection. However, as
Graustein (1967:262) noted, Timothy Conrad and
Richard Harlan reported on items from Nuttall’s

collections, and he became acquainted in New Bern
with Hardy B. Croom, a lawyer, plantation owner,
fellow botanist, and fossil enthusiast.

Croom (1835:168-171), in a letter dated 12 Sep-
tember 1833, was the first to report on the richly
fossiliferous deposits along the lower Neuse River,
“on the estate of Mr. Benners, occupying the north
bank of Neuse River, sixteen miles below New-
bem.” In his annotated list of specimens Croom
mentioned two as having been given to Mr. Nuttall.

Timothy Conrad (1835:107-110) reported on a
collection sent to him by Croom after the latter’s
publication of the same year, listing 67 species of
mollusks, and later (Conrad, 1 838:x[ 1 2]) stated, “on
the left bank of the Neuse, about 15 miles below
Newbem, in North Carolina, there is a vast collec-
tion of the rolled and water-worn bones . . . speci-
mens of which are preserved in my cabinet, most
of them a present from my kind friend, the distin-
guished Nuttall, who thoroughly explored this re-
markable locality.” He went on to quote Harlan’s

1984

RAY AND SANDERS- PLEISTOCENE TAPIRUS

291

note reproduced in part above, and Harlan (1842:
143) later alluded briefly to “a collection of fossils
obtained by Mr. Nuttall, in Newbern.”

Mitchell ( 1 842: 1 28) stated that “We have also the
grinder of an Elephant, found in the marl pits of the
late Lucas Benners, Esq., 16 miles below Newbern,
and other teeth not yet determined.”

Charles Lyell (1843:39) recorded that “Mr. Nut-
tall discovered, on the Neuse 15 miles below New-
bum, in South Carolina [corrected in reprint to
Newbern, in North Carolina; Lyell, 1844:323], a
large assemblage of mammalian bones. . . . Mr.
Conrad presented Mr. Lyell with the tooth of a horse
covered with barnacles, from this locality. Professor
Owen has examined it . . . .” In the account of his
second visit to the United States Lyell (1849:259)
noted that “fossil horses were found by Mr. Nuttall
on the banks of the Neuse, fifteen miles below New-
bern, in North Carolina,” and in a footnote, that
“Mr. Conrad intrusted me with Mr. Nuttall’s col-
lection, and Mr. Owen has found among them the
three species of Equidae . . . .” Some horse teeth
from the locality found their way to the BM(NH)
(Lydekker, 1886:89).

Although the Benners locality has fallen into lat-
terday obscurity, it clearly attracted great attention
in the 1830’s and 1840’s, and as late as 1923 (p.
358) Hay was able to say, “Doubtless the locality
in North Carolina, the most important to the stu-
dent of Pleistocene vertebrate palaeontology, is that
reported long ago on the northern shore of Neuse
River, 16 miles below Newbern.”

There is no evidence that Nuttall collected fossils
elsewhere in North Carolina, and the context of all
contemporary reports indicates that reference to
“Newbern” as a locality is merely shorthand for the
Benners locality, 16 mi downstream. The one men-
tion (Hays, 1852) of the source as the “bed of a
canal” in North Carolina can be dismissed as yet
another lapsus in view of the weight of evidence in
favor of the Benners estate. Some Pleistocene ver-
tebrate remains had been encountered in excava-
tions for the Clubfoot and Harlow canal on the south
side of the Neuse and slightly downriver, but very
little was salvaged, the date (1817, Mitchell, 1842:
128) was too early, and there is no hint that Nuttall
obtained material from there. Thus, the evidence
seems sufficiently conclusive to establish the type
locality for Tapirus haysii as the estate of Lucas
Benners, river bank and adjacent marl diggings on
north side of Neuse River, 16 mi below New Bern,

North Carolina, in what was then Craven Co., but
since 1872 has been Pamlico Co. (Corbitt, 1950:
169).

Records of real estate transactions housed in the
Craven County Courthouse, New Bern, show a 272-
acre tract of land known as the “Bluff of Beards
Creek” purchased by John Benners on 27 March
1767; a purchase of 100 acres on 16 December 1 790
on the north side of Neuse River on both sides of
the “old road that leads to Wilkinsons Point” and
referring to “Smith Branch” as one landmark; an
additional purchase in December 1792 of 228 acres
on the north side of the Neuse “being part of a patent
granted to Captain James Beard for 650 acres”; and
finally, on 1 May 1804 the purchase by Lucas Ben-
ners of 640 acres at the head of Beard Creek. Rich-
ards ( 1 950:4 1 ) also identified the Benners plantation
as lying on the “east side of Bairds (sic) Creek.”
Richards (1936:1634) was mistaken in supposing
that Mansfield’s locality 1 5 mi below New Bern was
near the Benners locality, as the two are on opposite
sides of the Neuse River.

Although excavations have not been available in
modem times, beginning in the early 1960’s enor-
mous numbers of vertebrate fossils, and lesser quan-
tities of invertebrates, have been collected from the
north bank and adjacent shelving bottom of the
Neuse River by P. J. Harmatuk and sons, Mrs. Thel-
ma Bennett, Raymond Douglas, Mr. and Mrs. Phil-
ip J. Kennel, and others. These collections have been
made for the most part from the vicinity of the
mouth of Beard Creek, at approximately 35°N,
76°52'W, on the Arapahoe 7.5-minute quadrangle,
USGS, and downstream some 2.3 mi (3.7 km) to
the vicinity of Smith Gut, approximately 34°59'N,
76°50'W, just upstream from Wilkinson Point, on
the Cherry Point 7.5-minute quadrangle, USGS. This
is undoubtedly the same segment of river in which
Nuttall collected. Among the thousands of speci-
mens recovered (some 1,400 horse teeth for ex-
ample) are a few of tapir, representing animals the
size of both T. veroensis and T. haysii. Thus, more
than 1 50 years after the holotype was collected, we
have available topotypes of Tapirus haysii.

Geologic age. — Simpson (1945:65) stated that “It
is much the most reasonable assumption that the
[type] specimen is from the Pleistocene, but even
this is not absolutely certain, and of course it cannot
be assigned to any particular part of that epoch.”
No reason was given for raising doubt as to its Pleis-
tocene age, and we are aware of no evidence for pre-

292

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Pleistocene tapirs of such large size in North Amer-
ica (unless all of the Blancan is regarded as Pliocene;
cf. Savage and Russell, 1983:345, fig. 7-4).

Although stratigraphic distinctions were recog-
nized as important from the beginning (Croom, 1835:
169) and considerable attention has been given to
Neogene biostratigraphy along the Neuse below New
Bern during the modem era, beginning with Dali
(1892), generally accepted answers to critical ques-
tions on number and age of discrete stratigraphic
units, faunal associations, stratigraphic relations on
opposite sides of the Neuse, and stratigraphic no-
menclature, are yet to be forthcoming.

The marl pits dug by Mr. Benners reached depths
up to 25 feet, or 10 feet below the river’s surface;
the upper levels produced shark teeth, bones of ma-
rine fish, and sea shells; the lower, terrestrial ver-
tebrates and sea shells of great variety (Croom, 1835:

1 69). In a collection from these pits sent by Croom
at that time, Conrad (1835:107-1 10) identified 67
species of mollusks, including a few indicative of
age no younger than the Yorktown Formation ac-
cording to Druid Wilson (personal communication).
Unfortunately, recent efforts to locate this collection
at ANSP have not been successful. Interestingly, the
Yorktown indicators were not included in a later
list by Conrad (1842:191-192), but that list appar-
ently was based on a collection made by Conrad
himself, and he noted that the older of the two “for-
mations” exposed in digging the marl was not ac-
cessible at the time of his visit as the pits had been
refilled. He mentioned also that Mr. Benners had
told him that all of the bones came from the younger
deposit. He further noted that many of the bones
are waterworn, black, and mineralized, and presum-
ably transported down the ancient Neuse River to
be deposited among marine shells. Foster (1857:
166), attributing his information to conversations
with Conrad, stated that the terrestrial mammal re-
mains came from an upper Tertiary stratum covered
by a deposit of Pleistocene shells up to 15 ft thick,
and that the much waterworn bones may have been
redeposited from an older stratum. Stephenson
(191 2:268, 289) regarded both levels at the Benners
locality as Pleistocene.

Unquestionable pre-Pleistocene fossils are known
in modem collections from the immediate area and
from the region. The large collections of vertebrate
fossils made by Mr. Harmatuk and others from the
Neuse River include marine vertebrates indistin-
guishable from those occurring in the Yorktown
Formation at the Lee Creek Mine and in the Su-

perior Stone Company Quarry at New Bern. This
latter locality exhibits a basal rubble of Yorktown
age including marine vertebrates similar in taxo-
nomic composition and in preservation (mutilated,
rolled, bored, encrusted, blackened). Marine mol-
lusks of Pliocene age are known from the south side
of the Neuse River across from the Benners locality,
either redeposited (Du Bar and Solliday, 1963:226)
or in place (Fallaw and Wheeler, 1969:52). Ward
and Blackwelder (1980:25, 57) recognized evidence
of their pre- Yorktown Eastover Formation, Cob-
ham Bay member (upper Miocene), at their locality
48, based on mollusks collected by Mr. Harmatuk
within the section of river of concern here.

Considering the restricted geographic area of the
Neuse River estuary below New Bern and the num-
ber of geologists who have devoted attention to it,
the interpretation of its geologic history seems in-
ordinately confusing. The undoubtedly real bio-
stratigraphic complexity is now buried beneath a
manmade overburden of biostratigraphic nomen-
clature, which continues to be deepened and re-
worked. Complete review of the extensive and con-
flicting literature is beyond the scope of this paper
and would in any case be inconclusive until the
critical field work and mature analysis have been
carried out.

For immediate purposes relating to the occur-
rence of tapirs and other terrestrial mammals at the
Benners Estate, there is no evidence of pre-Pleis-
tocene deposits in place and naturally exposed at
surface.

The Pleistocene beds here were assigned to the
Pamlico Formation of Stephenson (1912:286) by
Richards (1936:1634; 1 950:4 1 ) and would have been
included in the original concept of the Croatan For-
mation of Dali (1892:205-206, 209). Most recently
these beds have been referred to the Beard Creek
member (Mixon and Pilkey, 1976:1 6) of the Flanner
Beach Formation of Du Bar and Solliday (1963).
The Beard Creek member is equivalent to the Neuse
Formation of Fallaw and Wheeler (1969), both hav-
ing as type locality the bluff adjacent to the mouth
of Beard Creek. However, it now seems to be gen-
erally agreed that the name Neuse Formation should
be suppressed in deference to the name Flanner
Beach Formation (Wheeler et al., 1979:47; Black-
welder, 1981:12). Mixon and Pilkey (1976: fig. 4)
showed the base of the Flanner Beach Formation
(the base of their Quaternary, but not that of most
authors) at zero to 10 ft below sea level along the
Neuse River at the Benners locality. They regarded

1984

RAY AND SANDERS- PLEISTOCENE TAPIRUS

293

the Flanner Beach Formation as early Pleistocene,
whereas others (for example, McCartan et al., 1 982)
regarded it as late Pleistocene. Literature cited in
papers cited here will provide entree to the numer-
ous publications on the geology of the area. Ob-
viously, no conclusive statement about the source
bed or its exact age can be made as yet, but it may
be concluded tentatively that the terrestrial verte-
brates are derived from the Beard Creek member
of the Flanner Beach Formation of early or late (but
not latest) Pleistocene age.

The field context is such as to inspire confidence
that conclusive evidence lies fallow in the ground
of the Benners plantation adjacent to the Neuse Riv-
er. An excavation comparable to the marl pits of
1 50 years ago could yield a described and measured
section, with stratigraphically controlled samples of
vertebrates, macroinvertebrates, and microfossils.
Such a study not only would close the gap between
the meager, practically lost, promising beginnings
of the early nineteenth century and the extensive
(and expensive), relatively less fruitful work of the
modem era, but also would aid in resolving much
inconclusive contention.

Associated fauna. — The remains of quadrupeds identified
among the early collections, mostly Nuttall’s, included “teeth of
the great mastodon . . . the hoof, horns, and vertebrae of an elk
of great size and the teeth of an animal supposed to be the hyena”
reported by Croom (1835:169; the presence of hyena was ques-
tioned at the time by the editor and has not been verified since,
though of course quite possible). Harlan (in Conrad, 1 838:xi[ 1 3];
repeated by Harlan, 1842:143) identified “hippopotamus, mas-
todon, elephas, elk, deer, horse, sus, seal, cetacea, tortoise, shark
(several species), skate, snake, and fish.” Lyell (1849:259) re-
ported “ Equus curvidens. E. plicidens, and a third species of the
size of E. asinus ” identified by Owen from Nuttall’s collections.

No conclusive evidence has ever been presented of precolumbi-
an hippopotamus or Sus anywhere in North America. Otherwise,
with the exception of elk and snake, the modem collections from
the Neuse River correspond well with these early indications.
Preliminary study of the collections deposited in USNM by Mr.
Harmatuk yields the following terrestrial mammals: Megalonyx,
Eremotherium, Castoroides. Hydrochoerus, Mammut, Cuvieron-
iusl, Mammuthus, Tapirus spp., Equus spp., Nannipus, Platy-
gonus. a camelid, Odocoileus, and Bison. These are identified
from float collections and very well may not constitute a unit
fauna. Bruce MacFadden (personal communication) has identi-
fied as Nannipus phlegon some 38 teeth from the collection,
indicative of an age no younger than Blancan.

The same segment of river bottom and bank has yielded marine
vertebrates, including sharks, bony fishes, pinnipeds, and ceta-
ceans, at least most of which are indicative of an earlier fauna,
of Yorktown aspect.

On Harlan’s behalf it may be mentioned that teeth of Sus are
included in modem collections from this and other southeastern
coastal localities, and some of them are blackened and very sim-

ilar in appearance to teeth of known Pleistocene species. Further,
the modem collections from the Neuse River include fragments
of walrus tusks (thought to pertain to the Yorktown part of the
assemblage). Similar specimens well might have provided the
basis for inclusion of hippopotamus in Harlan’s list, not an egre-
gious error for the time.

Measurements.— Authors from Cope (for example, 1871:95)
onward have remarked that T. haysii was distinguished from T.
terrestris (=T. americanus ) and later from T. veroensis by “size
only,” presumably implying inadequacy. However, size alone is
sufficient if truly distinctive, and affords a convenient, testable
criterion especially useful for isolated teeth of tapirs, the stereo-
typed character of which may be strongly suspected of masking
species that would be readily separable on skulls. Thus, although
individual variation in tooth size (to say nothing of the difficulty
of correct placement in the dental series of many isolated teeth)
and generally unknown variation with time and geography un-
doubtedly swamp out real distinctiveness in size by introducing
spurious overlap from multiple sources, there yet seems to be
two clusters in size that are sufficiently extreme to militate against
assignment to a single extraordinarily variable species.

Simpson’s (1945) unified samples of T. copei and T. veroensis
showed almost no overlap in tooth measurements, but Lundelius
and Slaughter’s (1976) measurements of material referred to T.
veroensis from various localities in Florida and Texas ostensibly
resulted in expanding the observed size range of T. veroensis
upward to engulf that of T. copei more or less completely, es-
pecially for the lower dentition (Tables 1 and 2). On the basis of
their data and identifications the conclusion seems inevitable that
only one species can be recognized in North America. However,
Lundelius and Slaughter did not take that step. We have not been
able to substantiate their data for the upper limits of their ob-
served ranges. Further, it might appropriately be asked on what
basis the larger specimens were assigned to T. veroensis rather
than to T. copei (=T. haysii). In at least one instance a mandibular
fragment (UF 8225) was included in their table of specimens of
T. veroensis that had previously been identified as T. copei by
Ray (1964). Although the distinctions in size between teeth of
T. veroensis and T. haysii thus seem blurred at present, we feel
that size probably is still a usable criterion and prefer to await
confirmation of Lundelius and Slaughter’s findings. The time is
perhaps ripe for a new, unified statistical study of the teeth of
North American Pleistocene tapirs, with particular attention to
geographic and chronologic variation. There is for example, a
large sample from Melbourne, Florida, divided among MCZ,
USNM, and Amherst College, that has never been thoroughly
studied, and there is a still growing sample from the Cooper River,
South Carolina, that is suitable for quantitative study. A second
sample comparable to that from Port Kennedy is much to be
desired.

In addition to measurements given in Tables 1-3, dimensions
of selected individual teeth are given in the text where those teeth
are discussed.

Distribution.— As Simpson (1945:65) emphasized, citation of
the name T. haysii in the literature cannot be relied upon as
evidence that a large tapir is present at a given locality. This
applies in large part to the post- 1945 literature as well, in which
large tapirs generally are listed as T. copei. We have attempted
to compile a reasonably comprehensive record of distribution
(Fig. 3) based on direct examination of new and old specimens
and on critical evaluation of published accounts.

None of the Rancholabrean localities for T. copei listed by

294

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table l .—Dimensions (mm) of upper cheek teeth o/Tapirus veroensis and T. haysii. Asterisks (*) indicate possibly mixed samples
including deciduous predecessors or adjacent teeth. Parentheses indicate that the tooth is actually DP4, not P4.

Dimension

ChM PV4257
left/ right

Claiborne

Cave,

Tennessee,

left

(Bogan et al.,
1980)

UF/FGS
V277,
holotype
(Lundelius and
Slaughter,
1976)

T. veroensis,
Texas

observed range
(Lundelius and
Slaughter, 1976)

T. veroensis,

Florida

observed range
(Lundelius and
Slaughter, 1976)

T. veroensis,
Seminole field
observed range
(Simpson, 1945)

T. copei
Port Kennedy
observed range
(Simpson, 1945)

p1

L

16.2/16.2

19.2

17.5

19.2-20.4

17.5-20.8

16.9-20.4*

22.4-24.9

W

16.1/15.6

17.8

14.9

14.5-19.6

14.9-18.6

14.5-17.4*

19.6-21.5

p2

L

17.8/17.7

20.3

18.7

19.8-20.5

18.7-21.1

18.5-20.8

21.9-24.0

AW

20.0/20.0

20.0

20.9

21.1-23.9

19.6-23.2

19.3-22.9

25.5-26.5

PW

21.0/21.3

24.0

22.8

22.4-25.9

22.8-25.8

21.0-25.5

27.4-27.9

p3

L

18.9/19.1

21.1

19.1

20.6-23.5

19.0-22.0

19.3-22.2*

22.7-24.5

AW

22.7/22.6

24.0

24.5

25.8-27.6

22.0-26.3

22.8-27.6*

27.0-29.5

PW

22.4/22.2

25.0

24.1

25.0-27.6

24.1-26.5

23.0-27.5*

26.1-29.0

pj

L

19.6/19.4

(21.7)

20.7

21.0-24.5

20.0-22.9

19.3-22.2*

24.1-26.4

AW

24.3/23.9

(23.1)

26.9

29.4-30.3

24.2-28.6

22.8-27.6*

29.9-31.8

PW

23.5/22.9

(21.7)

26.2

29.1-29.3

22.5-28.9

23.0-27.5*

28.4-30.1

M1

L

21.1/21.2

22.8

22.4

22.6-25.6

20.2-23.8

19.6-27.1*

25.8-26.4

AW

24.7/24.7

25.7

26.8

25.2-29.2

24.2-28.4

22.8-30.9*

28.9-31.1

PW

23.0/23.2

23.3

24.0

24.0-27.2

22.3-26.2

21.3-27.7*

25.8-27.9

M2

L

24.0/23.6

25.6

25.2

25.7-27.4

23.5-27.0

19.6-27.1*

27.3-29.7

AW

27.4/27.1

27.7

27.8

30.3-31.1

26.3-31.1

22.8-30.9*

31.3-34.9

PW

23.9/23.7

24.4

26.1

26.5-29.0

23.8-28.4

21.3-27.7*

28.0-31.5

M3

L

—

—

25.1

23.7-28.5

23.5-26.1

22.4-26.0

26.8-29.2

AW

—

—

29.0

27.6-32.9

28.0-32.1

25.5-31.4

31.0-34.1

PW

-

-

24.6

22.3-29.1

23.1-28.1

20.4-24.8

26.5-29.0

Table 2. — Dimensions (mm) of lower cheek teeth o/Tapirus veroensis and T. haysii. Parentheses indicate uncertain measurements due
to breakage. Asterisks (*) indicate possible mixed samples including deciduous predecessors. Measurements of the Hanover Quarry
specimen, left ramus, were made by John E. Guilday on the original, not on the casts in CM and USNM.

Dimension

AAS

69-658,

Weona,

Arkansas

Lehner Site,
Arizona
(Lance, 1959)

USNM 24588,
Ladds, Georgia
left/right

USNM 214663,
Island Creek,
North Carolina
left/right

CM 30230,
USNM 244439
Hanover Quarry,
Pennsylvania

T. veroensis,
Florida

observed range
(Lundelius and
Slaughter, 1976)

T. veroensis,
Seminole Field
observed range
(Simpson, 1945)

T. copei,
Port Kennedy
observed range
(Simpson, 1945)

p2

L

—

(26.7)

— /22.8

— / —

26.0

22.3-28.5

21.1-26.9*

24.7-27.0

W

-

17.9

—/1 5.4

-/-

17.3

13.8-19.1

13.2-15.6*

15.5-17.6

p3

L

—

(26.7)

— / 1 9.9

19.6/20.2

23.1

20.0-24.5

20.6-20.9

23.0-25.1

AW

—

(16.8)

—/1 5.7

15.0/15.2

18.3

15.1-17.5

16.8-17.6

16.1-18.0

PW

-

21.4

—/1 6.8

1 6.6/ 1 6.6

20.4

16.0-21.4

18.3-18.9

17.8-20.2

DP

4 L

—

—

— / —

— / —

24.4

20.9-22.8

AW

—

—

— /—

— / —

19.0

15.2-16.5

PW

-

-

-/-

-/-

18.8

14.6-16.1

P<

L

—

29.7

20.4/20.3

19.6/20.2

—

20.2-25.2

22.0-24.6

24.1-24.9

AW

—

20.4

17.7/17.9

17.9/18.1

—

15.9-20.6

17.7-20.9

18.3-21.7

PW

24.3

(20.7)

18.0/18.4

18.4/18.3

-

15.9-22.0

19.5-22.2

19.5-22.8

M,

L

22.6

25.4

21.6/21.9

20.2/21.0

24.2

20.0-26.9

21.7-25.0

23.4-27.0

AW

21.4

20.8

18.0/18.6

17.9/17.8

21.5

17.0-22.0

17.6-19.3

19.8-22.9

PW

19.0

20.6

16.3/16.9

16.3/16.3

20.2

17.5-22.8

16.1-18.9

18.7-20.8

m2

L

26.0

(30.6)

23.6/24.1

23.4/23.5

29.1

22.5-29.0

22.6-26.6

27.4-30.8

AW

22.6

(22.6)

18.6/18.8

18.8/18.8

22.1

18.4-23.4

18.9-19.7

20.5-24.1

PW

21.4

(22.3)

17.4/17.7

17.4/17.5

21.9

17.8-22.8

18.0-19.7

18.3-22.8

m3

L

29.1

33.7

25.1/25.8

25.0/-

—

25.4-32.2

25.7-28.3

30.1-31.5

AW

23.0

24.2

18.6/19.1

19.1/19.2

—

19.0-23.1

19.5-20.5

21.7-23.8

PW

19.9

21.1

16.9/17.0

17.0/-

-

17.0-21.7

17.4-18.6

19.2-20.5

1984

RAY AND SANDERS- PLEISTOCENE TAPIRUS

295

Table 3.— Measurements (mm) of some skulls o/ Tapirus ver-
oensis. Dimensions A-H are those of Simpson (1945), I is that
of Hershkovitz (1954). Values of A-E and G-I for holotype are
after Simpson (1 945:75). Figures in parentheses are minima, ow-
ing to damage.

Dimension

UF/
FGS
V 277,
holo-
type

SCMC

ChM SC

PV4257 75.31.173
Cooper Cooper
River, River,

South South

Caro- Caro-
lina lina

Clai-

borne

Cave,

Ten-

nessee

A. Prosthion-foramen magnum

379

345

—

—

B. Prosthion-choana

199

189

—

—

C. Prosthion-orbit

174

168

—

—

D. Sagittal crest-basisphenoid

113

112

108

1 14

E. Postauditory breadth

122

(110)

117

—

F. Oblique nasal length

(68)

(60)

-

—

G. Nasal breadth

(75)

63

—

—

H. Maxillary diastema

44

43

-

-

I. Gnathion-nuchal crest

407

367

-

-

Kurten and Anderson (1980:293) is conclusive, and most are
doubtful or incorrect. Gazin (1950:403) listed both Tapirus cf.
haysii and T. veroensis for Melbourne, Florida, in his study of
the USNM collection of that fauna, but provided no measure-
ments, and neither we nor Lundelius and Slaughter (1976) have
found evidence in it for a tapir other than T. veroensis. The Moore
Pit, Texas, record (Slaughter, 1966:88) may prove to be valid,
but as yet it is somewhat equivocal in that the M2 (L 27.6, AW
ca. 21, PW 2 1 .7) on which it is based is within the expanded size
range for T. veroensis as conceived by Lundelius and Slaughter
(1976) as well as the observed range for T. copei (Table 2), and
the same beds produced a metapodial smaller than its homolog
in T. terrestris. The Mullen II, Nebraska, tapir was assigned to
Tapiruscf. excelsus, not T. copei, by Schultz et al. (1975:12). We
have been unable to find the El Paso Co., Texas, specimen, re-
ported by Richardson (1907, 1909). Leidy (1859:106), who was
well attuned to the significance of size in this context, stated that
the isolated molar from Natchez, Mississippi, "‘corresponds in
form and size with its homologue in the recent Tapir.” Simpson
(1945:63) reidentified as T. veroensis the tapir from Saber-tooth
Cave, Florida, that he earlier reported as Tapirus cf. haysii. The
Apollo Beach, Florida, specimen (Ray, 1964) seems properly
assigned to T. haysii. but is not demonstrably Rancholabrean.
Other, very late records, notably those for the Hancock Ditch,
Arkansas, and the Lehner Mammoth Site, Arizona, are not so
readily dismissed. These and other more or less reliably founded
records known to us are noted here and are mapped by number
in Fig. 3, proceeding generally east to west, starting with the type
locality.

Benners Estate, north bank of Neuse River, Pamlico Co., 16
mi below New Bern, North Carolina (Fig. 3, locality 1). ANSP
1 1504, the holotype, a right P4, L 26.3, AW 21.2, PW 21.6 (Fig.
2D). Contrary to Lundelius and Slaughter (1976:227), Simpson
did not demonstrate that the tooth "was probably within the size
range of T. veroensis .” In fact, it is a bit large for T. copei of Port
Kennedy, but Simpson’s (1945:66) expectation has been fulfilled
that the observed range in size of teeth for the species as a whole
would be greater than that in the homogeneous sample from Port

Kennedy. Topotypic specimens here assigned to T. haysii in-
clude: USNM 2 1 5062, left M2 or M3, L 28.9, AW 34. 1 . PW 30. 1 ,
Fig. 8E; USNM 215202, left M3 (cast), L 28.6, AW 33.5, PW
30.4; USNM 347270, left M2, L 29.8, AW 24.3, PW 23.0, Fig.
2A; USNM 215061, right M3, L 33.2, AW 24.0, PW 21.4, Fig.
2C; USNM 243692, right M2 (cast), L 33.3, AW 25.1, PW 23.0,
Fig. 2B. Other fragments of teeth, including a posterior loph of
a P4, are of appropriate size for T. haysii. In addition some
specimens in the assemblage are clearly referable to T. veroensis,
but it cannot be asserted that the two occurred together in life.

Fort Barnwell, Craven Co., North Carolina (Fig. 3, locality 2).
NCSM 1 32, right P4, L 26. 1 , AW 20.5, PW 20.5. Richards ( 1 950:
26, 42) reported “teeth” of Tapirus haysii along with remains of
mastodon and a cetacean from a Pleistocene muck overlying the
Yorktown Formation in marl pits on the property of Z. B. Broad-
way, one mi north of Fort Barnwell. We have been able to locate
only the single tooth in NCSM and none in ANSP.

Billy B. Fussell Co., Inc., marl pit, 1 mi south and 0.5 mi west,
of Rose Hill, Duplin Co., North Carolina, 34°48'00"N, 78°01'40"W
(Fig. 3, locality 3). The Pleistocene vertebrate fauna here has
been collected primarily by P. J. Harmatuk from overburden
being cleared in preparation for marl excavation. The fauna as
a whole is under study by R. E. Eshelman and C. E. Ray. Spec-
imens assigned to T. haysii include: USNM 306473, left DP1
(cast), L 21.3, W 21.3, Fig. 8A, B (original specimen); USNM
347321, left P4, L 24.8, AW 30.9 min., PW 30.6; USNM 347321,
left M1, L 26. 1 , AW 29. 1 , PW— ; USNM 347321, left M2, L 29.4,
AW 34.5, PW — ; USNM 347322, left M3, L 28.1, AW 33.1, PW
29.2; USNM 30521 3, left M\ L 28.2, AW 34.2, PW 29.0. USNM
306473 is in a fragment of bone revealing the vestige of a crypt
for P1 beneath the tooth. The three teeth, P4-M2, assigned a single
catalog number are thought to represent a single individual on
the basis of matching pressure facets, compatible wear, and hav-
ing been collected together on the same day in a small area of
spoil. USNM 347322, though collected at the same time by Mr.
Harmatuk, is from an older animal and is slightly different in
coloration. All of these teeth are suitable in size for T. haysii.
Two fragments of crowns from the same locality appear to rep-
resent smaller teeth, but their differing color and character suggest
milk teeth.

Myrtle Beach area, Horry Co., South Carolina (Fig. 3, locality
4). This is a classic area for beach collecting. Material comes from
the vicinity of Garden City Beach, Surfside Beach, and Myrtle
Beach. Little has been done by vertebrate paleontologists re-
garding provenience and faunal associations in these deposits,
but possibilities may exist for obtaining material in place. For-
mations of varied Pleistocene age occur in the area (McCartan
et ah, 1982).

Isolated teeth, mostly somewhat tumbled, polished, and frag-
mented, certainly span the size range of both T. veroensis and T.
haysii. Specimens more or less certainly vouching for occurrence
of T. haysii include: USNM 244383, right M3, L 32.3, AW 24.0,
PW 21.6; USNM 347323, right M,, L 30.8, AW 22.7, PW 21.6.

Savannah, Chatham Co., Georgia (Fig. 3, locality 5). Material
dredged from bottom of Savannah River. The specimens listed
below are among the many vertebrate fossils placed in museum
collections over a period of many years, at least from the 1930’s
to the 1960’s, by the late Ivan R. Tomkins, an avid and able
amateur naturalist, active mostly in the vicinity of Savannah.
Specimens of T. veroensis size are known also from dredgings at
Savannah, but the following are among those thought to represent
T. haysii: ChM 55.103.113, right M3?, L 29.3, AW 34.4, PW

296

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

27.5 min.; ChM 55.103.82, left M, or M2, L 29.4, AW 22.3, PW
21.1.

Phosphoria Mine, IMC, 3.5 mi SW of Bartow, Polk Co., Flor-
ida; See. 2, T 3 IS, R 24E, Bradley Junction 7.5-minute quad-
rangle, USGS (Fig. 3, locality 6). UF 40060, right M2, L 29.0,
AW 33.3, PW 30.3. This well preserved, slightly worn tooth with
roots intact was collected from spoil in a phosphate mine by Ben
Torres, Jr., in 1975.

Apollo Beach, Hillsborough Co., Florida (Fig. 3, locality 7).
UF 8225, fragmented mandible with left M, and right P2-M3;
from dredgings of uncertain age. Reported by Ray (1964). Lun-
delius and Slaughter! 1976:24 1 ) apparently saw only the left man-
dibular fragment, and regarded the tooth as M,.

There are other specimens from Florida that fall within the
size range of T. haysii, but the records thus far are unsatisfactory.
For example, Simpson (1945:64) noted a fragment of right man-
dibular ramus, AMNH 23110, with part of M, and unerupted
M, (L 30.5, AW 21.5, PW 19.6), in the Holmes collection, but
lacking locality data other than Florida. Another specimen is
NCSM 277, a moderately worn right lower tooth, probably M2,
L 30.6, AW 23.0, PW 21.9, from somewhere in Sarasota County.
Undoubtedly more specimens of large Pleistocene tapirs from
Florida are already in collections, or will come to light as new
faunas, especially of pre-Wisconsin age, are found. The richness
of the record in Florida and its southerly latitude make it the
most likely source of the material needed for significant improve-
ment of understanding not only of T. haysii but of fossil tapirs
generally.

Port Kennedy, Montgomery Co., Pennsylvania (Fig. 3, locality
8). This locality provided the classic ANSP collection, still the
only population sample of large Pleistocene tapirs from North
America, and the basis for Simpson’s (1945:66) Tapirus copei.
The fauna is early Irvingtonian in age.

Hanover Quarry, York Co., Pennsylvania (Fig. 3, locality 9).
In 1977 quarrymen recovered vertebrate fossils from a large fis-
sure encountered in the Hanover Quarry. The find was reported
to USNM from which it was referred to John E. Guilday at CM,
who arranged to borrow selected specimens for casting, photog-
raphy, and measurement. Remains of mastodon, horse, and tapir
are known to have been recovered. The tapir material consists
of a partial left mandibular ramus with P2, P3, DP4 M,, and M2,
and a right mandibular fragment with M,. The original specimens
were returned to the finders, who sold them to a private collector,
but casts are housed in CM (nos. 30230 and 30231) and USNM
(nos. 244439, represented in Fig. 2F, and 244440). Although left
and right mandibular rami are given separate numbers, they
almost certainly represent a single individual, judging from size,
details of ossification, and stage of tooth eruption and wear.

All comparable dental dimensions (Table 2) place the specimen
within the observed range for the Port Kennedy sample, and 1 2
of 1 4 measurements are beyond the observed range for Simpson’s
(1945) sample of T. veroensis. We assign the material tentatively
to T. haysii.

Pigeon Creek, near Evansville, Vanderburg Co, Indiana (Fig.
3, locality 10). ANSP 11505, left P4, L 22.1, AW 20.3, PW 22.9.
This tooth is deeply worn, with extensive interdental pressure
facets anteriorly and posteriorly that undoubtedly reduce L sig-
nificantly. On this basis and on that of its great PW (cf. Table
2), it is referred tentatively to T. haysii. The tooth was reported
by Leidy (1855:200), illustrated by him (1859: pi. 17, figs. 9 and
10) and the illustrations repeated by Hay (1912: figs. 38 and 39)

who mistakenly identified the specimen as the type in his figure
caption, but later (Hay, 1927:76) corrected his error.

A radiocarbon date of 9,400 ± 250 years (Rubin and Alex-
ander, 1958: 1484) has been cited repeatedly as the youngest date
for tapir remains in the Quaternary of North America. This date
was obtained from wood said to have been collected in 1870
from the same river alluvial stratum near Evansville from which
Lincke, at least 16 years earlier, collected vertebrate fossils in-
cluding the tapir tooth discussed here. This seems a tenuous
association on which to base any critical conclusion.

Hancock Ditch (ditch no. 39), near Weona, Poinsett Co., Ar-
kansas (Fig. 3, locality 1 1). AAS 69-658, left mandibular fragment
with posterior half of P4, and M,-M3 (Fig. 2E). The M, is deeply
worn and thus secondarily reduced in length. The specimen was
collected by Phyllis Morse on 30 May 1970 with mastodon and
possibly sloth from a clay deposit in a braided stream system
thought to be late Pleistocene in age (Dan F. Morse, personal
communication). The specimen was identified as tapir by the late
C. W. Hibbard and reported by Morse (1970), who also described
the field context.

The measurements of this specimen (Table 2) make its specific
assignment somewhat equivocal. The length of crown in all three
molars is too small for T. haysii, but all transverse dimensions
are quite large enough, and the bony ramus is very robust. We
assign the specimen tentatively to T. haysii. This and other spec-
imens make it very clear that much remains to be learned about
ranges in size and time of North American Pleistocene tapirs.

Mallory Sand and Gravel Pit, near Thedford, Thomas Co.,
Nebraska (Fig. 3, locality 12). Incomplete mandibular ramus with
roots of M2 and M3, recorded as Tapirus sp., “a very large species,
possibly T. copei,” by Schultz et al. (1975:13). They indicated
that early Pleistocene fossils had been collected at the site, but
that the horizon of the tapir was uncertain.

Holloman gravel pit, one mi north of Frederick, Tillman Co.,
Oklahoma (Fig. 3, locality 13). A fragment of right mandibular
ramus with DP3 (L 25, W 18.5), DP4 (L 26, W 21.4), P3 (L 27),
P4 (L 26), and M, (L 27.5, W 23.5), was reported and illustrated
by Hay and Cook (1930: 1 7-18; pi. 4, fig. 3). Their measurements
indicate an animal larger than any in the Port Kennedy sample
(cf. Table 2). Dalquest (1977:262) reported that the specimen is
now in the collection of Midwestern State University. The Hol-
loman local fauna is regarded as earliest Irvingtonian, comparable
in age to the Gilliland local fauna (Dalquest, 1977:256).

Gilliland local fauna, Knox Co., Texas (Fig. 3, locality 14).
Two partial mandibular rami of suitable size for T. copei were
reported by Hibbard and Dalquest (1966:39-40; fig. 5E), who
remarked on the similarity in geologic age of the Gilliland and
Port Kennedy faunas (both indicative of early Kansan, Irving-
tonian age) and on the wide geographic distribution of T. copei
in early middle Pleistocene time.

Campo Grande Arroyo, Hudspeth Co., Texas (Fig. 3, locality

1 5) . A weathered, partial palate of an old individual with deeply
worn, incomplete dentition was reported as Tapirus cf. copei by
Strain (1966:48-50; pi. 13), seemingly on good grounds. The
specimen was found in the lower part of the Camp Rice For-
mation, regarded as Aftonian, late Blancan, in age (Strain, 1966:
21).

Donnelly Ranch, Las Animas Co., Colorado (Fig. 3, locality

16) . Four upper teeth were referred to T. copei on the basis of
their large size by Hager (1974:13), who regarded the fauna as
of pre-Nebraskan, late Blancan age.

1984

RAY AND SANDERS -PLEISTOCENE TAPIRUS

297

Lehner Mammoth Site, near Hereford, Cochise Co., Arizona
(Fig. 3, locality 17). Lance (1959:37) identified as Tapirus sp. a
crushed left mandibular ramus with P3-M3, an isolated right P2,
and a fragment of upper molar, representing a very large tapir.
As Lance pointed out, the Lehner specimen exceeds in some
dental dimensions the upper limit observed in the Port Kennedy
sample of T. copei (cf. Table 2). However, it can be matched or
exceeded in all available dental measurements among specimens
here referred to T. haysii. Radiocarbon dates for the Lehner Site,
hovering about 1 1 ,000 years, provide a seemingly reliable late
date for this large tapir. Lance suggested an adaptation to drier
conditions to explain the occurrence of tapirs in the Quaternary
of the Southwest.

Bautista Creek badlands, northeast of Hemet, Riverside Co.,
California (Fig. 3, locality 18). As already noted, this is the type
locality of T. merriami, a very large tapir known from a single
fragment of mandible with fragments of two teeth (Frick, 1921),
which we regard as probably conspecific with T. haysii. The fauna
is Irvingtonian in age (Kurten and Anderson 1980:26).

Additional records, at least one already in other hands for
study, will be needed to bridge the geographic gaps in distribution
of large Pleistocene tapirs between the Pacific coast and the re-
mainder of the conterminous United States. Of course much
more and better material will be needed to resolve problems of
specific relationships, and better documentation and faunal as-
sociations will be needed to resolve time-stratigraphic distribu-
tion, which is puzzling at present.

Much of the limited evidence now available suggests that T.
haysii is a species at least especially characteristic of, if not limited
to, the early and middle Pleistocene (late Blancan and Irving-
tonian; Savage and Russell, 1983, include all of the Blancan in
the Pliocene). Most of its best-documented occurrences to date
fit this pattern and most other records are permissive. However,
there are a few records (notably the Hancock Ditch, Arkansas,
and Lehner Site, Arizona) that are both very late Pleistocene and
more or less conclusively identified. Hibbard and Dalquest ( 1 966:
40) and Hibbard et al. (1978:38) commented on the suggestively
early (pre-Yarmouth) distribution, and the latter authors also
suggested the possibility of niche partitioning, with T. haysii (=T.
copei) occupying drier, more open country, and T. veroensis,
wetter, more forested habitat. This would be compatible with
Lance’s ( 1959:39) suggested explanation of occurrence at the Leh-
ner Site. Smaller tapirs more or less certainly assigned to T.
veroensis are the common form in Wisconsinan faunas. If the
large tapir of the late Blancan and Irvingtonian is the same species
as that of the terminal Wisconsinan, where was it during most
of Rancholabrean time?

Taxonomic status.— The name Tapirus haysii was
widely and often loosely applied to North American
Pleistocene tapirs until Simpson (1945:66) conclud-
ed “that this type is essentially indeterminate and
that the species to which it belongs, and conse-
quently the species properly called T. haysii, cannot
at present be identified.” He then erected the new
species, T. copei, based on the sample from Port
Kennedy previously referred to T. haysii, and di-
agnosed it as follows: “essentially the species called

T. haysii by Cope (1899) and most later authors.
Larger than any recent species or than any other
North American Pleistocene species except T. mer-
riami. Significantly smaller than T. merriami. P1 a
large, robust, relatively transverse and complex tooth
with the protocone relatively far forward, a large,
heavily ridged basin between this and the ectoloph.
P2 advanced in molarization, difference in anterior
and posterior widths slight, protoloph fully devel-
oped.” He has been followed almost unanimously
in applying the name T. copei to remains of large
tapirs found subsequently in North America.

Simpson recognized the probable fallibility of both
tooth size and premolar characteristics in distin-
guishing species. As discussed elsewhere in this pa-
per, his expectations have been realized. The con-
tinuing poor representation of anterior upper
premolars of large tapirs has meant that size nec-
essarily has been the practical criterion by which T.
copei and T. veroensis have been distinguished in
every study since 1945. Its efficacy has been widely
doubted, but its application continues.

All workers on fossil and modern tapirs have been
appropriately impressed by their conservatism (see
Radinsky, 1965:69, 101). Their differentiation in
Quaternary time is little reflected in their dentition.
The modem occurrence of three species in northern
Colombia should instill caution in all paleontolo-
gists attempting to distinguish species on the basis
of isolated teeth and fragments of jaws. A strict con-
structionist could make the case that only T. ver-
oensis among described North American Pleisto-
cene tapirs is presently identifiable, and even that,
only by virtue of cranial characters.

Thus, in spite of the relatively rich sample, T.
copei could be faulted by standards similar to those
on which T. haysii was rejected, because skull char-
acters are not available. It would be convenient to
have better type specimens for many paleotaxa, but
stability is served best by conserving named taxa if
at all possible. General application of standards
comparable to those on which T. haysii was sup-
planted by T. copei would lead to rejection of many,
perhaps most, nineteenth century types in verte-
brate paleontology. Such summary rejection based
on rigorous criteria would result in chaos. Paleotaxa
are generally more or less provisional pending dis-
covery of new material. Skulls of large tapirs from
the east and skulls of any size from the far west will
be required to place the specific taxonomy on a sound
basis, and these will surely be forthcoming sooner

298

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

or later. Meanwhile, Tapirus haysii seems to be us-
able, pro tem.

Tapirus veroensis Sellards 1918

Skull.— As Simpson (1945:56) observed, Sellards
(1918) established the species securely on the basis
of the very fine holotype skull, now UF/FGS V 277.
Since then several skulls and partial skulls have been
found, including those of the closely related, prob-
ably conspecific T. excelsus. Each of these has stim-
ulated studies contributing in turn to progressively
better understanding of North American Pleisto-
cene tapirs (Simpson, 1945; Parmalee et al., 1969;
Lundelius and Slaughter, 1976). Most recently, Bo-
gan et al. (1980) have reported the major part of a
skull from a cave in Tennessee. Through the kind-
ness of its owner, Mr. Donn Claiborne, and of Paul
F. Parmalee, we have had this skull available for
further study in conjunction with previously unre-
ported new specimens from South Carolina.

In recent years collecting by divers in the rivers
of South Carolina in the vicinity of Charleston has
yielded numerous fine specimens of fossil verte-
brates, many of which have been deposited in the
collections of ChM and SCMC. The finest specimen
known to us, an essentially complete skull, now ChM
PV 4257, was found by Susan A. Wallace on the
bottom of the West Branch of the Cooper River in
Berkeley Co., South Carolina, approximately 20 mi
(32.2 km) due north of Charleston, during the sum-
mer of 1976 (Fig. 4). Miss Wallace, a Charleston
SCUBA-diving instructor, discovered the skull while
diving for fossils near the south bank of an eastward
course of the river at 33°4.4'N, 79°56.2'W (USGS
Kittredge 7. 5 -minute quadrangle), approximately 1 . 1
mi (1.8 km) southeast of Strawberry Landing. Ac-
cording to Miss Wallace, the specimen was found
in nearly 1 0 ft of water on a terrace that gently slopes
to the main channel of the river, which is bedded
in the Ashley Member (upper Oligocene) of the Coo-
per Formation, a highly-indurated, calcareous ma-
rine formation that underlies the entire Charleston
area. The terrace has resulted from subaqueous ero-
sion of overlying sediments from the surface of the
Cooper Formation between the main channel and
the south bank of the river.

Miss Wallace recalls that the skull was lying di-
rectly on the surface of the “marl” (the Cooper For-
mation), where it had come to rest after having been
washed from its matrix. The stratigraphic origin of
this specimen is highly uncertain. Recent studies by

the USGS indicate that in the immediate area of the
discovery the Cooper Formation is unconformably
overlain by early (?) Holocene sediments and that
the nearest Pleistocene deposits are those of the
Wando Formation (upper Pleistocene; McCartan et
al., 1980:114-115), which has been mapped at a
location 0.15 mi (0.2 km) upstream from the tapir
skull site (Robert E. Weems, personal communi-
cation). There being abundant evidence of Tapirus
in the Pleistocene of the eastern United States and
(to our knowledge) none from the Holocene of this
region, we might assume a priori that the skull was
washed out of the aforementioned Wando Forma-
tion deposits and subsequently transported by the
river currents downstream to its point of recovery.
However, Susan Wallace, who has logged many
hours of diving in this stream, has observed that the
currents are swiftest near the surface of the river and
are of such negligible velocity at the bottom that
objects as light as bottles usually remain at their
initial point of contact with the bottom. She con-
siders it quite unlikely that an object of the size of
the tapir skull would have been transported by bot-
tom currents over the distance from the Wando de-
posits upstream. Her opinion is supported by con-
sideration of the excellent condition in which the
skull has been preserved, showing no signs of having
been rolled about on the hard surface of the Cooper
Formation at the bottom of the river. Thus, it seems
almost certain that PV4257 came from a stratum
of the river bank in the immediate vicinity of its
discovery by Miss Wallace. That conclusion sug-
gests an early Holocene date for this specimen based
on USGS studies of the post-Cooper sediments in
the Charleston area. There is, however, a distinct
possibility, if not a probability, that a lens of Wan-
do-age (Sangamonian) sediments is concealed be-
neath the Holocene deposits bordering the river at
the tapir skull site and that it was that stratum that
yielded the specimen (R. E. Weems, personal com-
munication). Until the fact of the matter can be
determined we can state only that PV4257 is of late
Pleistocene or early Holocene age. Whatever its age,
we are extremely grateful to Susan Wallace for mak-
ing this splendid specimen permanently available
for scientific study and for her crucial information
about the stream in which she found it.

A well preserved braincase, also from the Cooper
River, was made available to us by Rudolph Mancke
from the SCMC collections, specimen number SC
75.31.173 (Fig. 5). SCMC also has several maxillae
and mandibles not utilized in this study, but which

1984

RAY AND SANDERS -PLEISTOCENE TAPIRUS

299

Fig. 4.— Skull of Tapirus veroensis from Cooper River, South Carolina, ChM PV4257, in dorsal (A), lateral (B), and ventral (C) aspects,
x0.4.

300

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

constitute a very useful population sample for future
study.

The Claiborne Cave (Fig. 6) and Cooper River
specimens provide the basis for further refinement
of our understanding of skull characters in T. ver-
oensis, in particular with regard to ontogenetic and
individual variation. There is now a fairly contin-
uous ontogenetic series of skulls available, beginning
with the holotype of T. excelsus, a juvenile retaining
all deciduous premolars, and with M1 in use and M2
unerupted (Simpson, 1945:71). A fragmented, in-
complete skull from Branford, Florida, UF 14056,
had P'-P3 in use, DP4 retained but well worn with
advanced P4 underneath, M1 slightly worn, and un-
worn M2 above the alveolar margin but not fully
erupted (this specimen is dentally younger than sug-
gested in the caption to Fig. IB of Lundelius and
Slaughter, 1976:228, and is recorded there under
the catalog number for an Itchtucknee River spec-
imen). The dentition of the Claiborne Cave, Ten-
nessee, specimen is at almost exactly the same stage
of wear and development (Fig. 8F). DP4 is the only
cheektooth in which the enamel has been breeched
by wear. The incompletely formed crown of M3 is
visible through a small window in its small crypt.
Next in dental ontogenetic stage are the skulls from
Crankshaft Cave, Missouri (Parmalee et al., 1969),
and from the Cooper River, South Carolina, re-
ported here (Fig. 7). These two skulls have perma-
nent P'-M2 fully erupted and worn, with M3 well
formed and visible through a large window in its
large crypt. Finally, the holotype of T. veroensis is
a mature skull with all adult teeth fully erupted and
worn. The skull from the Livingston Dam Site, Tex-
as, is similar in age or very slightly younger (Lun-
delius and Slaughter, 1976:228). This series is es-
pecially valuable for interpretation of development
of the temporal crests.

In the discussion of the skull that follows, primary
attention is given to the fine skull from the Cooper
River, ChM PV4257, with comparison to other
specimens and commentary on previous studies as
appropriate. Several excellent studies of Quaternary
tapir skulls have been published. Those found es-
pecially useful for our purposes are Hatcher (1896),
Sellards (1918), Simpson ( 1 945), Colbert and Hooi-
jer (1953), Hershkovitz (1954), and Lundelius and
Slaughter (1976). The work of Radinsky (for ex-
ample, 1 965), focused primarily on earlier tapiroids,
has been useful especially for the broader context.
In testimony to the perceptiveness of these investi-
gators, we have found few characters in the tapir

5 CM

Fig. 5.— Cranium of Tapirus veroensis from Cooper River, South
Carolina, SCMC SC 75.31.173, in dorsal (A), lateral (B), and
ventral (C) aspects.

skull that had not been thoroughly described or at
least touched upon by our predecesors. Thus, orig-
inality is claimed for little that follows, even though,
for brevity, prior work is cited generally only where
necessary, for example where our observations are
contradictory or supplementary. Discussion of skull
characteristics is organized according to the normal
aspects of the skull in which they are most clearly
revealed, in the sequence dorsal (Figs. 4A, 5A, 6A),
lateral (Figs. 4B, 5B, 6B), and ventral (Figs. 4C, 5C,
6C), in each case proceeding from anterior to pos-
terior.

Hershkovitz (1954:469) indicated that the living
Neotropical tapirs, with the possible partial excep-

1984

RAY AND SANDERS -PLEISTOCENE TAPIRUS

301

Fig. 6. — Partial skull of Tapirus veroensis from Claiborne Cave, Tennessee, private collection of Donn Claiborne, in dorsal (A), lateral
(B), and ventral (C) aspects.

tion of T. bairdii, have reached their definitive adult
size in terms of skull length by the time that M2 has
erupted. The only two skulls of T. veroensis on which
Hershkovitz’s measurement is possible do not fit

this pattern (Table 3). The Cooper River specimen,
with M2 in use is 367 mm long, and the holotype,
with M3 in use, is 407 mm long. In handling nu-
merous specimens of fossil (mostly fragmentary) and

302

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 7. — Dentition of Tapirus veroensis from Cooper River, South Carolina, ChM PV4257, in palatal aspect, x 1.

1984

RAY AND SANDERS- PLEISTOCENE TAP1RUS

303

modern tapirs through the years, one of us (Ray)
has formed the subjective impression that speci-
mens with all permanent teeth in use except M3 and
M3 are disproportionately well represented in col-
lections. This further suggests that this is a period
(young adulthood) in the life cycle either of unusual
vulnerability or of slower dental development (thus
representing a longer opportunity for sampling).

In dorsal aspect (Fig. 4A), the posterodorsally as-
cending process of the premaxilla terminates pos-
teriorly in an acute point above the anterior margin
of P1, as in the holotype. The details of shape and
posterior extent of this process are variable in the
living species, but are generally similar to T. ver-
oensis in T. terrestris, T. pinchaque, and T. indicus.
T. bairdii is dissimilar to all in having a shorter
posterior process and in having its medial border
oriented highly obliquely, resulting in a bluntly
transverse posterior termination. Hatcher (1896:
172-173) used these differences as key characters in
distinguishing bairdii (with dowi I) generically from
the other living tapirs.

The maxilla is exposed narrowly posteromedial
to the posterior process of the premaxilla, less broadly
than in the holotype. This is the “superior branch
of the maxillary” of Hatcher (1896:173), but its
characteristics do not seem to be as profoundly di-
agnostic as implied in his key.

There is no suggestion of a dorsal flange of the
maxilla for embracement of the vertical meseth-
moid cartilage as seen in extreme development in
T. bairdii, in which this flange rises dorsomedially
to roof over the entire bony muzzle posteriorly as
far as the infraorbital foramen, leaving only a uni-
formly narrow sagittal slit occupied by the meseth-
moid. This flange is fully formed even in very young
animals, and in older animals may coossify in part
with the mesethmoid (for example, ANSP 18873,
USNM 6938, 14217, and 283704). Colbert and
Hooijer (1953:84) mistakenly stated that this flange
is absent in American tapirs, whereas it finds its
extreme development in T. bairdii, in which the
flange develops a vertical, planar, medial border to
fit against the mesethmoid. The flange is modestly
developed in T. indicus and in T. augustus. In T.
indicus it is developed only anteriorly, medial to the
posterior process of the premaxilla, as a low, vertical
flange with a feathered or serrated edge. A similar
flange is weakly suggested in some individuals of T.
terrestris, in which a slight, feather-edged, upturned
flange is developed on the maxilla medial to the
posterior process of the premaxilla (for example,

USNM 239966 and 406847, neither superannuated
individuals). Most individuals of T. terrestris, how-
ever, show little or none of the maxilla medial to
the posterior process of the premaxilla. We have
seen no suggestion of a medial flange in any indi-
vidual of T. pinchaque, which most closely resem-
bles T. veroensis in this region.

Posterior to the termination of the premaxilla the
dorsomedial border of the maxilla is smoothly
rounded, directed medially, and posteriorly diver-
gent. It is more extensive in the holotype, possibly
related to greater individual age. T. indicus is most
like T. veroensis in this region, differing primarily
in the tendency to develop a very slight ventrally
directed edge, not a rolled edge. T. bairdii is of course
very different, as described above. In T. terrestris
and to a lesser degree in T. pinchaque this maxillary
border is rolled ventrally and the laterally adjacent
body of the maxilla is modestly inflated, to define
together a semi-enclosed chamber on either side of
the bony snout anterior to the infraorbital foramen.

A long, strong, posterodorsally ascending process
of the maxilla borders the narial aperture along its
anteroventral half, anterior to the orbit. This process
is widely exposed dorsaliy, medial to the anterior
supraorbital flange of the frontal and lateral to the
anteroventrally descending process of the nasal,
which borders the narial aperture along its postero-
dorsal half. The ascending process of the maxilla is
very similar to that of T. veroensis in T. terrestris,
T. pinchaque, and T. indicus, although in T. terres-
tris it tends to twist more strongly laterally in its
dorsal portion where it borders the descending pro-
cess of the nasal and forms the medial wall of the
lateral groove. T. bairdii is very different in that the
posterodorsally ascending process of the maxilla has
no broad, anterodorsally exposed surface, but is in-
stead a vertically oriented lamina lying almost en-
tirely inside the narial aperture, of which it forms
the lateral border along with the frontal. A descend-
ing process of the frontal takes the place of that of
the nasal, which is usually absent in T. bairdii (ob-
served on the right side only of one young individ-
ual, USNM 102522).

A detail of the ascending process of the maxilla
seemingly unique to T. veroensis is the presence of
two well marked creases extending from above the
infraorbital foramen posterodorsally, anterodorsal
and medial to the lacrimal bone. In T. indicus this
area is occupied by a single, wide trough, not two
grooves. In T. bairdii there is generally a single lat-
eral crease developed on the lacrimal and extending

304

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

posterodorsally between the frontal and the ascend-
ing process of the maxilla. There are no similar
creases in T. terrestris and T. pinchaque.

Everyone who has worked with tapir skulls has
devoted attention appropriately to the conspicuous
groove or channel which accommodates the dor-
solateral nasal diverticulum on each side in life. This
groove is surely important, but highly variable and
difficult to characterize. In the Cooper River spec-
imen it is very similar to that in the holotype of T.
veroensis. In plan, each groove resembles nothing
so much as an inverted bass clef of the musical staff.
The anterior outlet lies on the posteriorly ascending
process of the maxilla on either side of the cavernous
narial aperture. From there it passes posterolaterally
onto the supraorbital flange of the frontal bone,
thence posteromedially onto the dorsal table of the
frontal bone, on which lies its greatest part and on
which it swirls posteromedially and then anteriorly
to terminate on the dorsal surface of the nasal at its
posterior end. Efforts have been made to use this
scroll in taxonomic distinctions, usually based on
depth and position of the groove, proximity of ap-
proach to one another at the midline, and extent to
which the termination lies on the frontal or nasal
bone. In spite of great variation, some general fea-
tures or tendencies seem usable. In T. veroensis the
scrolls approach one another very closely at the mid-
line (but do not meet as Simpson, 1945:74, assert-
ed), especially in SCMC SC 75.31.173 (Fig. 5A). In
this they resemble T. bairdii and T. indicus, and
differ from T. terrestris and T. pinchaque, in which
they are widely separated.

The depth of the groove is highly variable in mod-
em tapirs, and somewhat so in the few specimens
of T. veroensis. Throughout its length it is somewhat
more deeply impressed in the holotype than in ChM
PV4257, in which it is shallowly but distinctly de-
marcated except at its anterointernal termination
on the nasal, where it fades away anteriorly. In SCMC
SC 75.31.173 it is strongly impressed as far as pre-
served, especially laterally, where it forms a deep
channel supraorbitally. T. bairdii exhibits extraor-
dinary variation in this region. One specimen,
USNM 179064, has no lateral groove at all, and
only a faint depression on the frontal and nasal where
the internal swirl should be. Other individuals col-
lected at the same time and in the same area have
a very deep groove. Two extremes of development
of the lateral groove may be observed in two skulls
taken at the same time and place in Costa Rica; one
a young adult with M2 erupted, USNM 14218, has

no lateral border at all, making the feature a lateral
shelf (more as in T. indicus ), not a groove; the other,
an adult with M3 in place, USNM 14219, has a high
lateral wall formed by the supraorbital flange of the
frontal, creating a deep lateral groove. The few sexed
specimens available give no suggestion of sexual
dimorphism in this character, nor does age seem to
be a factor. Presence or absence of the bony wall
simply may not be functionally significant.

T. indicus never has a lateral, supraorbital wall to
the lateral groove, which is developed as a supraor-
bital shelf. The groove terminates internally in T.
indicus in a very deep, rounded cup at the end of
the scroll, developed mainly on the posterior part
of the nasal.

The nasals are incomplete in ChM PV4257, but
the dorsoventral thinning anteriorly of the left nasal
suggests that very little is missing from its anterior
tip. To the extent preserved, it is very similar to the
nasals of the holotypes of T. veroensis and T. ex-
celsus. In all of these, the outline of the complete
paired nasals would resemble a stylized St. Valen-
tine’s Day heart. As Hatcher (1896: pi. 3, figs. A-
E) and Simpson (1945:49-50, 73) have shown, the
nasals of modem tapirs are exceedingly variable,
and difficult to use taxonomically. Our observations
confirm this at least for T. bairdii and T. terrestris.
In T. bairdii some individuals have only vestigial
nasals; some have asymmetrical nasals ossified from
multiple centers; some old individuals have long
nasals that blend indistinguishably into the ossified
mesethmoid (for example, USNM 6938, 14217, and
283704). In some individuals of T. bairdii there is
no indication of a nasal at all, its place being taken
apparently by an anterior projection of the frontal.
This condition was observed both in older (USNM
1 1 636 and 14485) and younger individuals (USNM
11281 and 266851) and in one apparent newborn
(USNM 21088). In view of the variability in this
region in T. bairdii, it would not be surprising if
those individuals had “floating” nasal ossifications
that were lost in preparation of the skulls. In T.
terrestris, nasal length varies greatly, at least from
80 mm (USNM 270353) to 1 18 mm (ANSP 2982)
in adults. Nevertheless, some tendencies may be
usable. For example, all T. pinchaque skulls known
to us have very long, slender nasals; all T. indicus
have uniformly short, lyriform nasals. The few ob-
servations possible thus far suggest that T. veroensis
consistently has shorter nasals than any other Amer-
ican tapir.

Hershkovitz (1954:470, 478, 489) described the

1984

RAY AND SANDERS- PLEISTOCENE TAPIRUS

305

“descending sigmoid process” of the nasal in the

living Neotropical tapirs. This is the process de-
scending anteroventrally along the upper part of the
lateral margin of the narial aperture, alluded to pre-
viously in discussion of the ascending process of the
maxilla. It is present and extends forward as far as
the anterior margin of the orbit in T. veroensis. It
is generally similar in T. pinchaque and T. terrestris,
although in at least one juvenile T. terrestris (ANSP
4725) it appears that the process comes from the
frontal, as it does in some T. indicus (ANSP 2985)
and normally in T. bairdii. In one juvenile T. bairdii
there is a short descending nasal process on the right
side only (USNM 102522).

The extent of the dorsal table of the frontal bones
exposed anteromedial to the anteriorly diverging
temporal lines or crests has been brought forward
as a taxonomic character. All available specimens
of T. veroensis resemble one another in this area,
and are most similar to T. pinchaque among living
tapirs. In both species the temporal lines of adults
converge posteriorly approximately at the midsagit-
tal frontoparietal boundary. From this point for-
ward, the temporal lines diverge more obliquely than
in T. indicus and much more so than in T. bairdii,
in which they are subparallel. In younger individuals
of T. veroensis in which the temporal crests have
not converged to approximate a sagittal crest this
dorsal table continues posteriorly on the parietals,
as it does through life in T. indicus and T. bairdii.
T. terrestris is grossly dissimilar in this region even
in juvenile animals, owing to its extraordinary and
precocious development of a high, thick, true sag-
ittal crest, which extends forward onto the frontals,
leaving only a very small dorsal frontal exposure
anteromedial to the divergence of the temporal crests.
Perhaps correlated with these differences, T. terres-
tris has its minimum postorbital width of the fron-
tals far anteriorly, whereas T. veroensis has it far
posteriorly, adjacent to the braincase, as is the ten-
dency in living tapirs other than T. terrestris.

The living species of tapirs differ greatly from one
another in characteristics of the sagittal crest or
parasagittal lines or crests. These features are well
known and thoroughly described and illustrated in
Hatcher (1896), Simpson (1945), and Hershkovitz
(1954).

In “normal” mammals, for example in many car-
nivores, it can be confidently assumed that temporal
crests will approach one another more closely with
increasing age of an individual and that a sagittal
crest, if any, will form and increase in accentuation

in adulthood. Modem tapirs are exceptional in that
juveniles tend to look like miniature adults not only
in general (Simpson, 1945:50), but specifically in
this feature. T. terrestris develops its strikingly high,
long, thick true sagittal crest early in life. In contrast,
T. bairdii and T. indicus never develop a sagittal
crest, but retain a low, broad dorsal table between
widely separated parasagittal crests throughout life.
Adult individuals of T. pinchaque have a low, rel-
atively short sagittal crest, developed only on the
parietals. Only recently has information been pub-
lished on the condition in youth, a single individual
with M2 already erupted (Lundelius and Slaughter,
1976:fig. 1C). In a young female skull of T. pin-
chaque (ANSP 19161) with all deciduous premolars
and M1 in place, the sagittal crest is fully formed
across the fused posterior half of the parietals, and
on the anterior half closely approximated parasag-
ittal crests are narrowly separated by a groove. In
an adult male (ANSP 1 9 1 60) with M3 near eruption
the parietals are not completely fused in the midline
and have a sagittal groove between them; the sagittal
crest is lower and shorter in anteroposterior extent
than in the younger female.

Simpson (1945) founded T. excelsus on the basis
of a juvenile skull which differed most profoundly
from the holotype of T. veroensis in its widely sep-
arated, subparallel temporal crests separated by a
broad intertemporal table. The species was named
on the basis of rigorous application of sound pro-
cedure, using analogy with the most closely related
taxa, in this case congeneric. Of the three adequately
known living relatives, one showed that, if a sagittal
crest was to develop at all, it did so very early, and
two showed that retention through adulthood of
widely separated temporal crests could be expected.
Only with subsequent accumulation and analysis of
specimens of various ages did it become clear that
this region of the skull in T. veroensis (and probably
in T. pinchaque) developed as in a “normal” mam-
mal, and not as in the well known species of Tapirus.
Lundelius and Slaughter (1976) thus made a con-
vincing case for regarding the holotype of T. excelsus
as a juvenile of T. veroensis, and our material cor-
roborates their interpretation. In the two specimens
from South Carolina the segments of the temporal
crests on the parietal bones approach one another
along the sagittal plane and virtually touch adjacent
to the open suture between the parietals, as in the
skull from Texas (Lundelius and Slaughter, 1976:
237; fig. 3B). In the holotype of T. veroensis the
parietals are coossified sagitally and the crest is es-

306

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

sentially sagittal. In successively younger individ-
uals described and illustrated by Lundelius and
Slaughter (1976) the crests are successively farther
apart, as would be expected. The skull from Clai-
borne Cave, Tennessee, fits this regression in that
its temporal crest condition and its dental age (Figs.
6A and 8F) are similar to those of the specimen
from Florida illustrated by Lundelius and Slaughter
( 1 976:fig. 1 B). Thus the known skulls of T. veroensis
show complete concordance in dental development
and temporal crest convergence. Some degree of in-
dividual variation undoubtedly will be evinced in
future finds.

Interparietal bones are notorious in mammals for
their variable incidence, size, number, and shape.
Their occurrence is difficult or impossible to deter-
mine in mature skulls. Among modern tapir skulls
studied in ANSP and USNM, none of the six skulls
of T. pinchaque show a demonstrable interparietal,
although ANSP 19160, an adult male with M3 near
eruption, seems likely to have shown one at an ear-
lier age. Lundelius and Slaughter (1976:fig. 1C) found
a free, triangular interparietal in a young T. pin-
chaque. Of 46 skulls of T. terrestris with the appro-
priate part of the skull intact, one (USNM 220915)
revealed the presence of an interparietal. This spec-
imen has DP' 4 in place, and M1 at the point of
eruption. The small, rounded interparietal is man-
ifested externally as a small irregular oval (17.1 mm
long, 1 1 mm wide) exposed on top of the already
well-formed, high, narrow sagittal crest near its pos-
terior end. A second, very young specimen, with
only DP1-3 in use (USNM 143861) shows a pair of
incomplete sutures, bilaterally situated, extending
posteromedially from each of the parietal-supra-
occipital sutures. These strongly imply the existence
of a discrete interparietal bone at a very early age.
Of 60 skulls of T. bairdii with the supraoccipital-
parietal region preserved, 10 show well-defined in-
terparietals. Of 10 T. indicus skulls available, one
(ANSP 12424), with P1-3 in place but unworn, DP4
and M1 worn, and M2 almost fully erupted, has a
completely separate, triangular interparietal. A sec-
ond skull (USNM 267510) has an open transverse
suture along the anterior edge of the supraoccipital
and an open oblique suture on the left side, with the
parietal, and continuous anteriorly with the sagittal
suture between parietals, but not matched by a cor-
responding suture on the right side. This may in-
dicate an independent interparietal at an earlier age,
but the sagittal sutures on the skull roof tend to
wander in T. indicus.

Thus, among living tapirs, interparietals were ob-

served in order of decreasing frequency in T. bairdii
(10 of 60), T. indicus (1 of 10, plus one possible),
T. terrestris (1 of 46, plus one probable), and T.
pinchaque (none of 6, one possible). These figures
do not reflect true incidence of the bone, as detection
is generally dependent on immaturity and open su-
tures.

Such is not the case in T. veroensis, in which every
known skull except the holotype shows a discrete
interparietal, and the holotype very likely had one
in youth as well (Lundelius and Slaughter, 1976:
237). These include the holotype of T. excelsus (with
two interparietals) described by Simpson (1945:74),
the Texas skull reported by Lundelius and Slaughter
(1976:237; fig. 3B), the specimen from Crankshaft
Cave, Missouri (Lundelius and Slaughter, 1976:237;
fig. ID), the specimen from Branford, Florida (UF
14056), the Claiborne Cave, Tennessee, skull (Fig.
6A), and both specimens from South Carolina (Figs.
4A and 5A). The bone is irregularly rhomboidal in
outline in some specimens (see Lundelius and
Slaughter, 1976, figs. 1A and 3B), but isosceles-tri-
angular in most (see Lundelius and Slaughter, 1976:
fig. ID, and our Figs. 4A, 5 A, 6 A). In the Claiborne
Cave specimen, the base of the triangle is 33.5 mm
long and each leg approximately 40 mm; in ChM
PV4257, 15.5 and 30 mm, respectively; in SCMC
SC 75.31.173, 17 and 25 mm; and in UF 14056, 28
and 34 mm. The high incidence of the interparietal
and retention of independence relatively late in on-
togeny would seem to be of some taxonomic value
for T. veroensis.

The lambdoidal crests are strongly developed and
posteriorly projecting in both specimens from the
Cooper River, more laterally projecting in the Clai-
borne Cave specimen and in the holotype. Their
profile in dorsal aspect is U-shaped or W-shaped,
varying from deep and narrow (Fig. 5A) to shallow
and wide (Fig. 6A). We have been unable to define
characters of taxonomic value in the lambdoidal
crests (see Simpson, 1945:74).

In lateral aspect (Fig. 4B) the third incisor is slight-
ly opisthodont in relation to the plane of the mo-
lariform toothrow (see Hershkovitz, 1954:470, 478,
489). As in T. pinchaque and most T. terrestris the
maxilla is not visible in lateral aspect above the
posterior ascending process of the premaxilla,
whereas it is visible as an elevated, serrated flange
in all T. indicus and weakly in some T. terrestris.
The blunt termination of the premaxillary process
in T. bairdii makes this relationship irrelevant in
that species.

The lateral border of the posterior premaxillary

1984

RAY AND SANDERS -PLEISTOCENE TAPIRUS

307

process descends anteroventrally to a point clearly
anterior to the alveolar margin of the canine, as in
all living tapirs except 7. indicus in which it essen-
tially intersects the lateral margin of the alveolar
border (see Hatcher, 1896:173; pi. 5, fig. 3), as it
does also in T. augustus (see Colbert and Hooijer,
1953: pi. 18, fig. 1).

The lacrimal is broadly exposed as a flattened
quadrangle on the face anterior to the orbit. The
dorsal border does not participate in the lateral
groove for the nasal diverticulum but is overlain by
the anterior extremity of the supraorbital flange of
the frontal. The anterior margin is applied to a thin
crest of the maxilla that interrupts in lateral aspect
the profile of the otherwise smooth margin of the
narial vacuity. This crest continues ventrally as a
line behind the infraorbital foramen, thence curving
posteroventrally to connect with a crest formed along
the ventral margin of the jugal. The lacrimal has a
broad, flattened preorbital process lying across the
middle of the orbital border of the facial part of the
lacrimal. Above this process, and partially con-
cealed by it in lateral aspect is a large lacrimal fo-
ramen; below it, and clearly visible in lateral aspect
is a second, smaller lacrimal foramen.

The lacrimal appears to be distinctive in T. ver-
oensis and in each living species of tapir, although
T. terrestris and T. pinchaque are most similar to
one another, and very different from all others. In
T. terrestris the facial exposure of the lacrimal is
high and narrow and rugose. Most individuals have
a slender, strongly developed, pointed or knobby
lacrimal process on the orbital border, and a second,
similar process, subequal in size, arising near the
anterior edge of the lacrimal. Typically there are two
rather small lacrimal foramina. The larger, more
dorsal one generally lies just inside the orbit, dor-
se media I to the preorbital lacrimal process, where
it may or may not be visible laterally. The smaller,
more ventral foramen lies ventral to the process and
is visible in lateral aspect. <

The lacrimal of T. pinchaque is similar to that of
T. terrestris in shape and extent of facial exposure
and in development of a spicular preorbital process.
There is a tendency to develop a second, anterior
process, and rugose sculpturing, but neither as
strongly marked as in T. terrestris. The lacrimal
foramina are similar in size and placement to those
of T. terrestris, but strikingly variable in the small
series available. Most individuals exhibit three fo-
ramina, of which two lie inside the orbital margin.

In 7. bairdii the lacrimal is extensively exposed
on the face, but the exposure is high and narrow,

and narrows conspicuously dorsally. There is a sin-
gle broad lacrimal process, and generally two foram-
ina—a smaller, more ventral one, usually visible in
lateral aspect on the orbital border below the pro-
cess; an extremely large dorsomedial one situated
inside the orbit and usually concealed from lateral
view behind the broad lacrimal process.

In T. indicus the lacrimal is only narrowly ex-
posed facially, and much of that exposure is occu-
pied by two rather large, shallowly interconnecting
foramina that are visible in lateral aspect.

In T. veroensis the dorsal profile of the cranium,
formed by the upper margins of the nasals, frontals,
parietals, and supraoccipital, is low and smoothly
curved. The nasals are set down slightly below the
level of the frontals, but their profiles merge without
an abrupt step. Behind the nasals the profile is very
low and gently convex throughout. In the holotype
the entire dorsal profile is more nearly rectilinear.
The dorsal profile of T. pinchaque is most like that
of T. veroensis, although flatter and with longer na-
sals. In both T. veroensis and T. pinchaque the dor-
sal profile is roughly parallel to the basicranial pro-
file (Figs. 4B, 5B, 6B).

In T. indicus and T. bairdii the nasals generally
are set down abruptly below the frontal profile, and
particularly in T. indicus the dorsal table of the fron-
tal is inflated, giving the skull a “forehead,” and
making that the highest point of the dorsal profile
above the basicranial line.

In T. terrestris the nasals are usually set well below
the level of the maximum elevation of the dorsal
profile, which is generally very high and strongly
convex or ascending posteriorly, in either case high-
ly elevated above the basicranium.

In T. veroensis the occlusal line of the cheek tooth
row ascends slightly anteriorly in relation to the ba-
sicranium, which is taken as the horizontal line. This
is the typical condition in all species examined other
than T. terrestris, in which the ascent tends to be
steeper, resulting in an upward flexure of the facial
upon the cranial part of the skull. This effect is ac-
centuated by the high, convex dorsal profile of the
cranium.

The postglenoid and mastoid processes of T. ver-
oensis converge, but do not touch, below the ven-
trally open, rounded externa! auditory meatus, de-
scribing a keyhole outline in lateral aspect. In T.
bairdii the meatus is rounded and generally con-
stricted ventrally in a narrow neck by complemen-
tary pointed processes from the postglenoid and
mastoid processes. In T. indicus the postglenoid
process curves posteriorly but the anterior edge of

308

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

the mastoid process is fairly straight and vertical,
leaving a rather wide ventral outlet from the audi-
tory meatus. T. augustus seems to be the most dis-
tinctive in this region, having the two processes
closely approaching or touching one another (Col-
bert and Hooijer, 1953:87). T. terrestris is highly
variable in this region. In many skulls the meatus
is widely open ventrally, whereas in others (for ex-
ample, USNM 281390) the mastoid process sends
a sharply pointed process forward to constrict the
opening severely.

We have found very little that is distinctive in the
ventral aspect of the skulls (Figs. 4C, 5C, 6C) other
than characters of the dentition (Figs. 7 and 8F).
The third incisors appear to be only slightly larger
than the first two or the canine, judging from the
alveoli of these latter teeth. The alveolus for the
canine however, is more nearly circular in cross-
section, whereas those of the incisors are elongated
labiolingually. Sellards (1918:60) felt that the palate
of T. veroensis was more concave in transverse sec-
tion than that of the living species, but his com-
parisons were with single specimens, and, as Simp-
son (1945:57) noted, this is a highly variable
character. The choanae terminate on a line with the
posterior half of M2, and a little farther posteriorly
in the holotype. The choanal margin on the ptery-
goids is slightly constricted posteriorly in the ho-
lotype (Sellards, 1918:60), but not in the Cooper
River skull.

The characteristics of P1 and P2 have been dis-
cussed in detail by Simpson (1945) and others. The
essential features are those of degree of molariza-
tion, in which T. veroensis in general resembles T.
terrestris. Hershkovitz (1954: fig. 60) illustrated typ-
ical Pi’s in T. terrestris and T. pinchaque, but noted
that P1 in some individuals of T. pinchaque is sim-
ilar to that of T. terrestris. One individual of T.
terrestris observed in this study, an adult female
(ANSP 12476), has the most highly molariform P1
of any specimen of tapir known to us. The tooth is
subquadrangular in outline and has a large, high
hypocone (?) almost as large as the protocone.

Measurements. — Problems in distinguishing T. veroensis and
T. haysii on the basis of size of molariform teeth have been
discussed under this heading for T. haysii. Dimensions for sam-
ples of each and for selected specimens are given in Tables 1-3.
Dimensions of other specimens, mostly isolated teeth, are given
in the text where the specimen in question is discussed. Addi-
tional statistical analyses of unified, well-documented samples of
dentitions, already available for T. veroensis and much needed
for T. haysii. undoubtedly will aid in clarification of temporal,
geographic, and taxonomic variation.

Distribution. — IX has not been feasible to reexamine by any
means all previously reported specimens or to seek out all new
or overlooked material. Nevertheless, the broad pattern of dis-
tribution of the smaller tapirs, of T. veroensis size, is reasonably
well known from the surveys by Hay (1923:203-210), Simpson
(1945:53-55), Schultz et al. (1975:17), and Bogan et al. (1980:
1 1, for Tennessee). Many records are based on isolated teeth that
are specifically identifiable not by morphology but by default,
that is, on the assumption that there is but one species of smaller
tapir in the Pleistocene of North America. Until a comprehen-
sive, critical review is undertaken, we have not felt that a new
map is warranted. Accordingly, we have concentrated on new,
overlooked, or obscure records known to us that either extend
the periphery of, or fill blanks in, the previously mapped distri-
bution. These records are discussed below in geographic order
from west to east, and north to south.

Carthage, Hancock Co., Illinois. ANSP 1 1 506, a right M3, L
25.9, AW 29.3, PW 25.9. The specimen consists of a moderately
worn, well preserved crown, lacking roots (Fig. 8C), attributed
to “Dr. G. M. Hall,” At the existing stage of wear, a posterior
pressure facet would be expected if the tooth were M2. It compares
closely to M3 in the holotype of T. veroensis in size, configuration,
and wear. In the Proceedings of ANSP for 1867 under “Donations
to the Museum” is listed “Hall, Geo. W., M.D. Fossil Tooth of
a Tapir, from Illinois.” Leidy ( 1 869:39 1 ) included Illinois among
the states in which Tapirus americanus was said to occur, pre-
sumably in allusion to this tooth. The specimen seems otherwise
not to have been noticed in the literature on North American
Pleistocene tapirs, nor to have been listed among the fossil mam-
mals of Illinois (Bader and Techter, 1959).

Hollidaysburg, Blair Co., Pennsylvania. Well preserved juve-
nile partial dentition identified as Tapirus veroensis. The tapir is
part of a rich late Wisconsinan fauna preserved in a fissure filling
in the same hillside as the classic Frankstown Cave fauna (Guil-
day, personal communication). The fauna is being collected and
studied by Dr. Shirley Fonda, and has been recorded in print as
yet only briefly (Fonda, 1982).

Gebhards Cave, near Schoharie, Schoharie Co., New York.
NYSM V-40, the major part of the crown of a left upper molar,
either M2 or M3, little or not at all worn in life but rolled and
polished postmortem, and lacking the basal lingual margin of the
enamel, L 24.0, AW ca. 26.6, PW ca. 24.5 (Fig. 8D). The tooth
and a supposed bone fragment were found in the bed of a small
stream in the cave by Dr. Richard Veenfliet, Jr., who brought
them to the New York State Museum. They were submitted to
W. K. Gregory and to C. L. Gazin, who identified the tooth as
tapir. The second object is not now to be found in the collections,
and probably was discarded, as there is a note with the tooth,
dated 31 October 1956, indicating that “the so called vertebra
is inorganic and is a piece of chert.” The tooth was donated to
NYSM in 1962 by Mrs. Milton Burgess, Dr. Veenfliet’s daughter.
The specimen was recorded by Moodie (1933:1 15) in an adden-
dum to a popular handbook, but has not previously found its
way into the literature on fossil tapirs.

Five Fathom Bank, 25 miSSE of Cape May, New Jersey. ANSP
15270, a fragmentary right mandibular ramus with alveoli of P4-
M3, with vestiges of their roots, in part freshly broken (Fig. 9B).
The size of the jaw is suggestive of T. veroensis but is inconclusive.
No meaningful measurements are possible as the alveolar mar-
gins are broken away. The specimen was dredged from the sea
bottom in the summer of 1956 by Preston Hawk of Rio Grande,
New Jersey, and recorded by Richards (1957), who also alluded
to it less explicitly later (Richards, 1959).

1984

RAY AND SANDERS -PLEISTOCENE TAPIRUS

309

Fig. 8.— Upper dentition of Tapirus haysii (A, B, E) and T. veroensis (C, D, F) in occiusal aspect (except B), x l. A and B) left DP1 in
occlusal and lingual aspects, original specimen whitened for photography, USNM 306473 (cast), Fussell’s Pit, North Carolina; C) ANSP
11506, right M3, Carthage, Illinois; D) NYSM V-4Q, left M2 or M3, Gebhards Cave, New York; E) USNM 215062, left M3, Neuse
River, North Carolina; F) private collection of Donn Claiborne, left P‘-P3, DP4, M1, M2, Claiborne Cave, Tennessee.

Neither Simpson (1945: fig. 5) nor Schultz et al. (1975: fig. 12)
mapped any localities for Pleistocene tapirs in West Virginia or
North Carolina, and but a single locality in Virginia, presumably
from Cope’s Wythe Co. collection, restudied by Guilday (1962).
Tapir material is uncommon in the otherwise rich Pleistocene
cave and fissure faunas of West Virginia and Virginia, and con-
sists of isolated and mostly fragmentary cheek teeth. From West

Virginia, Garton (1979:113) has recorded tapir from Buckeye
Creek Cave, Greenbrier Co., and from the Bowden Cave system,
Randolph Co. There is also an isolated tooth in the USNM from
Wormhole Cave, Pendleton Co. From Virginia, in addition to
the Wythe Co. occurrence, Clark (1938) recorded tapir from a
fissure in a quarry at Limeion, 6 mi south of Front Royal, Warren
Co. The material was identified at USNM by C. W. Gilmore but

310

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 9. — Partial mandibles of Tapirus veroensis in occlusal aspect, x 1. A) USNM 214663, left P3-M3, Island Creek local fauna. North
Carolina; B) ANSP 15270, partial alveoli and roots of P4-M3, continental shelf off New Jersey; C) USNM 24588, left and right I,-I3
and canines, right P,-M3, Ladds, Georgia.

1984

RAY AND SANDERS- PLEISTOCENE TAPIRUS

311

was not deposited in the collections. Isolated teeth, mostly frag-
ments, of T. veroensis size are known from at least three fissure
fillings— 4.5 mi south of Harper, West Virginia, in Highland Co.
(in the Strait Canyon fauna, under study by John Guilday at the
time of his death); 1 mi south of Edinburg, Shenandoah Co.,
collected in an abandoned stone quarry by Dr. and Mrs. L. S.
Bartell and Dr. Ray Gieseman, 1965; 4 mi east of Pearisburg,
Giles Co., collected in an abandoned quarry in the south bank
of New River by Robert Weems and Brenda Higgins, 1972.

From North Carolina, in addition to the material of Tapirus
haysii from the Neuse River, 16 mi below New Bern, there are
also isolated teeth and two incomplete maxillae referable to T.
veroensis. P. J. Harmatuk has also collected a fragment of a tapir
tooth from spoil along Mosley Creek, a tributary of the Neuse
River that forms a part of the Lenoir-Craven Co. line 8.5 mi east
northeast of Kinston; and several isolated teeth and fragments
from disturbed overburden at the N. C. Lime Company Quarry,
4 mi west-northwest of Comfort, Jones Co. The most important
material of T. veroensis from North Carolina is that collected by
Mr. Harmatuk from windrows of Pleistocene overburden at the
Ideal Cement Company Quarry, New Hanover Co., immediately
south of the junction of Island Creek with the Northeast Cape
Fear River (34°22'30"N, 77°50'00"W). The Pleistocene assem-
blage here has been designated the Island Creek fauna, and is
under study by R. E. Eshelman and C. E. Ray. The principal
specimens are left and right mandibular rami, USNM 214663,
not making contact across the symphysis but almost certainly
representing a single, rather small individual. Measurements of
the teeth are given in Table 2, and the left ramus is illustrated
in Fig. 9 A. A fragment of left ramus, USNM 2398 16, has a rather
large M3, L 29.4, AW 22.2, PW 18.8. The collection also includes
a partial right maxilla with P‘-P3 (USNM 244789), an isolated
right M2? (USNM 347326), a mandibular symphysis with in-
complete canine (USNM 239817), and a left astragalus (USNM
244749), all of which represent rather small individuals, certainly
not of T. haysii size.

In South Carolina and Georgia remains of tapirs of T. veroensis
size, mostly isolated teeth, are known from numerous coastal and
near-coastal occurrences, from beaches, river bottoms, and
dredgings. In addition to numerous teeth from the Myrtle Beach,
South Carolina, area, there is a left mandibular ramus with P4-
M3 (USNM 244462) from Litchfield Beach, approximately 20
mi southwest of Myrtle Beach, Georgetown Co. This specimen
was collected and donated by Roy Herrick in 1978. As noted
elsewhere, there are numerous specimens, mostly partial denti-
tions, in the ChM and SCMC collections from the Cooper River
where the skull described herein was found. Several specimens
were listed by Roth and Laerm (1980:18) from the Edisto Island
assemblage, some 21 mi southwest of Charleston. Isolated teeth
have been found in the dredgings from the Savannah River at
Savannah, Georgia, and from dredgings for interstate highway
construction near Brunswick. Tapirus cf. veroensis was recorded
among the mammals from Ladds, Bartow Co., Georgia, on the

DISCUSSION AI

Relationships among taxa of tapirs in the Qua-
ternary cannot be established confidently on the ba-
sis of material that does not include skulls. Thus the
only North American taxon that is adequately de-
fined as yet is Tapirus veroensis, founded on a fine

basis of a few fragments (Ray, 1967:142), but after that report
had been published, David Bailey, then a student at Shorter
College, found the major part of a mandible with all teeth on
both sides except left P2 and P3 (USNM 24588). Measurements
of the cheek teeth are given in Table 2, indicating that this is a
rather small individual. The right mandibular ramus and sym-
physeal region are shown in Fig. 9C.

Nearly 40 years ago Simpson (1945:63) was able to state that
localities for tapirs were essentially continuous over much of
Florida, and that is all the more accurate today. Webb (1974:18)
tabulated numerous localities for Florida, and Lundelius and
Slaughter (1976) studied specimens from several localities. It
should perhaps be put on record however, that several of the
Floridian localities in the latter paper are in error. Most impor-
tantly, neither of the two upper dentitions listed as topotypes in
their table 4 is from Vero; UF/FGS V4389 is from Rock Springs,
Orange Co., and UF/FGS V5447 is from Haile, Alachua Co.

Taxonomic status. — As Simpson (1945) indicat-
ed, Sellards (1918) firmly established the specific
distinctiveness of the skull of Tapirus veroensis. Our
study confirms this conclusion. Thus, whatever
eventuates regarding priorities, synonymies, and
nomenclature, there is unquestionably in the Qua-
ternary of North America a we); marked, wide-
spread, extinct species of tapir similar in size to the
living forms. It seems that tapirs, within the confines
of a rather limited and conservative repertoire of
characters, have recombined these characters in var-
ious ways in speciation, so that T. veroensis and the
four living species present a congery of variations
on the traditional tapirid themes. Supraspecific tax-
onomy is considered under Discussion and Conclu-
sions.

With regard to number of species, we suspect that
only a single species of late Pleistocene tapir of the
size of T. veroensis occurred in North America. With
regard to specific nomenclature, T. fossilis may be
suppressed as a forgotten name, and the character-
istics of 7 . californicus are so little known that its
status can scarcely be considered until additional
material is collected and studied.

With regard to subspecific nomenclature (possibly
utilizing the name T. excelsus, if its employment
eventually proves useful, it will be only after vari-
ation in time and space is much better documented
than it now is.

CONCLUSIONS

skull, supplemented by several referred skulls of var-
ious individual ages from various localities. They
demonstrate that T. veroensis is characterized by its
own combination of variations within the overall
conservative morphologic range of Tapirus ; that it

312

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

was a basic tapinne, without any extreme charac-
teristics such as the precociously high sagittal crest
of T. terrestris or the extensively ossified meseth-
moid cartilage embraced by high maxillary flanges
of T. bairdii. The available specimens indicate that
T. veroensis was very widespread in the contermi-
nous United States in late Pleistocene time, at least
from Texas and Missouri eastward through Ten-
nessee to South Carolina and Florida; that it may
have extended from coast to coast and north to with-
in the glaciated northern states, and possibly back
to earlier Pleistocene time, if more fragmentary
specimens are correctly referred.

All other nominal North American taxa and their
distribution in time must be regarded as provisional
pending recovery of well preserved, well docu-
mented skulls. Meanwhile these insecurely based
taxa need not be rejected, but should be retained
pending inevitable further discoveries. At present
our highly tentative conclusion is that there are two
species of Tapirus in the Pleistocene of North Amer-
ica—a smaller late (and probably earlier) Pleistocene
form including T. californicus, T. veroensis, T. ten-
nesseae, and T. excelsus\ and a larger, early to mid-
dle (and possibly latest) Pleistocene form including
T. haysii, T. merriami, and T. copei.

Supraspecific taxonomy of Quaternary tapirs can-
not be meaningfully considered without reference
to all adequately founded species, living and extinct.
Such consideration strongly recommends grouping
of all species within the single genus Tapirus, with-
out subgeneric divisions. Almost any desired group-

ing (with the possible exception of that uniting all
New World forms, see Simpson, 1945:40) could be
achieved by selection of the appropriate charac-
ters). Conversely, isolating each species in its own
monotypic genus or subgenus as favored by some
(see Hershkovitz, 1954:466; Colbert and Hooijer,
1953:90) defeats the grouping function of generic-
group nomenclature. In order to attain nomencla-
tural harmony within that scheme, T. veroensis
would require its own genus or subgenus, which we
regard as unwarranted.

If one accepts that the primary function of the
genus is grouping, not dividing, and considers all
adequately known Quaternary tapirs, then ironically
Megatapirus augustus and Tapirus indicus, of all
known species, provide the least justifiable case for
separation at the generic-group level. They are the
one pair of species among Quaternary tapirs that
are indubitably closer to one another than either is
to any other. Thus, they afford the one unarguable
opportunity for the generic group to fulfill its in-
tended function.

In the absence of any overriding argument in fa-
vor of generic group subdivisions, there is also a
practical value in retaining all within Tapirus, as
Simpson (1945:41) indicated. That arrangement
permits assignment of inadequately known species
such as T. haysii to a genus without implying more
than is known, and without making Tapirus or Tap-
irus ( Tapirus ) a receptacle for miscellany, as would
be inevitable under a divisive arrangement.

ACKNOWLEDGMENTS

As always, vertebrate paleontology begins in the field, with
recovery of significant specimens. This project, like so many in
our experience in the southeastern United States, owes its ini-
tiation to collectors who are not professional paleontologists.
Foremost in this case must be mentioned Susan Wallace who
found and donated to the Charleston Museum the best Pleisto-
cene tapir skull yet recovered in North America, rivaled only by
the holotype of Tapirus veroensis. Peter J. Harmatuk’s tireless
pursuit of fossils along the lower Neuse River has renewed interest
in this nearly forgotten classic collecting ground and provided
the catalyst for reconsideration of Tapirus haysii. Other collectors
who have provided key material and information include Mary
Agnew, Donnie Bailey, L. S. Bartell, Thelma Bennett, Raymond
Douglas, Fred Grady, Joe Hartsell, Roy Herrick, Mr. and Mrs.
Philip Kennel, Frances Malinow, M. Sturgis, and Brandy Vasi-
lew.

The following proprietors of specimens or collections have
made specimens available for study: Larry Agenbroad, Donn
Claiborne, Paul Connor, Howard Converse, Mary Dawson, Ar-

thur Harris, Ernest Lundelius, Bruce MacFadden, Rudolph
Mancke, Dan Morse, Paul Parmalee, Charles Schaff, Charles
Smart, Vince and Judy Schneider, Thomas Uzzell, Gay Vostreys,
and David Webb. They have also provided useful information,
as have Shirley Fonda, Lewis Lipps, Gary Morgan, Carol Spawn,
and Robert Weems.

The photographs were made by Victor Krantz, and the figures
prepared by Lawrence Isham.

Finally we wish to express our appreciation for the support of
this and other research provided through the years by the late
John Guilday. His help with this small project is exemplary of
the sustained professional performance that was seemingly in-
stinctive for him, but exceptional in the world. At irregular in-
tervals he passed along information and casts of new specimens
of fossil tapirs as they came his way, always unsolicited and with
no expectation of quid pro quo. His absence leaves a void in our
profession, but his memory provides a lasting model to be em-
ulated.

1984

RAY AND SANDERS- PLEISTOCENE TAPIRUS

313

LITERATURE CITED

Bader, R. S., and D. Techter. 1959. A list and bibliography
of the fossil mammals of Illinois. Nat. Hist. Miscellanea,
172:1-8.

Blackwelder, B. W. 1981. Stratigraphy of upper Pliocene and
lower Pleistocene marine and estuarine deposits of north-
eastern North Carolina and southeastern Virginia. U.S. Geol.
Surv. Bull., 1 502-B: 1 — 16.

Blackwelder, B. W., and L. W. Ward. 1983. Late Pliocene
and early Pleistocene Mollusca from the James City and
Chowan River formations at the Lee Creek Mine. Smith-
sonian Contrib. Paleobiol., in press.

Bogan, A. E., P. W. Parmalee, and R. R. Polhemus. 1980. A
review of fossil tapir records from Tennessee with descrip-
tions of specimens from two new localities. J. Tennessee
Acad. Sci., 55:10-14.

Carpenter, W. M. 1 842. Notice of an interesting fossil. Amer.
J. Sci. Arts, 42:390-391.

. 1846. Remarks on some fossil bones recently brought

to New Orleans from Tennessee and from Texas. Amer. J.
Sci. Arts, ser. 2, 1:244-250.

Clark, A. H. 1938. Some Pleistocene mammals from Warren
County, Virginia. Science, 88:82. (Reprinted in part, 1939,
Raven, 10:6-7.)

Colbert, E. H., and D. A. Hooijer. 1953. Pleistocene mam-
mals from the limestone fissures of Szechwan, China. Bull.
Amer. Mus. Nat. Hist., 102:1-134.

Conrad, T. A. 1835. Observations on the Tertiary strata of
the Atlantic coast. Amer. J. Sci. Arts, 28:104-1 11.

. 1838. Fossils of the medial Tertiary of the United States,

No. 1. Judah Dobson, Philadelphia, 52 pp. (Reprinted with
introduction and annotations by W. H. Dali, 1893, Wagner
Free Inst. Sci., Philadelphia.)

. 1842. Observations on a portion of the Atlantic Tertiary

region, with a description of new species of organic remains.
Bull. Proc. Nat. Inst. Promotion Sci., 2:171-194.

Cooper, W. 1831. Notices of Big-bone Lick. Monthly Amer.
J. Geol. Nat. Sci., 1:158-174, 205-217.

Cope, E. D. 1871. Preliminary report on the Vertebrata dis-
covered in the Port Kennedy bone cave. Proc. Amer. Philos.
Soc„ 12:73-102.

Corbitt, D. L. 1950. The formation of the North Carolina
counties, 1663-1943. State Dept. Archives and History, Ra-
leigh, 323 pp.

Croom, H.B. 1835. Some account of the organic remains found
in the marl pits of Lucas Benners, Esq. in Craven County,
N. C. Amer. J. Sci. Arts, 27:168-171.

Dall, W. H. 1892. Contributions to the Tertiary fauna of Flor-
ida, with especial reference to the Miocene silex-beds of
Tampa and the Pliocene beds of the Caloosahatchie River.
Part II. Streptodont and other gastropods, concluded. Trans.
Wagner Free Inst. Sci., 3:201-473.

Dalquest, W. W. 1977. Mammals of the Holloman local fauna.
Pleistocene of Oklahoma. Southwestern Nat., 22:255-268.

Du Bar, J. R., and J. R. Solliday. 1963. Stratigraphy of the
Neogene deposits, lower Neuse estuary. North Carolina.
Southeastern Geol., 4:213-233.

Emry, R. J., and R. W. Purdy. 1983. The holotype and would-
be holotypes of Hyracodon nebraskensis (Leidy 1850). Acad.
Nat. Sci. Philadelphia Notulae Naturae, in press.

Fallaw, W., and W. H. Wheeler. 1969. Marine fossiliferous
Pleistocene deposits in southeastern North Carolina. South-
eastern Geol., 10:35-54.

Fonda, S. 1982. Late Pleistocene vertebrates from a filled fis-
sure-cave in central Pennsylvania. Geol. Soc. Amer., NE.
SE. Sec., Abstracts with Programs, 14:18.

Foster, J. W. 1857. On the geological position of the deposits
in which occur the remains of the fossil elephant of North
America. Proc. Amer. Assoc. Adv. Sci., 10:148-169.

Frick, C. 1921. Extinct vertebrate faunas of the badlands of
Bautista Creek and San Timoteo Canon, southern California.
Univ. California Publ., Bull. Dept. Geol., 12:277-424.

Garton, E. R. 1979. Late Pleistocene and Recent mammal
remains from two caves at Bowden, West Virginia. Proc.
West Virginia Acad. Sci. for 1977, 49:1 10-1 16.

Gazin, C. L. 1950. Annotated list of fossil Mammalia associ-
ated with human remains at Melbourne, Fla. J. Washington
Acad. Sci., 40:397-404.

Gidley, J. W., and C. L. Gazin. 1938. The Pleistocene verte-
brate fauna from Cumberland Cave, Maryland. Bull. U.S.
Nat. Mus., 171:1-99.

Gillette, D. D., and E. H. Colbert. 1976. Catalogue of type
specimens of fossil vertebrates, Academy of Natural Sci-
ences, Philadelphia. Part II: terrestrial mammals. Proc. Acad.
Nat. Sci. Philadelphia, 128:25-38.

Graustein, J. E. 1967. Thomas Nuttall, naturalist, explora-
tions in America, 1808-1841. Harvard Univ. Press, Cam-
bridge, Massachusetts, 48 1 pp.

Guilday, J. E. 1962. Notes on Pleistocene vertebrates from
Wythe County, Virginia. Ann. Carnegie Mus., 36:77-86.

Hager, M. W. 1974. Late Pliocene and Pleistocene history of
the Donnelly ranch vertebrate site, southeastern Colorado.
Univ. Wyoming Contrib. Geol., Spec. Papers, 2(1975): 1-62.

Harlan, R. 1825. Fauna Americana: being a description of the
mammiferous animals inhabiting North America. Anthony
Finley, Philadelphia, i-x, 11-318 pp.

. 1835. Critical notices of various organic remains hith-
erto discovered in North America. Pp. 253-313, in Medical
and physical researches: or original memoirs in medicine,
surgery, physiology, geology, zoology, and comparative anat-
omy (R. Harlan), Lydia R. Bailey, Philadelphia, 653 pp.
(Reprinted from 1834, Trans. Geol. Soc. Pennsylvania,
1:46-1 12, non vid.).

. 1842. Notice of two new fossil mammals from Bruns-
wick Canal, Georgia; with observations on some of the fossil
quadrupeds of the United States. Amer. J. Sci. Arts, 43:141 —
144.

Hatcher, J. B. 1896. Recent and fossil tapirs. Amer. J. Sci.,
ser. 4, 1:161-180.

Hay, O. P. 1912. The Pleistocene Period and its Vertebrata.
Pp. 539-784, in E. Barrett, Indiana Dept. Geol. Nat. Res.
Ann. Rept. 1911, 36:1-873.

. 1914. The Pleistocene mammals of Iowa. Pp. 1-662,

in G. F. Kay, Iowa Geol. Surv. Ann. Rept. 1912, 23:i— xlviii,
1-662.

. 1920. Descriptions of some Pleistocene vertebrates

found in the United States. Proc. U.S. Nat. Mus., 58:83—
146.

. 1923. The Pleistocene of North America and its ver-

314

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

tebrated animals from the states east of the Mississippi River
and from the Canadian provinces east of longitude 95°. Car-
negie Inst. Washington Pub!., 322:1-499.

. 1927. The Pleistocene of the western region of North

America and its vertebrated animals. Carnegie Inst. Wash-
ington Publ., 322B: 1-346.

Hay, O. P., and H. J. Cook. 1930. Fossil vertebrates collected
near, or in association with, human artifacts at localities near
Colorado, Texas; Frederick, Oklahoma; and Folsom, New
Mexico. Proc. Colorado Mus. Nat. Hist., 9(2):4— 40.

Hays, I. 1834. Descriptions of the inferior maxillary bones of
mastodons, in the cabinet of the American Philosophical
Society, with remarks on the genus Tetracaulodon, &c. Trans.
Amer. Philos. Soc., 4:317-339. (Separate dated 1833.)

. 1852. [Remarks on a tooth of the fossil Tapir.] Proc.

Acad. Nat. Sci. Philadelphia, 6:53.

Hershkovitz, P. 1954. Mammals of northern Colombia, pre-
liminary report no. 7: tapirs (genus Tapirus ), with a system-
atic review of American species. Proc. U.S. Nat. Mus., 103:
465-496.

Hibbard, C. W., and W. W. Dalquest. 1966. Fossils from the
Seymour Formation of Knox and Baylor counties, Texas,
and their bearing on the late Kansan climate of that region.
Univ. Michigan, Contrib. Mus. Paleontol., 21:1-66.

Hibbard, C. W., R. J. Zakrzewski, R. E. Eshelman, G. Edmund,
C. D. Griggs, and C. Griggs. 1978. Mammals from the
Kanopolis local fauna, Pleistocene (Yarmouth) of Ellsworth
County, Kansas. Univ. Michigan, Contrib. Mus. Paleontol.,
25:1 1-44.

Kurten, B.. and E. Anderson. 1980. Pleistocene mammals of
North America. Columbia University Press, New York, 442
PP-

Lance, J. F. 1959. Faunal remains from the Lehner mammoth
site. Pp. 35-39, in The Lehner mammoth site, southeastern
Arizona (E. W. Haury, E. B. Sayles, and W. W. Wasley, eds.),
Amer. Antiq., 25:2-42.

Leffler, S. R. 1964. Fossil mammals from the Elk River For-
mation, Cape Blanco, Oregon. J. Mamm., 45:53-61.

Leidy, J. 1849. Tapirus Americanus fossilis. Proc. Acad. Nat.
Sci. Philadelphia, 4:180-182.

. 1852a. [Reference to a fossil tooth of a Tapir.] Proc.

Acad. Nat. Sci. Philadelphia, 6:106.

. 18526. [Remarks on Tapirus Haysii.] Proc. Acad. Nat.

Sci. Philadelphia, 6:148.

. 1853. The ancient fauna of Nebraska: or, a description

of remains of extinct Mammalia and Chelonia, from the
mauvaises terres of Nebraska. Smithsonian Contrib. Knowl-
edge, 6(7): 1-126.

. 1855. Notice of some fossil bones discovered by Mr.

Francis A. Lincke, in the banks of the Ohio River, Indiana.
Proc. Acad. Nat. Sci. Philadelphia, 7:199-201.

. 1858. [Remarks on a cast of a mastodon tooth, &c.]

Proc. Acad. Nat. Sci. Philadelphia, 10:12.

. 1859. Description of vertebrate fossils. Pp. 99-122, in

1858-1860, Post-Pleiocene fossils of South-Carolina (F. S.
Holmes, ed.), Russell and Jones, Charleston, 122 pp.

. 1869. The extinct mammalian fauna of Dakota and

Nebraska, including an account of some allied forms from
other localities, together with a synopsis of the mammalian
remains of North America. J. Acad. Nat. Sci. Philadelphia,
ser. 2, 7:i— vii + 8-472.

Lund, P. W. 1841. Blik paa Brasiliens dyreverden for sidste
jordomvaeltning. Tredie afhandling: fortsaettelse af patte-
dyrene. Kongelige Danske Videnskabemes Selskabs, Na-
turvidenskabelige og Mathematiske Afhandlinger, 8:21 7-272.

Lundelius, E. L., Jr. 1972. Fossil vertebrates from the late
Pleistocene Ingleside fauna, San Patricio County, Texas. Univ.
Texas Austin Bur. Econ. Geol. Rept. Investigations, 77:1 —
74.

Lundelius, E. L., Jr., and B. H. Slaughter. 1976. Notes on
American Pleistocene tapirs. Pp. 226-243, in Essays on pa-
laeontology in honour of Loris Shano Russell (C. S. Church-
er, ed.), Royal Ontario Mus., Life Sci. Misc. Publ., Athlon,

286 pp.

Lydekker, R. 1886. Catalogue of the fossil Mammalia in the
British Museum, (Natural History) Cromwell Road, S. W.
Part III. Containing the Order Ungulata, suborders Peris-
sodactyla, Toxodontia, Condylarthra, and Amblypoda. Brit-
ish Museum (Natural History), London, 186 pp.

Lyell, C. 1843. On the geological position of the Mastodon
giganteum and associated fossil remains at Bigbone Lick,
Kentucky, and other localities in the United States and Can-
ada. Proc. Geol. Soc. London 4:36-39. (Reprinted with mi-
nor changes, 1844, Amer. J. Sci. Arts, 46:320-323.)

. 1849. A second visit to the United States of North

America, vol. 1 . Harper and Brothers, Publishers, New York,
i-xii, 13-273 pp.

McCartan, L., J. P. Owens, B. W. Blackwelder, B. J. Szabo,
D. F. Belknap, N. Kriausakul, R. M. Mitterer, and J. F.
Wehmiller. 1982. Comparison of amino acid racemiza-
tion geochronometry with lithostratigraphy, biostratigraphy,
Uranium-series coral dating, and magnetostratigraphy in the
Atlantic coastal plain of the southeastern United States. Qua-
ternary Res., 18:337-359.

McCartan, L., R. E. Weems, and E. M. Lemon, Jr. 1980. The
Wando Formation (upper Pleistocene) in the Charleston,
South Carolina, area. In Changes in stratigraphic nomen-
clature by the U.S. Geological Survey, 1979 (N. F. Sohl and
W. B. Wright, eds.), U.S. Geol. Surv. Bull., 1502-A:110-
116.

Merriam, J. C. 1913. Tapir remains from late Cenozoic beds
of the Pacific coast region. Univ. California Publ., Bull. Dept.
Geol., 7:169-175.

Miller, W. E. 1971. Pleistocene vertebrates of the Los Angeles
basin and vicinity (exclusive of Rancho La Brea). Bull. Los
Angeles County Mus. Nat. Hist., Sci., 10:1-124.

Mitchell, E. 1 842. Elements of geology, with an outline of the
geology of North Carolina: for the use of the students of the
University. Publisher and place not indicated, 141 pp.

Mixon, R. B., and O. H. Pilkey. 1976. Reconnaissance geology
of the submerged and emerged coastal plain province, Cape
Lookout area, North Carolina. U.S. Geol. Surv. Prof. Paper,
859:1-45.

Moodie, R. L. 1933. A popular guide to the nature and the
environment of the fossil vertebrates of New York. New
York State Mus. Handbook, 12:1-122.

Morse, D. F. 1970. Preliminary notes on a recent mastodon
and tapir find in northeastern Arkansas. Arkansas Archeol.,
Bull. Arkansas Archeol. Soc., 1 1 :45 — 49.

Parmalee, P. W., R. D. Oesch, and J. E. Guilday. 1 969. Pleis-
tocene and Recent vertebrate faunas from Crankshaft Cave,
Missouri. Illinois State Mus. Rept. Investigations, 14:1-37.

1984

RAY AND SANDERS— PLEISTOCENE TAP1RUS

315

Peterson, O. A. 1926. The fossils of the Frankstown Cave,
Blair County, Pennsylvania. Ann. Carnegie Mus., 16:249-
314.

Radinsky, L. B. 1965. Evolution of the tapiroid skeleton from
Heptodon to Tapirus. Bull. Mus. Comp. Zoo!., 134:69-106.

Ray, C. E. 1964. Tapirus copei in the Pleistocene of Florida.
Quart. J. Florida Acad. Sci., 27:59-66.

. 1967. Pleistocene mammals from Ladds, Bartow Coun-
ty, Georgia. Bull. Georgia Acad. Sci., 25:120-150.

Richards, H. G. 1936. Fauna of the Pleistocene Pamlico For-
mation of the southern Atlantic coastal plain. Bull. Geol.
Soc. Amer., 47:1611-1656.

. 1950. Geology of the coastal plain of North Carolina.

Trans. Amer. Philos. Soc., new ser., 40:1-83.

. 1957. Fossil mammals from the New Jersey coast. Ann.

Bull. Cape May Geogr. Soc., 1 1:8-10.

. 1959. Pleistocene mammals dredged off the coast of

New Jersey. Bull. Geol. Soc. Amer., 70:1769 (abstract).

Richardson, G. B. 1907. [Elephas columbi, Equus complica-
tus, and Tapirus haysii from El Paso, Texas.] Science, n.s.,
25:32.

. 1909. Description of the El Paso District. U.S. Geol.

Surv., Geol. Atlas of the U.S., El Paso Folio, Texas, 166:1-
11.

Roth, J. A., and J. Laerm. 1980. A late Pleistocene vertebrate
assemblage from Edisto Island, South Carolina. Brimleyana,
J. North Carolina State Mus. Nat. Hist., 3:1-29.

Rubin, M., and C. Alexander. 1958. U.S. Geological Survey
radiocarbon dates IV. Science, 127:1476-1487.

Savage, D. E., and D. E. Russell. 1983. Mammalian paleo-
faunas of the World. Addison- Wesley Publishing Co., Read-
ing, Massachusetts, 432 pp.

Schultz, C. B.,L. D. Martin, and R. G. Corner. 1975. Middle
and late Cenozoic tapirs from Nebraska. Bull. Univ. Ne-
braska State Mus., 10:1-21.

Schultz, C. B., L. G. Tanner, F. C. W'hitmore, Jr., L. L. Ray,
and E. C. Crawford. 1963. Paleontologic Investigation at
Big Bone Lick State Park, Kentucky: a preliminary report.
Science, 142:1 167-1169.

Sellards, E. H. 1918. The skull of a Pleistocene tapir including

description of a new species and a note on the associated
fauna and flora. Florida Geol. Surv. Ann. Rept., 10:57-70.

Simpson, G.G. 1942. The beginnings of vertebrate paleontology
in North America. Proc. Amer. Philos. Soc., 86:130-188.

. 1945. Notes on Pleistocene and Recent tapirs. Bull.

Amer. Mus. Nat. Hist., 86:33-82.

Slaughter, B. H. 1966. The Moore Pit local fauna; Pleistocene
of Texas. J. Paleontol., 40:78-91.

Stephenson, L. W. 1912. Quaternary formations. Pp. 266-290,
in The coastal plain of North Carolina (W. B. Clark, B. L.
Miller, L. W. Stephenson, B. L. Johnson, and H. N. Parker,
eds.). North Carolina Geol. Econ. Surv., 3:1-552.

Strain, W. S. 1966. Blancan mammalian fauna and Pleistocene
formations, Hudspeth County, Texas. Bull. Texas Mem. Mus.,
10:1-55.

Ward, L. W\, and B. W. Blackwelder. 1980. Stratigraphic
revision of upper Miocene and lower Pliocene beds of the
Chesapeake Group, Middle Atlantic coastal plain. U.S. Geol.
Surv. Bull., 1482-D: 1-6 1 .

Warren, J. C. 1855. Description of a skeleton of the Mastodon
giganteus of North America. John Wilson and Son, Boston,
second ed., 260 pp.

Webb, S. D. 1974. Chronology of Florida Pleistocene mammals.
Pp. 5-31, in Pleistocene mammals of Florida (S. D. Webb,
ed.), Univ. Presses of Florida, Gainesville, 270 pp.

Wheeler, W. H., R. B. Daniels, and E. E. Gamble. 1979. Some
stratigraphic problems of the Pleistocene strata in the area
from Neuse River estuary to Hofmann Forest, North Car-
olina. Pp. 41-50, in Structural and stratigraphic framework
for the coastal plain of North Carolina (G. R. Baum, W. B.
Harris, and V. A. Zuilo, eds.), Carolina Geol. Soc. and At-
lantic Coastal Plain Geol. Assoc. Field Trip Guidebook, Oc-
tober 19-21, 1979, Wrightsville Beach, North Carolina, Geol.
Surv. Sec., Dept. Nat. Res. and Community Dev., Raleigh,

111 pp.

Winge, H. 1906. Jordfundne og nulevende Hovdyr ( Ungulata)
fra Lagoa Santa, Minas Geraes, Brasilien. Med Udsigt over
Hovdyrenes indbyrdes Slaegtskab. E Museo Lundii, 3( 1 ): 1 —
239.

Address (Ray): Department of Paleobiology, National Museum of Natural History, Smithsonian Institution,

Washington, D.C. 20560.

Address (Sanders): Department of Natural Sciences, Charleston Museum, Charleston, South Carolina 29403.

SANG AM ON A: THE FURTIVE DEER

Charles S. Churcher

DEDICATION

John E. Guilday was not enamoured of Sangamonafugitiva and wrote “I hope someone
someday comes up with both ends of this beast so we can really pin it down. . . . Ap-
parently its habits were such that it didn’t often get into cave deposits— and even when
it did it apparently checked its lower plate before doing so” (personal communication,
14 February 1977) and “Perhaps someday someone will come up with a good specimen
of Sangamona so that we can see what the beast looks like in its entirety. Someone
should make direct comparisons someday” (personal communication, 2 March 1978).
His attitude of mind is further revealed in his reference to it as “ Sangamona furtiva ”
(Guilday et al., 1969:64) and gives the title to this paper. I therefore dedicate this
consideration of the status of Sangamona fugitiva to my colleague and good friend John
E. Guilday, from whom I learnt much about Pleistocene vertebrates and with whom I
carried on a most cordial correspondence.

ABSTRACT

The cervid Sangamona fugitiva described by Hay (1921) from
deposits in Tennessee, Maryland, and Illinois is compared with
modem representatives of Cervus canadensis (wapiti), Odocoi-
leus virginianus (white-tailed deer), O. hemionus (mule deer) and
others. The type of S. fugitiva, a worn M2, is considered inde-
terminate. Hay’s other “typical” specimens from Maryland and
Illinois resemble elements of Odocoileus or Cervus. Additional
fossil specimens placed within Sangamona sensu lato from New

Mexico, Tennessee, Missouri, Nebraska, Iowa, and Pennsylvania
are also assessed. In most instances these specimens are associ-
ated with remains of Cervus, Odocoileus, Rangifer (caribou), and
Alces or Cervalces (moose) and may be assigned to one of these
taxa. The taxon Sangamona fugitiva Hay 1920 is therefore in-
valid, except for the type, which is undiagnostic and indeter-
minate to species and genus.

INTRODUCTION

Hay (1921:91, pi. 3, figs. 14-15) founded San-
gamona fugitiva on a left upper second molar
(USNM 8954) collected from the Pleistocene of a
cave at Whitesburg, Hamblen County, Tennessee,
in 1885 by Ira Sayles. In the same paper Hay at-
tributed to his new taxon specimens from Alton,
Illinois, collected by Hon. Wm. McAdams before
1883 and comprising two right mandibular frag-
ments with partial dentitions (USNM 9002-9003)
and a fragmentary and badly worn left upper molar
row (USNM 9096) which cannot be profitably com-
pared with the type specimen. He also included
within S. fugitiva cervid remains (USNM 9 1 92—
9193) from Cavetown, Maryland, collected by Dr.
Charles Peabody and Warren K. Moorehead. No

upper cheekteeth occur among these chiefly post-
cranial elements, but a well worn lower first or sec-
ond molar and a lower first incisor (USNM 9192)
are included.

Subsequent to the original description, other cer-
vid specimens have been assigned to Sangamona
fugitiva, few of which are directly comparable with
Hay’s (1921) type or “paratypic” materials de-
scribed in the same paper. The descriptions of ma-
terial given below include the type, paratypes, and
referred specimens that have been placed within S.
fugitiva by one or more authors. Associations with
other cervid taxa found in the same deposits are
noted.

ABBREVIATIONS AND MEASUREMENTS

The following abbreviations are used to indicate the institu-
tions or collections in which fossil specimens or recent compar-
ative skeletons are conserved:

ANSP Academy of Natural Sciences, Philadelphia, Pennsyl-
vania.

CM Carnegie Museum of Natural History, Pittsburgh, Penn-
sylvania.

FA Field Accession Number, Department of Mammalogy,
Royal Ontario Museum, Toronto, Ontario.

MCZ Museum of Comparative Zoology, Harvard University,
Cambridge, Massachusetts.

316

1984

CHURCHER- STATUS OF SANGAMON A

317

ROM Royal Ontario Museum (Department of Mammalogy),
Toronto, Ontario.

RW Red Willow County paleontological site, University of
Nebraska State Museum, Lincoln, Nebraska.

UNSM University ofNebraska State Museum (Division of Ver-
tebrate Paleontology), Lincoln, Nebraska.

USNM United States National Museum, now The National
Museum of the United States (Smithsonian Institution),
Washington, D.C.

All measurements are given in millimetres unless otherwise
indicated; ranges of variation are given as maxima, minima,
means, standard deviations and sample sizes (Max., Min., mean,
SD, and N); “e” indicates an estimated measurement, “ — ” a
measurement that cannot be obtained due to lack of reference
point(s), and blanks indicate absence of information.

MATERIAL CURRENTLY ASSIGNED TO SANGAMONA

Specimens Originally Assigned to
Sangamona fugitiva by Hay (1921)

1. Type. Left M2, complete but maturely worn
(USNM 8954). From a cave at Whitesburg, Ham-
blen Co., Tennessee. Hay, 1921:91, pi. 3, figs. 14-
1 5. Collected in 1885 by Ira Sayles. Associated with
Odocoileus virginianus and Cervus canadensis.

2. Right I, (USNM 9192), left M, or M2, distal
end of left radius, right scaphoid, right innominate
acetabulum, left malleolus, proximal half of astrag-
alus, right calcaneum, two (probably metatarsal) ses-
amoids, and first, second, and third (probably pedal)
phalanges (USNM 9193). From fissure deposits at
Cavetown, Washington Co., Maryland. Hay, 1921:
102-104. Collected by Dr. Charles Peabody and
W'arren K. Moorehead. Associated with Odocoileus
virginianus.

3. Two right mandibular fragments with maturely
worn cheekteeth; one with MrM2, remnants of P4,
and traces of mesial roots of M, (USNM 9002), and
one with M,-M3 (USNM 9003); and a very well
worn left upper dentition P3-M3 (USNM 9096).
From loess nodules from Alton, Madison Co., Il-
linois. Hay, 1921:1 1 1, pi. 5, figs. 5-6. Collected be-
fore 1883 by Hon. \¥m. McAdams. Associated with
Cervalces roosevelti and Rangifer muscatensis (=R.
tarandus).

Specimens Discovered Later and
Assigned to Sangamona fugitiva or
Sangamona sp.

4. Metapodial (ANSP 13931), humerus (ANSP
13930), right maxillae with P3~4 (ANSP 13592) and
P3-M! (ANSP 13591), left maxilla with P2-4 (ANSP
13578), right and left upper molars (ANSP 14065,
13582, respectively), right adult and immature den-
taries (ANSP 14060, 13579, respectively). From
Burnet Cave, Eddy Co., New Mexico. Schultz and
Howard, 1935:287, pi. 14, fig. 9 (ANSP 13931);
referred to Sangamona ? sp. Associated with Odo-

coileus virginianus, O. hemionus, Rangifer ? fricki n.
sp.

5. Two right P4’s, left P2, left p4 and right P3 (CM
8372-8375, 8377). From Armadillo Pit, Robinson
Cave, Overton Co., Tennessee. Guilday et al., 1969:
64, figs. 12b, d, e, f; referred to Sangamona furliva.
Associated with Odocoileus cf. virginianus.

6. Left metatarsal, left humerus, paired tibiae, and
distal half of left tibia assigned to “cf. Sangamona ”
and two left dentaries with P3-M3 and P4-M3 as-
signed to “cf. Sangamona or Odocoileus Haiti A From
Brynjulfson Cave No. 1, Boone Co., Missouri. Par-
malee and Oesch, 1972:40-43. Associated with
Odocoileus virginianus and cf. Alces or Cervalces.

7. Left antler fragment, pedicel and fragment of
frontal (UNSM 46326), distal ends of two right hu-
meri (UNSM 48534-48535), left metacarpal (USNM
48526), and distal end of right humerus (UNSM
46304). From gravel pits RW 101 and 102, south
bank of the Republican River, Red Willow Co., Ne-
braska. Corner, 1977:85. Associated with Odocoi-
leus sp. and Rangifer tarandus.

Specimens Reassigned to
Sangamona fugitiva from
Other Taxa

8. Right metatarsal lacking distal end, left hu-
merus lacking proximal epiphysis, and left radius
lacking distal end. From Dubuque, Dubuque Co.,
Iowa. Allen, 1876:49-50, assigned to Cervus whit-
neyi\ Kurten, 1975:508, referred to Sangamona. As-
sociated with Cervus canadensis.

9. Female skull fragments with reconstructed up-
per toothrows (CM 11044), limb and foot bones
(CM 11043). From Frankstown Cave, Blair Co.,
Pennsylvania. Peterson, 1926:257, assigned to Odo-
coileus hemionus ; Guilday et al., 1969:64, reas-
signed to Sangamona. Associated with Odocoileus
virginianus and Cervalces americanus.

10. Single P4 (CM 29501) and left P3 (CM 30060).
From Baker Bluff Cave, Sullivan Co., Tennessee.
Guilday et al., 1978:50, assigned to cf. Sangamona

318

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 1 . — Localities from which specimens assigned to Sangamona fugitiva s.l. have been reported.

1. Whitesburg, Hamblen Co., Tennessee. Type specimen; Hay, 1921.

2. Cavetown, Washington Co., Maryland. Paratypic specimens; Hay, 1921.

3. Alton, Madison Co., Illinois. Paratypic specimens; Hay, 1921.

4. Armadillo Pit, Robinson Cave, Overton Co., Tennessee. Guilday et al., 1969.

5. Brynjulfson Cave, Boone Co., Missouri. Parmalee and Oesch, 1972.

6. Republican River, pits RW 101 & 102, Red Willow Co., Nebraska. Comer, 1977.

7. Dubuque, Dubuque Co., Iowa. Kurten, 1975.

8. Frankstown Cave, Blair Co., Pennsylvania. Peterson, 1926.

9. Baker Bluff Cave, Sullivan Co., Tennessee. Guilday et al., 1978.

Burnet Cave, Eddy Co., New Mexico, and Peccary Cave, Newton Co., Arkansas, are not shown as the former represents Navahoceros
and the latter is unconfirmed.

fugitiva. Associated with cf. Cervus elaphus, Odo-
coileus virginianus and Rangifer tarandus.

Kurten and Anderson (1980:313) listed Peccary
Cave, Newton Co., Arkansas, as containing remains

of Sangamona fugitiva. However, Semken (1969)
and Davis (1969) did not mention Sangamona
among the mammals discussed. No other mention
of this record has been seen.

1984

CHURCHER-STATUS OF SANGAMONA

319

CONSIDERATION OF HAY’S TYPE AND ASSOCIATED SPECIMENS OF
SANGAMONA FUGITIVA (LOCALITIES 1-3)

The taxon, Sangamona fugitiva, was founded by
Hay ( 1 92 1 ) on an isolated second upper molar from
a cave at Whitesburg, Tennessee, for which the only
comparable fossil material known at present is that
from Frankstown Cave, Pennsylvania (Peterson,
1926). The upper molar row (USNM 9096) from
Alton, Illinois, is too worn and incomplete for any
of the qualitative characters observable in the type
to be compared and size comparisons are suspect
because of the small amount of enamel wall re-
maining. Wells (1959) has expressed reservations
on the use of single equid teeth as reliable taxonomic
specimens and suggested that entire molar rows or
extensive uniform collections of teeth only be con-
sidered as sound foundations for taxonomic iden-
tifications. Even then, because of variations within
species and individuals because of sex, age, and wear,
he considered that such identifications should be
used with caution. I believe that such caveata apply
equally to artiodactyl dentitions. Thus, Sangamona
fugitiva as founded on an isolated and overly worn
tooth as the type specimen is a nomen vanum in the
sense of Simpson (1945:27).

Hay (1921:91) described the characters of the type
(USNM 8954) of Sangamona fugitiva thus: “This
genus differs much from our other deer in the nearly
complete absence of the strong ribs which occupy
the outer faces of the lobes of the upper molars”
and diagnosed the species by “Styles, or ribs on
paracone and metacone absent or obsolete. Size in-
termediate between Virginian deer and wapiti.” Ex-
amination of the type specimen shows no rib on the
metacone and a very weak one on the paracone, and
the paracone style to be rounded and that on the
metacone bluntly pointed. No distal hypostyle is
present. The protocone is strong, with rugose enam-
el on the cingulum, and the hypocone is large. The
tooth appears proportionately more brachydont than
is usual in Odocoileus virginianus. There is a small
interlophar endostyle or pillar within the protocone-
hypocone valley as occurs in some O. virginianus.
The prefossette opens lingually past the endostyle,
the protocone joins the parastyle mesially, and the
hypocone joins the distobuccal comer of the meta-
cone. No plis occur within the fossettes, although
the distal end of the protoloph is slightly bifurcated.

The specimens from Cavetown, Maryland, are

mostly postcranial elements but include a right first
incisor and a first or second lower molar. Both of
these specimens are as unsuitable for specific iden-
tification as the upper molar from Whitesburg and
do not provide a reliable basis on which to found a
new taxon. Hay (1921:103) considered that “This
[molar] tooth and the bones are entirely too large
to have belonged to any known species of Odocoi-
leus and too small for any known species of Cervus.
The incisor (Cat. No. [USNM] 9 1 92) is considerably
larger than the corresponding one of the Virginian
deer” and “There is a rather strong tubercle ecto-
stylid or pillar at the mouth of the principal valley”
on the lower molar. These teeth and postcranial
elements are not illustrated, but Hay’s measure-
ments show the animal to be generally intermediate
in size between O. virginianus and C. canadensis,
but often closer to the former.

The specimens from Alton, Illinois, provide at
least two associated partial right mandibular tooth
rows by which the taxon may be characterized. Hay
(1921:111, pi. 5, figs. 5-6) stated of the M3’s “At
the outer mouth of the median valley of these teeth
there is a conspicuous accessory pillar [ectostylid].
The crowns of the lower molars are higher than in
Odocoileus. The inner faces of the lobes are flatter
than in Odocoileus, and they are, relative to the
length, much broader. They agree in size so well
with the upper tooth . . . found at Whitesburg, Ten-
nessee, and with the lower tooth found at Cavetown,
Maryland, that they are referred to that species. In
size, they agree well with the lower molar found at
Cavetown, Maryland, and are referred to S. fugiti-
va.” He provided few measurements of the speci-
mens and made no mention of the well worn upper
molar row (USNM 9096) also from Alton, Illinois.

The problem that arises from Hay’s (1921) de-
scriptions is how one can, with any degree of cer-
tainty, compare and conclude that an isolated upper
molar collected from a cave in Tennessee in 1885,
two partial right mandibular molar rows collected
from loess in Illinois before 1883, and an incisor
and lower molar collected in 1908 or later from a
fissure in Maryland belong within a single new genus
and species. The type specimen is isolated and can-
not be compared with the other specimens, and is
not compared with the worn molar row from Illi-

320

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 2.— Generalized cervid skeleton indicating major specimens assigned to Sangamona fugitiva. Not all specimens are shown. Shading
indicates elements assigned to S', fugitiva— light shading indicates a single specimen or element, dark shading indicates two or more
specimens of an element, and black indicates isolated teeth. Figures in parentheses indicate multiple specimens.

1. Type specimen from Whitesburg, Hamblen Co., Tennessee. (Hay, 1921).

2. Paratypic specimens from Cavetown, Washington Co., Maryland. (Hay, 1921).

3. Paratypic specimens from Alton, Madison Co., Illinois. (Hay, 1921).

4. Specimens from Armadillo Pit, Robinson Cave, Overton Co., Tennessee. (Guilday et al., 1969).

5. Specimens from Brynjulfson Cave, Boone Co., Missouri. (Parmalee and Oesch, 1972).

6. Specimens from pits RW 101 and 102, south bank of Republican River, Red Willow Co., Nebraska. (Comer, 1977).

7. Reassigned specimens from Dubuque, Dubuque Co., Iowa. (Kurten, 1975).

8. Reassigned partial skeleton from Frankstown Cave, Blair Co., Pennsylvania. (Peterson, 1926).

9. Specimens from Baker Bluff, Sullivan Co., Tennessee. (Guilday et al.. 1978).

10. Specimens from Burnet Cave, Eddy Co., New Mexico. (Schultz and Howard, 1935).

nois. The isolated incisor and postcranial elements
from Maryland are also unmatched. The isolated
first or second molar and the damaged mandibular
rows are comparable and could represent a mor-
phology sufficiently distinct to provide a basis for
the establishment of a new taxon. However, Hay
contented himself by noting that the lower teeth
from Illinois agree so well in size with the upper
tooth and isolated lower teeth that they are referred

to the same taxon. Hay (1921:91) diagnosed the
genus Sangamona thus:

"'Diagnosis— Upper molars of medium height,
broad, outer face of anterior lobe with feebly de-
veloped style; outer face of the hinder lobe deeply
concave and devoid of style. Lower molars rela-
tively broad; inner faces of front and hinder lobes
flat and with feebly developed style.”

He diagnosed the species S. fugitiva thus:

1984

CHURCHER- STATUS OF SANGAMON A

321

“Diagnosis— Styles, or ribs, on paracone and
metacone absent or obsolete. Size intermediate be-
tween the Virginian deer and the wapiti.”

Effectively, these diagnoses say the same thing,
that is, the upper molar teeth have small or reduced
styles and reduced or absent ribs, and the lower
molar teeth have flat lingual surfaces with feeble
styles or ribs. These conditions are not unknown in
Odocoileus but are rare. They are common in Ran-
gifer, but its tooth pattern is different from that of
the fossils. However, these conditions are more often
seen in small specimens of Cervus canadensis than
in Odocoileus.

In all three deposits (Whitesburg, Tennessee;
Cavetown, Maryland; Alton, Illinois) Sangamona
is found variously associated with Cervus canaden-
sis (Tennessee), Odocoileus virginianus (Tennessee
and Maryland), and Cervalces roosevelti and Ran-
gifer tarandus (Illinois), and thus the suspicion aris-
es that the specimens constitute elements of a well
known cervid that have been misidentified. Mis-
identification may have come about because Hay
had few specimens with which to compare his fossils
and because he may not have been aware of the
dental variants possible in these cervids.

In descriptions of fossil Sangamona Hay gives
comparative measurements of single specimens of
O. virginianus and/or C. canadensis. A reinvesti-
gation and remeasurement of original specimens and
extended comparative samples generating ranges of
measurements and mean values, provide insights
into the probable identities of the specimens attrib-
uted to Sangamona fugitiva by Hay and others.

Hay’s (1921) original measurements are uninfor-

mative except to show that Sangamona'’ s type and
associated specimens were marginally smaller in
mesiodistal and buccolingual diameters of the
cheekteeth than in Cervus canadensis. Comparisons
with more numerous specimens of Odocoileus show
the type upper molar to be similar in length but
broader, even though Sangamona lacks the well de-
veloped styles or ribs on the buccal faces of the
ectoloph (Table 1), and is considerably smaller than
comparable molars of Cervus.

The Alton Sangamona rami afford suites of mea-
surements for M2 and M3. Compared with those for
Odocoileus, the Sangamona teeth are always 2 to 3
mm larger in all dimensions. Compared with those
for Cervus, they fall just within or below the minima
for modern specimens.

The Cavetown postcranial elements are all more
massive than those of Odocoileus (Table 2), al-
though in some dimensions only marginally so, and
generally smaller than those of Cervus, but again in
some only marginally less.

These comparisons suggest that the type molar is
not only indeterminable qualitatively but quanti-
tatively, and could represent a large individual of
Odocoileus or a small one of Cervus, with both of
which it is associated in the cave. The Alton rami
bear teeth that are too large for Odocoileus but suit-
able for Cervus, from which they do not differ
strongly qualitatively except in the strength of the
lingual styles and ribs. The Cavetown postcranial
elements appear to represent an individual inter-
mediate in size between Odocoileus and Cervus, as
for the type.

COMMENTS ON OTHER SPECIMENS ASSIGNED TO SANGAMONA FUGITIVA

(LOCALITIES 4-7)

4. Burnet Cave, Eddy Co., New Mexico.

Schultz and Howard (1935) listed nine specimens
identified as “ Sangamona ? sp. Extinct Cervid” from
Burnet Cave, New Mexico. These comprise two right
maxillae with P3-4 and P3-M' (ANSP 13592, 13591,
respectively), a left maxilla with P2-4 (ANSP 13578),
left and right upper molars (ANSP 13582, 14065,
respectively), right adult and immature dentaries
(ANSP 14060, 13579, respectively), humerus (ANSP
13930) and partial metapodial (ANSP 13931), de-
riving from at least three individuals. Schultz and
Howard (1935:287) remarked only “Size interme-

diate between Odocoileus and Cervus. Probably equal
to Sangamona ” and refrained from giving mea-
surements or comparisons with living Odocoileus or
Cervus species, with Hay’s (1921) specimens of San-
gamona, or with their newly described and associ-
ated Rangiferl fricki.

However, Kurten (1975) considered the various
cervids from the Pleistocene of the southwestern
United States and Mexico, described as Rangiferl
fricki (Schultz and Howard, 1935), Sangamonal sp.
(Schultz and Howard, 1935), Cervus lascrucensis
(Frick, 1937), and Odocoileus halli (Alvarez, 1969),

322

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table {.—Measurements of the type specimens o/Sangamona fugitiva from Hay (1921) and by Churcher, and of comparable teeth of
Recent Cervus canadensis and Odocoileus virginianus. (L) or (R) indicate left or right side specimens, “e” indicates an estimated
measurement, “+ " a minimal measurement, SD = standard deviation, and "N” = number in samples.

Odocoileus virginianus

Churcher

Maxi-

Sangamona fugitiva Cervus canadensis mum —

Specimens and Mini-
measurements Hay Churcher Hay Churcher Hay mum Mean SD N

First lower incisor

Cavetown, Maryland

Specimens (USNM 9192) FA 342-4 9

Crown height

8

13

18.1

8.45-6.9

7.7

.56

Crown width (diameter)

7.5

9 Max.
6.7 Min.

14.1

8. 9-6. 8

7.6

1.18

Total length

32

50?

First or second lower molar

Cavetown, Maryland

Specimen

(USNM 9193 C [L])

Mesiodistal length

20

19.4

24

12.5

(vide infra)

Buccolingual width

14.8

15.0

15

9

Crown height

8.0

Ectostylid

present

present

First lower molar

Alton, Illinois USNM series

FA series 342-

9002

9003

9002

9003

Specimens

(R)

(R)

(R)

(R)

1 3

29

4 9

Mesiodistal length

15e

—

15. 8e

—

25.3

22.1

20

15.0-12.5

14.2

.82

8

Distal buccolingual

—

—

—

12.2

14.5

17.9

16.3

11.6-9.8

10.4

.77

8

width

Crown height

9 +

10.0

Ectostylid

present

absent

absent

present

present

present 3/8

Second lower molar

Mesiodistal length

18

20

19.4

23

24

27.7

28.7

26.0

16.9-15.6

16.2

.60

8

Mesial buccolingual

13

14.8

12.6

13.5

15.3

18.4

16.8

11.6-9.8

10.9

.69

8

width

Distal buccolingual

14.0

16.0

15

14.8

20.7

17.7

width

Crown height

11.5

14.5

Ectostylid

present

present

absent

absent

absent

present 1/8

Third lower molar

Total mesiodistal

length

22e

27

22e

28.5

36.1

38.5

35.3

21.1-19.3

20.3

.57

8

Mesiodistal length

18

20

19.2

21.5

28.7

27.4

26.0

16.8-14.3

15.6

.81

8

minus hypoconulid
Mesial buccolingual

13

13.8

16.0

16.2

19.2

17.6

1 1.0-9.5

10.5

.46

8

width

Distal buccolingual

12.1

13.4

15.8

19.7

16.9

width

Crown height

15

11.5

16.4

Ectostylid

present

absent

present

absent

absent

vestigial

1984

CHURCHER- STATUS OF SANGAMON A

323

Table 1. — Continued.

Sangamona fugitiva

Cervus canadensis

Odocoileus v
C

Maxi-

mum—

irginianus

hurcher

Mean SD

N

Specimens and
measurements

Hay

Churcher

Hay

Churcher

Hay mum

Second upper molar

Whitesburg, Tennessee

Specimens

(Type USNM 8954 [L])

FA 342-2 9

Maximum buccal mesio-

distal length

20

19.7

21

28.0

19.8-16.2

17.4

1.19

li

Mesiodistal length

over buccal

16

18.1

25.0

15.5-12.5

14.5

1.08

i i

cingulum

Mesial buccolingual

diameter over styles

22

21.6

26

28

18.4-14.9

16.5

1.45

i i

Distal buccolingual

20.5

20.4

25.3

19.0-15.3

16.8

1.58

i i

diameter over styles

Mesial buccolingual

19.3

27.8

18.4-14.2

15.8

1.32

li

diameter over ribs

Distal buccolingual

19.6

24.1

18.5-15.6

16.5

1.09

n

diameter over ribs

Crown height buccally

13.5

13.2-5.7

11.1

2.0

i i

Crown height lingually

9.1

12.5-4.9

10.7

2.0

i i

Interlophar endostyle

present 3/ 1 1

and placed them in a new taxon Navahoceros fricki.
Kurten (1975:537) characterized this monotypic ge-
nus as “Medium to large sized cervids with very
thick-set limb-bones and short metapodials; males
with simple forked antlers; females antlerless.” Kur-
ten and Anderson (1980:312) modify this to “large
deer, in the size range between mule deer and wapiti,
with stocky limbs, very short and heavy metapo-
dials, simply built, three-tined antlers, and molar
teeth with weakly developed ribs.” The observed
range of variation in the San Josecito Cave, Mexico,
sample encompasses that seen in the other speci-
mens and thus constitutes individual variation.

Kurten (1975) and Kurten and Anderson (1980)
described both Sangamona and Navahoceros as in-
termediate in size between (). hemionus and C. can-
adensis but that Sangamona has stilt-like metapo-
dials and long, slender limbs, whereas Navahoceros
has shortened metapodials and limb bones remi-
niscent of chamois and ibex. They also concluded
that the two species’ ranges did not overlap, with
Navahoceros being adapted to an alpine habitat and
Sangamona to the eastern regions, possibly as a
grazer on open ground.

5. Robinson Cave, Overton Co., Tennessee.

Guilday et al. (1969) listed three individuals of

Sangamona furtiva from the Armadillo Pit. These
are represented by two right P4’s (CM 8372-8373),
a left P2 (CM 8374), a left milk p4 (CM 8375) and
a right P3 (CM 8377). Two measurements are given
for each of these specimens (Table 3) and Guilday
et al. (1969:64, figs. 12b, d, e, 0 gave line diagrams
of one P4, and the P2, p4, and P3. While Hay (1921)
described no premolars in the original material, and
thus none of these teeth is comparable to any of the
paratypic specimens, he gave as a major character-
istic of the cheek teeth of Sangamona the absence
of well developed styles and ribs on the ectolophs
of upper molars and on the lingual faces of lower
molars. In the specimens reported by Guilday et al.,
1969:64) obvious styles and ribs are present, espe-
cially on the permanent upper premolars. The milk
fourth premolar (CM 8375) has stylids and ribs, and
twin ectostylids, typical of a cervid. Guilday and
Dr. Clayton E. Ray of The National Museum of the
United States (USNM) assigned these teeth to San-
gamona, although Guilday had previously identi-
fied them as Rangifer. Guilday suspected initially
that these teeth derived from Rangifer and that San-
gamona was a synonym for that genus. Measure-
ments of the Robinson Cave Sangamona teeth (Ta-
ble 3) show them to be closer in size to those of

324

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 2. — Measurements of the postcranial elements from Cavetown, Maryland, assigned to Sangamona fugitiva by Hay (1921) compared

with those of Cervus canadensis and Odocoileus virginianus.

Measurements

Sangamona fugitiva
Cavetown, Maryland
USNM 9193

Hay

C. canadensis

Churcher

O virginianus

Churcher

(ROM

Hay 14907 9)

ROM
25241 i

FA

342-1 3

Hay

Churcher

Radius, distal end

Maximum transverse diameter

48.6

58.5

63.2

35.1

Width above articulation less ulna

48

43.8

65

57.8

55.3

40

29.7

Transverse diameter of articulation

45

57.0

53.0

34.1

Thickness above articulation

31

32.4

45

44.5

43.1

27

26.2

Scaphoid

Proximodistal length

26

27

? 1 9

28.3

31.2

18.6

Anteroposterior width

32

32

26

35

38.0

24.2

Transverse diameter

17

? 1 9

18.8

22.4

13.1

Acetabulum

Anteroposterior diameter

45.8

56

59.8

33.6

Length

45

57

37

Malleolar

Anteroposterior horizontal diameter

26

25.4

33.1

31.9

19.3

Transverse diameter

17.4

16.3

17.3

9.0

Proximodistal diameter

27.8

26.1

24.0

16.2

Calcaneum (juvenile)

Preserved maximum length

±105

102 +

138

132.5

138.5

103

93.6

Height at articulation for fibula

44

43.5

56

49.2

50.7

32

30.5

Height over sustentaculum

37.1

43.0

40.5

25.5

Thickness at lateral (sustentacular)

process

33

32.5

40

41.0

35.5

30

23.4

Length of cuboid facet

30.5

33.1

36.3

20.6

Width of cuboid facet

12.2

15.0

14.8

8.8

Astragalus

Tibial (articular) width

30

30.7

37.4

36.5

26.1

Dorsoventral depth

26.1

34.6

34.6

21.1

Phalanx 1

Total length

66

66.5

68

62.8

65.1

53

53.2

Proximal height

28

28.5

35

30

31.4

21

21.3

Proximal width

23

22.5

27

23.7

25.0

17

17.4

Distal height

18

18.2

21

19.6

20.0

14

12.5

Distal width

19

18.6

26

24.4

24.3

15

12.8

Least shaft height

16.6

17.6

17.8

11.6

Least shaft width

17.0

20.3

19.1

12.3

Phalanx II

Total length

42

42.1

48

45.0

47.3

41

38.5

Proximal height

26

25.4

35

29.8

30.5

22

20.9

Proximal width

19

17.9

25

24.3

22.4

17

14.7

Distal height

24

24.5

29

27.0

28.5

18

17.0

Distal width

16

15.8

26

19.0

20.0

12

11.0

Least shaft height

19.3

20.5

21.3

14.1

Least shaft width

14.0

18.5

15.6

10.7

Phalanx III

Length

45e

38.4 +

50.3

36.6

Proximal height

30

31.3

32.3

20.2

Proximal width

18

18.1

10.0

12.3

1984

CHURCHER- STATUS OF SANGAMON A

325

Table 3.— Measurements of cheek teeth o/Sangamona fugitiva (CM 8372-8375, 8377) from Robinson Cave, Tennessee (after Guilday
et a!., 1969:65) compared with those of Recent Cervus canadensis, Odocoiieus virginianus, and Rangifer tarandus.

Tooth and
measurements

Sangamona furtiva
Tennessee

Cervus canadensis
British Columbia

Odocoiieus virginianus

Rangifer tarandus
Yukon

Ontario and Alberta

Ontario

p2

CM 8374

FA 342-2

Churcher Collection

ROM 14907

Churcher Collection

Mesiodistal length

17.5

20.5

11.8,

12.1, 15.4, 16.0

Buccoiingual width

13.7

17.3

10.7,

10.7, 14.0, 14.1

P4

CM 8372 CM 8373

Mesiodistal length

14.5 13.2

17.7

11.3,

10.7, 12.9, 12.8

10.2, 9.2

Buccoiingual width

18.5 18.4

22.5

12.3,

12.1, 17.8, 17.2

10.5, 13.1

P3

CM 8377

FA 342-3 FA 342-4

Mesiodistal length

13.1

18.8 18.3

11.1,

11.2, 11.8, 11.8

10.4

14.3

Buccoiingual width

9.1

13.0 12.3

7.7,

7.3, 7.2, 7.3

6.8

9.5

P<

CM 8375

Mesiodistal length

22.5

16.2,

15.8

22.5, 21.0, 22.0

Buccoiingual width

11.1

8.2,

8.1

7.6, 7+, 7.5

Cervus canadensis than to Odocoiieus virginianus
and are similar to those of Rangifer tarandus, except
in the buccoiingual diameter of p4.

The fauna from Robinson Cave includes Odo-
coiieus cf. virginianus, for which Guilday et al. ( 1 969)
identified a right P2 and P3, and it is likely that the
Sangamona teeth represent this deer because of
similar size and patterns of the cusps and lophs.

6. Brynjulfson Cave No. 1, Boone Co.. Missouri.

Parmalee and Gesch (1972) reported cf. Sanga-
mona sp. from Brynjulfson Cave No. 1, represented
by a left humerus, right and left tibiae, distal end of
left tibia, and a left metatarsal. Two left rami with
P4-M3 (Jaw A) and P3- M3 (Jaw B) (Parmalee and
Oesch, 1972:42-43, fig. 14, table 8; erroneously la-
beled P3 and P2-P3 in table 8) may belong to San-
gamona, although Dr. Clayton E. Ray is cited as
stating that they agree well with the two topotypic
rami of Odocoiieus halli. No teeth of Odocoiieus
are reported although remains of five individuals
(85 specimens) of O. virginianus and a left P1, a
cervical vertebra, a partial left calcaneum, a right
acetabulum, and the distal end of a right cubitus of
Alces americana or Cervalces are recorded.

None of the postcranial materials is comparable
to the paratypic specimens from Cavetown de-
scribed by Hay (1921; Table 4). The doubtfully as-
signed rami with molar series possess strongly ribbed
convex lingual surfaces and appear to differ from
the Alton rami in the characters of the molars (com-
pare Hay, 1 92 1 , pi. 5, figs. 5-6; Parmalee and Oesch,
1972:42, fig. 14). The pattern of the Brynjulfson P3
differs from that from Robinson Cave but, because
these patterns vary considerably in recent and ex-
tinct deer, this difference may not be significant. The

Brynjulfson Sangamona lower dentitions are all
smaller than those of C. canadensis (Table 5 and
compare with Table 1) and larger than those of O.
virginianus. Thus it appears that the Brynjulfson
dentitions are more properly assigned to O. halli,
as suggested by Dr. Ray, or to O. virginianus which
is known from the cave but not represented by den-
tal material.

The postcranial materials are not comparable to
Hay’s (1921) originally described materials but may
be compared with recent cervid elements. The hu-
merus is smaller than that of C. canadensis and
comparable to that of O. virginianus (Table 4). The
tibiae and metatarsal are intermediate in size be-
tween those of C. canadensis and O. virginianus.
The three postcranial elements may therefore be
parsimoniously assigned to Odocoiieus sp., which
supports the possible identity of the two rami.

7. Gravel Pits on the South Bank of the Repub-
lican River, Red Willow Co., Nebraska.

Comer (1977) reported a left antler pedicel (USNM
46326), two distal ends of right humeri (UNSM
48534-48535) from Pit RW 101 and a left meta-
carpal (UNSM 48526) and the distal end of a right
humerus (UNSM 46304) from Pit RW 102 from
the south bank of the Republican River.

Comer (1977:85) stated “The metacarpal is re-
ferred to Sangamona mainly on the basis of size,
being intermediate between mule deer and wapiti
(Kurten, 1975:508). Sangamona differs from Na-
vahoceros in having longer, more slender limbs. The
metacarpal is structurally similar to that of the wapi-
ti ( Cervus canadensis ), except the posterior vascular
groove is not as deep. The partial humeri were . . .
found to be similar in size and structure” to a hu-

326

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 4.— Measurements of the postcrania! elements o/‘ Sangamona "from Brynjulfson Cave, Missouri, after Parmalee and Oesch (1972:
41, table 7), of Cervus whitneyi, Odocoileus virginianus, and O. hemionus after Allen (1876:50), with corrected or additional measure-
ments of the same specimens and Recent Cervus canadensis, Odocoileus virginianus, and O. hemionus.

C. canadensis

Element and dimension

C. whitneyi

Dubuque

Iowa

“ Sangamona ”
Brynjulfson Cave
Missouri

ROM FA

25241 6 342-2 9

British Columbia

O. virginianus
MCZ 1733
Maine

O. hemionus
MCZ 1781
Wyoming

Humerus

Total length

Length from caput to medial condylar

—

236.0

308

310

220

227

surface

Transverse width of distal condylar

—

282

282

200

203

surfaces

48

49

62

68.8

38

42

Anteroposterior width over distal
medial condyle

Maximum anteroposterior width in

51

64.3

65.0

42

42

centre of shaft (A)

Lateral transverse width in centre of

29.0

40.5

36.0

24.6

—

shaft (B)

Calculated circumference in centre

—

23.0

33.5

29.5

19.8

—

of shaft (A + B)ir/2

85

81.7

1 16.4

218.4

73

76

Radius

Allen, 1876

Total length

—

230

242

Transverse width of proximal end

-

37

39

Transverse width of distal end

41

57.1

55.3

36

38

Least transverse diameter of shaft

29

24

25

Least circumference of shaft

80

65

68

Metatarsal

Total length

—

314.0

326

336

254

273

Transverse width of proximal end

33

28

29

Anteroposterior width of proximal end

36

40.0

47.3

47.1

27.9

32

Least transverse width of shaft

22

18

21

Transverse width in centre of shaft

19.5

28.8

26.2

15.4

Least circumference of shaft

67

58

66

Width of distal end

-

37.0

50.7

49.5

32.3

35

Tibia

Total length

376, 378, -

410

420

311.2

Width of proximal end

69.0, 72.0, -

86.2

91.2

56.2

Transverse width in centre of shaft

29,0, 29.0, 26.0

36.6

21.5

32.4

Maximum width of distal end

48.0, 48.0, 44.0

37.0

58.3

37.0

Table 5.— Measurements of the dental series from the "Sangamona " dentaries from Brynjulfson Cave, Missouri (after Parmalee and
Oesch, 1972:43, table 8) compared with those of Recent Cervus canadensis and Odocoileus virginianus.

“ Sangamona "
Brynjulfson Cave, Missouri

Cen’us canadensis
FA 342-2, 342-4, British Columbia

Dimension

P3

P4

M,

m2

M,

p2

p,

P4

M, M2 M3

Jaw A

Mesiodistal length

15.0

16.0

18.5

23.5

12.1

19.3

20.5

Measurements for lower molars of

Buccolingual width
Alveolar length of tooth row

-

11.5

101.0

13.5

14.0

13.5

9.7

116.7

12.3

14.7

Cervus canadensis and Odocoile-
us virginianus are given in Table
1.

Jaw B

Mesiodistal length

16.5

17.5

18.0

22.0

26.5

15.5

19.5

20.3

Buccolingual width
Alveolar length of toothrow

13.0

14.0 12.0

1 1 5 estimated

13.0

12.0

10.8

124.5

13.0

15.6

1984

CHURCHER- STATUS OF SANGAMONA

327

mcrus reported from Burnet Cave by Schultz and
Howard (1935:287) and considered as Navahoceros
fricki by Kurten (1975:507). Corner (1977:85) re-
ferred the partial humeri from Pit RW 101 to “ San -
gamona and not Navahoceros because there is no
evidence to suggest that both forms are present at
Red Willow, and the complete metacarpal can be
referred to Sangamona with some confidence.”

Comer (personal communication, 9 January 1978)
recognized that some Recent wapiti metacarpals ap-
proach that of the Red Willow specimen in size but
are narrower at the articulations, and also that the
humeri may “be referrable to the long-legged big-
horn sheep ( Ovis catclawensis).” Measurements of
the metacarpal (UNSM 48526) and of Recent C.
canadensis and O. virginianus metacarpals (Table
6) show that it more nearly agrees with those of the
former in all dimensions. The metacarpal probably
represents a small Cervus canadensis and the humeri
possibly Ovis catclawensis, which is also present at
RW 101.

“The antler fragment is much larger than one
would expect for the small Odocoileus ” (Comer,
1977:85) and “is intermediate in size between O.
hemionus and Cervus canadensis ” and “does com-
pare nicely to the Navahoceros from Slaughter Can-
yon Cave, New Mexico, which is mounted and on
display” in Lincoln, Nebraska. However, the antler

Table 6. — Measurements of the left metacarpal of "Sangamona"
from the Republican River, Red Willow Co., Nebraska, compared
with those of Recent Cervus canadensis and Odocoileus virgini-
anus.

Dimension

“ Sanga-
mona"
UNSM
48526
Nebraska

C. canadensis
British Columbia

ROM FA

25241 342-1

O. vir-
ginianus
ROM
14907
Ontario

Articular length

270

285

291

215.9

Transverse width of proxi-
mal end

45.3

47

50.3

28.9

Anteroposterior width of
proximal end

30.9

33.3

35.2

20.9

Midshaft transverse width

27.0

28.5

27.3

17.3

Minimum anteroposterior
shaft diameter

19.9

20.8

23.5

18.1

Transverse width of
distal end

45.2

49.3

51.0

30.7

Anteroposterior width of
distal end

—

33.3

34.8

20.8

pedicel (UNSM 46326)

is not

too

small

to have

derived from a small individual of Cervus canaden-
sis and likely represents the same taxon as the meta-
carpal.

The Red Willow County records of Sangamona
are thus suspect, and likely parlim represent Cervus
canadensis and parlim Ovis sp., possibly Ovis cat-
clawensis ( =0 . canadensis catclawensis).

COMMENTS ON SPECIMENS LATER REFERRED TO SANGAMONA FROM
OTHER CERVID TAX A (LOCALITIES 8-10)

8. Dubuque, Dubuque Co., Iowa; Cervus whit-
neyi.

Allen (1876) described Cervus whitneyi from Du-
buque, Iowa, from a fauna originally reported by
Wyman ( 1 862) as comprising Bos ( =Bison ), Cervus,
Mastodon ( =Mammut ), Megalonyx, Dicotyles
( =Platygonus ), and Canis and based on the cervid
specimens that Wyman identified as “Red Deer
(• Cervus virginianus ).” The cervid remains comprise
a left humerus lacking the proximal epiphysis, and
a left radius lacking the distal end, both imperfect
or damaged, and a right metatarsal lacking its distal
end (referred to as left and right by Allen, 1876:48,
50, respectively), all from a young animal. Wyman
(1862:412) described the humerus as “closely re-
sembling that of the red deer, and of intermediate
size between this and the humerus of a Caribou.”
Allen (1876:48) considered that these specimens

“evidently belonged to a species different from any
hitherto described, either extinct or living,” and
placed them in a new species “ Cervus Whitneyi, in
honor of their discoverer, Professor J. D. Whitney.”

Allen (1876:50) considered that the three cervid
elements indicated “a species of about the same
proportions as Cervus Virginianus [= Odocoileus
virginianus], but much larger, considerably exceed-
ing in size Cervus macrotis [=Odocoileus hemion-
us].” He gave comparative measurements of these
elements and those of “C. macrotis ” and “C. Vir-
ginianus ” which show the fossil form to be larger
than the recent species by between 1 0% and 1 7%
(Table 4).

The humerus is typically cervid in form, the ra-
dius is similar but the ulna is fused for nearly the
whole radial contact, as in O. virginianus but not in
O. hemionus. The metatarsal is similar to that of O.

328

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

virginianus, but the posterior groove continues more
distad and differs from that of O. hemionus in being
more laterally compressed distally, slenderer, and
rounder in section, as in C. canadensis.

Allen's (1876) trivial name “ whitneyi ” for these
Iowan specimens predates Hay’s (1921) description
of the genus Sangamona and its single species, S.
fugitiva. Thus the correct form of the name is San-
gamona whitneyi (Hay). However, Allen’s name has
seldom if ever been applied, except to indicate these
specimens, and Hay’s binomial has achieved a wide
currency. Kurten and Anderson (1980) prefer to re-
tain Sangamona fugitiva for Hay’s type materials
and other specimens assigned to it, and probably
for C. whitneyi also. The name Cervus whitneyi is
thus a nomen oblitum.

Kurten (1975:508) tangentially referred the Iowan
specimens to Sangamona. but without discussion.
Kurten and Anderson (1980:313) stated that San-
gamona fugitiva “is probably also represented by
remains from Dubuque, Iowa, which were described
as Cervus whitneyi by Allen (1876).” Kurten (per-
sonal communication) considered that the Dubuque
C. whitneyi elements typical of Sangamona. How-
ever, no specimens of humerus or metatarsal are
recorded by Hay (1921) from the type locality at
Whitesburg, Tennessee, or from Alton, Illinois, or
Cavetown, Maryland, and only the distal end of a
left radius from the latter, so comparisons are not
possible. No comparable elements are recorded from
Robinson Cave, Tennessee (Guilday et al., 1969),
but a left humerus and a left metatarsal are recorded
from Brynjulfson Cave, Missouri (Parmalee and
Oesch, 1972). Comparison of measurements of these
elements with those of modern O. virginianus and
C. canadensis show that the elements from the three
fossil sites resemble one another in size (Table 4).
The humerus from Brynjulfson Cave is slightly lon-
ger than in modern Odocoileus but similar in section
to that from Dubuque. The radial ends from Cave-
town (Table 2) and Dubuque (Table 4) are inter-
mediate in size between those of Odocoileus and
Cervus canadensis. The metatarsals from Brynjulf-
son Cave and Dubuque (Table 4) are smaller than
that from Red Willow Co. (Table 6) originally as-
signed to Sangamona by Corner (1977) but now
considered to derive from Cervus canadensis.

9. Frankstown Cave, Blair Co., Pennsylvania;
Odocoileus hemionus.

Peterson (1926) reported Odocoileus hemionus
from Frankstown Cave, Pennsylvania, based on
fragments of a skull (CM 1 1044) with complete den-

tition and a number of limb and foot bones (CM
1 1043). He (1926:257) stated “The bones are those
of an animal of considerably larger size than Odo-
coileus virginianus and most closely resemble the
recent species, O. hemionus. The inner basal cusps
of the [upper] molars . . . are in the present specimen,
small and rather irregularly developed, more nearly
like the recent mule deer.” The appendicular ele-
ments are appropriately sized to derive from a single
individual of this species. These remains were as-
sociated with elements of O. virginianus and Cer-
valces americanus.

The development of the inner basal cusps (en-
dostyles) between the lophs of the upper molars of
deer (and most artiodactyls) is highly varied and
unreliable for taxonomic use in Cervidae. Thus size
alone differentiates the Frankstown specimens from
those of the associated O. virginianus and suggests
O. hemionus.

Guilday et al. (1969:64) remarked that, “In the
process of comparing the specimens of Cervidae
from Robinson Cave, Tennessee, a partial skeleton
from Frankstown Cave, Pa. identified as mule deer,
Odocoileus hemionus (Peterson, 1926), was found
to be that of Sangamona .” Guilday et al.’s (1969)
statement was made when only Hay’s (1921) orig-
inal and Schultz and Howard’s (1935) Burnet Cave
reports and specimens were available. Thus com-
parisons were possible only with the worn type up-
per M2 from Whitesburg, Tennessee, the almost fea-
tureless and undescribed P3-M3 from Alton, Illinois,
the partial pes and other postcranial fragments from
Cavetown, Maryland, and the specimens from Bur-
net Cave, New Mexico, now assigned to Navaho-
ceros fricki. The comparative materials thus lacked
either valid points for qualitative distinction or were
distinctly different.

Guilday (personal communication, in precis) gave
the following details about the Frankstown Cave
cervid cranium: the skull is extremely fragmentary,
the preserved parts are the dorsal surfaces of the
frontals and parietals, portions of the occipitals with
intact condyles, two petrous temporals, and com-
plete upper tooth rows. The animal, although an
adult doe, is not small; it is markedly larger than a
whitetail. The molars are large with reduced buccal
ribbing, and the premolars are not only actually but
relatively larger than in O. virginianus. The skull is
so fragmentary that only two dimensions may be
measured: maximum width across the condyles,
61.0, and transverse diameter of foramen magnum,
26.8 mm. He also remarked that the upper dentition

1984

CHURCHER- STATUS OF SANGAMON A

329

was so strikingly dissimilar from that of Odocoileus
that he did not consider them closely related.

Examination of the Frankstown cranial speci-
mens allows little comparison as to size and mor-
phology, except to substantiate that the animal was
female and, while large, not outside the size range
for Odocoileus and towards the lower range of size
for C. canadensis.

10. Baker Bluff Cave, 13 km SE of Kingsport,
Sullivan Co., Tennessee.

Guilday et al. (1978) reported two isolated pre-
molars (P4, CM 29501; left P3, CM 30060) assigned
to “cf. Sangamona fugitiva Hay— ‘fugitive’ deer.”
They stated that the upper premolar was larger than
that of Odocoileus and about the size of that of large

Rangifer from the deposit. Both teeth have the weak
buccal ribbing that Hay (1921) considered diagnos-
tic of Sangamona. The P3 is “identical in cusp for-
mation with Cervus or Odocoileus and was referred
to Sangamona because of its intermediate size”
(Guilday et al., 1978:50).

Measurements of P4 (CM 29501) are mesiodistal
length, 14.0, buccolingual width, 15.8 mm, and of
P3 (CM 30060) are buccolingual width, 9.6 mm.
These measurements agree tolerably with those for
the Robinson Cave, Tennessee, specimens, but are,
of course, not comparable to the type molar from
Whitesburg, Tennessee (USNM 8954), or the lower
molars of the paratypes from Alton, Illinois (USNM
9002, 9003).

SUMMARY DISCUSSION

Each fossil occurrence attributed to Sangamona
has generated divergent views, but in few of them
have discussions or examinations of tooth size and
general morphology, or of the size of the whole an-
imal or its long bones, been made on a comparative
basis, or if they have been, then the bases for the
conclusions have not been stated. In part this is
probably because of the sparse and idiosyncratic
nature of Hay’s (1921) incomparable “type speci-
mens” and to the incomplete skeletal examples later
added or referred to Sangamona. When size alone
appears to be the ruling criterion for assignment to
a genus, it may well be unreliable, especially if the
taxa compared are not represented by ranges of size,
nor are all likely taxa considered.

Sangamona has thus “evolved” from an original
indeterminate tooth, supported by geographically
separated dental and postcranial samples, through
the additions of further dental and postcranial ele-
ments and two mandibular rami. When these
specimens are considered parsimoniously for com-
parability, it is evident that in only two sites (Bryn-
julfson Cave No. 1, Missouri, and Frankstown Cave,
Pennsylvania) are there possibly associated cranial
and postcranial elements, and in both instances these
are associated with verifiable Odocoileus virginianus
remains. No antler rack has been reported and the
pedicel (Red Willow Co., Nebraska) is likely Cervus
canadensis.

Martin and Guilday (1967:52) summed up the
taxonomic situation at the time when only Hay’s

(1921) type materials and Schultz and Howard’s
(1935) erroneous report were available. They stated
“An extinct deer about the size of the modern car-
ibou, Sangamona was widely distributed in N.A.
Remains have been found in both western and east-
ern U.S.A., but there is no evidence of contempora-
neity with man. The type specimen is a single tooth.
While additional fragmentary remains appear to up-
hold the validity of this late Pleistocene deer, a def-
inite description has yet to appear ” (my italics).

Thus Sangamona is known as a deer that lacks
antlers (possibly only females known?), tends to be
associated with other cervids in cave deposits, is
intermediate in size between O. hemionus or O.
virginianus and C. canadensis, possesses molars with
weak styles, and is built in other aspects like Odo-
coileus or Cervus. As stated before, size is an un-
reliable indicator of taxon when qualitative criteria
are absent, and thus the postcranial elements alone
provide little basis for allocation to a taxon other
than those already recovered in the deposit. Stylar
variation in North American cervid teeth is wide,
and anomalously heavy or nearly absent conditions
occur in modem populations of Odocoileus virgin-
ianus, Cervus canadensis, and Rangifer tarandus ex-
amined in the Royal Ontario Museum, Toronto, the
National Museum of Natural Science, Ottawa, and
the British Museum (Natural History), London.

The recovery of a cranium with portions of the
antler beams to the bez tines from the early post-
glacial (latest Pleistocene, 11,315 ± 325 B.P.) ofTo-

330

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 1 .—Summary of occurrences o/Sangamona fugitiva, identifications, associated cervid taxa, and revised identifications. + = present
in deposit, +cf= probable presence, +sp = genus present, species uncertain, and X = wrongly identified.

Site and original identification

Odocoileus

virgini-

anus

Odo-

coileus

hemionus

Cervus

canadensis

Rangifer

tarandus

A Ices alces or
Cervalces sp.

Revised identification

Whitesburg, Tennessee
Sangamona fugitiva- type

+

+

Indeterminate: ?large
Odocoileus or small
Cervus

Alton, Illinois
Sangamona fugitiva

+

+

Cervalces

roosevelti

Cervus canadensis

Cavetown, Maryland
Sangamona fugitiva

+

large Odocoileus or
small Cervus

Burnet Cave, New Mexico
Sangamona ? sp.

+

+

X

Rangifer ?
fricki

Navahoceros fricki—
Kurten, 1975

Robinson Cave, Tennessee
Sangamona furtiva

-f cf

Odocoileus virginianus

Brynjulfson Cave, Missouri
cf. Sangamona or Odocoileus halli

+

+cf

Alces or
Cervalces

Odocoileus lhalli

Republican River, Nebraska
Sangamona sp.

+ sp

+

Cervus canadensis and
Ovis catclawensis

Dubuque, Iowa
Cervus whitneyi

+ sp

Cervus canadensis

Frankstown Cave, Pennsylvania
Odocoileus hemionus

+

+

large Odocoileus or
small Cervus

Baker Bluff Cave, Tennessee
cf. Sangamona fugitiva

+

+

+

large Odocoileus or
small Cervus

ronto, named Torontoceros hypogaeus (Churcher
and Peterson, 1982), was greeted by some as pro-
viding the missing antlers of Sangamona. However,
the Toronto cranium has no points of comparison
with Hay’s (1921) original specimens and the
Frankstown Cave cranium is too fragmentary and
could not be reliably compared to the Toronto cra-
nium. Because the Toronto cranium has geological
significance in addition to its taxonomic informa-
tion, publication of the record was desirable. To
publish the cranium as Sangamona would only have
exacerbated the situation already existing by sug-
gesting information on the shape of the antlers, but
based on a specimen that had no reason except con-
venience to be assigned to Sangamona.

What is Sangamona fugitiva ? It appears to be
derived almost equally from Odocoileus (probably
O. virginianus) and Cervus canadensis remains (Ta-
ble 7). It was inadequately founded by Hay (1921)
and has been extended by attribution of unusual-
sized bones or cheek teeth with reduced styles from
seven other sites. Thus I respond to John Guilday’s
stated wishes that someone someday would make
direct comparisons and pin the beast down, and
conclude that Sangamona is mythological. Sanga-
mona fugitiva— the fugitive or furtive deer —is a
construct of diverse and disparate skeletal elements,
all of which derive from other and usually better
known taxa and, except in the minds of men, never
existed.

ACKNOWLEDGMENTS

I have first to thank Dr. R. L. Peterson of the Department of
Mammalogy, Royal Ontario Museum, Toronto for involving me
in the problem of the validity of Sangamona fugitiva by asking
me to identify the cranium with antlers we subsequently de-

scribed as Torontoceros hypogaeus (Churcher and Peterson, 1 982),
as this was possibly Sangamona. Without his unknowing im-
petus, this work would never have been completed. Dr. Bjorn
Kurten of the University of Helsinki, Finland, Dr. C. Richard

1984

CHURCHER - STATUS OF SANGAMON A

331

Harington of the National Museum of Natural Sciences, Ottawa,
Dr. R. George Comer of the University of Nebraska State Mu-
seum, Lincoln, Nebraska, and the late Mr. John E. Guilday all
made freely available their experience and knowledge of fossil
deer and Sangamona. However, the conclusions expressed here
are solely my responsibility. Dr. Juliet Glutton-Brock of the De-
partment of Mammals, and Dr. Alan W. Gentry of the Depart-
ment of Palaeontology, British Museum (Natural History), Lon-
don, England, and Drs. Harington and Peterson assisted me by

allowing me to use the collections in their care. I have also been
helped by the friendly advice of Dr. Thomas S. Parsons and Dr.
James A. Bums of this department, and by Mrs. Judy Hann-
Cfaemos, my technical assistant, who drew the diagrams and
typed an earlier draft of this paper, and by Ms. Margjka My-
chajlowycz, who helped proof the final version. The work was
financed by National Science and Engineering Research Council
Grant A 1760.

LITERATURE CITED

Allen, J. A. 1876. Description of some remains of an extinct
species of wolf and an extinct species of deer from the Lead
Region of the Upper Mississippi. Amer. J. Sci. Arts., Ser. 3,
11 (Whole No. 1 1 1):47— 50.

Alvarez, T. 1969. Restos fosiles de mamlferos de Tlapacoya,
Estado de Mexico (Pleistocene-Reciente). Misc. Publ. Mus.
Nat. Hist., Univ. Kansas, 51:93-112.

Chuxcher, C. S., and R. L. Peterson. 1982. Chronologic and
environmental implications of a new genus of fossil deer
from Late Wisconsin deposits at Toronto, Canada. Quater-
nary Res., 18:184-195.

Corner, R. G. 1977. A Late Pleistocene-Holocene vertebrate
fauna from Red Willow County, Nebraska. Trans. Nebraska
Acad. Sci., 4:77-93.

Davis, L. C. 1969. The biostratigraphy of Peccary Cave, New-
ton County, Arkansas. Proc. Arkansas Acad. Sci., 23:192-
196.

Frick, C. 1937. Homed ruminants of North America. Bull.
Amer. Mus. Nat. Hist., 69:1-699.

Guilday, J. E., H. W. Hamilton, and A. D. McCrady. 1969.
The Pleistocene vertebrate fauna of Robinson Cave, Overton
County, Tennessee. Palaeovertebrata, 2:25-75.

Guilday, J. E., H. W. Hamilton, E. Anderson, and P. W.
Parmalee. 1978. The Baker Bluff Cave deposit, Tennessee,
and the Late Pleistocene Faunal Gradient. Bull. Carnegie
Mus. Nat. Hist, 11:1-76.

Hay, O. P. 1921. Descriptions of some Pleistocene vertebrates
found in the United States. Proc. U.S. Natl. Mus., 58 (2328):
83-146.

Kurten, B. 1975. A new Pleistocene genus of American moun-
tain deer. J. Mamm., 56:507-508.

Kurten, B., and E. Anderson. 1980. Pleistocene mammals of
North America. Columbia Press, New York, xvii + 443 pp.

Martin, P. S., and J. E. Guilday. 1967. A bestiary for Pleis-
tocene biologists. Pp. 1-62, in Pleistocene extinctions: the
search for a cause (P. S. Martin and H. E. Wright, Jr., eds.),
Proc. VII Congress INQUA, Yale Univ. Press, New Haven,
x + 453 pp.

Parmalee, P. W.. and R. D. Oesch. 1972. Pleistocene and
Recent faunas from the Brynjulfson Cave, Missouri. Illinois
State Mus., Rept. Invest., 25:1-52.

Peterson, O. A. 1926. The fossils of Frankstown Cave, Blair
County, Pennsylvania. Ann. Carnegie Mus., 16:249-315.

Semken, H. A. 1969. Paleoecological implications of micro-
mammals from Peccary Cave, Newton County, Arkansas.
Geol. Soc. Amer., South Central Sect., Abstr. Prog., 2:27.

Simpson, G. G. 1945. Principles of classification and a classi-
fication of mammals. Bull. Amer. Mus. Nat. Hist., 85:1 —
350.

Schultz, C. B., and E. B. Howard. 1935. The fauna of Burnet
Cave, Guadalupe Mountains, New Mexico. Proc. Acad. Nat.
Sci. Philadelphia, 87:273-298.

Wells, L. H. 1959. The nomenclature of South African fossil
equids. S. African J. Sci., 55:64-66.

Wyman, J. 1862. Chapter VII, Observations upon the remains
of extinct and existing species of Mammalia found in the
crevices of the lead-bearing rocks, and in the superficial ac-
cumulations within the Lead Region of Wisconsin, Iowa and
Illinois. Pp. 421-423, in Report on the Geological Survey
of the State of Wisconsin (J. Hall and J. D. Whitney), 1:1-
437.

Address: Department of Zoology, University of Toronto, Toronto, Ontario, Canada MSS 1A1 and De-
partment of Vertebrate Palaeontology, Royal Ontario Museum, Toronto, Ontario, Canada MSS 2C6.

GEOLOGICAL SETTING AND PRELIMINARY FAUNAL
REPORT FOR THE ST-ELZEAR CAVE,
QUEBEC, CANADA

Pierre LaSalle

ABSTRACT

The St-Elzear cave is located in limestone terranes of Middle
Silurian age. Because of its vertical entrance, the cave was a trap
for large and small animals, mostly small mammals, the majority
of which are still living in the area today. The cave was presum-
ably opened to the surface only in post-glacial time. However,
until excavations are terminated, no definite age can be assumed.
It is virtually certain that the cave area was overridden by glacier

ice at some time in the past because of the presence of erratics
in the area surrounding the cave, north and south of it. The
erratics appear to have been superimposed on the rubble formed
in situ possibly in pre-glacial time. Uranium-thorium dates sug-
gest that the precipitation of calcite inside the cave was already
under way in pre-glacial time, probably during the Sangamon or
some older inter-glacial.

INTRODUCTION

The St-Elzear cave (65°2T30" long. W; 48°14'20"
lat. N), located near St-Elzear-de-Bonaventure,
Quebec, Canada (Fig. 1 ), was discovered (or redis-
covered?) in December 1976 by local residents dur-
ing a snowshoe trek. The author started working on
the talus of the cave entrance in the autumn of 1 977,
with the hope of discovering old sediments, possibly
older than the last glaciation. As many readers are
possibly not aware, the glacial history of the Gaspe
Peninsula has been the subject of a long debate which
is still going on vigorously, and evidence for mul-
tiple glaciations is still lacking in most of the area.
However, it is generally agreed that the Gaspe Pen-
insula has been overridden by continental glaciers
several times in the past (Mailhiot, 1919; Coleman,
1922;Alcock, 1935, 1944; McGerrigle, 1952; Grant,
1977; Lebuis and David, 1977). The dispute is more
with the nature of the glaciation and its extent during
Wisconsinan time (Grant, 1977) than with the fact
of the glaciation itself (Prest, 1970; Mayewski et al.,
1981).

A preliminary bone collection was made from the
talus at the foot of the cave entrance in the autumn
of 1977, and was sent for study and identification
to the late John E. Guilday. The author thought the
bones would be mostly from bats. To the great sur-
prise of both the author and John E. Guilday, the
bone collection included remains of the collared
lemming ( Dicrostonyx hudsonius Pallas). Further
studies by Guilday have revealed the presence of
the remains of more than 30 species in the talus
deposit of the St-Elzear cave; most of these species
are still living in the area today. A short report has
been published on the fauna (LaSalle and Guilday,
1980) and this matter will be dealt with briefly as
the studies are still in progress. Rather, the purpose
of the present paper is to describe the bedrock ge-
ology in the area surrounding the cave entrance, and
report on the glacial geology as it is known today.

BEDROCK GEOLOGY

For a review of the literature relevant to the bed-
rock geology of the area, the interested reader is
referred to Badgley (1956) and Bourque and La-
chambre ( 1 980). St-Elzear cave is located (Fig. 2) in
the La Vieille Formation of Middle Silurian age.
This formation is generally composed of three mem-
bers of nodular and algal limestone. The contact
with adjacent formations is gradual and determined
arbitrarily (50% nodular limestone or calcareous
mudstones; see Bourque and Lachambre, 1980:57).

The cave is located on the southern flank of a syn-
clinorium, and some minor faulting is associated
with the placement of the cave. This situation has
presumably favored groundwater circulation and
enhanced the transport of carbonate in solution.
Presently, the water table is below the cave floor,
and its level is linked in some way with the water
level in the stream flowing in the ravine below the
cave entrance.

As shown in Fig. 3, the cave, because of its vertical

332

1984 LaSALLE -ST-ELZEAR CAVE, CANADA 333

Fig. 1. — Glacial features in the area surrounding St-Elzear Cave, Gaspe Peninsula, Quebec.

334

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

65° 30’

48° 15’

48° 00'

Fig. 2. — Bedrock geology of the area surrounding St-Elzear Cave, Gaspe Peninsula, Quebec, after Bourque and Lachambre (1980).

1984

LaSALLE- ST-ELZEAR CAVE, CANADA

335

Key to Fig. 2.

entrance, is a trap for small and large animals. The
talus itself has accumulated by layers under the in-
fluence of gravity, and some of the material (besides
the bones) has been washed or blown in from the
outside. The talus is thus composed of various plant
and animal remains, together with blocks fallen from

GLACIAL

The glacial geology of the Gaspe Peninsula has
been the subject of discussions and controversy for
almost one hundred years. The first report dealing
specifically with the subject area of this paper was
published by Chalmers (1887; 1906). Subsequently,
several geologists (mostly bedrock geologists; see

the roof of the chamber set in a matrix of clayey
brownish material. Moonmilk is also present in small
masses or layers surrounding stones or other geo-
logical objects. A crude stratification can be ob-
served in the talus with a slope of approximately
35°.

GEOLOGY

Bourque and Lachambre, 1980) have commented
and added some new data concerning the glacial
geology of the area. Mailhiot (1919) thought that
Mont Albert had been glaciated, on the basis of its
geomorphology and the presence of a “few erratic
blocks of granite and hornblende schist . . . (Mail-

336

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

- 5

- 10

- 15

m

( approx.)

Fig. 3. — Schematic cross-section of the entrance and talus of the St-Elzear Cave, Gaspe Peninsula, Quebec.

hiot, 1919:1 46). Coleman ( 1 922) studied the glacial
geology of the Gaspe Peninsula and concluded that
it had been covered by local glaciers. Alcock (1935)
reported the presence of granite erratics on the north
shore of the Baie-des-Chaleurs, and also observed
glacial striations in the same area. McGerrigle (1952)
summarized the data available at the time of pub-
lication of his report (1952) and seems to have put
to rest one of the most persistent hypotheses put
forth by botanists, especially by Fernald (1925). The
latter thought that because of the presence of arctic
endemics, that is, tundra type plants in their con-
temporaneous floras, the higher peaks of the
McGerrigle Mountains in the northern part of the
Gaspe Peninsula (and other elevated regions of east-

ern North America) had been nunataks during gla-
ciations. As McGerrigle (1952:38) concluded, quot-
ing several authors, the nunatak theory is not needed
to explain the geographic distribution of the plant
endemics in the contemporaneous flora of the north-
ern highlands of the Gaspe Peninsula. Scoggan ( 1 950)
reviewed the question of the nunataks in Gaspe from
the point of view of a botanist and stated that “the
presence of endemics in the arctic-alpine flora of
eastern North America offers no positive evidence
in the support of the nunatak theory” (Scoggan, 1950:

1 1 ). Flint et al. ( 1 942), after a visit to the Table Top
Mountains, concluded that Mont Albert had been
buried by glacier ice at some time in the past. They
also believed that the angular mantle of debris on

1984

LaSALLE- ST-ELZEAR CAVE, CANADA

337

the Shickshocks resulted from weathering of the lo-
cal bedrock following the glaciation that had erased
all traces of continental or other glaciations (Flint
et al., 1942).

The idea that the ice sheet over-rode the Gaspe
Peninsula at some time in the past became firmly
established with the discovery of long distance er-
ratics on the plateau and on the southern flank of
the Gaspe Peninsula (McGerrigle, 1952; Prest, 1970).
Among those far travelled erratics were the follow-
ing lithologies: granite gneiss, specular hematite and
anorthosite (Alcock, 1935; McGerrigle, 1952; Badg-
ley, 1956; Skidmore, 1965); all of the erratics pre-
sumably coming from the north shore of the St-
Lawrence River, that is, from the Canadian Shield.
More recently, Lebuis and David (1977) have stud-
ied the western part of the Gaspe Peninsula and used
erratics with some success to decipher the glacial
history of that area. Their history (Lebuis and Da-
vid, 1977) shows glacial events related to the con-
tinental ice sheet and to local ice caps that existed
prior to the continental glaciation and also during
its waning phase. Hughes (1981) proposed a model
of continental glaciation that seems to explain well
many of the glacial or other geological features of
the Gaspe Peninsula, which David and Lebuis (1984,
in press) seem to have demonstrated at least for the
western part of the Gaspe Peninsula.

Hughes’ model (1981) calls for areas of glacier ice
with a frozen base and consequently no erosion and
areas with a melting base with deposition and ero-
sion. The latter statement is certainly a simplifica-
tion and does not explain everything, that is, that
glacier ice once it reached the higher plateau did not
seem to carry much debris beside the long-distance
erratics. This phenomenon is explained by the pres-
ence of a frozen ice base picking up very little local
debris and depositing almost none except long dis-
tance erratics (Hughes, 1981; David and Lebuis,
1984, in press). The erratics are picked up in the
higher part of the ice from terranes that stand in
relief in the topography (Hughes, 1981; David and
Lebuis, 1984, in press).

Indeed, there are very few accumulations of gla-
cial sediments on the southern plateau of Gaspe and
in general one finds fluvio-glacial deposits in the
large river valleys where they form the higher ter-
races. Those sediments can be ascribed to paragla-
cial sedimentation (Church and Ryder, 1972) and
were presumably emplaced in late-glacial time. Es-
kers and other fluvio-glacial deposits are found at

the southern edge of the plateau and in the lowlands,
at least in the area studied by the writer (see Fig. 1).

All geologists studying the bedrock geology have
recognized at various occasions and locations the
presence of long and short distance erratics in the
interior of the Gaspe Peninsula. During the summer
of 1983, the glacial geology of the area immediately
surrounding the St-Elzear cave, covering approxi-
mately four NTS Sheets (New Carlisle, 22 A/3; New
Richmond, 22 A/4; Lac McKay, 22 A/5; Riviere
Reboul, 22 A/6), was studied.

In that region, especially on the plateau, deposits
are generally thin. They consist of an accumulation
in situ of local disintegrated bedrock to which has
been added erratics of various dimensions (Fig. 4).
The in situ accumulation can also be referred to as
a rubble or a saprolite. In other places, on the plateau
and on limestone terranes, one finds only an accu-
mulation (generally 1 m thick) of fine (silt and clay
size) material, very sticky when wet, with erratics
semmingly resting on it. Hence, the long distance
erratics (volcanics or granites) seem to have been
superimposed on the material (saprolite or rubble)
already in place. In other places the rubble contains
striated angular pieces of bedrock. There are also
areas of bare bedrock, that is, with a remarkable
absence of erratics and rubble at the surface of bed-
rock. This is true of the limestone area northeast of
St-Jogues.

The situations described above may have been
the result of: A) in situ disintegration of previously
striated bedrock; this would imply that the rubble
was formed following deglaciation; B) in situ dis-
integration of elacial material (till or other glacial
diamicton) since the last glaciation; this is very un-
likely as true glacial deposits are rather rare on the
plateau; C) the rubble having been formed in pre-
glacial time under periglacial conditions preceding
the last glacial readvance. The rubble would then
have been buried under some sort of glacial dia-
micton (till or other) as has been observed in one
instance on the plateau on the southern flank of the
Gaspe Peninsula. The rubble (or saprolite) would
include debris formed by rock disintegration (in situ)
and soil forming processes in pre-glacial time. This
would be in agreement with the concept of a glacier
with a frozen ice base, thus preventing erosion of
the previously formed saprolite or rubble (David
and Lebuis, 1984, in press).

In places, one can also observe striated bedrock
undisturbed together with local rubble and long dis-

338

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 4. — Main sites of accumulation of Pleistocene mammal assemblages for eastern North America (after Hibbard et al., 1965).

tance erratics. This is particularly true of an area
north of the cave entrance (Lac Huard valley) where
abundant volcanic erratics can be observed with
local (or short distance transport) rubble lying on a
striated bedrock surface. In the area around the cave
itself, no erratics had been found by the writer until
this summer (1983).

Rubble1 and colluvium were observed in the area
of the cave entrance in the fall of 1 982 by the author

1 According to the AGI glossary, p. 620, the word “rubble” refers to “A loose mass,
layer, or accumulation of rough, irregular, or angular rock fragments broken from larger
masses usually by physical (natural or artificial) forces, coarser than sand (diameter
greater than 2 mm), and commonly but not necessarily poorly sorted; the unconsoli-
dated equivalent of breccia.”

1984

LaSALLE-ST-ELZEAR CAVE, CANADA

339

in company with Allen D. McCrady. In 1983, er-
ratics (volcanics from the Lac McKay area) were
found north and south of the cave entrance (Fig. 5),
and there is no doubt that glacier ice overrode the
area at sometime in the past. Grant (1977) has sug-
gested that ice during the Wisconsinan maximum
(some 18,000 years ago) did not entirely cover the
north shore of Chaleurs Bay. This can be true only
if the erratics (volcanics2 and granites) have been
scattered by glacier ice during some earlier glacia-
tions other than the Wisconsinan. As have many
authors who have studied the area (Badgley, 1956;
Skidmore, 1965), volcanics have been found by the
author scattered all over the area studied south of
the cave entrance right up to the New Carlisle mo-
raine. After a careful search (along favorable ex-
posures made by new lumber roads), no volcanics

2 A porphyritic facies of volcanics which is found outcropping in the Lac McKay
area.

of the Lac McKay type have been observed north
of Lac McKay. Hence, in the area studied there is
no evidence recorded to date, that would suggest a
northward ice flow, or glacial transport of erratics.

A careful search for volcanic erratics south of the
cave entrance and south of St-Elzear village has
shown that the concentration of volcanic erratics
decreases gradually from the outcrop area near Lac
McKay to the New Carlisle moraine. In the area
between St-Elzear and the coast (above the marine
limit) the surface material (excluding ice-contact
drift) can hardly be called a till, except in a few
isolated sections; it is generally a diamicton that is
loose, that is, not very compact, and composed of
local bedrock and long and short distance erratics.
Generally, south of the cave entrance and south of
the plateau limit (above the marine limit) volcanic
erratics form less than 1% of the material in the
cobble size range.

DISCUSSION

Several radiocarbon dates have been obtained on
bone material within the talus and at the surface of
the talus. They all give ages of about 5,000 B.P. or
younger. Hence, it can be suggested tentatively that

the opening of the cave entrance is post-glacial and
that glacier ice covered the area until about 12 to
13,000 B.P. (QU-83, 13,580 ± 350 B.P.; QU-84,
13,450 ± 470 B.P.; Lebuis and David, 1977). How-

340

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

ever, this is very tentative, and no precise dating of
the material in the talus has been obtained because
samples of charcoal are too small. Besides charcoal,
bones of small mammals could be used. Apart from
the reservations associated with bone radiocarbon
dating, this course of action has not been followed
because it would sacrifice too much bone material
before the whole collection has been fully studied.
Uranium-thorium dates have been performed on
speleothems by Mel Cascoyne formerly of Me Mas-
ter University, Hamilton, Canada (Roberge and
Gascoyne, 1978); those dates suggest that solution-
precipitation processes were active inside the cave
at some time between 100,000 and 225,000 B.P.,
possibly during the Sangamon or some other older
interglacials (Fulton et al., 1984, in press; Henning
etal., 1983).

The fact that a chamber with solution-precipita-
tion features presumably dating back to 225,000
years B.P. has not been eroded suggests that the
glacier was not very effective as an erosional agent
in that area (unless some of those chambers were at
some depth below the surface). It seems to support
the concept of a frozen ice base at least in some part
of the Gaspe Peninsula (David and Lebuis, 1984,
in press) during the last glaciation. It is also possible
that the opening of the cave entrance is due in part
to glacial erosion.

St-Elzear cave with its vertical entrance about 1 3
m deep constitutes an ideal trap for small and large
animals. The diameter of the funnel-shaped upper
part of the entrance is about 4 m. A large animal
such as a moose sliding down a steep snowy slope
had no chance of recovering. Hence, as Guilday
(1971) remarked, to compare cave faunas, not only
the type of entrance is important (its diameter and
its angle to the vertical) but also its latitudinal and
longitudinal location. Another important factor is
the mode of accumulation, that is, by free falling of
the animals or by accumulation of feces of raptors
feeding on particular species (Guilday, 1971). Com-
paring the faunas in the deposit of New Paris No.
4 and Clark’s Cave Guilday stated: “A casual size
analysis of entrapped animals from the two faunas
would give no indication of their differing modes of

The main problem in the area around the cave
and in the area studied by the writer during the
summer of 1983 is the differentiation of the glacial
material from the local rubble. As the opinion has

deposition. The proportion of small-to-large mam-
mals in a raptor deposit is due to the selection bias
of the birds of prey, and is independent of the time
involved in forming such a deposit. The proportion
of small-to-large mammals in a tumble-in trap, like
a sinkhole, will increase with time. The slower the
rate of infill, the greater the proportion of small-to-
large mammals, because of their overwhelming ma-
jority in the surrounding fauna. This majority in-
creases their probability of entrapment.” (Guilday,
1971:71-72).

Hence the differences between the two cave faunas
(New Paris No. 4 and Clark’s) are found at the level
of the number of individuals for each species be-
cause of raptor bias and in the presence or absence
of some species because of the size of the entrance
or latitudinal location. In the case of the St-Elzear
cave because of its large entrance, the large mam-
mals are present in the faunal deposit, but also in
the lower levels, one finds besides other small mam-
mals, Dicrostonyx hudsonius, which has been found
only in the deposit of New Paris No. 4 and St-Elzear
in eastern North America.

Dicrostonyx is found today only west of Hudson’s
Bay in New Quebec, beyond the limit of the boreal
forest. Microtus xanthognathus is present at St-El-
zear, Clark’s, New Paris No. 4, and at many other
sites reported by Guilday (1971 :64-67). The present
geographical distribution of these small mammals
compared with their distribution in late-Wisconsi-
nan time is not only linked with late-glacial (late-
Wisconsinan events; Guilday, 1971:64-67) but also
with the particular ecology of each species (Guilday,
1971:64-67). It is interesting to note that Microtus
xanthognathus has a present western distribution
that recalls that of Femald’s endemics (Fernald,
1925; Scoggan, 1950) of the high mountains of
northern Gaspe. Presumably, south of the Wiscon-
sinan ice limits, there was a mixture of species with
different ecological requirements, mixtures that are
not found today (Guilday et al., 1 977). Those species
were dispersed into new habitats by the changing
environments in late-glacial time, finally finding
presumably more permanent niches in post-glacial
time (Guilday, 1971).

been expressed before (Flint et al., 1942) that the
formation of the rubble is probably post-glacial this
opinion is supported by the presence of pieces of
striated local bedrock in the rubble in some places.

1984

LaSALLE- ST-ELZEAR CAVE, CANADA

341

The following hypotheses can thus be formulated
concerning the origin of the rubble:

1) the rubble has been formed before the glaciation
and not been eroded; in this case, glacial ele-
ments have been superimposed on it; the pres-
ence of a saprolite under a glacial diamicton as
observed in one instance would tend to support
the hypothesis;

2) the rubble has formed since the glaciation from
local bedrock; in certain places, angular pieces
of local bedrock in the rubble show striations
and this would suggest that bedrock was over-
ridden by glacier ice at some time in the past and
then broken in situ by weathering processes since
the last glaciation;

3) a third hypothesis (undoubtedly preposterous) is
that the rubble or saprolite has survived several
glaciations with each time new elements being
superimposed by glacial action with the final
product being a mixture of glacial elements and
rubble; some or all the erratics would have pro-
gressed towards the south as the result of suc-
cessive glaciations; if the model of Hughes (1981)
and David and Lebuis (1984, in press) is appli-
cable to the last glaciation, it should apply to the
other older ones as well, because it calls for a
frozen ice base with practically no erosion in
most areas on the plateau, it is hardly possible
for most of the erratics to have been incorporated
into the glacier more than once.

The occasional occurrence of erratics had been
reported before by others (Badgley, 1956, Alcock,
1935; Skidmore, 1965). However, the systematic
examination carried out during the past summer
removes any doubt as to the presence of glacier ice
sometime in the past on the plateaus in the New
Carlisle, New Richmond, Lac McKay, and Riviere
Reboul areas.

Concerning the faunal remains that have been
recovered in the talus, the author hesitates to com-

ment except to remark that the distribution of Mi-
crotus xanthognathus may not have been too dif-
ferent from that of the plant endemics of Fernald
(1925) in late Wisconsinan time. Its present appar-
ent distribution results either from a lack of infor-
mation on the real geographical distribution of the
species, or its distribution has become restricted in
post-glacial time because of unfavorable ecological
conditions. This would suggest that the elements of
the fauna recovered in deeper levels in the talus were
probably together south of the glacial limit during
glaciation as the Quaternary periglacial records of
Phenacomys would suggest (Guilday and Parmalee,
1972); they then migrated back north following de-
glaciation, and their present distribution is the result
of more favorable conditions for each particular
species in selected areas. Presumably, this is true of
Dicrostonyx hudsonius, which was also present in
the faunal remains of New Paris No. 4 (Guilday et
al., 1964) together with Microtus xanthognathus,
but whose ( Dicrostronyx ) southern limit today co-
incides with the northern limit of the boreal forest,
west of Hudson Bay. This would also be true of
Phenacomys as it is suggested by its periglacial rec-
ords and its present distribution in North America
(Guilday and Parmalee, 1972).

The cave entrance may have been opened to the
surface by glacial action, but more probably became
open to the surface in post-glacial time; however,
solution and precipitation inside the cave may have
occurred during the Sangamon or some older non-
glacial period (Henning et al., 1983; Fulton et al.,
1984, in press).

Henning et al. (1983) have shown that the for-
mation of travertine seems to be concentrated dur-
ing certain periods, mostly between 90,000 and
130,000 BP and since 15,000 years ago. The period
between 90,000 and 1 30,000 corresponds to zone 4
and 5 of Shackleton and Opdyke (1973) and to parts
of the Sangamon interglacial (Fulton et al., 1984).

ACKNOWLEDGMENTS

Allen D. McCrady, Norman Wuerthele, and Robert Cardillo
of the Carnegie Museum of Natural History in Pittsburgh, PA,
have participated in the excavation and their contribution is
gratefully acknowledged. The author is also very thankful to the
residents of St-Elzear-de-Bonaventure too numerous to be named,
who were very helpful during the excavations and have accom-
panied the author several times to the cave.

The author is also grateful to Michel Bouchard, Universite de
Montreal, who has commented on some aspects of this paper,
and to P. P. David, Universite de Montreal, who spent a few
days in the field with the author during the summer of 1983.
Published with permission of the deputy-Minister, Quebec De-
partment of Energy and Resources.

342

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

LITERATURE CITED

Alcock, F. J. 1935. Geologie de la region de la Baie des Cha-
leurs. Commission Geologique du Canada, Mem., 183:1-
165.

. 1944. Further information on glaciation in Gaspe. Trans.

Roy. Soc. Canada, 38:15-21.

Badgley, P. C. 1956. La region de New Carlisle. Ministere des
Mines, Quebec, Geol. Rept., 70:1-40.

Bourque, P. A. and G. Lachambre. 1980. Stratigraphie du
Silurien et du Devonien basal du sud de la Gaspesie. Min-
istere de l’Energie et des Ressources, Quebec, Spec, paper,
30:1-123.

Coleman, A. P. 1922. Physiography and glacial geology of Gaspe
Peninsula, Quebec. Bull. Geol. Surv. Canada, 34:1-52.

Chalmers, R. 1 887. Surface geology. Northern New Brunswick
and Southeastern Quebec. Geological and Natural History
Survey of Canada, Part M. Annual Report 1886, Dawson
Brothers, Montreal, 39 pp.

. 1906. Surface geology of eastern Quebec. Geol. Surv.

Canada; Ann. Report for the year 1904 (part A):250-263.

Church, M., and J. M. Ryder. 1972. Paraglacial sedimenta-
tion: A consideration of fluvial processes conditioned by
glaciation. Geological Soc. America Bull., 83:3059-3072.

David, P. P., and J. Lebuis. 1984. The last glacial maximum
and deglaciation of the western half of Gaspe Peninsula and
adjacent areas, Quebec, Canada; in press.

Fernald, M. L. 1925. Persistence of plants in glaciated areas
of boreal America. Memoirs of the Academy of Arts and
Sciences, XV(III):24 1-342.

Flint, R. F., M. Demorest, and A. L. Washburn. 1942. Gla-
ciation of Shick Shock Mountains, Gaspe Peninsula. Geol.
Soc. America Bull., 53:121 1-1230.

Fulton, R. J., P. F. Karrow, P. LaSalle, and D. R. Grant.
1984. Summary of Quaternary stratigraphy and history,
eastern Canada. In press.

Grant, D. R. 1977. Glacial style and ice limits, The Quaternary
Stratigraphic record, and changes of land and ocean level in
the Atlantic provinces, Canada. Geog. phys. et Quatemaire,
31:247-260.

Guilday, J. E. 1971. The Pleistocene history of the Appala-
chian mammal fauna. In The distributional history of the
biota of the southern Appalachians, Part III (P. C. Holt, ed.)
Research Monogr. Virginia Polytechnic Inst. State Umv.,
Blacksburg, Virginia, 4.

Guilday, J. E., P. S. Martin, and A. D. McCrady. 1964. New
Paris No. 4. A Pleistocene cave deposit in Bedford County,
Pennsylvania. Bull. Nat. Speleol. Soc., 26:121-194.

Guilday, J. E., P. W. Parmalee, and H. W. Hamilton. 1977.

The Clark’s cave bone deposit and the late Pleistocene pa-
leoecology of the Central Appalachian Mountains of Virgin-
ia. Bull. Carnegie Mus. Nat. Hist., 2:1-87.

Guilday, J. E., and P. W. Parmalee. 1972. Quaternary peri-
glacial records of voles of the genus Phenacomys Merriam
(Cricetidae: Rodentia): Quaternary Research, 2:170-175.

Henning, G. J., R. Grun, and K. Brunnacker. 1983. Speleo-
thems, travertines and paleoclimates. Quaternary Res. 20:
1-29.

Hibbard, C. W„ D. E. Ray, D. E. Savage, D. W. Taylor, and
J. E. Guilday. 1965. Quaternary mammals of North Amer-
ica. Pp. 509-526, in Quaternary of the United States (H. E.
Wright and D. G. Frey, eds.), Princeton Univ. Press, 922

pp.

Hughes, T. J. 1981. Numerical reconstruction of paleo-ice
sheets. Pp. 222-261, in The last great ice sheets (G. H. Den-
ton and T. J. Hughes, eds.), John Wiley and Son, New York.

Lasalle, P„ and J. E. Guilday. 1980. Caveme de St-Elzear-
de-Bonaventure. Ministere de l’Energie et des Ressources,
Quebec, DPV-750, 31 pp.

Lebuis, J., and P. P. David. 1977. La stratigraphie et les evene-
ments du Quatemaire de la partie occidentale de la Gaspesie,
Quebec. Geog. Phys. Quatemaire, 21:275-296.

Mailhiot, A. 1919. Geology of Mount Albert, County of Gaspe,
P.Q. Report on Mining Operations in the Province of Quebec
for 1918, pp. 146-151.

Mayewski, P. A., G. H. Denton, and T. J. Hughes. 1981. Late
Wisconsin ice sheets of North America. Pp. 67-178, in The
last great ice sheets (G. H. Denton and T. J. Hughes, eds.)
John Wiley and Son, New York.

McGerrigle, H. W. 1952. Pleistocene glaciation of Gaspe Pen-
insula. Trans. Royal Soc. Canada, 46:37-51.

Prest, V. K. 1970. Quaternary geology of Canada. Pp. 676-
764, in Geology and economic minerals of Canada. Eco-
nomic Geology Report No. 1, Fifth ed.

Roberge, J., and M. Gascoyne. 1978. Premiers resultats de
datations dans la grotte de St-Elzear, Gaspesie, Quebec. Geog.
Phys. Quatemaire 32:287.

Scoggan, H. J. 1950. The flora of Bic and the Gaspe Peninsula,
Quebec. Bull. Nat. Mus. Canada, 1 15:1-309.

Shackleton, N. J., and N. D. Opdyke. 1973. Oxygen isotope
and paleomagnetic stratigraphy of equatorial Pacific core
V28-238: Oxygen isotope temperatures and ice volumes of
a 105 year and 106 year scale. Quat. Res., 3:39-55.

Skjdmore, W. B. 1965. Region d’Honorat-Reboul, Comte de
Bonaventure. Ministere des Richesses Naturelles, Quebec,
Rapport Geologique, 107:1-36.

Address: Ministere de l’Energie et des Ressources, 1620 Boulevard de l’Entente Quebec City, Quebec G1S
4N6, Canada.

1984

LaSALLE- ST-ELZEAR CAVE, CANADA

343

APPENDIX I

Summary of Faunal Remains from
Caverne de St-Elzear-de-Bonaventure,
1977-1978, by the late John E. Guilday

This report covers vertebrate and invertebrate re-
mains excavated from Caverne de St-Elzear-de-
Bonaventure by Dr. Pierre LaSalle, Geological Ex-
ploration Service, Department of Energy and Re-
sources, Province of Quebec, assisted by A. D.
McCrady, Section of Vertebrate Fossils, Carnegie
Museum of Natural History, Pittsburgh, Pennsyl-
vania, during the field seasons of 1977 and 1978.
All items were from the talus cone at the base of
the entrance shaft. They are currently housed at Car-
negie Museum of Natural History. Insect remains
were identified by Dr. George E. Ball, The Univer-
sity of Alberta; reptiles by Dr. Thomas Van Deven-
der, University of Arizona, Tucson; birds by Dr.
Paul W. Parmalee, University of Tennessee, Knox-
ville; seeds by Dr. C. R. Gunn, Plant Taxonomy
Laboratory, U.S.D.A., Beltville, Maryland.

A minimum of 532 invertebrates and 4,679 ver-
tebrates were recovered; one beetle, 531 land snails,
181 amphibians (three species), two reptiles (one
species), four birds (four species), and 4,503 mam-
mals (34 species). The vertebrate remains (97.9%)
consisted primarily of skeletal elements of small
mammals (smaller than a hare)— 74.3% of these were
from 14 species of small rodents and 24.5% from
seven species of small insectivores. See faunal list,
Table 1.

The minimum numbers of individual animals per
level were derived from bone or tooth counts. These
are only close approximations; a left lower jaw from
one level and a right lower jaw from an adjacent
level may conceivably have belonged to the same
individual, but have been recorded as two animals
when tallied stratigraphically, and some skeletons
were reduced by talus abuse to unidentifiable frag-
ments.

Because the stratigraphic levels of the 1977 ex-
cavation differ somewhat from those of the 1978
excavation due to talus slope, and because strati-
graphic levels were arbitrary rather than natural, the
minimum numbers of individuals from both years
were lumped into four stratigraphic zones that ap-
peared to have some biological justification. This
insured that the minimum number of individuals
from each zone would be large enough for the rel-
ative fluctuations between zones to have some sta-

tistical validity, and minimized the chances of
counting the same animal more than once from scat-
tered remains (the skeleton of a single black bear
cub, for instance, occurred throughout a vertical
depth of at least 77 cm).

The four stratigraphic zones were characterized
by the presence or absence of certain distinctive
genera of small mammals and were organized from
the 1977 and 1978 stratigraphic levels as follows:

Zone 1: surface of talus and general cave floor.
Zone 2: 1-49 cm (1977) plus 1-60 cm (1978), talus
only. Phenacomys and Dicrostonyx ab-
sent.

Zone 3: 49-1 19 cm (1977) plus 60-80 cm (1978).

Phenacomys present, Dicrostonyx absent.
Zone 4: 119-168cm(1977)plus80-120cm(1978).

Both Phenacomys and Dicrostonyx pres-
ent.

All amphibians, reptiles, and birds from the de-
posit are present in the area today. One species of
mammal, the yellow-cheeked vole, Microtus xan-
thognathus, is presently a native of Alaska and Kee-
watin and has not been recorded east of Hudson’s
Bay. It was, however, a common rodent throughout
the Appalachian Mountain region from Tennessee
north at least to Pennsylvania during the late Wis-
consinan, in boreal faunas that predated the glacial
recession. Its presence at St-Elzear (one individual
so far) suggests that this now western taiga species
survived to a relatively late date in the Gaspe. Five
other species of mammals have not been reported
from the Recent mammal fauna of the Gaspe. The
Ungava collared lemming, Dicrostonyx hudsonius,
is now confined to the tundra of northern Ungava.
The Arctic hare, Lepus arcticus, another tundra
species, reaches its southern limit in Newfoundland.
Both occur rarely in Zone 4, suggesting that lower
levels may hold further evidence of ecological
changes. The heather vole, Phenacomys interme-
dins, a boreal forest rodent confined to Zones 3 and
4, is not found south of the Gulf of St. Lawrence
today. The other two species hitherto unreported
from the Gaspe, the Arctic shrew, Sorex arcticus,
and the least weasel, Mustela nivalis, have been found
in adjacent areas and their former presence was not
unexpected. The Gaspe shrew, Sorex gaspensis, and
(to the best of my knowledge) the Arctic hare have
not hitherto been reported in the fossil state.

344

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 1 .—Faunal list for Caverne de St-Elzear-de-Bonaventure, Quebec, 1977-1978 field collections.

Stratigraphic levels

Zone Zone Zone Zone Total

Scientific name Common name MMC 12 3 4 MNI

Vertebrates

Class Mammalia, order Insectivora, family Talpidae
Condylura cristata (Linnaeus)

Family Soricidae
Sorex arcticus Kerr
Sorex cinereus Kerr
Sorex fumeus Miller
Sorex gaspensis Anthony
Sorex palustris Richardson
Microsore x hoyi (Baird)

Blarina brevicauda (Say)

Order Chiroptera, family Vespertilionidae
Myotis lucifugus (LeConte)

Order Lagomorpha, family Leporidae
Lepus americanus Erxleben
Lepus arcticus Ross

Order Rodentia, family Sciuridae
Marmota monax (Linnaeus
Tamias striatus (Linnaeus)

Tamiasciurus hudsonicus (Erxleben)
Glaucomys sabrinus (Shaw)

Family Cricetidae
Peromyscus maniculatus (Wagner)

Family Arvicolidae
Clethrionomys gapperi (Vigors)
Synaptomys cooperi Baird
Synaptomys borealis (Richardson)
Phenacomys intermedius Merriam
Dicrostonyx hudsonius (Pallas)

Microtus pennsylvanicus (Ord)

Microtus chrotorrhinus (Miller)

Microtus xanthognathus (Leach)

Family Zapodidae
Napaeozapus insignis (Miller)

Zapus hudsonius (Zimmerman)

Family Castoridae
Castor canadensis Kuhl

Family Erethizontidae
Erethizon dorsatum (Linnaeus)

Order Carnivora, family Ursidae
Ursus americanus Pallas

Family Mustelidae
Mustela nivalis Linnaeus
Mustela erminea Linnaeus
Gulo gulo (Linnaeus)

Order Artiodactyla, family Cervidae
Alces alces (Linnaeus)

Rangifer tarandus (Linnaeus)

Star-nosed mole

7

Arctic shrew

Masked shrew

117

Smoky shrew

29

Gaspe shrew

16

Water shrew

6

Pigmy shrew

23

Short-tailed shrew

54

Little brown bat

14

Snowshoe hare

24

Arctic hare

—

Woodchuck

23

Chipmunk

48

Red squirrel

70

Northern flying squirrel

7

Deer mouse

319

Red-backed vole

423

Southern bog lemming

3

Northern bog lemming

1

Heather vole

—

Ungava collared lemming

-

Meadow vole

158

Rock vole

22

Yellow-cheeked vole

—

Woodland jumping mouse

159

Meadow jumping mouse

25

Beaver

3

Porcupine

11

Black bear

6

Least weasel

_

Ermine

20

Wolverine

—

Moose

1

Caribou

10

—

1

—

—

1

1

3

6

10

44

150

311

315

820

3

9

17

14

43

—

1

4

8

13

—

—

1

4

5

5

23

52

46

126

8

13

41

24

86

1

2

5

1

9

7

3

3

9

22

—

—

—

1

1

1 1

4

3

4

22

—

—

2

2

4

1

—

1

3

5

1

1

3

2

7

1

4

10

18

33

57

235

366

470

1,128

7

40

19

17

83

36

125

239

105

505

—

—

22

34

56

—

—

—

7

7

10

48

37

36

131

16

279

432

383

1,210

—

—

—

1

1

14

39

37

30

120

1

1

1

_

3

1 1

i

i

1 1

2

2

3

18

5

1

1

2

9

2

1

1

4

3

2

1

-

1

4

7

1

1

1

2

1 1
1

1984

LaSALLE -ST-ELZEAR CAVE, CANADA

345

Scientific name

Table 1 . — Continued.

Stratigraphic levels

Zone

Zone

Zone

Zone

Total

Common name

MMC

1

2

3

4

MNI

Mammal total (cave)

Class Aves, order Passeriformes, family Sittidae
Sitta canadensis Linnaeus Red-breasted nuthatch

Family Corvidae

Cyanocitta cristata (Linnaeus) Blue jay

Family Fringillidae

Loxia cf. leucoptera Gmelin White-winged crossbill ?

Species, indeterminate

Class Amphibia (data incomplete), order Caudata, family Ambystomidae
Ambystoma laterale Hallowed Blue-spotted salamander

Order Salientia, family Bufonidae
Bufo americanus Holbrook American toad

Family Ranidae

Rana palustris LeConte Wood frog

Invertebrates

Class Gastropoda (incomplete data)
Iriodopsis sp.

Anguispira sp.

Succinca sp.

Land snails
Land snails
Land snails

Other species of land snails to be identified
Class Insecta, order Coleoptera, family Carabidae
Pleroslichus stypicus Land beetle

MMC = Modem mammal collection (i.e., specimens of living species in museum mammal collections).

351 985 1,614 1,553 4,503

1

1

2

14 104 20 37 175

4

The cave fauna excavated to date suggests boreal
forest conditions at least as severe as those of today,
with an indication of nearby tundra conditions dur-
ing lower level times. There are some striking dif-
ferences between the relative composition of the
Recent mammalian fauna and that of the cave de-
posit, as well as internal changes within the deposit
itself that point to more boreal conditions. For in-
stance, a survey of Recent mammals collected in
the Gaspe, in North American museum and uni-
versity collections, shows that the deer mouse, Pero-
myscus maniculatus, is one of the Peninsula’s com-
monest small mammals— 20.4%. Elowever, in the
four zones of the deposit its percentage occurrence
varies from 0.3% to 1.16%. This woodland rodent
is rare in northern boreal forests but its high per-
centage today may also be due to ecological changes
brought on by timbering practices. The percentage
of the red-backed vole, Clethrionomys gapperi, a

characteristic boreal forest vole, increases with depth
from 17.6% in Zone 1 to 30.5% in Zone 4. The
northern bog lemming, Synaptomys borealis, is
known from the Recent fauna of the Gaspe by a
single individual trapped on Mt. Jacques Cartier,
and now in the Canadian National Museum of Nat-
ural History. Nevertheless, it was one of the more
common small mammals at all levels of the deposit,
ranging from 6.8% to 14.85% — a total of 505 in-
dividuals from all levels. The masked shrew, Sorex
cinereus, more common in boreal regions, increases
relatively with depth. It comprises 7.5% of the Re-
cent fauna, but 13.6% to 20.45% in the lowest levels
of the deposit. Relative numbers of the woodland
jumping mouse, Napaeozapus insignis, however,
show a reduction in numbers with depth. It com-
prises 10.19% of the Recent fauna, but decreases
with depth from 4.3% to 1.94% in the cave deposit,
again suggesting increasingly harsh conditions cor-

346

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

related with stratigraphic depth. This is also sug-
gested by the appearance of Dicrostonyx hudsonius
and Lepus arcticus, true tundra forms, in Zone 4.

Upon the initial discovery of the cave, skeletons
of moose (five adults and two calves), one caribou
calf, five black bears, 1 1 porcupines, and two wol-
verines were found scattered about the cave floor.
Many of their bones showed signs of rodent gnawing
and at least one had been heavily gnawed by a large
carnivore, probably bear or wolverine, so that all
animals were not killed by the initial fall, but sur-
vived to die of hunger and exposure. Many of the
bones were broken and the breakage patterns of some
of the large bones will be studied by Dr. Gary Haynes,
Department of Anthropology, Smithsonian Insti-
tution. He is trying to differentiate breakage patterns
associated with early man as opposed to those that
can occur under natural conditions. There is no evi-
dence of man at the cave, so that the cave collection
will serve as an uncontaminated sample of natural
breakage patterns. Some of the moose bones have
acquired an unusual polish with rounded surfaces

on broken edges that an archaeologist would have
called man-made. Obviously some natural mecha-
nism is at work here that requires further study.

Based upon tooth wear and replacement, four of
the entrapped moose were from three to four years
old (at least two were bulls bearing antlers). One was
a yearling about 1.5 years of age, and two calves
were between 4-5 months of age. Assuming that
most moose calve in June, these fell in in September
or October. The age of the single caribou calf sug-
gests this as well.

None of this large collection of vertebrate re-
mains, unique in eastern North America north of
the Wisconsinan terminal moraine, has been studied
in detail. Research efforts to date have been con-
centrated primarily on sorting, identifying, and cat-
aloguing this large collection. The evidence to date
for biologic change with depth is strong and further
excavation will almost certainly lead to increasing
our knowledge of the late Pleistocene/early Holo-
cene history of the area.

APPENDIX II

Radiocarbon dates pertinent to talus history of the St-Elzear cave.

Lab no.

Material

Excavation levels

Age BP

QU-714

Bone fragments of moose (Alces alces )

St-Elzear 1

410 ± 120

QU-715

Wood

St-Elzear 2

Modem

QU-717

Bone fragments of moose (Alces alces)

St-Elzear 4

4,400 ± 130

QU-745

Bone fragments of moose (Alces alces)

St-Elzear 24

4,390 ± 120

QU-978

Wood

78-1

940 ± 120

GX-7016

Bone fragments of moose (Alces alces)

Surface of talus

5,110 ± 150
(13C corrected)

APPENDIX III

230Th/234U ages (modified from Roberge and Gas-
coyne, 1978).

Lab no.

Material

Age

Ka » 103

+ 1 a

Ka x 103

— 1 a
Ka x 103

77023-1

Stalagmite

223.9

30.7

25.0

77027-1

Stalagmite

204.8

13.2

12.0

77039-1

Stalagmite

152.8

20.8

18.0

77026-1

Stalagmite

135.9

7.6

7.2

77040-2

Stalagmite

135.3

9.6

8.9

77040-1

Stalagmite

129.9

6.2

5.9

77029-1

Stalagmite

102.5

7.6

7.1

MEADOWCROFT ROCKSHELTER AND THE PLEISTOCENE/
HOLOCENE TRANSITION IN SOUTHWESTERN
PENNSYLVANIA

J. M. Adovasio, J. Donahue, R. C. Carlisle, K. Cushman,

R. Stuckenrath, and P. Wiegman

ABSTRACT

Meadowcroft Rockshelter (36WH297) is a deeply stratified
multicomponent site in Washington Co., southwestern Pennsyl-
vania. The 1 1 well-defined stratigraphic units identified at the
site span at least 1 6,000 and probably 1 9,000 years of intermittent
occupation by human groups representing all the major cultural
stages/periods now recognized in the prehistory of northeastern
North America. Throughout the temporal sequence, the site served
as a locus for hunting, collecting, and food processing activities
which involved the seasonal exploitation of the Cross Creek Val-
ley and contiguous uplands. Presently, Meadowcroft Rockshelter
represents the earliest well-dated evidence of human beings in
the New World and demonstrates the longest occupation se-
quence in the Western Hemisphere.

All of the extant ecofactual information including macrofaunal,
microfaunal, macrofloral, and microfloral remains as well as var-
ious categories of geological/geomorphoiogical data suggest that
from ca. 9,300 or 9,000 B.C. (ca. 1 1,250 or 10,950 B.P.) to the
present, the environment of Cross Creek was essentially modem
in aspect. It also appears that the Late Pleistocene environment
in that part of Cross Creek immediately adjacent to the rock-
shelter was not radically different from that of today. Given these
conditions, it appears that the Pleistocene/Holocene transition
in this portion of Pennsylvania was a low amplitude event of
relatively short duration.

INTRODUCTION

John Guilday and Paul Parmalee (1982:163) ob-
served in commenting on the archaeofauna from
Meadowcroft Rockshelter that . . no other North
American archaeological site has yielded the re-
mains of so many vertebrate species.” Indeed, the
analysis and interpretation of the nearly one million
element vertebrate faunal assemblage from Mead-
owcroft was not only the most extensive scrutiny of
archaeologically recovered faunal remains that John
Guilday (and his co-workers) had ever undertaken;
it was also the last large archaeofaunal assemblage
he examined. Given the size and diversity of this
assemblage, the analysis phase occupied much of
John’s time as well as that of those by whom he was
assisted from mid- 1975 to 1980 when the exhaus-
tive report (Guilday et al., 1980) on this material
was completed for ultimate incorporation into the
final Meadowcroft synthetic volume (Adovasio et
al., n.d. b).

Clearly, John expended incredible time and pa-
tience on examining and recording the details of the
Meadowcroft fauna and in assessing the potential
of that data base for furthering our understanding
of the prehistory and paleoecology of southwestern
Pennsylvania. Unfortunately for all of us, John will
not be able to witness the ultimate fruition of his
countless hours of observation and attention to de-

tail or to participate directly, and for the credit he
so richly deserves, in the synthesis of all the Mead-
owcroft data sets.

At times, particularly at the beginning of our as-
sociation, John was uncomfortable with what the
Meadowcroft faunal materials seemed to reflect.
They sometimes did not square with the “big pic-
ture” or with the ecological interpretations and re-
constructions to which he subscribed and which were
based on his personal examination of faunal remains
from many other sites. Nevertheless, John unwaver-
ingly adhered to a cardinal rule of science— to follow
where the data lead and to keep an open mind. In
the final analysis, this approach is John’s greatest
legacy to a new generation of faunal analysts, and
at the end, it allowed him to attain the synthesis
that comes when the logical conclusions of hard
scientific observation fly in the face of what is es-
tablished, what is perceived to be true or simply
what one would like to believe is so.

One of the critical facets of the Meadowcroft eco-
factual record tantalizingly touched upon, but by no
means resolved, by the analysis of the vertebrate
faunal assemblage alone is a subject that was dear
to John’s scholarly heart— the nature of the Pleis-
tocene-Holocene “transition” not only in Pennsyl-
vania, but in eastern North America as a whole. The

347

348

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

interpretations that follow represent our current un-
derstanding of the Meadowcroft/Cross Creek data
base as it applies to that theme. Of course, these

interpretations are subject to revision before the fi
nal Meadowcroft/Cross Creek publication (Ado
vasio et al., n.d. b) appears.

GENERAL SETTING

Location and General Geology

Meadowcroft Rockshelter (36WH297) is a stratified, multi-
component site 48.27 air km (78.84 km via road) southwest of
Pittsburgh, Pennsylvania, and 4.02 surface km northwest of Av-
ella in Washington Co., Pennsylvania (Fig. 1). The site is on the
north bank of Cross Creek, a small tributary of the Ohio River
which lies some 12.16 km to the west. The exact location of the
site is 40°17T2"N; 80°29'0"W (U.S.G.S. Avella, Pennsylvania,
7.5' Quadrangle).

Meadowcroft Rockshelter is oriented approximately east to
west and has a southern exposure. It is ca. 15.06 m above Cross
Creek and 259.9 m above sea level. The area protected by the
extant overhang is ca. 65 m2, and the overhang itself is ca. 13m
above the modem surface of the site. In addition to water in
Cross Creek, springs are abundant near the rockshelter. The pre-
vailing wind is west to east across the mouth of the rockshelter.
This nearly continuous ventilation quickly clears both smoke and
insects.

Physiographically, Meadowcroft is on the Unglaciated Appa-
lachian or Allegheny Plateau west of the Valley and Ridge Prov-
ince of the Appalachian mountains, and northwest of the Ap-
palachian Basin. The rockshelter is formed beneath a cliff of
Morgantown-Connellsville sandstone, a thick fluvial or channel
sandstone within the Casselman Formation (Flint, 1955) of Penn-
sylvanian age.

Physiography

Meadowcroft Rockshelter is in a maturely dissected topograph-
ic region. More than one-half of the 14,164.3 ha encompassed
by the Cross Creek watershed are in valley slopes; upland and
valley bottoms are in the minority. Maximum elevations in the
Cross Creek drainage are generally above 396 m. At the divides
on the east, elevations are above 426 m. Elevations at stream
level are 310 m at Rea on the South Fork, 276 m at Avella, and
193 m normal pool level at the confluence with the Ohio River.

The main stem of Cross Creek flows for ca. 31.3 km. The
maximum north-south watershed width is approximately 1 5 km.
The prevailing stream pattern is dendritic, and numerous small
creeks and runs supply the main stem of Cross Creek. The steep
gradient headwaters and tributaries of Cross Creek have their
sources in the hills of central Washington Co., Pennsylvania.
Cross Creek and its major tributaries ofNorth, Middle, and South
Forks have an average gradient of 0.4%. Most of the gradient is
in the upper watershed where the streams are small and of low
volume. The drainage is northwest to west toward the West Vir-
ginia/Ohio border and the Ohio River.

The Cross Creek drainage pattern is very asymmetric; the
northern tributaries are much shorter than those on the south.
The drainage area south of Cross Creek is therefore much larger
than that to the north. This condition is probably the result of a
drainage pattern superimposed on a 3°-5° regional dip.

Adjacent to Cross Creek are four other major watersheds. Har-
mon Creek to the north parallels Cross Creek as does Buffalo

Creek to the south. Both are quite similar to Cross Creek in
structure, and both flow into the Ohio River. On the northeast
is Raccoon Creek. This is a north-flowing stream which joins the
Ohio River just west of the confluence with the Beaver River.
To the east and southeast, the large Chartiers Creek watershed
flows north and joins the Ohio River near its inception at Pitts-
burgh.

Present area topography developed during the latter Pleisto-
cene when increased precipitation and runoff caused extensive
downcutting. The area was unaffected directly by glacial ice as
the mapped Wisconsinan Kent moraine (Woodfordian age) is
some 50 km north of the site, whereas the later Lavery till is ca.
83 km to the north.

Contemporary Climate

The present continental climate of the study area is character-
ized by a wide seasonal temperature range and by moderate pre-
cipitation that falls principally during the warmer parts of the
year. The yearly temperature in Washington Co. averages 10.6°C
with extremes ranging from -28.9°C during the winter months
to 32.2°+C in July and August. The frost-free growing season
averages 150 days. Precipitation averages approximately 1,016
mm per year with about 560 mm falling during the growing
season. The amount of precipitation correlates closely with ele-
vation and roughness of terrain, particularly on windward slopes.
Temperatures tend to be lower in hilly areas than in more level
places due to the effects of elevation and air drainage. Night
temperatures are generally colder, and day temperatures are slightly
higher on the valley bottoms than on hilltops.

Contemporary Flora

Vegetative cover in the Cross Creek watershed today is a mo-
saic of second growth forest, abandoned fields, and pastures,
agricultural tracts, and residential areas. Variation in these plant
communities is promoted by the degree and duration of distur-
bances as well as by degree and exposure of slope. Human dis-
turbance is now the principal determinant of vegetative cover.

Forest once covered essentially all of the watershed. Several
forest types were common, including mixed oak on the hilltops
and steep, south-facing slopes. Mixed mesophytic forest covered
the more mesic north-facing slopes and headwater coves. Riv-
erine vegetation spread along the alluvial flood plains.

Woodland now covers approximately 40% of the area. Little
of this is old growth, and only a few stands approach the typical
composition of virgin forest. Most of the area was heavily de-
forested during the early and middle 1800s when the land was
first settled and cleared for cultivation. Areas allowed to re-es-
tablish forest have been logged repeatedly with resulting second-
ary forests.

Where woodlands have matured for a long period, they are
beginning to resemble original stands (Wiegman, 1977). Mixed
oak forests are common on the drier ridgetop sites and south-
facing slopes, and they are typical on thin, stoney soils of mod-

1984

ADOVASIO ET AL. — MEADOWCROFT ROCKSHELTER

349

l l

Okm 10km

Fig. 1. — General view of the Upper Ohio Valley and contiguous regions showing maximum southward extension of the Wisconsinan
glacial advance (after Johnson, 1981).

erate fertility. Dominant species are white oak ( Quercus alba)
and red oak ( Quercus rubra). Associated species include red ma-
ple ( Acer rubrum), sugar maple (Acer saccharum). hickories (Car-
ya sp.), black birch (Betula lenta). American beech (Fagus gran-
difolia), tulip tree (Liriodendron tulipifera), chestnut oak (Quercus
prinus). black oak (Quercus velutina), and scarlet oak (Quercus
coccinea).

Vegetative layers below the canopy are sparsely populated.
Typical shrubs include juneberry ( Amelanchier arborea), maple-
leaved viburnum ( Viburnum acerfolia), flowering dogwood (Cor-
nus florida), poison ivy (Rhus radicans), summer grape ( lit is
aestivalis), and occasional small patches of mountain laurel (Kal-
mia latifolia). Individual plants are scattered or occur in small
groups creating a rather open understory.

350

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

The herbaceous layer is equally poor in species diversity and
population. During the spring, early saxifrage (, Saxifraga virgin-
iensis), rupanemone ( Anemonella lhalictroides ), and starry cam-
pion (Silene stellata) are scattered over the rocky surface. Later
dominants include shinleaf (Pyrola elliptica ), spotted wintergreen
(Chimaphila maculata ), tick trefoil ( Desmodium sp.), and black
snakeroot ( Sanicula marilandica). Grasses are also found in the
dry oak woods and include linear leaf panicum (Panicum lin-
earifoliwn ), bushy panic grass (Panicum dichotomum), and Can-
ada brome grass (Bromus purgans).

On moister north-facing slopes and headwater coves, mixed
mesophytic forest is found. These stands occur on rich, well-
drained, fertile soils with deep humus layers. Woodland domi-
nants include sugar maple, American beech, tulip tree, white oak,
red oak and white basswood ( Tilia heterophylla). There are many
other species in this forest which are common, but not dominant.
These include red maple, yellow birch (Betula allegheniensis),
hickories, white ash (Fraxinus americana), black walnut (Juglans
nigra), cucumber tree (Magnolia acuminata), black gum (Nyssa
sylvatica), wild black cherry (Prunus serotina), and eastern hem-
lock ( Tsuga canadensis).

The shrub layer is generally denser and more diverse than that
of the mixed oak type. Prominant species include juneberry,
flowering dogwood, witch-hazel (Hamamelis virginiana), spice-
bush (Lindera benzoin), pawpaw (Asimina triloba), Virginia
creeper (Parthenocissus quinquefolia), wild plum (Prunus amer-
icana), and scattered occurrences of redbud (Cercis canadensis)
on neutral or alkaline soils.

The herbaceous layer contains a myriad of species and is es-
pecially rich during the prevemal and vernal periods. Trillium
is a prominent family with the large trillium (Trillium grandijlo-
rum) the most common. Other species include numerous ex-
amples of violets ( Viola sp.), blue phlox (Phlox divaricata), trout
lily (Erythronium americanum), larkspur (Delphinium tricorne),
spring beauty (Claytonia virginica), toothwarts (Dicentra laci-
niata and Dicentra diphylla), solomon’s-seal (Polygonatum bi-
florum), and jack-in-the-pulpit (Arisaema atrorubens). After the
canopy has closed there are fewer flowering species, but the gen-
eral dense coverage remains.

The most prevalent forest type in the Cross Creek watershed
today is mixed mesophytic. It most often occurs on steep slopes
adjacent to larger streams. Here, the land is unsuitable for cul-
tivation, and repeated logging provides a source of disturbance.
In less accessible areas, the forest is well-developed maturing
second growth.

Riverine forests grow on the flood plains of lower Cross Creek.
Sycamore (Piatanus occidentalis) and black willow (Salix nigra)
are common in small groves and in bands adjacent to the stream.
These woodlands blend with the mesophytic, and it is often dif-
ficult to draw distinct boundaries. Riverine forests share many
species with mixed mesophytic forests including sugar maple,
American beech, and white basswood. Associated species re-
stricted to the flood plain are box elder (Acer negundo), occasional
sweet buckeye (Aesculus glabra), black walnut, and silver maple
(Acer saccharinum).

Shrub species under the forest canopy are also quite similar to
those of the mixed mesophytic areas but are often quite dense.
Species ordinarily restricted to this area are elder (Sambucus
canadensis), ninebark (Physocarpus opulifolius), various willows
(Salix sp.), hoptree (Pletea trifolia), bladdemut (Staphylea tri-
folia), silky dogwood (Cornus amomum), and smooth alder (Al-
nus serrulata). Other tree-covered areas are often the result of

significant vegetative alteration. Many abandoned fields now have
extensive thickets of young black locust (Robinia pseudoacacia)
or hawthorn (Crataegus sp.). Old fence rows are now lines of
trees with black locust and black cherry as dominant species.

Strip mines are common throughout the Cross Creek wa-
tershed. Where reclamation has not been attempted, old spoil
banks often have patches of black locust. More recent mines often
have been regraded and planted with a variety of pines and spruces.

Abandoned fields and meadows are common in the Cross Creek
watershed. They exist in various stages of succession depending
on previous use and amount of time since abandonment. Wood-
land recovery occurs less rapidly in cultivated fields than in pas-
tures. Moisture is a prominent controlling factor in woodland
recovery; moister areas more rapidly achieve tree community
development.

Old fields follow a general rejuvenation pattern. A variety of
grasses and herbaceous plants first appear followed by thickets
of woody shrubs, briars and young shade-intolerant trees such
as red maple, black cherry, aspen (Populus sp.) and tulip tree. As
the primary trees mature, shade-tolerant plants such as oak, sugar
maple, beech, cucumber, hickories and gum begin growth. This
final stage produces permanent forest cover.

Hackberry (Celtis sp.) was commonly recovered as a macro-
fossil at Meadowcroft (Cushman, 1982). Present distribution in
the area apparently is less extensive than formerly. Celtis occi-
dentalis and Celtis tenui folia occur with two varieties (Celtis oc-
cidentalis var. pumila) and (Celtis tenuifolia var. georgiana). In
the mixed mesophytic forest, Celtis occidentalis is an infrequent
small tree. Only four hackberry specimens from Washington Co.,
Pennsylvania, are in the Carnegie Museum of Natural History
Herbarium. Two Celtis occidentalis specimens were collected along
Mingo Creek, near Riverview, Pennsylvania, in 1939 and 1942.
Two others were collected from Little Pine Creek in 1927 and
from Chartiers Creek west of Hendersonville in 1957. The most
common plants in the immediate vicinity of Meadowcroft Rock-
shelter are listed in Table 1 .

Contemporary Fauna

The area near Meadowcroft Rockshelter supports a wide range
of terrestrial and avian Carolinian fauna. This assemblage, how-
ever, is a pale reflection of that in the area’s deciduous oak/
hickory forests after the end of the Wisconsinan glaciation. As
late as the beginning of the 18th century, elk (Cervus elaphus),
black bear (Ursus americanus), mountain lion (Felis concolor),
wild cat (Lynx rufus), timber wolf (Canis lupus), fisher (Martes
pennanti ), otter (Lutra canadensis), beaver (Castor canadensis ),
wild turkey (Meleagris gallopavo), and passenger pigeon (Ecto-
pistes migratorius) could be found in the hills of western Penn-
sylvania.

Hunting and rapid replacement of forest by farms and pastures
soon altered the area’s faunal population. Today, thanks chiefly
to game conservation laws and secondary forest growth, the fau-
nal assemblage has regained some of its former abundant char-
acter (Guilday, 1977). The white-tailed deer (Odocoileus virgin-
ianus), for instance, is once again found near the rockshelter but
is the only large game species of the immediate vicinity. Presently,
white-tailed deer, cottontail rabbit (Sylvilagus Jloridanus), gray
and fox squirrel ( Sciurus carolinensis and S. niger), ringneck
pheasant (Phasianus co/chicus), bobwhite quail ( Colinus virgin-
ianus), ruffed grouse (Bonasa umbellus), muskrat (Ondatra zi-
bethicus), and mink (Mustela vison) constitute the principal game

1984

ADOVASIO ET AL. — MEADOWCROFT ROCKSHELTER

351

Table 1 . — Checklist of contemporary arboreal flora in the im-
mediate vicinity of Meadowcroft Rockshelter*

Taxa

Common name

Occur-

rence**

Acer saccharum

sugar maple

c

Acer rubrum

red maple

c

Fagus grandifolia

American beech

c

Quercus rubra

red oak

c

Quercus alba

white oak

c

Fraxinus americana

white ash

c

Ulmus americana

white elm

c

Ulmus rubra

slippery elm

c

Tsuga canadensis

eastern hemlock

c

Ostrya virginiana

hop-hornbeam

c

Platanus occidentalis

sycamore

c

Tiha americana

basswood

c

Sassafras albidum

white sassafras

c

Prunus serotina

wild black cherry

c

Betula nigra

red birch

c

Juniperus virginiana

red cedar

R

Carya tomentosa

mockemut hickory

R

Asimina triloba

pawpaw

U

Hamamelis virginiana

witch-hazel

U

Cercis canadensis

redbud

U

* From Carlisle et al., 1982:12, table 1.

'* C = Common. R = Rare. U = Understory.

species of the area (Grant, 1973). The relative incidence of these
species is plotted in Table 2.

Unlike many larger mammals, smaller mammals often sur-
vived the end of the Wisconsinan glaciation and the subsequent
deforestation of the Historic period. During the 1973-1978 ex-
cavations, the field crews observed many representatives of this
group including: gray squirrel, fox squirrel, gray fox (Urocyon
cinereoargenteus), raccoon ( Procyon lot or), and a large variety of
rodents. Some mammals actually benefitted from extensive clear-
ing and are now more widespread than they were aboriginally.
These include: cottontail rabbit, woodchuck ( Mar mot a monax),
opossum ( Didelphis virginianus), striped skunk (Mephitis me-
phitis), and red fox ( Vulpes vulpes).

Although many animals found farther north or within the
mountains of west-central Pennsylvania do not occur near Mea-
dowcroft (for example, porcupine, Erethizon dorsatum, and
snowshoe hare, Lepus americanus), the area does support an
abundant and varied mammalian assemblage, which includes
these additional species— chipmunk (Tamias striatus), southern
flying squirrel (Glaucomys volans), meadow vole ( Microtus penn -
sylvanicus), woodland or pine vole (Microtus pinetorum), south-
ern bog lemming (Synaptomys cooperi), white-footed mouse
(Peromyscus leucopus), meadow jumping mouse (Zapus hudson-
ius), short-tailed shrew (Blarina brevicauda), least shrew (Cryp-
totis parva), smoky shrew (Sorex fumeus), hairy-tailed mole (Pa-
rascalops breweri), least weasel (Mustela nivalis), long-tailed weasel
(Mustela frenata) and several species of bats.

Domesticated animals such as the horse (Equus caballus), cow
( Bos taurus), sheep (Ovis aries), pig (Sits scrofa), Norway rat
(Rattus norvegicus), house mouse (Mus musculus), feral dog (Canis
familiaris) and common cat (Felis domesticus) are, not unex-
pectedly, also found in the area today.

Table 2.— Relative incidence of principal game species in the Cross
Creek watershed as of December 1973.*

Species

evaluated

Main stem
Cross
Creek**

South Fork
Cross
Creek**

North Fork
Cross
Creek**

Middle Fork
Cross
Creek**

Deer

L

L

M

L

Rabbit

M

M

M

M

Squirrel

L

L

M

L

Pheasant

S

S

S

S

Quail

L

L

L

L

Grouse

L

L

M

M

Muskrat

H

M

M

M

Raccoon

H

M

M

M

Mink

L

L

L

L

Fox

L

L

L

L

* After Grant,
** L = Low. M

1973.

= Moderate. H =

Heavy. S = Population dependent on stocking.

Table 3 . — Checklist of contemporary fish in the Cross Creek wa-
tershed*

Taxa

Common name**

Cyprinidae

Campostoma anomalum

Stoneroller

Clinostomus elongatus

Redside dace

Cyprinus carpio

Carp

Ericumba buccata

Silverjaw minnow

Nocomis micropogon

River chub

Notropis atherinoides

Emerald shiner

Notropis chrysocephalus

Striped shiner

Notropis stramineus

Sand shiner

Notropis rubellus

Rosyface shiner

Pimephales notatus

Bluntnose minnow

Pimephales promelas

Fathead minnow

Rhinichthys atratulus

Blacknose dace

Semotilus atromaculatus

Creek chub

Catostomidae

Catostomus commersoni

White sucker

Hypentelium nigricans

Northern hog sucker

Centrarchidae

Amplophites rupestris

Rock bass

Lepomis cyanellus

Green sunfish

Lepomis gibbosus

Pumpkinseed

Lepomis macrochirus

Bluegill

Micropterus dolomieui

Smallmouth bass

Micropterus salmoides

Largemouth bass

Percidae

Etheostoma blennoides

Greenside darter

Etheostoma caeruleum

Rainbow darter

Etheostoma flabellare

Fantail darter

Etheostoma nigrum

Johnny darter

Cottidae

Cottus bairdi

Mottled sculpin

* From Carlisle et al., 1982:15, table 3.

** Common names follow the list recommended by the American Fisheries Society
Special Publication 6 (1970).

352

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

The avian fauna reflects continuing adjustment to changes in
land utilization and vegetative cover. In general, forest-dwelling
species have diminished in contrast to open-country forms since
the onset of Euro-American settlement. As previously men-
tioned, the ruffed grouse and the introduced ringneck pheasant
are important upland game birds. Man also has introduced the
rock dove (Columba livia ), English sparrow (Passer domesticus),
and the starling (Sturnus vulgaris) to the inventory of birds which
in aboriginal times included great numbers of migratory water-
fowl, swans, ducks, and geese. The transient and resident avifauna
now includes at least 30 species.

Terrestrial and riverine reptiles include black snake (Coluber
constrictor), garter snake (Thamnophis sirtalis), snapping turtle

(Chelydra serpentina), and box turtle (Terrepene Carolina). Am-
phibians include various plethodontids, toads (Bufo americanus),
tree frogs (Rana clamitans), and bullfrogs (Rana catesbiana).

Although not a large stream. Cross Creek is a direct tributary
of the Ohio River, and many fish species were undoubtedly avail-
able aboriginally. The present riverine fauna is somewhat de-
pauperate though not nearly so depleted as generally assumed.
Polluted by raw sewage and mine effluvia, Cross Creek never-
theless supports a restricted fauna. Freshwater mussels are almost
wholly absent, but a minimum of 26 fish species occur in small
to moderate numbers (Table 3). Current field observations sup-
port the view of Cooper (1972) who noted general scarcity of
game fish in the Cross Creek watershed.

EXCAVATION PROCEDURES

The excavation procedures employed at Meadow-
croft Rockshelter are thoroughly detailed in other
publications on the site (e.g., Adovasio et al., 1975,
1977a. 1917b, 1978a; Adovasio et al., 19786, 1979-
1980a, 1979-19806; Adovasio, 1982). During the
466 working days of the 1973-1978 projects, ap-
proximately 60.5 m2 of surface area inside the drip-
line and ca. 46.6 m2 outside the dripline were ex-
cavated. Over 230 m3 of fill were removed. Nearly
all of the excavation was conducted with trowels or
smaller instruments, and excellent vertical and hor-
izontal controls over the artifactual and “ecofac-
tual” assemblage were able to be maintained.

As explained in many other Meadowcroft pub-
lications, 1 1 natural strata were distinguished during
excavation. The earliest stratum is termed Stratum
I; the most recent is Stratum XI. Two of the eleven
strata (VI and X) occur only inside the dripline; the
other strata are continuous across the site. In the
interest of space, the reader is referred to Stucken-
rath et al. (1982) for details on the 1 1 Meadowcroft
strata. A composite profile of the site stratigraphy
is presented in Fig. 2.

RADIOCARBON CHRONOLOGY

One hundred Meadowcroft samples were sub-
mitted for radiocarbon assay to the Radiation Bi-
ology Laboratory of the Smithsonian Institution. In
all but two cases, the charcoal was derived from
firepits, firefloors, or charcoal lenses within the de-
posits. The exceptions are portions of completely
carbonized simple plaited basketry fragments (Ado-
vasio et al., 1977a; Stile, 1982). To date, 70 of the
samples have been processed. Twenty-two samples
were too small to count. The results of the assays
are presented in absolute stratigraphic order in Ta-
ble 4, and the dates are plotted in Fig. 3.

The initial human occupation of the rockshelter
is positively ascribable to the 1 5th millennium B.C.;
the latest radiocarbon date on purely aboriginal ma-
terials is A.D. 1,265 ± 80 (685 B.P.). The deepest
microstrata within Stratum Ha have produced two
radiocarbon dates with limited cultural material in
excess of 17,000 B.C. (18,950 B.P.) suggesting an

even earlier initial occupation. Cross-dated lithics
and ceramics from Strata VIII-XI indicate contin-
ued occupation or utilization of the site into the
Historic period.

The radiocarbon sequence is remarkably consis-
tent with the observed stratigraphy. Stratum Ila is
the deepest and oldest culture-bearing depositional
unit. For analysis and discussion purposes, Stratum
Ila is subdivided into three subunits of unequal
thickness labeled upper, middle, and lower Stratum
Ila. Each of these subunits is bracketed by major
roof spalling episodes, and each is well-dated by
radiocarbon assay (Table 5). Upper Stratum Ila has
a terminal date of 6,060 ±110 B.C. (8,010 B.P.)
from the uppermost living or occupation floor with-
in this subunit and a date of 7, 165 ± 1 15 B.C. (9,1 15
B.P.) from a slightly deeper occupational surface
within the unit. At the base of upper Stratum Ila is
a substantial roof spalling episode that marks the

1984

ADOVASIO ET AL. - MEADOWCROFT ROCKSHELTER

353

£

o

-C

Q-

OX)

a

£

o

U

00

£

boundary between this subunit and middle Stratum
Ha. While the top of the roof spalling event that
separates upper from middle Stratum I la is undated,
an assay of 9,350 ± 700 B.C. (1 1,300 B.P.) is avail-
able from directly beneath the roof spalling event
at the top of middle Stratum Ila. Hence, for all
intents and purposes, upper Stratum Ila dates ca.
9,000-6,000 B.C. (ca. 10,950-7,950 B.P.) and is of
Holocene age.

Middle Stratum Ila, sealed from upper Stratum
Ila by the roof spalling episode described above, is
also initiated by a roof spalling episode. Directly
beneath the latter roof spall is a date of 10,850 ±
870 B.C. (12,800 B.P.). Middle Stratum Ila is there-
fore bracketed by dates ranging ca. 1 1 ,000-9,000
B.C. (12,950-10,950 B.P.); it is of terminal Pleis-
tocene age.

Lower Stratum Ila, which lies beneath the roof
spalling episode that separates this subunit from
middle Stratum Ila, has seven additional radiocar-
bon dates ranging from 17,650 ± 2,400 B.C. (19,600
B.P.) to 1 1,290 ± 1,010 B.C. (13,240 B.P.). The 18th
millennium B.C. date constitutes the deepest date
from the rockshelter that is associated with mate-
rials of indisputable human manufacture and also
marks the onset of human utilization of this locality.

The maximum excavated depth of Stratum Ila
varies between 70 cm and 90 cm in different por-
tions of the rockshelter. At the interface of lower
Stratum Ila and underlying Stratum I are several
lenses of charcoal that have produced radiocarbon
dates in the 20th and 29th millennia B.C. range.
These dates are not associated with any cultural
materials. Furthermore, they are separated from the
deepest occupational floors within Stratum Ila by a
considerable thickness of sterile deposits (Adovasio
et al„ 1980:588-589).

No questions have arisen about the post- 10,000
B.C. (1 1,950 B.P.) Meadowcroft dates or their cul-
tural associations. Some of the Stratum Ila dates,
though collected and analyzed under identical con-
ditions, have stimulated considerable discussion (for
example, Haynes 1977, 1980; Mead, 1980; Din-
cauze, 1981). The discussion has centered on the
suggestion that the radiocarbon dates from the old-
est levels at the site, specifically from middle and
lower Stratum Ila, are “too old” because of the in-
jection of “dead” carbon in the form of coal particles
or so-called organic solubles (Haynes, 1977, 1980).
However, the evidence for particulate (Adovasio et
al., 1978a) and nonparticulate (Adovasio et al.,
1980, 1981) contamination of the Stratum Ila sam-

354

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table A. — Radiocarbon chronology from Meadowcroft Rockshelter as of August 1983.“

Stratum

(field designation)

Provenience/description

Lab designation

Date

XI (F-3)

Charcoal from firepit, mid-

SI-3013

A.D. 1,775

± 50

die one-third of stratum

X (F-25)

Charcoal from firepits

Samples too small

—

to process

IX (F-9)

Charcoal from firepit, up-

SI-2363

A.D. 1,265

± 80

per one-third of stratum

VIII (F-12)

Charcoal from firepit

SI-3023

A.D. 1,320

± 100

VII (F- 13)

Charcoal from firepits.

SI-2047

A.D. 1,025

± 65

middle one-third of stra-

SI-3026

A.D. 660

± 60

turn

VI (F-63)

Charcoal from firepits and

Samples too small

—

lenses

to process

V (F-14)

Charcoal from firepits, up-

SI-3024

A.D. 285

± 65

per one-third of stratum

SI-3027

A.D. 160

± 60

SI-3022

A.D. 70

± 65

SI-2362

125

± 125 B.C.

SI-2487

205

± 65 B.C.

IV (F-16)

Charcoal from firepits, up-

SI-2051

340

± 90 B.C.

per one-third of stratum

SI- 1674

375

± 75 B.C.

SI-2359

535

± 350 B.C.

SI-3031

705

± 120 B.C.

Charcoal from firefloor,

SI-1665

865

± 80 B.C.

middle one-third of stra-

turn

Charcoal from firepit, mid-

SI-1668

870

± 75 B.C.

die one-third of stratum

Charcoal from firepits/fire-

SI- 1660

910

± 80 B.C.

floors, lowest one-third
of stratum

SI-2049

1,100

± 85 B.C.

III (F- 18)

Charcoal from firepits, up-

SI-2066

980

± 75 B.C.

per one-third of stratum

SI- 1664

1,115

± 80 B.C.

SI-2053

1,140

±115 B.C.

SI-3030

1,150

± 90 B.C.

SI-2046

1,165

± 70 B.C.

Charcoal from firepit, mid-

SI-1679

1,305

±115 B.C.

die one-third of stratum

Charcoal from firepits/fire-

Samples too small

—

floors, lowest one-third
of stratum

to process

lib (F-46 up-

Charcoal from firepit, up-

SI- 1681

1,260

± 95 B.C.

per)

per one-third of stratumb

Carbonized basketry frag-

SI-1680

1,820

± 90 B.C.

ment, upper one-third of

stratum

Charcoal from firepits.

SI-2063

2,000

± 240 B.C.

middle one-third of stra-

SI-2058

2,020

± 85 B.C.

turn

SI-2054

2,055

± 85 B.C.

SI-2356

2,430

± 500 B.C.

SI-1685

2,870

± 85 B.C.

SI-2358

4,340

± 355 B.C.

Cultural period

Historic

Late Prehistoric

early Late Woodland

Middle Woodland

Early Woodland

Late Archaic

1984

ADOVASIO ET AL. — MEADOWCROFT ROCKSHELTER

355

Table A. — Continued.

Stratum

(field designation)

Provenience/description

Lab designation

Date

Cultural period

Charcoal from firefloor.

SI-2055

4,720

± 140 B.C.

lowest one-third of stra-

turn

Charcoal from firepit, low-

SI-2056

3,350

± 130 B.C.

Middle Archaic

est one-third of stratum

Charcoal from firepits/fire-

Samples too small

—

floors, lowest one-third
of stratum

to process

Ila (F-46 low-

Charcoal from firepits, up-

SI-2064

6,060

± 1 10 B.C.

Early Archaic

er)

per one-third of stratumc

SI-2061

7,125

±115 B.C.d

Charcoal from firepit/fire-

SI-2491

9,350

± 700 B.C.

floor, middle one-third
of stratum

Charcoal from firepits,

SI-2489

10,850

± 870 B.C.

Paleo-Indian

lowest one-third of stra-

SI-2065

11,290

± 1,010 B.C.e

turn

SI-2488

11,320

± 340 B.C.

SI-1872

12,975

± 620 B.C/

SI-1686

13,170

± 165 B.C.

SI-2354

14,225

± 975 B.C.

Charcoal concentration.

SI-2062

17,150

± 810 B.C.8

Paleo-Indian?

lowest level within stra-

turn

Carbonized fragment of cut

SI-2060

17,650

± 2,400 B.C.

bark-like material, possi-
ble basketry fragment,
lowest level within stra-

turn

Charcoal concentration.

DIC-2187

19,120

± 475 B.C.

base of stratum and di-
rectly above the Stratum
I/IIa interface

I/IIa interface

Charcoal from lenses at the

SI-2121

19,430

± 800 B.C.h

No cultural

interface

SI-1687

28,760

± 1,140 B.C.h

associations

I (F-85)
(Omega Unit)

Birmingham Shale

—

—

* All dates are uncorrected and are listed in absolute stratigraphic order.
b Provenience incorrectly listed as “upper x/i" in Adovasio et al., 1975:16, table 3.
c Provenience incorrectly listed as “middle V3” in Adovasio et al., 1975:16, table 3.

d This date has previously been listed incorrectly as 7,165 ± 115 B.C. in various publications (e.g., Adovasio et al., 1977a:33, table 7, Stuckenrath et al., 1982:80, table 2; 83,
table 3, among others).

e Date listed as 1 1,300 ± 1,000 B.C. in Adovasio et al., 1975:16, table 3 after rounding.
r Date incorrectly listed as 12,900 ± 200 B.C. in Adovasio et al., 1975:16, table 3.

8 Date incorrectly listed as 17,150 ± 801 B.C. in Stuckenrath et al., 1982:83, table 3.

h Dates from the Strata I/IIa interface were originally and tentatively ascribed to the “upper (?) V3” of Stratum I in Adovasio et al., 1975:16, table 3.

pies is unconvincing. Of importance for understand-
ing this point, the last remaining charcoal sample
from lowest Stratum I la was submitted to Dicarb
Radioisotope Company. The sample came from a
locus immediately below dates SI-2062 at 17,150
±810 B.C. (19,100 B.P.) and SI-2060, the cut bark
basketry at 17,650 ± 2,400 B.C. (19,600 B.P.) but

above sample SI-2 121, the charcoal with no cultural
associations at 19,430 ± 800 B.C. (21,380 B.P.).
This charcoal sample was submitted to Dicarb with
virtually no “background” information other than
the fact that it was from somewhere in Cross Creek
and “was probably older than 9,000 years.” The
sample was pretreated with sodium hydroxide and

356

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 3. — Plot of the Meadowcroft Rockshelter radiocarbon dates showing one standard deviation (from Stuckenrath et al., 1982:81,
fig- 2).

hydrochloric acid, that is, not using the nitration
pretreatment employed by the Smithsonian labo-
ratory. Liquid scintillation counting of benzene
rather than methane gas assay (which is standard
practice at the Smithsonian) was used. Despite the
different approaches, DIC-2187 provided a date in

stratigraphic order of 19,120 ± 475 B.C. (21,070
B.P.). This appears in Fig. 3 as the hollow rectangle
in the lower left of the plot. There was no evidence
for any contamination of the sample (Irene Stehli,
1982, personal communication).

1984

ADOVASIO ET AL. — MEADOWCROFT ROCKSHELTER

357

Table 5. — Radiocarbon chronology from Strata I and Ila at Meadowcroft Rockshelter as of August 1983.“

Stratum

(field designation)

Provenience/description

Laboratory

designation

Date

Cultural period

Ila (F46)

Charcoal from firepits, upper

SI-2064

6,060

± 1 10 B.C.

Early Archaic

one-third of stratum6

SI-2061

7,125

±115 B.C.C

Charcoal from firepit/firefloor

SI-2491

9,350

± 700 B.C.

middle one-third of stratum

Charcoal from firepits, lowest

SI-2489

10,850

± 870 B.C.

Paleo-Indian

one-third of stratum

SI-2065

11,290

± 1,010 B.C.d

SI-2488

11,320

± 340 B.C.

SI-1872

12,975

± 620 B.C.e

SI- 1686

13,170

± 165 B.C.

SI-2354

14,225

± 975 B.C.

Charcoal concentration, lowest

SI-2062

17,150

±810 B.C/

Paleo-Indian?

level within stratum

Carbonized fragment of cut

SI-2060

17,650

± 2,400 B.C.

bark-like material, possible
basketry fragment, lowest
level within stratum

Charcoal concentration, base of

DIC-2187

19,120 ± 475 B.C.

stratum and directly above
the Stratum I/IIa interface

I/IIa Interface

Charcoal from lenses at the

SI-2121

19,430

± 800 B.C.8

No cultural

interface

SI-1687

28,760

± 1,140 B.C.8

associations

I (F85) (Omega Unit)

No samples

No dates

No cultural

associations

0 All dates are uncorrected and are listed in absolute stratigraphic order.

b Provenience incorrectly listed as “middle 'A” in Adovasio et al., 1975:16, table 3.

1 This date has previously been listed incorrectly as 7,165 ± 115 B.C. in various publications (for example, Adovasio et al., 1977a:33, table 7; Stuckenrath et al., 1982:80, table
2, 83, table 3, among others).

d Date listed as 1 1,300 ± 1,000 B.C. in Adovasio et al., 1975:16, table 3 after rounding.

' Date incorrectly listed as 12,900 ± 200 B.C. in Adovasio et al., 1975:16, table 3.
f Date incorrectly listed as 17,150 ± 801 B.C. in Stuckenrath et al., 1982:83, table 3.

8 Dates from the Strata I/IIa interface were originally and tentatively ascribed to the “upper(?) ‘A” of Stratum I in Adovasio et al., 1975:16, table 3.

THE DATA BASE

Several of the many discrete data sets from Mead-
owcroft Rockshelter are important both for paleoen-
vironmental reconstruction at and near the site and
for the Mid-Atlantic region as a whole. They are
also important for elucidating the nature of the
Pleistocene/Holocene transition both in the general
area and at the rockshelter. These data sets include
vertebrate and invertebrate faunal as well as floral
remains buttressed by detailed geological, geochem-
ical, and sedimentological information. Extensive
discussion of all the Meadowcroft data to date is
provided in the appropriate chapters of Carlisle and
Adovasio (1982).

The diverse Meadowcroft data sets range from
poor to very good in the scope of information that
they offer for understanding the paleoenvironment
and the process of paleoenvironmental change at
the rockshelter throughout the Holocene. Several
data sets, however, are substantially incomplete for
the critical Pleistocene/Holocene interface. Unfor-
tunately, these include two of the potentially most

revealing and diagnostic components, the faunal and
floral remains.

Faunal Remains

Vertebrate faunal remains constitute the single
most commonly encountered set of inclusions in the
rockshelter. Over 1 15,166 bones or fragments were
individually examined of the total of nearly one
million recovered bones. Remains of at least 5,634
individual vertebrates, representing 1 5 1 taxa were
identified: 149 to the species level — 66 birds, 44
mammals, 26 reptiles, 8 fish, and 5 amphibian
species are present (see, Adovasio et al., 1979-19806:
108-109). Over 90% of the remains are from dis-
integrated digestive pellets regurgitated by raptorial
birds, primarily owls, that formerly roosted on the
cliff face of the rockshelter during the time that the
deposits were building. Southern flying squirrel, the
extinct passenger pigeon, and toad ( Glaucomys vo-
lans, Ectopistes migratorius, Bufo sp.) account for
68% of all the identified vertebrates. Some idea of

358

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

raptor activity at the site can be gathered from the
fact that 44% of all the individual vertebrates from
the site constitute but a single species, the mouse-
sized southern flying squirrel — at least 2,503 indi-
viduals. Approximately 7% of the bone collection
is probably attributable to Indian activity.

Bone preservation at all levels within the present
dripline of the rockshelter is good, but remains of
large and medium -sized vertebrates have been re-
duced to bone fragments or isolated teeth. In all,
23% of all fragments are charred, a figure that in-
creases to 93% outside the dripline. No charring
patterns were noted. All species were involved in
what appears to have been the result of a random,
accidental burning of fragments in the aboriginal
hearth substrata; they were not necessarily burned
as a result of aboriginal food preparation techniques.

The faunal reconstruction of the bone-bearing se-
quence at Meadowcroft advanced by Guilday and
Parmalee (1982: 1 70) is that of a temperate Carolin-
ian biota typical of southwestern Pennsylvania until
the 1 9th century.

In addition to the overall composition of the
Meadowcroft fauna there is intraspecific evidence
for temperate conditions as well. Late Pleistocene
specimens of southern flying squirrel from New Paris
Sinkhole No. 4, Pennsylvania, are larger on average
than Recent Pennsylvania material, whereas south-
ern bog lemming specimens are smaller (Guilday et
al., 1964). They were deposited under boreal con-
ditions, and the size characteristics are believed to
have been caused by physiological adaptation. Mea-
surements of southern flying squirrel humeri and
southern bog lemming first lower molars were com-
pared at different strata in the Meadowcroft deposit
down to and including Stratum Ha to see if there
was any indication of time-related size changes that
might suggest climatic change. In both cases the
results were negative, and the sample parameters
agree with those of Recent material at all levels in-
side the dripline, suggesting temperate conditions
throughout.

Of the 21 species of non-aquatic vertebrates that
reached their lowest stratigraphic level inside the
dripline (dated at ca. 9,350 B.C.; 11,300 B.P.), the
presence of box turtle, timber rattlesnake, bobwhite,
turkey, eastern mole, and pine or woodland vole
implies that temperate Carolinian conditions pre-
vailed in this part of the Upper Ohio Valley as early
as 1 1,300 years ago.

The interpretation of the meager faunal remains
in older levels from outside the dripline is not as

clear. With the exception of the charred base of a
white-tailed deer antler, all material is highly frag-
mented. Only four vertebrate species could be iden-
tified from associations dated at 9,350 B.C. (1 1,300
B.P.) or older in strata outside the dripline — white-
tailed deer, eastern chipmunk, southern flying squir-
rel, and passenger pigeon. Charred fragments of toad,
colubrid snake and deer mouse ( Peromyscus sp.) are
also present. Deer remains were not identified below
the 14,225 ± 975 B.C. (16,175 B.P.) level in lower
Stratum Ila. Passenger pigeon was found at 1 1,290 ±
1,010 B.C. to 13,170 ± 165 B.C. (13,240 to 15,120
B.P.) levels in lower Stratum Ila. One charred post-
cranial element of southern flying squirrel does oc-
cur at 17,150 ± 810 B.C. (19,100 B.P.) or older.
The chipmunk remains may be suspect, as the bones
are uncharred and the animal is an avid burrower,
but white-tailed deer, southern flying squirrel, and
passenger pigeon suggest, although they do not nec-
essarily demand , a temperate setting. These species
do, or did, range north rarely and marginally into
Canadian zone situations, so that their ecological
interpretation in these levels would be equivocal in
the absence of the accompanying botanical record.

Unfortunately, the invertebrate faunal record from
Meadowcroft Rockshelter is also more-or-less mute
for the older levels of the site (pre-9,350 B.C.) as
well. Although not common, mollusk remains are
represented throughout the long occupational se-
quence (Lord, 1982). A minimum of 35 terrestrial
and three aquatic gastropod species as well as 1 1
naiad (freshwater mussel) species have been iden-
tified. All recovered species are present in the area
today, although several mussel species have been
exterminated from the segment of Cross Creek ad-
jacent to the rockshelter.

The gastropod fauna indicates that the local en-
vironment has remained essentially stable for at least
the past 11,000 years before which time mollusks
are poorly represented in the site’s deposits. The
vegetative and climatic regime reflected by the ex-
tant molluscan assemblage is characterized by a sta-
ble woodland, predominantly of oak and hickory —
quite moist and shady but not extremely dense.

Floral Remains

Floral remains are the second most abundant class
of material recovered from Meadowcroft Rock-
shelter. Included are moderately large sections of
tree trunks and limbs, with and without bark, mi-
nute seeds and seed coats, fruits, charcoal, and small
amounts of pollen. Vegetal remains are represented

1984

ADOVASIO ET AL. — MEADOWCROFT ROCKSHELTER

359

in all occupation levels, including Stratum Ha, and
therefore span approximately 16,000 years of the
site’s history (Volman, 1981:4).

Paleoenvironmental reconstructions incorporat-
ing botanical data are incomplete for much of the
Northeast. This is due to inadequate sampling of
archaeological and natural deposits and/or to poor
site deposition conditions (for example, abrasive soil
texture, excessive water or extreme pH values). Con-
sequently, few sites in the Northeast have produced
quantities of well-preserved botanical remains.

The comments on Meadowcroft faunal preser-
vation are also applicable in large measure to the
floral assemblage. Over 97% of it is from upper
Stratum Ila and above and is therefore of Holocene
or modern aspect. Portions of this larger assemblage
have been discussed elsewhere (Adovasio et al.,
1911 a, 1 979-1 980&; Adovasio and Johnson, 1981).
A modest amount of floral material, usually charred,
comes from middle and lower Stratum 1 la and in-
cludes deciduous forest elements; this is in keeping
with the Holocene character of the fauna.

The availability of both macrofloral elements and
pollen from Meadowcroft is ideal and provides in-
formation on local and regional vegetation. Most
macrobotanical remains are not susceptible to long-
distance transport. Their presence in archaeological
sites frequently indicates that the plant source was
at or near the place of deposition. Taxa not repre-
sented in the cultural assemblage of a site were not
necessarily absent from the local flora, however. In
contrast, airborne pollen can originate some dis-
tance from the place of deposition, and most (but
not all) pollen is a regional vegetation marker (Vol-
man, 1981:4).

Approximately 10% of all macrobotanical ma-
terials from Meadowcroft were analyzed. This sam-
ple consists of more than 30,000 separate plant
pieces. The remaining 90% of the material was ex-
amined to insure that the studied remains included
adequate samples of all elements and taxa preserved
at the site. Twelve complete sample columns also
were taken. The sediment in these columns was pro-
cessed in a hydrogen peroxide solution to separate
the heavy (larger seeds and nutshell) from the light
(smaller seeds and charcoal) fractions. All floral re-
mains from these columns have been analyzed (Vol-
man, 1981:13).

Among specimens of wood and charcoal, the
chronological range and taxa diversity suggest that
the vegetation at and near Meadowcroft was mixed
conifer-hardwood from the time of initial human

occupation onward. The two genera from the oldest
plant-containing sediments, ca. 14,500 years ago,
are Quercus sp. and Carya sp. Both genera remain
in evidence to the present day (Volman, 1981:90).
Two other genera abundant in the fossil record ( Cel -
tis sp. and Primus sp.) are no longer extant with any
frequency and decrease in numbers from the earliest
through more recent strata. Celtis sp. could have
been a relatively common species in the prehistoric
forests around Meadowcroft. Present data suggest
that this plant source may have provided food to
both the human and animal populations of the area.

Pinus sp. appears in the prehistoric macrobotan-
ical record about 11,000 years ago and remains to
the present. It is particularly abundant in the period
from about 3,000 years ago to the Historic period
when, apparently, it was decimated by logging. Pi-
nus sp. is a pioneer species intolerant of shade. Its
presence in relative abundance may indicate that
the forest surrounding Meadowcroft from ca. 1 0,000
years ago to the present was in the process of re-
generation following some disturbance such as fire,
storm damage, or land clearance (Volman, 1981:
91). Juglans sp., although a very shade tolerant
species (Harlow and Harrar, 1969), persists from
about 9,000 years ago to recent times in compara-
tively well-represented amounts. It is likely that Ju-
glans sp. retained a dominant or co-dominant po-
sition in the forest canopy during its occupance in
the Meadowcroft area.

The only genera represented through time with
greater frequency than Juglans sp. are Quercus sp.
and Tsuga sp., both of which are tolerant of a variety
of light and moisture conditions. Quercus sp. is found
at every time period at Meadowcroft and probably
was (as it is presently) a major component of the
mixed conifer-hardwood forest in the vicinity of the
site. Tsuga sp. suggests the onset of possibly cooler,
moister conditions, but its tolerance of marginal
conditions renders this interpretation conjectural.
Co-representation with Pinus sp. in comparatively
large amounts at ca. A.D. 1025 and A. D. 1775 leads
one to believe that the forest canopy may then have
been partly or mainly Pinus sp. with Tsuga sp. a
potential understory component.

Carya sp., Ulmus sp., Fraxinus sp., and Fagus
sp. parallel each other in comparative amounts and
frequency of representation through time. These
genera are mesophytic and tolerant, although Fagus
sp. is not usually found on dry soils (Harlow and
Harrar, 1969). The greatest amount of Fagus sp.
recovered from the fossil evidence corresponds with

360

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

the greatest amount of Tsuga sp., which also prefers
moist habitats. The periods ca. A.D. 1025 and A.D.
1775 at Meadowcroft may thus have been slightly
wetter than previous times, based on the presence
and apparent relative abundance of these two genera
(Volman, 1981:93).

Nutshells of Quercus sp. and Carya sp. date to at
least 9,000 years ago at Meadowcroft, and Juglans
sp. nutshells are known from 6,000 year old levels.
This supports the interpretation of the wood and
charcoal data that mixed conifer-hardwood forests
surrounded Meadowcroft at least since this time.
Walnut trees probably were abundant for a long time
in the prehistoric forest near the site. This is also
reflected to some extent in the wood and charcoal
remains as Juglans sp. occurs in small quantities at
various times but is present in increased amounts
in Strata VII1-XI.

Juglans sp. nutshell occurs in every time period
represented at Meadowcroft, but its wood and char-
coal are absent from the earliest strata. The greatest
amounts of Juglans sp. nutshell parallel the largest
number of Juglans sp. wood and charcoal specimens
in most of the recent strata. Carya sp. nutshell oc-
curs in greater relative frequency in the fossil record
than does Carya sp. wood or charcoal.

Quercus sp. wood and charcoal occurs from the
earliest levels at Meadowcroft, although acorn shell
does not. The lack of acorns from the lowest levels
probably can be attributed to a failure on the part
of the aboriginal inhabitants of the site to collect
them. Most of the recovered acorns are not carbon-

ized, and they increase in quantity from Stratum V
upward.

Fruits and seeds other than nutshell from Mead-
owcroft increase through time in numbers and di-
versity; however, the numbers recovered per stra-
tum are low. The relative increase in weedy annuals
since about 1,000 B.C., especially Amaranthus sp.,
suggests some increase in land clearance and/or dis-
turbance. Pollen data from the same period corrob-
orate this observation.

Fruit and seed remains from earlier time periods
are, unfortunately, very scant. However, Nyssa sp.
stones from at least 13,000 years ago argue that the
environment in the vicinity of Meadowcroft then
was at least as warm as that today, if not slightly
warmer. Meadowcroft lies at the extreme margin of
this plant’s northernmost modern range (Volman,
1981:95; Fowells, 1965).

Although by no means common in the lowest
occupational levels at the site, the extant Meadow-
croft floral remains do extend to levels older than
9,350 B.C., the horizon below which faunal data are
meager. Vegetation in the proximity of the site ap-
pears to have been mixed conifer-hardwood
throughout the ca. 16,000 years represented by the
floral remains. The presence in earliest time periods
of Quercus sp., Juglans sp. and Carya sp. argues
against boreal or tundra conditions near the site
during late glacial times. Significantly, all species
present in the fossil record are also present in the
extant vegetation of the area.

GEOLOGICAL EVIDENCE

The faunal and floral data from Meadowcroft are
limited in those strata that span the Pleistocene/
Holocene boundary. Geological data for this critical
time period, however, are abundant, continuous and
not subject to interpretational distortions arising
from sample size alone (Beynon, 1981; Beynon and
Donahue, 1982). Much of this geological informa-
tion has been presented in earlier publications (for
example, Beynon and Donahue, 1982) but seldom
discussed in the context of elucidating the Pleisto-
cene/Holocene boundary in the study area. It is
therefore useful to review here salient facts about
the geology of sandstone rockshelters in general and
of Meadowcroft Rockshelter in particular.

Detailed geoarchaeological examination of a
number of sandstone rockshelters in eastern North
America including Meadowcroft (Adovasio et al.,
n.d. a\ Carlisle and Adovasio, 1982; Vento et al.,
1980; Adovasio, 1982 (compiler); Yedlowski et al.,
1982) has demonstrated the presence of a more-
or-less “standardized” geological sequence in the
formation of such sites and in the geological mech-
anisms for the accumulation of sediments. The
rockshelters are developed within medium-bedded
to massively bedded sandstones typically underlain
by shales. Rockshelter development proceeds by flu-
vial downcutting through the sandstone unit(s) and
then via undercutting of the less resistant shale
thereby forming a re-entrant. In the case of Mead-

1984

ADOVASIO ET AL. — MEADO WCROFT ROCKSHELTER

361

LONGITUDINAL PROFILE OF CROSS CREEK

W e

Fig. 4. — Longitudinal profile for Cross Creek from its confluence with the Ohio River on the west to its headwaters on the east.
Uppermost line shows the elevation of uplands in the region, indicating that Cross Creek had ca. 51.8 m of relief in Nebraskan time.
Five discontinuous terraces, Nebraskan, Kansan, Illinoian, early Middle Wisconsinan, and Late Wisconsinan in age, are indicated (after
Beynon, 1981:210, fig. 43).

owcroft Rockshelter, the presence of a series of
Pleistocene terraces along the Ohio River and Cross
Creek can be used to establish the timing and mech-
anism of the initial development of the rockshelter
(Benyon, 1981; Benyon and Donahue, 1982; Ray,
1974).

Progressive downcutting of Cross Creek can be
documented by study of the terraces that developed
along the Cross Creek drainage and by correlating
them with terraces on the Ohio River (Fig. 4). As
indicated in Fig. 4, Cross Creek in pre-Pleistocene
time was a tributary of the north-flowing Monon-
gahela-Beaver drainage system. The valley was wide,
gently sloping and had not yet eroded into the Mor-
gantown-Connellsville Sandstone in which the rock-
shelter ultimately was developed. Initial erosion of
the sandstone occurred during the Yarmouth Inter-
glacial. However, it was not until the Sangamon
Interglacial that shale undercutting beneath the
sandstone and the initial development of the rock-
shelter re-entrant occurred (Fig. 5).

Fluvial sedimentation during early Middle Wis-
consinan time was responsible for deposition of a
silty clay at the Stratum I/IIa interface. When Cross
Creek resumed active downcutting in Middle Wis-
consinan time, the rockshelter was enlarged but was
not again affected by fluvial erosion and deposition.
The Late Wisconsinan terrace of Cross Creek is ca.
3 m to 6 m below the top of Stratum I at Meadow-
croft. Thus, at any point after middle Late Wiscon-
sinan time (ca. 21,300 B.P.) the site was available
for potential occupation and open to colluvial sedi-
mentation.

All of the Meadowcroft sediments that post-date
the deposition of the Stratum I/IIa interface silty
clay are essentially of colluvial origin. Detailed field
and laboratory examination of textural properties,
mineralogy, and sedimentary structures of this col-
luvial pile indicate that it is derived via gradual but
continuous erosion from a relatively limited num-
ber of sources. These include grain-by-grain attri-
tion from the rockshelter ceiling and walls, rain-

362

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

NEBRASKAN

Z-and Sh 7 -Z-Z

ILL INOIAN

KANSAN

EARL Y/M !D~ W I SCONS! NAN

LATE WISCONSINAN PRESENT

Fig. 5. — Sequential cross sections of Cross Creek at Meadowcroft Rockshelter. Progressive deposition and downcutting through Pleis-
tocene terraces are illustrated (after Beynon, 1981:205, fig. 42; 213, fig. 44; 216, fig. 45; 218, fig. 47; 221, fig. 49; 225, fig. 51).

induced sheetwash from upland surfaces above the
rockshelter, and rock fall (Beynon and Donahue,
1982:42-48). None of the other principal sediment
source mechanisms for sandstone rockshelters in
eastern North America (that is, stream action, aeo-
lian deposition, or cultural activity) appear to have
been of any importance for the accumulation of the
relatively thick, generally poorly sorted Meadow-
croft sediment pile.

Culturally derived sediments and modified sedi-
ments in the form of archaeological features are, of
course, present in the Meadowcroft sequence from
near basal Stratum Ila and proceeding upward
through Stratum XI; however, this source consti-
tutes a minor contribution to the total sediment pile.

When Stratum Ila cultural deposition began. Cross
Creek was well below the basal occupational floor
of the rockshelter. This fact, and information from
intensive examination of the Stratum Ila sediments,
indicate that flooding, stream deposition, and re-
working of cultural deposits did not occur.

Neither dune nor loess sequences exist in the
Meadowcroft sedimentological record. Dune fields
are obviated by the humidity of the area, and loess
deposits are not abundant anywhere in this part of
Pennsylvania. Paleosols have not been discerned at
Meadowcroft, and given the continuous nature of
sedimentation at the site, it is not possible that they
could have formed. In sum, once the Meadowcroft
colluvial pile began to accumulate above the Stra-

1984

ADOVASIO ET AL. — MEADOWCROFT ROCKSHELTER

363

turn I/IIa interface, sedimentation quite unrelated
to Cross Creek was continuous, proceeding to the
present without interruption or erosional episodes
(Beynon and Donahue, 1982:43).

Differences in strata composition at Meadowcroft
are not due to climatic fluctuations such as those
posited for the Pleistocene/Holocene boundary;
rather, these differences arise from changes in the
overall configuration of the rockshelter itself. The
physical appearance of the rockshelter therefore ef-
fectively molds and conditions the character and
composition of the sediment pile at any point in
time as it has throughout the geological history of
the site above the Stratum I/IIa interface.

Attempts to understand the interplay between
natural and cultural factors in the development of
Meadowcroft as a physical entity require recognition
of the fact that once emplaced, colluvial sediments
within the rockshelter were never reworked by water
action. The readily identified presence of a distinct
paleo-dripline within the colluvial sediments from
lowest Stratum I la upward through the sediment
pile to the present floor of the site traces the gradual
northward migration of the cliff (that is, the pro-
gressive retreat of the overhang) through time. It
also, and, perhaps more graphically than any other
indicator, testifies against any major reworking of
the sediment pile, a process that would have eroded
and erased the dripline. The position of the dripline
is documented by fluctuations in calcium carbonate
levels as well as by scanning electron microscopy of
quartz grains from both inside and outside the rock-
shelter overhang. Discussions of this aspect of the
Meadowcroft investigations are available in Ado-
vasio et al. (1977a), Beynon (1981), Beynon and
Donahue (1982), and Adovasio et al. (1983).

Detailed examination of continuous sediment
column samples from various loci at Meadowcroft
(for example, Adovasio et al., 1977a: 14, fig. 7) as
well as long-term scrutiny of modern attrition and
sheetwash processes and products have permitted
the systematic discrimination of sediments by size
category and source within each stratum. The results
of these examinations permit very detailed recon-
struction of the entire depositional history of the
site (Adovasio et al., 1977a; Beynon and Donahue,
1982; Adovasio et al., 1983). A summary of that
part of the overall sequence bracketing the Pleis-
tocene/Holocene transition is germane at this junc-
ture for the additional interpretative potency it lends
to the somewhat more restricted floral and faunal
evidence.

The late Pleistocene and early Holocene at Mead-
owcroft are encompassed within Stratum Ila. As
noted above, Stratum Ila is a relatively thick unit
which varies from 70 cm to 90 cm in maximum
excavated depth. It is divided into three subunits of
unequal thickness each of which is bracketed by
well-dated roof falls. Lower Stratum Ila dates from
ca. 19,430 ± 800 B.C. (21,380 B.P.) to 10,850 ±
870 B.C. (12,800 B.P.) and is late Pleistocene in age.
Middle Stratum Ila, it has been observed, dates from
ca. 1 1,000-9,000 B.C. (12,950-10,950 B.P.) and is
terminal Pleistocene in age. Upper Stratum Ila dates
to the early Holocene, ca. 9,000-6,000 B.C. (ca.
10,950-7,950 B.P.).

During this long period, all three sediment sources
contributed to the sediment pile at Meadowcroft,
though in very unequal measure. In the entire lower
Stratum Ila sequence, sedimentation was princi-
pally grain-by-grain attrition; rock fall episodes oc-
curred only when the rockshelter roof became suf-
ficiently unstable to spall off larger blocks. Modern
attrition samples collected and analyzed for a con-
tinuous five year period suggest that cold and in-
creased moisture may slightly increase the release
of individual sand grains, but they certainly do not
drastically accelerate the sedimentation rate from
this source. During the deposition of lower Stratum
Ila, there were no entrants in the overhang, and
sheetwash could not contribute to the sediment pile.

Middle Stratum Ila also accumulated principally
via grain-by-grain attrition punctuated by limited
spalling, and this period reflects the first indications
of the accumulation of limited amounts of sheet-
wash on the western margin of the site. The sheet-
wash was admitted through an entrant developed
by a partial roof collapse termed the Old Roof Fall.
Only a portion of the roof collapsed, but a large
enough access route was created to elevate the sedi-
mentation rate in that part of the rockshelter near
the collapse.

Holocene-age upper Stratum Ila accumulated, as
did middle and lower Stratum Ila, through a com-
bination of attrition, limited rock spalling and the
addition of some sheetwash to the western edge of
the site. The sedimentation rates for upper and mid-
dle Stratum Ila are virtually identical, and both are
only slightly elevated on average over the rates for
lower Stratum Ila.

It is clear that during most of Stratum Ila time at
Meadowcroft, the principal sediment source was
grain-by-grain attrition. The sedimentation rate re-
mained essentially the same except in those parts

364

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 6. — Calculated rates of sedimentation for Strata lla-Xl at Meadowcroft Rockshelter*

Stratum

Duration
(in radiocarbon
years)

Average thickness

Rate of sedimentation

Dominant and
subordinate sources

VIII-XI

925 years

35 cm

38 cm/ 1,000 years

sheetwash

VII

365 years

40 cm

1 10 cm/ 1,000 years

sheetwash and rock fall

(average value)

VI

375 years

100 cm

267 cm/ 1,000 years

rock fall and sheetwash

V

625 years

30 cm

48 cm/ 1,000 years

sheetwash and attrition

IV

760 years

55 cm

72 cm/ 1,000 years

sheetwash and attrition

(average value)

III

200 years

50 cm

250 cm/ 1,000 years

attrition, sheetwash

and rock fall

lib

4,700 years

40 cm

8.5 cm/1 ,000 years

attrition and rock fall

Ila

13,000 years

90 cm

6.9 cm/ 1,000 years

attrition and rock fall

Total

Average sedi-

20,950 years

440 cm

21.0 cm/ 1,000 years

mentation value

* After Stuckenrath et al., 1982:89, table 4.

of the site affected by the creation of the western
entrant. Although variation in rock fall frequency
and location in Stratum Ha time may well have been
controlled to some extent by climatic factors, rock
fall was certainly never the most important con-
tributor to the sedimentary history of the site.

Sedimentation rate and source are guided by fac-
tors that key on the configuration of the overhang
in a virtually direct, one-to-one correlation. Where
the overhang remains intact, gradual sedimentation
takes place via grain-by-grain attrition and roof
spalling alone. When entrants develop in the roof,
sedimentation rates increase proportionately due to
the influx of sheetwash. Despite the entrant of mid-
dle Stratum Ila age, sedimentation rate at the site
did not change appreciably at the Pleistocene/Ho-
locene boundary, which is marked at Meadowcroft
by the middle/upper Stratum Ila interface, ca.
9,000 ± 500 B.C.

Haynes (1982:97) has suggested, . . that there
is a pronounced stratigraphic break in alluvial
successions throughout the conterminous United
States that marks . . . the end of late Pleistocene
fauna, ... a major change in stream regimen . . .
and a significant change in climate as reflected in
the plant record.” No comparable break is docu-
mented at Meadowcroft nor, we believe, in any oth-
er closed site of this time horizon (for example, Fort
Rock Cave, Oregon; Wilson Butte Cave, Idaho).

One other observation on the sedimentation rate(s)
reflected in Stratum Ila at Meadowcroft is worth a
comment here. Sedimentation rates for the various

Meadowcroft strata were calculated using the pres-
ent radiocarbon chronology and the average thick-
ness of each depositional unit (Table 6). As Table
6 shows, in those strata where grain-by-grain attri-
tion is the predominant or only sediment source,
sedimentation rate is, as Dincauze (1981 :4) has put
it, “exquisitely slow.” Conversely, in strata where
sheetwash from one or both major roof fall entrants
is operative, sedimentation is quite rapid.

To ascertain approximately how slow sedimen-
tation can be in areas affected by attrition alone,
samples from the 5 m square attrition trap on the
roof of the excavations from 1974 through 1978
were again scrutinized. The trap was divided into
five 1 m by 5 m strips, and the amount of sediment
falling on each of these strips was assessed. The
samples from each strip generally yielded total
weights ranging from 0.5 g to 10 g per day. Average
values measured approximately 2 g per day per strip.
The sediment collected in the trap consisted of quartz
grains with a mean diameter of about 0. 1 5 mm and
a density of 2.65 g/cm3. Using this information, an
approximate if rough calculation can be made for
modem sedimentation rate. With a fall of 2.65 g for
one day, a volume of 1 cm3 accrues. Over a 5 m2
area (that is, one of the 1 m by 5 m strips comprising
the attrition trap), this results in a sediment thick-
ness of 2.0 x 10~5 cm. This in turn equals a rate of
7.3 cm/ 1,000 years. As attrition sediments have po-
rosities of 15% to 25%, the actual sedimentation
rate would increase to 8.75 cm/ 1,000 years. Burial
of these same sediments causes compaction and re-

1984

ADOVASIO ET AL. — MEADOWCROFT ROCKSHELTER

365

suits in an apparent decrease in rate. The significant Ha, where grain-by-grain attrition was the dominant

point here is that the modern rate is perfectly com- sediment source,

patible with the measured rates in lower Stratum

OVERVIEW

The biological evidence from Meadowcroft Rock-
shelter presents a straightforward picture of a tem-
perate “modem” biota. There is no problem of in-
terpretation or correlation with other regional sites,
paleontological and archaeological, that date from
less than ca. 9,350 B.C. (1 1,300 B.P.). The Meadow-
croft fauna agrees in content with archaeological
faunas from 43 other regional sites ranging in cul-
tural ascription from Archaic to Late Prehistoric
(Guilday et al., 1980). All of these sites present a
typical “Carolinian” fauna that one would expect
to occur in the area today in the absence of European
colonization.

This temperate fauna can be traced at other sites
as far back as 7,290 ± 1,000 B.C. (9,240 B.P.) at
Hosterman’s Pit, Centre Co., Pennsylvania (Guil-
day, 1967). Other regional sites of some antiquity
that contain a Recent fauna include Sheep Rock
Shelter, Eluntingdon Co., Pennsylvania, 6,970 ± 320
B.C. (8,920 B.P.) (Michels and Smith, 1967) and
New Paris Sinkhole No. 3, Bedford Co., Pennsyl-
vania, 6,620 ± 145 B.C. (8,570 B.P.) (Guilday et
al., 1964),

Correlating the Meadowcroft data from older
levels at the site with evidence from other mid-
Appalachian sites is not as simple, however. Several
writers, notably Haynes (1980) and Mead (1980)
have suggested that the faunal and floral data from
middle and lower Stratum Ila are climatically/eco-
logically inappropriate for the time period indicated
by the radiocarbon dates. As amply noted in earlier
Meadowcroft publications, however, the entire fau-
nal sample from levels older than 9,350 ± 700 B.C.
(1 1,300 B.P.), that is, from middle Stratum Ila and
below, is limited to 1 1 identifiable specimens from
among a total of 278 fragments. Only white-tailed
deer, eastern chipmunk, southern flying squirrel, deer
mouse, passenger pigeon, colubrid snake, and toad
have been positively identified. As a group, these
taxa suggest temperate conditions provided that they
co-occurred near the rockshelter and were not trans-
ported some distance to the site by humans or rap-
tors. Because each species might have occurred mar-
ginally in a boreal context (judging from modern

ranges) their climatic implications would be ambig-
uous were it not for the botanical evidence which
suggests that black gum (Nyssa sp.) occurred as early
as 13,000 B.P., oak ( Quercus sp.) and hickory ( Car -
ya sp.) as early as 14,500 B.P., and walnut/butternut
(Juglans sp.) as early as 16,000 years ago (Volman,
1981:90, 93-94). Once again, these data suggest
temperate conditions at the rockshelter. The ver-
tebrate and invertebrate faunal evidence (including
specific and, in the case of southern bog lemming
and southern flying squirrel, infraspecific data) in
concert with the floral remains therefore present a
mutually supportive picture of the local environ-
ment which Meadowcroft’s human occupants ex-
perienced at the close of the Pleistocene. That pic-
ture in general, the data lead us to believe, was not
one substantively different from that of the present
day. Unfortunately, preservation of fauna and flora
at Meadowcroft prior to approximately 1 1 ,000 years
ago is not good, but the information at hand reflects
no radical ecological reorganization at the site itself.

In this regard it is important to note three fre-
quently overlooked facts about Meadowcroft, the
full ecological implications of which are as yet far
from clear:

1) The Cross Creek drainage lies relatively far
south (ca. 50 km) of the mapped Kent moraine,
the maximum southern extent of the Wisconsinan
ice. For much of the later Pleistocene human his-
tory of the site (that is, after ca. 12,830 B.C. or
so), the discontinuous and only relatively recently
mapped Lavery till, ca. 83 km north of Meadow-
croft, marks the southern edge of the ice.

2) Meadowcroft occurs in a topographical set-
ting which in one recent year had 40 to 50 more
frost-free days than did the higher elevations or
drainages contiguous to it. Indeed, modern Cross
Creek has a more “southerly” temperature regime
than any other drainage in the area. If this is now
the situation, it may well have been the case pre-
historically. In any event, as far as indications of
late Pleistocene/Holocene environmental change
at Meadowcroft Rockshelter are concerned, the

366

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

biotic record is silent. The fact that the Cross
Creek drainage trends generally east-west rather
than north-south may also help to explain its en-
during temperate ecology as may the fact that the
rockshelter itself faces south, away from the gla-
cial front. Rather than an “either/or” ecological
situation, however, it is not unreasonable to as-
sume that Meadowcroft represents a floral and
faunal mosaic perhaps not atypical or unique
among other sites away from the glacial front.

3) Consistent with the greater number of frost-
free days at Meadowcroft is its lower elevation
(259.9 m) when compared not only to some other
areas in the Cross Creek drainage but to well-
studied paleontological sites in Pennsylvania and
elsewhere. Hosterman’s Pit stands at 377.9 m
(Guilday, 1967:231). Not surprisingly, the mam-
malian faunal assemblage associated with its
7,290 ± 1 ,000 B.C. (9,240 B.P.) radiocarbon date
is of Recent aspect (Guilday et al., 1977:80-81,
table 23). As noted previously. New Paris Sink-
hole No. 3, Pennsylvania, also has Recent fauna
associated with its 6,620 ± 145 B.C. (8,570 B.P.)
date. Leaving aside considerations of locational
differences in comparison to Meadowcroft, these
sites do not illuminate the characters of a Pleis-
tocene/Holocene transition fauna in Pennsylva-
nia.

New Paris Sinkhole No. 4, only a few meters
from New Paris No. 3, however, has a 9,350 ±
1,000 B.C. (1 1,300 B.P.) date (at the 4.6 m level)
and a fauna dominated by “boreal small rodents
and insectivores” (Guilday et al., 1977:80). This
is during middle Stratum Ila times at Meadow-
croft (see Tables 4 and 5). It is important to note
that New Paris No. 4 stands at 465 m above sea
level (Guilday et al., 1977:79), that is, ca. 205 m
above Meadowcroft’s elevation. Furthermore,
New Paris No. 4 is very near the foot of the Al-
legheny Front (Johnson, 1981:89). Its predomi-
nantly (but not completely) boreal fauna is there-
fore not unanticipated. Clearly, the utility of the
data from this site for reconstructing the contem-
poraneous environment at Meadowcroft (on the
Unglaciated Allegheny Plateau and at much lower
elevation to name but two of the most obvious
differences) is circumscribed. With some small
deviations, the New Paris No. 4 fauna is “iden-
tical” (Guilday et al., 1977:77) to that at poorly
dated but probably late Wisconsinan/early Ho-
locene Clark’s Cave in Bath Co., Virginia, which
stands at 448 m, or 188 m above Meadowcroft’s

elevation (Guilday et al., 1977:6, fig. 2). Though
farther south than Meadowcroft, Clark’s Cave was
a short distance from the Allegheny Front.

Baker Bluff Cave in Sullivan Co., Tennessee, at
450 m elevation (Guilday et al., 1978:5) dates to
as early as 1 7, 1 50 ± 850 B.C. ( 19, 100 B.P.) when
deposition may have begun during the recovery
phase of the Connersville Interstadial (Guilday et
al., 1978:55). The lower fauna from this site sug-
gest more temperate deciduous/coniferous con-
ditions that thereafter disintegrated to boreal co-
niferous parkland (Guilday et al., 1978:77).

For matters of ecological reconstruction, it is
an important fact that New Paris No. 4, Clark’s
Cave, and Baker Bluff Cave all contain fauna with
mixed boreal, mid-continental, and eastern tem-
perate elements though a general north to south,
boreal to temperate dine is supported (Guilday
etal., 1978:57). Pollen data (Davis, 1969 a, 19696;
Miller, 1973) suggest that essentially modern
mixed coniferous/deciduous forest characterized
the Glaciated Allegheny Plateau — north of Mead-
owcroft and at much higher elevation — by ca.
7,000 B.C. (8,950 B.P.) (Johnson, 1981:92).

The importance of data from these and other
sites for interpreting the Meadowcroft informa-
tion cannot be underestimated. By the same to-
ken, however, climatic reconstruction for these
sites, all of which are at much higher elevations
(again, to name only one important variable) than
Meadowcroft, cannot be used directly or vicari-
ously to mold or condition our understanding of
the possibly localized temperate climatic, vege-
tative, and faunal montage that characterized that
site. Indeed, Meadowcroft’s temperate setting may
have been one of its primary attractions to ab-
original populations.

The geological data from Meadowcroft indicate
that whatever the exact linkage between climatic
variation and sedimentation at the site, the end of
the Pleistocene and the onset of the Holocene is not
signaled by any remarkable or, indeed, any percep-
tible interlude in the site’s depositional history. Sedi-
mentation rates at the Pleistocene/Holocene bound-
ary apparently were not essentially different from
those occurring at the site today.

The Meadowcroft geological, floral, and faunal
data summarized above clearly lend support to the
hypothesis that the Pleistocene/Holocene transition
was not accompanied by any dramatic event or se-
ries of events witnessed in the extant ecofactual or

1984

ADOVASIO ET AL.-MEADOWCROFT ROCKSHELTER

367

depositional record at this exceptional site for which
so many lines of evidence exist. These data sets
benefit from the comprehensiveness of their collec-
tion, processing, and evaluation and imply that late
Pleistocene conditions in areas south of the Wis-
consinan glacial front were neither as uniform nor
as inexorably harsh as some previous reconstruc-
tions for the Northeast and Mid-Atlantic areas might
suggest. In the immediate vicinity of Meadowcroft,
at least, there is very little to support the notion that
late Pleistocene climatic conditions or the faunal/
floral correlates of those conditions differed appre-
ciably from circumstances that characterized the area
in the early Holocene or, indeed, today. This does
not mean that late Pleistocene/early Holocene eco-
logical associations were the same in all details or
that conditions were identical to those at present. It

seems, rather, that the local or microclimatic pa-
rameters and the indigenous biota were simply not
substantially different. In sum, the authors share
John Guilday’s succinct characterization and inter-
pretation of the Pleistocene/Holocene transition at
Meadowcroft. Specifically, any environmental
changes that occurred during the long span of time
represented in the Meadowcroft faunal sequence took
place within a mast forest context and apparently
were of such a low order that the biota was not
seriously disturbed at the site. The great majority
of the faunal sequence post-dates the Wisconsinan/
Holocene environmental change from boreal to
temperate conditions, and it suggests that the period
of transition may have been earlier, very rapid and
relatively short in this part of western Pennsylvania.

ACKNOWLEDGMENTS

The excavations at Meadowcroft Rockshelter and throughout
the Cross Creek drainage as well as the analysis of recovered
materials were conducted under the auspices of the former Ar-
chaeological Research Program (now the Cultural Resource Man-
agement Program) of the Department of Anthropology, Univer-
sity of Pittsburgh. The incipient 1973 field project in addition to
the 1977 and 1978 field seasons were directed by J. M. Adovasio.
The 1 974-1976 field projects were co-directed by J. M. Adovasio
and J. D. Gunn. Analysis of all recovered materials is under the
ultimate direction of J. M. Adovasio. Computer studies of the
Meadowcroft/Cross Creek data base are supervised by J. D. Gunn
at the University of Texas, San Antonio, and by R. Drennan at
the University of Pittsburgh.

Generous financial and logistic support for the 1973-1978 ex-
cavations and analyses was provided by the University of Pitts-
burgh, the Meadowcroft Foundation, the National Geographic
Society, the National Science Foundation, the Alcoa Foundation,
the Buhl Foundation, the Leon Falk Family Trust, and Messrs.
John and Edward Boyle of Oil City, Pennsylvania.

Radiocarbon assays were supplied via the Radiation Biology
Laboratory, Smithsonian Institution, and in one instance by Di-
carb Radioisotope Company. Line drawings for the figures used
in this contribution are by K. Adkins, R. L. Andrews, and R.
Stuckenrath. The manuscript version of this contribution was
typed by G. LoAlbo Placone, Department of Anthropology, Uni-
versity of Pittsburgh.

LITERATURE CITED

Adovasio, J. M. 1982. Multidisciplinary research in the North-
east: one view from Meadowcroft Rockshelter. Pennsylvania
Archaeologist, 52(3—4): 57—68.

(compiler). 1982. The prehistory of the Paintsville Res-
ervoir, Johnson and Morgan counties, Kentucky (by J. M.
Adovasio, R. C. Carlisle, W. C. Johnson, P. T. Fitzgibbons,
J. D. Applegarth, J. Donahue, R. Drennan and J. L. Yed-
lowski). Ethnology Monogr., 6:1-1074.

Adovasio, J. M., R. C. Carlisle, K. Cushman, J. Donahue, J.
E. Guilday, W. C. Johnson, K. Lord, P. W. Parmalee, R.
Stuckenrath, and P. Wiegman. n.d. a. Paleoenviron-
mental Reconstruction at Meadowcroft Rockshelter, Wash-
ington County, Pennsylvania. In Environment and extinc-
tions: man in late glacial North America (J. I. Mead and D.
J. Meltzer, eds.). Peopling of the Americas Series, Center for
the Study of Early Man, Univ. Maine, Orono, in preparation.
Adovasio, J. M., J. Donahue, R. Stuckenrath, and J. D. Gunn.
1981. The Meadowcroft Papers: a response to Dincauze.
Quart. Rev. Archaeol., 2:14-15.

Adovasio, J. M., J. Donahue, K. Cushman, R. C. Carlisle, R.
Stuckenrath, J. Gunn, and W. C. Johnson. 1983. Evi-
dence from Meadowcroft Rockshelter. Pp. 163-190, in Early
man in the New World (R. Shutler, Jr., ed.), Sage Press,
Beverly Hills, California, 223 pp.

Adovasio, J. M., J. D. Gunn, J. Donahue, and R. Stuckenrath.
1975. Excavations at Meadowcroft Rockshelter, 1973-1974:
a progress report. Pennsylvania Archaeologist, 45(3): 1-30.

. 1977a. Meadowcroft Rockshelter: retrospect 1976.

Pennsylvania Archaeologist, 47(2-3): 1-93.

. 19776. Progress Report on the Meadowcroft Rock-
shelter—A 16,000 year chronical. Pp. 37-159, in Amerinds
and their paleoenvironments in northeastern North America
(W. S. Newman and B. Salwen, eds.), Ann. New York Acad.
Sci., 288:1-570.

. 1978a. Meadowcroft Rockshelter. 1977: an overview.

Amer. Antiq., 43:632-651.

Adovasio, J. M., J. D. Gunn, J. Donahue, R. Stuckenrath, J.
Guilday, and K. Lord. 19786. Meadowcroft Rockshelter.

368

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Pp. 1 40- 180, in Early man in America from a circum-Pacific
perspective (A. L. Bryan, ed.), Occas. Papers Dept. Anthro-
pology, Univ. Alberta, 1:1-327.

. 1979-1 980a. Meadowcroft Rockshelter— retrospect

1977 (Part 1). North American Archaeologist, 1(1): 3—44.

. 1979-19806. Meadowcroft Rockshelter— retrospect

1977 (Part 2). North American Archaeologist, 1 (2):99— 137.

Adovasio, J. M., J. D. Gunn, J. Donahue, R. Stuckenrath, J.
Guilday, and K. Volman. 1980. Yes Virginia, it really is
that old: a reply to Haynes and Mead. Amer. Antiq., 45:
588-595.

Adovasio, J. M., and W. C. Johnson. 1981. The appearance
of cultigens in the Upper Ohio Valley: a view from Meadow-
croft Rockshelter. Pennsylvania Archaeologist, 5 1( 1— 2):63—
80.

Adovasio, J. M. et al. (order of other authors not yet determined),
n.d. 6. The archaeology of Meadowcroft Rockshelter and
the Cross Creek drainage. University of Pittsburgh Press,
Pittsburgh, in preparation.

Beynon, D. E. 1981. The geoarchaeology of Meadowcroft
Rockshelter. Unpublished Ph D. dissert., Dept. Anthropol-
ogy, Univ. Pittsburgh, 283 pp.

Beynon, D. E., and J. Donahue. 1982. The geology and geo-
morphology of Meadowcroft Rockshelter and the Cross Creek
drainage. Pp. 31-52, in Meadowcroft: collected papers on
the archaeology of Meadowcroft Rockshelter and the Cross
Creek drainage (R. C. Carlisle and J. M. Adovasio, eds.),
Dept. Anthropology, Univ. Pittsburgh, 270 pp.

Carlisle, R. C., and J. M. Adovasio (eds.). 1982. Meadow-
croft: collected papers on the archaeology of Meadowcroft
Rockshelter and the Cross Creek drainage. Dept. Anthro-
pology, Univ. Pittsburgh, 270 pp.

Carlisle, R. C., J. M. Adovasio, J. Donahue, P. Wiegman, and
J. E. Guilday. 1982. An introduction to the Meadowcroft/
Cross Creek Archaeological Project, 1973-1982. Pp. 1-30,
in Meadowcroft: collected papers on the archaeology of
Meadowcroft Rockshelter and the Cross Creek drainage (R.
C. Carlisle and J. M. Adovasio, eds.), Dept. Anthropology,
Univ. Pittsburgh, 270 pp.

Cooper, E. L. 1972. Biological survey of headwater portion of
Cross Creek watershed in Washington County, Pennsylva-
nia. State College, Pennsylvania.

Cushman, K. A. 1982. Floral remains from Meadowcroft
Rockshelter, Washington County, southwestern Pennsyl-
vania. Pp. 207-220, in Meadowcroft: collected papers on
the archaeology of Meadowcroft Rockshelter and the Cross
Creek drainage (R. C. Carlisle and J. M, Adovasio, eds.).
Dept. Anthropology, Univ. Pittsburgh, 270 pp.

Davis, M. B. 1969a. Palynology and environmental history
during the Quaternary Period. Amer. Scientist, 57:317-332.

. 19696. Climatic changes in southern Connecticut re-
corded by pollen deposition at Rogers Lake. Ecology, 50:
409-442.

Dincauze, D. F. 1981. The Meadowcroft papers. Quart. Rev.
Archaeol., 2:3-4.

Flint, N. K. 1955. Geology and mineral resources of Somerset
County, Pennsylvania. Pennsylvania Geological Survey
County Report C56A, 267 pp.

Fowells, H. A. 1965. Silvics of forest trees of the United States.
United States Dept. Agriculture Handbook, 271:1-762.

Grant, K. E. 1973. Work plan for the Cross Creek watershed.
Manuscript on file, Soil Conservation Service, Harrisburg.

Guilday, J. E. 1967. The climatic significance of the Hoster-

man’s Pit Local Fauna, Centre County, Pennsylvania. Amer.
Antiq., 32:231-232.

. 1977. Fauna. In Meadowcroft Rockshelter: retrospect

1976, by J. M. Adovasio, J. D. Gunn, J. Donahue, and R.
Stuckenrath. Pennsylvania Archaeologist, 47(2-3):7-9.

Guilday, J. E., H. W. Hamilton, E. Anderson, and P. W.
Parmalee. 1978. The Baker Bluff Cave deposit, Tennessee,
and the late Pleistocene faunal gradient. Bull. Carnegie Mus.
Nat. Hist., 1 1 : 1-67.

Guilday, J. E., P. S. Martin, and A. D. McCrady. 1964. New
Paris No. 4: A Pleistocene Cave deposit in Bedford County,
Pennsylvania. Bull. Nat. Speleol. Soc., 26:121-194.

Guilday, J. E., and P. W. Parmalee. 1982. Vertebrate Faunal
Remains from Meadowcroft Rockshelter, Washington
County, Pennsylvania: Summary and Interpretation. Pp. 1 63—
174, in Meadowcroft: collected papers on the Archaeology
of Meadowcroft Rockshelter and the Cross Creek drainage
(R. C. Carlisle and J. M. Adovasio, eds.). Dept. Anthropol-
ogy, Univ. Pittsburgh, 270 pp.

Guilday, J. E., P. W. Parmalee, and H. W. Hamilton. 1977.
The Clark’s Cave bone deposit and the late Pleistocene pa-
leoecology of the central Appalachian Mountains of Virginia.
Bull. Carnegie Mus. Nat. Hist., 2:1-87.

Guilday, J. E., P. W. Parmalee, and R. C. Wilson. 1980.
Vertebrate faunal remains from Meadowcroft Rockshelter
(36WH297), Washington County, Pennsylvania. Manu-
script on file, Dept. Anthropology, Univ. Pittsburgh, 140 pp.

Harlow, W. M., and E. S. Harrar. 1969. Textbook of den-
drology. McGraw-Hill Book Co., New York, 512 pp.

Haynes, C. V. 1977. When and from where did man arrive in
northeastern North America: a discussion. Pp. 165-166, in
Amerinds and their paleoenvironments in northeastern North
America (W. S. Newman and B. Salwen, eds.), Ann. New
York Acad. Sci., 288:1-570.

. 1980. Paleoindian charcoal from Meadowcroft Rock-
shelter: is contamination a problem? Amer. Antiq., 45:582-
587.

. 1982. Pleistocene-Holocene boundary in the United

States; alluvial stratigraphy and geochronology. American
Quaternary Association Program and Abstracts, Seventh
biennial conference, Univ. Washington, Seattle.

Johnson, W. C. 1981. Archaeological review activities in Sur-
vey Region IV, northwestern Pennsylvania: year end report
of the Regional Archaeologist for the period September 1 979
through August 1980. A report Prepared for the Pennsyl-
vania Historical and Museum Commission by the Cultural
Resource Management Program, Univ. Pittsburgh, Pitts-
burgh, Pennsylvania, Under the Supervision of J. M. Ado-
vasio, in Accordance with the Provisions of Service Purchase
Contract 645987.

Lord, K. 1 982. Invertebrate faunal remains from Meadowcroft
Rockshelter, Washington County, southwestern Pennsyl-
vania. Pp. 186-206, in Meadowcroft: collected papers on
the archaeology of Meadowcroft Rockshelter and the Cross
Creek drainage (R. C. Carlisle and J. M. Adovasio, eds.),
Dept, of Anthropology, Univ. of Pittsburgh, 270 pp.

Mead, J. I. 1980. Is it really that old? A comment about the
Meadowcroft “Overview.” Amer. Antiq., 45:579-582.

Michels, J. W., and I. F. Smith (eds.). 1967. A preliminary
report of archaeological investigations at the Sheep Rock
Shelter Site, Huntingdon, Pennsylvania. Pennsylvania State
Univ., State College, Pennsylvania, 1:1-426; 2:427-943.

Miller, N. G. 1973. Late-glacial and post-glacial vegetation

1984

ADOVASIO ET AL. — MEADOWCROFT ROCKSHELTER

369

change in southwestern New York State. Bull. New York
State Mus. Sci. Service, 420:1-102.

Ray, L. L. 1974. Geomorphology and Quaternary geology of
the glaciated Ohio River Valley— a reconnaissance study.
U.S. Geological Survey Professional Papers, 826:1-74.

Stile, T. 1982. Perishable artifacts from Meadowcroft Rock-
shelter, Washington County, southwestern Pennsylvania. Pp.
130-141, in Meadowcroft: collected papers on the archae-
ology of Meadowcroft Rockshelter and the Cross Creek
drainage (R. C. Carlisle and J. M. Adovasio, eds.), Dept.
Anthropology, Univ. Pittsburgh, 270 pp.

Stuckenrath, R., J. M. Adovasio, J. Donahue, and R. C. Car-
lisle. 1 982. The stratigraphy, cultural features and chro-
nology at Meadowcroft Rockshelter, Washington County,
southwestern Pennsylvania. Pp. 69-90, in Meadowcroft: col-
lected papers on the archaeology of Meadowcroft Rock-
shelter and the Cross Creek drainage (R. C. Carlisle and J.
M. Adovasio, eds.), Dept. Anthropology, Univ. Pittsburgh,
270 pp.

Vento, F. J., J. M. Adovasio, and J. Donahue. 1980. Exca-
vations at Dameron Rockshelter ( 1 5J023A), Johnson Coun-
ty, Kentucky. Ethnology Monogr., 4:1-235.

Volman, K. C. 1981. Paleoenvironmental Implications of Bo-
tanical Data from Meadowcroft Rockshelter, Pennsylvania.
Unpublished Ph.D. dissert., Graduate College, Texas A&M
Univ., 225 pp.

Weigman, P. G. 1977. Flora. In Meadowcroft Rockshelter:
retrospect 1 976, by J. M. Adovasio, J. D. Gunn, J. Donahue,
and R. Stuckenrath. Pennsylvania Archaeologist, 47(2-3):
5-7.

Yedlowski, J. L., J. M. Adovasio, J. Donahue, R. C. Carlisle,
K. Cushman, H. B. Rollins, and J. FI. Schwartz. 1982.
Archaeological data recovery at Three Rockshelters in the
Tombigee River Multi-Resource District, Alabama and Mis-
sissippi. An Interim Report Prepared for the U.S. Depart-
ment of the Interior, National Park Service, Mid-Atlantic
Region Under the Supervision of J. M. Adovasio. Principal
Investigator, Under Contract Number C-54028(8Q).

Address (Adovasio): Professor and Chairman of Anthropology; Director, Cultural Resource Management

Program, Department of Anthropology, University of Pittsburgh, Pittsburgh, Pennsylvania 15260.

Address (Donahue): Professor of Geology and Planetary Sciences and Anthropology, University of Pitts-

burgh, Pittsburgh, Pennsylvania 1 5260.

Address (Carlisle): Research Assistant Instructor, Assistant to the Chairman, Department of Anthropology

and Editor, Cultural Resource Management Program, University of Pittsburgh, Pittsburgh, Pennsylvania
15260.

Address (Cushman): 3262 Cripple Creek Trail, Boulder, Colorado 80303.

Address (Stuckenrath): Radiation Biology Laboratory, Smithsonian Institution, Washington, D.C. 20560.

Address (Wiegman): Western Pennsylvania Conservancy, 316 Fourth Avenue, Pittsburgh, Pennsylvania

15222.

HISTORICAL BIOGEOGRAPHY OF FLORIDA
PLEISTOCENE MAMMALS

S. David Webb and Kenneth T. Wilkins
ABSTRACT

The Pleistocene history of land mammal distributions in Flor-
ida can be drawn from that state’s rich fossil record. The present
essay interprets the geographic affinities of mammalian species
from five rich Quaternary faunas as well as the modem fauna of
Florida. Brief geographic accounts of the various mammal species
are presented. The results are tabulated and interpreted according
to a simple model which recognizes three routes by which land
mammals could enter the Florida peninsula. The northern route
accounted for between 4% and 10% of successive Quartemary
faunas and is at its highest in the modem mammalian fauna. The
Gulf Coastal Corridor has been far more important, transmitting
between 1 7% and 34% of successive mammalian faunas, but is
at its lowest in the modem fauna. The northern species apparently

arrived more frequently within interglacial intervals (10.9% as
compared with 5.9% of glacial faunas), whereas the Gulf Coastal
Corridor species entered more abundantly during glacial intervals
when the corridor was much wider (27.6% of glacial faunas as
compared with 22.2% of interglacial faunas). Many Pleistocene
species that occurred in Florida and around the Gulf Coastal
Plain also extended northward up the Atlantic Coastal Plain to
South Carolina. The classic view of Florida as an “ice-age winter
resort” for northern or temperate taxa during glacial intervals is
incorrect. Throughout the Pleistocene, about two-thirds of Flor-
ida’s mammal fauna was also widespread in the southeastern
United States and beyond. Endemic Florida species were ex-
tremely rare.

INTRODUCTION

Two circumstances render Florida a favorable
place to study the historical biogeography of mam-
mals. First, the state has produced a rich record of
Pleistocene mammals. And secondly its location as
a peninsula in the southeastern corner of the United
States limits the potential routes for terrestrial fau-
nal connections. Thanks especially to John Guil-
day’s work to the north and to that of Claude Hib-
bard, Ernest Lundelius, and others to the west, it is
possible to determine the probable geographic his-
tory of many Florida Pleistocene mammals.

In this essay, we propose a simple model of how
Florida’s Pleistocene environments may have af-
fected its terrestrial biota during alternating glacial
and interglacial intervals. We then recount the known
distributional histories of Florida mammal groups.
And finally we attempt crudely to quantify these
diverse histories in terms of a few common dispersal
patterns. In this manner, our simple model approx-
imates the complex history of Pleistocene mammal
distributions in Florida.

THE MODEL

In the Pleistocene, as at present, Florida was a
south-facing cul-de-sac with its toe in the tropics.
The northern opening into the peninsula extends
350 kms from the mouth of the Apalachicola River
on the west to the north-flowing St. Johns River on
the east at a latitude of about 30°. As indicated in
Fig. 1, non- volant biota could enter the peninsula
by any one of three more or less distinct routes.
First, they could enter from the west along the Gulf
Coastal Plain, crossing the Apalachicola, Ochlock-
onee and Suwannee rivers. Secondly, they could en-
ter from the north by any of several relatively un-
impeded avenues. It should be noted that although
these northern approaches are often referred to as
“highlands,” that is a peculiarly Floridian perspec-

tive and the nearest truly “Appalachian” environ-
ments lie at least 240 kms farther north across south-
ern Georgia. Or thirdly, terrestrial biota could enter
the peninsula from the northeast along the Atlantic
Coastal Plain.

The greatest environmental changes that affected
the peninsula during the Pleistocene epoch resulted
from sea-level changes of worldwide (eustatic) scope.
During glacial time-intervals the sea fell some 140
m with respect to the present elevation of the pen-
insula. During interglacial intervals, the sea rose as
little as 3 m and as much as 40 m above its present
level, but progressively less in successive interglacial
times. The piezometric surface (or regional water
table level) tracked these sea-level cycles in a general

370

1984

WEBB AND WILKINS — FLORIDA PLEISTOCENE BIOGEOGRAPHY

371

PLEISTOCENE CORRIDORS INTO
THE FLORIDA PENINSULA

GULF COASTAL NORTHERN ATLANTIC COASTAL

tit

non- glacial
area

glacial area

Fig. 1 . — A geographic model indicating three more or less distinct corridors by which non-volant biota could enter the Florida peninsula.
During non-glacial intervals the peninsula had approximately its present outline, whereas during glacial intervals an expanded Gulf
Coastal Corridor (stippled area) emerged from the Gulf of Mexico.

way, so that the available surface water surely de-
clined as sea-level dropped, especially in the central
peninsula where permeable Paleogene limestone
beds predominate. The most remarkable feature of
the Florida landscape during glacial intervals was
the major addition (essentially doubling) of lowland

area as the shallow bottom of the Gulf of Mexico
was exposed (Fig. 1).

In our simple model we do not attempt to cal-
culate the redistribution of particular habitats dur-
ing glacial and interglacial intervals. Palynological
studies indicate that the peninsula tended to become

372

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

more xeric during the last full glacial and presum-
ably during earlier glacial intervals as well (for ex-
ample. Watts, 1971). On the other hand, these same
studies encourage a cautious attitude toward ex-
trapolating widespread changes from a limited num-
ber of local samples. Delcourt and Delcourt (1983)
recognized “minimal change in the composition of
terrestrial plant communities and environmental

conditions over the past 20,000 years . . . between
29°30' and 33°N. latitude,” due to a persisting mar-
itime tropical airmass over Florida and adjacent
areas. In this study we have not invoked any Pleis-
tocene climatic or ecological shifts in the Florida
peninsula, even though we acknowledge that sub-
humid habitats probably expanded proportionally
over the enlarged peninsula during glacial epochs.

FAUNAL SUCCESSION

In Table 1 we present a sampling of the Pleisto-
cene faunal succession in Florida. Many other local
faunas could be listed; a number of others appear
in Webb (1974) and Kurten and Anderson (1980);
but we have selected five of the richest samples.
They span the entire Pleistocene and are rather
clearly assignable to glacial or interglacial intervals.
We have also included the Recent fauna because it
is both well-sampled and representative of a non-
glacial interval. We will summarize the distributions
of all of these species in an effort to test our simple
biogeographic model. First, however, we offer a few
specific comments on these selected faunas.

The Inglis local faunal was collected from a fissure
in the Inglis Limestone at present sea level about 1
km inland from the Gulf of Mexico (Webb, 1974).
Although the entire vertebrate fauna has not been
described, the rich samples of unusual taxa have
inspired the following special studies; Frazier, 1981,
on Erethizon ; Berta, 198 1, on Chasmaporthetes, and,
in press, on Smilodon\ Wilkins, 1984, on Geomys\
and Meylan, 1982, on the herpetofauna. The fauna
correlates rather closely with the Curtis Ranch local
fauna in Arizona which has been shown by magneto-
stratigraphic and radioisotopic studies to be 1 .9 mil-
lion years old (Johnson et al., 1975). Because the
Inglis site lies at or just below present sea level but
has no marine fossils or sediments it is correlated
with a glacial interval, presumably the Nebraskan
(Webb, 1974; Kurten and Anderson, 1980).

The Coleman II A fauna was obtained from al-
ternating beds of sands and clays filling a sinkhole
in the late Eocene Ocala Limestone in Sumter Co.,
Florida (Martin, 1974a). The site is 27 m above
present sea level. Martin’s analysis of this fauna
indicates a latest Irvingtonian age and suggests that
it was deposited during a glacial period (probably
Kansan) when sea level was at least as low as at

present. This fauna appears to correlate most closely
with the Cumberland Cave fauna of Maryland.

Two Rancholabrean faunas are used in this paper
to represent the Sangamonian interglacial interval.
Deposits at both Arredondo IIA (Alachua Co.) and
Reddick IA (Marion Co.) accumulated in fissures in
the Eocene Ocala Limestone (Kurten and Anderson,
1980). These central peninsula sites lie at about 25
m above present sea level. The mammalian faunas
of these sites have not yet been comprehensively
treated. Recent curation of the extensive microfauna
has yielded a more diverse faunal list (Table 1) for
the Arredondo IIA site than indicated in Webb
(1974). Gut and Ray (1963) described the geological
context of the Reddick IA deposits and listed 53
mammalian taxa.

Weigel (1962) produced a list of vertebrates from
the Vero deposits in Indian River Co. Recent cu-
ration of unsorted Vero material collected by Weigel
has resulted in addition of three taxa to his list—
Mustela vison, Felix atrox, and Platygonus com-
pressus. These deposits gained notoriety due to as-
sociation of extinct vertebrate fossils with human
bones and artifacts (Sellards, 1917). Later Wiscon-
sinan (Rancholabrean) and Holocene fossils (in strata
2 and 3, respectively) occur in sediments filling a
former stream channel. Bone-bearing layers are un-
derlain by a Pleistocene marine shell marl of the
Anastasia Formation (Sellards, 1916). Deposition
occurred during times of reduced sea level in the
last glacial interval. Other coastal Florida sites con-
temporaneous with Vero include Melbourne and
Seminole Field (Kurten, 1965).

The modem mammalian fauna of Florida in-
cludes approximately 53 native terrestrial species
(Stevenson, 1976; Hamilton and Whitaker, 1979;
Hall, 1981). Although not yet documented from
Florida, the swamp rabbit ( Sylvi/agus aquaticus ) is

1984

WEBB AND WILKINS- FLORIDA PLEISTOCENE BIOGEOGRAPHY

373

Table l.— List of mammals from five Pleistocene sites and from the modern fauna of Florida. Geographic affinities denoted by letter: N,

Northern: V, Varied; T, Tropical ; W, Western; E. Endemic.

Taxon

Inglis

IA

Coleman

IIA

Arredondo

IIA

Reddick

IA

Vero

Modem

Affinity

Didelphis virginiana

X

X

X

X

X

T

Scalopus aquaticus

X

X

X

X

X

X

V

Cryptotis parva

X

X

X

X

X

X

V

Sorex longirostris

X

V

Blarina carolinensis

X

X

X

X

X

X

V

Desmodus stocki

X

X

X

T

Myotis austroriparius

X

X

X

X

V

Myotis grisescens

X

N

Myotis keenii

X

N

Myotis sodalis

X

N

Myotis sp.

X

V

Pipistrellus subflavus

X

X

X

V

Eptesicus fuscus

X

X

V

Lasiurus borealis

X

X

N

Lasiurus seminolus

X

X

T

Lasiurus cinereus

X

N

Lasiurus intermedius

X

X

X

X

T

Nycticeius humeralis

X

X

V

Plecotus rafinesquii

X

V

Tadarida brasiliensis

X

X

X

T

Eumops glaucinus

X

X

T

Dasypus bellus

X

X

X

X

X

T

Dasypus novemcinctus

X

T

Kraglievichia paranense

X

T

Holmesina septentrionalis

X

X

X

T

Glyptotherium arizonae

X

T

Megalonyx leptostomus

X

T

Megalonyx sp.

X

X

X

T

Eremtherium mirabile

X

T

Glossotherium chapadmalensis

X

T

Glossotherium harlani

X

X

T

Tamias striatus

X

N

Tamias aristus

X

N

Sciurus carolinensis

X

X

X

X

V

Sciurus niger

X

V

Sciurus sp.

X

X

X

V

Glaucomys volans

X

X

X

X

X

V

Geomys propinetis

X

w

Geomys pinetis

X

X

X

X

X

V

Thomomys sp.

X

N

Castor canadensis

X

N

Erethizon kleini

X

T

Erethizon dorsatum

X

N

Hydrochoerus holmesi

X

X

X

X

T

Pitymys cf. mcnowni (=aratai)

X

N

Pitymys pinetorum

X

X

X

X

N

Microtus pennsylvanicus

X

X

N

Neofiber alleni

X

X

X

X

X

V

Ondatra cf. idahoensis

X

N

Synaptomys australis

X

X

X

N

Oryzomys palustris

X

X

X

X

T

Reithrodontomys humulis

X

X

X

X

X

X

V

Peromyscus polionotus

X

X

X

X

V

Peromyscus gossypinus

X

X

X

X

V

Peromyscus floridanus

X

X

X

X

E

Peromyscus sp.

X

X

X

X

X

V

Ochrotomys nuttalli

X

X

X

X

V

374

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table [. — Continued.

Taxon

Inglis

IA

Coleman
1 1 A

Arredondo

IIA

Reddick

IA

Vero

Modem

Affinity

Sigmodon curtisi

X

w

Sigmodon bakeri

X

E

Sigmodon hispidus

X

X

X

X

T

Neotoma n. sp.

X

V

Neotoma floridana

X

X

X

X

V

Lepus alleni

X

X

w

Sylvilagus floridanus

X

X

X

V

Sylvilagus sp.

X

X

X

X

V

Urocyon sp.

X

X

V

Urocyon cinereoargenteus

X

X

X

X

V

Vulpes vulpes

X

X

N

Canis edwardi

X

V

Canis lupus

X

V

Canis latrans

X

X

X

w

Canis dims (—ayersi)

X

X

X

V

Canis rufus

X

V

Tremarctos floridanus

X

X

X

T

Ursus americanus

X

X

X

V

Arctodus pristinus

X

X

V

Procyon n. sp.

X

V

Procyon lotor

X

X

X

X

V

Mustela frenata

X

X

V

Mustela vison

X

X

V

Lutra canadensis

X

X

V

Spilogale putorius

X

X

X

X

X

V

Mephitis mephitis

X

X

X

V

Conepatus leuconotus

X

X

X

T

Trigonictis macrodon

X

V

Felis cf. inexpectata

X

V

Felis onca

X

X

X

X

V

Felis concolor

X

X

V

Felis pardalis

X

T

Felis atrox

X

V

Felis rufus

X

X

X

X

X

X

V

Smilodon gracilis

X

V

Smilodon floridanus

X

V

Homotherium serum

X

X

V

Chasmaporthests ossifragus

X

V

Mammuthus sp.

X

X

X

V

Mammut americanum

X

X

X

V

Tapirus sp.

X

V

Tapirus veroensis

X

X

X

V

Equus sp.

X

X

X

X

V

Platygonus bicalcaratus

X

V

Platygonus cumberlandensis

X

V

Platygonus compressus

X

X

X

V

Mylohyus sp.

X

X

X

V

Mylohyus fossilis ( =nasutus )

X

V

Hemiauchenia macrocephala

X

X

X

X

V

Hemiauchenia small species

X

T

Palaeolama mirifica

X

X

X

X

T

Odocoileus virginianus

X

X

X

X

X

X

V

Capromeryx arizonensis

X

w

Ovibovinae sp. ( =1Euceratherium sp.)

X

N

Bison antiquus

X

X

X

V

Bison bison

X

V

1984

WEBB AND WILKINS- FLORIDA PLEISTOCENE BIOGEOGRAPHY

375

thought to have recently invaded the panhandle re-
gion. The total fauna might be further increased
through inclusion of the star-nosed mole ( Condylura
cristata ) whose southernmost known occurrence is
the Okefenokee Swamp in southern Georgia. Trap-

ping in extreme northern Florida in the southern
extension of this swamp will likely yield Condylura.
The modern fauna is interpreted here as represen-
tative of an interglacial period of relatively high sea
level.

PLEISTOCENE MAMMAL HISTORY

We attempted to determine the geographic affin-
ities of each species in Table 1. We inferred these
affinities from the known geographic range of each
species or its sister-group during the same or im-
mediately preceding mammal age. We recognized
the following three distinctive distributional cate-
gories: northern (N), western (W), and tropical (T).
In practice we were unable to distinguish between
species that were distributed northward into Ap-
palachian sites and those that were distributed
northeastward along the Atlantic Coastal Plain. We
also recognized two other distributional categories,
namely endemic (E), and various (V), although these
did not help sharpen our routes by which Pleistocene
land mammals reached Florida. Before compiling
the results of this tabulation, we may briefly recount
the apparent Pleistocene history of Florida’s major
land mammal groups.

Marsupials

Didelphis appeared during the Irvingtonian, but
not in the early Inglis IA local fauna, and was a
constant, often abundant presence thereafter in
Florida. The reported occurrence in the late Blancan
at Santa Fe I (Kurten and Anderson, 1980) is prob-
ably incorrect, based on Rancholabrean material
mixed in the river site with Blancan specimens. Pre-
sumably it came from Neotropical origins by way
of the Gulf Coast. Guilday (1958) noted its late or
post-Pleistocene spread northward.

Bats

Fourteen species comprise Florida’s modern bat
fauna. By at least the Rancholabrean (for example,
Reddick) Florida Pleistocene bat faunas displayed
contributions from both the north ( Myotis sp., Lasi-
urus borealis, Lasiurus cinereus ) and from the Neo-
tropics ( Desmodus , Lasiurus seminolus, Lasiurus
intermedins, Eumops, and Mormoops). The other
bat species in Table 1 are rather widespread in North
America and thus of uncertain geographic affinities.

The only bat present in the Rancholabrean of
Florida which is not present in the state today is the
vampire bat, Desmodus stocki, from Reddick and
Arredondo (Gut, 1959; Jones, 1958; Hutchinson,

1 967). Affinities of vampire bats lie within the Neo-
tropical Realm where all three extant vampire species
now occur. Their invasion of Florida presumably
coincided with colonization by the Neotropical
megafauna, whose blood surely comprised their diet.
The earliest record for Desmodus (in Florida and
elsewhere) is Inglis IA. Desmodus probably entered
Florida along the Gulf Coast corridor during a gla-
cial interval such as during Inglis IA or perhaps
earlier times. The genus persisted in the more inland
sites during the Sangamonian Interglacials, but is
unknown from the Florida Wisconsinan despite the
Wisconsinan presence of a megafauna seemingly ad-
equate for the vampire’s diet.

Insectivores

None of Florida’s insectivore species have narrow
enough geographic ranges to provide useful results
for this survey. The apparent late Pleistocene arrival
of Sorex longirostris at Haile XIB is noteworthy,
but does not clearly indicate any particular geo-
graphic route (Webb, 1974).

SCIURIDS

The distributions of tree squirrels (for example,
Sciurus species, Glaucomys volans) occurring in the
Quaternary of Florida are so widespread as to be
inconclusive as regards geographic affinities. How-
ever, both western and northern connections are
indicated by a ground squirrel and two chipmunks.
The eastern chipmunk ( Tamias striatus) now occurs
restrictedly in Okaloosa Co. in the Florida panhan-
dle in relictual woodland habitats characteristic of
the Appalachian Mountains and other boreal en-
virons (Jones, 1978). Similar northern affinities may
reasonably be accorded to the Arredondo II fauna

376

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

on the basis of presence of a mandible (UF 57101)
tentatively referred to T. aristus, the extinct “no-
blest” chipmunk, previously known only from its
Sangamonian type locality (Ladds, in northern
Georgia) and distinguished from T. striatus by its
30% larger size (Ray, 1965#).

Florida’s western connection is evinced by the
occurrence of a ground squirrel, Spermophilus sp.,
in the Haile XIVA fauna (Alachua Co., Florida).
This fauna is thought to have been deposited in the
early Wisconsinan, a glacial interval when lowered
sea levels opened the Gulf Coast Corridor as a west
to east dispersal avenue. Careful study of this ma-
terial suggests it to be distinct from Spermophilus
spilosoma, S. tridecemlineatus , and S. mexicanus
(Martin, 19746).

Pocket gophers

Two genera of pocket gophers occurred in Florida
during the Pleistocene. The record of Geomys and
more-or-less continuous, whereas that for Thomo-
mys is spotty and of shorter duration. Florida fossil
deposits record nearly two million years of evolu-
tion of the genus Geomys. A new species, described
by Wilkins (1984), characterized the early Irving-
tonian Inglis IA and mid-Irvingtonian Haile XVIA
sites. This more primitive species presumably
evolved into the extant G. pinetis by late Irvington-
ian times (Coleman IIA).

The route by which Geomys entered Florida is
uncertain, but either or both of two scenarios seem
plausible. The first has pocket gophers travelling
along the Gulf Coast corridor at times of lower sea
level (such as characterized the time of Inglis IA
deposition) when sandy marine terraces were emer-
gent. Throughout the Cenozoic, the Mississippi Riv-
er has surely been a significant barrier to many ter-
restrial species. Dispersal of pocket gophers across
the Mississippi and other such rivers (for example,
Apalachicola River in the Florida panhandle) was
likely passive, via changes in river courses rather
than active. Honeycutt and Schmidly (1979) dis-
cussed the significance of oxbowing in pocket gopher
dispersal and gene flow. Crossing of rivers flowing
over and beneath karst topography (for example
Suwannee and other rivers of the Florida peninsula)
probably occurred in several ways. Pocket gophers
could swim through very shallow waters during
moderate seasonal dry periods or even travel across
the dry river beds on the ground surface during ex-
treme droughts. The preceding scenarios character-
ize the moister interglacial periods when higher pi-

ezometric surfaces led to surface flow of rivers.
Conversely, substantial lowering of water table levels
during glacial intervals resulted in considerable un-
derground movement of rivers through karst lime-
stone and, thereby, abandonment of surficial river
beds in many areas. At such times, pocket gopher
dispersal was surely unimpeded by these dried river
beds.

Alternatively, Geomys could have reached Flor-
ida via a midwestem route running through the In-
terior Lower Plateau situated west of the Appala-
chian Mountains and east of the Mississippi River.
Parmalee and Klippel (1981) summarized late Wis-
consinan and early Holocene records of G. cf. bur-
sari us in this gap between extant G. bursarius in
Illinois and Indiana and G. pinetis in Alabama and
Georgia. As Guilday et al. (1971) suggested, it is
indeed likely that Geomys pocket gophers were con-
tinuously distributed from the Midwestern plains
into the southeastern coastal plains during the late
Pleistocene (and perhaps at other earlier times).
However, their contention that differentiation of the
two extant species occurred since the Wisconsinan
glaciation is unlikely in light of recent work (Wil-
kins, 1984) wherein the chronological range of G.
pinetis was extended to the late Irvingtonian (late
Kansan to early Illinoian Coleman IIA). Further-
more, Heaney and Timm (1983) argued that G. pi-
netis possesses several primitive features, thereby
indicating an early (pre-late Irvingtonian) split from
the G. bursarius species group. Because of the dif-
ficulty in distinguishing G. pinetis and G. bursarius
on the basis of postcranial elements, specific iden-
tification of gopher material from late Pleistocene
and early Holocene sites in the present hiatus region
(as well as a better understanding of the evolutionary
relationships and history of these species) must await
dental and craniometric analyses.

Until recently, the only published record of Pleis-
tocene Thomomys pocket gophers in the southeast-
ern United States was Simpson’s (1928) description
of T. orientalis from the Sangamonian Sabertooth
Cave deposit in west-central peninsular Florida
(Citrus Co.). Wilkins (in press) reported an ad-
ditional Sangamonian locality (Rock Springs, Or-
ange Co.) for Thomomys sp. in central peninsular
Florida. Re-examination of microfauna at the Flor-
ida State Museum Vertebrate Paleontology Collec-
tion revealed one Thomomys sp. mandible from the
late Irvingtonian Coleman IIA site in Sumter Co.,
Florida. A fourth Florida Pleistocene record for
Thomomys sp. is Williston IIIB (Levy Co.), a San-

1984

WEBB AND WILKINS -FLORIDA PLEISTOCENE BIOGEOGRAPHY

377

gamonian site collected in 1973 by Robert A. Mar-
tin. Hence, the temporal range of Thomomys sp. in
Florida extends from the late Irvingtonian into the
middle Rancholabrean.

Thomomys could possibly have reached Florida
by either the Gulf Coastal or Midwestern routes
proposed above for Geomys pocket gophers. Another
alternative, however, is dispersal southward along
the Appalachian Mountains. The presence of fossil
Thomomys potomacensis in Maryland (Cumber-
land Cave, middle Irvingtonian) and West Virginia
(lower levels of Trout Cave, late Irvingtonian) at
times slightly antedating the earliest Florida record
(late Irvingtonian Coleman II A) suggests the Ap-
palachians as the most likely of the three possible
dispersal corridors.

Microtines

Representation of microtines in Florida’s modern
fauna ( Pitymys pinetorum, Microtus pennsylvani-
cus, and Neofiber alleni) is less diverse than found
in the state’s Rancholabrean deposits (five species).
Although Microtus pennsylvanicus was relatively
widespread in Florida during the late Pleistocene, it
has only recently been added to the list of living
Florida mammals. An isolated population located
in the coastal saltmarshes of Levy Co. apparently
represents a relict population which has endured
since the late Pleistocene moderation of the Florida
climate (Woods et al., 1982). The depauperate mi-
crotine fauna at the early Irvingtonian Inglis IA site
suggests that four of the five Florida species may
not have reached the peninsula until the late Ir-
vingtonian. In fact, Reddick comprises the earliest
known record anywhere for Pitymys pinetorum. The
record of all five microtine genera in Florida is es-
sentially continuous from the late Irvingtonian
through the Rancholabrean. This includes Ondatra
as well, which is represented by the extant O. zi-
bet hicus in several Sangamonian and Wisconsinan
deposits (Webb, 1974) although this temporal dis-
tribution is not evident in Table 1 . Synaptomys last
appears in Florida in the mid-Holocene Devil’s Den
site dated about 8,000 y.b.p. (Martin and Webb,
1974) and Ondatra continued until some 3,000 years
ago.

Neofiber has been known in Florida more or less
continuously at least since Irvingtonian time. Dur-
ing the early and medial Pleistocene these round-
tailed muskrats were by no means Florida endemics,
but ranged far to the west and through the southern
Great Plains. Frazier (1977) attributed their very

late Pleistocene retreat to a relictual distribution in
Florida to the advent of prolonged freezing winter
temperatures during the last (and most severe) gla-
cial epoch.

Voles, lemmings, and other microtines are gen-
erally presumed to have boreal affinities and to in-
dicate cooler and moister climatic regimes. Repen-
ning (1983) summarized the biogeography of the
genus Pitymys in North America and traced their
invasion of the continent via Beringia. Dispersal of
microtines into Florida was undoubtedly from the
north, although whether it was via the Appalachians
or along lowland corridors east or west thereof is
uncertain.

Histricognath Rodents

The Inglis IA fauna reveals already in the earliest
Pleistocene the presence of the two hystricognath
families that continuously occupied Florida up to
the Recent, namely the porcupines (Erethizontidae)
and the capybaras (Hydrochoeridae).

The distinctive small species of Erethizon, E.
kleini, is unique to the Irvingtonian of Florida, but
before Rancholabrean times gave way to the extant
species, E. dorsatum. The final latest Pleistocene
records occur at Seminole Field and possibly New
Port Richey (Frazier, 1981). A pre-Pleistocene
species of hydrochoerid, Neochoerus dichroplax de-
scribed by Ahearn and Lance (1980) is supplanted
by a more progressive species of Neochoerus and by
the smaller living genus Hydrochoerus. Both of these
large hydrochoerid rodents survived along the Gulf
Coastal Plain in the southeastern United States until
the latest Pleistocene. The demise of Erethizon in
the southeast has been attributed by Ray and Lipps
(1970) to the activity of man.

Both of these families of hystricognath rodents
surely moved north from South America in the late
Pliocene. Late Blancan records of hydrochoerids are
known from low temperate latitudes in Mexico, Ar-
izona, and Florida (Ahearn and Lance, 1980). Sim-
ilar aged records of Erethizon are scattered from
Mexico to Idaho (Frazier, 1981). Their most prob-
able avenue into Florida is from the south or the
west via the Gulf Coastal Corridor. The apparent
elimination of hydrochoerids from the Great Basin
by late Irvingtonian time may be related to increas-
ing aridity in that region. By contrast two genera of
hydrochoerids survived, in apparent sympatry,
through the late Pleistocene in Florida and the
southeastern coastal plain.

378

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Hares and Rabbits

An abundance of Sylvi/agus floridanus is a regular
feature of nearly all Pleistocene faunas in Florida,
and the geographic distribution of that species is
exceedingly broad. More notable are the abundant
records of a large hare referable to Lepus alleni in
the Inglis IA and Coleman IIA local faunas. These
“jackrabbits” strongly indicate a broad coastal scrub
or savanna connection between Florida and their
western ranges. It is perhaps not surprising that these
faunas represent early glacial intervals.

Edentates

The history of how edentate species entered Flor-
ida is relatively easy to deduce. Surely they origi-
nated ultimately from South America. Their most
probable route is by circling the Gulf of Mexico
through tropical and subtropical latitudes. In the
Inglis local fauna six species represent six different
families (three pilosan and three cingulate), includ-
ing the earliest record of any megatheriid in North
America. The other five taxa do not represent oldest
records but apparently reached Florida during the
late Blancan (late Pliocene) age, where they are first
known in the Sante Fe River IA local fauna of prob-
able glacial age (Webb, 1974; Robertson, 1976).

A seventh edentate genus apparently reached
Florida during the late Irvingtonian. McDonald
(1984) reported Nothrotheriops from the Coleman
IIA local fauna, and explained its immigration by
postulating a corridor of scrub-like vegetation along
the Gulf of Mexico to semiarid western sites from
which it was previously known (Lull, 1 929). There
is little doubt that the Coleman IIA local fauna rep-
resents a low-water glacial interval, possibly the
Kansan (Webb, 1974; Kurten and Anderson, 1980).

Since the Irvingtonian, the only additional eden-
tate to reach Florida was the modern Dasypus no-
vemcinctus which was accidentally introduced by
human meddling, but has since spread eastward
around the northern Gulf of Mexico and into Flor-
ida on its own. Clearly this was accomplished very
recently during the present interglacial— a time when
Gulf Coast Corridor is relatively narrowed. Its
northward distribution has been limited by winter
cold and its westward distribution by available
moisture (Humphrey, 1974).

Carnivores

In general, the Florida records of Pleistocene car-
nivoran species were parts of widespread distribu-

tions. For example, Webb (1974) showed that the
large sample of Smilodon californicus from the Ran-
cho La Brea tar pits in California could not be dis-
tinguished from the sample of Smilodon floridanus
in Rancholabrean sites in Florida. And Kurten (1966)
found a close resemblance between the large sample
of Tremarctos floridanus from Florida and the
smaller sample of spectacled bears from southern
California. In view of such geographic continuity,
it is difficult to determine direction of dispersal. The
fact that Tremarctos is confined to low latitudes
throughout its known (Pleistocene) record suggests
that it either originated in Florida or extended its
range there from a Neotropical source.

The most interesting geographic pattern with re-
spect to Florida’s Pleistocene camivorans is their
ultimate retreat to the American tropics. Hog-nosed
skunks, jaguarundis, jaguars, ocelots, spectacled
bears, and possibly margays, are all Florida Pleis-
tocene species, now confined to tropical latitudes or
Neotropical ranges. The northernmost limits of these
taxa vary somewhat, but spectacled bears and jag-
uars have been recorded from the late Pleistocene
in Tennessee (Guilday and Irving, 1967; Guilday
and McGinnis, 1972). The ocelot, jaguarundi, and
possibly margay, however, appear to have had ranges
consistently restricted to latitudes below about 30°
(Gillette, 1976; Ray et al., 1963; Ray, 1964).

In summary, it is difficult to determine the direc-
tion of immigration of Florida’s Pleistocene carni-
vores, but the present geographic affinities of many
are toward the New World tropics, and a few were
essentially tropical.

Ungulates

Most of Florida’s Pleistocene ungulates are also
distributed widely across North America. As with
some of the camivorans, however, a number share
special ties to the present Neotropical Fauna. These
species include llamas, tapirs, peccaries, certain deer,
and certain proboscideans. Among Pleistocene lla-
mas, for example, the commoner genus, Hemiau-
chenia is widespread from temperate North Amer-
ica through much of South America, whereas the
rarer genus, Palaeolama is confined in North Amer-
ica to the Coastal Plain of South Carolina, Georgia,
Florida, and Texas and also occurs in southern Cal-
ifornia (Webb, 1974). Simpson (1929) recognized
the South American cervid genus Blastocerus in the
late Pleistocene of Florida, although this remains to
be confirmed by additional discoveries. An instruc-
tive case is the distribution of Cuvieronius, a pro-

1984

WEBB AND WILKINS -FLORIDA PLEISTOCENE BIOGEOGRAPHY

379

gressive gomphotheriid, distinguished by its tusk-
less, brevirostrine mandible. In the medial
Pleistocene it ranged north into the southern Great
Plains, but by late Pleistocene it occurred only
along the coastal plain from South Carolina through

Florida to Texas and is otherwise confined to Me-
soamerica and South America. Several ungulate taxa
including Tapirus , Palaeo/ama, and Cuveronius had
essentially Neotropical and Coastal Plain distribu-
tions by the late Pleistocene.

RESULTS

Summary of Table 1 provides an estimate of pro-
portional geographic affinities of the six Pleistocene
and modem faunas (Table 2). From 61% to 68% of
the taxa from each fauna have Pleistocene and/or
modem ranges not conclusively indicating routes or
directions of movement into Florida. Endemism
(that is, occurrence only in Florida during the entire
time interval under consideration) of these faunas
is low, ranging from zero to two species or up to
5.1%. Peromyscus floridanus is endemic in four fau-
nas; this species and Sigmodon bakeri are the only
endemic mammals in the Coleman IIA fauna of
latest Irvingtonian age.

The remaining one-third of the species in well-
sampled Quaternary faunas has proved amenable
to our interpretations of faunal affinities. We have
recognized these species as invaders either from the

west (having tropical or western affinities) or from
the north. Species with western ties comprise neg-
ligible portions of these faunas except for Inglis IA
which boasts four western taxa (8.5% of the fauna).
Neotropical species comprise a much larger part of
all-aged Florida faunas (between 1 5 and 28% of the
successive faunas in Table 1). Both western and
Neotropical species presumably entered Florida via
the Gulf Coastal corridor, they therefore are summed
under the heading “Gulf Coastal Corridor” in Table
2. Species with northern geographic affinities were
seldom as numerous as the Neotropical contingent
except in the present fauna which is notably high in
northern taxa and notably low in Neotropical species.
Nevertheless, the northern group is regularly rep-
resented during the Pleistocene and accounts for
between 4% and 10% of each fauna.

DISCUSSION

Our results largely contradict the traditional view
of Florida as a southern cul-de-sac into which north-
ern fauna retreated during glacial intervals. In fact,

we can find no example that justifies Colbert’s (1942)
labelling of Florida as an “Ice Age Winter Resort.”
Presumably such an example would show temperate

Table 2 .—Summary of geographic affinities of mammalian species in six Quaternary faunas in Florida. Paired table entries are numbers

of species and corresponding percentages of each fauna.

Geographic affinities

Inglia IA
(glacial)

Coleman IIA
(glacial)

Arredondo II
(interglacial)

Reddick

(interglacial)

Vero

(glacial)

Modem

(interglacial)

Tropical

12

7

8

15

14

8

25.5%

17.9%

20.0%

27.8%

26.4%

15.1%

Western

4

1

—

1

1

1

8.5%

2.6%

1.9%

1.9%

1.9%

Northern

2

3

4

2

3

10

4.3%

7.7%

10.0%

3.7%

5.7%

18.9%

Varied

29

29

27

35

35

33

61.7%

66.7%

67.5%

64.7%

66.0%

62.2%

Endemic

—

2

1

1

—

1

5.1%

2.5%

1.9%

1.9%

Total taxa

47

39

40

54

53

53

Gulf Coastal Corridor

16

8

8

16

15

9

(Tropical + Western)

34.0%

20.5%

20.0%

29.7%

28.3%

17.0%

380

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

species surviving glacial intervals in Florida and
then returning northward during interglacials. On-
datra comes close to fitting this pattern, but is not
restricted to Florida at any time. John Guilday, more
than a decade ago, pointed out an important dis-
tinction between the strong boreal influences reach-
ing Appalachia during glacial intervals and the rel-
atively limited northern influence that reached
Florida (Guilday, 1971).

According to our data there is a moderate pattern
of northern species distributing southward into
Florida during the Pleistocene. The best examples
are found, as might be expected, among microtine
rodents. All five microtine species that reached Flor-
ida during the Pleistocene appear in medial or late
Irvingtonian time. When it is fully studied the me-
dial Irvingtonian site, Haile XVI A, will shed im-
portant light on the schedule of microtine first ap-
pearances. It is noteworthy that Ondatra and
Synaptomys have retreated northward during Re-
cent time and Microtus hangs on only by a minute
relictual population. Erethizon has also retreated
northward from Florida, following northern conifer
belts. On the other hand, several very late Pleisto-
cene arrivals in Florida such as Vulpes vulpes, Sci-
urus niger, and Sorex longirostris probably have
northern origins. The number of truly boreal forms
is very limited compared to even the low end of the
gradient studied by Guilday et al. (1978). According
to our tabulations, the average proportion of Florida
land-mammal species having northern affinities is
higher during interglacial epochs (10.9%) than dur-
ing glacials (5.9%). This may be an artifact, based
especially on the strong influence of northern groups
in the Recent fauna.

Much the strongest pattern of land mammal dis-
tribution in Florida’s Pleistocene record is the Gulf
Coastal Corridor (Table 2). This corridor transmit-
ted immigrant species from source areas that ap-
parently lay both southward in the Neotropical
Realm and westward at latitudes around 30° North.
We have not always been able clearly to distinguish
the western from the tropical patterns, and retain
the impression that the two patterns are in many
taxa interrelated. The sloth genus Not hr other tops,
for example, could be counted as western on the
basis of the most familiar late Pleistocene records
in western United States, but it could equally well
be considered Neotropical on the basis of its ulti-
mate origin in early Pleistocene time. Although the
western pattern appears weaker than the tropical
pattern in Pleistocene mammals, a very strong west-

LAND MAMMAL DISTRIBUTIONS
INTO THE FLORIDA PENINSULA

A. Glacial Period

B. Interglacial Period

Fig. 2. — Preliminary calculations of apparent land-mammal dis-
persals into the Florida Peninsula (arrow widths proportional to
percentage of Florida species) A) during glacial periods, B) during
interglacial periods. The percentage of newly appearing taxa from
each corridor is averaged over several glacial or interglacial in-
tervals (see text).

1984

WEBB AND WILKINS- FLORIDA PLEISTOCENE BIOGEOGRAPHY

381

ern pattern is evident in terrestrial lower vertebrates.
The Inglis squamate reptile fauna, recently de-
scribed by Meylan (1982), consists of 31 species, of
which 15 are xeric-adapted forms with western
counterparts, another three have circum-Gulf dis-
tributions, and two, Rhineura and Ophisaurus, are
Florida relicts with former western ranges. Together
western and Neotropical taxa entered Florida via
the Gulf Coastal Corridor.

The strongest surge of new immigrants from the
west and south appears to have arrived in the early
Pleistocene. Possibly the first glacially lowered sea
levels provided the strongest impetus for range ex-
tensions along the greatly broadened Gulf Coastal
Corridor. The fact that the isthmian connection be-
tween South and North America had just formed
also augumented the wave of new immigrants en-
tering the southern end of the corridor, but this
explains much more about the mammal faunal in-
flux than it does about the herpetofaunal influx
(Webb, 1976; cf. Meylan, 1982).

The actual record of Pleistocene mammals in the
Gulf Coastal Plain westward and southward from
Florida is sparse except in the latest Pleistocene,
principally because the Mississippi Embayment per-
sistently buried itself under a thick mass of sedi-
ments. The Ingleside fauna from Galveston, Texas,
provides the richest late Pleistocene sample from
the western Gulf of Mexico and is almost identical
to Florida faunas of comparable age (Lundelius,
1972). This is the best direct evidence of the west-
ward continuity of the Gulf Coastal Corridor with
its maximum exposure during a glacial interval. The
optimum conditions for burial of fossil faunas along
the coastal plain probably developed during marine
transgressive phases toward the end of each glacial
interval when backfilling of stream and karst sys-
tems was most prevalent. While it is difficult to test
this idea in early and medial Pleistocene chronol-
ogies, it does seem to be confirmed by the prevalence
of late Wisconsinan sites during the last deglaciai
hemicycle.

The least important corridor for introducing land
mammals into peninsular Florida is the Atlantic
Coastal Plain. Indeed we were unable with certainty
to distinguish any species that might have used that
corridor rather than a more inland northern route.
Thus the distinction we made a priori between these
two potential routes in our Fig. 1 did not materialize
in practice. On the other hand, this route exhibited
great continuity with the Gulf Coastal Corridor—
virtually all Pleistocene land mammals that entered
Florida from the west or Neotropics also extended
their ranges at least a short distance northward along
the Atlantic Coastal Plain. In recent times expansion
of the nine-banded armadillo’s range into Georgia
exemplifies this pattern. In the Pleistocene such
Neotropical groups as glyptodonts, chlamytheres,
armadillos, several groups of sloths, and capybaras,
reached the coastal plain of South Carolina (Ray,
1 96 5Z?; Roth and Laerm, 1980). Likewise a number
ofcarnivorans and ungulates that had extended their
ranges into the Neotropical region via Florida and
the Gulf Coastal Plain maintained a foothold along
the Atlantic Coastal Plain into late Pleistocene time;
examples include Tremarctos, Tapirus, Mylohyus,
and Palaeolama. The giant tortoises of the genus
Geochelone also followed this route. Florida thus
played a pivotal role in linking the Atlantic Coastal
Plain with the Gulf Coastal Plain.

The role of the Gulf Coastal Corridor was poten-
tially much greater during glacial intervals when sea
levels were lower and the extent of the coastal plain
vastly increased. In general this effect seems to be
corroborated by our limited data set in Table 2; for
the mean proportion of western and Neotropical
species during glacial intervals (27.6%) exceeds that
for interglacial (22.2%) (Fig. 2). The modern fauna
has by far the lowest influence from Neotropical and
western taxa. The early glacial interval represented
by the Inglis I A local fauna, evidently played the
most profound role in establishing a major Neo-
tropical and western faunal contingent in the south-
eastern coastal plain region.

ACKNOWLEDGMENTS

First, we wish to acknowledge John Guilday’s abiding influence
both as a meticulous scientist and as a profoundly sensitive per-
son. Among the others whose insights we have shared over the
years we thank particularly Clayton Ray, Bjom Kurten, Ernest
Lundelius, Elaine Anderson, Robert Martin, and Greg Mc-

Donald. Our Florida Pleistocene work has been supported by
grants from the National Geographic Society and the National
Science Foundation.

Contribution no. 228 from the Vertebrate Paleontology Lab-
oratory, Florida State Museum.

382

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

LITERATURE CITED

Ahearn, M. E., and J. F. Lance. 1980. A new species of Neo-
choerus (Rodentia: Hydrochoeridae) from the Blancan (late
Pleistocene) of North America. Proc. Biol. Soc. Washington,
93:435-442.

Berta, A. 1981. The Plio-Pleistocene hyaena Chasmaporthetes
ossifragus from Florida J. Vert. Paleo., 1:341-356.

. In press. The sabrecat, Smilodon gracilis from Florida

and a discussion of its relationships (Felidae; Mammalia).
Florida State Mus. Bull. Nat. Sci.

Colbert, E. H. 1942. Ice age winter resort. Nat. Hist., 50: 16-

21.

Delcourt, P. A., and H. R. Delcourt. 1983. Late Quaternary
vegetational dynamics and community stability reconsid-
ered. Quaternary Res., 19:265-271.

Frazier, M. K. 1977. New records of Neofiber leonardi (Ro-
dentia: Cricetidae) and the paleoecology of the genus. J.
Mamm., 58:368-373.

. 1981. A revision of the fossil Erethizontidae of North

America. Bull. Florida St. Mus., Biol. Sci., 27:1-76.

Gillette, D. G. 1976. A new species of small cat from the late
Quaternary of southeastern United States. J. Mamm., 57:
664-476.

Guilday, J. E. 1958. The prehistoric distribution of the opos-
sum. J. Mamm., 39:39-43.

. 1971. The Pleistocene history of the Appalachian

mammal fauna. Pp. 233-262, in The distributional history
of the biota of the southern Appalachians. Part III: Verte-
brates (P. C. Holt, ed.), Virginia Polytech. Inst., Research
Div. Monogr., 4:1-306.

Guilday, J. E., H. W. Hamilton, E. Anderson, and P. W.
Parmalee. 1978. The Baker Bluff Cave Deposit, Tennes-
see, and the late Pleistocene Faunal Gradient. Bull. Carnegie
Mus. Nat. Hist., 1 1:1-67.

Guilday, J. F., H. W. Hamilton, and A. D. McCrady. 1971.
The Welsh Cave peccaries ( Platygonus ) and associated fau-
na, Kentucky Pleistocene. Ann. Carnegie Mus., 43:249-319.

Guilday, J. E., and D. C. Irving. 1967. Extinct Florida spec-
tacled bear Tremarctos floridanus (Gidley) from central Ten-
nessee. Bull. Natl. Speleol. Soc., 29:149-162.

Guilday, J. E., and H. McGinnis. 1972. Jaguar (Panthera
onca) remains from Big Bone Cave, Tennessee and east cen-
tral North America. Bull. Nat. Speleol. Soc., 34:1-14.

Gut, H. J. 1959. A Pleistocene vampire bat from Florida. J.
Mamm., 40:534-538.

Gut, H. J., and C. E. Ray. 1963. The Pleistocene vertebrate
fauna of Reddick, Florida. Quar. J. Florida Acad. Sci., 26:
315-328.

Hall, E. R. 1981. The Mammals of North America. 2 vols.,
2nd ed. 1181 pp. J. Wiley and Sons, New York.

Hamilton, W. J., and J. O. Whitaker. 1979. Mammals of the
eastern United States. Cornell Univ. Press, Ithaca and Lon-
don, second ed. 346 pp.

Heaney, L. R., and R. M. Timm. 1983. Relationships of pocket
gophers of the genus Geomys from the central and northern
Great Plains. Univ. Kans. Mus. Nat. Hist., Misc. Publ., 74:
1-59.

Honeycutt, R. L., and D. J. Schmidly. 1979. Chromosomal
and morphological variation in the plains pocket gopher
Geomys bursarius, in Texas and adjacent states. Occas.
Papers Mus., Texas Tech Univ., 58:1-54.

Humphrey, S. R. 1974. Zoogeogaphy of the nine-banded ar-
madillo (Dasypus novemcinctus) in the United States. Bio-
science, 24:457-462.

Hutchinson, J. H. 1967. A Pleistocene vampire bat (Desmodus
stocki) from Potter Creek Cave, Shasta County, California.
Contrib. Mus. Paleont., Univ. California, 3:1-6.

Johnson, N. M., N. D. Opdyke, and E. H. Lindsay. 1975.
Magnetic polarity stratigraphy of Pliocene-Pleistocene ter-
restrial deposits and vertebrate faunas. San Pedro Valley,
Arizona. Bull. Geol. Soc. Amer., 86:5-12.

Jones, C. 1978. Eastern chipmunk, Tamias striatus (Linnaeus).
Pp. 35-36, in Mammals, Vol. 1, Rare and Endangered Biota
of Florida (J. N. Layne, ed.), Univ. Presses Florida, Gaines-
ville, 52 pp.

Jones, J. K., Jr. 1 958. Pleistocene bats from San Josectio Cave,
Nuevo Leon, Mexico. Univ. Kansas Publ., Mus. Nat. Hist.,
9:389-396.

Kurten, B. 1965. The Pleistocene Felidae of Florida. Bull.
Florida State Mus., Biol. Sci., 9:215-273.

. 1966. Pleistocene bears of North America: 1. Genus

Tremarctos, spectacled bears. Acta Zool. Fennica, 1 15:1 —
120.

Kurten, B., and E. Anderson. 1980. Pleistocene mammals of
North America. Columbia Univ. Press, New York, 442 pp.

Lull, R. S. 1929. A remarkable ground sloth. Mem. Yale Pea-
body Mus., 3(2): 1-39.

Lundelius, E. L., Jr. 1972. Fossil vertebrates from the late
Pleistocene Ingleside Fauna, San Patricio County, Texas.
Bur. Econ. Geol., Univ. Texas, Rept. Invest., 77:1-74.

Martin, R. A. 1974a. Fossil mammals from the Coleman IIA
fauna, Sumter Co. Pp. 35-99, in Pleistocene mammals of
Florida (S. D. Webb, ed.), Univ. Florida Presses, Gainesville,
270 pp.

. 19746. Fossil vertebrates from the Haile XIVA fauna,

Alachua Co., Pp. 100-1 13, in Pleistocene mammals of Flor-
ida (S. D. Webb, ed.), Univ. Florida Presses, Gainesville,
270 pp.

Martin, R. A., and S. D. Webb. 1974. Late Pleistocene mam-
mals from the Devil’s Den Fauna, Levy County. Pp. 114-
145 in Pleistocene mammals of Florida (S. D. Webb, ed.),
Univ. Florida Presses, Gainesville, 270 pp.

McDonald, H. G. 1984. An Irvingtonian record of Nothro-
theriops in Florida. In The biology of edentates (Edentata:
Nothrotheriinae) (G. Montgomery, ed.), Smithsonian Inst.
Press, in press.

Meylan, P. 1982. The squamate reptiles of the Inglis IA Fauna.
Bull. Florida State Mus., Biol. Sci., 27:1-85.

Parmalee, P. W., and W. E. Klippel. 1981. A late Pleistocene
population of the pocket gopher, Geomys cf. bursarius in the
Nashville Basin, Tennessee. J. Mamm. ,60:38 1-835.

Ray, C. E. 1964. The jaguarundi in the Quaternary of Florida.
J. Mamm., 45:330-332.

. 1965a. A new chipmunk, Tamias aristus, from the

Pleistocene of Georgia. J. Paleontol., 39:1016-1022.

. 19656. A glyptodont from South Carolina. Charleston

Mus. Leaflet, 27:1-12.

Ray, C. E., and L. Lipps. 1970. Southerly distribution of por-
cupine in eastern United States during late Quaternary time.
Bull. Georgia Acad. Sci., 28(2):24.

Ray, C. E„ S. J. Olsen, and H. J. Gut. 1963. Three mammals

1984

WEBB AND WILKINS- FLORIDA PLEISTOCENE BIOGEOGRAPHY

383

new to the Pleistocene fauna of Florida, and a reconsider-
ation of five earlier records. J. Mamm., 44:373-395.

Repenning, C. A. 1983. Pitymys meadensis Hibbard from the
valley of Mexico and the classification of North American
species of Pitymys (Rodentia: Cricetidae). J. Vert. Paleo., 2:
471-482.

Robertson, J. S. 1976. Latest Pliocene mammals from Haile
XVA, Alachua County, Florida. Bull. Florida State Mus.,
Biol. Sci., 20:11 1-186.

Roth, J. A., and J. Laerm. 1980. A late Pleistocene vertebrate
assemblage from Edisto Island, South Carolina. Brimleyana,
3:1-29.

Sellards, E. H. 1916. Human remains and associated fossils
from the Pleistocene of Florida. Florida Geol. Surv., 8th
Ann. Kept, pp. 121-160.

. 1917. On the association of human remains and extinct

vertebrates at Vero, Florida. J. Geol., 25:2-4.

Simpson, G. G. 1928. Pleistocene mammals from a cave in
Citrus County, Florida. Amer. Mus. Novitates, 328:1-16.

. 1929. Pleistocene mammalian fauna of the Seminole

Field, Pinellas County, Florida. Bull. Amer. Mus. Nat. Hist.,
56:561-599.

Stevenson, H. M. 1976. Vertebrates of Florida. Univ. Presses
Florida, Gainesville, 607 pp.

Watts, W. A. 1971. Postglacial and interglacial vegetation his-
tory of southern Georgia and central Florida. Ecology, 52:
676-690.

Webb, S. D. 1974. Pleistocene mammals of Florida. Univ. Flor-
ida Presses, Gainesville, 270 pp.

. 1976. Mammalian faunal dynamics of the Great Amer-
ican Interchange. Paleobiology, 2:220-234.

Weigel, R. D. 1962. Fossil vertebrates of Vero, Florida. Florida
Geol. Surv., Spec. Publ., 10:1-59.

Wilkins, K. E. 1984. Evolutionary trends in Florida Pleistocene
pocket gophers (genus Geomys) with description of a new
species. J. Vert. Paleont., 3:166-181.

. in press. Pleistocene mammals from the Rock Springs

local fauna, central Florida. Brimleyana.

Woods, C. A., W. Post, and C. W. Kilpatrick. 1982. Microtus
pennsylvanicus (Rodentia: Muridae) in Florida: A Pleisto-
cene relict in a coastal saltmarsh. Bull. Florida State Mus.,
Biol. Sci., 28:25-52.

Address: Florida State Museum, University of Florida, Gainesville, Florida 32611.

TWO IRVINGTONIAN (MEDIAL PLEISTOCENE) VERTEBRATE
FAUNAS FROM NORTH-CENTRAL KANSAS

Ralph Eshelman and Michael Hager

ABSTRACT

Two Irvingtonian localities, the Courtland Canal and Hall Ash
local faunas, are reported from Jewell County, north-central Kan-
sas. The Courtland Canal local fauna consists of eight aquatic
and eight terrestrial mollusc, five fish, three amphibian, nine
reptilian, seven bird, and 17 mammalian species including the
rare Soergelia mayfieldi. The paleoclimatological and paleogeo-
morphological interpretations of the locality suggest a deposi-
tional environment of a slow-moving stream or larger river slough
with a wooded border and relatively dry sandy prairie nearby.
The Courtland Canal local fauna is regarded as middle Irving-
tonian and comparable to early Kansan and late Aftonian age
High Plain faunas such as the Sappa, Mullen I, and Rock Creek.

The Hall Ash local fauna was recovered from beneath the

Hartford Ash (dated at 0.74 mybp) and consists of six aquatic
and sixteen terrestrial mollusc, four amphibian, one reptilian,
and 1 1 mammalian species. The paleoclimatological and paleo-
geomorphological interpretations of the site suggest a deposi-
tional environment of a prairie pond with prairie areas nearby.
The Hall Ash local fauna is regarded as middle-late Irvingtonian
and comparable to but possibly slightly older than late Kansan
age High Plains faunas such as the Vera and Cudahy.

Collectively, both faunas suggest a paleoclimate of milder win-
ters, cooler summers, and more effective moisture. We know of
no environmental regime where these presently disharmonious
species could live sympatrically today.

INTRODUCTION

The Courtland Canal local fauna was discovered
while prospecting exposures along the Courtland Ir-
rigation Canal on 22 May 1973, during field work
on the White Rock local fauna (Eshelman, 1975).
The locality contained abundant fossil bone pro-
truding from a paleo-channel fill deposit exposed
along an irrigation canal. A Cam's jaw, elements of
horse, proboscidian, camel, antelope, bovid, numer-
ous birds, herps, fish, and molluscs were lying on
the outcrop surface. Approximately 400 pounds of
matrix were collected and washed for possible mi-
crofauna. Although the initial microsamples were
not promising, complete jaws of Ondatra and Bla-
rina were later recovered.

Preliminary study of the material suggested a
Kansan age. It wasn’t until the summer of 1 978 that
a field crew was organized and serious collection of
the locality began. Between 29 May and 5 June,
nearly 5 tons of matrix were washed and picked
from the Courtland Canal site and nearly 2 tons
from a newly found locality, the Hall Ash site. The
latter site was discovered while prospecting an aban-
doned ash pit within sight of the Courtland Canal
locality. The matrix was washed using the method
of Hibbard (1949 A). Some of the clayey matrix at
the Courtland Canal site required three cycles of
drying and washing before adequate breakdown of
the matrix was suitable for picking.

GEOLOGY

The Courtland Canal site is exposed in a paleo-channel stream
fill bisected by the Courtland Irrigation Canal on Myren Inter-
mill’s property in the EVi, SE'A, Sec. 8, T1S, R6W, Webber, IVi
min. quad, topographic map, Jewell Co., Kansas, 1969, at an
elevation of approximately 1,630 ft (Fig. 1). The section from
bottom to top (Figs. 2 and 3) consists of interbedded and inter-
fingered silts (Unit A), coarse limonitic sand (Unit B), a dark
organic gley paleosol (Unit C), and a mottled reddish sandy-clay
and clayey sand (Unit D). Unit D may be equivalent to the
Loveland Formation. The gley paleosol has a dip 8° west and a
strike of N 52°E. The majority of the specimens were collected
from Unit A in light tan silts containing caliche nodules on the
northeast side of the canal over a horizontal distance of approx-
imately 20 m. All of the microfauna was collected from Unit A
within a 1 square m between 1.6 and 2.3 m above the road at

approximately the level where a skull of Soergelia was excavated.
On the other side of the canal, a crushed proboscidian skull and
antilocaprid and peccary material were recovered. Specimens not
recovered from Unit A were discovered in spoil piles on either
side of the canal in matrix believed to be equivalent to Unit A.

The Hall Ash site is located in the northeast comer of an
abandoned ash pit, 0.9 km southeast of the Courtland Canal
locality on the Harry Hall property in the EVi, SW'/t, Sec. 9, T1S,
R6W, Scandia N.W. IV2 min. quad, topographic map, Jewell Co.,
Kansas, 1969, at an elevation of approximately 1,620 ft (Fig. 1).
The section consists of a lower greenish gray silt (Unit A), possibly
equivalent to the Sappa Formation, a bedded ash mixed with
silts (Unit B) identified by fission-track as 0.7 million years old,
suggesting the Hartford Ash and finally, an overlying red-molted
clay with pebbles (Unit C) possibly equivalent to the Loveland

384

1984

ESHELMAN AND HAGER — IRVINGTONIAN FAUNAS FROM KANSAS

385

98

o

Fig. 1. — Locality map of Jewell Co., north-central Kansas showing locations of Courtland Canal and Hall Ash sites. Arrows point to
test pit and Webber Road localities.

Formation. See Figs. 4 and 5 for section drawing and photo of below the quarry floor. The bottom of Unit B was not reached,

site. Unit C lies above the ash over the entire quarry site. The Along Webber Road, located between and south of the Courtland

ash (Unit B) is only exposed on the south wall of the quarry. The Canal and Hall Ash localities, at NW‘A, SW'A, Sec. 1 6, T 1 S, R6 W,

ash is believed to have accumulated in a paleovalley or some Webber lxh min. quad, topographic map, Jewell Co., Kansas,

form of depression represented by Unit A. This exact relationship 1 969, this same ash is exposed (see Fig. 1 ). Here the relationships

could not be substantiated at the ash site due to quarrying activity of Unit A and B are clearly visible. The ash is exposed over an

and extensive spoil heaps covering much of the quarry; but this 18 m interval along the road. See Figs. 6 and 7 for geological

is the relationship shown along Webber Road discussed below. section and photo of site. The above same geological relationships

All the fossils were collected from the northeast comer of the are also visible in what appears to be a test pit 0.8 km south of

quarry in undisturbed sediments of Unit A. Apparently the ash the Hall Ash site. Based on these two exposures we interpret the

has pinched out here and the silts of Unit A represent the edge same stratigraphic sequence for the Hall Ash site even though
of this assumed basin. In the southeast comer of the quarry the the lower relationship is not visible anywhere in the pit. The Hall

ash reaches a thickness of 3.96 m plus another probed 1.8 m Ash fauna is therefore interpreted to be pre-ash in age. The Hall

386

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 2.— (A) Overall view of Courtland Canal locality, figure standing at spot where Soergelia skull was collected. (B) Detail view of
Courtland Canal locality, contact of Unit A and B at level of raised hand.

Ash locality is discussed by Fishel and Leonard (1955:108 and
plate 10A, we believe the locality is more accurately located in
Sec. 9, not Sec. 1 6) and referred to as Pearlette volcanic ash in
the Sappa member of the Meade Formation. Zeller (1968) and
Eshelman (1975) reviewed the Pleistocene geology of the region;
in summary, the deposits are generally referred to as Kansan age
sediments of the Sappa Formation.

Age Determination of the Hall Ash

An ash sample was collected from the Hall Ash site and sub-
mitted to John Boellstorff for a fission-track age determination

based on glass shards. These samples were tested providing a
mean age of 0.706 ± 0.017 (see Miller and Eshelman, in press,
table 3 for fission-track analysis of site). This date compares best
to the Hartford Ash (0.74 ± 0.09) rather than to the more prev-
alent Pearlette Ash restricted (0.61 ± 0.04) (Boellstorff, 1978:
table 1-1). The type Hartford Ash is located at Minnehaha Co.,
South Dakota. In addition, the Hartford Ash has been identified
at the Little Sioux or County Line Section of Harrison Co., Iowa,
where Gerry Paulson recovered an unpublished mammalian fau-
na (see Guilday and Parmalee, 1972).

AGE AND CORRELATION

The presence of Ondatra annectens and Synap-
tomys (Mictomys) cf. S. meltoni from both the
Courtland Canal and Hall Ash local faunas suggest
an Irvingtonian age for both faunas (see also the
taxonomic discussion of Ondatra). Recovery of cf.
AUophaiomys and Titanotylopus from the Court-

land Canal local fauna suggests a somewhat earlier
age than the Hall Ash local fauna where only Mi-
crotus paroperarius is present. The presence of the
Hartford Ash (0.74 mybp) immediately above the
Hall Ash local fauna also supports an Irvingtonian
age.

1984

ESHELMAN AND HAGER -IRVINGTON! AN FAUNAS FROM KANSAS

387

GEOLOGIC SECTION AT
INTERMILL SITE,
COURTLAND CANAL SITE

Fig. 3. — Stratigraphic section of exposures on east side of Irrigation Canal on Myren Intermill property, Courtland Canal locality.

The Little Sioux local fauna from Harrison Co.,
Iowa, worked by Gerald Paulson but unpublished
(see Guilday and Parmalee, 1972) was also re-
covered from beneath the Hartford Ash and pos-
sesses a very similar fauna (personal communica-
tions, Gerald Paulson, 1 1 January 1980, and Miller
and Eshelman, in press). Below the ash, Paulson
recovered Synaptomys (Mictomys) meltoni, Micro-
tus paroperarius, Tamias, Sciurus, and a Sorex like
S. dispar but slightly larger, and a snail fauna pre-
dominated by Vertigo, a cool climatic indicator.
Above the ash, Paulson recovered Geomys, a large

Blarina, Synaptomys (Mictomys) sp., and again M.
paroperarius, but the snail fauna exhibited a more
southern distribution.

The presence of Microtus paroperarius at both the
Hall Ash and County Line sites, beneath the 0.7
mybp Hartford Ash, concurs with Van der Meulen’s
(1978:143) suggestion that M. paroperarius migrat-
ed into North America prior to 0.7 million years
ago. Van der Meulen (1978:135) notes that the pres-
ence of Pitymys meadensis in the Cumberland Cave
local fauna regarded as Middle Irvingtonian may
mean M. paroperarius immigrated before P. mead-

388

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 4.— (A) Overall view of Hal! Ash locality looking northwest with Courtland Canal locality behind trees in left background at arrow.
(B) Detail view of Hall Ash locality looking southeast, figure standing where Hall Ash local fauna collected.

ensis into North America. Repenning (1983:480)
notes Allophaiomys and early Microtus are present
in early Irvingtonian faunas of North America dated

at about 1.9 mybp, but that Pitymys immigrated
less than 1 .2 mybp. Assuming that there are no eco-
logical or sampling explanations, the absence of P.

1984

ESHELMAN AND HAGER- IRVINGTONIAN FAUNAS FROM KANSAS

389

GEOLOGIC SECTION AT
SE CORNER

meadensis from the Hall Ash local fauna would
therefore suggest a pre-Cudahy, pre-Cumberland
Cave age.

The Sappa local fauna of Harlan Co., Nebraska
(Schultz and Martin, 1 970) is somewhat of an enig-
ma as it was recovered from beneath the “S” type
Pearlette ash dated at 1.2 my bp, but it compares
most favorably with the younger late Kansan Age
faunas such as the Cudahy (Kurten and Anderson,
1980:32; Eshelman and Hibbard, 1981:325) from
beneath the “O” type Pearlette ash dated at 0.6

mybp. At least part of this disparity is that ash
dates above faunas five only minimum age dates,
and the amount of time between faunal disposition
and ash deposition cannot be determined. In fact,
Cudahy, Hall Ash, and Sappa may all be nearer or
older than 1.2 mybp, but not necessarily directly
correlative. The Sappa is still under study and future
work may help to clarify this relationship.

In summary, the Courtland Canal local fauna is
regarded as middle Irvingtonian and comparable to
early Kansan and late Aftonian age High Plain fau-

390

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 6. — Detail of exposures at Webber Road Ash locality showing Hartford Ash (Unit B, white layer) sandwiched between Unit A at
bottom and Unit C at top.

nas such as the Sappa, Mullen I, and Rock Creek, rable but possibly slightly older than the late Kansan

respectively, whereas the Hall Ash local fauna is age Vera and Cudahy local faunas,

regarded as middle-late Irvingtonian and compa-

TAXONOMY

Leslie Fay of Michigan State University prepared
pollen samples collected from Units A and C from
both the Courtland Canal and Hall Ash sites. Un-
fortunately, the samples were practically barren,
providing only one elm pollen grain from Unit C
of the Hall Ash site, and one each of elm, grass, and
pine from the Courtland Canal site. Fay (letter to
Hager, 18 January 1979) states, “The few grains
recovered were battered, indicating some trans-
port.” Oxidation of the sediments is probably re-
sponsible for the lack of preserved pollen. These
results are not surprising as pollen is rare from pre-
Illinoian vertebrate localities from the Pleistocene
of the High Plains (Kapp, 1970).

The molluscan faunule of the Hall Ash site (Miller
and Eshelman, in press) consists of 17 terrestrial
and six aquatic species, all collected from the Hall
Ash site. The faunule is unusual in that it contains
the first fossil occurrence of Allogona profunda from
the Pleistocene of Kansas and probable unde-
scribed species of Gastrocopta cf. G. near armifera.

Individuals of terrestrial taxa outnumber aquatic
individuals 366 to 14. Miller also identified the mol-
luscs from the Courtland Canal site. Eight aquatic
and eight terrestrial species were identified.

Gerald Smith, of the Museum of Paleontology,
University of Michigan, made preliminary identi-
fication of the fish remains. The herpetofauna has
been completed by Rogers (1982) who reports all
species have been identified from other Pleistocene
localities and only the land tortoise Geochelone is
extinct. David Steadman, Department of Vertebrate
Zoology, Smithsonian Institution, made prelimi-
nary identification of the birds. A total of 24 mam-
malian taxa were identified by the senior author.
The Courtland Canal local fauna is represented by
a total of 4 1 vertebrate and 1 6 invertebrate taxa and
the Hall Ash local fauna by 16 vertebrate and 24
invertebrate taxa. A complete list of all taxa known
to date, from both the Courtland Canal and Hall
Ash localities, is found in Table 1.

1984

ESHELMAN AND HAGER- 1RV1NGTONIAN FAUNAS FROM KANSAS

391

GEOLOGIC SECTION AT
WEBBER ROAD ASH
LOCALITY

Fig. 7. — Stratigraphic section of exposures along east side of Webber Road at the Webber Road Ash locality.

COURTLAND CANAL LOCAL FAUNA

Class Mammalia
Order Insectivora
Family Soricidae
Sorex sp.

Material. — Courtland Canal locality: USNM 304243, left pos-
terior mandible fragment with ascending ramus.

Remarks. — The specimen is that of a large Sorex.
Much of the detail is lost due to wear and breakage.

In general shape, the specimen resembles Sorex
(Neosorex) megapalustris Paulson from the Cudahy
local fauna. Comparison with S. cudahyensis Hib-
bard is difficult due to the lack of teeth. It differs
from Microsorex pratensis in that the ascending ra-
mus is more pointed and elongated and the form of
the mandibular foramen is deeper and more exca-
vated.

392

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 1 . — Fossils from the Courtland Canal and Hal I Ash local-
ities.

Taxa

Court-

land Hall

Canal Ash

Mollusca

Aquatic

Ferrissia cf. fragilis
Ferrissia rivularis

+

Gyraulus parvus

+

Helisoma anceps

+

Helisoma trivolvis

+

Physa sp. (broken)

+

Physa gyrina
Pisidium compression

+

Planorbula campestris
Sphaerium sp.
Sphaerium simile

+

Stagnicola caperata
Valvata tricarinata

+

Terrestrial

Carychium exiguum
Allogona profunda
Stenotrema leai
Zonitoides nitidus
Euconulus fulvus
Discus cronkhitei

+

Cionella lubrica
Helicodiscus parallelus

+

Nesovitrea electrina
Succinea ovalis
Hawaiia minuscula

+

Zonitoides arboreus

+

Gastrocopta cf. G. armifera
Gastrocopta armifera

+

Vallonia gracilicosta

+

Vallonia pulchella

+

Pupilla sinistra
cf. Succinea
Deroceras aenigma

+

Class Osteichthyes
Order Cypriniformes
IMinytrema sp.

+

Notropes spp.

+

Semotilus cf. atromaculatus

+

Order Silunformes
Ictalurus melas

+

Order Perciformes
Lepomis cf L. humilis

+

Class Amphibia
Order Urodela
Ambystoma tigrinum
Ambystoma maculatum
Order Anura
Rana pipiens

+

Rana catesbeiana

+

Bufo hemiophrys

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

Table {. — Continued.

Court-

land Hall

Taxa

Canal

Ash

Class Reptilia
Order Chelonia

Geochelone sp.

+

Chrysemys picta

+

Chrysemys scripta

+

Order Squamata

Heterodon nasicus

+

Nerodia sipedon

+

Thamnophis proximus

+

Thamnophis radix

+

Thamnophis sirtalis

+

Columber constrictor

+

Agkistrodon contortrix

+

Class Aves
Order Anseriformes

Aserineae

+

cf. Anas crecca

+

size of Anas discors

+

size of A. clypeata

+

Order Galliformes

cf. Tympanuchus

+

Order Charadriiformes
Scolopacidae

size of Capella gallinago

+

Order Passeriformes

cf. Muscicapidae (Turdinae)

+

Class Mammalia
Order Insectivora

Sorex aff. S. cinereus

+

Sorex sp.

+

Microsorex cf. M. pratensis

+

Blarina sp.

+

Order Edentata

IGlossotherium harlani

+

Order Carnivora

Lutra cf. L. canadensis

+

Canis latrans

+

Order Rodentia

Spermophilus cf. S. richardsonii

+

+

Spermophilus (Spermophilus) sp.

+

Cynomys sp.

+

Geomys sp.

+

+

Castoroides cf. C. ohioensis

+

Phenacomys cf. P. intermedius

+

cf Allophaiomys sp.

+

Microtus paroperarius

+

Ondatra annectens

+

+

Synaptomys ( Mictomys ) cf S. meltoni

+

+

Zapus sandersi

+

Order Perissodactyla

Equus sp.

+

1984

ESHELMAN AND HAGER- IRVINGTONIAN FAUNAS FROM KANSAS

393

Table 1. — Continued.

Taxa

Court-

Sand

Canal

Hall

Ash

Order Artiodactyla

Platygonus cf. P. vetus

+

Titanotylopus sp.

+

Antilocapridae gen. et sp. unident.

+

Soergelia mayfieldi

+

Order Proboscidea

Gomphotheriidae gen. et sp. unident.

+

Blarina sp.

Material. — Courtland Canal locality: USNM 304244, right
mandible with incisor and P4-M3.

Measurement. — Length from tip of incisor to tip of condyle,
12.97 mm; length from tip of incisor to posterior border of M3,
8.68 mm; MrM3 length, 4.61 mm; M,-M2 length, 3.58 mm; P4
1.38 by 1.07 mm; M,, 2.13 by 1.37 mm; M2, 1.63 by 1.17 mm;
M3, 1.22 by 0.91 mm (note the incisor appears to be slightly
extended from the alveolus).

Remarks. — The specimen is smaller than all me-
dial Pleistocene specimens; M,-M2 length of Cu-
dahy specimen 3.8 mm, and width of M, of Cudahy
specimen (Paulson, 1961:134) 1.4 mm; M,-M3
length of three Kanopolis specimens (Hibbard et al.,
1978:16) 4.8 mm, 5.0 mm, and 5.1 mm. Hibbard
(1956:163) mentions a Blarina from the late Blan-
can age Dixon local fauna as a large short-tailed
shrew. Eshelman (1975:25) also reports a large Bla-
rina from the similar age White Rock local fauna,
having an M, length of 2.5 mm and width of 1.5
mm. Martin (1973:53) lists a new Blarina species
(unnamed) from the Nebraskan? age Java local fau-
na. Martin gives no discussion or measurements for
this taxon.

The external cingulum on M,-M2 is more pro-
nounced than in Recent specimens. This character
is also noted by Hibbard et al. (1978:16) for a den-
tary (UM 61263) recovered from a Cudahy local
fauna equivalent (UM-K3-71) in Kansas. The lin-
gual cingulum on M2 does not appear to be better
developed than Recent specimens as Paulson (1961:
134) noted for a specimen (UM 36802) from the
Cudahy local fauna.

Graham and Semken (1976:437) suggest the phena
represented in living populations of Blarina are sim-
ilar to those recovered in the late Kansan age Cum-
berland Cave local fauna. Whether the Courtland
Canal specimen can be attributed to one of these
phena, an extinct species as Martin (1973) suggests,
or a combination of one of the above coupled with

its positive Bergman’s response, cannot be deter-
mined with the material in hand.

Order Rodentia

Family Sciuridae

Spermophilus cf. S. richardsottii (Sabine)

Material.— Courtland Canal locality: USNM 304232, right M,
or M2.

Measurements. — I.?)! mm (estimated) by 2.96 mm.

Remarks. — The isolated molar is broken along
the anterior surface and is badly worn. It compares
best to Recent specimens of Spermophilus richard-
sonii. Also recorded from the Cudahy (Paulson,
1961:134), this is the earliest record for S. richard-
sonii. Richardson’s ground squirrel presently ranges
north and west of the Courtland Canal locality in
South Dakota, Wyoming, and Colorado (Hall and
Kelson, 1959:339). Later age Illinoian and Ran-
cholabrean fossil localities are also south of the pres-
ent range (Kurten and Anderson, 1980:214).

Family Geomyidae
Geomys sp.

Material. —Courtland Canal locality: USNM 304245, right
dentary with I; USNM 304246, right dentary with I; USNM
304268, left dentary with I. P4-M,; USNM 304247, six upper
incisor fragments, five lower molars, two upper molars, P4, P4,
M3.

Remarks. — Paulson (1961:137) reported Geomys
tobinensis from the Kansan age Cudahy fauna of
southwest Kansas. The material reported here is too
fragmentary for specific determination. No material
of Thomomvs is recognized, although Paulson (1961:
137) and Hibbard et al. (1978:22) have recovered
specimens from the Cudahy and Kanopolis faunas,
respectively.

Family Castoridae
Castoroides cf. C. ohioensis Foster

Material.— Courtland Canal locality: USNM 304237, four in-
cisor fragments; USNM 304238, proximal head of femur; USNM
304239, proximal half of ulna.

Measurements. — Maximum width of incisor 25.6 mm.

Remarks.— The incisor fragments may represent
one or more incisors. Each exhibits enamel crinkling
or crenulation. This character plus large size differ-
entiates the specimens from Paradipoides stova/i,
Rinker and Hibbard (1952:99), and Procastorides
sweeti Barbour and Schultz (1937:7), which are
smaller and have no enamel crinkling (see also
Woodburne, 196 1:82, table III). The Courtland Ca-

394

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

nal specimens compare well to other Castoroides
material in the USNM collections. Castoroides
ohioensis is also reported from the Yarmouth age
Kanopolis local fauna of Kansas (Hibbard et al.,
1978:23) and several Illinoian faunas from the Great
Plains (Hibbard, 1970:table 6).

Family Cricetidae
Subfamily Microtinae

cf. Allophaiomys sp.

Material. — Courtland Canal locality: USNM 304254, right
mandible fragment with M, and M2, left mandible fragment with
M2 and M„ right mandible fragment with incisor fragment, two
right M,s, three left M,s, right M2, left M2, two left M3s, four
right M‘s, four left M‘s, right M2, two left M2s.

Remarks. — In general, the material matches the
descriptions given by Hibbard (1944:729) and Paul-
son (1961:1 48) for Pitymys ilanensis which Hibbard
recovered from the Cudahy local fauna. The M2s in
the left and right mandible fragments show a slight
variation in that the first and second triangles are
confluent and not closed as in all other material
observed. The anterior loop or cap of M, is a simple
crescent-shape which generally corresponds to Van
der Meulen’s (1978:107) morphotype la. This is the
same configuration for Pitymys from the Wathena
and Kentuck faunas and is more primitive than the
trefoiled loop of the conard Fissure and Cudahy
specimens (Van der Meulen, 1978:1 18). The Pity-
mys llanansisMxs from the Conard Fissure virtually
all belong type lb. The mean length measurements
of the first lower molar corresponds best to the Ken-
tuck population (Table 2) but the range of variation
overlaps all populations noted by Van der Meulen
(1978:table 2). The Courtland Canal sample is also
more elongate than Pitymys reported by Hibbard et
al. (1978:table 3) for the younger Kanopolis local
fauna of Yarmouth age. Geographical as well as tem-
poral differences certainly account for some of this
variation.

The following evolutionary lineage for Pitymys in
North America is drawn from Zakrzewski in Hib-
bard et al. (1978:29), Van der Meulen (1978: 1 1 8),
and Repenning (1983:476): Allophaiomys sp. > Al-
lophaiomys gaildayi > Pitymys Ilanensis > P. cum-
berlandensis > P. ochrogaster. Martin (1975) cor-
relates the Java fauna of South Dakota with the
Kentuck and assigns both microtids to Allophaio-
mys cf. A. pliocaenicus Kormos. Van der Meulen
(1978) believes the Kentuck and Wathena speci-
mens possibly represent a new species of Allophaio-

Table 2.— Molar measurements (mm) of Allophaiomys sp. from
Courtland Canal locality.

Length

Width

Tooth

N

Mean

Observed range

N

Mean

Observed range

M,

6

2.99

2.85-3.24

4

1.27

1.20-1.33

m2

4

1.86

1.73-1.99

4

1.17

1.12-1.24

m3

3

1.58

1.47-1.78

3

0.88

0.83-0.92

M1

8

2.64

2.29-2.97

10

1.44

1.26-1.57

M2

3

1.83

1.81-1.85

3

1.14

1.10-1.22

mys. Whether the Courtland Canal material is re-
ferable to Allophaiomys or Pitymys and a possible
new species cannot be determined without better
comparative material. Eshelman has followed Re-
penning (1983) purely for convenience and cannot
satisfactorily defend either generic placement with-
out extensive further study. Allophaiomys is also
reported from the Nash local fauna (Eshelman and
Hibbard, 1 98 1:323) and Fyllan Cave, Texas (Kurten
and Anderson, 1980:258).

Ondatra annectens (Brown)

Material. —Courtland Canal locality: USNM 304252, adult left
mandible with M,-M3; USNM 304253, right M1, left M2, left
M3.

Measurements. — Left mandible, M„ 5.92 by 2.44 mm; M2,
3.58 by 2.47 mm; M3, 3.10 by 2.08 mm; right M1, 4.40 by 2.00
mm; left M2, 3.36 by 2.26 mm; left M3, 3.37 by 2.14 mm.

Remarks.— The dentine tract height of M, (see
Eshelman, 1975:39 for terminology) is 1.8 mm; the
tooth is damaged near the position of measurement.
This tract height corresponds to an intermediate
position between Kansan and Illinoian age Ondatra
measurements as plotted by Nelson and Semken
(1970:fig. 2). The length versus width measurement
falls within the scatter of the Cudahy fauna as plot-
ted by Nelson and Semken (1970:fig. 1).

Nelson and Semken (1970) were able to demon-
strate that the M, length/width ratio could be used
to distinguish between warm and cool faunas. This
study was substantiated by corresponding ratio dif-
ferences between Recent northern and southern
specimens and fossil glacial and interglacial speci-
mens. The M, ratio from the Courtland Canal site
(2.43) corresponds to the cooler glacial faunas of
Nelson and Semken’s graph ( 1 970:fig. 7). Therefore,
a glacial age of the Courtland Canal site is suggested
by this method. Ondatra annectens has also been
identified from Cumberland Cave, Java, Trout Cave,
and Vera (Kurten and Anderson, 1980:266).

1984

ESHELMAN AND HAGER — IRVINGTGNIAN FAUNAS FROM KANSAS

395

Table 3.— Molar measurements (mm) for Synaptomys (Micto-
mysj cf. S. meltoni from the Courtiand Canal and Hall Ash site.

Length

Width

Tooth

N

Mean

Observed range

N

Mean

Observed range

M‘

5

3.05

2.79-3.38

5

1.32

1.23-1.45

M3

2

1.90

1.82-1.98

2

0.92

Q.8S-0.96

M,

6

2.77

2.59-2.94

6

1.35

1.26-1.47

m2

2

1.94

1.85-2.02

2

1.14

1.08-1.19

m3

2

1.61

1.46-1.75

2

1.03

0.93-1.12

Synaptomys (Mictomys) cf. S. meltoni Paulson

Material— Courtiand Cana! locality: USNM 304248, right
mandible with I, M,, and M2 fragment; USNM 304249, left
mandible with M2 and M3; USNM 304250, four right M's, two
left M's, right M2, left M3, right VI % four right M,s, one left M,,
one right Mz.

Remarks. — The enamel of the molars is differ-
entiated into thick and thin portions as noted by
Paulson (1961:142). There is no distinct, cement-
filled labial re-entrant on the M3s as is typical in
Synaptomys borealis (Richardson). The first triangle
of the Ms has a convex rather than a concave pos-
terior wall.

Both mandibles are damaged around the basal
capsular process, but the process terminates near
the posterior margin of the M3. This character is
noted by Hibbard (1952:10) and Paulson (1961:1 44)
as being characteristic of S. kansasensis from the
Kansan age Kentuck fauna. However, the incisor
measurement (1.01 mm) and molar measurements
(Table 3) are most closely similar to those of S.
meltoni reported by Paulson (1961:143) from the
Kansan age Cudahy fauna.

This suggests that the incisor basal capsular pro-
cess position is too variable to be used reliably as a
specific character. If two distinct species are present,
S. meltoni is characterized by smaller, compara-
tively more square, first lower molars (less elongate
and wider), whereas S. kansasensis is characterized
by slightly larger, more rectangular, first lower mo-
lars (more elongate and narrower). Additional ma-
terial may explain at least some of this variation as
due to clinal change such as recognized by Hibbard
(1963:211) and Eshelman (1975:37) for the subge-
nus Synaptomys.

One of the right upper molars has a re-entrant
angle along the face of the anterior loop. This con-
dition has also been reported by Hibbard (1944:738)
in a specimen of 1 Pity my s from the Tobin local
fauna, except the re-entrant also contains cement.

A Recent specimen, Pitymys nemoralis from Doug-
las Co., Kansas, also processes a re-entrant on the
anterior loop without cement (Hibbard, 1944:738).
Synaptomys meltoni is also reported from the Sappa
local fauna (Schultz and Martin, 1970).

Order Edentata
Family Mylodontidae

‘IGlossotherium harlani (Owen)

Material— Courtiand Canal locality: USNM 304251, badly
fragmented left tibia.

Remarks.— The tibia is large in comparison to
Rancholabrean age specimens noted by Stock (1925:
table 88) for specimens from Rancho La Brea, Cal-
ifornia, and even larger than those given by Lun-
delius (1972:table 14) for specimens from Ingleside,
Texas. This greater size is inconsistent with the gen-
eral evolutionary trend of increased size and stout-
ness of build (Kurten and Anderson 1980:143). The
greatest thickness of the distal extremity is 1 10 mm;
the greatest width of distal extremity, 141 mm; the
least width of shaft, 1 1 1 mm; and least thickness of
shaft, 54 mm.

Family Canidae
Cams cf. C. latrans Say

Material. — Courtiand Canal locality: USNM 304242, left man-
dible with Mj and M

Measurements. — M: , 24.0 by 9.4 mm; M2, 10.3 by 7.1 mm;
depth of mandible at anterior edge of M„ 19.4 mm; distance
from anterior edge of P; to posterior edge of M3, 83.4 mm;
minimum depth of mandible between P3 and P4, 16.2 mm.

Remarks.— The mandible is approximately the
size of that of a large coyote. However, the ascending
ramus on the anterior border is more rounded, the
vertical border of the mandible more rounded, and
the posterior crest of the ramus is not as hooked as
in Recent Canis latrans examined in the USNM
collections. The width of the M, is broader than two
late Blancan Borchers specimens (9.0 mm) reported
by Getz (1960:363) and the late Pleistocene type of
C. I. harriseooki (8.0 mm) reported by Slaughter
(1961:503). All of the above parameters fall within
the range of variation reported by Nowak (1979:
appendix C) for Recent male populations or late
Pleistocene specimens of C. latrans.

Lutra canadensis (Schreber)

Material.— Courtiand Canal locality: USNM 304229, proxi-
mal half of right humerus.

Remarks. — The specimen compares identically
with Recent humeri in the USNM collections, as

396

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

well as with fossil humeri from the Irvingtonian
Cumberland Cave local fauna assigned to Lutra
parvicuspis Gidley and Gazin. Other than tooth
characters, Gidley and Gazin (1933:349) note only
that this extinct otter is somewhat larger than Lutra
canadensis. Hall (1936:75) synonymized L. parvi-
cuspis with L. canadensis lataxina. Gidley and Ga-
zin (1938:42) again argued that L. parvicuspis is a
distinct species. Kurten and Anderson (1980:158)
refer L. parvicuspis to the extant species.

Order Perissodactyla
Family Equidae

Equus sp.

Material.— Courtland Canal locality: USNM 304223, left up-
per M1? fragment; USNM 304221, right lower molar fragment;
USNM 304222, left lower third premolar fragment?; USNM
304224, two upper incisors; USNM 304225, distal two-thirds of
left metatarsal; USNM 304226, distal end of metapodial; USNM
304227 proximal two-thirds of phalanx; and USNM 304228, left
ramus fragment with roots of DP2.

Remarks.— The upper incisors have cups as does
Equus niobrarensis Hay and Equus laurentius Hay,
both from the Hay Springs local fauna. The right
lower molar and left third lower premolar fragment
are the size of specimens identified as Equus sp.
from Hay Springs in the USNM collections. Both
metatarsal fragments and the phalanx are the size
of Equus excelsus specimens from Niobrara River
in the USNM collection. All of the material com-
pares well to Equus specimens from the early middle
Pleistocene of the High Plains.

The right lower molar fragment, USNM 304221
is heavily worn and measures 23.0 by 15.8 mm. The
metatarsal fragment, USNM 304225, has a greater
length and width measurement of 52.5 by 39.0 mm
on the distal end. The remaining material is too
scrappy for measurement.

Order Artiodactyla

Family Tayassuidae

Platygonus cf. P. vetus Leidy

Material. — Courtland Canal locality: USNM 304240, left man-
dibular fragment with DP,-P4; USNM 304241, left P2, left P4?
fragment, right P4?, fragments of at least three upper molars and
two canine fragments.

Measurements. — DP,, 8.8 by 5.8 mm; DP3, 10.4 by 7.5 mm;
DP4, 17.8 by 10.2 mm; LP2, 12.4 by 11.5 mm; LP4?, 12.6 by
1 5.0 mm.

Remarks.— All of the upper teeth were recovered
from within a one square meter area and are be-
lieved to be representative of one individual. The

upper premolars and deciduous lower premolars
compare favorably in size to teeth of Platygonus
vetus from Cumberland Cave, Maryland (medial Ir-
vingtonian), in the USNM collections. A minimum
of two individuals are indicated. Platygonus is also
reported from the following High Plains, Kansan
age faunas— Gilliland local fauna (Hibbard and
Dalquest, 1 966:3 1), Holloman local fauna (Hay and
Cook, 1930), and Cudahy fauna (Paulson, 1961:151).
Kurten and Anderson (1980:298) have synony-
mized P. cumberlandensis and P. vetus.

Family Camelidae
Titanotylopus sp.

Material.— Courtland Canal locality: USNM 304229, left M,
fragment; USNM 304217, left metacarpal; USNM 304218, meta-
tarsal proximal end fragment.

Remarks.— The metacarpal measures 439 mm in
length, proximal transverse width, 83 mm; proximal
anteroposterior width 54 mm; mid-length trans-
verse width, 47 mm; and mid-length anteroposte-
rior width, 44 mm. The length of the metacarpal is
approximately 50 mm longer than Came/ops spec-
imens reported by Webb (1965:table 1 1).

Skinner and Hibbard et al. (1972:114) report a
metacarpal length of 400.00 (F:AM 24900) from the
Sand Draw local fauna referred to Titanotylopus
spatulus (Cope). Meade (1945) reports a metacarpal
length for four specimens of T. spatulus from the
Blanco fauna from 395 to 497, proximal width 85
to 103 and distal width ranging from 104.5 to 148
mm. Hibbard and Riggs (1949) report the following
measurements for T. spatulus from the Keefe Can-
yon local fauna for two metacarpals: length, 448,
455; width proximal end, 103 and 106; and width
distal end, 136 and 140 mm. Barbour and Schultz
(1934) report a metacarpal length of 425 for T.fricki
synonymized with T. spatulus from the Broadwater
local fauna. To our knowledge, no metacarpals are
known that have been assigned to T. nebraskensis
(Barbour and Schultz). In addition, an incisor,
USNM 304219, may belong to this taxon.

Family Antilocapridae

Genus and species indeterminate

Material. — Courtland Canal locality; USNM 304255, left max-
illary fragment with M1 and M2.

Measurements. — M1 , 12.3 by 9.3 mm; M2, 14.7 by 10.2 mm.

Remarks. — The molars compare favorably in size
and morphology with the extant Antilocapra ameri-
canus (Ord) and the extinct southwestern form

1984

ESHELMAN AND HAGER — IRVINGTONIAN FAUNAS FROM KANSAS

397

Fig. 8 .—Soergelia cf. 5. mayfieldi. (A) Dorsal-ventral view of occipital region and horn cores. (B) Ventral-anterior view of cranium
posterior and horn cores.

Stockoceros onuarosagris (Roosevelt and Burden),
(Colbert and Chaffee, 1939:14). The molars are ap-
proximately 25% smaller than S', conklingi (Stock).
Tetrameryxl knoxensis (Hibbard and Dalquest,
1960), known only from horn cores, and Capro-
meryx sp. were recovered from the Late Kansan
Gilliland local fauna of Knox County, Texas. Paul-
son (1961:151) reports an unidentified antilocaprid
from the Cudahy fauna.

Family Bovidae

Soergelia cf. S. mayfieldi (Troxell)

Material.— C ourtland Canal locality: USNM 304216, poste-
rior one-third of cranium, one-half of the left horn core, most of
the right horn core and most of the occipital region (Fig. 8).

Description.— The horn cores are flattened slightly more dorsal
ventrally than anteroposteriorly as reported by Troxell (1 9 1 5) for

the type specimen. The burrs on the horn core base do not appear
as pronounced as in the type specimen. The horn cores curve up
and outward from the cranium and then are directed forward
nearly horizontally. This differs from Kurten and Anderson's
(1980:332) description of Soergelia mayfieldi where the horn
cores curve outward and downward and then somewhat forward.
The Courtland Canal horn cores turn prominently forward. Har-
rington (letter to Eshelman, 5 January 1974) makes the following
observations on the Courtland Canal specimen: The region of
the frontals, the horn core burrs and the angle of rise of the horn
cores all compare similarly to the European Soergelia. The strik-
ing difference is the greater hom core length in the Courtland
Canal specimen. The horn cores appear to rise forward over a
greater distance from the burrs before their downward deflection.
In European Soergelia the horn cores usually are directed slightly
back prior to the forward curve and slope downward more
abruptly. Finally, the hom cores taper down much more rapidly
in European specimens. The foramen magnum agrees with the
type Soergelia mayfieldi in that the opening is on the dorsal
surface of the occiput.

398

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 4. — Cranial measurements (mm) o /'Soergelia mayfieldi /row Rock Creek, Courtland Canal, and Soergelia elisabethae/row Basel,

Germany.

Soergelia mayfieldi

Soergelia elisabethae

Measurement1

Courtland Canal

Rock Creek2

D 3763

D 4593

Width between bases of horn cores
Greatest diameter of left horn core burr

65

92

75

85

75.6

77.3

Least diameter of left horn core burr

78

74

65.9

59.8

Greatest diameter of right horn core burr

93

—

—

80.3

Least diameter of right horn core burr

78

—

—

62.8

Diameter of left horn core burr

285

—

—

232

Diameter of right horn core burr

274

—

239

229

Width of basi-occipital

58

55

—

—

Right horn core length on upper curve, burr to
existing tip-4

330

_

182

165

Right horn core length on lower curve, burr to
existing tip4

305

_

Right horn core length, from tip to upper base

268

—

-

-

Distance between skull mid-line to beginning of
horn core burr

40

40

Distance from mid-line of skull to tip of right
horn core

304

194

198

Greatest width of auditory opening

154

-

-

-

Width of condyles

89

-

-

-

Depth, occipital crest to top of foramen magnum

67

-

—

—

Depth, occipital crest to lower border of foramen
magnum

94

Greatest width of foramen magnum opening

40

-

—

34.4

Greatest height of foramen magnum opening

24

-

-

27.9

Angle of posterior divergence of horn core from
cranium

79°

Angle of proximal horn core rise

21°

-

-

-

1 Measurements after Skinner and Kaisen, 1 947 :fig. 1.

2 Rock Creek, Texas, Troxell, 1915.

3 Measurements by Leo Carson Davis from specimens in the Naturhistorisches Museum, 4051 Basel, Angnstinergasee 2, Germany; specimen D376 type specimen Siissenbom;
D459 cast of original destroyed during WW1I from Kapellenberg bei Rastenberg.

4 Right horn core missing; estimated 30 mm of tip.

Measurements from Table 4 clearly indicate the horn cores of
the Courtland Canal specimen are larger than the Rock Creek
specimens. Sexual dimorphism most certainly played a role in
this primitive musk oxen and may suggest that the Kansas spec-
imen was a male.

Remarks. — The Courtland Canal specimen is
probably the most complete cranium of Soergelia
yet discovered in North America (Nelson and Neas,
1980:227). Harrington was the first to propose the
synonomy of Preptoceras under Soergelia. Remains
of Soergelia are rare in North America. Aside from
the two specimens already mentioned, a third oc-
currence of uncertain age is reported by Harrington
(1977) from the Old Crow River locality 1 1 A of the
Yukon. Harrington (1980:46) suggests Soergelia is
a biostratigraphic indicator of Kansan age deposits
in North America. The Courtland Canal specimen
does not dispute this claim, although Kurten and
Anderson (1980) suggest that the Rock Creek local

fauna may be late Aftonian and that Soergelia spans
a large part of the Irvingtonian.

Order Proboscidea
Family Gomphotheriidae

Genus and species indeterminate

Material.— Courtland Canal locality: USNM 304230, molar
fragment.

Remarks.— The fragment is too scrappy for iden-
tification. Mr. Myren Intermill (personal commu-
nication, July 1973) states that when the Canal was
dug, several large “elephant” bones were recovered
and taken to the Kansas State University at Man-
hattan. During the field season of 1973, a badly
crushed gomphothere skull was discovered on the
west bank. During the 1978 field season the skull
could not be relocated and is presumed eroded into
the canal.

1984

ESHELMAN AND HAGER — IRVINGTONI AN FAUNAS FROM KANSAS

399

HALL ASH LOCAL FAUNA

Class Mammalia
Order Insectivora
Family Soricidae

Microsorex cf. M. pratensis

1944. Microsorex pratensis Hibbard, Bull. Geol. Soc. Amer.,
55:722.

1980. Sorex (Microsorex) pratensis Diersing, J. Mamm., 61:76.
1980. Microsorex pratensis Kurten and Anderson, 1980.

Material — Hall Ash locality: USNM 304256, right jaw frag-
ment with P4-M3 associated incisor in jaw fragment.

Measurements.— The specimen measures 9.05 mm from tip
of incisor to tip of condyle (9.2 mm in the Cudahy specimen);
M]-M3 length is 3.18 mm (3.2 mm for two Cudahy specimens);
length is 2.42 mm; P4, 1.32 by 0.46 mm; M,, 1 .40 by 0.85 mm;
M2, 1.20 by 0.84 mm; M3, 0.82 by 0.65 mm (Note— the specimen
has subsequently been broken and the P4 lost).

Remarks. — Viewed dorsally, the lingual side of
the ramus posterior to the M3 is arched abruptly
labtad. This character serves to separate this form
from those of the other small shrews, Sore x cinereus
Kerr and S. cudahyensis Hibbard of the Cudahy
fauna (Paulson, 1961:134). The mental foramen is
positioned just posterior to the metaconid of the
M,. The ramus is heavier than 5”. cinereus where
the mental foramen is directly below the metaconid
of M,. The internal temporal fossa is more squared
ventrally and is larger as in S. cinereus. The incisor
has one tubercle, whereas Paulson (1961:133) states
that S. pratensis from the Cudahy local fauna has
two well-defined tubercles, and the suggestion of a
third. The specimen is about the size of Sorex (Mi-
crosorex) hoyi but the dentition is slightly heavier.
The variation noted above in the number of incisor
tubercles and the shape of the internal temporal
fossa preclude positive identification.

Sorex aff. cinereus Kerr

Material — Hall Ash locality: USNM 304257, right mandib-
ular fragment with M,-M3.

Measurements. — M,-M3 length, 3.09 mm; M,-M2 length, 2.29
mm; Mb 1.3! by 0.64 mm; M2, 1.12 by 0.62 mm; M3, 0.87 by
0.47 mm.

Remarks.— The specimen is the size of Sorex ci-
nereus, but the dentition is narrower. Paulson (1961:
128) reports the presence of S. cinereus from the
Cudahy local fauna, but states that the fossils are
generally more robust than normal from Recent
specimens, but overlap does occur. Hibbard (1956),
Skinner and Hibbard et al. (1972), and Eshelman
(1975:24) all report a small but slightly more robust

Sorex from the Dixon, Sand Draw, and White Rock
local faunas, respectively. Eshelman (1975) states
an M2, V31973 from the Dixon local fauna mea-
sures 1.1 by 1.7 mm; this is obviously a typograph-
ical error and should read 1.1 by 0.7 mm. The Hall
Ash site specimen does not appear to belong to this
form, but is closely related to S. cinereus. The spec-
imen is smaller than S. cudahayensis Hibbard. The
mental foramen is just anterior to the metaconial
on M|, whereas it is typically more posterior in S.
ci nereus (Hibbard, 1944:719).

Three isolated ascending rami fragments (USNM
304258) appear to belong to this same taxon, but
no positive assignment is attempted.

Order Rodentia
Family Sciuridae

Spermophilus cf. S. richardsonii (Sabine)

Material. — Hall Ash locality: USNM 304235, left M1 or M2.

Measurements. — 2.36 by 3.64 mm.

Remarks. — The molar has a comparatively nar-
rower basin between the paracone and metacone
than is found in Spermophilus franklinr, in this char-
acter the molar compares best to S. richardsonii. A
lower molar (USNM 304267) may belong to this
taxon, but no assignment is attempted.

Spermophilus ( Spermophilus ) sp. Cuvier

Material- Hall Ash locality: USNM 304233, left M1 or M2;
USNM 304234, left M, or M2.

Measurements. — USNM 304233, 2.14 by 3.48 mm; USNM
304234, 2.19 by 3.22 mm.

Remarks. — The narrow triangular nature of the
upper molar suggests that these teeth belong to the
subgenus Spermophilus. In comparison to Recent
specimens, the best fit is with Spermophilus beldin-
gi. Spermophilus richardsonii, also within this sub-
genus, is known from the Cudahy fauna (Paulson,
1961:134); the upper molars, however, are more
broadly triangular in this species than in the fossil
teeth. Spermophilus beldingi is extant far to the west
in the montane basin region of Idaho and Nevada,
whereas S. richardsonii ranges nearer to the north
and west in South Dakota, Wyoming, and Colorado
(Hall and Kelson, 1959). The teeth do not appear
to be those of S. franklini or S. richardsonii.

Cynomys sp.

Material. — Hall Ash locality: USNM 304236, left P4.

Measurement. — 3.35 by 3.87 mm.

400

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Remarks. — The tooth, in early stage of wear, com-
pares best with Cynomys ludovicianus (Ord). The
specimen is larger than the late Blancan C. hibbardi
Eshelman or C. cf. C. vet us Hibbard (Eshelman,
1975) and the extant of C. gunnisoni (Baird). Cyn-
omys niobrarius Hay, C. leucurus Merriam, and C.
ludovicianus are nearer in size.

Family Geomyidae
Geomys sp.

Material. — Hall Ash locality: USNM 304262, P4, M\ isolated
molar.

Remarks. — See remarks for this taxon in Court-
land Canal local fauna.

Family Cricetidae
Subfamily Microtinae

Phenacomys cf. P. intermedius Merriam

Material. — Hall Ash locality: USNM 305261, right M,.

Remarks. —The tooth is rooted, lacks cement, and
falls within the occlusal variation noted by Guilday
and Parmalee (1972:171) and Howell (1926) for
Phenacomys. The internal re-entrant angles are
deeper than the external ones, but are not as typi-
cally pronounced as in the late Pleistocene and living
forms. In this character, the tooth in question is
more similar to Pliophenacomys primaevus Hib-
bard (1938). The tooth measures 2.29 by 1.03 mm.
The mean of M, length measurements given by Hib-
bard (in Skinner and Hibbard et al., 1972:103) for
six specimens of Pliophenacomys primarvus is 2.86
mm. Guilday et al. (1964:table 20), however, found
a range of variation from 1 .8 to 3.3 mm with a mean
of 2.67 mm for Phenacomys cf. intermedius from a
late Pleistocene fauna.

Hibbard (1944:739) reported a M2 questionably
referred to Phenacomys from the Kansan Age Wil-
son Valley faunule. This report was confirmed in
Hibbard’s Cudahy faunal lists of 1949:1421 and
1970:419. Phenacomys sp. is also reported from
Gerald R. Paulson’s unpublished Kansan Age Little
Sioux faunule, Lincoln Co., Iowa (Guilday and Par-
malee, 1972:fig. 1) as well as from Cumberland and
Trout Caves. Steward (1978:45) reports P. inter-
medius from the late Pleistocene Trapshoot local
fauna of Rooks County, Kansas. With statistically
significant numbers and more complete material, a
new Kansan age species may be warranted.

Microtus paroperarius Hibbard

Material. — Hall Ash locality: USNM 30423 1 , left jaw with M
M2, two right jaws with M,, two edentulous left jaws, six right

Table 5 .—Molar measurements (mm) o/ Microtus paroperarius
Hibbard from the Hall Ash locality.

Tooth

Length

Width

N

Mean

Observed range

N

Mean

Observed range

M,

9

2.80

2.63-2.97

9

1.04

0.97-1.09

m2

8

1.51

1.34-1.57

8

0.93

0.83-0.98

M'

8

2.13

1.96-2.35

13

1.17

1.07-1.29

M2

10

1.67

1.40-1.80

10

1.04

0.95-1.14

M3

4

1.87

1.74-1.97

4

0.94

0.83-0.99

M,s, six left M2s, seven right M,s, two left M2s, twelve right M's,
two left M's, seven right M2s, four left M2s, two right M3s, two
left M3s.

Description. — Paulson (1961:144) indicates that in the Cudahy
fauna approximately 20% of the M,s of Al. paroperarius have a
closed fifth triangle. None of the Hall Ash specimens show this
apparently variable character. Paulson ( 1 96 1 : 1 44) also states that
the “fourth labial re-entrant is normally present and developed
enough to contain cement in more than half of the specimens.”
Of the Hall Ash specimens, approximately twenty percent exhibit
cement in the fourth labial re-entrant.

Remarks. — Paulson (1961:145) states that a spec-
imen of M. paroperarius recovered from beneath
the “Pearlette ash” (type O) in Valley Co., Nebraska,
is 1 1% smaller than the average specimen from the
type locality. Additional specimens from Russell and
Lincoln Co., Kansas (Hibbard, 1944:737, 740), are
intermediate in size. Paulson (1961:145) suggests
that these size differences are geographical (negative
Bergmann’s Response) and not species differences.
The Hall Ash specimens (see Table 5) are also small-
er than the Cudahy specimens [M, length: 2.80 mm
(2.63-2.97 mm) versus 2.96 mm (2. 6-3. 4 mm) and
M, width 1.04 mm (0.97-1.09 mm) versus 1.07 mm
(1.0-1. 2 mm)], supporting Paulson’s conclusion. M.
paroperarius is also identified from the Conard Fis-
sure, Mullen I, and Vera local faunas.

Ondatra annecteus (Brown)

Material. — Hall Ash locality: USNM 304264, left M, posterior
fragment.

Remarks.— See remarks for this taxon in Court-
land Canal local fauna.

Synaptomys (Mictomys) cf. S. meltoni Paulson

Material. — Hall Ash locality: USNM 304263, right M3.

Remarks.— See remarks for this taxon in Court-
land Canal local fauna.

Family Dipodidae
Zapus sandersi Hibbard

Material. — Hall Ash locality: USNM 304259, right maxillary

1984

ESHELMAN AND HAGER -IRVINGTONIAN FAUNAS FROM KANSAS

401

fragment with P3-M2; USNM 304260, right mandibular fragment
with M,-M3.

Measurements. — USNM 304259, M1, 1.43 by 1.01 mm; M2,
1.24 by 0.97 mm; USNM 304260, M„ 1.50 by 0.94; M2, 1.39
by 0.95; M3, 0.79 by 0.70 mm.

Remarks. — Both specimens are in stage three of
wear (Klingener, 1963) and probably represent the
same individual. The anteroconid is not broader
posteriorly as typical material of similar age as-
signed to Zapus s. sandersi, but is more similar in
this character to the earlier subspecies form Z. s.
rexroadensis (Klingener, 1963:258). Therefore, no
subspecific assignment is made here and the validity
of such an assignment is questioned.

The lower first molar exhibits a posterior anter-
oconid notch as seen in the Cudahy fauna (Paulson,
1 96 1 :fig. 4B and C). The second lower molar is wider
than in the Recent forms as noted by Paulson (1961).
The measurements agree with those given by Paul-
son (1961) and are smaller than those noted by Esh-
elman (1975:44) from the late Blancan age White
Rock fauna. The greatest occlusal length of M,-M3
(3.68) is greater than that given for Z. sandersi from
the Yarmouth age Kanopolis local fauna (3.43 mm)
(Hibbard et al., 1 978). Zapus sandersi is also known
from the Sanders, Sappa, and Java local faunas.

PALEOECOLOGY

The molluscan fauna from the Hall Ash site (Mil-
ler and Eshelman, in press) exhibits a predominance
of extant species with northern (39%) and eastern
(26%) affinities. Combined with the total absence of
snails with southern distributions, a climate of cool-
er summer temperatures than now characterize the
area is inferred (Miller et al., 1980).

Miller (personal communication, 19 May 1983)
states the following:

“The Courtland Canal molluscan assemblage is evenly dis-
tributed between aquatic and terrestrial taxa, although 73% of
the individuals belong to aquatic species. The aquatic molluscs
include Ferrissia rivularis, Sphaerium simile, Valvata tricarinata,
and Helisoma anceps species which require permanent water
habitats. F. rivularis is typically found in streams and rivers with
some type of hard substrate. S. simile seems to prefer bodies of
water with little or no current action. None of the S. simile show
any sign of abrasion and one individual retains the articulated
valves. The molluscs suggest that the depositional environment
was either a small, slow moving stream or slough of a larger river.
The terrestrial snails include species that could live in a grassland
prairie situation with scattered trees or shrubs to provide shelter
and some leaf humus.”

The amphibians and reptiles from both faunas
were published by Rogers (1982) who states:

“The Courtland Canal Fauna represents a permanent prairie
pond ( Thamnophis radix, T. sirtalis ) with shallow, quiet water
with a muddy bottom (Chrysemys scripta ) bordered by a wooded
hillside ( Agkistrodon contortrix ) and not far removed from a
relatively dry, sandy prairie area (Heterodon nasicus). The Hall
Ash Fauna represents a prairie pond ( Thamnophis radix ) bor-
dered by an area of moist ground ( Ambystoma maculatus). Prairie
areas would have been nearby ( Geochelone ).”

Three species of ducks, one goose, and a sand-
piper-like shore bird, all suggestive of a standing
water habitat, were recovered from the Courtland

Canal site. In addition an open country grouse from
the same locality indicates a nearby grassland.

The Hall Ash site contains no fish or aquatic tur-
tles. Individuals of terrestrial snails outnumbered
aquatic individuals 276 to 13. Yet the Courtland
Canal site contains significant fish and aquatic turtle
remains, as well as clams, Lutra and Casloroides,
all absent from the Hall Ash locality. The muskrat
Ondatra and frogs are found in both localities,
whereas salamanders are found only in the Hall Ash
site.

Geomorphologically and sedimentologically, the
Courtland Canal locality appears to represent a pa-
leo-stream channel fill, whereas the Hall Ash site
suggests a shallow depression, such as a prairie pond.
The faunal remains support such an interpretation
and suggest further that the Hall Ash locality may
have had intermittent water and/or a moist grassy
habitat. None of the fauna except Agkistrodon sug-
gest wooded areas, but instead a grassy, low meadow
environment near or adjacent to standing water.

Cool or cold climate indicators such as Zapus,
Synaptomys meltoni, Phenacomys (see Kurten and
Anderson, 1980:258), and Ondatra annectans (see
taxonomic discussion) support the mollusc sugges-
tions of cooler temperatures. The lack of lizards and
cricetids and the abundance of microtids also agree
with this interpretation. The only disharmonious
taxon is Geochelone represented by one carapace
fragment recovered on the surface of the Hall Ash
locality. The burrowing habits of the extinct small
Geochelone are unknown, but as Preston (1971)
points out, the small species of Geochelone often
occurs in association with the large, non-burrowing
species (see also Holman, 1972:65). Living mem-

402

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

bers of the genus Geochelone are confined to open
sub-tropical or tropical regions where they feed on
succulent plants (Brattstrom, 1961). Whether they
could burrow to escape freezing weather or not, the
winters must have been milder than exist in north-
central Kansas today (Hibbard, 1960). It is possible
that this specimen is intrusive or that the tortoise
is a late “hold out” from the Yarmouthian. This
interpretation would also suggest a younger age for
the Hall Ash local fauna than for the Courtland
Canal local fauna.

Rogers (1982:177) in interpreting the paleoecol-
ogy of the amphibians and reptiles from the two
faunas states:

“Species in the Courtland Canal and Hall Ash faunas that do
not occur in Jewell Co., Kansas, today support the idea of a
greenhouse effect on the Kansan paleoclimate of Jewell Co., Kan-
sas. A comparison of species ranges with isolines of average solar

radiation and vegetation maps indicates that Chrysemys scripta,
Agkistrodon contortrix, Ambystoma maculatus, and Bufo hem-
iophrys are species that occur in areas with more water or more
cloud cover than is found in Jewell Co., Kansas, today. With the
increased moisture and moderated temperatures that would re-
sult from the greenhouse effect, this “northern” species and these
“southern” species could exist together.”

As numerous authors have hypothesized (Gra-
ham, 1976; Hibbard, 1960; Eshelman, 1975; Lun-
delius, 1976), more equable climates, with milder
winters and cooler summers and more effective
moisture are necessary for these seemingly dishar-
monious and allopatric species to have lived sym-
patrically in the past. Unless the physiologies of
these animals have changed, there is no environ-
mental regime known today which can serve as an
analog for these past faunal associations (Lundelius,
1974).

ACKNOWLEDGMENTS

The authors would like to thank Messrs. Myren Intermill and
Harry Hall for allowing us to collect on their property. The State
of Kansas graciously allowed the field crew to camp and wash
matrix at Lake Lovewel! State Park.

The following persons are acknowledged for their expertise,
for without their assistance this paper would be significantly in-
complete: John Boellstorff, research geologist, Nebraska Geolog-
ical Survey, Lincoln, Nebraska, fission-track ash dating; Leslie
Fay, The Museum, Michigan State University, pollen analysis;
Barry Miller, Department of Geology, Kent State University,
Kent, Ohio, molluscs; Gerald Smith and Nancy Neff, Museum
of Paleontology, University of Michigan, Ann Arbor, Michigan,
fishes; Karl Rogers, Division of Scientific and Technological
Studies, Adams State College, Alamosa, Colorado, amphibians
and reptiles; David Steadman, Department of Vertebrate Zool-
ogy, Smithsonian Institution, Washington, D.C., birds; Jessica

Barbour, E. H., and C. B. Schultz. 1934. A new giant camel
Gigantrocamelus fricki. Bull. Univ. Nebraska State Mus.,
2:17-27.

. 1937. An early Pleistocene fauna from Nebraska. Amer.

Mus. Novitates, 942:1-10.

Boellstorff, J. 1978. North American Pleistocene stages re-
considered in light of probable Pliocene-Pleistocene conti-
nental glaciation. Science, 202:305-307.

Brattstrom, B. H. 1961. Some new fossil tortoises from west-
ern North America with remarks on zoogeography and pa-
leoecology of tortoises. Paleont., 35:543-560.

Colbert, E. H., and R. G. Chaffee. 1939. A study of “Tetra-
meryx” and associated fossils from Papago Springs Cave,
Sonoita, Arizona. Amer. Mus. Novitates, 1034:1-21.

Diersing, V. E. 1980. Systematics and evolution of pygmy
shrews (Subgenus Microsorex) of North America. J. Mamm.
61:76-101.

Eshelman, R. E. 1975. Geology and paleontology of the early

Harrison, geologist. Department of Exhibits, National Museum
of Natural History, Smithsonian Institution, Washington, D.C.,
camels; C. R. Harrington, National Museum ofNatural Sciences,
Ottawa, Canada, Soergelia\ and L. Carson Davis, University of
Southern Arkansas, Soergelia measurements from Europe.

Augustana College, Rock Island, Illinois, provided Hager with
a faculty research grant that provided funds for the field crew
during the summer of 1978. All other expenses were borne by
the investigators of the project. Michael Carleton, Division of
Mammals, and Clayton Ray, Division of Paleobiology, National
Museum ofNatural History, Smithsonian Institution, Washing-
ton, D.C., kindly allowed access to their collections and use of
research facilities.

Finally, the authors would like to acknowledge the field assis-
tance of John Herman and Evelyne Eshelman.

E CITED

Pleistocene (late Blancan) White Rock Fauna from north-
central Kansas. Pp. 1-60, in Studies on Cenozoic paleon-
tology and stratigraphy, Claude W. Hibbard Memorial Vol.
4, Univ. Michigan Papers Paleont., 13.

Eshelman, R. E., and C. W. Hibbard. 1981. Nash Local Fauna
(Pleistocene: Aftonian) of Meade County, Kansas. Contrib.
Mus. Paleont., Univ. Michigan, 25:317-326.

Fishel, V. C„ and A. R. Leonard. 1955. Geology and ground-
water resources of Jewell County, Kansas. Bull. Kansas Geol.
Surv., 1 1 5:1-1 52.

Getz, L. L. 1960. Middle Pleistocene carnivores from south-
western Kansas. J. Mamm., 41:361-365.

Gidley, J. W„ and C. L. Gazin. 1933. New mammalia in the
Pleistocene fauna from Cumberland Cave. J. Mamm., 14:
343-347.

. 1938. The Pleistocene vertebrate fauna from Cumber-
land Cave, Maryland. Bull. U.S. Nat. Mus., 171:1-99.
Graham, R. W. 1976. Late Wisconsin mammalian faunas and

1984

ESHELMAN AND HAGER- IRVINGTONI AN FAUNAS FROM KANSAS

403

environmental gradients of the eastern United States. Pa-
leobiology, 2:343-350.

Graham, R. W., and H. A. Semken. 1976. Paleoecological
significance of the short-tailed shrews ( Blarina ), with a sys-
tematic description of Blarina ozarkensis. J. Mamm., 57:
433-49.

Guilday, J. F... P. S. Martin, and A. D. McCrady. 1 964. New
Paris No. 4: A Pleistocene cave deposit in Bedford County,
Pennsylvania. Nat. Speleo. Soc. Bull., 26:121-194.

Guilday, J. E., and P. W. Parmalee. 1972. Quaternary Peri-
glacial records of voles of the genus Phenacomys Meniam
(Cricetidae: Rodentia). Quaternary Research, 2:170-175.

Hall, E. R. 1936. Mustelid mammals from the Pleistocene of
North America. Carnegie Inst. Washington Pub!., 473:41-
119.

Hall, E. R., and K. R. Kelson. 1959. Mammals of North
America. The Ronald Press, New York, 1 :xxxi + 1-546 + 90;
2: viii + 547- 1083 + 90.

Harrington, C. R. 1977. Pleistocene mammals of the Yukon
Territory. Unpublished Ph.D. dissert., Univ. Alberta, Ed-
monton, 1060 pp.

. 1980. Faunal exchanges between Siberia and North

America: Evidence from Quaternary land mammal remains
in Siberia, Alaska, and Yukon Territory. Canadian I. An-
throp., 1:45-49.

Hay, O. P., and H. J. Cook. 1930. Fossil vertebrates collected
near, or in association with, human artifacts at localities near
Colorado, Texas; Frederick, Oklahoma; and Folsom, New
Mexico. Proc. Colorado Mus. Nat. Hist., 9:4-40.

Hibbard, C. W. 1938. An upper Pliocene fauna from Meade
County, Kansas. Trans. Kansas Acad. Sci., 40:239-265.

. 1944. Stratigraphy and vertebrate paleontology of Pleis-
tocene deposits of southwestern Kansas. Bull. Geol. Soc.
Amer., 55:707-745.

. 1949. Pleistocene vertebrate paleontology in North

America. Bull. Geol. Soc. Amer., 60:1417-1428.

. 1949A. Techniques of collecting microvertebrate fos-
sils. Contr. Mus. Paleont, Univ. Michigan, 8:7-19.

. 1952. Vertebrate fossils from Late Cenozoic deposits

of Central Kansas. Univ. Kansas, Paleontol. Contrib., 2:1-
14.

. 1956. Vertebrate fossils from Meade Formation of

southwestern Kansas. Papers Michigan Acad. Sci., Arts, Let-
ters, 41:145-203.

. 1960. An interpretation of Pliocene and Pleistocene

climates in North America. Michigan Acad. Sci. Ann. Kept.,
62:5-30.

. 1963. A late Illinoian fauna from Kansas and its cli-
matic significance. Papers Michigan Acad. Sci., Arts, Letters,
48:187-221.

. 1970. Pleistocene mammalian local faunas from the

Great Plains and Central Lowland provinces of the United
States. Pp. 395-433, in Pleistocene and Recent environ-
ments of the central Great Plains (W. Dort, Jr., and J. K.
Jones, Jr., eds.), Univ. Press Kansas, Lawrence, 433 pp.

Hibbard, C. W., and W. W. Dalquest. i960. A new antilo-
caprid from the Pleistocene of Knox County, Texas. J.
Mamm., 41:20-23.

. 1966. Fossils from the Seymour Formation of Knox

and Baylor counties, Texas, and their bearing on the Late
Kansan climate of that region. Contrib. Mus. Paleont., Univ.
Michigan, 21:1-66.

Hibbard, C. W., and E. S. Riggs. 1949. Upper Pliocene ver-
tebrates from Keefe Canyon, Meade County, Kansas. Bull.
Geol. Soc. Amer., 60:829-860.

Hibbard, C. W., R. J. Zakrzewski, R. E. Eshelman, G. Edmund,
C. D. Griggs, and C. Griggs. 1978. Mammals from the
Kanopolis Local Fauna, Pleistocene (Yarmouth) of Ells-
worth County, Kansas. Contrib. Mus. Paleont., Univ. Mich-
igan, 25: 1 1-44.

Holman, J. A. 1972. Herpetofauna of the Kanopolis local fauna
(Pleistocene: Yarmouth) of Kansas. Michigan Academician,
5( 1 ):87— 98.

Howell, A. B. 1926. Voles of the genus Phenacomys. N. Amer.
Fauna, 48: 1-66.

Kapp, P. O. 1970. Pollen analysis of pre-Wisconsin sediment
from the Great Plains. Pp. 143-155, in Pleistocene and Re-
cent environments of the Central Great Plains (W. Dort, Jr.,
and J. K. Jones, Jr., eds.), Univ. Press Kansas, Lawrence,
433 pp.

Klingener, D. 1963. Dental evolution of Zapus. J. Mamm.,
44:248-260.

Kurten, B., and E. Anderson. 1980. Pleistocene mammals of
North America. Columbia Univ. Press, New York, 442 pp.

Lundelius, E. L., Jr. 1972. Fossil vertebrates from the late
Pleistocene Ingleside fauna, San Patricio County, Texas. Rept.
Investigations Bur. Econ. Geol., Univ. Texas, 77:1-74.

. 1974. The last fifteen years of faunal change in North

America. Pp. 141-1 60, in History and prehistory of the Lub-
bock Lake Site (C. C. Black, ed.), Museum J., 15:1-160.

. 1976. Vertebrate paleontology of the Pleistocene: an

overview. Geoscience and Man, 13:45-59.

Martin, R. A. 1973. The Java local fauna, Pleistocene of South
Dakota: A preliminary report. Bull. New Jersey Acad. Sci.,
18:48-56.

. 1975. Allophaiomys Kormas from the Pleistocene of

North America. Pp. 97-100, in Studies on Cenozoic pa-
leontology and stratigraphy, Claude W. Hibbard Memorial
Vol. 3, Univ. Michigan Papers Paleont., 12:1-143.

Meade, G. E. 1945. The Blanco fauna. Univ. Texas Publ., 4401:
509-556.

Miller, B. B., R. E. Eshelman, and M. W. Hager. 1980. A
molluscan faunule collected beneath the Hartford!?) Ash (0.7 1
MYBP), Jewell County, Kansas. Abstract, Geol. Soc. Amer.,
north-central section, 14th Ann. meeting, Indiana Univ.,
Bloomington, 12:54.

Miller, B. B., and R. E. Eshelman. In press. Pleistocene Mol-
luscs From Jewell County, Kansas Associated With The
Hartford Ash (0.706 MYBP). j. Paleont.

Nelson, M. E., and J. Neas. 1 980. Pleistocene musk oxen from
Kansas. Trans. Kansas Acad. Sci., 83:215-229.

Nelson, R. S., and H. A. Semken. 1970. Paleoecological and
stratigraphic significance of the muskrat in Pleistocene de-
posits. Bull. Geol. Soc. Amer., 81:3733-3738.

Nowak, R. M. 1979. North American Quaternary Canis.
Monogr. Mus. Nat. Hist., Univ. Kansas, 6:1-154.

Paulson, G. R. 1961. The mammals of the Cudahy fauna.
Papers Michigan Acad. Sci., Arts, Letters, 46:127-153.

Preston, R. E. 1971. Pleistocene Turtles from the Arkalon
Local Fauna of Southwestern Kansas. J. Herpetol., 5:208-
211.

Repenning, C. A. 1983. Pityrnys meadensis Hibbard from the
Valley of Mexico and the classification of North American

404

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

species of Pitymys (Rodentia: Cricetidae). J. Vert. Paleo.,
2:471-482.

Rinker, G. C., and C. W. Hibbard. 1952. A new beaver and
associated vertebrates, from the Pleistocene of Oklahoma.
J. Mamm., 33:98-101.

Rogers, K. L. 1982. Herpetofaunas of the Courtland Canal
and Hall Ash Local Faunas (Pleistocene: Early Kansan) of
Jewell County, Kansas. J. Herpetol., 16:174-177.

Schultz, C. B., and L. D. Martin. 1970. Quaternary mam-
malian sequence in the Central Great Plains. Pp. 341-353,
in Pleistocene and Recent Environments of the central Great
Plains (W. Dort, Jr., and J. K. Jones, Jr., eds.), Univ. Press
Kansas, Lawrence, 433 pp.

Skjnner, M. F., C. W. Hibbard, et al. 1972. Early Pleistocene
pre-glacial and glacial rocks and faunas of north-central Ne-
braska. Bull. Amer. Mus. Nat. Hist., 148:1-148.

Skjnner, M. F., and O. C. Kaisen. 1947. The fossil Bison of
Alaska and preliminary revision of the genus. Bull. Amer.
Mus. Nat. Hist., 89:1-256.

Slaughter, B. H. 1961. A new coyote in the Late Pleistocene
of Texas. J. Mamm., 42:503-509.

Steward, J. C. 1978. Mammals of the Trapshoot local fauna.
Late Pleistocene of Rooks County, Kansas. Proc. Nebraska
Acad. Sci., 88th Ann. meeting, 14-15 April, pp. 45-46.

Stock, C. 1925, Cenozoic gravigrade edentates of western North
America, with special reference to the Pleistocene Megalon-
ychidae and Mylodontidae of Rancho La Brea. Carnegie Inst.
Washington Publ., 331:1-206.

Troxell, E. L. 1915. A fossil ruminant from Rock Creek,
Texas, Preptoceras mayfieldi sp. nov. Amer. J. Sci., 40:479-
482.

Van der Meulen, A. J. 1978. Microtus and Pitymys (Arvi-
colidae) from Cumberland Cave, Maryland, with a com-
parison of some new and old world species. Ann. Carnegie
Mus., 47:101-145.

Webb, S. D. 1965. The osteology of Camelops. Bull. Los Angeles
Co. Mus. Sci., 1:1-54.

Woodburne, M. O. 1961. Upper Pliocene geology and verte-
brate paleontology of part of the Meade Basin, Kansas. Pa-
pers Michigan Acad. Sci., Arts, Letters, 46:61-101.

Zeller, D. E. 1968. The stratigraphic succession. Bull. State
Geol. Surv. Kansas, 189:1-81.

Address (Eshelman): Calvert Marine Museum, Solomons, Maryland 20688 and Research Associate, De-
partment Paleobiology, Smithsonian Institution, Washington, D.C. 20560.

Address (Hager): Museum of the Rockies, Montana State University, Bozeman, Montana 59719.

PALEOECOLOGY OF A LATE WISCONSINAN/HOLOCENE
MICROMAMMAL SEQUENCE IN PECCARY CAVE,
NORTHWESTERN ARKANSAS

Holmes A. Semken, Jr.

ABSTRACT

Thirty-five taxa of rodents and insectivores, represented by
1,942 individuals, were recovered from seven levels within three
fossiliferous stratigraphic units in Peccary Cave, northwestern
Arkansas. Radiocarbon dates indicated that the fossils accu-
mulated between 16,700 and 2,290 years ago. The two lower
units also contained fossil remains typical of the late Wisconsinan
megafauna; those in the uppermost unit were predominately
modem species.

Statistical analysis of the associated micromammals reveals
that neither Late Wisconsinan nor Holocene climates were static,
that fauna! turnover at the end of the Pleistocene was not cata-
strophic and that the modem community is a depauperate re-
siduum developed by individualistic species response to climatic
change. First, circa 16,700 years ago, the local community was

dominated by individuals characteristic of a cool steppe with
coniferous forest patches, (2) this gave way to a mixed-forest
parkland association, (3) four dry boreal forest mammals then
disappeared, and individuals representing both meadow and de-
ciduous forest species increased in numbers to equal coniferous
ecotypes in the parkland, (4) next, two additional boreal species
were lost and deciduous parkland species became dominant, (5)
megavertebrate extinction followed and was contemporaneous
with a dramatic increase in both deciduous forest and temperate
prairie elements, (6) finally, deciduous forest species prevailed
but local enclaves for both prairie and boreal ecotypes remained.
The modem closed deciduous forest must post-date 2,290 years
B.P.

INTRODUCTION

Interpretation of Pleistocene
Local Faunas

Since the recognition of multiple Pleistocene gla-
ciations, the prediction of synchronous cooling in
temperate regions has been based partially on the
discovery of fossils of both arctic and boreal animals
in mid-latitudes, for example, Ovibos, the barren
ground muskox (Kitts, 1953). Conversely, deposits
with remains of subtropical taxa, for example, Neo-
fiber, the water rat (Blair, 1958), north of their pres-
ent range were referred to interglacial stages. This
concept is especially well-grounded when all taxa in
a local fauna are representative of a discrete region
(area of sympatry, Stephens, 1 960) either to the north
or south of their occurrence as fossils. A local fauna
of this nature is regarded as harmonious and can be
“explained by a comparatively simple environmen-
tal shift” (Hibbard and Taylor, 1960:37). This biome
migration concept was designated the “Cliseral Shift
Hypothesis” by Graham (1979). Harmonious local
faunas, for example, the Robert (Schultz, 1967),
Cragin Quarry (Hibbard and Taylor, 1960), and El-
kader (Woodman, 1982) support the contention that
existing North America biomes have antiquity and
probably fluctuated north and south in response to
glacial movements (Blair, 1958).

However, harmonious glacial-age local faunas
are rare and most others, for example, Natural

Chimneys (Guilday, 1962), contain several species
that now are allopatric. Some, like the Sandahl
(Semken, 1966), have a harmonious mammalian
component but other classes, for example, molluscs
(Miller, 1970), are disharmonious. These dishar-
monious local faunas, which include Peccary Cave,
do not have a modem analog and, as a result are
more difficult to interpret. Four hypotheses have
been proposed. (1) The specimens were mixed as a
result of reworking and thus constitute an “assem-
blage” rather than a local fauna. The Kentuck “as-
semblage” of McPherson Co., Kansas, contains both
Mictomys (northern bog lemming) and Sigmodon
(cotton rat), which now are separated from each
other by at least 625 mi (1,000 km) (Hall, 1981),
and it was explained in this manner (Hibbard, 1 952;
Semken, 1966). (2) Guilday (1962) considered slow
fossil accumulation over a period of biome migra-
tion (cliseral shift) to explain the joint occurrence
of boreal taxa ( Phenacomys cf. ungava, Sorex arc-
ticus, Synaptomys borealis, and others) with tem-
perate forms (Crypt otis parva, Neotoma florid ana,
Tamias striatus, and others) in the Natural Chim-
neys local fauna. (3) Alternatively, Guilday (1962)
suggested that the Natural Chimneys association may
be a product of owls “grazing” on life zones devel-
oped on Appalachian topography during glacial ad-
vance. Patterned ground on Appalachian summits

405

406

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

supported the previous existence of “tundra” (Guil-
day et al., 1964). Under this model a temperate
forest and its associated fauna were restricted to
valley floors with Canadian and boreal life zones
occupying intermediate elevations. This interpre-
tation is viable for montane regions but is difficult
to invoke in areas with low relief (for example,
Schultze Cave, Dalquest et al., 1969). (4) Dishar-
monious local faunas may represent climates unlike
any known today. This non-uniformitarian inter-
pretation was considered initially by both Braun
(1955) and Hibbard (1960) but rarely was taken
seriously until the 1970’s.

Dalquest (1965) discussed the disharmonious
Howard Ranch local fauna (16,775 ± 565 years
B.P.) where topographic relief is insufficient for the
life zone concept. He invoked Hibbard’s (1955)
equability model in which muted seasonal extremes
and a reduction in the number of severe cold fronts
were utilized to account for the disharmonious oc-
currence of presently allopatric taxa. Hibbard (1955)
reasoned that the northern geographic limit of an-
imals with southern distributions is largely con-
trolled by the coldest days in winter and that the
southern margin of animals with northern affinities
is determined by the hottest days in summer. Re-
duction of seasonal extremes, caused by warmer
winters and cooler summers, would permit many
presently allopatric species to coexist.

Hibbard (1960) expanded the concept of equa-
bility to explain climatic conditions associated with
many Pleistocene local faunas, the Illinoian climate
of Meade Co., Kansas, compared to that of today
“being generally cooler, the difference mostly or en-
tirely in the summers.” Hibbard (1960:23) further
noted that “climatic zoning as known at the present
with extreme low winter temperatures of the Interior
is in large part due to the strong continentality of
the climate.” Slaughter (1965) subsequently sug-
gested that megafaunal extinction at the end of the
Wisconsinan was explained best by increased tem-
perature extremes during the birthing season from
a more continental post-glacial climate. Dalquest et
al. (1969) proposed that a simple reduction in the
number of cold fronts (“northers”) on the Texas
plains, because of a reduced climatic gradient, would
result in overall warmer winters during glacial ad-
vance. Increased climatic extremes associated with
the onset of Holocene climate would limit repro-
duction of species which require cool summers to
now boreal regions, and thus reduce species diver-
sity in temperate latitudes. The modern temperate
fauna of the United States was regarded by Semken

(1974) and Martin and Webb (1974:138) as “im-
poverished,” created by the northward withdrawal
of animals now considered as either boreal or arctic
in affinity. Graham (1976) emphasized that relative
frequencies of boreal, deciduous, and steppe species
were more balanced in late Wisconsinan than mod-
em faunas of the eastern United States because the
boreal species had immigrated into regions occupied
by temperate animals in response to cooler sum-
mers. Cooler summers did not materially affect tem-
perate species in mid-latitudes so these animals did
not participate in a cliseral shift to the south. Thus,
many late Pleistocene faunas are not only dishar-
monious but are characterized by increased species
diversity. Because each boreal species responded in-
dividualistically (Graham, 1976) to glacial climates,
some dispersed further to the south than others. This
created a gradient with decreasing numbers of boreal
species in lower latitudes. Both Graham (1979) and
Guilday et al. (1978) attribute this increase in species
diversity to more equable climates. This change in
faunal composition permitted both Graham (1979)
and Martin and Neuner (1978) to independently
delineate similar Wisconsinan faunal complexes or
provinces, none of which are directly analogous to
any modern biome.

Most Wisconsinan local faunas, with the excep-
tions mentioned above, are disharmonious. How-
ever, the equable nature of associated climates is
not uniformly applicable. A cluster of local faunas
recently described from Iowa [Craigmile and Wau-
bonsie (Rhodes, in press), Brayton (Dulian, 1975),
Eagle Point (Rosenberg, 1983) and Elkader (Wood-
man, 1982)] and Wisconsin [Moscow (Foley, in
press)] all suggest that Wisconsinan winters were
substantially colder in this region. Cooler summers,
which are compatible with the equability model,
also were predicted.

The late Wisconsinan/Holocene micromammal
sequence from Peccary Cave in northwestern Ar-
kansas, which contains up to 35 species in a single
level, also is characterized by high species diversity
(Table 1). It not only lies between the equable local
faunas of the southern United States and the “cold
steppe” (Rhodes, in press) local faunas of the upper
Mississippi valley, but also provides insight into the
nature of changing climatic patterns associated with
deglaciation.

Location and Previous Investigations

A diverse vertebrate fauna, the Peccary Cave local
fauna, was collected by the Department of Geology,

1984

SEMKEN — PECCARY CAVE PALEOECOLOGY

407

University of Arkansas, from Peccary Cave, on Ben’s
Branch of Cave Creek, SW lA, SE XA, Sec. 22, R.
19W., T. 15N., Newton Co., Arkansas. Excavations
were conducted from 1967 to 1969 under the aus-
pices of National Science Foundation Grant
GB6762; James H. Quinn and Charles R. Me-
Gimsy, principal investigators. Initial excavations
by Jack McCutheon, cave discoverer, and L. Carson
Davis were summarized by Davis (1969). Semken
(1969) provided a short list of small mammals iden-
tified in the early stages of faunal analysis. Addi-
tional material was collected by the University of
Iowa in the summers of 1971 and 1972. Quinn (1972)
elaborated on the regional geomorphology, cave
sedimentation, and the large mammals discovered
during excavation. A typical late Pleistocene me-
gamammal faunal list consisting of Symbos sp., Cer-
valces sp., Tapirus terrestus, Platygonus sp., Dasy-
pus bellus, Cams dims, Sangamona sp., Mammut
sp., Mammuthus sp., Odocoileus virginianus, O. aff.
hemionus, Homo sp., and possibly E quits was re-
corded from the cave. The stratigraphic position of
these specimens was not published, but radiocarbon
dates (Quinn, 1972) range from 2,230 ± 180 (I-
4828); 2,980 ± 180 (1-3477); 4,290 ± 1 10 (1-3894)
and 4,050 ± 190 (1-5392) years B.P. on charcoal;
9,510 ± 140 years B.P. (1-4830) on Mesodon shell;
to 16,700 ± 250 years B.P. (1-5262) on bone col-
lagen; and indicate that a late-glacial through post-
glacial sedimentary sequence was deposited in the
cave. Davis (1973) interpreted the Peccary Cave
herpetofauna in relation to this transition and pre-
dicted a warming then drying trend from the bio-
stratigraphic sequence.

Stratigraphy

Davis (1973), one of the site excavators, defined
four lithologic units in Peccary Cave. Unit D, the
oldest, is an unfossiliferous yellow sand and clay
disconformably underlying unit C, a fossiliferous
light reddish-brown (2.5 YR 4/6) clay grading up-
ward into dark reddish-brown (2.5 YR 3/6-4) clay.
Unit B, a fossiliferous dark red-brown (5 YR 3/3)
clay, is conformable with Unit C but disconform-
able with overlying Unit A, a fossiliferous friable
light-brown (5 YR 4/4) matrix with masses of mam-
mal feces. The duration of either hiatus is not known.

The radiocarbon dates, 16,700 to 2,230 years B.P.,
can be supported on subjective evidence. The large
mammal component has not been completely ex-
amined, but it includes at least the eight extinct taxa
listed above (Quinn, 1972). Both demonstrate that
some of the fill is Pleistocene in age. An open sink-
hole with a developing talus cone (Davis, 1973),
occupation by owls, woodrat nests, and Archaic Tra-
dition artifacts indicate that sediment has continued
to accumulate through the Holocene in the cavern.

The present source of sediment, primarily from
the sinkhole, does not explain the bulk of fossilif-
erous matrix comprising lithologic units A, B, and
C. Quinn (1972) reasoned that these essentially hor-
izontal but channelled sediments, all of which con-
tain fish and crayfish remains, were best explained
by periodic flooding from Ben’s Branch into a hor-
izontal entrance. Subsequent excavation exposed this
entrance and a charcoal date (1-4828) from the up-
permost alluvium closing it suggests that major sedi-
mentation ceased about 2,230 years B.P. Thus, these
cave deposits clearly transcend the Pleistocene/Ho-
locene boundary.

METHODS

Excavation

Twenty-three trenches of varying lengths, depending on con-
figuration of the cave, were excavated by square and level during
the grant period (Davis, 1973, fig. 1). All were excavated in one
foot intervals except those in Trenches 4, 8, 13, and 18 where
smaller breakdown clast size permitted six inch divisions. All
matrix was washed in the manner described by Hibbard (1949)
and McKenna (1962) and the concentrate was picked for fossils
(Davis, 1973).

Numerical Evaluation

Trenches 8 and 1 3 were selected for analysis of insectivore and
rodent remains because of their smaller excavation intervals (six
inches), high bone concentration, trench juxtaposition, and pres-
ence of the younger stratigraphic units (A and B). The discon-

formity between Units A and B, not visible until the profile dried,
bisected level 3 in Trenches 8 and 13. Unfortunately, the sample
was excavated by arbitrary level prior to delineation of the dis-
conforrnity and that sample can not be separated into its natural
stratigraphic components.

Trench 15, excavated in one foot intervals, was added as it
also contained a high concentration of bone, was stratigraphically
below Trenches 8 and 13 (Unit C), and test samples appeared to
have a greater proportion of boreal taxa than that found in
Trenches 8 and 13. The contact between Units B and C is con-
formable, and lower Unit B may be represented in uppermost
level i of Trench 15 (T 15-1).

The minimum number of individuals for each taxon was es-
tablished for each square, usually on the maximum number of
right or left first lower molars. A Chi-square analysis demon-
strated that there was no significant difference between adjoining

408

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

squares of each level in Trenches 8 and 1 3, and that the combined
sample could be utilized for this analysis. Bias due to undulating
stratigraphic contacts was minimal except in level 3 (T8/13-3)
where statistical tests were neutral and the Unit A-B discon-
formity was developed.

Tests for faunal trends within the stratigraphic sequence (Ta-
bles 1 and 2) utilize the Freeman-Tukey Deviate, an exploratory
statistical technique similar to C/n-square. The Freeman-Tukey
Deviate, however, provides a method of examining the direction
and significance of the relationship between actual and expected
counts on a cell-by-cell basis. The analysis is conducted in the
following manner: ( 1 ) tabulate minimum numbers of individuals
(actual count, AC) for each taxon per level in a row and column
format (Table 1); (2) total both columns and rows; (3) calculate
the percent abundance (PA) of each taxon (row total of Table 1)
in proportion to the grand total of individuals identified; (4)

compute the expected number (EN) of individuals in each cell
assuming that the species were uniformly distributed in all levels
in proportion to their percent abundance in the entire sample;
this is done by multiplying each column total (CT) by the percent
abundance (PA) of a given taxon for the entire sample (EN =
PA x CT); (5) the Freeman-Tukey Deviate is obtained by the
following equation: F-T D = V4AC + 2 - \/4EN + 1. The
constants are added to insure a minimal value for each cell. The
resulting deviate is a measure of the significance of the difference
between the actual value and that expected. The sign notes the
direction of the difference, a minus indicates that fewer specimens
than expected were recovered, a plus indicates that more than
expected numbers were present. A measure of 0.5 is the lowest
significant deviate. The greater the deviate, the greater the dis-
parity, and hence the significance, between the actual and ex-
pected values.

SYSTEMATIC DISCUSSIONS

Introduction

Specimens used in this analysis, those from
Trenches 8, 13, and 15, are cataloged into the Pa-
leontological Repository, Museum of Natural His-
tory, University of Iowa (SUI) by agreement with
the University of Arkansas. Postcranial elements
from these trenches and all other vertebrate re-
mains, collected under NSF Grant GB6762, the bulk
of the Peccary Cave sample, is or will be deposited
at the University of Arkansas (UA) at Fayetteville.
Collections from Trench 24, excavated in 1971 and
1972 with University of Iowa funds, also are housed
in Iowa City (Megivern, 1982). All specimens of a
given taxon are cataloged by trench. Thus, all Sorex
arcticus from Trench 8 (T8) bear one number, those
from Trench 13 (T13) another, but all are reposited
by level and square in capsules with detailed pro-
venience. Measured specimens are designated by
either alphabetical or numerical subscripts.

The traditional listing of each species by taxo-
nomic hierarchy has been condensed. However, the
diagnostic characters identifying each species are
noted in the following paragraphs because some
identifications are more subjective than implied by
a taxonomic list (Table 1).

Moles

Both Scalopus aquaticus (eastern mole) and Con-
dylura cristata (star-nosed mole) are present in Pec-
cary Cave. Scalopus, characterized by robust pro-
portions, shallow re-entrant angles between the
trigonidand talonid, and two prominent lingual cusps
on the lower molars, is readily separated from the
diminutive Condylura with its deep re-entrants and

additional lobate median cusp. The Parascalops
(hairy-tailed mole) lower molar, not identified in
Peccary Cave, exhibits five lingual cusps, deep re-
entrants, and both a smaller and more delicate con-
figuration.

Shrews

Eight species of shrews are present in Peccary Cave.
Soricid mandibles greatly outnumber upper dental
elements in Peccary Cave and the key developed by
Guilday ( 1 962) was the most useful taxonomic guide.
The few rostra sufficiently complete to be identified
by Junge and Hoffmann’s (1981) criteria were as-
signed to comparable taxa. The pygmy shrew ( Sorex
hoyi, Junge and Hoffmann, 1981) was separated from
the masked shrew (S. cinereus including 5. haydeni )
by the characteristically reduced entoconid in the
former; S. arcticus (arctic shrew) exhibited both a
Sorex- type articular condyle and a post-mandibular
foramen; and S. palustris (water shrew) was distin-
guished on a deep hypoconid valley, large size, and
absence of a post-mandibular foramen. Cryptotis
parva (least shrew) displayed both a distinct con-
cavity between the articular facets of the condyle
and a reduced m3 talonid. Blarina (short-tailed
shrew) was more massive than any other Peccary
Cave shrew. A bivariate plot (Fig. 1) of the molar
row length against median body depth suggests that
there is character displacement by size in the Pec-
cary Cave soricids. Thus, most fragmentary soricid
remains also can be separated by size.

The Blarina cluster from T15-1 on the soricid
plot (Fig. 1) is almost as large as that representing
the five smaller species in the Peccary Cave sample.

1984

SEMKEN- PECCARY CAVE PALEOECOLOGY

409

Table l.— Actual (below diagonal) and expected (above diagonal) counts of rodents and insectivores from Peccary Cave Trenches 8+13

and 15. The columns become younger from right to left.

TRENCH 8 + 13 TRENCH 15

410

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 1. — Bivariate plot of median body depth versus molar row length of Peccary Cave soricid left mandibles. The species of Blarina
are designated by the linear index of Graham and Semken (1976).

Size variation within the Blarina sample also is much
greater than that for any single modern subspecies
(Graham and Semken, 1976). This implies either
that Blarina was more variable circa 12,000 years
ago or that more than one species of Blarina is pres-
ent. Statistical inspection of a previously examined
sample reveals a trimodel distribution and appli-
cation of a linear index (Graham and Semken, 1 976,
Fig. 3) separated the Peccary Cave specimens into
three groups. The largest group (Graham and
Semken, 1976, Fig. 3) was comparable in size to B.
b. brevicauda, the intermediate was similar to B. b.
kirtlandi (and B. b. hy/ophaga) and the smallest
equaled B. b. carolinensis. These phena must have
been behaving as species to be sympatric near the
cave (Graham and Semken, 1976).

Two Blarina phena previously regarded as sub-
species were demonstrated to be species by Geno-
ways and Choate (1972) who captured two “sub-
species” without morphological intergradation in the
same trap line in Nebraska. Subsequently, the sym-
patric distribution of B. brevicauda and B. hvloph-
aga was mapped across southern Iowa and northern
Missouri (Jones, 1982). Similarly, B. brevicauda and
B. carolinensis are sympatric in Georgia and South

Carolina and B. hylophaga and B. carolinensis co-
occur in Texas, Louisiana, and Arkansas (Jones,
1982). Thus, large, intermediate, and small species
of Blarina currently are recognized in eastern North
America. Although B. hylophaga is the size of B. b.
kirtlandi, the two are not the same taxon. Karyologic
data indicates that B. b. kirtlandi is a small subspe-
cies of B. brevicauda. Morphologically, however, the
B. b. kirtlandi cluster (Jones, 1982, fig. 3) shows
greater separation from that of B. brevicauda than
from either cluster of the karyologically distinct
taxa— B. hylophaga or carolinensis.

B. hylophaga presently resides in the Peccary Cave
vicinity, B. carolinensis is recorded within 75 mi
(120 km) to the east and B. brevicauda has been
collected approximately 250 mi (400 km) to the
northeast of the cave. For this reason the interme-
diate-sized phenon of Blarina from Peccary Cave,
originally regarded as B. kirtlandi by Graham and
Semken (1976), is here redesignated B. hylophaga
to be compatible with both recent biogeography and
assignment of similarity sized specimens in other
Middle Mississippi Valley local faunas (Jones, 1982).

Jones (1982) pooled suites of cranial characters
taken from modern Blarina into five categories rep-

1984

SEMKEN — PECCARY CAVE PALEO ECOLOGY

resentative of modern taxa and searched 30 Ran-
cholabrean local faunas for analogies. She did not
detect three phena of Marina in any single site, but
recognized the presence of two at such geographi-
cally diverse sites as Crankshaft Cave, Missouri;
Kline Cave, Texas; Baker Bluflf Cave, Tennessee;
New Paris #4, Pennsylvania; and Ladds, Georgia
among others. Klippel and Farm a ice (1982a) iden-
tified three phena of Marina in Stratum 5 (Wiscon-
sinan) of Cheek Bend Cave and noted that these
phena were parapatric into the Holocene of central
Tennessee. Therefore, as at Peccary Cave, multiple
Marina taxa are present in many other Wisconsinan
local faunas.

Squirrels

Seven species of squirrels are present in the Pec-
cary Cave local fauna (Table 1). Remains of the
woodchuck (Marmota monax), which weighs be-
tween 2.3 and 4.6 kg when alive, are readily sepa-
rated by size from those of the next largest identified
sciurid, the 0.45 to 1.4 kg fox squirrel ( Sciurus cf.
niger). All species of Sciurus are long-snouted and
the diastema is thus longer than that of the short-
faced red squirrel ( Tamiasciurus hudsonicus ). Sci-
urus, also is distinctly larger than Tamiasciurus, lacks
accessory cuspids present in Tamiasciurus, and has
a mandibular length clearly greater than mandibular
depth. The fox squirrel (S. niger ) and grey squirrel
(S. carolinensis) both exhibit east/west dines with
S. carolinensis increasing and S. niger decreasing in
size toward the west. They approach the same size
at their common western extreme (Purdue, 1980).
Thus, size will not necessarily separate the two on
the plains. Clinal shifts during the Holocene, dem-
onstrated by Purdue (1980), complicate size as a
taxonomic character. However, S. carolinensis usu-
ally exhibits both a tiny peg-like P3, absent in S.
niger, and a two-rooted p4. Fox squirrels generally
have three-rooted p4’s. Neither Sciurus maxillary
has a P3 alveolus and four of the six p4’s are three-
rooted. Thus, S. niger characters dominate the Sci-
urus sample, but S. carolinensis also might be pres-
ent.

The three small forest squirrels are separated by
the measurements of complete mandibles (Fig. 2).
The alveolar row lengths of the eastern chipmunk
( Tamias striatus) and the southern flying squirrel
( Glaucomys volans) are almost identical in the Pec-
cary Cave sample. For its size, however, T. striatus
has a longer diastema than G. volans. The least chip-
munk ( Eutamias minimus) has a significantly small-

41 1

er alveolar length than either T. striatus or G. volans.
None of the Peccary Cave Eutamias retain a com-
plete diastema, so a sample from Shield Trap, Mon-
tana (Geppert, 1984), was plotted for comparison.
The two complete molar rows (diastema estimated)
from Peccary Cave assigned to Eutamias (Fig. 2)
best compare to the Shield Trap sample. Reasonably
distinctive dental morphologies confirm this diag-
nosis. The six Glaucomys specimens from Peccary
Cave have an alveolar row length that ranges from
6.2 to 6.8 mm, which compares to G. volans (6.2
to 7.3 mm), and are smaller than Late Pleistocene
G. sabrinus (7.0 to 8.7 mm) in the eastern United
States (Guilday et al., 1977). The alveolar row length
of the 21 Peccary Cave Tamias (5.9 to 6.9 mm)
compares well to that of 1 7 recent and 1 1 4 late
Holocene specimens (5.8 to 6.9 mm) from Penn-
sylvania (Guilday et al., 1978). Because T. striatus
exhibits a negative Bergman’s response (Guilday et
al., 1978), the small size of the Peccary Cave Tamias
for its latitude apparently reflects a response to a
colder climate.

The thirteen-lined ground squirrel ( Spermophilus
tridecemlineatus), known from complete mandibles
and maxillae, is distinct from the other chipmunk-
sized squirrels because it has rhombohedral instead
of square molars and high anterior lophids.

Gophers

Geomys bursarius (plains pocket gopher) is a con-
spicuous component throughout Peccary Cave. It
was identified by the presence of two-grooved upper
incisors, rootless ovoid molars, equal p4 lophid
widths, and parallel anterior and posterior faces on
the p4 re-entrants.

Mice

Taxonomy of fossil Peromyscus is difficult be-
cause of the degree of inter- and intraspecific dental
variation. Hooper (1957) and Guilday and Handley
(1967) attempted to identify molars of Peromyscus
by the degree of development of both accessory styles
(-ids) and lophs (-ids). The relative abundance of
these structures is fairly constant within each species
but up to seven species (Guilday and Handley, 1 967)
share identical values. The mean percentages be-
tween other species differ by less than 5%, this is
less than the intraspecific variance. For this reason
Peromyscus, a common component of most late
Pleistocene local faunas, frequently is listed as Pero-
myscus sp. Two common and widespread modern
species, P. leucopus (white-footed mouse) and P.

412

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

E

E

7.

T T T
T T

G T T T
T T

G G G T
T T

T

T T

cn

c

<D

O

Ql.

3

2

I

0

9

8

T T T
G T

T

o

o

0)

>

<

E E
E

E E E E E E
E E
E E
EE

G

G

G G G

G/aucomys
^ volans

T T T

T T

IT T T

T T

ITT
T T
T T T

T

T am ias
striatus

E

^ £ E utamias

£ minimus

E % E E E = Shield Trap

EE E - Peccary Cave

E E
E E

G G

G G G G
0123456

3.

E

E

E

E E E
E E E

E E E E E E
78901 23

T

T T T

T T T T T

TTTTTTTTTTTT
45678901234567890 12

4.

5.

6.

Diastema Length (mm)

Peccary Cave, Small Squirrels

Fig. 2. — Bivariate plot of alveolar row versus diastema length of fossil G/aucomys, Tamias, and Eutamias from Peccary Cave. The
control sample for Eutamias was taken from Shield Trap (Geppert, 1984).

maniculatus (deer mouse) have reasonably distinct
ml’s. Fifty percent of the former and 22% of the
latter exhibit mesostylids and mesolophids. Guilday
and Handley (1967) also observed that in young
specimens, P. leucopus typically exhibited a well
developed, symmetrical anteroconid with a deep an-
terior re-entrant and that the P. maniculatus antero-
conid was structureless and characterized by a re-
duced buccal surface giving it an asymmetrical
configuration. Peccary Cave ml’s with wholly con-
sistent characters were assigned to species. Those
with either excessive wear or mixed character com-
binations were assigned to Peromyscus sp. Both the
P. leucopus and P. maniculatus groups may be ar-
bitrary. Five species of Peromyscus presently in-
habit northwestern Arkansas ( P . leucopus, P.
maniculatus, P. boylii and P. [Ochrotomys] nuttalli).
Any or all might be present in the fossil Peromyscus
sample but none of the distinctive Ochrotomys mo-

lars were found. Neither could any specimens be
attributed to Reithrodontomys.

The Onychomys leucogaster (northern grasshop-
per mouse) maxillary was identified by a long al-
veolar row (4.6 mm) and the sharp (90°) inflection
between the anterior and dorsal borders of the zy-
gomatic plate. Teeth of the rice rat ( Oryzomys pa-
lustris) were separated from other Peccary Cave cri-
cetids by a first upper and lower molar length greater
than 1.8 mm, a rudimentary accessory root on both
first molars, an anterocone (id) width nearly equal
to the mean tooth width, and low cusps aligned
opposite to (rather than diagonal to) each other. The
posterior alveolus of the first upper molar of Ory-
zomys also is distinct because it is round, half the
width of the tooth row, and offset to the labial side
of the maxillary (Satorius-Fox, 1982).

The upper and lower molars of Neotoma (wood-
rat) are readily distinguished from those of other

1984

SEMKEN- PECCARY CAVE PALEOECOLOGY

413

cricetids by their rooted, cementless, arvicoline-like
form with four elements. The Peccary Cave Neot-
oma ml’s typically exhibited the flattened lingual
triangle of the anterior loop that separates TV. flor-
idana (eastern woodrat) from other species of Neot-
oma (Hibbard and Taylor, 1960; Patton, 1963;
Semken, 1966). The Peccary Cave mbs are larger
than TV. f osagensis, the subspecies presently living
in northwestern Arkansas, and have accessory cusps
in the re-entrants of m2. These cusps are lacking in
the 40 examined TV. f osagensis but are found on
both the first and second molars of TV. ozarkensis
(Brown, 1908), an Irvingtonian woodrat from the
nearby Conard Fissure. The large size of the Peccary
Cave Neotoma may relate to “giantism” character-
istic of many late Wisconsinan species.

Voles

The muskrat (Ondatra zibet hicus) was identified
by a ml length in excess of 7 mm, rooted molars
with reticulate cement and high dentine tracts. Both
the red-backed vole ( Clethrionomys gapperi) and
the heather vole (Phenacomys intermedins) also have
rooted molars but the ml’s are less than 4 mm in
length. Phenacomys was separated from Clethrion-
omys by the absence of cement, tiny labial triangles,
and closed first and second triangles in the former.
The Clethrionomys ml has cement, equal labial and
lingual triangles, and a confluent first and second
triangle.

The “rootless” arvicolines also are common in
Peccary Cave. The bog lemming ( Synaptomys ) low-
er molars, with lingual re-entrants extending across
the tooth, could readily be divided into those with
labial triangles ( S . cooperi) and those without ( S .
borealis). This distinction can not be applied to up-
per molars because combinations of these characters
are present in both species. S. cooperi presently ex-
hibits a negative Bergman’s response and Guilday
et al. (1978) record the clinal nature of Pleistocene
samples, predicted by Hibbard (1963), in the Ap-
palachians. They also note that there is a paucity of
sites between the largest Pleistocene specimens in
Tennessee and a larger morph, described as S. aus-
tralis, in Florida and suggested that intermediate
forms might be discovered. S. australis- sized spec-
imens have been collected with S. cooperi- sized fos-
sils only in the Ladds assemblage, Georgia (Ray,
1967; Kurten and Anderson, 1980), a site where the
fauna may be mixed. The 16 Peccary Cave Syn-
aptomys ml’s (mean = 2.73, OR 2. 5-2.9 mm) are

smaller than those of S. australis (mean = 3.5, OR
3. 3-3. 9 mm) but are slightly larger than the largest
known eastern Pleistocene S. cooperi sample from
Robinson Cave (mean = 2.65, OR 2. 3-2. 9), Ten-
nessee (Guilday et al., 1978, Table 12). Thus, the
intermediate-sized Peccary Cave sample along with
those of the 1 4,800 year B.P. Waubonsie local fauna,
Iowa (mean = 3.0, OR 2. 6-3. 3 mm), recorded by
Rhodes (in press) demonstrate that S. australis is
not as distinct in size as previously thought and that
the cline extends to the west as well as to the south
(Rhodes, in press).

The distinction between Microtus (Pitymys) pi-
netorum (woodland vole) and M. (Pedomys) och-
rogaster (prairie vole) is difficult using the dental
battery and the two frequently are not differentiated
(Guilday et al., 1978). There are, however, at least
four proposed techniques for separating these taxa.
Patton (1963) devised a method utilizing the m3;
specimens with nearly closed second and third tri-
angles are Pedomys, open ones are Pitymys. This
configuration appears diagnostic most (70%) of the
time, but it is not generally applicable to the Peccary
Cave sample because of the large numbers of iso-
lated ml’s from which the minimum number of in-
dividuals was obtained. The technique does confirm
the diagnosis for mandibles with complete tooth
rows. Johnson (1972), working with both the Glen-
wood local fauna, Iowa, and a recent sample, de-
vised a technique using the width of the “isthmus”
separating the fifth triangle of the ml from the an-
terior loop; widths greater than 0.2 mm being Pe-
domys and. those smaller, Pitymys. Martin and Webb
(1974) noted that in Pitymys the sixth re-entrant
angle of the ml is deep and that the anterior border
of the fourth triangle slopes posteriorly. In Pedomys
the re-entrant is shallow and the anterior border of
the fourth triangle is at a right angle to the axis of
the tooth. Van der Meulen (1978) pointed out that
only the medial part of the wide, shallow re-entrants
characteristic of Pedomys are directed anteriorly
(morphotype 1) and that the entire narrow, deep re-
entrants of Pitymys are directed anteriorly (mor-
photype 4). He also observed that the enamel is the
same thickness on both sides of the triangle in Pi-
tymys (morphotype 4) but thin on the posterior bor-
der of Pedomys triangles (morphotype 1). All of
these techniques have merit and some are related,
for example, the deep re-entrants of Pitymys noted
by both Martin and Webb (1974) and Van der Meu-
len (1978) will produce the narrow “isthmus” of
Johnson (1972). All combinations of these charac-

414

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

ters have been observed in recent skulls of Pitymys
from North Carolina and Pedomys from Kansas and
are gradational between the two taxa; however, these
distinctions are reasonable generalizations. The Pec-
cary Cave sample was divided using the technique
of Johnson (1972) because all specimens could be
objectively assigned, knowing that some, especially
those approaching 0.2 mm in width, are incorrectly
diagnosed. Generally, the distinctive features noted
by the other three methods were consistent with the
assignment.

Microtus xanthognathus (yellow-cheeked vole)
initially was recorded from Peccary Cave by Hall-
berg et al. (1974). Their sample, based on five M3’s
over 2.6 mm in length with cementless third labial
re-entrants (Guilday and Bender, 1960), has since
been augmented by two more M3’s and three left
and two right ml’s from Trench 1 5. All of the latter
exceeded 3.55 mm in length and clearly are sepa-
rated from Peccary Cave specimens of M. pennsyl-
vanicus, which do not exceed 3.2 mm. M. xantho-
gnathus counts (Table 1) are based on M3’s.

The Peccary Cave meadow vole ( Microtus penn-
sylvanicus) sample includes many ml’s with five or
more closed triangles. This triangle count is char-
acteristic of a variety of species of Microtus (Guil-
day, 1982), but distinctive M2’s and M3’s of M.
pennsvlvanicus are present in proportion to the ml’s.
At one time, any Microtus ml collected in central
North America with five or more triangles was as-
signed to M. pennsvlvanicus. Davis (1975) first chal-
lenged this assignment in the late Pleistocene Jones
local fauna, Meade Co., Kansas. He counted 67 ml’s
with five closed triangles (M. pennsylvanicus type)
and six ml’s with three closed triangles (M. ochro-
gaster type). However, only 12 of 44 M2’s were of
the distinctive M. pennsylvanicus five element type.
The remainder were of the four element M. och-
rogaster type. Thus, the majority of four element
M2’s, Davis reasoned, were associated with five ele-
ment ml’s and had to belong to a Microtus other
than M. pennsylvanicus. This animal was dubbed
the “phantom” microtine because no specimen could
definitely be assigned to it. Subsequently, Stewart
(1978) identified the skull of M. montanus, which
has both a five element ml and a four element M2,
from the late Pleistocene Trapshoot local fauna in
central Kansas. Guilday ( 1 982) noted that the ml of
the meadow vole in the eastern United States also
is inseparable from that of the yellow-nosed vole
(M. chrotorrhinus) which also has a four element

M2. The yellow-nosed vole does have distinctive
M3’s, with two lingual cementum-filled re-entrants
on the posterior loop. Usually the meadow vole only
has one.

The number of five versus four element Microtus
M2’s in Peccary Cave are in proportion to the num-
ber of ml’s with five (M. pennsylvanicus type) and
three (M. ochrogaster type) closed triangles. Only
one M. chrotorrhinus type M3 is present (T15) in
Peccary Cave. This pattern is recorded by Guilday
(1982) in up to 5% of any modem M. pennsylvanicus
sample and therefore is an expected variant in a
large fossil M. pennsylvanicus sample. Few, if any,
of the ml’s assigned to M. pennsylvanicus could be-
long to either M. montanus or M. chrotorrhinus. It
is conceded that either M. montanus or M. chro-
torrhinus may be a rare species in the Peccary Cave
fauna, but M. pennsylvanicus must account for most
of the five closed triangle ml’s.

Five Microtus ml’s with a dental pattern char-
acteristic of the root vole (M. oeconomus) are known
from Peccary Cave. Each is characterized by four
closed triangles but the anterior loops of two are
more complicated than observed in two recent M.
oeconomus ml’s. Simple loops, present in the other
three specimens, also are found in the Peccary Cave
M. pennsylvanicus sample. Moreover, Semken
(1966) examined 398 recent M. pennsylvanicus ml’s
and found either four closed triangles or aberrant
dentitions in 1% of the specimens. Thus, five of 379
M. pennsylvanicus specimens (1.3%) with an M.
oeconomus pattern are expected in a meadow vole
sample of this size. The tundra dwelling root vole
has not been identified in other late Wisconsinan
local faunas from temperate North America. How-
ever, its presence south of Canada in Pleistocene
deposits can not be dismissed because the tundra-
dwelling collared lemming ( Dicrostonyx ) has been
identified in 1 7 sites in the contiguous United States.
Woodman (1982) also recorded both Spermophilus
parryi (arctic ground squirrel) and M. cf. miurus
(singing vole), both tundra species, as well as Di-
crostonyx in the 20,500 year B.P. Elkader local fauna
of northeastern Iowa. Thus, tundra components def-
initely were present at low latitudes and M. oecono-
mus might be expected in Wisconsinan deposits. As
the Peccary Cave four closed triangle ml sample
could represent aberrant individuals, they are cat-
aloged as M. pennsylvanicus, oeconomus pattern.
All are from Trench 15, levels 1 and 2, where the
other Microtus molars are most abundant.

1984

SEMKEN- PECCARY CAVE PALEOECOLOGY

415

PALEOECOLOGY

Moles

Biogeography

Scalopus aquaticus (eastern mole) presently is
found in the Peccary Cave region, but Condylura
cristata (star-nosed mole) ranges no closer to Pec-
cary Cave than 550 mi (880 km). Where Scalopus
and Condylura are sympatric, the star-nosed mole
occupies low, wet ground near open water and the
eastern mole is found in moist but well-drained sandy
loam (Banfield, 1974).

Biostratigraphy

The eastern and star-nosed moles comprise 1.8%
of the Peccary Cave small mammals (MNI). This
relative abundance is similar to that noted in both
Baker Bluff Cave, 1.6%, and Clark’s Cave, 0.9%,
(Guilday et al., 1977, 1978). In the Wisconsinan
Appalachian sites, for example, Baker Bluff Cave,
no one species exceeds 50% of the talpid sample.
Conversely, Holocene sites, for example, Sheep Rock
Shelter (Guilday and Parmalee, 1965) are domi-
nated (>85%) by one species. This ratio also is
reflected in recent owl pellet accumulations (Guil-
day et al., 1977). Both Scalopus and Condylura are
common in the two lowest levels, T15-1 & 2 (Table
1). Only one Condylura specimen is recorded in the
upper five levels (T8/13). The mucky ground nec-
essary for the star-nosed mole, which primarily feeds
on aquatic insects, apparently was severely reduced
above the two lowest levels in response to the tran-
sition to post-glacial climate. If mole distribution is
an index, only the two lower levels would be Pleis-
tocene. A similar stratigraphic dicotomy is found in
New Paris #4 (Guilday et al., 1964) where Condy-
lura is found below the 5.5-m level and Parascalops
breweri (hairy-tailed mole) is found above.

Shrews

Biogeography

Blarina brevicauda (northern short-tailed shrew)
today has the northernmost distribution of the three
Blarina identified in Peccary Cave. It ranges into
both the eastern deciduous and mixed conifer for-
ests and extends west into the gallery forests of the
Great Plains. Forest cover of some nature in nec-
essary for its survival. B. carolinensis (southern
short-tailed shrew) is an Austroriparian form that
is found in the southeastern portion of the United
States. Either B. hylophaga (western short-tailed

shrew [west]) or B. kirtlandi (Kirtland’s short-tailed
shrew [east]) separate the two. Thus, the interme-
diate-sized forms lie in a “temperate” zone between
the cooler adapted B. brevicauda and warmer adapt-
ed B. carolinensis. Long-tailed shrews ( Sorex ) prefer
mesic environments but most also occupy dry, open
associations (Banfield, 1974). Sorex palustris ( water
shrew), which now lives at least 700 mi (1,120 km)
north of the cave, is an exception and it requires an
aquatic environment. The immigration of S. arc-
ticus (arctic shrew) into northwestern Arkansas rep-
resents a southern dispersal of 600 mi (960 km),
that of S. hoyi (pygmy shrew) is 510 mi (820 km),
and S. cinereus (masked shrew) is 350 mi (560 km)
south of its modem range. Peccary Cave contains
the southwestern most known fossils for both S.
hoyi and S. arcticus. S. cinereus is common in the
Pleistocene of Texas and fossils of S. palustris have
been recovered from a deposit in northcentral Tex-
as. Cryptotis parva (least shrew), presently a resident
of northwestern Arkansas, has a center of distri-
bution in the southeastern United States where it
prefers open grass-covered areas, marshes, and glades
(Banfield, 1974). It is commonly associated with
boreal species in the Pleistocene of the eastern United
States and Guilday et al. (1978:27) noted that the
presence of this temperate form seems anomalous
to the boreal context indicated by its associates.
However, the northern limit of Cryptotis is within
50 mi (80 km) of and may be narrowly sympatric
with the southern limits of the boreal shrews in east-
central Wisconsin. It, along with B. carolinensis,
which is dramatically disjunct from the area of sym-
patry (Fig. 3), does indicate that rainfall was suffi-
cient for tree growth but insufficient to generate a
closed forest.

Biostratigraphy

Although all of the soricids (7% of the local fauna)
are associated on one or more of the seven tested
Peccary Cave levels, they are not randomly distrib-
uted and clearly are most abundant in two lowest
levels (Table 1; Fig. 4). This may be directly related
to the higher diversity of prey if the comparison of
the number of mammals (mean = 33) in the two
lowest levels (T15-1 & 2) to that (mean = 21) in the
five upper levels (T8/13) is an index. Cryptotis has
been identified only in the lowest two levels of the
cave (T15), Sorex cinereus is found in the lower

416

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 3.— Area of sympatry for the Trench 15, level 1 or 2 faunules. Peccary Cave local fauna, Newton Co., Arkansas. Ecoregions after
Bailey (1981).

four levels (T15, T8/13-4 & 5), and the other Sore x
are sporadically represented but present throughout
the section. Blarina brevicauda is most abundant at
the bottom of the section but is dramatically reduced
in numbers in the upper levels (Fig. 4). B. hylophaga
is common in the lower levels, diminishes, and then
becomes dominant in the middle portion of the
sampled units. B. carolinensis, also most common
in the lower portion of the section, is absent in T8/

13-5 and gradually increases upward into level two
(T8/13-2). Its upper mode lies above that of the
intermediate-sized B. hylophaga. This gradual re-
placement of one Blarina phenon by another through
time is better exhibited and statistically confirmed
in Table 2. B. brevicauda is found in substantially
greater than expected numbers low in the section,
B. hylophaga dominates in the middle but is under
represented both earlier and later, and B. caroli-

1984

SEMKEN- PECCARY CAVE PALEOECOLOGY

417

Peccary Cave Shrews
Percent Composition per level

Fig. 4. — Relative abundance of soricids (%) in each level of Trenches 8+13 and 15, Peccary Cave local fauna, Newton Co., Arkansas.

nensis has greater than expected abundance both
low and high in the column. The replacement pat-
tern of one phenon of Blarina by another demon-
strates both that the morphotypes are behaving as
species and that the soricids are adjusting individ-
ualistically as proposed by King and Graham (1981)
rather than as members of an “organismic” com-
munity.

The change in relative dominance of Blarina in
Peccary Cave from large to small species through
time suggests that temperatures generally increased
during deposition of the examined portion of the
section. The conditions could have been either rel-
atively mesic if the intermediate-sized phenon is
kirtlandi or, almost certainly, xeric if hylophaga is
present. General summer warming also is implied
by the upward reduction in all species of Sorex, each
of which now lives only to the north of Arkansas.
It may be significant that S. palustris, a semiaquatic,
is absent in the portion of the column that is dom-
inated by the xeric adapted B. hylophaga.

Squirrels

Biogeography

The thirteen-lined ground squirrel ( Spermophilus
tridecemlineatus), now restricted to the prairies of

central North America, has been recovered east of
its present range in nine Wisconsinan sites in the
Appalachian region (Semken, 1983, fig. 12-9). Suit-
able habitat also was present in the upper Missis-
sippi/middle Missouri River region because it is in
the Craigmile and Waubonsie local faunas of south-
western Iowa (Rhodes, 1982), Brayton local fauna
of west-central Iowa (Dulian, 1975), Moscow Fis-
sure of southwestern Wisconsin (Foley, in press),
Crankshaft Cave in eastern Missouri (Parmalee et
al., 1969), Bat Cave (Hawksley et al., 1973) and
Brynjulfson Caves (Parmalee and Oesch, 1972) in
central Missouri. It clearly was distributed widely
during the Wisconsinan and indicates that open
ground was more widespread in the eastern United
States during late-glacial time than at present (Guil-
day et al., 1978:55). It also occurs in the Peccary
Cave local fauna (Table 1, Fig. 5).

As in the Appalachian sites, the thirteen-lined
ground squirrel in Peccary Cave is associated with
a variety of forest squirrels including the fox squirrel
( Sciurus cf. niger), woodchuck ( Marmota monax),
red squirrel ( Tamiasciurus hudsonicus), southern
flying squirrel ( Glaucomys volans), eastern chip-
munk ( Tamias striatus), and least chipmunk {Eu-
tamias minimus). Except for the absence of the

418

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 2 Freeman- Tukey deviates for rodents and insectivores from Peccary Cave Trenches 8+13 and 15. The signs (+ or —) indicate
the direction of the actual from the expected count. Any deviate over 0.5 is increasingly significant.

TRENCH 8+13 TRENCH 15

TAXON

1

2

3

4

5

1

1

2

F-TD=V4 count+ 2 — V4expected count + 1

Scalopus aquaticus D

+ 0.2

+ 0.3

+ 0 8

+ 18

+ 1.6

-0.8

-1.3

Condyluro cristata B

-0.2

+ 0 8

-0.2

-0.2

+ 0.1

/

-0.1

+0.4

Cryptotis parva D

-0.2

-0.2

-0.2

-0.2

+0 1

+ 1.4

-0.3

Sorex cinereus B

-1.8

-1.7

-1.7

+ 0.1

+ 0.2

+ 1.6

-0.2

Sorex palustris B

+ 0.8

+0.8

-0.2

-0.2

+ 1.1

-0.8

-0.3

Sorex arclicus B

-0.9

+ 0.2

-0.8

-0.8

-0.3

+ 1.0

+ 0.3

Sorex hoyi B

-0.1

-0.1

-0.1

+ 1.1

+ 0.2

zo

+ 0.4

+ 0. 1

Blanna brevicauda D

-1.6

-2.0

-0.8

-0.8

+ 0. 1

+ 1.3

+0.6

Biarina hylophaga D

+ 0.8

+ 1.0

+3 8

+3 0

+0.4

1

-1.4

-4.6

Blarina carolinensis D

+ 0.3

+ 1.0

-0.3

-0.3

-0.5

\

+ 1.3

-1.8

Tamias striatus D

+ 2 4

+ 1.2

+ 1.2

-1.2

-0.5

\

+0.1

-2 8

Eutamias minimus B

+ 0.2

+ 0.2

+0.2

+0.2

+0.2

+0.4

+ 0.6

Marmota monax D

+ 2.3

+ 1.4

+ 0.8

+ 0.8

+0.5

-2.3

-1 5

S. tridecemlineatus S

-0.4

+0 6

-0.4

-0.4

-0.1

-0.4

+ 1.3

Sciurus cf. niger D

+ 1.5

+ 2.0

+ 1.1

+ 16

+ 0.5

-2.0

-2.2

Tomiasciurus hudsonicus B

-0.4

-0.3

-0.3

-0.3

-0.1

+0.5

+ II

Glaucomys volans D

+ 1.3

+ 0 7

-0.3

+ 0.7

-0.1

-0.6

-0.1

Geomys bursarius S

-1.2

+ 0.2

-0.6

+0.2

+ 0.4

+ 1.1

-0.6

Castor canadensis

+ 1.3

+ 0.2

+ 0.2

+ 0.2

+0.2

+0.4

-0.4

Erethizon dorsatum B

+ 1.3

+ 0.7

-0.3

-0.3

+ 1.1

-0.6

-0.1

Peromyscus cf leucopus D

+ 1.6

+ 2 4

+ 2.4

+ 0.3

+0.8

-2.4

-1.9

Peromyscus cf. maniculatus

+ 1 1

+ 0.2

+ 1.7

+ 1.7

+ 1.6

- 1.1

-2.4

Peromyscus sp.

+ 4 8

+ 3.6

+ 3.2

+ 0.8

+ 0.5

-0.4

-9.8

Onychomys leucogaster s

+ 0.2

+ 1.3

+ 0.2

+ 0.2

+ 0.2

+ 0.2

+ 0.2

Neotoma floridana D

+ 0.4

+ 1.0

-0.4

- 1. i

+ 1.0

+ 0.6

-0.9

Oryzomys palustris D

+ 1.0

+0.1

-0.7

+ 1.1

+ 0.2

-0.5

- 1.0

Ondatra zibethicus

-0.2

-0.2

+ 0 8

+ 0.8

-0.1

z

-0.1

-0.3

Microtus pinetorum D

+ 1.1

-0.2

+ 0.7

-0.2

-1.7

/

+ 2.1

-3.1

Microtus ochrogaster S

-1.5

+ 0.2

+ 0.1

+ 0.5

-0.3

+ 0.4

-0.1

Microtus pennsylvanicus B

-6 7

-5.6

-5.3

-1.6

- 1.3

\

V

+ 0.2

+ 6.0

Microtus xanthognathus B

-0.1

+ 0.1

-0.1

+ 0.1

+0.2

- 0.3

+ 1.3

Synaptomys cooperi B

+ 0.1

+ 0 1

+ 0 8

+ 2.8

+ 18

-1.8

-1.5

Synaptomys borealis B

+ 0.6

-0.4

+ 0.6

-0.4

+ 1.0

/

+ 0.1

+ 0.3

Phenacomys intermedius B

-0.9

-0.8

-0.8

-0.8

-0.3

(

-0.5

+ 2.4

Clethrionomys gapperi B

-4.3

-5.7

-4.1

-3.0

- 1.8

)

-1.2

+6.2

Zapus hudsonlus B

+0 . 8

+0.8

-0.2

-0.2

-0. 1

1

-0.8

+ 0.4

1984

SEM KEN -PECCARY CAVE PALEOECOLOGY

419

U L MNI

Peccary Cave Squirrels
Percent Composition per level

Fig. 5. — Relative abundance of sciurids (%) in each level of Trenches 8+13 and 15, Peccary Cave local fauna, Newton Co., Arkansas.

northern flying squirrel ( Glaucomys sabrinus ), the
species composition of the Peccary Cave sciurids
(Table 1) is identical to that of the Appalachian local
faunas (Guilday et al., 1978).

Peccary Cave records of both the least chipmunk
{Eutamias minimus) and the red squirrel ( Tamias -
ciurus hudsonicus) extend the Wisconsinan range of
these animals into northwestern Arkansas, over 550
mi (880 km) south of their present southern non-
montane limit in Wisconsin. Eutamias occupies a
variety of habitats but boreal parkland is preferred
and closed forests are avoided. Tamiasciurus also
is characteristic of boreal parkland but it will occupy
both closed conifer and mixed conifer/hardwood
forests. The other squirrels in Peccary Cave are den-
izens of the eastern deciduous forest but they do
range into the southern mixed conifer/hardwood
forest. All of the Peccary Cave sciurids coexist today
in the central Wisconsin area of sympatry (Fig. 3).

Biostratigraphy

Overall, squirrels comprise 4.5% of the Peccary
Cave local fauna, a misleading figure because there
is a substantial change in relative abundance up-
section (Fig. 5). Squirrels comprised 2.8% of the
local fauna in the lowest level (T15-2) and progres-
sively increased through 3.0, 3.5, 5.6, 6.3, 10.6, to

12.5% in the uppermost level (T8/13-1). Excluding
Spermophilus, the percentages gradually increase
from 1.9% in T15-2 to 12.5% in T8/13-1. This in-
dicates increasing forest density during deposition.
The significance of this progression is shown by the
Freeman-Tukey deviates (Table 2). The eastern
chipmunk ( Tamias striatus), fox squirrel ( Sciurus
cf. niger), woodchuck {Mar mot a monax), and
southern flying squirrel {Glaucomys vo/ans), all with
centers of distribution in the eastern deciduous for-
est are under represented at the bottom of the sec-
tion (T15-2) and over represented at the top (T8/
13-1). Conversely, the thirteen-lined ground squir-
rel {Spermophilus tridecemlineatus), a prairie eco-
type, is present in much greater than expected num-
bers in the lowest level (T15-2) and is under
represented above. The only forest squirrels that are
present in greater than expected numbers low in the
section are the red squirrel Tamiasciurus hudsoni-
cus) and least chipmunk {Eutamias minimus ), both
of which center in the boreal forest. Thus, the sci-
urids suggest that a boreal parkland, with a decid-
uous forest component, was in the vicinity of Pec-
cary Cave during deposition of T 15-2. The steppe
aspect of this diminished by T15-1 and conifers
became much less important by the beginning of
T8/13-5 time. Mature deciduous forest elements
were progressively more important during each de-

420

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

positional phase and resembled modern climax for-
ests at the top sampled interval. The reduction in
relative abundance (Fig. 5) of the fox squirrel, which
prefers open forest, in level T8/13-1 probably re-
flects closure of the deciduous forest.

Gophers

Biogeographv

The plains pocket gopher (Geomys bursarius )
presently is found in sandy soils as close as 40 mi
(64 km) to the southwest of Peccary Cave. It is ex-
cluded from the Ozark Highlands because of both
shallow soil and dense forest growth. The gopher,
which feeds on the roots, stems and tubers overlying
its burrow, is restricted to regions sufficiently open
for lush surface growth. It is widely distributed in
prairies but also is common in parkland situations.
The discovery of Geomys as far east as Welsh Cave,
central Kentucky (Guilday et al., 1971), and Cheek
Bend Cave in central Tennessee (Parmalee and
Klippel, 1981a), where a closed canopy forest now
exists, indicates that open ground extended much
further east during late Wisconsinan time than it
does at present. The presence of the closely related
but “isolated” population of Geomys pinetis (south-
eastern pocket gopher) in southern Georgia, Ala-
bama, and northern Florida prompted Guilday et
al. ( 1 97 1 ) to suggest that the pocket gopher and hence
open ground was ubiquitous in the southeastern
United States during the late Pleistocene. The oc-
currence of the thirteen-lined ground squirrel (Sper-
mophilus tridecemlineatus ) in late glacial deposits
of Pennsylvania, Virginia, Kentucky, Tennessee, and
Florida (Semken, 1983) supports this conclusion.
This eastern and southern range extension of Geo-
mys suggests that less precipitation than that of to-
day characterized the late Pleistocene of the south-
eastern and south-central United States. This does
not imply either aridity or extensive prairies. The
number of tree squirrels and boreal species from
each of these local faunas preclude this possibility.
Precipitation, while reduced, must have been more
effective. A widely distributed parkland is compat-
ible with each local fauna. The Peccary Cave local
fauna differs from those in Iowa and Wisconsin in
that Thomomys talpoides (northern plains pocket
gopher), which also extended its range well to the
east during the Wisconsinan (Rhodes, in press; Fol-
ey, in press; Rosenberg, 1983) is not present.

Biostratigraphy

Geomys bursarius is the fourth most common
species (6% of the local fauna) and is present in all

levels in Peccary Cave (Table 1). Initially it is under
represented in the section (T15-2) but the situation
is reversed in T 15-1 (Table 2). It is present in near
expected numbers in all of the upper levels but is
rare in T8/13-3 and especially T8/13-1. Its low
abundance in the upper levels is attributed to forest
closure. This is supported by increased numbers of
forest species, especially sciurids, in this part of the
section. However, the predominance of grassland
species in T15-2 precludes its scarcity in this level
being attributed to a closed forest. Level T1 5-2 also
is the most “boreal” of all and probably is best
described in terms of a cool steppe, a condition not
attractive to Geomys.

Mice

Biogeography

The new world rats and mice, represented by four
genera and 584 individuals (30% of the total fauna),
comprise the second largest family identified in Pec-
cary Cave. Collectively, the white-footed mice
( Peromyscus ) and woodrats ( Neotoma ) accounted
for 580 of these individuals. The rice rat (Oryzomys)
and grasshopper mouse ( Onychomys ), known from
three and one individuals respectively, comprise the
remainder.

When Peromyscus leucopus (white-footed mouse)
and P. maniculatus (deer mouse) are sympatric, the
former occupies woodlands and the latter more open
brushy areas. The deer mouse will reside in any dry-
land habitat while the white-footed mouse will select
more mesic situations. Both feed on seeds, nuts, and
insects and indicate deciduous vegetation. The
northern grasshopper mouse ( Onychomys leucogas-
ter), relatively rare in its modern range, is a carni-
vore that is restricted to the Great Plains. Where
short-tailed shrews are absent, grasshopper mice be-
come the major small carnivore. The single speci-
men of Onychomys, which occurs within 110 mi
(176 km) of Peccary Cave, indicates a more open
environment in Arkansas at that time.

Neotoma floridana (eastern woodrat) is the sec-
ond most common mammal, after the meadow vole
(Microtus pennsylvanicus) in Peccary Cave. The
eastern woodrat presently is common in the Ozarks
where bluffs, caves, and fissures, their preferred hab-
itat, are well developed. They also frequent wooded
terrain and build nests against fallen logs, in hollow
trees, or in burrows (Sealander, 1979). At the present
time the eastern woodrat lives no further north than
central Missouri in the Mississippi Valley but fossils
have been recovered from post-altithermal cave fills
(Eshelman, 1971) and deposits near bluffs (Fay,

1984

SEMKEN- PECCARY CAVE PALEOECOLOGY

421

1980) in Iowa. These “outliers” of Neotoma have
disappeared and suggest that factors other than de-
glaciation are responsible for its present distribu-
tion. Western species of the genus range north to
the Yukon and the eastern woodrat is found as far
north as both South Dakota and New York. There-
fore, the restricted northern limit of this southeast-
ern form in the Mississippi valley at present prob-
ably is related to something other than cold winters.

Meadow jumping mice ( Zapus hudsonius ) now
range approximately 100 mi (160 km) to the north
of Peccary Cave. These mice, which feed on seeds,
insects, and fruits, prefer meadow and edge zone
habitats for nests and feeding.

The rice rat (Oryzomys palustris) also is rare in
Peccary Cave, but it occurs sporadically throughout
the column (Table 1). This suggests that this semi-
aquatic, deciduous forest biome inhabitant previ-
ously, as now, was only locally present in north-
western Arkansas (Sealander, 1979). It indicates the
presence of wet meadows with dense vegetation bor-
dering lentic environments at the time of deposition.

Biostratigraphy

The relative abundance of Peromyscus changes
substantially in the section (Table 2). Collectively,
Peromyscus , a browser, is strongly under repre-
sented in the lowest levels (T15-1 & 2) and over
represented in the upper levels (T8/ 13-all) of the
cave. Peromyscus, along with the sciurids (Fig. 5,
Table 2), becomes dominant upsection at the ex-
pense of the arvicolines. Both P. cf. leucopus (white-
footed mouse), which prefers dry deciduous wood-
lands and P. cf. maniculatus (deer mouse), which
characteristically inhabits more open brushy areas
are rare in T15. P. cf. maniculatus then becomes
abundant in the lower levels of T8/13. While it is
still common in the upper levels of T8/13, P. cf.
leucopus replaces it as the most common of the two.
It is difficult to interpret the pattern precisely be-
cause both of these mice occupy a variety of habi-
tats, but a transition from open terrain in T15-2
through a scrub forest (T8/13-5, 4, & 3) to a dry,
closed deciduous forest (T8/13-1 & 2) is apparent.
While the overall complexion of the T8/13-2 fau-
nule clearly is that of a forest association, it is sig-
nificant that the grasshopper mouse ( Onychomys
leucogaster), thirteen-lined ground squirrel ( Sper -
mophilus tridecemlineatus ), and prairie pocket go-
pher ( Geomys bursarius), all prairie ecotypes, ex-
hibit increases in abundance in this level. This may
be related to altithermal conditions recorded else-
where in the Ozarks (Semken, 1983).

Neotoma reflects at least 10.5% of the fauna in
any level and may represent almost 20% of the in-
dividuals (T8/13-5). Its actual, with respect to ex-
pected, numbers (Table 2) fluctuate greatly and show
no apparent pattern. This suggests that local rather
than regional influences primarily affected the pop-
ulation. The presence of the rice rat ( Oryzomys pa-
lustris) in Peccary Cave is compatible with the
meadow habitat suggested by the numbers of voles
but its southeastern distribution is incongruous with
the boreal center of distribution of most of the as-
sociated voles and shrews. The occurrence of the
rice rat demonstrates, along with Blarina carolinen-
sis, B. hylophaga, and Neotoma floridana, that tem-
perate species did not vacate Arkansas when the
boreal fauna immigrated. These four species have
two areas of sympatry (Fig. 3) on the prairie-forest
border of the southern plains. Both are strictly dis-
harmonious with the common sympatry for the
squirrels, voles, and insectivores, less Blarina, (Fig.
3) but substantiate a parkland environment for the
late Wisconsinan around the cave. Both Peromyscus
leucopus and P. maniculatus are found in the cri-
cetid sympatries. Reithrodontomys (harvest mouse),
a common component of Holocene local faunas, is
not identified in the Peccary Cave sample. It appears
to have been rare during the Pleistocene east of the
Great Plains.

Voles

Biogeographv

The most abundant mammal in the Peccary Cave
local fauna (Table 1) is the meadow vole ( Microtus
pennsylvanicus). It is recorded as both more nu-
merous and widespread in eastern and central
North America (Zakrzewski, in press) during the
Pleistocene when it ranged south to central Texas
(Lundelius, 1967), Louisiana, and Florida (Martin,
1968). This animal, which avoids closed forests
(Banfield, 1974), indicates that expanses of suitable
grass cover (for example, meadows, mesic prairies,
or forest edge situations) were more widespread than
at present. Martin (1968) predicts that both 5°F
cooler July temperatures and increased moisture to-
ward the southwest were necessary for the meadow
vole to colonize the Gulf coastal plain. The meadow
vole digs shallow burrows in the summer but resides
in spherical grass nests constructed under snow in
winter. It is active all winter, as are all of the Peccary
Cave voles, and consumes cached food or green
grass bases growing under the snow (Banfield, 1974).
Its modem southern distribution roughly parallels
isograds on snow cover maps (Visher, 1954) and it

422

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

may reflect increased Wisconsinan snow cover in
Arkansas.

Hall (1981) mapped the northern bog lemming
( Synaptomys borealis ), which lives in the boreal for-
est, and the southern bog lemming ( S . cooperi), a
mixed forest form, as being nearly allopatric today.
Although they now occur together only in portions
of southeastern Manitoba, eastern Quebec, and
northern New Brunswick (Banfield, 1974), the two
were widely sympatric and co-occurred as far south
as Peccary Cave, Arkansas, and Baker Bluff Cave,
Tennessee, during the Pleistocene. S', cooperi- like
forms ( S . australis) ranged south into both Florida
and Texas (Lundelius, 1967) during late Wisconsi-
nan time. During late glacial time in the northeast-
ern United States, the northern bog lemming was
both common and more abundant than the southern
bog lemming (Guilday et al., 1978). This ratio is
reversed to the south in both Peccary and Baker
Bluff caves and S. borealis was not present further
south in the Cheek Bend local fauna, Tennessee
(Klippel and Parmalee, 1982 b), or in any central
Texas locality (Lundelius, 1 967). Both species ranged
hundreds of kilometers further south during the
Wisconsinan but neither abandoned their present
unglaciated northern range. The isolating mecha-
nisms geographically separating the two at present
were either reorganized or different during the Pleis-
tocene. Both animals, as the name implies, were
associated with bogs and meadows of the Arkansas
Pleistocene.

The heather vole ( Phenacomys intermedius), now
restricted to the boreal forest of Canada in eastern
North America, was nearly ubiquitous east of the
Rocky Mountains during the late Wisconsinan.
Guilday and Parmalee (1972) recorded 16 extra-
limital Pleistocene localities with Phenacomys, one
of which (Peccary Cave) was approximately 800 mi
(1,280 km) south of its present limits. Since then
specimens have been recovered in Cheek Bend Cave,
Tennessee, which also is 800 mi to the south, by
Parmalee and Klippel ( 1 9826) and 600 mi (960 km)
south of its modern range in northcentral Nebraska
(Voorhies, 1981). This distribution was generally
harmonious with the Pleistocene range of the north-
ern bog lemming. Dry, open forest with a coniferous
component and an understory of heaths (dwarf birch,
willow, and blueberry) is the prime habitat of the
heather vole. This environment must have been
more widespread during late glacial time in order
for this browsing arvicoline to colonize now tem-
perate regions of the eastern United States.

The red-backed vole, Clethrionomys gapperi, rep-

resented by 196 individuals, is the third most com-
mon species in Peccary Cave (10% of the entire
fauna, 23% of all voles), as it is in most other late
Wisconsinan local faunas. Clethrionomys comprises
17.6% of all voles in Baker Bluff Cave, 10% in Car-
rier Quarry Cave, and 14% in Robinson Cave (Guil-
day et al., 1978), all of which are in eastern Ten-
nessee at approximately the same latitude as Peccary
Cave. Klippel and Parmalee (19826) record the red-
backed vole as representing 1 2% of all voles in Pleis-
tocene portions of the Cheek Bend local fauna. It
also is in the Pleistocene of Texas (Lundelius, 1967)
but is not recorded by Webb (1974) in the Pleisto-
cene of Florida. Clethrionomys, which now is re-
stricted to the uppermost reaches of the Mississippi
River in central North America, is a common vole
in the mixed conifer/deciduous forest, especially
where the floor is littered with logs and stumps. It
also frequents areas with a brushy understory in
forest edge situations, and aspen bluff/shrubby cou-
lees on the prairie (Banfield, 1974). It feeds pri-
marily on the petioles of broad leafed forbs and
shrubs in summer and on heaths which remain green
under the snow in winter. Unlike Phenacomys,
Clethrionomys will graze to supplement its diet.

Microtus xanthognathus, the yellow-cheeked vole,
presently is restricted to the northern boreal forest
and bordering tundra west of Hudson Bay. It pri-
marily is sylvan but does construct runways from
its burrow to sphagnum bogs (Banfield, 1974). This
taxon, which feeds on horsetails and lichens, is widely
distributed in the late Wisconsinan of the United
States. It is known from Peccary Cave, Arkansas
(Hallberg et al., 1974), and Baker Bluff Cave, Ten-
nessee (Guilday et al., 1978). Both sites are at least
1,500 mi (2,400 km) south of its modern range. It
is common in sites north of these two localities in
both the upper Mississippi Valley (Rhodes, in press;
Foley, in press; Woodman, 1982; Rosenberg, 1983)
as well as three Appalachian sites reviewed by Guil-
day et al. (1978). The presence of this vole in Ar-
kansas reinforces the interpretation of boreal forest
in the late Pleistocene of northwestern Arkansas.

The prairie vole, Microtus ochrogaster, does not
inhabit the immediate vicinity of Peccary Cave to-
day but it has been collected 50 mi (80 km) to the
northwest. M. ochrogaster is restricted to dry grass-
lands where there is both sufficient grass to cover
its runways and soil for burrows (Banfield, 1974).
It will locate in shrubby terrain but never enters
wooded regions. Arkansas populations of the prairie
vole presently are restricted to prairie enclaves, for
example, Grand Prairie, but it is predicted to have

1984

SEMKEN — PECCARY CAVE PALEOECOLOGY

423

“2 C

?4.4

6.3

J

0.9

0.5

3.7

2.3

0.6

0.3

1.4

MNI

23

19

27

43

0.6 355

3.0 u0.8 370

Peccary Cave Arvicolines
Percent Composition per level

Fig. 6. — Relative abundance of arvicolines (%) in each level ofTrenches 8+13 and 1 5, Peccary Cave local fauna, Newton Co., Arkansas.

been more widely distributed in Arkansas in the past
(Sealander, 1 979). The prairie vole’s declining abun-
dance upsection in Peccary Cave (Table 2; Fig. 6)
and its absence in the vicinity today suggest that
Sealander (1979) is correct.

The woodland vole, Microtus pinetorum, is the
only arvicoline found in the Peccary Cave local fau-
na that still inhabits the region. As the name implies,
these voles live in woodlands, especially dry, mature
deciduous forests, where they construct burrows in
litter on the forest floor. They feed (from the un-
derside) on bulbs and tubers as well as green leaves,
nuts, and fruits which fall to the ground. These semi-
fossoria! forms rarely enter fields and never do so
when prairie voles are present. They are found
throughout the Peccary' Cave section, represent ap-
proximately 16% of all voles, and clearly were not
forced out of the region by late Wisconsinan cli-
mates. Their continued presence demonstrates that
dry deciduous woodland survived in the area when
the boreal elements immigrated into northwestern
Arkansas during the Wisconsinan. The woodland
vole also is common in all of the known late Pleis-
tocene faunas from the Appalachians.

Biostratigraphy

Arvicoline rodents, represented by nine species
and 852 individuals, comprise 44% of the local fau-
na and are the most abundant animals in Peccary

Cave. This relative abundance, which is in con-
cordance with that at New Paris #4 (41%), Clark’s
Cave (47%) and Baker Bluff Cave (47%) in the Ap-
palachians (Guilday et a!., 1978), but greater than
that at Brynjulfson #1 (28%) in the Ozarks (Par-
malee and Oesch, 1972). Almost half of the Peccary
arvicolines, 379 individuals, are meadow voles ( Mi-
crotus pennsyivanicus) which now range south only
to northern Missouri, approximately 300 mi (480
km) to the north. It is followed, in decreasing order
of abundance (Table 1, Fig. 6), by the red-backed
vole ( Clethrionomys gapperi), woodland vole (M.
pinetorum), prairie vole (M. ochrogaster), southern
bog lemming ( Synaptomys cooperi), heather vole
(. Phenacomys intermedins), northern bog lemming
(S. borealis), muskrat ( Ondatra zibet hicus), and yel-
low-cheeked vole ( M . xanthognathus).

The meadow vole ( Microtus pennsyivanicus)
comprises 50% of the T15-2 voles (Fig. 6). It pro-
gressively declines to 1 3% in relative abundance in
the uppermost level (T8/13-1). The red-backed vole,
Clethrionomys gapperi, initially representing 31%
of the voles, exhibits a corresponding decline up-
section. Collectively, the patterns (Table 2; Fig. 6)
of these two species indicate a progressive dimin-
uation in importance of both boreal meadow and
boreal forest associations. The dominance of boreal
habitats in the lowest levels (T 1 5 - 1 & 2) is supported
by both the heather vole, Phenacomys intermedins,

424

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

and the yellow-cheeked vole, M. xanthognathus,
which are sympatric today only in northern Canada
(Fig. 3). These two species have not been identified
above the lowest two levels, are less abundant in
T15-1 than in T15-2, and suggest that the environ-
ment transgressed from an open, dry mixed forest
(Canadian) parkland with an understory of heaths
to a parkland like that near the confluence of the
boreal, deciduous forest, and prairie biomes (Fig.
3). Reduction in relative abundance or loss of these
species above T15-1 indicates that the mixed park-
land gradually became more temperate (deciduous)
in aspect through deposition of T8/13-3.

The decline in abundance or loss of the four above
species was accompanied by an immediate but short-
lived (T8/13-5) increase in both Synaptotnvs bo-
realis and 5. cooperi. S. borealis, now confined to
the boreal forest, was rare but continued to survive
in the cave area later than the other strictly boreal

species. This probably was because of locally fa-
vorable conditions. Its southern counterpart, S.
cooperi, now characteristic of the Canadian (Lau-
rentian) mixed forest, continued as a major but pro-
gressively less abundant element upsection.

These boreal voles were replaced in a near recip-
rocal pattern by voles of the temperate prairie, Mi-
crotus ochrogaster, and deciduous forest, M. pine-
torum. Thus, there was a gradual replacement of
boreal meadow by prairie species among the grazing
voles and a transition from predominately conifer
(for example, M. xanthognathus ), through mixed
(for example, Synaptomvs cooperi), to predomi-
nately deciduous forest (for example, M. pinetorum)
species among the browsing voles in the parkland.
The closed deciduous forest of the Ozark Uplands
today apparently was not completely established un-
til after 2,290 years B.P. when major deposition
ceased in the cave.

SUMMARY AND CONCLUSIONS

Paleoecology

Peccary Cave microvertebrates are not randomly
distributed throughout the section. Arvicolines con-
stitute 64.2% of the total faunule in the lowest level
(T 1 5-2) but decline to 16.9% in the uppermost level
(T8/13-1). Squirrels, primarily deciduous forest
species, conversely expand from 2.8% in T15-2 to
12.5% in T8/13-1. The distribution of micromam-
mals in the Peccary Cave sequence indicates both
that Wisconsinan climates were not monolithic
chronologically and that faunal turnover at the end
of the Pleistocene in northwestern Arkansas was not
catastrophic, but typified by both individualistic and
progressive responses. First, faunal composition
changed from one dominated by individuals char-
acteristic of cool steppe with dominantly coniferous
mixed forest inliers to one of a mixed forest park-
land. This was followed in order by: (2) extirpation
of four dry boreal forest species, (3) an increase in
both meadow and forest species to balanced rep-
resentation between deciduous and boreal conifer-
ous ecotypes in the parkland, (4) loss of two more
boreal species and dominance of deciduous park-
land ecotypes, (5) mega vertebrate extinction and a
dramatic increase in both deciduous forest and tem-
perate prairie elements, (6) development of the east-
ern deciduous forest with local enclaves of prairie
and boreal conifer stands, and finally (7) gradual

closure of the deciduous forest to a density ap-
proaching modern conditions. A completely mod-
em faunal analogue is not present in the section and
must post-date 2,290 years B.P.

Eight taxa are over-represented significantly in the
lowest level (T15-2). These are Blarina brevicauda,
Eutamias minimus, Spermophilus tridecemlinea-
tus, Tamiasciurus hudsonicus, Microtus Pennsyl-
vaniais, M. xanthognathus, Phenacomys interme-
dius, and Clethrionomys gapperi. All, except B.
brevicauda are boreal and, of these, only T. hud-
sonicus will occupy closed forests. Each is typical of
either open parkland or meadow/prairie associa-
tions. Fourteen taxa are clearly under-represented
in T 15-2: Scalopus aquaticus, B. hylophaga, B. car-
olinensis, Tamias striatus, Marmota monax, Sci-
urus cf. niger, Geomys bursarius, Peromyscus cf.
leucopus, P. cf. maniculatus, P. sp., Neotoma flor-
idana, Oryzomys palustris, M. pinetorum, and Syn-
aptomys cooperi. With the exception of G. bursarius
(steppe) and S. cooperi (boreal but in mixed forest)
all have distributions centered in the deciduous for-
est. Thus, the T15-2 faunule represents a cool, gen-
erally dry steppe, but one unlike that recorded by
Rhodes (in press) in southwestern Iowa. The mixed
forest aspect of the parkland was better developed
in Arkansas than in Iowa. High species diversity,
including a minor but diverse deciduous forest com-

1984

SEMKEN- PECCARY CAVE PALEOECOLOGY

425

munity within the T15-2 faunule and the dishar-
monious sympatry demonstrate that this parkland
is not analogous to any existing boreal ecosystem.
The flora probably resembled that presently occur-
ring in the tension zone (Curtis, 1959) of central
Wisconsin (Fig. 3), the area of sympatry, but grass-
land rather than forest dominated the landscape.
Boreal conifers, understory heaths, and moss were
more important than deciduous elements. Precipi-
tation was low enough to permit extensive devel-
opment of an open parkland but was sufficiently
effective to maintain meadows ( Synaptomys ),
marshes ( Sorex palustris), and ponds ( Ondatra zi-
bethicus) as well as forest groves (Sciurus and Tami-
asciurus) in favorable topographic locations. Sum-
mer temperatures must have averaged at least 1 2°F
cooler than those in Arkansas today (Fig. 3) to per-
mit immigration of the boreal component. Winter
temperatures may have been similar to those pres-
ently in Central Wisconsin or slightly warmer. They
were not significantly colder than those in Wisconsin
now because many of the temperate deciduous for-
est species living in Arkansas today (for example,
Neotoma and Blarina carolinensis) also resided in
the area during the Wisconsinan.

The most striking change in relative abundance
of species (Table 2) occurred between levels T15-2
and 115-1. Eleven species reversed their position
as either under (— ) or over (+) represented by at
least 1.5 Freeman-Tukey units (Table 2). Cryptotis
parva, Sorex cinereus, Blarina carolinensis , Sper-
rnophilus tridecemlineatus, Geomys bursarius,
Neotoma floridana, and Microtus pinetorum, all of
which now range in temperate regions were present
in greater than expected numbers in T 1 5- 1 . M. xan-
thognathus, Phenacomys intermedins, Clethriono-
mys gapped, and Zapus hudsonius, all boreal species,
became substantially reduced. At the same time 10
species moved at least 0.5 Freeman-Tukey units
toward equivalence between their actual and ex-
pected abundance ( Seal opus aquaticus, B. hylopha-
ga, Tamias striatus, Glaucomys volans, Peromyscus
cf. maniculatus, P. sp., Oryzomys palustris, Tami-
asciurus hudsonicus, Erethizon dorsatum, and M.
pennsylvanicus). The direction of faunal turnover is
clear; deciduous forest forms either changed from
under- to over-representation or became less “de-
ficient” in the sample; boreal and grassland eco-
types, except Geomys bursarius, were of reduced
importance. The T15-1 faunule suggests that the
parkland became more heavily wooded, brush oc-
curred with heaths in the understory, and deciduous

trees occupied a greater proportion of the landscape
than previously (T15-2). This implies warmer sum-
mers, more effective precipitation (first Oryzomys )
and less severe winters than during T1 5-2 time. The
T15-1 faunule also has the highest species diversity
(35 species) of any level and the most equal balance
between boreal (43%), steppe (9%), and deciduous
forest (37%) species. It represents an “equable” fau-
na as described in the Introduction and suggests that
equable Wisconsinan climates apparently are more
characteristic of late Wisconsinan time.

There possibly is a hiatus between T 1 5- 1 and T8/
13-5, but a veneer of Unit B deposits at the top of
T1 5-1 and a less dramatic change in relative abun-
dance than between the lower two levels suggest that
the time lag is minor. T8/13-5 is characterized by
the loss of four boreal species (. Eutamias minimus,
Tamiasciurus hudsonicus, Microtus xanthognathus,
and Phenacomys intermedins). With the exception
of both species of Synaptomys, Erethizon dorsatum
and Sorex palustris, all other boreal species are found
in reduced numbers. Because Erethizon survived in
Arkansas until historic time, its boreal classification
is an artifact of history and it should not be consid-
ered a “boreal” indicator. The other three taxa are
associated with wetlands and their increase is in
accord with predicted moisture increases shown by
loss of the four dry conifer parkland indicators. De-
ciduous forest species increase dramatically in T8/
3 3-5 with substantial expansion in proportions of
Scalopus aquaticus, Marmota monax, Sciurus cf.
niger, Peromyscus cf. leucopus, P. cf. maniculatus,
and Neotoma floridana. Only M. pinetorum and
Glaucomys volans of the deciduous forest group de-
crease in relative numbers. This may relate to re-
structuring of the forest from predominantly conif-
erous to deciduous and increased distance between
surviving deciduous trees. The P. cf. maniculatus
peak in T8/1 3-5 suggests an increase of brush in the
understory. This is supported by reduction of three
taxa— two prairie species, Geomys bursarius and
Spermophilus tridecemlineatus and the grazing Mi-
crotus pennsylvanicus, which tolerates cool, tem-
perate grasslands. Wetlands also occupied a greater
area and may have restricted available habitat for
grazing rodents. The climate of T8/ 13-5 time un-
doubtedly had still warmer summers and increas-
ingly effective rainfall compared to that of the un-
derlying T15-1. Winters probably were warmer.

The number of individuals associated with boreal
environments continued to decrease in T8/13-4.
Only Sorex hoyi and Synaptomys cooperi of this

426

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

group increased in relative abundance but the for-
mer along with S. cinereus last appeared in this
level. Deciduous forest forms (for example, Glau-
comys and Sciurus ) continued to become more
abundant but boreal conifers, indicated by Cleth-
rionornys, still were locally present. The increase in
both the prairie vole (M. ochrogaster) and woodland
vole (M. pinetorum) denotes overall temperate con-
ditions. Continued increases in summer tempera-
ture and precipitation permitted the development
of a continuous deciduous forest in the area; conifers
became restricted to small enclaves.

Because the A/B disconformity bisects level T8/
13-3, it undoubtedly contains a mixture of speci-
mens from both stratigraphic units. Nevertheless
the last extinct megavertebrates are associated with
this level and are most likely Unit B specimens.
There is no evidence of any major climatic change
in the T8/ 1 3-3 micromammal assemblage other than
a sharp reduction in the relative abundance of Mi-
crotus pennsylvanicus.

Level T8/13-2, dated 4,290 years B.P. is com-
pletely within Unit A. The faunule basically rep-
resents that of a temperate deciduous forest. Both
Microtus ochrogaster and M. pinetorum, temperate
species, become more abundant than M. pennsyl-
vanicus for the first time. T8/1 3-2 is unique because
Onychomys leucogaster, a steppe indicator, is con-
fined to this level. The two other steppe species
recorded in Peccary Cave, Spermophilus tridecem-
lineatus, and Geomys bursarius , also exhibit a mode
in this level. These animals are not sufficiently nu-
merous to suggest major prairie encroachment but
evidence for either balds or upland prairie enclaves
is better developed here than in the other T8/13
levels. Because T8/13-2 immediately post-dates the
Altithermal, the prairie indicators may represent a
remnant of more widespread prairie present during
the Unit A/B hiatus. Evidence of prairie encroach-
ment to the east during the Altithermal also is sug-
gested by fossil mammals recovered from both the
Cherokee Sewer site in Iowa (Semken, 1980) and
Rodgers Shelter in Missouri (Wood and McMillan,
1976).

The T8/13-1 faunule has the most modem aspect
of the seven examined faunules and represents a
nearly closed deciduous forest. Boreal indicators still
are present but the seven species only account for
8% of the individuals in the faunule. They undoubt-
edly were relictual but their presence in the post-
Altithermal of Arkansas suggests that the Ozarks
were a refugium for boreal species late into the Ho-

locene. Individuals of the two steppe species, Mi-
crotus ochrogaster and Geomys bursarius, both pres-
ently extirpated, comprise 6% of the fauna. The
remainder are either widespread species, primarily
Peromyscus, or deciduous forest taxa. The wood-
land vole ( M. pinetorum ), eastern chipmunk (Tami-
as striatus), woodchuck ( Marmota monax), western
short-tailed shrew ( Blarina hylophaga) and fox
squirrel ( Sciurus niger) are most common. The
modem deciduous forest of the area, which does not
host either steppe or boreal species, developed
sometime after 2,290 year B.P.

The Pleistocene/Holocene Boundary in
Peccary Cave

If the traditional biostratigraphic concept of me-
gavertebrate extinction is used, the Pleistocene/Ho-
locene boundary in the Peccary Cave sequence, is
best placed at the disconformity separating Units A
and B (T8/13-3). Extinct megavertebrates are com-
mon in both Units B and C but are rare in the
overlying Unit A. However, Davis ( 1973) notes that
all large mammal remains are sparse in Unit A and
believes that the paucity is partly attributable to
closure of the horizontal entrance by alluviation.
Four post-5,000 year B.P. radiocarbon dates, one
(1-3894) from T8/13-2, support middle to late Ho-
locene age for Unit A. If these dates from Unit A
are correct and megavertebrate extinction is chron-
ologically precise. Units B and C are Pleistocene and
Unit A is Holocene. The Unit A/B disconformity
then represents an early to mid-Holocene hiatus.
The sparse remains of extinct megavertebrates in
Unit A are reworked if this division is correct. Quinn
(1972:95-96), however, believes that some extinct
species became “entombed at the very end of the
Altithermal” in Peccary Cave and that the Ozarks
provided a post- 1 0,000 year B.P. refugium for them.

Micromammal relative abundance shows no ob-
vious discontinuities (Figs. 4-6) separating the
Pleistocene from the Holocene. Biostratigraphic
boundaries usually are based on ( 1 ) last appearances,
(2) first appearances and/or (3) coincident changes
in relative abundance between units. Taken in order:
(1) Eight of 33 species recorded in the lowest Peccary
Cave level (T15-2) are not known from the upper-
most level (T8/13-1). Five of these ( Microtus xan-
thognathus, Phenacomys intermedius, Eutamias
minimus, Tamiasciurus hudsonicus, and Cryptotis
parva ) disappear in the interval between levels T1 5-
1 and T8/13-5. All, except C. parva, are boreal and
their loss is expected. The interim between T8/13-

1984

SEMKEN- PECCARY CAVE PALEOECOLOGY

427

4 and 3 is a second datum marked by the last ap-
pearance of Sorex hoyi and S. cinereus, also boreal
species. Ondatra zibethicus is not recorded above
T8/13-3 and finally, two boreal species, S. arcticus
and Condylwa cristata (Table 1) disappear above
T8/13-2. Major loss of the boreal components oc-
curs between T 1 5- 1 and T8/ 1 3-5, but extirpation is
a continuing process throughout the section. (2) Only
four species in the Peccary Cave section are not
present in T15-2. Oryzomys palustris and Castor
canadensis are first recorded in T15-1. Onchomys
leucogaster appears in T8/13-2 (Table 1) and No-
tiosorex crawfordi (desert shrew), which occupies
northwestern Arkansas today, is recorded in other,
presumably younger. Unit A sediments of other
trenches. Thus, appearance of new taxa is limited
and indicates that the modem fauna of Arkansas is
largely an impoverished residue of a more diverse
Wisconsinan community. (3) Ten species (Figs. 4-
6) achieve maximum relative abundance in T15-2,
two in T15-1, four in T8/13-5, three in T8/13-4,
two in both T8/ 1 3-3 and 2 and one in T8/ 13-1. The
differences in relative abundance are much greater
between T15-1 and 2 than between any other two
levels. The difference across the Unit A/B discon-
formity is much less than expected for a 5,000 year
hiatus separating Pleistocene and Holocene depos-
its. Faunal turnover in the Peccary Cave section
(Figs. 4-6, Table 2) is highly individualistic and
generally reflects a progressive transition between
late Wisconsinan (T 1 5-2) and post-Altithermal (T8/
13-1) time. This confirms gradual environmental
change, following an abrupt change between T15-2
and 1 , interpreted above for the Peccary Cave sec-
tion.

To examine the boundary question from another
perspective, the faunal counts (Table 1) for sciurids
(Fig. 7) and arvicolines (Fig. 8) were normalized (to
130 individuals per level to minimize sample size
differences) and replotted to reflect species density
(percent distribution throughout the entire fauna
rather than an individual level). While these graphs
appear similar to the relative abundance per level
plots (Figs. 4-6), they weigh the distributions in-
dependently and graphically reflect the Freeman-
Tukey deviates (Table 2). The distribution of Mi-
crotus pinetorum is illustrative. This taxon gradually
comprises a greater proportion of each faunule up-
section (Fig. 6) but its absolute abundance (density)
is fairly stable (Fig. 8) through time. The arvicolines
as a family were almost decimated (Fig. 8) after
deposition of T 15-1. However, the proportions of

each species changed gradually (Fig. 6), both across
the break and within the radically thinned ranks of
voles (Fig. 8) in the upper five levels. This suggests
that the specific habitat of each vole, except for the
extirpated yellow-cheeked and heather voles, re-
mained in the Peccary Cave vicinity but occupied
substantially less area. The animals in these reduced
niches then responded in a patterned manner (T8/

1 3) to the factor(s) which created the most abrupt
biostratigraphic change, presumably the new cli-
matic regime discussed in the previous section. The
sharpest inflection for the sciurids appears between
T15-1 and T8/13-5 immediately after, and perhaps
in response to, the major reduction in the number
of voles. A gradual increase in deciduous forest
squirrels upsection is almost reciprocal to the de-
mise of the arvicolines.

The position of the Pleistocene/Holocene bound-
ary in the Peccary Cave section is difficult to estab-
lish, perhaps because of its transitional nature. Based
on the loss of megamammals and radiocarbon dates,
the most logical boundary is in the disconformity
that bisects T8/ 13-3. This position can be supported
faunally by the last appearance of both Sorex ci-
nereus and S. hoyi, an abrupt decline in the abun-
dance of Microtus pennsylvanicus, an increase in
numbers of Tamias striatus, the first appearance of
Onychomys leucogaster, and an increase in numbers
of two other steppe species ( M . ochrogaster and
Spermophilus tridecemlineatus). The majority of last
appearances, however, are concentrated in the in-
terval between T 1 5- 1 and T8/ 1 3-5 where five species,
including four boreal forms disappear. T8/1 3-5 also
is marked by a sharp increase in the importance of
sciurids. Although the major change in density from
dominance of voles (boreal) to sciurids (deciduous)
occurs between T15-1 and T8/13-5, major faunal
reorganization occurred earlier, between T15-2 and
T15-1. The greatest change in relative abundance
of species, confirmed by dramatic differences in sign
of the Freeman-Tukey deviates, is between these
two levels and, as noted in the paleoecology section,
the only “reversal” in paleoecological trend is re-
corded here. The first appearance of two species
{Oryzomys palustris and Castor canadensis ) also oc-
curs at this time.

Thus, in Peccary Cave, community reorganiza-
tion was the first event to mark the Pleistocene/
Holocene transition, maximum replacement of voles
by sciurids followed, and finally the Wisconsinan
megavertebrates disappeared. Modern closed decid-
uous forest conditions post-date T8/ 13-1. The exact

428

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 8

Peccary Cave Squirrels
Percent Total Composition (Normalized)

U L

0.1

0.2

Peccary Cave Arvicolines
Percent Total Composition (Normalized)

Fig. 7 (above).— Sciurid density (%), normalized to 130 MNI, in Trenches 8+13 and 15, Peccary Cave local fauna, Newton Co.,
Arkansas.

Fig. 8 (below).— Arvicoline density (%), normalized to 130 MNI, in Trenches 8+13 and 15, Peccary Cave local fauna, Newton Co.,
Arkansas.

position of the boundary then becomes a matter of
definition. Climatic change probably triggered the
first event. This initiated a series of individualistic
responses marked by failure of most boreal small
mammals to reproduce in the area, their replace-

ment by deciduous forest species and ultimately in
mega vertebrate extinction. The small mammals
continued to adjust well into late Holocene (post-
4,000 years B.P.) time. This scenario supports pro-
gressive but ultimately radical climatic change as

1984

SEMKEN- PECCARY CAVE PALEOECOLOGY

429

the prime motivator for Wisconsinan megaverte-
brate extinction as proposed by Graham and Lun-
delius (in press). The Peccary Cave chronofauna,
along with that recorded by Guilday et al. (1978) at
Baker Bluff Cave, Graham (1976) at Friesenhahn
Cave, Klippel and Parmalee ( 1 982 a) at Cheek Bend,

and Guilday et al. (1964) at New Paris, also dem-
onstrates that Wisconsinan climates were not mon-
olithic but characterized temporally as well as geo-
graphically, by biologic, ecologic, and climatic
variability.

ACKNOWLEDGMENTS

The author wishes to thank the late James H. Quinn, Univer-
sity of Arkansas, for bringing the Peccary Cave local fauna to his
attention; Jack McCutcheon, property owner, both for the dis-
covery and recognition of the significance of the fossils; and to
L. Carson Davis, Southern Arkansas University who, along with
McCutcheon, excavated, waterscreened, and processed the ver-
tebrate remains.

For professional assistance 1 thank: Michael D. Carleton, Uni-
versity of Michigan; L. Carson Davis, Southern Arkansas Uni-
versity; Russell W. Graham, Illinois State Museum; the late John
E. Guilday, CMNH; Robert S. Hoffmann, University of Kansas;

Emmet T. Hooper, University of Michigan; Ernest L. Lundelius,
University of Texas; Larry D. Martin, University of Kansas;
Richard S. Rhodes II, University of Iowa; Gerald R. Smith,
University of Michigan; and Michael R. Voorhies, University of
Nebraska. Financial assistance for excavation was provided by
NSF Grant GB6762. The manuscript was prepared on the word
processor by Rebecca R. Bush; figures were drafted by Joyce E.
Chrisinger.

1 want to especially thank Richard S. Rhodes II for his diligent
work on the manuscript and both he and L. Carson Davis for
their continued support of this project.

LITERATURE CITED

Bailey, R. G. 1981. Ecoregions of North America (map). USD1
Fish and Wildlife Service, Washington, D.C.

Banfield, A. W. F. 1974. The mammals of Canada. Univ.
Toronto Press, Toronto, xxv + 438 pp.

Blair, W. F. 1958. Distributional patterns of vertebrates in the
southern United States in relation to past and present en-
vironments. Pp. 433-568, in Zoogeography (C. L. Hubbs,
ed.), American Assoc. Adv. Sci., Publ, 51.

Braun, E. L. 1955. Phytogeography of unglaciated eastern
United States and its interpretation. Botanical Rev., 21:297-
375.

Brown, E. L. 1908. The Conard Fissure, a Pleistocene bone
deposit in northern Arkansas: with a description of two new
genera and twenty new species of mammals. Mem. Amer.
Mus. Nat. Hist., 9:157-208.

Curtis, J. T. 1959. The vegetation of Wisconsin. Univ. Wis-
consin Press, Madison, 657 pp.

Dalquest, W. W. 1965. New Pleistocene formation and local
fauna from Hardeman County, Texas. J. Paleontol., 39:63-
79.

Dalquest, W. W., E. Roth, and F. Judd. 1969. The mammal
fauna from Schulze Cave, Edwards County, Texas. Bull.
Honda St. Mus., 13:206-276.

Davis, L. C. 1969. The biostratigraphy of Peccary Cave, New-
ton County, Arkansas. Proc. Arkansas Acad. Sci., 23:1 92—
196.

. 1973. The herpetofauna of Peccary Cave, Newton

County, Arkansas. Unpublished Masters thesis, Univ. Ar-
kansas, Fayetteville, 85 pp.

. 1975. Late Pleistocene geology and paleoecology of the

Spring Valley Basin, Meade County, Kansas. Unpublished
Ph.D. dissert., Univ. Iowa, Iowa City, 170 pp.

Dulian, J. J. 1975. Paleoecology of the Brayton local biota,
late Wisconsinan of southwestern Iowa. Unpublished Mas-
ters thesis, Univ. Iowa, Iowa City, 50 pp.

Eshelman, R. E. 1971. The paleoecology of Willard Cave, Del-
aware County, Iowa. Unpublished Masters thesis, Univ. Iowa,
Iowa City, 72 pp.

Fay, L. P. 1980. Mammals of the Garrett Farm and Pleasant
Ridge local biotas (Holocene), Mills County, Iowa. Pp. 74-
75, in Amer. Quat. Assoc. Sixth Biennial Meeting Abs.,
Orono, Maine.

Foley, R. L. In press. Late Pleistocene (Woodfordian) verte-
brates from the Driftless Area of southwestern Wisconsin,
the Moscow local fauna. Illinois State Mus. Repts. Invest..
39.

Genoways, H. H., and J. R. Choate. 1972. A multivariate
analysis of systematic relationships among populations of
the short-tailed shrew (genus Blarina ) in Nebraska. Syst.
Zool., 21:106-1 16.

Geppert, T. J. 1984. Small mammals of Shield Trap, East Pryor
Mountains, Montana. Unpublished Masters thesis, Univ.
Iowa, Iowa City, 45 pp.

Graham, R. W. 1976. Late Wisconsinan mammalian faunas
and environmental gradients of the eastern United States.
Paleobiology, 2:343-350.

. 1979. Paleoclimates and late Pleistocene faunal prov-
inces in North America. Pp. 49-69, in Pre-Llano Cultures
of the Americas: problems and paradoxes (R. L. Humphrey
and D. Stanford, eds.), Washington Anthro. Soc., Washing-
ton, D.C., 1 50 pp.

Graham, R. W., and E. L. Lundelius, Jr. In press. Coevolu-
tionary disequilibrium and Pleistocene extinction. In Qua-
ternary extinctions (P. S. Martin and R. G. Klein, eds.), Univ.
Arizona Press, Tucson.

Graham, R. W., and H. A. Semken, Jr. 1976. Paleoecological
significance of the short-tailed shrew (Blarina), with a sys-
tematic discussion of Blarina ozarkensis. J. Mamm., 57:
443-449.

Guilday, J. E. 1962. The Pleistocene local fauna of the Natural

430

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Chimneys, Augusta County, Virginia. Ann. Carnegie Mus.,
36:87-122.

. 1982. Dental variation in Microtus xanthognathus, M.

chrotorrhinus, and M. pennsylvanicus (Rodentia: Mamma-
lia). Ann. Carnegie Mus., 51:211-230.

Guilday, J. E., and M. S. Bender. 1960. Late Pleistocene
records of the yellow-cheeked vole, Microtus xanthognathus
(Leach). Ann. Carnegie Mus., 35:315-330.

Guilday, J. E., H. W. Hamilton, E. Anderson, and P. W.
Parmalee. 1978. The Baker Bluff Cave deposit, Tennessee,
and the late Pleistocene faunal gradient. Bull. Carnegie Mus.
Nat. Hist., 1 1:1-67.

Guilday, J. E., H. W. Hamilton, and A. D. McCrady. 1971.
The Welsh Cave peccaries ( Platygonus ) and associated fau-
na, Kentucky Pleistocene. Ann. Carnegie Mus., 43:249-320.

Guilday, J. E., and C. O. Handley, Jr. 1967. A new Pero-
myscus (Rodentia: Cricetidae) from the Pleistocene of Mary-
land. Ann. Carnegie Mus., 39:91-103.

Guilday, J. E., P. S. Martin, and A. D. McCrady. 1964. New
Paris No. 4: A Pleistocene cave deposit in Bedford County,
Pennsylvania. Bull. Nat. Speleo. Soc., 26:121-194.

Guilday, J. E., and P. W. Parmalee. 1965. Animal remains
from the Sheep Rock Shelter (36 Hu 1), Huntingdon County,
Pennsylvania. Pennsylvania Arch., 35:34-49.

. 1972. Quaternary periglacial records of voles of the

genus Phenacomys Merriam (Cricetidae: Rodentia). Qua-
ternary Res., 2:170-175.

Guilday, J. E., P. W. Parmalee, and H. W. Hamilton. 1977.
The Clark’s Cave bone deposit and the late Pleistocene pa-
leoecology of the central Appalachian Mountains of Virginia.
Bull. Carnegie Mus. Nat. Hist., 2:1-87.

Hall, E. R. 1981. The mammals of North America. John Wiley
and Sons, New York, l:xv + 1-600, + 90\ 2:vi + 601-1 181
+ 90.

Hallberg, G. R., H. A. Semken, Jr., and L. C. Davis. 1974.
Quaternary records of Microtus xanthognathus (Leach), the
yellow-cheeked vole, from northwestern Arkansas and
southwestern Iowa. J. Mamrn., 55:640-645.

Hawksley, O., J. F. Reynolds, and R. L. Foley. 1973. Pleis-
tocene vertebrate fauna of Bat Cave, Pulaski County, Mis-
souri. Bull. Nat. Speleo. Soc., 35:61-87.

Hibbard, C. W. 1949. Techniques of collecting microvertebrate
fossils. Contrib. Mus. Paleont., Univ. Michigan, 8:7-19.

. 1952. Vertebrate fossils from Late Cenozoic deposits

of Central Kansas. Univ. Kansas, Paleont. Contrib., Verte-
brata, 2:1-14.

. 1955. The Jinglebob interglacial (Sangamon?) fauna

from Kansas and its climatic significance. Contrib. Mus.
Paleont., Univ. Michigan, 12:179-228.

. 1960. An interpretation of Pliocene and Pleistocene

climates in North America. Ann. Rept. Michigan Acad. Sci.,
Arts and Letters, 62:5-30.

. 1963. A late Illinoian fauna from Kansas and its cli-
matic significance. Papers Michigan Acad. Sci., Arts, Letters,
48:181-221.

Hibbard, C. W., and D. W. Taylor. 1960. Two late Pleistocene
faunas from southwestern Kansas. Contrib. Mus. Paleont.,
Univ. Michigan, 16:1-223.

Hooper, E. T. 1957. Dental patterns in mice of the genus Pero-
myscus. Misc. Publ. Mus. Zook, Univ. Michigan, 99:1-59.

Johnson, P. C. 1972. Mammal remains associated with Ne-
braska phase earth lodges in Mills County, Iowa. Unpub-
lished Masters thesis, Univ. Iowa, Iowa City, 71 pp.

Jones, C. A. 1982. Paleontology and systematic relationships
of short-tailed shrews (Genus Blarina). Unpublished Mas-
ters thesis, Fort Hays State Univ., Hays, Kansas, 210 pp.

Junge, J. A., and R. S. Hoffmann. 1981. An annotated key to
the long-tailed shrews (Genus Sorex) of the United States
and Canada, with notes on middle American Sorex. Occas.
Papers Mus. Nat. Hist., Univ. Kansas, 94:1-48.

King, F. B., and R. W. Graham. 1981. Effects of ecological
and paleoecological patterns on subsistence and paleoenvi-
ronmental reconstructions. Amer. Antiq., 46:128-142.

Kitts, D. B. 1953. A Pleistocene musk ox from New York and
the distribution of the musk-oxen. Amer. Mus. Novitates,
1607:1-8.

Klippel, W. E., and P. W. Parmalee. 1982a. Diachronic vari-
ation in insectivores from Cheek Bend Cave and environ-
mental change in the Midsouth. Paleobiology, 8:447-458.

. 19826. The paleontology of Cheek Bend Cave, Maury

County, Tennessee. Phase II report to the Tennessee Valley
Authority, Knoxville.

Kurten, B., and E. Anderson. 1980. Pleistocene mammals of
North America. Columbia Univ. Press, New York, 442 pp.

Lundelius, E. L. 1967. Late-Pleistocene and Holocene faunal
history of central Texas. Pp. 287-319, in Pleistocene ex-
tinctions (P. S. Martin and H. E. Wright, Jr., eds.), Yale
Univ. Press, New Haven, 453 pp.

Martin, L. D., and A. M. Neuner. 1978. The end of the
Pleistocene in North America. Trans. Nebraska Acad. Sci.,
6:1 17-126.

Martin, R. A. 1968. Late Pleistocene distribution of Microtus
pennsylvanicus. J. Mamm., 49:265-271.

Martin, R. A., and S. D. Webb. 1974. Late Pleistocene mam-
mals from the Devil’s Den Fauna, Levy County. Pp. 1 14-
145, in Pleistocene mammals of Florida (S. David Webb,
ed.), Univ. Florida Press, Gainesville, x + 270 pp.

McKenna, M. C. 1962. Collecting small fossils by washing and
screening. Curator, 3:221-235.

Megivern, K. J. 1982. Paleoclimatic significance of the Pleis-
tocene Insectivora and Rodentia of Trench 24, Peccary Cave,
Newton County, Arkansas. Unpublished Masters thesis,
Univ. Iowa, Iowa City, 56 pp.

Miller, B. B. 1970. The Sandahl molluscan fauna (Illinoian)
from McPherson County, Kansas. Ohio J. Sci., 70:39-50.

Parmalee, P. W., and W. E. Klippel. 1981a. A late Pleistocene
population of the pocket gopher, Geomys cf. bursarius, in
the Nashville Basin, Tennessee. J. Mamm., 62:831-835.

. 19816. A late Pleistocene record of the heather vole

(Phenacomys intermedius) in the Nashville Basin, Tennes-
see. J. Tennessee Acad. Sci., 56:127-129.

Parmalee, P. W., and R. D. Oesch. 1972. Pleistocene and
Recent faunas from the Brynjulfson Caves, Missouri. Illinois
State Museum, Rept. Investigation, 25:52 pp.

Parmalee, P. W„ R. D. Oesch, and J. E. Guilday. 1 969. Pleis-
tocene and Recent vertebrate faunas from Crankshaft Cave,
Missouri. Illinois State Museum, Rept. Investigation, 14:1—
37.

Patton, T. H. 1963. Fossil vertebrates from Miller’s Cave,
Llano County, Texas. Bull. Texas Memorial Mus., 7:1-41.

Purdue, J. R. 1980. Clinal variation of some mammals during
the Holocene in Missouri. Quaternary Res., 13:242-258.

Quinn, J. H. 1972. Extinct mammals in Arkansas and related
C14 dates circa 3,000 years ago. 24th International Geol.
Cong., 12:89-96.

1984

SEMKEN- PECCARY CAVE PALEOECOLOGY

431

Ray, C. E. 1967. Pleistocene mammals from Ladds, Bartow
County, Georgia. Bull. Georgia Acad. Sci., 25:120-150.

Rhodes, R. S., II. In press. Paleoecology and regional paleo-
climatic implications of the Farmdalion Craigmile and
Woodfordian Waubonsie mammalian local faunas, south-
western Iowa. Illinois State Mus. Repts. Invest., 40.

Rosenberg, R. S. 1983. The paleoecology of the late Wiscon-
sinan Eagle Point local fauna, Clinton County, Iowa. Un-
published Masters thesis, Univ. Iowa, Iowa City, 70 pp.

Satorius-Fox, M. R. 1982. Paleoecological analysis of micro-
mammals from the Schmidt Site, a Central Plains Tradition
village in Howard County, Nebraska. Univ. Nebraska Div.
Archaeological Res. Technical Rpt., 83-13, 87 pp.

Schultz, G. E. 1967. Four superimposed Late-Pleistocene ver-
tebrate faunas from southwest Kansas. Pp. 321-336, in Pleis-
tocene extinctions (P. S. Martin and H. E. Wright, Jr., eds.),
Yale Univ. Press, New Haven, 453 pp.

Sealander, J. A. 1979. A guide to the Arkansas mammals.
River Road Press, Conway, Arkansas, 313 pp.

Semkjen, H. A., Jr. 1966. Stratigraphy and paleontology of the
McPherson Equus beds (Sandahl local fauna), McPherson
Co., Kansas. Contrib. Mus. Paleont., Univ. Michigan, 20:
121-178.

. 1969. Paleontological implications of micromammals

from Peccary Cave, Newton County, Arkansas. Geol. Soc.
America, South Central Sec. Abs., p. 27.

. 1974. Micromammal distribution and migration dur-
ing the Holocene. Amer. Quaternary Assoc. Abs., Madison,
3:25.

. 1980. Holocene climatic reconstructions derived from

the three micromammal bearing cultural horizons of the
Cherokee Sewer Site, northwestern Iowa. Pp. 67-99, in The
Cherokee excavations (D. C. Anderson and H. A. Semken,
Jr., eds.). Academic Press, New York, 277 pp.

. 1983. The Holocene mammalian record of the eastern

and central United States. Pp. 182-207, in Late-Quaternary
environments of the United States: The Holocene (H. E.
Wright, Jr., ed.), Univ. Minnesota Press, Minneapolis, 277

pp.

Slaughter, B. H. 1965. Animal ranges as a clue to late Pleis-
tocene extinction. Pp. 155-167, in Pleistocene extinctions
(P. S. Martin and H. E. Wright, Jr., eds.), Yale Univ. Press,
New Haven, 453 pp.

Stephens, J. J. 1960. Stratigraphy and paleontology of a late
Pleistocene basin. Harper County, Oklahoma. Bull. Geol.
Soc. Amer., 71:1675-1702.

Stewart, J. D. 1978. Mammals of the trapshoot local fauna.
Late Pleistocene of Rooks County, Kansas. Proc. Nebraska
Acad. Sci., Abs. for 88th Annual Meeting.

Van der Meulen, A. J. 1978. Microtus and Pitymys (Arvi-
colidae) from Cumberland Cave, Maryland, with a com-
parison of some new and old world species. Ann. Carnegie
Mus., 47:101-145.

Visher, S. S. 1954. Climatic atlas of the United States. Harvard
Univ. Press, Cambridge, 403 pp.

Voorhies, M. R. 1981. A fossil record of the porcupine (Er-
ethizon dorsatum) from the Great Plains. J. Mamm., 62:
835-837.

Webb, S. D. (ed.). 1974. Pleistocene mammals of Florida. Univ.
Florida Press, Gainesville, x + 270 pp.

Wood, W. R., and R. B. McMillan. 1976. Prehistoric man
and his environments. Academic Press, New York, 271 pp.

Woodman, N. 1982. A subarctic fauna from the late Wiscon-
sinan Elkader site, Clayton County, Iowa. Unpublished Mas-
ters thesis, Univ. Iowa, Iowa City, 56 pp.

Zakrzewskj, R. J. In press. The fossil record of Microtus in
North America. In The Biology of New World Microtus (R.
A. Tamann, ed.). Spec. Publ. Amer. Soc. Mamm.

Address: Department of Geology, University of Iowa, Iowa City, Iowa 52242.

LATE PLEISTOCENE AND EARLY RECENT MAMMALS
FROM FOWLKES CAVE, SOUTHERN
CULBERSON COUNTY, TEXAS

Walter W. Dalquest and Frederick B. Stangl, Jr.

ABSTRACT

Stratified deposits from Fowlkes Cave in the Chihuahuan Des-
ert of Trans-Pecos Texas yielded abundant fossil remains of late
Pleistocene vertebrates from a 20 cm thick layer sandwiched
between seemingly barren layers. The mammalian fossils include
both species adapted to boreal habitat (Sorex vagrans, S. palus-
tris, Eutamias cinericollis, Microtus mexicanus ) and forms typ-
ical of arid desert habitat (Dipodomys merriami, D. spectabilis,
Peromyscus eremicus, Onychomys leucogaster, O. torridus), all
from a single stratum that must have accumulated in a relatively
short time. Similar associations have been reported from caves
in southern New Mexico but were thought to have resulted from

postmortum mixing of materials. We suggest that, because the
Guadalupe Mountains, 100 km to the north, are known to have
been glaciated, the improbable association found at Fowlkes Cave
resulted from glaciation. Meltwater streams permitted the exis-
tence of cool meadows and thickets while, back from the oasis-
like streamsides, desert conditions existed and desert mammals
lived. The now barren desert ranges, the Delaware and Apache
mountains, formed an avenue by which some alpine mammals
emigrated from the Guadalupe Mountains to the Davis Moun-
tains where they are now isolated.

INTRODUCTION

Culberson Co. (Fig. 1) is large (3,787 sq mi), and
lies in extreme western, or Trans-Pecos Texas. Most
of the county'consists of desert flats, scrub, grass-
land, and rugged hills and mesas, typical of the Chi-
huahuan Desert. In the northwestern corner of the
county the Guadalupe Mountains, a southward ex-
tension of the southern Rocky Mountains, rise to
8,700 ft elevation and support a montane flora and
fauna. The Recent mammals of the Guadalupe
Mountains have interested mammalogists since
Bailey ( 1 905) collected specimens there in 1901. For
more recent accounts see Genoways et al. (1979).
The Recent mammals of the desert lowlands of Cul-
berson Co. are less well-known but were treated by
Davis and Robertson ( 1 944) and by Sehmidly ( 1977).
Numerous other accounts deal, at least in part, with
the mammals of Culberson Co. and are cited in the
bibliographies of the four papers mentioned.

The geology of Culberson Co. is complicated but
much of the exposed surface rock consists of Per-
mian limestones in which caves are relatively com-
mon. Similar cave-forming limestones occur in ad-
jacent New Mexico and extend nearly across the
entire southern border of that state. Carlsbad Cav-
ern, for example, lies in this Permian limestone se-
ries. For a summary of some 25 early Recent and
late Pleistocene mammalian faunas from caves in
Trans-Pecos Texas and southern New Mexico (see
Harris, 1974).

The cave faunas reported to date, from the Gua-
dalupe Mountains and southern New Mexico, occur
in a band approximately 1 00 km from north to south
and 400 km from east to west, and many lie at
relatively high elevations, where modern conditions
are more mesic than those typical of the Chihuahuan
Desert.

In 1979, Mr. J. M. Fowlkes of Pecos, Texas, told
us of a sinkhole-type cave on his large ranch (70
sections, 44,800 acres) north of Kent, Texas. This
cave lies 1 00 km south of the caves in the Guadalupe
Mountains and southern New Mexico, where Pleis-
tocene and early Recent mammalian faunas have
been described. It seemed possible that fossils from
Fowlkes Cave might contribute important infor-
mation concerning an area south of the previously
known caves, at lower elevation, and in what is now
arid, typical Chihuahuan Desert. With the cooper-
ation of Mr. Fowlkes, we were able to obtain a quan-
tity of the black, silty surficial deposits in the cave,
and of the upper part of the underlying late Pleis-
tocene sediments. Both samples yielded abundant
bones of mammals, birds, and reptiles. The fossils
of lower vertebrates from the Pleistocene deposit
will be reported elsewhere by other workers. Only
the early Recent and Pleistocene mammals are dealt
with here.

432

1984

DALQUEST AND STANGL- FOWLKES CAVE LOCAL FAUNA

433

Fig. 1. — Map of Culberson Co., Texas, and major physiographic features of the region.

434

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

DESCRIPTION OF FOWLKES CAVE AND ITS SURROUNDINGS

Fowlkes Cave lies approximately 10 km north of Kent (a gas-
oline station and general store plus a few buildings north of
Interstate Highway 10). The exact location of the cave is not
detailed at the request of the landowner. It lies near the edge of
the southern termination of the barren, limestone hills termed
the Apache Mountains, which are a southern extension of the
Delaware Mountains. The Delaware Mountains, in turn, consti-
tute a southern extension of the Guadalupe Mountains. Just to
the south are the Davis Mountains, where montane conditions
exist and where the flora and, in part, the fauna, is similar to that
of the Guadalupe Mountains. Zoogeographically, Fowlkes Cave
is especially important in that its location is in the desert lowlands
between the Guadalupe Mountains to the north and Davis Moun-
tains to the south, and almost on the flanks of the high ridge
between these mountain ranges. Montane mammals that occur
in the Davis Mountains today presumably reached the fir-oak-
pine woodlands of the Davis Mountains from the Guadalupe
Mountains in the Pleistocene, when ecological conditions be-
tween the two mountain ranges were suitable for the existence
of these species (for speculation see Schmidly, 1977:17-18). The
chain of hills (Delaware Mountains-Apache Mountains) would
be a logical pathway between the two montane highlands, and
Fowlkes Cave might preserve a sample of the mammalian em-
igrants. Study to date, reported here, suggests that this is true.

Fowlkes Cave is a sinkhole (Fig. 2) whose entrance is located
on the slopes of a limestone hill (Fig. 3). From the cave mouth
to the aluvial fan below measures approximately 90 m. The slope,

from cave mouth to aluvial fan, is approximately 45 degrees.
The slopes of the entire hill is so steep that only grasses and a
few desert shrubs such as cacti and juniper bushes, with roots
anchored in crevices in the limestone, manage to survive on the
hill. Below the hill the aluvial fans support a rich flora of yuccas,
cacti, and other desert shrubs. The aluvial fans grade into creosote
bush flats (Fig. 4). Several kilometers away are dry washes and
fringing vegetation that might, in a cooler climatic cycle, have
supported riparian marshes and permanent running water.

The cave entrance measures approximately 2 m2. The opening
drops 3 m to a shallow step-like shelf. Beneath this is a sheer
drop of 6 m to a platform 1 m wide, followed by another sheer
drop of 5 m to the debris-filled upper chamber. From this point
one can scramble 2 m down the slopes of a debris cone to the
lower chamber. Both upper and lower chambers are on the same
axis, the upper extending northwest and the lower southeast. The
upper chamber is filled to a considerable but unknown depth by
dust, trash, cave-swallow droppings and nests, and limestone
fragments that have spalled from the cave walls. Its dimensions,
from a point below the entrance shaft, are approximately 22 by
12.5 m. A small pit in the back leads to other chambers and
passages that were explored in part but do not play a part in this
work. The lower chamber measures approximately 25.5 by 8 m
from a point below the entrance shaft, but part of the chamber
is nearly filled to the level of the upper chamber by debris that
collected beneath the entrance shaft. Stalactites, stalagmites, par-
tial cave pearls, cave coral, and other intricate shapes and struc-

Fig. 2. — Sinkhole entrance to Fowlkes Cave.

1984

DALQUEST AND STANGL- FOWLKES CAVE LOCAL FAUNA

435

Fig. 3.— Sinkhole entrance (arrow) to Fowlkes Cave, as seen from alluvial fan. Distance from foreground figure to cave is approximately

90 m.

tures of carbonate line the walls and ceilings. Although bats have
occupied the cave in the past, only cave swallows seem to have
lived there in late years.

Mr. Fowlkes stated that in past years the cave was mined for
bat guano, and evidence of this work, in the form of wooden
planks and scrap metal, litter the slope near the cave, while ropes,
cables, and ladders were found in the cave itself. The mining,
however, seems to have been very limited and confined to digging

in the upper chamber. Probably the product was swallow guano
rather than bat guano. Partly for this reason, we excavated only
in the lower chamber. The sediments here were undisturbed, save
for one hole approximately a 1 m2 that penetrated the surficial
black silts near the middle of the chamber floor. A yard-square
piece of ‘/2-inch mesh hail screen lay here, apparently left by
unsuccessful artifact hunters in past years.

436

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 4. — View from Fowlkes Cave entrance, to the northeast, of alluvial fans and desert flats. Dark areas of vegetation in distance fringe
currently dry watercourses.

MATERIALS AND METHODS

A rectangle 10 by 20 ft was laid off in the lower chamber, along
the southern wall of the cave and with the eastern end extending
into the debris cone at the base of the shaft. Matrix was actually
gathered only from the western 10 ft length of the area, however,
though exploratory pits were made at several places in the eastern
end. The black surficial silt, less than 1 ft deep at the western
extremity but more than 16 inches deep at the eastern end, closer
to the debris cone, was collected. The actual weight of the material
was probably near 1,000 pounds. The silt was shoveled into
buckets and, with pulleys and ropes, lifted out of the cave and
passed down a steel cable to a truck on the alluvial fan below.
There the buckets were loaded into burlap bags and the empty
buckets pulled up to the cave mouth to be refilled.

When the surficial, black Recent materials had been removed,
the yellow Pleistocene sediments were exposed. Unlike most cave
deposits, these were stratified. There was very little mud or clay-
like material in the Pleistocene deposits. A trench showed an
upper layer (1) of approximately 35 cm thickness, composed of
irregular limestone fragments and pebbles, bits of stalactitic lime-
stone, cave pearls, and crystals of water-clear calcite. Average
diameter of the rock fragments, exclusive of stalactite fragments,
was between 1 5 and 25 mm. Beneath this was a layer (2) averaging
approximately 20 cm thick, composed of smaller fragments of
similar materials, between 5-15 mm in diameter. Layers 1 and
2 did not seem to have a distinct boundary separating them, but
the nature of the materials made this difficult to determine. Be-
neath Layer 2 was a third layer (3) of perhaps 30 cm in thickness,
composed of finer materials, the consistency of coarse sand. This

layer was set off distinctly from the lower surface of Layer 2.
Beneath Layer 3, calcareous structures, partly “cave coral” from
the cave walls and partly travertine structures formed locally, in
the sediments, were interlocked so intricately that deeper digging
was difficult and investigation has not been done to date.

An exploratory excavation was made in the Pleistocene sedi-
ments where the artifact hunters had dug through the black silts.
However, though fossils were abundant, they were incorporated
into successive, parallel layers of travertine rock, from which
their removal was judged to be too difficult and expensive to
attempt.

The coarse, gravel-like materials of Layer 1 were examined
carefully in several places but no fossil bones were detected.
Therefore, we attempted to dig away and discard all of Layer 1.
We cannot be certain that we were successful at this, for the lower
limits of this layer were not sharply demarked. Because fossils,
if present at all, were very rare here, we doubt that contamination
of Layer 2 occurred.

Layer 3 also seemed to be barren of fossils. Layer 2, the pro-
ductive layer, was removed from the surface of Layer 3 with ease,
for the two layers are sharply demarked. Sediments of Layer 3
were examined in many places but no fossils of any kind were
found. Limb bones, vertebra, or jaws of even small mice should
have been readily apparent in the sand-like materials, even in
the light of a gasoline lantern.

In contrast to layers 1 and 3, Layer 2 contained abundant
vertebrate fossils. Quite early in the excavation, the jaw of a
diminutive antilocaprid, Capromeryx, was found in place near

1984

DALQUEST AND STANGL- FOWLKES CAVE LOCAL FAUNA

437

the middle of the layer. This extinct species was sufficient to
establish the Pleistocene age of the layer, and deeper excavation
was thought undesirable at that time. Thirty buckets filled with
Pleistocene sediments, with estimated total weight in excess of
1,000 pounds, together with remains found in Recent silts, form
the basis of this report.

The early Recent silts were washed and sorted in the customary
fashion (Hibbard, 1949) and presented no problems. The half-

ton of Pleistocene matrix was also washed but this resulted in
the elimination of only about 200 pounds of the finely-divided
material. However, the coarse, clean, washed matrix was reduced,
with the aid of gold-pans, to approximately 50 pounds of con-
centrate containing almost all of the teeth, bones and bone frag-
ments. This concentrate was sorted under jewelers’ lenses to ob-
tain the teeth and jaws used in this report.

SYSTEMATIC ACCOUNTS OF PLEISTOCENE SPECIES

Of the 42 taxa listed from the Pleistocene deposits
of the Fowlkes Cave local fauna, the two ungulates
(Mylohyus and Capromeryx cf. C. furcifer) are ex-
tinct. Of the 40 extant forms, a pocket mouse (Per-
ognathus) can not be identified to species and a har-
vest mouse ( Reithrodontomys fulvescens ) is only
tentatively identified. Thirty-eight species have been
identified with reasonable certainty. In the accounts
of species of the Pleistocene fauna that follow, the
geographic range and present day habitat require-
ments of each species are given, if pertinent, and
the basis of identification is explained. Numbers in
parentheses in each account are those of the Mid-
western State University Collection of Fossil Ver-
tebrates and pertain to material mentioned in Table
1 . Scientific and vernacular names, and systematic
arrangement of species are taken from Jones et al.
(1982).

Sorex vagrans Baird

Vagrant Shrew ( 1 1 928, 1 1 929)

The vagrant shrew does not occur in Texas today
but is found in the mountains of south-central New
Mexico (Hall, 1981), and ranges southward onto the
Mexican Plateau. It occurred in Eddy Co., southern
New Mexico, in post-Pleistocene times (Findley,
1965). In New Mexico it is found in greatest num-
bers in streamside thickets, marshes, and meadows.
It occupies woodlands where the soil is soft and
humid. In New Mexico, it is an alpine form that
does not occur at low elevations.

Sore x palustris Richardson

Water Shrew (1 1947)

Water shrews range across much of Canada and
mountainous areas of the United States, but today
do not occur south of the mountains of western-
central Arizona and northern New Mexico. In the
late Pleistocene, the species lived on what is now
the Mesquite Plains of the southern Texas Panhan-
dle (Dalquest 1 965). Typically, the water shrew lives

along the margins of small, swift, cold-water streams
in alpine habitat but is often found in marshes and
wet meadows beside brooks and rills.

Notiosorex cranfordi (Coues)

Desert Shrew (1 1930)

This must have been an abundant species near
Fowlkes Cave in the late Pleistocene, for it is rep-
resented by more than 100 fossil specimens. The
desert shrew ranges from California to Arkansas and
from southern Colorado to central Mexico. It has a
broad tolerance to environmental conditions, and
is found from dry deserts to tropical forests. Its pre-
ferred habitat seems to be arid brushland, and it is
rarely trapped by collectors. However, one individ-
ual was captured under a rock in typical Chihuahuan
desert habitat in Jeff Davis Co., about 15 km south
of Fowlkes Cave.

Myotis lucifugus (LeConte)

Little Brown Bat (11931)

The only specimen is a lower jaw fragment but
preservation of the fossil is good. Size and dentition
match specimens of the living Myotis lucifugus oc-
cultus Hollister. Hall (1981) indicates that the only
record of Myotis lucifugus from Texas is a specimen
from Fort Hancock, a site well east of Fowlkes Cave,
but the species ranges widely in New Mexico. It is
typically a woodland bat, but wanders into other
habitat in its nocturnal hunting activities. Presence
near Fowlkes Cave today, as a transient, would not
be unexpected.

Myotis velifer (J. A. Allen)

Cave Myotis (1 1933)

The cave myotis lived in Fowlkes Cave in the
recent past, for many bones were discovered on the
surface of the black silts of the cave floor and were
scattered through the surficial silts. Fossil bones were
common in the late Pleistocene sediments as well.
Jaws from the Pleistocene part of the cave average

438

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table l.— Modern, early Recent, and Pleistocene Fauna of Fowlkes Cave. Note: Column 1 includes species known to live in the vicinity
of the cave today. Column 2 lists species whose remains were found in the early Recent deposits of the cave. Column 3 indicates Pleistocene

records from the cave.

Species

1

2

3

Basis for Pleistocene record

Sorex vagrans

+

2 lower jaws with teeth

Sorex palustris

+

1 maxillary with teeth

Notiosorex crawfordi

+

+

+

2 rostra, 14 maxillaries, 1 14 lower jaws

My otis lucifugus

+

1 lower jaw with M,-M3

Myotis velifer

+

+

+

1 rostrum, 2 maxillaries, 27 lower jaws

Pipistrellus hesperus

+

Eptesicus fuscus

+

+

7 lower jaws

Antrozous pallidus

+

+

Mormoops megalophyla

+

Tadarida braseliensis

+

Homo sapiens

+

+

Lepus californicus

~r

+

+

Several fragments of limb bones

Sylvilagus floridanus

+

1 lower jaw with teeth

Sylvilagus audubonii

+

+

+

2 lower jaws, premaxillary with incisors

Marmota flaviventris

+

5 cheek teeth

Eutamias cinericollis

+

Lower jaw fragment with M,

Ammospermophilus interpres

+

Spermophilus spilosoma

+

Lower jaw fragment with M,; isolated M,1

Spermophilus variegatus

+

+

+

Isolated teeth

Cynomys ludovicianus

+

1 M3

Thomomys bottae

+

+

+

5 skulls, 15 palates, 71 lower jaws

Pappogeomys castanops

+

+

+

4 upper incisors, 4 cheek teeth

Perognathus hispidus

+

+

+

1 skull, 116 lower jaws, many maxillaries

Perognathus sp.

+

+

3 lower jaws

Perognathus flavus

+

+

+

657 lower jaws, many maxillaries

Dipodomys ordii

+

+

+

107 lower jaws

Dipodomys merriami

+

+

+

14 lower jaws

Dipodomys spectabilis

+

+

+

77 lower jaws

Reithrodontomys megalotis

+

+

+

84 lower jaws

Reithrodontomys monlanus

+

+

+

19 lower jaws

Reithrodontomys fulvescens

+

+

2 tentatively referred lower jaws

Peromyscus eremicus

+

+

+

55 lower jaws

Peromyscus maniculatus

+

+

+

16 lower jaws

Peromyscus leucopus

+

+

+

10 lower jaws

Peromyscus boylii

+

22 lower jaws

Peromyscus pectoralis

+

+

+

6 lower jaws

Peromyscus difficilis

+

+

3 lower jaws

Onychomys leucogaster

+

183 lower jaws

Onychomys torridus

+

+

+

6 lower jaws

Sigmodon hispidus

+

+

+

1 palate, 44 maxillaries, 189 lower jaws

Neotoma micropus

+

+

+

23 lower jaws

Neotoma albigula

+

+

+

33 lower jaws

Neotoma mexicana

+

1 lower jaw

Microtus mexicanus

+

2 lower jaws

Erethizon dorsatum

+

1 incisor fragment

Canis latrans

+

+

Bassariscus astutus

+

+

Felis rufus

+

Maxillary fragment with teeth

Dicotyles tajacu

+

Mylohyus sp.

-j-

Tusk fragment

Capromeryx cf. C. furcifer

+

Lower jaw

Odocoileus hemionus

+

1984

DALQUEST AND STANGL-FOWLKES CAVE LOCAL FAUNA

439

larger than Recent specimens and probably repre-
sent Myotis velifer magnimolaris Choate and Hall,
a race now found in Kansas, Oklahoma, and north-
central Texas (Dalquest and Stangl, in press).

The cave myotis ranges from Kansas southward
over much of Mexico, with a distinct preference for
arid habitat. Its presence in Fowlkes Cave is not
surprising. There are numerous records of the species
from late Pleistocene caves in Texas.

Eptesicus fuscus (Beauvois)

Big Brown Bat (1 1934)

The big brown bat is widespread, ranging from
the tropical jungles of southern Mexico to southern
Canada and from coast to coast. We did not capture
living specimens at Fowlkes Cave, but it may occur
there, at least at times. Eptesicus fuscus is not usually
considered a desert bat but we have specimens taken
in typical Chihuahuan desert habitat in Presidio Co.,
Texas. There are numerous records of the big brown
bat from the Pleistocene of Texas.

Lepus californicus Gray

Black-tailed Jackrabbit (1 1935)

Black-tailed jackrabbits range widely over much
of western United States and Mexico, and are com-
mon about Fowlkes Cave today. However, the only
fossils recovered in the cave include the proximal
end of a femur, distal half of a tibia, distal end of a
humerus, and part of an inominate. The sutures of
the bones are closed but they are all a bit smaller
than bones of modem jackrabbits from north-cen-
tral Texas.

The scarcity of jackrabbit remains in the Pleis-
tocene sediments of Fowlkes Cave probably results
from the habits of the species. They prefer level areas
of ground, with a cover of grass or scattered clumps
of brush. They would avoid the steep, barren hillside
where the mouth of Fowlkes Cave is located. Adults
are too large to be brought into the cave by barn
owls. Lepus californicus has been recorded from al-
most all late Pleistocene cave deposits in Texas where
the microvertebrate fossils have been studied.

Sylvilagus floridanus (Allen)

Eastern Cottontail (1 1936)

The lower jaw fragment contains the M2 and the
alveoli of the other cheek teeth. The M2 is 2.2 mm
long, the trigonid is 2.8 mm wide, and the talonid
is 2. 1 5 mm wide. Alveolar length P2-M3 is 1 3.8 mm.
These measurements are too great for Sylvilagus

audubonii and are like those of the few available
specimens of S. f robustus from the Davis Moun-
tains, to the south of Fowlkes Cave. Some authors
consider the Davis Mountains cottontail to belong
to a distinct species, S. robustus Bailey.

The Davis Mountains cottontail lives in the pine-
oak woodland of the Davis Mountains and the Gua-
dalupe Mountains. It ranges into the oak-juniper
belt where there is ample cover, but does not enter
the desert habitat like that surrounding Fowlkes Cave
today. The presence of the single specimen shows
that the range of the eastern cottontail extended
completely across the gap between the ranges of the
present populations in the late Pleistocene. The sin-
gle fossil also indicates existence of areas of thicker
brush or woodland in the late Pleistocene than occur
near Fowlkes Cave today.

Sylvilagus audubonii (Baird)

Desert Cottontail (1 1937-1 1939)

In addition to the material listed in Table 1, there
are numerous isolated cheek teeth and some limb
bones that probably belong to this species. Many of
the teeth are of immature animals. The desert cot-
tontail is a small species but adult individuals are
probably too large to be readily overcome by barn
owls, and the fossils of adult cottontails are probably
of animals that fell into the cave trap.

Eutamias cinericollis (J. A. Allen)

Gray-collared Chipmunk (11941)

Chipmunk is represented by a lower jaw fragment
with the M2. The specimen is referred to the species
that lives today in the Guadalupe Mountains, but
the reference is tentative. Other species occur in the
mountains of New Mexico.

Eutamias cinericollis inhabits montane habitat
and does not descend to the desert. It probably ranged
into the Fowlkes Cave area in the late Pleistocene
along cool, ice-fed streams, along with the vagrant
shrew, water shrew, voles, and other montane
species. There are records from the late Pleistocene
cave deposits in the Guadalupe Mountains and
southern New Mexico.

Marmota flaviventris (Audubon and Bachman)

Yellow-bellied Marmot (1 1940)

The five cheek teeth are clearly of a single indi-
vidual. They are P4, M1, M,, and right and left M2’s,
with roots closed but almost unworn.

Marmots today occur no closer to Fowlkes Cave

440

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

than the mountains of northern New Mexico, and
their range extends northward almost to the arctic
region. There are numerous late Pleistocene records
from southern New Mexico and southward as far
as Nuevo Leon, Mexico (Stock, 1942). The yellow-
bellied marmot seems to be an obligatory inhabitant
of rocky areas, cliffs and talus. In the western parts
of the range of the species (Nevada, Oregon, etc.),
marmots occur in desert habitat but they do not
occur today in Chihuahuan Desert.

Spermophilus spilosoma Bennett

Spotted Ground Squirrel (1 1942)

A lower jaw fragment has the alveoli of P4-M3
but only M2 present. There is also an isolated M1,
tentatively referred. Alveolar length of the tooth row
is 7.8 mm, and the M2 measures 1.9 by 1.9 mm.
The isolated tooth measures 2.1 by 1.8 mm. The
jaw ramus is not narrowed at the site of P4, as is the
similar-sized antelope ground squirrel, Ammosper-
mophilus. The jaw and teeth differ from Ammosper-
mophilus in other details also.

Spermophilus spilosoma lives in prairie and des-
ert habitat from South Dakota to central Mexico. It
shows definite preference to sandy soils, which are
almost lacking about Fowlkes Cave today but must
have been present in the late Pleistocene.

Spermophilus variegatus (Erxleben)

Rock Squirrel (1 1943)

One enamel cap of an upper molar had not erupt-
ed, and is definitely of the rock squirrel. However,
there are three well-preserved upper and one lower
cheek teeth, and several worn or broken teeth, that
are only tentatively referred to this species. The
enamel pattern and size of these teeth is like that of
the rock squirrel, but the lingual halves of the crowns
are markedly narrower than in S. variegatus. Nu-
merous specimens of the modern forms from Trans-
Pecos Texas and the Edwards Plateau were studied,
and none had teeth like these fossils. In shape, the
teeth resemble those of S. columbianus but are much
larger than that northern species. S. beecheyi (Rich-
ardson) and S. annulatus Audubon and Bachman
are large species but their teeth do not have the
narrowed, triangular shape like that of the fossils.
The fossils show several different stages of wear, and
represent several different individual squirrels. They
probably represent a slightly aberrant population of
S. variegatus. If rock squirrels live or lived in an
area, their remains may be expected in the cave
deposits. Fossils are absent in the late Pleistocene

deposits of caves on the Edwards Plateau, though
they are common in the early Recent surficial silts
of the same caves. The species is recorded from the
caves in the Guadalupe Mountains (Williams Cave,
Ayer, 1937; Upper Sloth Cave, Logan and Black,
1979) but seems to be absent from the well-studied
caves of southern New Mexico (Harris, 1974). Ap-
parently rock squirrels lived in the Trans-Pecos area
in the late Pleistocene but did not extend their range
eastward to the Edwards Plateau and northward to
southern New Mexico until the Recent.

Cynomys ludovicianus (Ord)

Black-tailed Prairie Dog ( 1 1 944)

The prairie dog is represented in the cave fauna
by a single worn M3. This diurnal, large squirrel
would not be prey for the bam owl, and the tooth
must have arrived in the cave in some fashion other
than most of the other fossils.

Prairie dogs lived in Culberson Co. until quite
recently, but we found no evidence of their existence
near Fowlkes Cave today. Doubtlessly they have
been exterminated by man. There are records of the
prairie dog from caves in southern New Mexico,
and from Pratt Cave, in the Guadalupe Mountains
(Lundelius, 1 979), but fossils of this species are usu-
ally rare in caves. The rodents are active only by
day, and are too large for the cave-inhabiting owls,
like the bam owl, to carry. They live on level grass-
lands, usually far from the rugged terrain where caves
occur.

Thomomys bottae (Eydoux and Gervais)

Botta’s Pocket Gopher (1 1953-1 1956)

Remains of Thomomys are abundant in the cave
and include nearly complete skulls, numerous pal-
ates, and many lower jaws. Isolated teeth are present
by the thousands. Obviously the small pocket go-
pher was a favored prey of the barn owls in the late
Pleistocene as it is today. The cave record shows
that it was abundant in the past, both in the Pleis-
tocene and early Recent, but it is a rare resident near
the cave today. We trapped one specimen about 2
km from the cave mouth.

Schmidly (1977) notes that the common pocket
gopher of the deeper soils of the Trans-Pecos is Pap-
pogeomys castanops, whereas Thomomys bottae is
found on thinner soils of hillsides and mountains,
usually in the vicinity of lecheguilla plants. It has
been suggested (Reichman and Baker, 1972) that
Pappogeomys is replacing Thomomys in the Trans-
Pecos. This seems to be true at Fowlkes Cave, where

1984

DALQUEST AND STANGL- FOWLKES CAVE LOCAL FAUNA

441

remains of Thomomys are abundant in late Pleis-
tocene and early Recent sediments. We set traps
near the cave, where gopher workings were observed
near lecheguilla plants on thin soils of hillsides, and
expected to capture Thomomys. Instead, the traps
took only Pappogeomys. Our only Thomomys was
trapped in deep soils, and a Pappogeomys was taken
in a trap set 30 meters away. Pappogeomys remains
are rare in the cave sediments, but the species today
lives about the cave almost everywhere that gophers
can exist. Changes in the environment caused by
cattle or other activities of man may be responsible
for the increasing scarcity of Thomomys in Culber-
son Co.

It is possible that the cave record may be decep-
tive with regard to Pappogeomys. The animals are
quite large and their burrows are often deeper in the
ground than those of Thomomys. Owls may be less
able to capture the larger species. We are unable to
explain the ability of the barn owl to capture Tho-
momys so easily, in view of the almost completely
fossorial habits of the rodents. However, exami-
nation of regurgitated pellets of barn owls collected
from areas where Thomomys live, almost invariably
show Thomomys to be a common food item. What-
ever feature of the habits of Thomomys that makes
them susceptible to predation by bam owls may not
operate for Pappogeomys.

Pappogeomys castanops (Baird)

Yellow-faced Pocket Gopher (1 1957)

Fossils are rare and include four upper incisors,
one P4 and three P4’s. In contrast, Thomomys is
represented by nearly one hundred skulls and jaws.
Although fragmentary, the fossils show definitely
that Pappogeomys lived about Fowlkes Cave in the
late Pleistocene. There are records from cave de-
posits in the Guadalupe Mountains and southern
New Mexico as well, but most suggest relatively late
Pleistocene or early Recent age. Pappogeomys are
found from Kansas and Colorado southward deep
into Mexico. It is not known from the Pleistocene
cave deposits of the Edwards Plateau, but does occur
there in early Recent cave sediments. Presumably,
Pappogeomys invaded Texas in the late Pleistocene.

Genus Perognathus

Five species of pocket mice occur in the Trans-
Pecos today, but one of these (Perognathus nelsoni )
is not recorded from Culberson Co. It may, of course,
have lived there in the past. We trapped specimens
of the large species, Perognathus hispidus, and the

diminutive species, P. flavus, less than 1 km from
the cave. We did not find the other species (P. nel-
soni, P. intermedins, and P. penicillatus), although
special efforts were made to catch them.

Identification of fossil Perognathus jaws presents
special problems. No characters of the enamel pat-
tern seem reliable in separating species, and any
enamel pattern is transient because surface patterns
of the teeth wear down swiftly and change with age
and wear. Relative size of P4 versus M3 may be
helpful but is of little practical value for the Fowlkes
Cave fossils, because most have lost the M3. Trans-
verse breadth of the lower incisor and alveolar length
of the tooth row are the only effective means of
identifying the fossils, and these are of limited value
in separation of some species.

Alveolar length of the lower tooth row was de-
termined in 25 adult, randomly-chosen specimens
of the five species of pocket mice. Measurements
were taken with an occular micrometer, with the
jaw held at approximately a 40° angle, and viewed
from the lingual side. Extremes of measurements
were: P. hispidus, 4.0-4.75 mm; P. nelsoni, 3.35—
3.70 mm; P. intermedins, 3.25-3.70 mm; P. peni-
ciliiatus, 3.10-3.90 mm; P. flavus, 2.65-3.30 mm.

The incisor breadth is readily measured with a
caliper. In P. hispidus the breadth exceeds .65 mm;
in the other species the breadth is less than .6 mm.
P. hispidus is readily identified by the breadth of the
incisor. In the other species, however, there is so
much overlap that the character is of no value. Sur-
prisingly, the little P. flavus sometimes has the
breadth of the incisor greater than that of some P.
peniciiiatus.

Perognathus hispidus lower jaws are therefore sep-
arated with confidence on the basis of incisor breadth
and alveolar length of tooth row. The tiny P. flavus
can not be identified with certainty, but more than
90% of the jaws of this species have the length of
the tooth row less than 3.15 mm. More than 90%
of the P. penicillatus examined have the tooth row
more than 3.15 mm, as do all P. nelsoni and P.
intermedins. Thus, lower jaws with the alveolar
length of the tooth row 3.15 mm or less are referred
to P. flavus with the reservation that a very few small
(tooth row 3. 1-3.2 mm) P. penicillatus might be
included. No firm method of separating fossils of
the three medium-sized species of pocket mice was
found.

Alveolar length of the tooth row of Perognathus
tends to be greater in aged animals, where the bone
often retreats from the roots of the teeth when these

442

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

are excessively worn. This is apparent only in jaws
with greatly worn teeth. It was also noted that mea-
surements of alveolar length of the tooth row tended
to be slightly greater in jaws lacking P2 and M3.
Perhaps the loss of teeth involves slight but imper-
ceptible damage to the alveoli.

Perognathus hispidus Baird

Hispid Pocket Mouse (1 1945)

Remains of this species are abundant in the Pleis-
tocene and early Recent deposits of the cave. It was
clearly a favored food of the barn owls.

The hispid pocket mouse has a geographic range
that extends from North Dakota southward deep
into Mexico, and includes much of Texas. It is pri-
marily a prairie or grassland mouse. Its abundance
in the past may be due to the apparent absence of
competing medium-sized species of pocket mice. P.
hispidus is known from many late Pleistocene cave
deposits, including those in the Guadalupe Moun-
tains and southern New Mexico.

Perognathus flavus Baird

Silky Pocket Mouse (1 1947)

The silky pocket mouse is the most abundant
species of mouse in the cave fauna, represented by
657 complete to fragmentary lower jaws and nu-
merous palates and maxillary fragments.

Perognathus flavus occurs from Wyoming to
southern Mexico and from western Arizona to the
Gulf Coast of Texas, including all of the Trans-
Pecos. It has a broad range of tolerance to environ-
mental conditions and inhabits prairies and deserts,
level ground and rocky, broken land. No particular
significance can be placed on its occurrence in
Fowlkes Cave. Wherever found, it seems to be a
favored prey of barn owls.

Perognathus sp. (1 1946)

Three lower jaws are from medium-sized pocket
mice. They may be of any of the following: Perog-
nathus intermedius, P. nelsoni, or P. penicillatus.

Perognathus nelsoni and P. intermedius prefer
stony soils and are rarely found far from rocks and
rough, broken land. P. nelsoni seems not, at this
time, to occur north of the Davis Mountains, where-
as P. intermedius has been found only along the
western border of Culberson Co. The present ab-
sence of these species in the vicinity of Fowlkes Cave
and the areas to the north and east, may be signif-
icant in view of the scarcity of remains of medium-

sized pocket mice in the Pleistocene sediments of
the cave.

Perognathus penicillatus occupies loose, sandy
soils, and its range includes virtually the entire Trans-
Pecos. No suitable habitat for the desert pocket
mouse was found within several kilometers of the
cave, and trapping in the sandy areas that we did
find, yielded no specimens of P. penicillatus. Sandy
soils certainly existed closer to the cave in the late
Pleistocene, in view of the abundance of other sand-
dwelling species (for example, Dipodomys ordii,
Onychomys leucogaster) in the cave fauna. It is
probable, on the basis of modem distribution, that
the three fossil jaws listed belong to P. penicillatus.
However, where P. penicillatus is found today, it is
usually abundant, whereas P. intermedius and P.
nelsoni are rarely abundant in any area. If P. peni-
cillatus lived near Fowlkes Cave at all, it should
have been represented by numerous fossil jaws. The
extent of predation on this species by barn owls
seems not to have been investigated in detail.

Genus Dipodomys

Three species of kangaroo rats occur near Fowlkes
Cave today— the large Dipodomys spectabilis, me-
dium-sized D. ordii, and small D. merriami. The
three are readily separated by size of the lower jaw
and incisor. In D. spectabilis the breadth of the in-
cisor is greater than 1.0 mm, with no overlapping
of the measurements of those of D. ordii. Incisor
breadth of D. ordii is .8 to .95 mm; in D. merriami
it is less than .8 mm. Although no actual overlap in
measurements was found in specimens from the
Trans-Pecos, it is possible that some slight overlap
may occur.

Alveolar length of tooth row is greater than 5.6
mm in Dipodomys spectabilis, 4.8-5. 5 mm in D.
ordii, and less than 4.75 in D. merriami.

The breadth of the incisor is the best character to
use in separation of fossil jaws. The incisor usually
remains in the fossil jaws and is readily measured
with a caliper. If the incisor is missing, the alveolar
length of the tooth row is available. Unfortunately,
the M3 of the fossil Dipodomys is often missing, and
usually the thin bone bordering the alveolus pos-
teriorly is broken away as well. It is usually impos-
sible to then estimate the position of the posterior
border of the alveolus. Dipodomys spectabilis can
then be identified by the large size of cheek teeth,
and D. merriami can be separated from D. ordii by
smaller molars.

1984

DALQUEST AND STANGL- FOWLKES CAVE LOCAL FAUNA

443

Dipodomys ordii Woodhouse

Ord’s Kangaroo Rat (1 1948)

Ord’s kangaroo rat is found in desert areas of the
west from southern Canada to southern Mexico,
including all of the Trans-Pecos. It shows a strong
preference for loose, sandy soils and rarely occurs
on clay soils or stony ground. The abundance of this
species at Fowlkes Cave in Pleistocene deposits sug-
gests that loose, sandy soils must have existed near
the cave in the late Pleistocene. No suitable habitat
for the species was noted around the cave today.
However, one specimen was taken in a trap less than
1 km from the cave, along with several Merriam
kangaroo rats. It was not recognized as D. ordii until
it was being prepared as a study specimen. A small
population survives in the area, though suitable hab-
itat is several kilometers away. Whether the speci-
men represents a remnant of the Pleistocene pop-
ulation or is a more recent invader can not be known.

Dipodomys spectabilis Merriam

Banner-tailed Kangaroo Rat (1 1950)

The banner-tailed kangaroo rat is found from
northern New Mexico to central Mexico, and over
much of the Trans- Pecos. It inhabits the high desert
grasslands but occurs also on mesquite flats and
aluvial fans. It does not seem to like the extensive
creosote bush flats with their stony soils, or steep
hillsides. It is moderately common about Fowlkes
Cave, and a number of specimens were trapped near
the cave.

Dipodomys merriami Meams
Merriam’s Kangaroo Rat (1 1949)

Merriam’s kangaroo rat has a geographic range
extending from northern Nevada to central Mexico
and California to western Texas. Its preferred hab-
itat in Texas is the creosote bush flats, but it has a
relatively great range of tolerance to conditions. D.
merriami and D. spectabilis are commonly sym-
patric. Wherever found, D. merriami is an indicator
of true, arid, desert conditions. It is the common
kangaroo rat near Fowlkes Cave today. Its relatively
scarcity (14 jaws versus 107 D. ordii and 77 D.
spectabilis) indicates a change in environment be-
tween the late Pleistocene and modem times.

Genus Reithrodontomys

Three species of harvest mice are sympatric in
Trans-Pecos Texas today —Reithrodontomys meg-

alotis, R. montanus, and R. fulvescens. R. montanus
can readily be separated from the other two forms
by the alveolar length of the tooth row— less than
3.0 mm in R. montanus, but greater than 3.0 mm
in R. megalotis and R. fulvescens. Although R. ful-
vescens is larger and heavier-bodied than R. meg-
alotis, there is some overlap in measurements of
breadth of lower incisor, alveolar length of M,-M3,
and length of M,. In jaws retaining a M, in the
proper stages of wear, R. megalotis exhibits an “C”
pattern to the enamel, while the pattern in R. ful-
vescens appears as a “S.”

Reithrodontomys megalotis (Baird)

Western Harvest Mouse (11912)

The geographic range of Reithrodontomys meg-
alotis extends from southern Canada to southern
Mexico and includes much of the United States west
of the Mississippi. It is found throughout the Trans-
Pecos, and is common about Fowlkes Cave today.
Its habitat extends from hot desert to alpine mead-
ows, and grassy plains and prairies to cliff's and talus
slides. The species is not a reliable indicator of past
climate. It has been found in the Pleistocene fauna
of Pratt Cave, Culberson Co. (Lundelius, 1979).

Reithrodontomys montanus (Baird)

Great Plains Harvest Mouse (11911)

Reithrodontomys montanus is a grassland mam-
mal, usually found where the grass is relatively short
but dense enough to offer overhead cover. It is most
often found in moderately arid climate where soils
are well-drained. It occurs in the Trans-Pecos today,
but usually is sparsely distributed and rare. We
trapped one specimen near Fowlkes Cave. The 18
jaws recorded from Fowlkes Cave (versus 84 jaws
of R. megalotis) suggests that R. montanus was rel-
atively more abundant near the cave in the late
Pleistocene, and that grasslands were present closer
to the cave than at present. We know of no records
of R. montanus from the caves in the Guadalupe
Mountains or southern New Mexico.

Reithrodontomys cf. fulvescens J. A. Allen

Fulvous Harvest Mouse (1 1913)

Two lower jaws in the cave collection have the
alveolar length of the lower tooth row and breadth
of the lower incisor greater than the maximum ex-
pected in Reithrodontomys megalotis (3.25 and .56
mm respectively). These measurements are within
the lower limits for R. fulvescens. Both, unfortu-

444

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

nately, lack the diagnostic M3. They are hesitantly
referred to R.fulvescens, but may be unusually large
late Pleistocene representatives of R. megalotis.

The geographic range of Reithrodontomys fulves-
cens extends from Kansas and Missouri southward
through Mexico to Nicaragua. It is rare in the Trans-
Pecos, seemingly confined to the eastern half of the
area. It has not been recorded from Culberson Co.
but has been taken in Reeves Co., 50 mi to the east
(Schmidly 1977). The fulvous harvest mouse is a
grassland species that prefers denser and taller grass
than does R. montanus. The two species rarely occur
together. There are no records of R. fulvescens from
the Pleistocene cave deposits in the Guadalupe
Mountains or southern New Mexico.

Genus Peromyscus

This genus presents special problems in identifi-
cation in that seven species are known to live today
in Culberson Co. (Schmidly, 1977) and another
( Peromyscus crinitus Merriam) has been recorded
from Dry Cave, Eddy Co., New Mexico (Harris,
1974), slightly more than 100 mi to the north. P.
maniculatus, P. leucopus, P. eremicus, and P. pec-
toralis occur within 1 km or so of Fowlkes Cave,
and P. boylii, P. difficilis, and P. truei occur at higher
elevations, in the northern part of the county. Cri-
teria for the identification of lower jaws of these
eight species, based on Recent specimens, have been
discussed elsewhere (Dalquest and Stangl, 1983).
Only a brief summary is given here.

The incisor-base capsule is rated in three grades—
Grade 0 (only a slight bulge when seen from directly
above). Grade 1 (strong bulge when seen from di-
rectly above but not forming a blunt, finger-like pro-
jection), and Grade 2 (capsule so strongly developed
as to form a blunt, finger-like projection). This char-
acter serves to separate Peromyscus maniculatus, P.
leucopus, P. crinitus, and most P. eremicus from the
other four forms concerned here.

The anteroconid of M, is also rated in three
grades— Grade 0 (not divided). Grade 1 (narrowly
divided), and Grade 2 (separated by a broad anterior
valley). Other complications of the enamel pattern
of M, found to be of value include the presence or
absence of ectostylids, mesostylids, ectolophids, and
mesolophids. Measurements of the alveolar lengths
of M,-M3 serve to identify Peromyscus difficilis and
aid in identification of other forms.

Measurements and characters overlap broadly in
some species but, in combination, permit identifi-
cation of nearly all jaws of some species, some in-

dividuals of other species, and tentative identifica-
tion of still others.

Peromyscus crinitus is readily separated from the
other concerned species. The incisor-base capsule is
Grade 2, anteroconid not divided, lophids and sty-
lids absent on M,, and alveolar length of MrM3 is
3.5 mm or less. The incisor-base capsule is uniquely
narrow. It would be possible to confuse this species
with lower jaws of large Reithrodontomys if the in-
cisor-base capsule were severely damaged.

Only four specimens of Peromyscus truei were
available from Texas. All had the incisor-base cap-
sule Grade 0, two had divided anteroconids (Grade
2), one had an ectostylid, and another had an ec-
tolophid. Alveolar length of the lower cheek-tooth
row ranged from 3.85 to 4.20 mm. Specimens from
Arizona, Utah, and westward usually had M,’s with
deeply divided anteroconids and strongly developed
stylids and lophids. The Texas mice appear to be
more variable and to have more simple M,’s than
those to the west. Four specimens are too few to
characterize the species but it would appear that it
might be confused with large individuals of P. pec -
toralis or P. boylii, or even small individuals of P.
difficilis. No fossils are referred to P. truei.

All specimens of Peromyscus from Fowlkes Cave
that were utilized were more or less fragmentary
lower jaws. Another 22 jaws are moderately well-
preserved but either are too worn or have critical
features too damaged to be identified. These have
been referred to Peromyscus sp. (1 1922).

Peromyscus difficilis has the incisor-base capsule
Grade 0, anteroconid divided (90%), stylids and lo-
phids usually present (86% and 82%), and the al-
veolar length of M,-M3 is 4.1 mm or greater. Only
the very largest P. pectoralis might be mistaken for
P. difficilis, having the M^M., length 4.1 mm.

Peromyscus boylii and P. pectoralis have the in-
cisor-base capsule Grade 0 and the M,-M3 length
almost the identical. However, lower jaws are readi-
ly separated because P. boylii has relatively few sty-
lids (29%) and almost never has lophids in the major
valleys of M,. P. pectoralis usually has stylids (86%)
and lophids (82%). The anteroconid of M, is divided
in P. boylii in about 60% of the specimens studied,
but in P. pectoralis the anteroconid is almost always
divided, and is usually more deeply and strongly
divided than in P. boylii. Thus, P. boylii and P.
pectoralis can be separated with confidence in al-
most all instances.

Separation of these two species from P. eremicus
is also a problem, for P. eremicus may have a Grade

1984

DALQUEST AND STANGL- FOWLKES CAVE LOCAL FAUNA

445

0 incisor capsule. More than a third of the specimens
examined (39%), resemble P. boylii and P. pectoralis
in this respect. Because P. eremicus usually (82%)
lacks the divided anteroconid, it differs from P. pec-
toralis, where the anteroconid is almost always di-
vided. Moreover, the 18% of P. eremicus that do
have the anteroconid divided, it is always Grade 1 .
In P. pectoralis, it is Grade 2 in 61% of the Recent
specimens studied. Thus, P. pectoralis is separated
from P. eremicus with confidence, although it is
possible that a very few specimens might be wrongly
identified by the listed characters.

There are no good characters that will reliably
separate the lower M, of P. boylii from those P.
eremicus that have Grade 0 incisor-base capsules.
Although the alveolar length of the lower tooth row
overlaps broadly in the two species (3.55-3.95 mm
versus 3.65-4.20 mm), approximately 90% of the
specimens of P. eremicus have the measurement
3.85 mm or less, whereas 79% of the P. boylii have
the length 3.85 mm or greater. This is used in the
present identifications.

Identification of those species with well-devel-
oped incisor-base capsules ( Peromyscus inanicu-
latus, P. leucopus, and P. eremicus) is more difficult.
However, P. leucopus can be separated from the
other two species in almost all instances. P. leucopus
is larger, and the alveolar length of the lower tooth
row, although it overlaps that of the other species
broadly ( leucopus 3.55-4.15 mm; maniculatus 3.42-
4.00 mm; eremicus 3.55-3.95 mm), almost 90% of
the P. eremicus have the tooth row less than 3.85
mm, and 96% of the P. maniculatus have it less
than 3.85 mm, whereas 79% of the P. leucopus have
the tooth row 3.95 mm or greater. Further, all P.
leucopus have undivided anteroconids and lack lo-
phids in the Ml5 whereas 21% of the P. maniculatus
and 5% of the P. eremicus have divided anteroco-
nids and 29% of the P. eremicus possess lophids on
the M/s. Still further, almost all (96%) of the P.
leucopus have the incisor-base capsule Grade 2,
whereas 36% of the P. maniculatus and 39% of the
P. eremicus have the incisor-base capsule Grade 1 .

As a check, all fossil jaws identified as Peromyscus
leucopus because they had Grade 2 incisor-base cap-
sules and the alveolar length of the M,-M3 measured
3.95 or more, were checked for divided anteroco-
nids and presence of lophids. None were found.

Separation of Peromyscus maniculatus from those
P. eremicus that have Grade 1 incisor-base capsules
is more difficult. Jaws with incisor-base capsules
Grade 2 and tooth rows measuring less than 3.95

mm are considered to be P. maniculatus, though a
very few small P. leucopus could be included or even
some P. eremicus. There are average differences be-
tween Recent specimens of the two species (for ex-
ample, 92% of P. eremicus have lophids present on
Mj versus only 7% in P. maniculatus), but we find
no character that can be of reliable use. In Recent
material, only 60% of the P. eremicus resemble P.
maniculatus in having Grade 1 or Grade 2 incisor-
base capsules. There are 19 jaws of this type, and
55 jaws referred to P. eremicus because they have
Grade 0 incisor-base capsules but alveolar length of
lower tooth row 3.85 mm or less. If only P. eremicus
were included in the collection, these figures would
be reverse, providing that the percentages were the
same in the late Pleistocene as they are today.

There is evidence of the presence of Peromyscus
maniculatus in the Fowlkes Cave collection, on the
basis of specimens with Grade 2 incisor-base cap-
sules but there are 19 jaws that might represent either
P. maniculatus or P. eremicus.

With the foregoing reservations, it is thought that
six species of Peromyscus are represented in the cave
collection.

Peromyscus eremicus (Baird)

Cactus Mouse (11919)

Peromyscus eremicus ranges from southern Ne-
vada and Utah southward through the deserts of
Mexico, and from western Texas to California. It is
found in suitable habitat throughout the Trans-Pe-
cos. It inhabits the hot desert flats and rocky hillsides
above the creosote bush flats, where cover in the
form of cactus, yuccas, or other plants or rocks oc-
cur. It is commonly associated with the Merriam
kangaroo rat and southern grasshopper mouse. The
few records of this species from the late Pleistocene
are open to question in view of the difficulty of
identification.

Peromyscus maniculatus (Wagner)

Deer Mouse (11914)

This species is almost ubiquitous, ranging over
most of temperate North America and penetrating
into almost every habitat where small mammals
survive. The present inhabitant of Culberson Co.,
including the Guadalupe Mountains (Schmidly
1977) is P. m. blandus Osgood, a pale race usually
considered a desert mouse. Today it is abundant
about Fowlkes Cave, inhabiting the desert flats and
the lower margins of the aluvial fans, but rarely

446

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

enters rocky land or hillsides where other species of
Peromyscus occur. Because the population of the
Guadalupe Mountains as well as that of the desert
about Fowlkes Cave are both referable to P. m.
blandus, that is probably the race that lived near the
cave in the late Pleistocene.

Peromyscus leucopus (Rahnesque)
White-footed Mouse (11915)

Peromyscus leucopus is found over much of east-
ern United States, from southern Canada to south-
ern Mexico, and westward to Arizona, including
virtually all of Texas. The desert race, P. 1. tornillo
Meams, prefers denser cover than most desert
mammals, and is most common in thickets along
streambeds and dry washes, but it enters rocky areas
as well. There are few records of P. leucopus from
the late Pleistocene of western Texas and southern
New Mexico, probably due to difficulty in identifi-
cation.

Peromyscus boylii (Baird)

Brush Mouse (11918)

Peromyscus boylii ranges from northern Califor-
nia and Utah to southern Mexico, and western Tex-
as to California. In Trans-Pecos, Texas it is found
mostly at higher elevations, in the pinyon-juniper
belt and upward into the pine forests. In some lo-
calities, however, as in southern Brewster Co., it is
found in the denser woody vegetation fringing can-
yons at low elevation, not far from the Rio Grande.
The presence of rocky cliffs or talus in rather dense
brushland or woods seems to be a requirement for
the species.

There is no habitat at present near Fowlkes Cave
suitable for P. boylii, but it is found in the Davis
Mountains to the south and the Guadalupe Moun-
tains to the north. Its presence in the cave fauna
indicates heavier cover on the hills about the cave
in the late Pleistocene. The only record for the late
Pleistocene of southern New Mexico is that of Harris
(1974), and this record is quieried.

Peromyscus pectoralis Osgood
White-ankled Mouse (11917)

The comparative scarcity, six only, of jaws refer-
able to P. pectoralis is surprising, for it occurs near
the cave today. The species ranges from Texas and
southeastern New Mexico southward into Mexico.
Preferred habitat is moderately dense vegetative
cover growing in and about cliffs, rocky talus, or
rough, broken land. In the Trans-Pecos, Texas it is

found with P. boylii in the pinyon-juniper belt and
downward into desert habitat on rocky hills. Near
Fowlkes Cave it was trapped almost exclusively at
the bases of juniper trees on rocky hills. There seem
to be no records of this species from late Pleistocene
deposits in Culberson Co. or southern New Mexico.

Peromyscus difficilis (J. A. Allen)

Rock Mouse (11916)

Peromyscus difficilis ranges from northern Colo-
rado to southern Mexico, in high desert habitat. Past
records from the Trans-Pecos are all from the flanks
of mountain ranges at relatively high elevations.
Preferred habitat is cliffs, talus, and rocky slopes.
We did not find this species near Fowlkes Cave but
suspect that it may yet survive on the higher lime-
stone hills of the Apache Mountains. Its presence
in the cave fauna suggests conditions essentially
similar to those about the cave today. The rock
mouse has been listed by Harris (1974) from Dry
Cave, Eddy Co., New Mexico, and by Lundelius
(1979) from Pratt Cave, northern Culberson Co.

Onychomys leucogaster (Wied-Neuwied)

Northern Grasshopper Mouse (1 1958)

The northern grasshopper mouse is readily sep-
arated from Onychomys torridus by strongly devel-
oped rather than poorly developed incisor-base cap-
sule and greater length of lower tooth row and teeth
(see Carleton and Eshelman, 1979). Onychomys leu-
cogaster ranges from southern Canada to northern
Mexico, through arid regions of the western United
States. It has a strong preference for sand and sandy
soils, and its abundance in the cave sediments in-
dicates the presence of sandy soils nearby which are
absent at the present time.

At present, Onychomys leucogaster does not oc-
cur near Fowlkes Cave but does live east of the Pecos
River, to the east, and in the extreme western part
of Trans-Pecos Texas. Its absence near the cave
probably results from increased aridity in the post-
Pleistocene. O. leucogaster has been reported from
Pratt Cave, in northern Culberson Co., by Lundelius
( 1 979) and from late Pleistocene cave sites in south-
ern New Mexico (Harris, 1974).

Onychomys torridus (Coues)

Southern Grasshopper Mouse (1 1959)

Three of the six fossil jaws are somewhat damaged
and certain identification is not possible. The other
three are definitely Onychomys torridus. This long-
tailed grasshopper mouse is the present inhabitant

1984

DALQUEST AND STANGL- FOWLKES CAVE LOCAL FAUNA

447

of the area about Fowlkes Cave and most of the
Trans-Pecos. For identifying characters, see fore-
going account. O. torridus ranges from central Ne-
vada southward over the hotter deserts of the west-
ern United States and Mexico.

This species of grasshopper mouse is less confined
to sandy habitat, and is found on the stony soils of
the creosote bush flats and the varied cover of alu-
vial fans, but avoids hills and broken land. The
comparative scarcity of the species in the cave fauna
suggests that it was just invading the area from the
south when the fossil-containing deposit was form-
ing. Lundelius (1979) found both O. torridus and
O. leucogaster in the early Recent deposits of Pratt
Cave, in northern Culberson Co., and postulated
that the relatively temperate McKittrick Canyon,
where Pratt Cave is located, may have served as a
refuge for O. leucogaster into the Holocene. The
fauna of Fowlkes Cave substantiates this, for the
replacement of O. leucogaster by O. torridus in
southern Culberson Co. was underway in the late
Pleistocene.

Sigmodon hispidus Say and Ord
Hispid Cotton Rat (1 1923-1 1926)

Present-day distribution of the cotton rat is from
Nebraska to Panama and Arizona to the Atlantic.
Wherever it occurs, it is, in its proper habitat of
dense grass and weeds, often abundant. It is an im-
portant item in the diet of the bam owl, and this
doubtless accounts for its abundance in Fowlkes
Cave. Cotton rats occur throughout the Trans-Pe-
cos. Some were trapped in tall, dense grass and weeds
along a dry wash less than 1 km from the cave.
Sigmodon, probably S. hispidus, has been recorded
from late Pleistocene cave deposits in Eddy and
Grant cos., New Mexico (Harris, 1974), and the
post-Pleistocene deposits in Pratt Cave, northern
Culberson Co. (Lundelius, 1979).

A related species, Sigmodon ochrognathus Bailey,
is found in the Davis Mountains, and specimens
were taken some 30 mi south of Fowlkes Cave. In
the Guadalupe Mountains of northern Culberson
Co., S. ochrognathus does not occur, but Microtus
mexicanus, a vole, is found instead. M. mexicanus
is not found in the Davis Mountains. Both species
occupy similar habitat— the scattered, sometimes
extensive, areas of grass, weeds, and succulent vege-
tation at or above the pine-oak belt on the moun-
tains. S. ochrognathus is the ecological vicar of M.
mexicanus, the former a southern form with tropical

affinities, whereas the latter is a northern species
derived from boreal faunas.

It seemed important to determine whether any of
the Sigmodon material from Fowlkes Cave might
represent S. ochrognathus, rather than S. hispidus.
We could find no reliable characters that would sep-
arate jaws or teeth of the two species, and Martin
(1979) reached identical conclusions. Lundelius
(1979) referred all Sigmodon from Pratt Cave to S.
hispidus on the basis of large size. However, al-
though S. hispidus averages larger than S. ochro-
gnathus, and some extremely large S. hispidus are
larger than the largest .S', ochrognathus available,
there is a great deal of overlap in measurements
between the two species. The largest Sigmodon spec-
imens in the Fowlkes Cave collection can only be
S. hispidus but some of the smaller fossils could be
S. ochrognathus.

Because Sigmodon hispidus currently occurs near
Fowlkes Cave, and Microtus mexicanus, a Pleisto-
cene resident of the region, is not known to occur
in sympatry with S. ochrognathus, all the Sigmodon
material is referred to S. hispidus.

Genus Neotoma

Three species of woodrats ( Neotoma micropus, TV.
albigula, and TV. mexicana) occur in the Trans-Pecos
today, and another TV. cinerea (Ord), has been found
in late Pleistocene and early Recent cave deposits
in northern Culberson Co. and southern New Mex-
ico. The four species may be separated as follows.
Neotoma micropus is relatively large. The width of
the second lophid of M, is greater than 1.94 mm
(Dalquest et al., 1969; not a completely reliable
character according to Lundelius, 1979) and there
are no strong dentine tracts on the anteroexternal
side of M, (Lundelius, 1979).

Neotoma albigula is a small woodrat. The second
lophid of Mj usually measures less than 1.94 mm
and there are no dentine tracts on the anteroexternal
side of M, (Lundelius, 1979). Although there is some
overlap in measurements of this character, it prob-
ably serves in the majority of instances to separate
N. albigula from M. micropus. Because a number
of specimens have the breadth of the second lophid
distinctly greater than 1.94 mm, and a number of
others have the breadth distinctly less, it is quite
certain that both Neotoma micropus and TV. albigula
are represented in the fauna.

Neotoma mexicana is the smallest woodrat of the
four species here considered. None of the 20 TV.
mexicana specimens from the Davis Mountains in

448

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

the Midwestern State University collection had the
breadth of the second lophid of M, greater than 1.75
mm, whereas none of the specimens of TV. albigida
examined had the breadth of the lophid less than
1.90 mm. Moreover, in TV. mexicana the M, has
well-developed dentine tracts extending from one-
quarter to one-third of the height from root to top
of crown in the unworn tooth. These are absent or
vestigial in TV. albigula (Lundelius, 1979). This
species can be separated from other Trans-Pecos
woodrats with certainty.

Neotoma cinerea is a large, stout-bodied woodrat,
as known from skulls from northern New Mexico.
The breadth of the second lophid of M, is greater
than 1 .94 mm, and there are well-developed dentine
tracts on the anteroextemal side of M, (Lundelius,
1979). Although TV. cinerea has been reported from
Pleistocene and Post- Pleistocene cave deposits in
northern Culberson Co. (Logan and Black, 1979;
Lundelius, 1979) and southern New Mexico (Harris,
1974), this species was absent in Fowlkes Cave.

There are 30 maxillary fragments, eight lower jaws
lacking M,, and many isolated teeth (11930) that
have been identified only as Neotoma sp.

Neotoma micropus Baird

Southern Plains Woodrat (1 1927)

The geographic range of Neotoma micropus ex-
tends from Kansas southward into northeastern
Mexico, and from the western two-thirds of Texas
westward to Colorado. It is found throughout the
Trans-Pecos in suitable habitat such as desert flats
and the more gentle slopes of aluvial fans. It seems,
in the Trans-Pecos, to avoid the cliffs and talus where
N. albigula lives. It does require cover of cactus,
trees, and yuccas, where it constructs stick nests on
the surface of the ground. It is usually common and
prominent wherever found. It has been recorded
from late Pleistocene and early Recent deposits of
northern Culberson Co. and southern New Mexico
(Harris, 1974; Logan and Black, 1979; Lundelius,
1979).

Both Neotoma micropus and N. albigida occur
about Fowlkes Cave today, and N. micropus is by
far the most abundant of the two, probably because
its habitat, the desert lowlands, is so extensive com-
pared to the rocky cliffs and talus required by N.
albigida. In the collection of fossils, it is N. albigida
that is most common. Today, N. albigida occurs in
rocky areas about the cave, and even has lived in

the cave itself, whereas N. micropus is found no
closer than several hundred meters from the cave
on the aluvial fans below. The relative numbers of
fossils of the two species does not indicate that N.
albigida was more abundant in the late Pleistocene
but rather that N. micropus is a large, heavy rat that
must be near the upper limit of size of prey that a
bam owl can overcome and carry. This is supported
by the fact that only four of the TV. micropus spec-
imens are of adults, and the others are all more or
less immature. In contrast, the bulk of the TV. al-
bigida fossils are of adults.

Neotoma albigula Hartley

White-throated Woodrat (1 1928)

The white-throated woodrat ranges from south-
ern Utah and Colorado southward into central Mex-
ico and from central Texas to California. It occurs
throughout the Trans-Pecos. Favored habitat is
crevices in rocky outcrops, cliffs, and talus slides,
and less commonly in vegetation, such as patches
of dense cactus. There are records of the species
from late Pleistocene caves in southern New Mexico
(Harris, 1974) and Pleistocene and early Recent caves
in northern Culberson Co. (Logan and Black, 1979;
Lundelius, 1979). Neotoma albigida and TV. micro-
pus are often sympatric in the Trans-Pecos, but, as
mentioned in the account of TV. micropus, occupy
mutually exclusive habitat. The presence of TV. al-
bigula in the cave fauna only indicates arid condi-
tions and the presence of rocky habitat nearby.

Neotoma mexicana Baird

Mexican Woodrat (11929)

This is rare in the cave fauna, represented by a
single lower jaw. Neotoma mexicana is primarily a
southern, even tropical, woodrat, but in the north-
ern parts of its range it is found only at higher ele-
vations in mountains. It ranges from the mountains
of northern Colorado southward to Guatemala and
from the Trans-Pecos to Arizona. It occurs in the
Davis Mountains and the Guadalupe Mountains, in
rocky habitat and talus slides. The bulk of its range
in Texas is in the pine-oak belt, but specimens were
trapped in the pinyon-juniper belt 33 km south of
Fowlkes Cave, and it doubtless occurs even closer.
It certainly does not occur at present in the desert
habitat near the cave.

1984

DALQUEST AND STANGL- FOWLKES CAVE LOCAL FAUNA

449

Microtus mexicanus (Sassure)

Mexican Vole (11932)

The Mexican vole is found in alpine habitat from
southern Colorado southward through mountains
of Arizona, New Mexico, and western Texas, and
the highlands of Mexico almost to the Guatemala
border. It enters Texas only in the Guadalupe
Mountains, where it is locally common in grassy
areas and in succulent vegetation at elevations of
7,800 ft or greater (Schmidly, 1977).

The two jaws from Fowlkes Cave are well-pre-
served and identification is confident. The presence
of the fossils is enough to establish the presence of
cool grasslands near the cave in the late Pleistocene.

Erethizon dorsatum (Linnaeus)

Porcupine (11933)

Material includes one incisor fragment. The na-
ture and angle of the occlusal surface and cross-
sectional shape show that a porcupine rather than
a beaver or marmot is involved. Porcupines range
from northern Mexico into Canada and virtually
over the entire United States. The species is com-
mon in the Trans-Pecos, and doubtless lives near
Fowlkes Cave today though we found no specimens.
The tooth probably washed into the cave or was
brought in by woodrats.

Felis tufas Schreber
Bobcat (11935)

Bobcats range over most of temperate North
America southward to southern Mexico. They are
found throughout the Trans-Pecos. There are rec-
ords from late Pleistocene cave deposits in southern
New Mexico (Harris, 1974). The specimen from
Fowlkes Cave is of a kitten, and is the only evidence
of carnivore in the Pleistocene fauna.

Mylohyus sp.

Long-nosed Peccary (1 1936)

Two fragments of tusk were found separately, and
when fitted together formed part of the base of a
peccary tusk. The root canal was almost closed when
the animal died. Wrinklings on the surface of the
fragment closely resemble those on isolated tusks of
Prosthenops from the early Hemphillian Ocote local
fauna of Mexico, and Prosthenops is the presumed
ancestor of Mylohyus. No tusks of Mylohyus are
available for comparison. The cross-sectional shape
of the tusk base is different from that of tusks of
Platygonus or the modem Tayassu.

Capromeryx cf. futcifer (Matthew)

Extinct Antilocaprid (1 1937)

The lower jaw was discovered in place, during
excavation, in approximately the center of the Layer
2 sediments, and was sufficient to establish the Pleis-
tocene age of this layer. The teeth are the proper
size for Capromeryx furcifer but, without the diag-
nostic horn core, identification to species is only
tentative. C. furcifer was presumably a grassland
species, and its presence in the cave suggests the
presence nearby of prairies that are not found in the
near vicinity today.

The specimen consists of a ramus broken across
just in front of the posterior alveolus of P2 and just
back of M3. Much of the ventral border of the jaw
is missing also. P2 is lost but P3-M3 are present and
well-preserved. The posterior part of the root of M3
is exposed, showing the height of the third lobe to
be 30.8 mm. The fossil is of a young-adult animal.
Alveolar length of P3-M3 is 49.1 mm. Antilocaprid
teeth, especially the molars, are wedge-shaped and
increase in length at the occlusal and alveolar levels
with increased age.

STRATIFICATION OF DEPOSITS

Of the 42 taxa listed from Fowlkes Cave, two are
extinct, and a pocket mouse and the fulvous harvest
mouse are only tentatively identified. Thirty-eight
extant species, however, are identified, and the dis-
tribution and habitat requirements of these species
are well known. In view of our knowledge of the
distribution and habitat requirements of these 38
species today, the association of some of the species
found at Fowlkes Cave seems absurd. Nowhere to-

day would one find Dipodomys merriami, Onych-
ornys torridus, and Peromyscus eremicus, forms typ-
ical of hot deserts, living in proximity to such alpine
species as Sorex vagrans, Sorex palustris, and Mi-
crotus mexicanus. The latter forms would be ex-
pected along a trout stream among mountain peaks
or far to the north.

However, the fossils were all found in a single 20-
cm (8-inch) thick layer of sediments sandwiched

450

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

between thicker layers that appeared to be barren
of fossils. The stratigraphic evidence indicates that
these species did live in association.

Harris (1970, 1974) encountered a fauna in Dry
Cave, New Mexico, composed of species very sim-
ilar to those making up the Fowlkes Cave local fau-
na. Because he did not himself collect the fossils first
obtained from Dry Cave, and some of the collec-
tions made later could have been subject to post-
mortum mixing, Harris assumed that the fauna of
Dry Cave represented a mixture composed of cold-
climate elements, derived from a period of glacial
advance, and a warm-climate element derived from
an interstadial time.

The Fowlkes Cave fauna seems not to have been
subject to postmortum mixing. There were tens of
thousands of bones and bone fragments in Layer 2,
sandwiched, between layers of sediment that were
barren of fossils. Had mixing occurred, there would
have been fossils in the layers above and below Lay-
er 2, and there seemed not to have been. The strat-
ification in the Fowlkes Cave sediments is unusual.
Most Pleistocene cave deposits consist of unstrati-
fied red clay rather than stratified layers of limestone
pebbles and fragments.

No record was kept of the level within Layer 2 at
which bones were imbedded. Fossils did seem most
abundant in the lowest part of the layer, and larger
bones, such as limb bones of large rodents, lay at
various angles to each other, although most seemed
to lie parallel to the surface of Layer 2. It seems
impossible that there was appreciable stratification
within Layer 2, and if there was, it would have been
next to impossible to detect. It seems equally un-
likely that appreciable postmortum mixing might
have occurred after Layer 3 was deposited and be-
fore Layer 1 was formed. In short, we think Layer
2 must have been deposited during one relatively
brief period of time.

The stratification of the Fowlkes Cave deposits
may have resulted from climatic conditions. The
finely-divided, sand-like, materials of Layer 3 may
have been formed by more effective shattering of

limestone and spalling of the cave lining in the cold
of a glacial advance. The coarser but still fine ma-
terials of Layer 2 may represent less-fine fragmen-
tation at the terminal part of the glacial advance,
and the coarse materials of Layer 1 might be the
result of erosion during still-warmer climate im-
mediately following the end of the Pleistocene cli-
mate. The Recent, of course, is represented by the
black, surficial silts.

We doubt that there was standing or running water
in Fowlkes Cave that might have stratified sedi-
ments. The cave chambers containing the fossils are
many meters above the level of the aluvial fans
below, and fractures in the limestone would have
drained water away swifty. Today, except for oc-
casional drips, the cave is dry. Snail fossils are only
moderately common in the cave sediments, but a
collection was made. Identification by Dr. Richard
W. Fullington, Dallas Museum of Natural History,
included: Glyphvalini identata paucilirata (More-
let), Hawiia minuscula (Binney), Pupoides albilabris
(Adams), Helicodiscus singleyanus (Pilsbry), Suc-
cinea sp., Gastrocopta pel/icuda hordeacella (Pfeif-
fer), Gastrocopta pentodon (Say), and Gastrocopta
armifera (Say).

Dr. Fullington notes (personal communication)
that all of the listed species are land snails and all
occur in Culberson Co. today, although Gastrocopta
indentata, G. armifera, and Helicodiscus singley-
anus live only in isolated colonies high in the Gua-
dalupe Mountains. Dr. Fullington finds it interesting
that none of the xeric or western species of Oreo-
hi/ix, Holospira, and Ashmunella are represented.

Absence of standing water in Fowlkes Cave seems
highly probable, both because the cave would have
been well-drained, and because none of the snails
in the fauna are aquatic species. The only feature
that could conceivably be responsible for the strat-
ification in the cave is climate. It is assumed that
colder climate caused greater fragmentation of lime-
stone debris and the calcite lining materials in the
cave than occurred during warmer cycles.

TAPHONOMY

Bones of mammals found in caves probably
reached their resting places by four main routes: ( 1 )
carcasses were brought into the cave as prey of pred-
atory birds or mammals; (2) the animals flew, walked,
fell or climbed into the cave; (3) mammals died

nearby and their bones washed or fell into the cave;
(4) bones were brought into the cave as nesting ma-
terial by woodrats.

Fowlkes Cave would seem to be an ideal pitfall
trap, for the opening lies level with the surrounding

1984

DALQUEST AND STANGL- FOWLKES CAVE LOCAL FAUNA

451

rock. However, it seems not to have been an effec-
tive trap for mammals. We did find the mummified
bodies of two snakes on the surface debris. These
had fallen into the cave, perhaps in attempting to
catch cave swallows. However, there were no bones
of deer or other large mammals present on the sur-
face debris, unlike most caves of the sinkhole type.
Further, the Pleistocene fauna is strikingly deficient
in ungulates and carnivores (1 Capromeryx jaw, 1
peccary tooth, 1 bobcat jaw fragment), and bones
of adult rabbits are few. Ungulates, carnivores, and
rabbits are the forms expected to be trapped in a
pitfall cave.

Although the entrance of Fowlkes Cave is large
today, it is possible that it was much smaller in the
late Pleistocene. However, even the early Recent
sediments yielded relatively few remains thought to
have been trapped in the cave. It is more probable
that the steep limestone hillside at the site of the
cave was a place where few ungulates, rabbits, or
carnivores would travel. Jackrabbits, for example,
rarely leave level or gently sloping ground, and the
few jackrabbit bones found in the Pleistocene sed-
iments in the cave may have belonged to a single
individual. Also, the bare rock surrounding the cave
entrance may have made the entrance so obvious
as to warn approaching animals of their danger.
Lastly, just within the opening to the vertical shaft
is a ledge that may have prevented many animals
from falling deeper into the cave trap.

Whatever the reasons, Fowlkes Cave seems not
to have sampled the large mammals that doubtless
lived in the vicinity in the past. The few bones of
larger animals present (antilocaprid jaw, peccary
tusk, porcupine tooth, and perhaps even the marmot
teeth) probably were washed into the cave after they
died nearby, or were brought into the cave by wood-
rats.

In modem times, some mammals, such as bats,
doubtless reside in the cave by day. The mum-
mified body of a Myotis velifer and numerous scat-
tered bones of the same species were found on a bed
of bat guano in a side chamber, and the articulated
skeleton of a ringtail (Bassariscus astutus) was found
nearby. Remains of woodrats were commonly noted
on the surface debris. Ringtails and woodrats, as
well as Peromyscus, are agile climbers and probably
enter and leave the cave at will, and did so in the
past as well. Only the most skillful climbing animals
would be able to scale the vertical walls of the en-
trance shaft.

The bulk of the fossils from Fowlkes Cave were

brought to the cave by barn owls. These owls, after
capturing and devouring prey, retire to sheltered and
protected sites, such as caves or old buildings, to
roost. When the flesh and most connective tissues
of their meal has been digested, the remaining ma-
terial, consisting of bones, teeth and hair, is regur-
gitated and falls to the ground beneath the roost.

Hibbard (1941) noted that owls might have been
the source of fossils he recovered from Pleistocene
terrace deposits in Kansas, and others (for example,
Dalquest et al., 1969; Carleton and Eshelman, 1979)
have attributed the fossils of small mammals found
in Pleistocene caves to barn owls. Evidence suggests
that barn owls brought almost all of the Fowlkes
Cave fossils into the cave. Although no bam owls
were seen in the cave while we were working there,
they resided in the cave in the recent past for the
mummified body of one was recovered from the
surface debris. Doubtless they lived there in the late
Pleistocene as well. The mammal fossils recovered,
with the exceptions noted, are all of small forms,
that could be overcome and eaten by the owls. In
the case of the southern plains woodrat and cotton-
tail rabbits, the bulk of the specimens are of im-
mature individuals. Note that the owls need not
carry the complete carcasses of mammals to the
cave, although they might do so as food for nestlings,
but only eat their prey where captured and return
to the cave with the bones in their stomachs. The
bones would later be deposited in pellets on the cave
floor.

When the fossil-bearing sediments were screen-
washed (see Materials and Methods), small clusters
of bones were often noted when the matrix began
to separate in the water. These clusters of bones were
the size of and almost certainly were ancient owl
pellets. There was no cement holding the bones to-
gether, and apparently they remained together only
because the bones were so interlocked (as in fresh
owl pellets). Most disintegrated immediately, but by
moving swiftly, a few were recovered with only the
outer-most bones falling away. These have been pre-
served (11943, 11944) as fossil owl pellets. Thus,
there is quite strong proof that the bulk of the small
mammal fossils in Fowlkes Cave were brought to
the cave by owls.

The identification of the source of the fossils as
bam owl pellet-derived is important in understand-
ing the composition of the Fowlkes Cave local fauna.
Owls are certainly, to a large extent, opportunistic
hunters and take whatever prey is available. Thus,
one might expect prey species that live in rocky

452

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

habitat, aluvial fans, and on steep hillsides, to be
most common in the fauna, because steep hills and
the aluvial fans are closest to the cave. The creosote
bush flats are farther away, though great areas of
such habitat are within 1 km or so of the cave.
Species that inhabit riparian habitat, however, were
probably 1 0 or more km distant from the cave, even
in the late Pleistocene (see Fig. 4). Thus species that
lived in streamside meadows and marshes (wan-
dering shrew, water shrew, vole) would be expected
to be rare in the cave fauna, as they are.

It is assumed that bam owls prey most heavily
on the small mammals that are most abundant near-
est their roosts. The owls seem to have preference
for certain kinds of prey, at least in any given locality
(see Goyer et al., 1981), or certain species of mam-
mals may simply be easier for the owls to capture

because they tend to move about in the open rather
than under cover. Owls hunting from perches find
the rocky hillsides and aluvial fans offer such perch-
es in abundance, whereas the creosote bush flats are
almost devoid of suitable perches. It is most likely,
however, that the owls fly no farther than they must
to obtain their prey. In times of relative scarcity of
desert rodents, the Pleistocene owls may have flown
to the distant streamside meadows to feed on shrews
and voles, but rarely would waste the time and en-
ergy to fly so far unless it were necessary. We are
informed (personal communication) by Dr. Keith
A. Arnold, Texas A&M University, that bam owls
are known to regularly fly 10 mi or more to hunt,
and to hunt by hawk-like quartering in areas barren
of perches.

PALEOECOLOGY

Thirty-eight of the 42 taxa of mammals identified
fall into three groups: (1) Chihuahuan desert species,
that live near the cave today; (2) alpine types, that
today are found in the Canadian or Hudsonian life-
zones of the Davis, Guadalupe, or southern Rocky
mountains; and (3) species found in both desert and
alpine habitat in Texas and southern New Mexico
(Table 2). The latter forms are of less value in in-
terpretation of paleoecology and paleoclimate at the
cave site.

The placement of some of these species might be
argued. For example, Peromyscus boylii does occur
in sites in the desert where conditions are suitable,
but these may be relict populations. Peromyscus dif-
ficilis is, over much of its range in Mexico, a mouse
of talus slides, cliffs, and rocky sites on the desert.
In Texas and New Mexico, however, records are all
from high elevations.

The desert-species composition from Fowlkes
Cave shows a definite difference from the modem
fauna. Pappogeomys castanops is rare as a fossil,
vastly outnumbered by Thomomys bottae. The re-
verse is true in the modem fauna. It is possible that
the scarcity of Pappogeomys is related to the ability
of bam owls to capture this gopher. Dipodomys or-
dii, rare in the modem fauna near the cave, out-
numbers D. merriami (107 to 14) in the fossil fauna.
D. merriami is abundant about the cave at present.
The abundance of D. ordii is thought to indicate the
presence of sandy soils nearby in the late Pleisto-
cene. Onychomys leucogaster, not found in most of

the Trans-Pecos today, is the common grasshopper
mouse in the fossil fauna, outnumbering O. torridus
183 to six. Again, presence of extensive sandy soils
in the past is here indicated.

Of the alpine species listed, some would require
a somewhat more temperate climate, with more
green vegetation growing throughout the year, and
presence of more junipers or other trees, before they
could exist on the rocky hillsides and aluvial fans
near the cave. Probably only a moderate increase
in rainfall, more equitably distributed throughout
the year, would permit these species to live in the
vicinity of the cave today. We would include here
Myotis Iucifugus, Sylvilagus floridanus, Eutamias
cincericollis, Marmota flaviventris, Peromyscus boy-
lii, P. difficilis, and Neotoma mexicana. Other than
Marmota flaviventris, these species are known to
live in the Davis or Guadalupe mountains, at no
great distance from Fowlkes Cave.

Marmota flaviventris now occurs no farther south
than the mountains of northern New Mexico, south-
ern Utah, and southern California (Hall, 1981).
There are late Pleistocene records from sites far south
of its present range, including: Arizona (Meade and
Phillips, 1981), southern New Mexico (Harris, 1974),
Trans-Pecos Texas (Logan and Black, 1979), and
Nuevo Leon, Mexico (Stock, 1942). Lundelius(1979)
reported marmot remains found in association with
bone dated by carbon- 1 4 at slightly more than 2,000
years B.P., showing that marmots existed in the
Trans-Pecos until quite late in the Recent age. Even

1984

DALQUEST AND STANGL- FOWLKES CAVE LOCAL FAUNA

453

Table 2 .—Paleoecological affinities of mammalian species iden-
tified in Fowlkes Cave local fauna.

Species

Chihuahuan

desert

Alpine

habitat

Generalized

Sorex vagrans

+

Sorex palustris

+

Notiosorex crawfordi

+

My otis lucifugus

+

Myotis velifer

+

Eptesicus fuscus

+

Sylvilagus floridanus

+

Sylvilagus audubonii

+

Lepus californicus

+

Eutamias cinericollis

+

Marmota flaviventris

+

Spermophilus spilosoma

+

Spermophilus variegatus

+

Cynomys ludovicianus

+

Thomomys bottae

+

Pappogeomys castanops

+

Perognathus hispidus

+

Perognathus flavus

+

Dipodomys ordii

+

Dipodomys merriami

+

Dipodomys spectabilis

+

Reithrodontomys montanus

+

Reithrodontomys megalotis

+

Peromyscus maniculatus

+

Peromyscus leucopus

+

Peromyscus eremicus

+

Peromyscus boylii

+

Peromyscus pectoralis

+

Peromyscus difficilis

+

Onychomys leucogaster

+

Onychomys torridus

+

Sigmodon hispidus

+

Neotoma albigula

+

Neotoma micropus

+

Neotoma mexicana

+

Microtus mexicanus

+

Erethizon dorsatum

+

Felis rufus

+

Totals

17

10

n

a conservative outline of geographic range con-
structed from these records indicates a vastly greater
range in the late Pleistocene, including all of Arizona
and New Mexico, Trans-Pecos Texas, much of Chi-
huahua, and parts of Coahuila, Durango, Zacatecas,
Nuevo Leon, and San Luis Potosi.

The extermination of marmots over so great an
area must have been primarily due to the increased
aridity in post-Pleistocene times. Harris ( 1 970) stat-
ed “If the mountains of Arizona and southern New
Mexico are excepted, marmots occur in the west
where there is sufficient winter and early spring pre-

cipitation to support green plants during the critical
spring-early summer drought period, regardless of
elevation. At present, approximately 2 inches of
winter precipitation seems to be sufficient in the
southern Rockies.” Harris feels that marmots were
eliminated from Arizona and southern New Mexico
(and by implication from Texas and Mexico as well)
by droughts of moderate extent but at critical pe-
riods in the late winter and early spring.

Increasing aridity of the early Recent is doubtless
also responsible for the local elimination of the east-
ern cottontail (Sylvilagus floridanus robust us), chip-
munk, brush mouse, rock mouse, and Mexican
woodrat, as well as the marmot. Some of these species
live today on the slopes of the Davis Mountains, at
elevations higher than those at Fowlkes Cave, and
where green or woody vegetation is more common.
Moisture, not elevation or temperature, appears to
be the limiting factor.

Ecologically, the most unusual members of the
Fowlkes Cave local fauna are the vagrant shrew,
water shrew, and Mexican vole. These forms de-
mand mesic conditions, and the water shrew seems
to be confined to the vicinity of cool streamlets of
clear, swiftly-moving water. Water shrews do enter
marshy areas and damp meadows when these are
not far removed from running water. Wandering
shrews prefer damp meadows and marshes but enter
woodlands if the soil is soft and preferably damp.
The Mexican vole is a mouse of grassy meadows,
preferably of short, dense grass, but enters areas of
taller grass, weedy areas and even the margins of
dense brushland if the shrubs are low. The habitat
demands of these species can scarcely have changed
in the past 10,000 years, and since they are present
in the cave, habitat suitable for them must have
occurred nearby, within the (10-mi) maximum
hunting-range of the bam owl. Because these species
are rare in the fauna, this habitat must have been
near the extreme range of the owls. Because the
habitat was certainly riparian, cold, running water
and fringing damp meadows and willow thickets
must have been present where the dry washes now
exist.

Glaciation of the Guadalupe Mountains occurred,
and the ice must have been melting when the Fowlkes
Cave fauna lived in the vicinity. Only this will ac-
count for the presence of Sorex and voles in the
fauna. Swift streams of cool water might also have
brought sandy terraces into existence, which have
since been eroded away. The Ord kangaroo rat and
short-tailed grasshopper mouse are sand-dwelling

454

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

species, abundant in the Pleistocene fauna. Cool,
swift streams with fringing meadows do not contra-
dict the presence of Chihuahuan Desert habitat back
from the streams. Where permanent-water streams
flow today in deserts, oasis-like conditions com-
monly exist. In some places in the Great Basin, such
sites are refugia for mesic-habitat mammals such as
voles. Nowhere today are found, so far as we are
aware, cool streams through Chihuahuan Desert
where wandering shrew, water shrew, or Mexican
vole are brought into near contact with desert mam-
mals, but somewhat equivalent conditions do exist.

The reconstructed ecological conditions near
Fowlkes Cave during the period when the fossils
were deposited are as follows. Cool streams with
fringing borders of willows and meadowlands ex-
isted where there are now only dry washes, 10 km
or more from the cave. The hillsides and aluvial
fans supported somewhat more vegetation, includ-

ing more junipers and perhaps some other trees. The
creosote bush flats existed as today, but soils were
sandy on terraces closer to the streams. Some of the
typical desert mammals, such as Merriam kangaroo
rat and long-tailed grasshopper mouse, were just
entering the area. Precipitation need not have been
much greater but was perhaps more equitably dis-
tributed. Summer temperatures were probably little
lower than those today and winters could have been
but little cooler. The Merriam kangaroo rat and long-
tailed grasshopper mouse remain active in winter,
and apparently cannot tolerate extended periods of
freezing conditions. The Ord kangaroo rat, at least,
apparently becomes dormant in freezing weather,
and is thus able to survive in the colder deserts of
the north. The scarcity of the Merriam kangaroo rat
and long-tailed grasshopper mouse suggest that
cooler climate of the late Pleistocene was just giving
way to the hot, desert climate of modern conditions.

ACKNOWLEDGMENTS

We are deeply indebted to J. M. Fowlkes for allowing us to
work in Fowlkes Cave, to collect Recent mammals on his ranch,
and for numerous other kindnesses. Edward Matelski is due our
thanks for his aid in collecting the fossil-bearing sediments, and
for his efforts in washing and sorting of the materials collected.
Guy Nelson and Gregory Zolnerowich also aided in the work
involved in collecting and transporting the fossil matrix. Dr.

Robert J. Baker permitted us the use of specimens in The Mu-
seum, Texas Tech University, and Dr. David J. Schmidly of
Texas A&M University made certain pertinent specimens avail-
able. We wish to thank Dr. Norman Homer, Midwestern State
University, for taking the photographs. (Craig Hood and Gregory
Zolnerowich reviewed an earlier draft of this manuscript. Shirley
Neunaber typed the manuscript in its final form.

LITERATURE CITED

Ayer, M. Y. 1937. The archeological and faunal material from
Williams Cave, Guadalupe Mountains, Texas. Proc. Acad.
Nat. Sci. Philadelphia, 88:599-618.

Bailey, V. 1905. Biological survey of Texas. N. Amer. Fauna,
25:1-222.

Carleton, M. D., and R. E. Eschelman. 1979. A synopsis of
fossil grasshopper mice, genus Onychomys, and their rela-
tionships to Recent species. Mus. Zool., Univ. Michigan,
Papers on Paleo., Hibbard Memm. Vol., 7:1-63.

Dalquest, W. W. 1965. A new Pleistocene formation and local
fauna from Hardeman County, Texas. J. Paleo., 39:63-79.

Dalquest, W. W., E. Roth, and F. Judd. 1969. The mam-
malian fauna of Schulze Cave, Edwards County, Texas. Bull.
Florida State Mus., 13:205-276.

Dalquest, W. W., and F. B. Stangl, Jr. 1983. Identification
of seven species of Peromyscus from Trans-Pecos Texas by
characters of the lower jaws. Occas. Papers Mus., Texas Tech
Univ., 9:1-12.

. In press. The taxonomic status of Myotis magnimolaris

Choate and Hall.

Davis, W. B., and J. L. Robertson, Jr. 1944. The mammals
of Culberson County, Texas. J. Mamm., 25:254-273.

Findley, J. S. 1965. Shrews from Hermit Cave, Guadalupe
Mountains, New Mexico. J. Mamm., 46:206-210.

Genoways, H. H., R. J. Baker, and J. E. Cornely. 1979. Mam-
mals of the Guadalupe Mountains National Park, Texas. Pp.
271-332, in Biological investigations in the Guadalupe
Mountains National Park, Texas (H. H. Genoways and R.
J. Baker, eds.), Nat. Park Serv., Proc. Trans. Ser., 4:xvii +
1-422.

Goyer, N., A. L. Barr, and A. R. P. Journet. 1981. Bam Owl
pellet analysis in northwestern Harris County, Texas. South-
western Nat., 26:202-204.

Hall, E. R. 1981. The mammals of North America. John Wiley
and Sons, New York.

Harris, A. H. 1 970. The Dry Cave mammalian fauna and late
pluvial conditions in southeastern New Mexico. Texas J.
Sci., 22:3-27.

. 1974. Wisconsin age environments in the northern Chi-
huahuan Desert: evidence from the higher vertebrates. Pp.
23-52, in Transactions of the symposium on the biological
resources of the Chihuahuan Desert Region (R. H. Waver
and D. H. Riskind, eds.), Nat. Park. Serv., Trans. Proc.,
3:xxii + 1-658.

1984

DALQUEST AND STANGL- FOWLKES CAVE LOCAL FAUNA

455

Hibbard, C. W. ! 94 1 . The Borchers Fauna, a new Pleistocene
interglacial fauna from Meade County, Kansas. Bull. Kansas
St. Geol. Surv., Univ. Kansas Publ., 38:197-220.

. 1949. Techniques of collecting microvertebrate fossils.

Contrib. Mus. Paleo., Univ. Michigan, 8:7-19.

Jones, J. K., Jr., D. C. Carter, H. H. Genoways, R. S.
Hoffmann, and D. W. Rice. 1982. Revised checklist of
North American mammals north of Mexico. Occas. Papers
Mus., Texas Tech Univ., 80:1-22.

Logan, L. E., and C. C. Black. 1979. The Quaternary verte-
brate fauna of Upper Sloth Cave, Guadalupe Mountains
National Park, Texas. Pp. 141-158, in Biological investi-
gations in the Guadalupe Mountains National Park, Texas
(H. H. Genoways and R. J. Baker, eds.), Nat. Park Serv.,
Proc. Trans. Ser., 4:xvii + 1-422.

Lundelius, E. L. 1979. Post-Pleistocene mammals from Pratt
Cave and their environmental significance. Pp. 239-258, in

Biological investigations in the Guadalupe Mountains Na-
tional Park, Texas (H. H. Genoways, and R. J. Baker, eds.),
Nat. Park Serv., Proc. Trans. Ser., 4:xvii + 1-422.

Martin, R. A. 1979. Fossil history of the rodent genus Sig-
modon. Evol. Monogr., 2:1-36.

Meade, J. I., and A. M. Phillips. 1981. The late Pleistocene
and Holocene fauna and flora of Vulture Cave, Grand Can-
yon, Arizona. Southwestern Nat., 26:257-288.

Reichman, O. J., and R. J. Baker. 1972. Distribution and
movements of two species of pocket gophers (Geomyiidae)
in an area of sympatry in the Davis Mountains, Texas. J.
Maram., 53:21-33.

Schmidly, D. J. 1977. The mammals of Trans-Pecos Texas.
Texas A&M Univ. Press, College Station, 225 pp.

Stock, C. 1942. The cave of San Josecito, Nuevo Leon, Mexico.
Bull. Geol. Soc. Amer., 52:1822.

Address: Department of Biology, Midwestern State University, Wichita Falls, Texas 76308.

Present address (Stangl): Department of Biological Sciences, Texas Tech University, Lubbock, Texas 79408.

A LATE PLEISTOCENE MAMMALIAN FAUNA FROM CUEVA
QUEBRADA, VAL VERDE COUNTY, TEXAS

Ernest L. Lundelius, Jr.

ABSTRACT

Deposits from Cueva Quebrada, Val Verde County, Texas,
contain a mammalian fauna consisting of chiropterans, Spi/ogale
sp.. Mephitis mephitis, Canis sp., Urocyon cinereoargenteus, Bas-
sariscus cf. astutus, Arctodus simus, Ammospermophilus in-
terpres, Thomomys bottae, Pappageomys caslanops, Perognathus
sp., Baiomys taylori, Onychomys leucogaster, Peromyscus sp.,
Neoloma sp., Sylvilagus sp., Lepus cf. calif ornicus, Equus cf.
scotti, Equus francisci, cf. Camelops, Navajoceros fricki, Stock-
oceros sp., and Bison sp.

Much of the bone is charred and extensively broken. Only five

specimens show damage that can be readily attributed to car-
nivore activity. Both spiral and rectilinear fractures are present.
Although few cut marks were seen on the bones, the general
breakage pattern is consistent with human activity.

The fauna has only one extant species, Baiomys taylori, that
is not found in the Amistad area today. It suggests a more humid
climate at the time of accumulation of the deposits.

The composition of the fauna indicates a late Pleistocene age
which is confirmed by three radiocarbon dates of 12,280 ± 170
B.P., 13,920 ± 210 B.P., and 14,300 ± 220 B.P.

INTRODUCTION

Cueva Quebrada is located on the north side of a
small tributary canyon to the Rio Grande River
about 2 mi north of the mouth of the Pecos River,
1 8 mi (28 km) west of Comstock, Val Verde County,
Texas, at 29°44'N by 101°24'W (Figs. 1, 2). It is
located immediately east of a large shelter known
as Conejo Shelter, the archaeology of which was
studied by Alexander (1974). The excavations in
Cueva Quebrada were done under the supervision
of Alexander in connection with the Texas Archae-
ological Salvage Project’s (TASP) contract with the

National Park Service in the Amistad Reservoir. It
is recorded in the TASP files as 41 VV 162A.

The archaeological material found at this site was
negligible, but a considerable amount of faunal ma-
terial of Pleistocene age was recovered from the low-
er units. This material provides new information on
the Pleistocene fauna of southwest Texas. The con-
dition of the bones and their distribution in the cave
raises questions as to what were the agent(s) re-
sponsible for bringing the bones into the cave and
what caused the burning.

ACKNOWLEDGMENTS

I thank the following people for assistance in this study: Ms. available his original measurements of Equus metapodials from

Melissa Winans prepared the computer generated maps and scat- Dry Cave, New Mexico; Mr. Elton Prewitt provided information

ter diagrams (Figs. 8-1 1), and made available her measurements on the stratigraphy of the deposits. The figures of the teeth and

on the metapodials of Equus samples from San Josecito Cave, bones were drawn by Mrs. Pam Westerby. My wife Judith Lun-

Chanmng, Texas, and Rock Creek, Texas; Dr. Dee Ann Story delius edited the manuscript. Financial assistance was provided

and Dr. Solveig Turpin helped with finding and interpreting the by the Geology Foundation of the University of Texas at Austin,

available records of the excavation; Dr. Arthur Harris made

MATERIALS AND METHODS

The cave was excavated as an archaeological site according to blackened but with organic material still present; unbumed bones

standard archaeological methods. Provenience data on some of showed no evidence of scorching. All specimens with horizontal

the bones was kept in terms of precise horizontal grid coordinates grid and vertical control were used to construct maps with the

and depth below a datum. Other bone material was recorded only aid of a computer mapping routine to determine any patterns of

from a given square and stratigraphic unit. All specimens were occurrence of burning and/or taxonomic units,

scored for degree of burning— burned was defined as having all The materials listed for each taxon are not necessarily all spec-

the organic material burned which left the bone light blue-gray imens of that taxon, but are those that are most useful for iden-

or white in color and often deformed; charred was defined as tification. The detailed provenience of each specimen is also

456

1984

LUNDELIUS — CUEVA QUEBRADA FAUNA

457

Fig. 1. — Photo of Cueva Quebrada and Conejo Shelter. Cueva Quebrada is the small opening to the right and slightly below the larger
Conejo Shelter.

omitted. A complete list with provenience is available from the
Vertebrate Paleontology Laboratory, Texas Memorial Museum,
University of Texas, Austin.

Fig. 2. — Map showing the location of Cueva Quebrada.

Measurements were taken with dial calipers and a microscope
reticle. Abbreviations are: TMM— Texas Memorial Museum, M
Recent Skeletal collection of Vertebrate Paleontology Labora-
tory, TNHC— Texas Natural History Collection of Texas Me-
morial Museum, TAMU— Texas A&M University. The material
from Cueva Quebrada is catalogued under TMM 41238.

STRATIGRAPHY

The cave is small, measuring 30 ft by 15 ft with
the long axis oriented north-south (Fig. 3). A num-
ber of stratigraphic units were recognized (Fig. 4)
but notes with descriptions of these units are not
available. Layer I at the top was composed of white
limestone dust. It was completely devoid of artefacts
and bone. Layers II through IV were also composed
of limestone dust but were various shades of gray
and brown presumably related to the amount of
finely divided charcoal mixed with the limestone
dust. Layer V was a light yellow color. The strati-
graphic layers were approximately horizontal but

458

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 3. — Map of Cueva Quebrada showing the excavation areas. A-A' is the line of section shown in Fig. 4. One inch equals 6.3 ft.

the surfaces dividing them were frequently uneven.
The cross section (Fig. 4) shows that the surface of

Layers III and IV were eroded in places. All layers
have limestone spalls of various sizes.

AGE

The presence of several extinct taxa in the de- dates (Valastro et al., 1977, 1979). They are as fol-

posits of Cueva Quebrada indicates a late Pleisto- lows:

cene age. In addition there are three radiocarbon

Depth below _ j 41 VV 162- A

datum (feet): y'® a ^

I | West Wall of Stratum Block

LUNDELIUS — CUEVA QUEBRADA FAUNA

459

1984

Fig. 4. — North-south section along line A-A' of deposits in Cueva Quebrada.

460

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

TX-879 12,280 ± 170 on charcoal from units I
and II, 1.04-1.19 m below datum.

TX-880 1 3,920 ± 2 1 0 on one piece of wood from
Unit IC, top of dark layer, 1.37 m below datum,
immediately above a broken femur of Equus
scotti (TMM 41238-317).

TX-881 14,300 ± 220 on one piece of wood .18
m north of TMM 41238-317, 1.37 m below
datum.

These dates are consistent with the late Pleisto-
cene age indicated by the fauna. Two of the dates,
TX-880 and TX-881, overlap at ±1.96 standard
deviations and are not significantly different. The
third, TX-879, does not overlap the other two at
1 .96 standard deviations and is slightly younger. All
the dated samples are from levels that contain abun-
dant bone and would seem to be good indicators of
the age of the bone.

TAPHONOMY

The fauna of Cueva Quebrada is not extensive,
but the condition of the bones and the sparse evi-
dence of human involvement raise questions as to
the agent or agents responsible for bringing the bones
into the cave. Much of the bone is burned, some so
thoroughly as to have lost all organic material and
to have been plastically distorted as is seen in much
of the Bison bone from Bonfire Shelter in the same
region (Dibble and Lorrain, 1968). Burning, es-
pecially in caves, is usually thought of as indicating
human activity. As pointed out by Dibble and Lor-
raine (1968) large accumulations of organic material
undergoing decomposition may generate enough heat
to start combustion. In contrast to the situation at
Bonfire Shelter, where there is evidence that large
numbers of Bison were driven over the cliff and fell
into the cave, the maximum number of large ani-
mals represented by the bones from Cueva Que-
brada is approximately nine, which would not seem
to provide enough organic material to initiate spon-
taneous combustion. There is also no evidence that
all these animals were introduced into the cave at
the same time. On the contrary, the general lack of
articulation and the distribution of burned and
charred bone in the deposits suggests a number of
separate episodes of burning.

An examination of the maps showing the hori-
zontal distribution of burned, charred and unburned
bone for which there was precise data shows little
concentration of burned or charred bones. There is
a slight indication of increased burning and charring
of bones of Equus in the central part of the cave.
This is consistent with the observation of E. Prewitt
(personal communication) that there were no hearths
or fire pits.

The possibility that a packrat nest was ignited
either by natural or human agency should not be
overlooked. These nests are sometimes very large
(up to 2 m in height and diameter) and represent a
considerable amount of fuel (Van Devender, per-
sonal communication).

The agent or agents responsible for bringing the
bones into the cave cannot be identified certainly.
The bones of the small animals, such as rodents,
probably were brought in by owls. They are gener-
ally unbroken, which is characteristic of bones de-
rived from owl pellets. A number of the bones of
the larger animals such as horse are not broken or
the fragments were not widely dispersed and it is
possible to reassemble them. The skull of Arctodus
was burned and then fractured but was found intact.
The thoracic and lumbar vertebrae of Arctodus were
articulated, a foot of Equus scotti was articulated,
and the phalanges of a foot of Stockoceros were
articulated.

Many of the breaks in the long bones show a spiral
fracture pattern indicating breakage as fresh bones.
Some of the horse metapodials show rectilinear frac-
tures. All of the skull material of horse, bison, and
antelope is extensively broken.

The postmortem alteration of bones by humans
and animals has recently been considered by a num-
ber of people (Bonnichsen, 1973, 1979; Brain, 1981;
Haynes, 1980, 1983). Both experimental and field
studies have identified criteria that indicate carni-
vore damage to bones. When these criteria are ap-
plied to the bones from Cueva Quebrada, a few seem
to show features indicating carnivore damage. Three
bones show what appear to be tooth marks. A horse
first phalanx (TMM 4 1 238-322) has conical depres-
sions on both the anterior and posterior faces that
have the shape and size of the canine of a wolf (Fig.
5A). Fragments of the surface bone that were driven
into the depression are still present. The proximal
end of a femur of E. scotti (TMM 4 1238-290) shows
a similar depression on the posterior surface just
below the head (Fig. 5B). These depressions are sim-
ilar to puncture marks made by a wolf on the femur
of a bison figured by Haynes (1983: fig. 1). There
are also grooves that are U shaped in cross section
on the ball of the femur that resemble those pro-
duced by carnivore teeth. If Arctodus was a carni-

1984

LUN DELIUS — CUE V A QUEBRADA FAUNA

461

vore as Kurten (1967) suggests, it is possible that
the depressions on the horse phalanx and femur
were made by the carnassials of that animal. The
protoconid of the M, is approximately the right size
and shape.

Two horse tibias (TMM 41238-152, 412) show
damage resembling that reported by Haynes (1980,
1983) to the tibias of moose by wolves (Fig. 5). In
one (TMM 41238-152), the cnemial crest and the
anterior part of the proximal articular surfaces are
removed. The other (TMM 41238-412) lacks the
proximal end, and the posterior part of the distal
end is removed. No grooves that might have been
made by carnivore teeth are visible on these two
bones.

The distal end of a humerus of Stockoceros has a
large hole in the epicondylar area and a deep conical
depression on the inside of the medial wall of the
epicondylar fossa. This latter depression seems to
have been made by a more slender tooth than the
one that made the depressions in the other two bones.
This suggests that more than one carnivore was in-

ENVIRONMENTAL

The Pleistocene fauna from Cueva Quebrada has
only a few extant taxa that are useful in a paleoeco-
logical analysis. The relatively poor representation
of the smaller animals of the fauna has made species
identification difficult for many taxa. Most of the
extant taxa have a wide distribution in the southwest
today and are still living in Val Verde County. One
exception is Baiomys taylori which is now found in

volved in chewing the bones. Although Arctodus is
the only carnivore represented by skeletal material
in the cave deposits that is large enough to have
made the large depressions, the late Pleistocene fau-
na included large canids (Canis lupus and Canis
dims) which were widespread in North America and
were almost certainly members of the fauna of the
Amistad region. There is no guarantee that any of
these predators were responsible for bringing the
animals into the cave, but they could have scav-
enged the carcasses found there.

All the bones were examined for cut marks, par-
ticularly near the ends, that might indicate human
butchering activity. Only one was found. The prox-
imal end of the horse femur (TMM 4 1 238-290) has
three linear V shaped grooves that resemble grooves
on bones known to have been made by tools (Fig.
5C). Thus the breakage pattern, the surface alter-
ations and the burning, suggest that both man and
animals were responsible for the presence and con-
dition of the bones in the cave.

INTERPRETATION

more mesic areas to the east in Texas and at higher
elevations to the south in Mexico.

The extinct fauna, with two species of horse, an
antilocaprid, and Bison, suggests prevalence of open
country, grasslands or savannah, on the uplands.
The canyons, as now, probably supported brushy
vegetation and provided habitats for many non-
grassland and savannah species.

SPECIES ACCOUNTS

Class Mammalia
Order Chiroptera

Material.— An incomplete rostrum (TMM 4 1 238-476); a num-
ber of isolated teeth.

Discussion. — The bat material is inadequate for
reliable identification.

Order Carnivora
Family Mustelidae

Spilogale sp.

Material — Right edentulous juvenile mandibular ramus (TMM
41238-571).

Description. — The alveoli for all the post incisor
teeth are present. The alveoli for the M , show clearly
that this tooth had two small lateral roots under the

midpoint of the tooth. This distinguishes it from
comparable sized species of Mustela in which the
M i has only one very small median root on the labial
side. The horizontal ramus is very shallow and the
surface of the bone is spongy indicating that it is
from a young animal. The distinction between S.
gracilis and S. putorius cannot be made on the basis
of the material reported here.

Discussion. — The spotted skunk Spilogale gracilis
is widely distributed in Trans-Pecos Texas where it
is usually found in rocky and brushy conditions
(Schmidly, 1977).

Mephitis mephitis (Schreber)

Material. — Left mandibular ramus with M,, roots of P4 alveo-
lus for M, (TMM 41238-178); left mandibular ramus with roots
of P4, M, (TMM 41238-406).

462

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table {. — Dental and mandibular measurements of fossil and Recent specimens of Urocyon cinereoargenteus and Vulpes macrotis.

Specimen

Length

P4

Length

M,

Anterior width
M,

Posterior width
M,

Length talonid
M,

Length

m2

Depth mandible
at midlength
M,

Urocyon cinereoargenteus

Cueva Quebrada

TMM 41238-252

7.2

—

—

—

—

—

12.7

TMM 41238-366

-

11.5

5.5

5.3

4.5

6.1

12.4

Recent

M-1662

7.5

11.6

4.7

4.8

4.7

6.2

11.3

M-998

7.1

13.1

4.9

5.0

4.7

6.3

10.9

M-2063

7.7

11.4

4.9

4.4

4.2

6.6

11.6

M-3453

6.9

12.6

4.8

5.1

5.0

6.7

10.7

M-4822

7.3

12.3

5.3

5.3

4.8

7.4

12.1

M-1200

7.7

12.2

Vulpes macrotis

4.5

4.2

3.8

11.3

M-1412

7.7

11.8

4.7

4.1

3.6

5.1

10.5

M-1269

7.7

11.7

4.5

4.1

3.3

5.6

10.6

M-1910

8.4

12.6

4.8

4.5

3.4

5.6

11.8

Description.— The M, of TMM 41238-178 is
deeply worn especially on the lingual side where the
wear extends below the enamel. The trigonid is long-
er than the talonid and the talonid is relatively nar-
row as in Mephitis and in contrast to Conepatus in
which the talonid is longer than the trigonid and is
very much wider than the trigonid. Specimen TMM
4 1 238-406 has the crown of M , and P4 broken away.
It is assigned to M. mephitis on the basis of size and
the presence of well developed lateral roots under
the center of M,. Mustela vison, which has an M,
slightly smaller than that of M. mephitis, has a very
small central root only on the labial side.

Discussion.— Mephitis mephitis is present today
in the vicinity of Cueva Quebrada. It occupies a
wide variety of habitats in Trans-Pecos Texas but
is least common in rough rocky terrain (Schmidly,
1977).

Family Canidae
Canis sp.

Material. — One first phalanx (TMM 41238-201).

Description. — The phalanx is similar in size and
proportions to those of living specimens of Canis
latrans which is common in Val Verde County to-
day.

Urocyon cinereoargenteus (Schreber)

Material. — Fan of a left horizontal ramus with part of P2, P4,
alveoli for P,, P3 and M, (TMM 41238-252); part of a right
horizontal ramus with M,. M2 alveoli for P4 and M, (TMM
41238-366).

Description. — The specimens from Cueva Que-
brada are similar to U. cinereoargenteus in both size
and morphology. The cuspule directly posterior to
the main cusp of the P4 is relatively larger than that
of Vulpes macrotis. The talonid of the M, is broad
and has equidimensional entoconid and hypoconid,
whereas in V. macrotis the hypoconid is larger than
the entoconid and the talonid is relatively narrower
and shorter (Table 1). The size of the P4 and the
depth of the mandible are closer to the correspond-
ing dimensions of U. cinereoargenteus than to Vulpes
vulpes, which is larger, or to V. macrotis which is
smaller.

Discussion. — Urocyon cinereoargenteus is wide-
spread in Texas today including the Trans-Pecos
region where it is usually found in rocky areas in
association with pinyon-juniper forests (Schmidly,
1977).

Family Procyonidae
Bassariscus cf. astutus (Lichtenstein)

Material. — Left calcaneum (TMM 41238-560).

Description. — The calcaneum of Bassariscus is
distinguishable from those of similar sized animals
such as Mephitis, Conepatus, Spilogale, and Mus-
tela vison on the basis of both size and morphology.
It is smaller than the corresponding bone in Mephitis
and Conepatus and is larger than that of Spilogale.
The proximal articular facet extends posteriorly over
the dorsal surface of the calcaneal tuber, which is
not the case in any of the skunks. The calcaneum

1984

LUNDELIUS — CUEVA QUEBRADA FAUNA

463

of Mustela vison is the same size as that of Bassa-
riscus and also has a proximal articular facet that
extends posteriorly onto the dorsal surface of the
calcaneal tuber but the form of the dorsal articular
facet differs. The part of the dorsal articular facet
that lies on the dorsal surface of the calcaneal tuber
is concave in Bassariscus and convex in Mustela
vison and the anterior portion of the facet is not
abruptly turned vent rally as in Bassariscus astutus.
The calcanei of these two animals can be further
distinguished by the relatively longer pre-proximal
articular facet length in B. astutus. This is approx-
imately one of the half total length of the calcaneum
in B. astutus and one third the total calcaneum length
in M. vison. In all these characters the specimen
from Cueva Quebrada resembles B. astutus.

Discussion. — Bassariscus astutus is widely dis-
tributed in Texas today and is present in the vicinity
of Cueva Quebrada. Dalquest et al. ( 1 969) have not-
ed the general absence of B. astutus from Pleistocene
faunas of central Texas. The one exception is a re-
cord from the Pleistocene Red Fill unit of Longhorn
Cavern, Burnet County, Texas (Semken, 1961).
These authors have suggested that B. astutus has
moved eastward into central Texas in post-Pleis-
tocene time. It is known from a number of Pleis-
tocene faunas in the southwest such as Rancho La
Brea, California (Akersten et al., 1979), Smith Creek
Cave, Nevada (Miller, 1979), Rampart Cave, Ari-
zona (Meade, 1981), Upper Sloth Cave in the Gua-
delupe Mountains of West Texas (Logan and Black,
1979); Burnet Cave and Dry Cave in eastern New
Mexico (Harris, 1977). Except for the Longhorn
Cavern record, the Cueva Quebrada occurrence is
the easternmost Pleistocene record for this species.

According to Schmidly ( 1 977) the most important
factor influencing the distribution of B. astutus is
the presence of rocky areas such as cliffs and can-
yons. These situations were surely present on the
Edwards Plateau during the Pleistocene as they are
today and yet Bassariscus was apparently absent.

Family Ursidae
Arctodus simus (Cope)

Material. — Skull (TMM 41238-72); ten thoracic and lumbar
vertebrae (TMM 41238-249); right femur (TMM 41238-347).

Description.— The Arctodus material from Cueva
Quebrada is being described and compared with
other Arctodus material by Kurten, Lundelius, and
Johnson and the reader is referred to that paper for
more detailed information. The skull is fairly com-

plete but has been burned and is distorted (Fig. 6).
The teeth have been essentially destroyed by the
burning and no characters can be discerned on them.
Some idea of tooth size can be obtained from the
alveoli in the skull. The skull is proportionately
broader across the canines and orbits than in Ursus
as noted by Kurten (1967). This skull is morpho-
logically similar to other known skulls but is some-
what small. In most dimensions it is within the range
reported by Kurten (1967) but others are below the
observed range (Table 2). Some of the dimensions
probably have been affected by the distortion caused
by the burning but there seems to be no doubt that
the skull is at the small end of the size range of this
species. Kurten (1967) presents evidence for con-
siderable sexual dimorphism in Arctodus simus with
the females being noticeably smaller than the males.
On this basis the small size of the Cueva Quebrada
specimen probably indicates that it is a female.

The femur is long and slender with a wide distal
end. Although it is longer than the femur of Ursus
americanus it is not relatively more massive as might
be expected. Kurten (1967) has noted this charac-
teristic of the femur of Arctodus. The femur is slight-
ly small in comparison to other North American
Arctodus simus material. An examination of Table
3 shows that the Cueva Quebrada femur is slightly
shorter than other femora reported by Kurten ( 1 967).

Discussion. — Remains of this large extinct bear
have been found over a large part of North America
(Kurten and Anderson, 1 980) but are not common.
Kurten (1967) has suggested that this large bear was
predominantly carnivorous on the basis of its skull
shape, primarily its short, broad rostrum which is
similar to that of the large cats. If this interpretation
of the mode of life is correct, Arctodus could have
been an important agent in bringing other animals
into the cave (see the section on taphonomy).

Order Rodentia
Family Sciuridae

Ammospermophilus interpres (Merriam)

Material. — Left upper molar (M1 or M2) (TMM 41238-730);
right M3 (TMM 41238-731).

Description. — The upper molar is broadly trian-
gular. The anteroloph does not join the protocone
by turning abruptly posteriorly as in Spermophilus
spilosoma, S. tridecemlineatus, and .S', mexicanus.
The metaconule shows no tendency to fuse with the
metacone and the metaloph does not join the pro-

464

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 5. — Bones showing carnivore and possible human damage. A) Anterior surface of first phalanx of E quus francisci (TMM 41238-
322) showing perforations caused by carnivores; B) proximal end of left femur of E. scotti (TMM 41238-290) showing a perforation
caused by a carnivore tooth; C) ball of left femur of E. scotti (TMM 41238-290) showing possible cut marks from a tool and grooves
from carnivore gnawing; D) right tibia of Equus scotti (TMM 41238-152) with cnemial crest missing; E) right tibia of Equus scotti
(TMM 41238-412) showing damage to distal end.

tocone. The mesostyle was apparently small and has
been largely removed by wear.

The tooth is larger than the molars of Eatamias
cinereicollis, E. quadrivittatus, and Tamias striatus,
and smaller than the molars of the three species of
Spermophilus mentioned above. It is the same size
as the M1 and M2 of three specimens of Ammosper-

mophi/us interpres from Brewster County, Texas
(Table 4). The M3 is rounded in occlusal view, wear
has removed many of the features of the crown. The
anteroloph is close to the metaloph. The upper M1
or M2 is more similar to the upper molars of this
species than to any other sciurid and is tentatively
assigned to it. The M3 is the same size as those of

1984

LUNDELI US — CUE V A QUEBRADA FAUNA

465

Table 2. — Skull measurements of Arctodus simus from Cueva
Quebrada and other North American specimens

Measurements

TMM

41238-72

Observed range
of North
American
Arctodus simus 1

Basal length

343

330-440

Condylobasal

366

352-463

Extreme length

371

373-521

Palatal length

207

185-260

Zygomatic width

246

222-364

Rostral width at canines

101 est.

100-136

Width over M2

118

110-150

Interorbital width
Width over postorbital

112

117-153

processes

Width over postorbital

161

147-205

constriction

107 est.

98-107

Postorbital height

113

128-187

Width of nasal opening

69

76-107

1 From Kurten, 1967, Table 5.

A. interpres, but the extensive wear makes a detailed
comparison impossible. Its assignment to A. in-
terpres is uncertain.

Discussion. —Ammospermophiius interpres is cur-
rently found in the area of Cueva Quebrada in Val
Verde County, where it is an inhabitant of canyons
with bare rocks and cliffs (Schmidly, 1977).

Family Geomyidae

T homo my s bottae (Eydoux and Gervais)

Material. — Right mandibular ramus with incisor, Vi2 and al-
veoli for P4, M„ and M3 (TMM 4 1 238-439); left P4 (TMM 4 1 238-
722); left PJ (TMM 41238-732).

Description.— The anterior lobe of the P4 is longer
than wide and has dentine tracts on both sides. The
posterior lobe of the P4 is oval with dentine tracts
on both sides. The M > is strongly narrowed lingually

Tabie 3 .—Measurements of femora of Arctodus simus from Cueva
Quebrada and other North American specimens.

Measurements

TMM

41238-248

Observed range
of North American
Arctodus simus1

Greatest length

508

513-678

Greatest proximal width

123

122-165

Caput diameter

63

61-77

Least transverse width
of shaft

42

41-62

Greatest distal width
over epicondyles

104

108-137

1 From Kurten, 1967, Tabie 18.

Table A.— Measurements of upper molars of fossil and Recent
specimens of Ammospermophiius interpres.

Specimens

Tooth

Length

(mm)

Width

(mm)

Cueva Quebrada

TMM 41238-730

M1 or M2

1.74

2.03

TMM 41238-731

M3

1.85

1.87

Recent

TNHC 4002

M‘

1.63

1.74

M2

1.63

2.03

M3

1.72

1.90

TNHC 4005

M1

1.74

2.00

M2

1.79

2.05

M3

1.87

1.96

TNHC 4003

M1

1.70

2.03

M2

1.74

2.07

M3

1.85

1.98

to produce a tear drop shape. The dentine tract on
the labial side is much wider than the one on the
lingual side and is the site of a shallow groove on
this side of the tooth. The P4 is bilobed with the
anterior lobe more flattened than in a specimen of
T. bottae from Dona Ana County, New Mexico, and
the posterior lobe is less constricted labially.

Discussion. — This species occurs today in Val
Verde County where it lives under a wide variety
of environmental conditions (Davis, 1960; Schmid-
ly, 1977).

Pappageomys castanops (Baird)

Material. — Right mandibular ramus with incisor and P4 (TMM
41238-353); right mandibular ramus with incisor (TMM 41238-
405); left mandibular ramus with alveoli for incisor, P4, NT 3,
probably a juvenile (TMM 41238-566); upper incisor (TMM
41238-616); right edentulous mandibular ramus with alveoli for
incisor and P4, M,_3 (TMM 41238-553); right mandibular ramus
with incisor, alveoli for P4, M, (TMM 41238-552); four upper
incisors (TMM 41238-407).

Description. —The P4 has dentine tracts on the
antero-external and the antero-internal faces of the
anterior lobe and on both the internal and external
edges of the posterior lobe. The upper incisors are
unisulcate.

Discussion. — Pappageomys castanops occurs
throughout Trans-Pecos Texas (Schmidly, 1977) and
is recorded in the vicinity of Langtry (Hall and Kel-
son, 1959). This species is reported by Schmidly
(1977) to occupy areas of sandy loam with a min-
imum depth of 6 to 8 inches.

466

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 5 . — Measurements of premolars of fossil and Recent spec-
imens of Perognathus.

P

P4

Specimens

Length

(mm)

Pos-

terior

width

(mm)

Length

(mm)

Anterior

width

(mm)

Pos-

terior

width

(mm)

Cueva Quebrada

TMM 41238-675

.73

.85

.91

.73

.88

TMM 41238-738

-

-

.60

.54

.65

Perognath us flavus

TMM M-2666

.685

.93

—

—

—

TMM M-887

.65

.91

—

—

—

TMM M-6059

.65

.85

—

—

—

TMM M-6060

.57

.85

—

—

—

TMM M-6061

.65

.86

—

—

—

TMM M-6062

.685

1.02

—

—

—

TMM M-6063

.73

.89

—

—

—

TMM M-2666

.72

.91

—

—

—

TMM M-6064

—

—

.52

.51

.57

TMM M-6065

—

—

.51

.49

.60

TMM M-6066

—

—

.54

.47

.59

TNHC 6114

.55

.85

.59

.54

.64

TNHC 6115 worn

.73

.86

.55

.46

.60

TNHC 407

.72

.96

.60

.52

.60

Perognathus penicillatus
TNHC 4039 .85

1.02

.76

.55

.73

TNHC 4040

.75

1.09

.73

.57

.75

TMM M- 1994

.98

1.06

.76

.62

.86

TMM M-1993

.91

1.14

.73

.65

.81

Perognathus intermedius
TNHC 2558 .83

1.14

.67

.65

.80

TMM M-3730

.82

1.06

.80

.57

.78

Perognathus nelsoni

TNHC 3256 worn

1.02

1.06

.76

.62

.73

TNHC 3257

.89

1.03

.68

.60

.73

TNHC 3261

.89

1.14

.70

.68

.78

Family Heteromyidae

Perognathus sp.

Material. — Left maxilla with P4 (TMM 41238-675); right man-
dible fragment with P4 (TMM 41238-738); left mandible frag-
ment with P4 (TMM 4 1 238-737); two molars (TMM 41238-71 1,
728).

Description. — The P4 consists of three major cusps,
one anterior and two posterior that form a trans-
verse loph with wear. In size and morphology the
P4 is similar to that tooth in P. flavus (Table 5). It
differs from the P4 of P. penicillata, P. intermedius
and P. nelsoni in being smaller (Table 5).

The two P4’s are somewhat different in size but
both have four cusps. One, TMM 41238-737, is
comparable in size to the P4 of P. penicillatus , P.
nelsoni and P. intermedius. One other, TMM 41238-

Table 6.— Measurements of teeth of Onychomys leucogaster from
Cueva Quebrada.

Specimens

Tooth

Length

Anterior

width

Posterior

width

TMM 41238-741

M,

1.81

.95

1.18

TMM 41238-744

m2

1.55

1.16

1.18

TMM 41238-783

m2

1.65

1.32

1.17

TMM 41238-757

M1

2.04

1.25

1.29

TMM 41238-758

M1

1.85

1.05

1.25

738, is comparable in size to the P4 of P. flavus
(Table 5). The material is inadequate to allow a
positive specific assignment. The different size of
the P4’s suggests that two species might be repre-
sented.

Discussion. — Six species of Perognathus are found
today in Trans-Pecos Texas. Two of them, P. flavus
and P. nelsoni, are known to occur in Val Verde
County and two more, P. hispidus and P. penicil-
latus might have occurred there in the past in view
of their present distributions. The three that fall into
the size category of the Cueva Quebrada specimens,
P. flavus, P. nelsoni, and P. penicillatus occupy a
variety of environments in this region (Schmidly,
1977).

Family Cricetidae
Baiomys taylori (Thomas)

Material. — Right M, (TMM 41238-776).

Description. — The M , is similar to a series of M , ’s
of recent B. taylori from San Patricio County, Texas,
in size and morphology. This is one of the smallest
of the North American cricetid rodents. The size
(length 1.36 mm, maximum width .82 mm) is slight-
ly larger than 4 specimens from San Patricio County,
Texas (length 1.19 to 1.24 mm, maximum width
.76 to .81 mm). The tooth is simple with no acces-
sory cuspules and the major cusps are relatively high
and inclined forward.

Discussion. — This species is not found in the Val
Verde County area today. Its primary distribution
in Texas is the Gulf Coastal Plain and south Texas.
The two records closest to Val Verde County are
Boerne, Kendall County, Texas, and six miles south-
west of Geronimo, Coahuila, Mexico (Hall and Kel-
son, 1959:660). The area of Texas where it occurs
today has a more humid climate than does Val Verde
County. Its presence in the Pleistocene deposits of
Cueva Quebrada implies more effective moisture
for that area in the past.

1984

LUNDELIUS — CUEVA QUEBRADA FAUNA

467

Onychomys cf. leucogaster

Material. -Left M1 (TMM 41238-757); right M' (TMM 41238-
758); right M2 (TMM 41238-744); left M, (TMM 41238-771).

Description. — The teeth all have the simple pat-
tern with high cusps and open valleys that are char-
acteristic of this genus. The assignment of the
Onychomys material from Cueva Quebrada to O.
cf. leucogaster is based on size. The dimensions of
the teeth from Cueva Quebrada are closer in value
to those given by Carleton and Eshelman (1979) for
O. leucogaster than for O. torridus (Table 6).

Discussion. — O. leucogaster is recorded from
northeastern Terrell County and from the Edwards
Plateau to the east (Hall and Kelson, 1959; Schmid-
ly, 1977).

Peromyscus sp.

Material. — Nine left M’’s (TMM 41238-718, 740, 742, 743,
759, 772 through 775); one right M1 (TMM 41238-777); one
right M, (TMM 41238-760); two right M2’s (TMM 41238-778,
779).

Discussion.— The specific identification of indi-
vidual Peromyscus teeth is difficult and unreliable.
The Cueva Quebrada sample shows some variation
in the size and complexity of the Mps which may
indicate the presence of more than one species.

Neotoma sp.

Material — Left mandible fragment with incisor (TMM 41238-
554); right edentulous mandible (TMM 41238-533); two left M’’s
(TMM 41238-724, 746); right M1 (TMM 41238-750); two left
M2’s (TMM 41238-725, 745); left M3 (TMM41238-739); right
M3 (TMM 41238-751); five left M,’s (TMM 41238-726, 747,
753, 727, 748); left M, (TMM 41238-723).

Description.— Three species of Neotoma occur to-
day in Trans-Pecos Texas, N. albigula, N. micropus,
and N. mexicanus (Schmidly, 1977). In addition N.
floridana occurs in eastern Texas up to the edge of
the Edwards Plateau with an isolated record near
Rock Springs in Edwards County (Goldman, 1910)
and N. cinerea occurs in the mountains of northern
New Mexico (Hall and Kelson, 1959). N. cinerea
has been reported in a mid-Holocene fauna from
the southern Guadelupe Mountains of Hudspeth
County, Texas (Lundelius, 1979). Both of these lat-
ter species might be expected to have occurred far-
ther south during the Pleistocene when the climate
was cooler and/or wetter than present.

Neotoma cinerea and N. mexicana can be readily
distinguished from the other three species on the
basis of the presence of a dentine tract on the antero-
external side of the M,. The differentiation of the

Table 1 .—Measurements of teeth of Neotoma sp. from Cueva
Quebrada.

Teeth and
specimens

Length
at base

Width
at base

Width of
second loph

M1

TMM 41238-724

3.28

2.33

—

TMM 41238-746

3.42

2.38

-

M2

TMM 41238-745

2.64

2.05

—

TMM 41238-725

2.68

2.23

—

TMM 41238-735

2.54

2.34

-

M3

TMM 41238-739

1.84

1.83

—

TMM 41238-751

1.95

1.70

-

M,

TMM 41238-748

3.06

1.78

1.78

TMM 41238-753

3.01

1.92

1.92

TMM 41238-747

3.50

1.91

1.91

TMM 41238-726

3.28

1.89

1.89

m2

TMM 41238-723

2.70

1.97

-

other three species on the basis of individual teeth
and mandibles is difficult. Dalquest et al. (1969)
found no qualitative dental characters that separate
N. floridana from N. micropus and the teeth do not
differ in size. Dalquest et al. (1969) reported that N.
albigula and N. micropus can be separated on the
basis of the width of the second loph of the M,.
They report that in all Texas specimens of N. al-
bigula they measured, the width of the second loph
of the M, was less than 1 .94 mm and in N. micropus
this measure was greater than 1.94 mm. The sam-
ples they studied showed no overlap. M. Winans
(pers. comm.) has found that in samples of 72 TV.
micropus and 68 TV. albigula from Texas this mea-
surement showed extensive overlap and indicates
that this character is not reliable when large samples
from varied geographic areas are compared. This is
also true when specimens of both species from Trans-
Pecos Texas are compared.

Four M,’s are available from Cueva Quebrada.
None have the dentine tract on the antero-labial side
which rules out their assignment to either TV. cinerea
or TV. mexicana. The dimensions and qualitative
characters of the Neotoma teeth from Cueva Que-
brada do not allow a specific identification (Table
7).

Discussion. — TV. micropus is known today as a liv-
ing animal in western Val Verde County; TV. albigula
is recorded 7.5 mi ( 1 1 .5 km) to the west of Val Verde
County (Schmidly, 1977) and the nearest known

468

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

occurrence of N. floridana is 80 mi (123 km) to the
east. N. floridana might be expected to have ex-
tended its range farther west during the Pleistocene
but more material will be needed to demonstrate its
presence.

Order Lagomorpha
Sylvilagus sp.

Material. — Two upper molars (TMM 41238-754, 755); two
right P3’s (TMM 41238-756, 789); left P: (TMM 41238-790);
fragment of left maxillary (TMM 41238-618); fragment of left
mandible (TMM 41238-572); proximal ends of two left humeri
(TMM 41238-559, 398); ventral ends of two right scapulae (TMM
41238-585, 562); two partial pelves (TMM 41238-473, 403);
distal ends of two right femora (TMM 41238-622, 648); proximal
ends of two right tibias (TMM 41238-455, 623); distal end of
one right tibia (TMM 41238-626); three right calcanei (TMM
41238-240, 480, 627).

Description. — The material assigned to Sylvilagus
is unmistakable in both size and morphology as to
its generic assignment but is totally inadequate to
permit specific identification.

Discussion. — Two species of Sylvilagus, S. flori-
danus and Y. audubonii occur today in Val Verde
County.

Lepus cf. californicus (Gray)

Material. — Left mandibular ramus with P2, P3 alveolus for M,
(TMM 41238-499); partial right edentulous right maxilla (TMM
41238-590); partial right ilium (TMM 41238-386).

Description.— The material assigned to Lepus cal-
ifornicus is within the size range of modern speci-
mens of that species. The measurements of the den-
tition are P3 L. 3.76 mm, W. 3.30 mm; P4 L. 3.00
mm, W. 3.49 mm. The morphology of the P3 is also
like that of living specimens of L. californicus. Lepus
townsendi has been reported from Pleistocene fau-
nas from Schultze Cave in Edwards County, Texas,
80 km (50 mi) to the east of Cueva Quebrada
(Dalquest et al., 1969), Dry Cave Eddy County, New
Mexico, 419 km (262 mi) to the north of Cueva
Quebrada (Harris, 1970) and Burnet Cave in the
Guadelupe Mountains, 395 km (247 mi) north of
Cueva Quebrada. A comparison of the Cueva Que-
brada specimen with two specimens of L. townsendi
from Minnesota (TMM M-3347) shows that the lat-
ter species has fewer crenulations of the enamel in
the re-entrant angles of the P3. Recent specimens of
L. californicus from Texas have more crenulated
enamel in the re-entrants of the P3. The Cueva Que-
brada material is assigned to L. californicus on this
basis.

Discussion. — L. californicus is found in the area

of Cueva Quebrada today. Pleistocene records of L.
townsendi mentioned above are well to the south of
its southernmost occurrence today and raise the
question as to the extent of its southward extension
at that time. Dry Cave is located at an altitude of
1 ,300 m (4,200 ft) which is 900 m higher than Cueva
Quebrada as well as being farther north. However
Schultze Cave is only 300 m higher than Cueva
Quebrada and is no farther north. If the specimen
from Schultz Cave is L. townsendi there is a real
possibility that it also reached Val Verde County
during late Pleistocene time. Additional material
from more localities will be necessary to settle this
question.

Order Perissodactyla
Family Equidae

The taxonomy of the Pleistocene horses is cur-
rently confused. Although it is usually possible to
assign horse material from one locality to one or
more groups it is difficult to decide on the name or
names that should be applied. This is the case with
the horses from Cueva Quebrada. There are clearly
two horses represented, a large form and a small
one, and most of the material can be assigned to
one or the other on the basis of size.

Equus cf. scotti Gidley

Material.— Three right M2’s (TMM 41238-64, 68, 103); frag-
ments of upper molars (TMM 41238-433); left upper premolar
(TMM 41238-99); right upper premolar (TMM 41238-105); two
right upper molars (TMM 4 1 238-62, 60); two upper molars (TMM
41238-61, 191); right horizontal ramus with P3-M3 (TMM 41238-
2); right P, (TMM 41238-74); left M3 (TMM 41238-102); three
left lower premolars (TMM 41238-65, 83, 66); two right lower
premolars (TMM 41238-104, 278); four right lower molars (TMM
41238-70, 67, 255, 253); two left lower molars (TMM 41238-
197, 258); one unworn lower cheek tooth (TMM 41238-196);
distal ends of two right humeri (TMM 41238-225, 194); right
radius (TMM 4 1 238- 195); distal end of right femur (TMM 4 1238-
37); proximal end of left femur (TMM 41238-290); right tibia
(TMM 41238-152); left tibia (TMM 41238-159); distal ends of
two left tibias (TMM 41238-47, 23); distal end of one right tibia
(TMM 41238-153); right metacarpal (TMM 41238-48); left
metacarpal (TMM 4 1 238-4 1 ); two right metatarsals (TMM 4 1238-
88, 42); three right calcanei (TMM 41238-139, 96, 138); two
right astragali (TMM 41238-79, 158); one left astragalus (TMM
41238-205); left articulated pes with all tarsals and proximal end
of metatarsal (TMM 41238-142 through 151); fifteen first pha-
langes (TMM 41238-322, 90, 303, 305, 109, 14, 160, 304, 202,
301, 45, 193, 17, 302, 169); six second phalanges (TMM 41238-
15, 174,44, 166, 175, 192); three third phalanges (TMM 41238-
316, 461, 256).

Description.— The material assigned to this taxon
is from a heavily built horse about the size of a riding

1984

LUNDELIUS — CUEVA QUEBRADA FAUNA

469

Table 8.— Measurements of teeth of Equus scotti from Cueva
Quebrada.

Specimens and teeth

Measurements

Upper teeth

Length

along

ectoloph

Width
normal to
para-mesostyle

Width
normal to
meso-metastyle

TMM 41238-62 Pm

32.0

28.2 est.

30.6

TMM 41238-105 Pm

30.7

29.9

29.0

TMM 41238-99 Pm

32.1

28.3

31.1

TMM 41238-60 M

27.1

28.3

27.2

TMM 41238-61 M

30.9

25.3

26.2 est.

TMM 41238-191 M

31.4

27.6

28.8

Lower teeth

Anterior

Posterior

Length

width

width

TMM 41238-64 P2

44.4

—

—

TMM 41238-68 P2

37.9

14.3

16.4

TMM 41238-74 P2

34.1

13.3

15.8

TMM 41238-102 M3

31.8

13.7

13.2

TMM 41238-65 Pm or M

31.0

17.2

18.5

TMM 41238-66 Pm or M

31.4

19.0

—

TMM 41238-253 Pm or M

33.7

15.4

14.0 est.

TMM 41238-197 Pm or M

27.7

15.5

14.9

TMM 41238-70 Pm or M

31.5

17.1

15.2

TMM 41238-83 Pm or M

28.2

—

18.6

TMM 41238-255 Pm or M

28.9

14.7

15.0

TMM 41238-104 Pm or M

31.7

17.6

18.0

TMM 41238-278 Pm or M

—

16.5

—

TMM 41238-67 Pm or M

33.7

15.1

14.0

TMM 41238-2 P3

30.3

15.5

17.7

TMM 41238-2 P4

27.0

18.3 est.

18.0

TMM 41238-2 M,

27.1

16.7

—

TMM 41238-2 M2

26.8

1 5.6 est.

15.0

TMM 41238-2 M3

-

-

14.0

horse. The upper premolars and molars are large,
very close to the dimensions given by Dalquest
(1964) for specimens assigned by him to E. scotti
from three localities in north Texas including the
type locality in Briscoe County, Texas. The enamel
pattern is comparable in complexity (Fig. 7). The
three specimens figured by Dalquest show some
variation in the fossettes and the development of
the pli caballin. The latter tends to be better devel-
oped on the premolars in the three specimens illus-
trated by Dalquest (1964) and the same is true of
the Cueva Quebrada material.

The lower premolars and molars are also large
and high crowned when unworn (Table 8). All the
P3-M3’s have open U-shaped linguaflexids with flat-
tened metastylids (Fig. 7). In the one specimen in
which the position of each tooth can be established
(TMM 41238-2), the ectoflexids of P3 and P4 are

blunt and do not enter the metaconid-metastylid
isthmus. The ectoflexids of M, and M2 are narrower
than those of P3 and P4 and extend into the meta-
conid-metastylid isthmus. The ectoflexid of M3 ex-
tends to the level of, but not into, the isthmus.

The postcranial material that can be assigned to
this taxon with some confidence consists of limb
and foot bones. Scatter diagrams show the meta-
podials match those from the type locality of E.
scotti in Briscoe County, Texas, in both length and
width (Figs. 8, 9). The horse phalanges from Cueva
Quebrada fall into two size groups. The larger size
group agrees with phalanges of E. scotti and are
assigned to that taxon (Tables 9-12).

Discussion. — There is little agreement on the tax-
onomy of the large late Pleistocene horses. In the
southern Great Plains region these horses have been
assigned to a number of species. Stock and Bode
(1937) assigned large horse teeth from Blackwater
Draw, New Mexico, to E. exceisus. Quinn (1957)
recognized E. caballus cabaUus from two localities
(Blackwater Draw, New Mexico and Lubbock, Tex-
as), E. caballus from Blackwater Draw and E. mid-
landensis from three localities (Blackwater Draw, J.
O. Baggett Ranch in Ector County, Texas, and
Scharbaur Ranch in Midland County, Texas). Lun-
delius (1972) assigned the large horse material from
the gray sand unit at Blackwater Draw to E. scotti
on the basis of morphological similarity. Harris and
Porter (1980) assigned much of the larger horse ma-
terial from the Dry Cave, New Mexico fauna and
the Blackwater Draw material to E. niobrarensis
and differentiated it from E. scotti on the basis of
its smaller size.

However, a scatter diagram of length versus trans-
verse width of the distal articular surface of meta-
tarsals of large horses from various late Pleistocene
localities shows a clustering of specimens from
Blackwater Draw, Knox County, Texas, and Cueva
Quebrada (Fig. 9). This cluster includes the type of
E. scotti and other specimens from the type locality
in Briscoe County, Texas. Part of the Dry Cave
sample of metatarsals assigned to E. niobrarensis
by Harris and Porter ( 1 980) plot with the specimens
from Rock Creek, Blackwater Draw, Scharbaur
Ranch, and Cueva Quebrada and are considered to
be the same taxon. One Dry Cave specimen (UTEP
31-64) assigned to E. scotti by Harris and Porter
(1980) on the basis of its size, is somewhat longer
(304 mm). Although it plots outside the cluster of
the other large horses, it would not extend the range
beyond expectation if the total ranges of the other

470

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 6. — Dorsal view of skull of Arctodus simus (TMM 41238-72) from Cueva Quebrada. x.5.

Fig. 7.— Teeth of Equus scotti from Cueva Quebrada. A) Right P2-M3 (TMM 41238-2); B) right upper premolar (TMM 41238-105); C)
left upper molar (TMM 41238-61); D) left upper molar (TMM 41238-191); E) right upper molar (TMM 41238-62); F) right lower
molar (TMM 41238-67). x.75.

1984

LUNDELIUS — CUEVA QUEBRADA FAUNA

471

Fig. 8. — Metapodials of equids from Cueva Quebrada. A) Right third metatarsal of Equus francisci (TMM 41238-46); B) right third
metatarsal of Equus francisci (TMM 4 1 238-227); C) left third metacarpal of Equus francisci (TMM 4 1238-629); D) right third metatarsal
of Equus scotti (TMM 41238-42); E) right third metacarpal of Equus scotti (TMM 41238-48). x.55.

samples shown in Fig. 8 are considered. The plot of
the distal width versus trochlear diameter shows
more heterogeneity (Fig. 10). The Cueva Quebrada
specimens are closely associated with E. scotti.

An examination of the scatter diagram of the length

versus distal width and distal width versus trochlear
diameter of the metacarpals leads to the same con-
clusion, although there seems to be more variation
in the distal width (Figs. 11, 12) and less clear cut
differentiation of the Dry Cave sample. The large

472

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 9.— Measurements of major limb bones of Equus scotti
from Cueva Quebrada.

Proximal

Bones and specimens Length width Distal width

Femur

TMM 41238-37 - - 102

Tibia

TMM 41238-47 - - 80

TMM 41238-23 - - 83 (min.)

Humerus

TMM 41238-255 - - 86

TMM 41238-194 - - 85

Radius

TMM 41238-195 353 88 84

metacarpals from Cueva Quebrada are associated
with those of E. scotti from Rock Creek and the
larger metacarpals from Dry Cave. The teeth of the
large horse from Cueva Quebrada show no signifi-
cant differences in either enamel pattern or size from
those of E. scotti from Briscoe County, Texas (Table
8, Fig. 7).

It is not clear which names should be applied to
the large late Pleistocene horses and a thorough in-
vestigation of this problem is beyond the scope of
this study. E. caballus is inappropriate as it was
applied by Linnaeus (1758) to the domestic horse.
The status of the other species is uncertain. The
oldest name is E. excelsus but the type consists of
a maxilla with P4-M3. The type of E. scotti is a skull,
jaws, and part of a skeleton. Topotype material con-

sisting of several individuals is available. The type
of E. midlandensis consists of left and right man-
dibular rami, P1, P3-M‘, a metatarsal III, and pha-
langes of fore and hind foot believed to come from
one individual.

All the material assigned to the various species
listed above, with the possible exception of E. ex-
celsus, appear to form a relatively homogeneous
group in size and morphology in so far as they are
presently known. The best known type material is
that of E. scotti and the Cueva Quebrada large horse
is assigned to that taxon. If future work demon-
strates that E. excelsus and E. scotti are synony-
mous, E. excelsus will become the proper name.

The type material of E. scotti is from the Irving-
tonian Rock Creek local fauna in Briscoe County,
Texas, which is about .75 million years old. It is
realized that the application of this name to material
of late Rancholabrean age implies the continuity of
one species over this period of time. This is well
within the range of species longevity given by Kur-
ten (1968) for Pleistocene perissodactyls of Europe.

Equus francisci Hay

Material. — Left maxillary fragment with PJ-M’ (TMM 41238-
1 57); right upper premolar (TMM 41238-639); right upper molar
(TMM 41238-100); left upper cheek tooth (TMM 41238-220);
right M3 (TMM 41238-485); right lower cheek tooth (TMM 41238-
75); right ilium (TMM 41238-232); right humerus (TMM 41238-
234); right pelvis (TMM 41238-231 ); two left radii (TMM 41238-
226. 131); distal part of right radius (TMM 41238-285); distal
end of left radius (TMM 41238-313); two right femora (TMM
41238-39, 229); distal half of left femur (TMM 4 1 238-38); distal
ends of two right tibias (TMM 41238-19, 342); distal ends of

Table 10 .—Measurements of metacarpals of Equus scotti from Cueva Quebrada.

Measurements

TMM

41238-48

TMM

41238-41

TMM

41238-13

TMM

41238-78

Length

237

237

231

—

Transverse width of proximal end

58.9

56.2

47.7

—

Antero-posterior width of proximal end
Transverse diameter of posterior articular facet

—

36.8

35.6 (min.)

33.6 (min.)

for the magnum

32.5

33.6

37.0

31.1

Transverse width of facet for magnum

43.9

41.9

46.2

40.7

Transverse width of shaft at midpoint

38.6

38.5

36.7

36.1

Antero-posterior diameter of shaft at midpoint
Transverse width of distal end above the

28.8

29.0

27.4

27.7

articular facet

48.8

51.2

49.2

—

Transverse width of distal articular surface
Antero-posterior diameter of inner distal articular

48.5

51.8

52.4 est.

—

facet

Antero-posterior diameter of central ridge of distal

31.8

35.7

33.1

—

articular surface

39.9

41.5

38.3

-

1984

LUNDELIUS — CUEVA QUEBRADA FAUNA

473

Table 11 .—Measurements of metatarsals of Equus scotti from
Cueva Quebrada.

Measurements

TMM

41238-88

TMM

41238-42

Length

291

289

Transverse width of proximal end

59.5

53.0

Antero-posterior width of proxi-
mal end

52.7

51.2

Transverse diameter of posterior
articular facet

29.3

28.3

Transverse width of facet for
ectocuneiform

51.5

48.9

Transverse width of shaft at
midpoint

40.2

37.1

Antero-posterior diameter of
shaft at midpoint

35.5

33.4

Transverse width of distal end
above the articular surface

51.0

Transverse width of distal
articular surface

50.7

Antero-posterior diameter of
inner distal articular facet

39.4 est.

32.0

Antero-posterior diameter of
central ridge of distal
articular surface

38.0

two left tibias (TMM 41238-207, 208); two right metatarsals
(TMM 4 1 238-46, 227); one left metatarsal (TMM 4 1 238- 1 ); two
left metacarpals (TMM 41238-1 54, 629); distal ends of two meta-
tarsals (TMM 41238-228, 660); proximal end of right metatarsal
(TMM 41238-87); distal ends of two metacarpals (TMM 41238-
40, 49); one left and two right calcanei (TMM 41238-1 18, 457,
165); two right astragali (TMM 41238-52, 312); seven first pha-
langes (TMM 41238-16, 4, 6, 161, 630, 692, 3); distal end of
first phalanx (TMM 41238-425); six second phalanges (TMM
41238-299, 176, 424, 168, 438, 213); one third phalanx (TMM
41238-315).

Description.— The upper dentition is much less
complex than that of E. scotti (Fig. 1 3). The fossettes
of the available teeth have virtually no secondary
plications. The pli caballin is small on both pre-
molars and molars. The protocones of the upper
teeth are short (except for TMM 41238-220) and
most have grooves on their lingual faces. In size and
qualitative characters the upper teeth from Cueva
Quebrada are similar to the dentition of the type of
E. francisci (TAMU 2518) reported by Lundelius
and Stevens (1970).

Only one lower cheek tooth can be confidently
assigned to this taxon (TMM 41238-75). It is small-
er than most of the lower teeth assigned to E. scotti.
Several teeth of the latter species are very close to
the same length (TMM 41238-2, M, and M2, TMM
41238-83, 255, 258) but are noticeably wider (Ta-

Table 12. — Measurements of phalanges of Equus scotti from
Cueva Quebrada.

Phalanges
and specimens

Length

Proximal

width

Distal

width

Mid

width

First phalanx

TMM 41238-90

—

—

46.4

30.5

TMM 41238-301

—

—

43.0

35.7

TMM 41238-322

—

—

44.7

30.1

TMM 41238-45

80.4

58.4

47.9

36.8

TMM 42138-14

81.7

54.3

47.8

35.8

TMM 42138-17

83.0

—

—

35.5

TMM 42138-160

81.1

51.5

45.8

35.6

TMM 42138-202

—

—

—

34.7

TMM 42138-193

74.9

58.5

45.3

33.3

TMM 41238-305

78.2

59.1

45.1

36.6

TMM 41238-302

78.8

58.4

—

36.9

TMM 41238-304

79.5

58.5

46.2

36.2

TMM 41238-169

-

56.9

-

-

Second phalanx

TMM 41238-175

36.7

54.6

—

—

TMM 41238-44

38.7

55.3

51.1

49.0

TMM 41238-166

38.3

47.2

—

41.0

TMM 41238-15

38.6

55.1

50.6

47.6

TMM 41238-192

39.3

54.6

48.7

44.6

TMM 41238-174

40.3

57.5

51.0

47.3

bles 8, 1 3). The Cueva Quebrada specimen is some-
what longer than any of the lower teeth of the type
specimen of E. francisci but is close to the same
width of that specimen (Lundelius and Stevens, 1 970:
table 1). Some of the length difference may be the
result of the difference in the stage of wear of the
two specimens. The Cueva Quebrada specimen is
less deeply worn. Its antero-posterior length at the
midpoint of its crown height is 23.7 mm which is
close to the widths of the lowers of the type of E.
francisci. The ectoflexid just reaches the metaconid-
metastylid isthmus and a small pli caballinid is pres-
ent. Both these characters differ from the type of E.
francisci in which the ectoflexid does not reach the
metaconid-metastylid isthmus and no pli caballinid
is present. The metaconid-metastylid groove is open
but is V-shaped as in the type of E. francisci.

The limb bones are smaller and more gracile than
those of E. scotti and the metapodials are long and
slender (Tables 14-16). A scatter diagram of the
length versus distal width of the metatarsals shows
that two of the three specimens from Cueva Que-
brada are grouped with the type of E. francisci and
specimens of E. quinni (Slaughter et al., 1962) and
a large sample of Equus sp. from Channing, Texas
(Fig. 9). One metatarsal (TMM 4 1 238-227) is closer

474

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 1 3.— Measurements of teeth o/Equus francisci from Cueva
Quebrada.

Measurements

Upper teeth

Specimens and teeth

Length along
ectoloph

Width
normal
to para-
mesostyle

Width
normal
to meso-
metastyie

TMM 41238-157 P4

23.1

24.9

25.2

TMM 41238-157 M1

22.1

24.4

23.5

TMM 41238-157 M2

22.5

23.3

24.9

TMM 41238-157 M-1

23.7

20.1

21.8

TMM 41238-639 Pm

24.3 (min.)

24.8

—

TMM 41238-100 M

21.7

23.2

24.3

TMM 41238-220 M

-

21.1

-

Lower teeth

Anterior

Posterior

Length

width

width

TMM 41238-75 Pm

or M

27.8

1 1.6

12.0

in length to metatarsals from the Slaton fauna as-
signed to E. conversidens by Dalquest (1967), the
Cedazo fauna from Aguascalientes, Mexico assigned
to E. conversidens by Mooser and Dalquest (1975),
the more slender metatarsals from the Dry Cave
fauna assigned to E. conversidens by Harris and Por-
ter (1980) and a large sample from San Josecito
Cave, Nuevo Leon, assigned to E. conversidens by
Stock (1950, 1953). The Cueva Quebrada metatar-
sal is narrower distally than all of the others with

Table 14.— Measurements of major limb bones of Equus franci-
sci from Cueva Quebrada.

Bones and
specimens

Length

Proximal

width

Distal

width

Femur

TMM 41238-39

310

—

80

TMM 41238-229

315

108

82

TMM 41238-38

-

-

83

Tibia

TMM 41238-159

345

81 (min.)

62

TMM 41238-152

357

91 (min.)

—

TMM 41238-208

—

—

65

TMM 41238-207

—

—

59

TMM 41238-342

-

-

52

Humerus

TMM 41238-234

232

80

66

Radius

TMM 41238-226

315

—

65

TMM 41238-130

—

—

69

TMM 41238-180

-

-

67

Table 15 .— Measurements of metacarpals of Equus francisci from
Cueva Quebrada.

Measurements

TMM

41238-628

TMM

41238-154

Length

223

237

Transverse width of proximal end
Antero-posterior width of proxi-

44.4

—

mal end

Transverse diameter of posterior

29.3

—

articular facet for the magnum
Transverse width of facet for

28.7

—

magnum

Transverse width of shaft at

33.3

—

midpoint

Antero-posterior diameter of

27.2

29.4

shaft at midpoint
Transverse width of distal end

24.1

23.7

above the articular facet
Transverse width of distal

39.1

39.1

articular surface
Antero-posterior diameter of

39.7

—

inner distal articular facet
Antero-posterior diameter of
central ridge of distal articu-

27.7

27.2

lar surface

30.4

30.4

the exception of the Cedazo material which is short-
er. The proportions of the distal end (trochlear di-
ameter versus distal width) show much more ho-
mogeneity in the Cueva Quebrada metatarsals (Fig.
10) and plots with the San Josecito sample.

The two small metacarpals do not show the same
amount of variation as the metatarsals. They are
shorter than those from Channing, Texas, and more
slender than those from San Josecito but they are
similar to two from Dry Cave (Figs. 11, 12).

The phalanges assigned to this taxon are slender
and show no evidence of heterogeneity (Table 17).

Discussion. — The taxonomy of the small Pleis-
tocene horses is still uncertain. There is no agree-
ment on the number of species or on the characters
or the limits of variation that might characterize the
different species. All of the small horse material from
Cueva Quebrada indicates the presence of only one
species with the possible exception of the somewhat
short metatarsal (TMM 41238-227). If only one
species is represented the three metatarsals represent
the range of length for this species. This is not ap-
preciably greater than the ranges shown by two other
much larger samples of metatarsals of comparable
sized horses, one from Channing, Texas, and one
from San Josecito Cave, Nuevo Leon, Mexico (Fig.
9). The Cueva Quebrada sample would be inter-

DISTRL WIDTH

LUNDELIUS — CUEVA QUEBRADA FAUNA

475

1984

O

o

en_

in

a

a

in_

o

a

a

o

r^_

''T

a

a

cri-

cn

a

a

in_

on

0 = SAN J05ECIT0
0 = ROCK CREEK
A= BLACKWATER DRAW
+= CHflNNING
X = DRY CAVE
0= E. MIDLANDENSIS
K= E. FRANC I 5C I
X = VV162A - SCOTT I

X= VV162A - FRANCISCI x

iz= SLATON

Z= CEDAZO x

Y = E. QUINNI © $

□ 0

0 0

□ 0
X

□ 0$o

0 0
0

Ittp

o

o

Z

O

O

O

X

o

Ox

X

A

*

+

+ + +
+ +

++ n
+ + + +

t +

+

+ + ++
++

a

o

1 1 — — ‘ 1 1 — i 1 1 1 1 1

°235. 00 242.00 249.00 256.00 263.00 210. 00 277.00 284.00 291.00 298.00 305.00

LENGTH

Fig. 9. — Scatter diagram of length versus distal width of metatarsals of Pleistocene Equus.

Table 16.— Measurements of metatarsals of Equus francisci from Cueva Quebrada.

Measurements

TMM

41238-227

TMM

41238-1

TMM

41238-46

TMM

41238-228

TMM

41238-87

Length

258

284

290

—

—

Transverse width of proximal end

41.8

—

45.8

-

44.6

Antero-posterior width of proximal end

39.8

—

—

—

-

Transverse diameter of posterior articular facet

19.2

—

—

—

—

Transverse width of facet for ectocuneiform

39.8

—

40.0

—

40.6

Transverse width of shaft at midpoint

28.6

27.8

31.3

29.2

—

Antero-posterior diameter of shaft at midpoint
Transverse width of distal end above the articular

28.4

27.7

28.7

30.4

—

surface

38.9

38.5

40.9

41.7

—

Transverse width of distal articular surface
Antero-posterior diameter of inner distal articular

39.6

39.6

41.3

40.5

—

facet

Antero-posterior diameter of central ridge of distal

27.8

29.2

29.4

29.0

—

articular surface

30.0

3 1 .3 est.

33.6 est.

31.4

-

476

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

a:

i—

ai
1 1
Q 1

H= SAN JGSECITO
0= ROCK CREEK
A= BLRCKWRTER DRRW
+= CHRNNING
X = ORT CRVE
0 = E. M I DLRN0EN5 I S
K= E. FRRNCISCI
* = VV162R - SCOTTI
X= VV162R - FRRNCISCI
&= SLATON
Z= CEDRZO
y= E. QUINNI

<»>

X

Xx

X

°^0. 00 24.00 28.00 32.00 36.00

TROCHLEAR DIAMETER

40. 00

44. 00

Fig. 10. — Scatter diagram of distal width versus trochlear diameter of metatarsals of Pleistocene Equus.

mediate in size. If two species are represented the
other small horse material shows no clear evidence
of this. A single species appears more probable and
is adopted here.

Which name should be applied to the Cueva Que-
brada material, is not clear. Lundelius and Stevens
(1970) demonstrated that the type of E. francisci
had long, slender metapodials. In their study of the
Cedazo fauna from Aguascalientes, Mooser and
Dalquest (1975) associated somewhat stockier
metapodials with dentitions they believed were re-
ferable to E. conversidens and metapodials similar
in proportions to those of E. francisci and Cueva

Quebrada specimens were associated with denti-
tions they believed could be assigned to E. tau. If
these associations are correct, E. francisci would be
a junior synonym of E. tau.

Order Artiodactyla
Family Camelidae

cf. Camelops

Material — Distal end of a third or fourth metapodial (TMM
41238-326).

Description. — The form and size of the articular
surface match those of Camelops from other late

1984

LUNDELIUS— CUEVA QUEBRADA FAUNA

477

x

t—

CD

_l ^
CE

tn

0= 5RN JOSECITO
0= ROCK CREEK
A= BLflCKWRTER DRRN
+= CHRNNING
X = DRT CRVE
0= E. H ! DLRN0EN5 I 5
X= E. FRRNCISC1
X= VV162R - SCOT T I
X= VV162R - FRRNC1SC1
■& = SLRTON
Z= CEDRZO
y = E. QUINN1

☆ X-&

0 §

□ &

a

□a

x *

a "sSr

++

++

++

"viiO.'cn 207. 00 21 4. 00 221 . 00 228. 00 235. 00 242. 00 249. 00 256. 00 263. 00

LENGTH

Fig. 1 1. — Scatter diagram of length versus distal width of metacarpals of Pleistocene Equus.

270. 00

Pleistocene localities. The crest on the articular sur-
face is confined to the ventral part. The width of
the articular surface is 40.0 mm, the antero-poste-

Table 17 .—Measurements of phalanges of Equus francisci from
Cueva Quebrada.

Phalanges and
specimens

Length

Proximal

width

Distal

width

Mid

width

First phalanx

TMM 41238-3

72.4

41.7

33.8

25.2

TMM 41238-16

76.3

40.5

35.1

26.3

TMM 41238-6

87.6

45.3

36.6

26.7

TMM 41238-4

81.7

48.2 est.

35.2

27.6

TMM 41238-161

78.1

41.4 est.

34.4

26.2

TMM 41238-630

73.1

43.6

33.7

26.5

TMM 41238-692

86.2

-

-

28.3

Second phalanx

TMM 41238-438

34.6

42.2

39.4

35.2

TMM 41238-299

33.4

37.9

33.7

31.1

TMM 41238-213

34.8

41.0

37.0

32.5

TMM 41238-168

33.8

38.5

37.6

33.5

TMM 41238-176

26.8

36.1

34.1

30.6

TMM 41238-424

-

41.4

-

-

rior diameter is 42.3 mm. This is slightly below or
at the lower end of the size range of Camelops given
by Webb (1965). The differences are slight and are
probably not significant in view of the small size of
the sample available to Webb.

Discussion. — Camelops sp. is a widespread large
camel in late Pleistocene faunas in the western part
of North America (Kurten and Anderson, 1980).
Remains of a large camel, probably Camelops , are
known from Bonfire Shelter in Val Verde County,
Texas (Frank, 1968).

Family Cervidae
Navajoceros fricki Kurten
Material. — Proximal part of a left femur (TMM 41238-362).

Description. — The femur is intermediate in size
between those of Odocoileus hemionus and Cervus
canadensis. The length of the Cueva Quebrada spec-
imen cannot be determined but the length of the
femur given by Kurten (1975), of 328 mm would
not be an impossible estimate. The proximal width
is 89 mm.

478

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

O

o

ai-

in

o

a

in_

in

a

a

in

a

a

r-’_

I ^

cc

cn

i— i CD
□ °

O

o

aL

cn

24. 00

0

O

Q Cfc,

0

*

*

X

0

*

* 2

X

□

X

X

0

>$<

00 a

z

a

g 0

X

ac3 0

+

□

t^++*

0= SRN JQ5ECIT0
0= ROCK CREEK
A= ESLRCKWRTER DRRH
+= CHRNNING
X = DRT CRVE
0 = E. M I DLRNDENS I S
K= E. FRRNCISCI
*= VV162R - SCOTTI
X= VV162R - FRRNCISCI
■&= SLR T ON
Z= CEDRZO
Y= E. QUINNI

28. 00

32. 00 36. 00 40. 00

TROCHLEAR DIAMETER

44. 00

7s. 00

Fig. 12. — Scatter diagram of distal width versus trochlear diameter of metacarpals of Pleistocene Eauus.

Discussion. — This large deer has been reported
from a number of localities in the Rocky Mountains
from Wyoming to Mexico (Kurten and Anderson,
1980). The Cueva Quebrada record is the eastern-
most occurrence known in the United States. Al-
though Val Verde County is located well east of the
Rocky Mountains as such, the rocky, steep terrain
associated with the canyons in the region of Cueva
Quebrada apparently provided the proper environ-
ment.

Family Antilocapridae
Stockoceros sp.

Material. — Fragment of a right frontal with basal part of an-
terior horn core and base of posterior horn core (TMM 41238-
462); right M3 (TMM 41238-1 1); right mandibular fragments
with P4-M2 (TMM 41238-9); right P4 (TMM 41238-466); two
left M,’s (TMM 4 1 238- 1 2, 10); lower incisor (TMM 4 1 238-410);
distal ends of one right and two left humeri (TMM 41238-216,
215, 513); one juvenile left radius with distal epiphysis gone
(TMM 41238-172); distal part of juvenile right radius with epiph-

Fig. 13. — Teeth of E quus francisci from Cueva Quebrada. A) Left PJ-M3 (TMM 41238-157); B) right upper premolar (TMM 41238-
639); C) right lower molar or premolar (TMM 41238-75). x .75.

1984

LUNDELIUS — CUEVA QUEBRADA FAUNA

479

Table 1 8. — Measurements of teeth of Stockoceros sp. from Cueva
Quebrada.

Specimens

Teeth

Length

Anterior

width

Posterior

width

Width of
fourth lobe

TMM 41238-1 1

M3

17.3

9.0

7.8

—

TMM 41238-466

P3

9.7

5.1

4.9

—

TMM 41238-9

P4

9.5

5.0

5.0

—

TMM 41238-9

M,

10.1

4.7

5.4

—

TMM 41238-9

M2

11.5

6.2

6.3

—

TMM 41238-10

m3

—

6.9

6.3

4.8

TMM 41238-12

m3

19.4

6.9

6.8

4.7

ysis gone (TMM 41238-112); proximal part of right radius (TMM
41238-658); proximal end of left ulna (TMM 41238-173); part
of shaft of a left femur (TMM 4 1 238- 171); shaft and distal epiph-
yses of right femur (TMM 41238-5); shaft and distal end of
juvenile right tibia (TMM 41238-107); proximal part of one right
and one left tibia (TMM 41238-286, 642); distal parts of right
and left tibia (TMM 41238-257, 21); distal epiphysis of a left
tibia (TMM 41238-368); two right and one left astragali (TMM
41238-369, 370, 611); proximal end of right metacarpal (TMM
41238-101); distal epiphysis of a metapodial (TMM 41238-2 1 7);
distal end of a metapodial (TMM 41238-641); articulated pair
of first, second and third, phalanges (TMM 41238-7); eight first
phalanges (TMM 41238-222, 54, 111, 356, 245, 58, 110, 602);
distal end of a first phalanx (TMM 4 1 238-395); proximal end of
first phalanx (TMM 41238-123); eleven second phalanges (TMM
41238-291, 508, 56, 55, 57, 1 13, 394, 472, 292, 294, 279).

Description.— The fragment of the frontal bone
has the lower part of the anterior horn core and the
base of the posterior horn core. The anterior edge
of the anterior horn core is broken away so the entire
cross section of the horn core cannot be determined,
but it appears to have been oval as in Stockoceros
onusrosagris (Skinner, 1942). The breakage of the
base of the posterior horn core makes it impossible
to determine the angle between the two horn cores
but it appears to have been approximately that shown
for Stockoceros onusrosagris by Skinner (1942:fig.
10). The prominent pit at the base of the horn cores
at the junction with the supraorbital process stated
by Skinner ( 1 942) to be characteristic of Stockoceros
is discernable on the Cueva Quebrada specimen but
is not obvious because of the breakage of the su-
praorbital process. The supraorbital foramen is lo-
cated immediately anterior to the core base.

The dentition of the Cueva Quebrada antilocaprid
is similar to that of Stockoceros onusrosagris in both
size and morphology (Table 1 8). The M3 has a small-
er metacone that is seen on the M3 of Antilocapra
americana, a point of similarity with S. onusrosa-
gris, and a prominent heel is present on the external

Table 19 .— Measurements of major limb bones of Stockoceros
sp. from Cueva Quebrada.

Bones and
specimens

Proximal width

Distal width

Humerus

TMM 41238-215

—

30.6

TMM 41238-513

-

30.7

Radius

TMM 41238-172

30.4

—

TMM 41238-658

29.3

-

Tibia

TMM 41238-107

—

29.5

TMM 41238-21

—

29.5

TMM 41238-368

—

31.2

TMM 41238-642

48.5

—

TMM 41238-257

-

27.2

Metacarpal

TMM 41238-101

22.6

-

side of the posterior end. This character is variable
in both A. americana and S. onusrosagris (Skinner,
1942). The lower dentition is similar to that of
Stockoceros onusrosagris from Papago Springs Cave.
The P4 consists of two lobes, a large anterior one
and a much smaller posterior one. The posterior
fold is closed by the union of the hypoconid and
entoconid to form a posterior median fossette. The
anterior median fold is closed by the junction of the
paraconid and metaconid to form a median fossette
in the anterior lobe. The form of the two P4's from
Cueva Quebrada is not as molariform as those fig-
ured by Skinner (1942:fig. 16). They are more like
those figured by Colbert and Chaffee (1939:fig. 7).
The M3’s are three lobed teeth with weakly devel-
oped posterior ridges. In neither of these specimens
does the posterior ridge approach the stage of a fourth
lobe shown in some specimens from Papago Springs
Cave (Skinner, 1942). The size of the dentition is
close to that of the Papago Springs Cave sample and
is smaller than that of A. americana (Table 18). The
postcranial material is too incomplete to provide
measurements that would allow comparison with
those of S’, onusrosagris from Papago Springs Cave
given by Roosevelt and Burden (1934), Colbert
and Chaffee (1939), and Skinner (1942). Compari-
son with a skeleton of A. americana shows the Cueva
Quebrada specimens to be somewhat smaller. Mea-
surements of some postcranial elements are given
in Table 19, 20.

Discussion. — Antilocaprids of the genus Stocko-

480

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 20. — Measurements of phalanges of Stockoceros sp. from
Cueva Quebrada.

Phalanges and
specimens

Length

Depth

large

dist.

art.

surface

Proximal

width

Proximal

depth

First phalanges

TMM 41238-58

—

12.5

—

—

TMM 41238-54

40.4

10.1

1 1.2

1 5. 1 (min.)

TMM 41238-356

40.1

12.6

12.0(min.)

15.9

TMM 41238-1 1 1

39.9 (min.) 1 2.3

—

—

TMM 41238-245

39.1

11.5

11.5

14.9

TMM 41238-222

39.9

12.2

11.9

15.7

TMM 41238-7

44.1

11.3

1 1 .7 (min.)

16.8

TMM 41238-7

44.4

12.2

12.3

16.9

TMM 41238-110

—

12.7

13.1

16.0

TMM 41238-123

—

—

13.2

16.8

TMM 41238-602

-

10.7

-

-

Second phalanges

TMM 41238-291

25.1

12.4

—

—

TMM 41238-508

26.9

12.3

1 1.6

14.9

TMM 41238-279

26.2

1 1.6

10.8

14.0

TMM 41238-56

27.1

10.1

10.5

—

TMM 41238-55

25.4

1 1.8

10.9

14.2

TMM 41238-57

24.3

11.0

10.3

1 3.4 (min.)

TMM 41238-394

24.2

12.4

10.7

14.2

TMM 41238-472

27.0

11.6

1 1.4

14.7

TMM 41238-292

26.4

12.8

11.1

14.1

TMM 41238-294

25.0

11.1

10.1

13.4

ceros are known from a number of late Pleistocene
localities in the southwestern United States and
Mexico. Two species S', conklingi and S. onusros-
agris have been named but the differences between
them are slight and Kurten and Anderson (1980)
have suggested that detailed quantitative studies on
all samples may clarify their relationship.

Family Bovidae
Bison sp.

Material.— Two partial horn cores (TMM 41238-22); distal
end of a metapodial (TMM 41238-447); first phalanx (TMM
41238-432).

Description. — The two horn cores, probably from
the same individual are too large to be from a mod-
ern bison. They are the right size for either B. an-
tiquus or B. occidentalis, but the material is too
incomplete to allow a specific identification.

Discussion. — Bison remains are not common in
Pleistocene deposits in Trans-Pecos Texas, but are
common farther north on the Great Plains. Bison
remains are abundant in deposits in Bonfire Shelter
in the same area (Dibble and Lorrain, 1968). It can
be shown there that on at least three occasions Paleo
Indians drove a herd of bison over the cliff and into
the shelter.

LITERATURE CITED

Akersten, W. A., R. L. Reynolds, and A. E. Tejada-Flores.
1979. New mammalian records from the late Pleistocene
of Rancho La Brea. Bull. Southern California Acad. Sci., 78:
141-143.

Alexander, R. K. 1974. The archaeology of Conejo Shelter:
A study of cultural stability at an archaic rockshelter site in
southwestern Texas. Unpublished dissert., Univ. Texas,
Austin, 340 pp.

Bonnichsen, R. 1973. Some operational aspects of human and
animal bone alteration. Pp. 9-24, in Mammalian osteology:
North America (B. M. Gilbert, ed.), Missouri Archaeol. Soc.,
Spec. Publ., Columbia, 337 pp.

. 1979. Pleistocene bone technology in the Benngian

refugium. Nat. Mus. Man Merc. Ser., Archaeol. Surv. Can-
ada Paper, 89:1-297.

Brain, C. K. 1981. The hunters or the hunted? An introduction
to African cave taphonomy. Univ. Chicago Press, Chicago,
365 pp.

Carleton, M. D., and R. E. Eshelman. 1979. A synopsis of
fossil grasshopper mice, genus Onychomys, and their rela-
tionships to Recent species. Univ. Michigan, Papers Pale-
ont., 21:1-63.

Colbert, E. H., and R. G. Chaffee. 1939. A study of Tetra-
meryx and associated fossils from Papago Springs Cave, Son-
oita County, Arizona. Amer. Mus. Novit. 1034:1-21.

Dalquest, W. W. 1964. Equus scotti from a high terrace near
Childress, Texas. Texas J. Sci., 16:350-358.

. 1967. Mammals of the Pleistocene Slaton local fauna

of Texas. Southwestern Nat., 12:1-30.

Dalquest, W. W., E. Roth, and F. Judd. 1969. The mammal
fauna of Schulze Cave, Edwards County, Texas. Bull. Florida
State Mus., Biol. Sci., 13:205-276.

Davis, W. B. 1960. The mammals of Texas. Bull. Tex. Game
and Fish Comm., 27:1-252.

Dibble, D. S., and D. Lorrain. 1968. Bonfire shelter: a strat-
ified bison kill site, Val Verde County, Texas. Misc. Papers,
Texas Memorial Mus., 1:9-138.

Frank, R. M. 1968. Identification of miscellaneous faunal re-
mains from Bonfire Shelter. Pp. 1 33-134 Appendix, in Bon-
fire Shelter: a stratified bison kill site, Val Verde County,
Texas (D. S. Dibble and D. Lorrain), Misc. Papers, Texas
Memorial Mus., 1:9-138.

Goldman, E. A. 1910. Revision of the wood rats of the genus
Neotoma. N. Amer. Fauna, 34:1-124.

Hall, E. R., and K. R. Kelson. 1959. Mammals of North
America. The Ronald Press Co., New York, l:xxx + 1-546
+ 79; 2: viii + 547-1083 + 79.

Harris, A. H. 1970. The Dry Cave mammalian fauna and late
pluvial conditions in southeastern New Mexico. Texas J.
Sci., 22:3-27.

1984

LUNDELIU S — CUE V A QUEBRADA FAUNA

481

. 1977. Wisconsin age environments in the northern Chi-

huahuan Desert: Evidence from the higher vertebrates. Pp.
23-52, in Transactions of the Symposium on the Biological
Resources of the Chihuahuan Desert Region, United States
and Mexico (R. H. Wauer and D. H. Riskind, eds.). National
Park Service. Proc. Trans. Ser., 3:ixxii + 1-658.

Harris, A. H., and L. S. W. Porter. 1980. Late Pleistocene
horses of Dry Cave, Eddy County, New Mexico. I. Mamm.,
61:46-65.

Haynes, G. 1 980. Evidence of carnivore gnawing on Pleisto-
cene and Recent mammalian bones. Paleobiology, 6:34 1 —
351.

. 1983. A guide for differentiating mammalian carnivore

taxa responsible for gnaw damage to herbivore limb bones.
Paleobiology, 9:164-172.

Kurten, B. 1967. Pleistocene bears of North America. 2. Genus
Arctodus , short-faced bears. Acta. Zool. Fennica, 1 17:1-60.

. 1968. Pleistocene mammals of Europe. Aldine Publ.

Co., Chicago, 317 pp.

— . 1975. A new Pleistocene genus of American mountain

deer. j. Mamm., 56:507-508.

Kurten, B., and E. Anderson. 1980. Pleistocene mammals of
North America. Columbia Univ. Press, New York, 442 pp.

Linnaeus, C. 1758. Systema Naturae. 10th ed. Stockholm, v.

I, 824 pp.

Logan, L. E., and C. C. Black. 1979. The Quaternary verte-
brate fauna of Upper Sloth Cave, Guadelupe Mountains
National Park, Texas. Pp. 141-158, in Biological investi-
gations in the Guadelupe Mountains National Park, Texas
(H. H. Genoways and R. I. Baker, eds.), National Park Ser-
vice Proc. Trans. Ser., 4:xvii + 1-442.

Lundelius, E. L., Jr. 1972. Vertebrate remains from the Gray
Sand. Pp. 148-163, in Blackwater Locality No. 1, A strat-
ified early man site in eastern New Mexico (J. J. Hester, ed.),
Publ. Fort Burgwin Research Center, 8:1-239.

. 1979. Post-Pleistocene mammals from Pratt Cave and

their environmental significance. Pp. 239-257, in Biological
Investigations in the Guadelupe Mountains National Park,
Texas (H. H. Genoways and R. J. Baker, eds.), National Park
Service Proc. Trans. Ser., 4:xvi + 1-442.

Lundelius, E. L., Jr., and M. S. Stevens. 1970. Equus francisci
Hay, a small stilt legged horse, Middle Pleistocene of Texas.

J. Paleo., 44:148-153.

Meade, J.J. 1981. The last 30,000 years of faunal history within
the Grand Canyon, Arizona. Quat. Res., 15:31 1-326.

Miller, S. J. 1979. The archaeological fauna of Smith Creek
Canyon. Pp. 273-329, in The Archaeology of Smith Creek
Canyon, eastern Nevada (D. R. Tuohy and D. L. Randall,
eds.), Nevada State Mus. Anthro. Papers, 17:273-329.

Mooser, O., and W. W. Dalquest. 1975. Pleistocene mam-
mals from Aguascalientes, central Mexico. J. Mamm., 56:
781-820.

Quinn, J. H. 1957. Pleistocene Equidae of Texas. Bur. Eco.
Geol., Univ. Texas, Rept. Invest., 33:5-51.

Roosevelt, Q., and J. W. Burden. 1934. A new species of
antilocaprine, Tetrameryx onusrosagris, from a Pleistocene
cave deposit in southern Arizona. Amer. Mus. Novit., 754:
1-4.

Schmidly, D. J. 1977. The mammals of Trans-Pecos Texas.
Texas A&M University Press, College Station, 225 pp.

Semken, H. A. 1961. Fossil vertebrates from Longhorn Cavern,
Burnet County, Texas. Texas j. Sci., 13:290-310.

Skinner, M. 1942. The fauna of Papago Springs Cave, Ar-
izona and a study of Stockoceros with three new antiloca-
p nines from Nebraska and Arizona. Bull. Amer. Mus. Nat.
Hist, 80:143-220.

Slaughter, B., W. W. Crook, Jr., R. K. Harris, D. C. Allen,
and M. Seifert. 1962. The Hill-Shuler local faunas of the
Upper Trinity River, Dallas and Denton Counties, Texas.
Univ. Texas Bur. Econ. Geol., Rept. Invest., 48:1-75.

Stock, C. 1950. 25,000-year-old horse; the skeleton of an ice
age horse makes a return trip to Mexico. Eng. Sci. Monthly,
14:16-17.

. 1953. El caballo pleistocene { Equus conversidens leoni)

de la cueva de San Josecito, Aramberri, Nuevo Leon. Mem.
Congr. Cient. Mexico, 3:170-171.

Stock, C., and F. D. Bode. 1937. The occurrence of flints and
extinct animals in pluvial deposits near Clovis, New Mexico,
Part III— Geology and vertebrate paleontology of the late
Quaternary near Clovis, New Mexico. Proc. Acad. Nat. Sci.
Philadelphia, 88:219-241.

Valastro, S., E. M. Davis, Jr., and A. Varela. 1977. Uni-
versity of Texas at Austin, Radiocarbon Dates XI. Radiocar-
bon, 19:280-325.

. 1979. University of Texas at Austin, Radiocarbon Dates

XIII. Radiocarbon, 21:257-273.

Webb, S. D. 1 965. The osteology of Camelops. Bull. Los Angeles
County Mus., Sci., 1:1-54.

Address: Department of Geoiogical Sciences, University of Texas at Austin, Austin, Texas 78712.

ALASKAN MEGABUCKS, MEGABULLS, AND MEGARAMS:
THE ISSUE OF PLEISTOCENE GIGANTISM

R. Dale Guthrie

ABSTRACT

Pleistocene large mammals from Alaska appear to be larger
than their living counterparts. This seems to be the case for
cervids ( Rangifer , Cervus, and Alces ) and bovids (Ovis. Ovibos.
and Bison). The circumstances of this larger body size are ex-
amined and it is concluded that the primary reason behind such
large bodied, large homed or antlered, individuals was two-fold
(1) they belonged to populations which were kept understocked
by high winter mortality and (2) experienced a long peak in
nutrient availability during the growth season. The length of this
seasonal peak in nutrient availability is seen as critical to studies

of body size changes. Wild sheep are used as an exemplary case
and are examined in greater detail than other species.

Large non-ruminant ungulates seem to respond differently than
ruminants to the changing conditions from mid-Pleistocene, to
late Pleistocene, through the Holocene. This may be due to their
more conservative life histories. Pleistocene ground squirrels
(Spertnophilus) in Alaska seem to have responded differently than
ungulates in their body size changes. Their special life history
features, particularly hibernation and reproduction may account
for this difference.

INTRODUCTION: THE ISSUE OF CAUSES UNDERLYING BODY SIZE CHANGES

We are accustomed to thinking in typological
symbolic images when reconstructing past creatures.
For example, “how did a wooly mammoth look?”
Yet we realize that there was considerable variation
at any one time and particularly through time. How
a wooly mammoth looked depends on which wooly
mammoth one picks and from what time period.
My main emphasis in this paper is that the resources
available to wooly mammoths and their ilk influ-
enced these differences. We see subtle versions of
these influences today even in our own species.
French women bom before World War II, are two
inches (50 mm) shorter than their counterparts bom
in the 1960’s (Zeldin, 1983).

Because we can easily study the differences among
contemporary populations of wild mammals, they
can help us picture how these interactions of avail-
able resources and body size must work through
time. The large black bull buffalo, Syncerus cajfer,
of East Africa is twice the size of its small subspecific
effeminately colored reddish-brown counterpart in
West Africa. The horn size differences are even more
exaggerated. Likewise, the small Florida white-tailed
deer, Odocoileus virgianus, are almost unrecogniz-
able as conspecific to the subspecies which grows to
immense proportions in the northern com-belt
states.

Wildlife managers understand the basic devel-
opmental and ecological reasons why these differ-
ences in individual “quality” exist (for example,
Schmidt and Gilbert, 1978). Animals reared on
overstocked ranges of poor forage quantity and qual-

ity during the growth-seasons are considerably
smaller than the same genotypes reared on high
quality understocked ranges just as with French girls.

One could conclude the intuitively obvious, that
if these differences persist for a long enough time
the balances of natural selections will also be skewed
and genetic changes will ultimately emerge. Poor
growth-season conditions select for smaller animals
and the opposite for good growth conditions. For
large mammals, at least, there is a wealth of data
and experience which supports these conclusions.
Reconstructing what is meant by poor and good
growing seasons as one goes back through time,
however, is no easy task, and what one might in-
tuitively consider poor “quality” growing condi-
tions may not be so to ungulates. This is particularly
true in the cold north.

It is difficult to imagine that glacial conditions, so
severe as to rid Alaska of virtually all trees, would
result in a richer growth season for ungulates than
the warmer wetter conditions of the postglacial, but
that indeed seems to be the case.

Unglaciated Alaska was an arm of Asia during
most of the Pleistocene, and was a peculiarly dif-
ferent place than the now ubiquitous damp spruce
woodlands and wet tundra (Hopkins et al., 1982).
A herbaceous mammoth steppe extended from Eu-
rope across northern Asia to Alaska. The harsh cli-
matic conditions of increased aridity and coolness
along with different seasonal patterns of temperature
and moisture created this xeric herbaceous vegeta-
tion that supported an amazing array of large mam-

482

1984

GUTHRIE- PLEISTOCENE GIGANTISM

483

mal species. In other studies (for example, Guthrie,
1982) I have discussed the paleoclimatic signifi-
cance of these Alaskan large mammal faunas, and
the possible conditions leading to their local and

regional extinctions (Guthrie, 1984). In this study I
wish to look to another dimension, that of body size
changes.

SEASONAL RESOURCE AVAILABILITY AND BODY SIZE

By way of brief introduction to some selection
dynamics of body size, let me say that mammals
and many other organisms have an intrinsic growth
rate potential which tends to level-off or stop at a
predetermined size. Those organisms exhibiting this
“plateaued” pattern of growth are called “deter-
minate.” All ungulates are relatively determinate in
growth pattern, with sexes showing similar growth
rates (males often slightly higher). Female growth
usually reaches a plateau— and slows to a stop—
earlier than the male; this creates a marked sexual
dimorphism among adults. The earlier plateau of
the females is usually due to an adaptation in which
growth energy is normally diverted to reproduction.

The various evolutionary and developmental
forces which affect growth rate and the point at which
determinate growth is reached (body size) are com-
plex and quite contextual. I have argued, however,
that the most important factor is the quality and
quantity of resources during the growth season
(Guthrie, 1984).

There are a number of diverse selection forces
which would increase body size— social advantages,
physiological benefits, predatory or antipredatory
benefits, and others. However, there are costs to
being bigger in a competitive world. The benefits of
larger body size are counterbalanced by absolute
limitations of essential basic resources. The larger
body size variants may experience increased debili-
tation and mortality from insufficient nutrients or
energy, exposure to predation from spending more
time searching for food, and other diverse, contex-
tual stresses in this same vein. When a higher quality
and quantity of food becomes available, animals
usually respond first by moving closer to their max-
imum genetic size potential (see Bell, 1971; Hanley,
1980, for a discussion). Then, given sufficient time,
that genetic potential is extended upward, that is,
there is selection for greater body size.

Thus, the character of the growth season is of
maximum importance. However, the time spent
outside the growth season is important as well, be-
cause this is the time of major adult mortality which,
in turn, influences the degree of competition for the

succeeding growth season’s resources. Reductions
in the available quality and quantity of resources,
either through scarcity or through competition,
would select for a reduced growth rate and/or post-
pone growth to subsequent growth seasons. In-
creased quality of available resources would select
for increased growth rate. We thus end up with sev-
eral potential ways to increase body size in relation
to changing resources (Fig. 1).

There are ultimately three evolutionary avenues
to increased body size. I will argue that one of these
was the major element in ungulate gigantism during
the Pleistocene.

The first avenue is an improvement in food quality
and a concomitant ability to use this “bonus” qual-
ity for a more rapid growth within any one growing
season. The second often accompanies a decline in
food quality (and reduced mortality) and compen-
sates by retaining the ability to growth over more
years; instead of stopping growth at, say, the end of
the second year, it continues on to the fourth. The
third evolutionary avenue to larger body size nor-
mally accompanies a lengthening of the season when
growth resources are available. Thus an animal is
able to grow for a longer time within each annual
growth season. These different patterns are por-
trayed in Fig. 1 and discussed in more detail below.

Large Body Size Due to Higher Peak in
Resource Quality

For herbivores, young vegetation is normally very
palatable, that is, it has low concentrations of fiber,
phytoliths, and antiherbivory compounds. It is also
high in growth materials for the herbivore — soluable
carbohydrates, phosphorus, nitrogen, potassium, and
others — which are in easily digestable forms.

However, the maximum capability of ungulates
to use growth nutrients is usually around or below
the spring peak in vegetational quality, so that ad-
ditional dietary quality during this time is not readi-
ly usable (that is, ungulates have intrinsic volume
and assimilation limits to their capacities as well as
metabolic limits). Plains, mountain, and tundra un-
gulates are generally mobile; they follow early plant

484

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

More Rapid Growth Extended Annual Growth Extended Seasonal Growth

Fig. 1. — There are several different ways to increase body size. From an original condition (lower part of figure) one can either increase
growth rate (Alternative 1), or the same rate can be maintained for more growth seasons (Alternative 2), or the length of the annual
growth season can be extended (Alternative 3). Although growth rates were possibly somewhat greater, the third alternative seems to
be the major factor in explaining the general phenomenon of Pleistocene gigantism. The zones numbered 1, 2, and 3 represent growth
seasons.

growth topographically or geographically— thus ex-
tending the effective duration of the maximum con-
sumption of quality resources. Ungulates can find
peak food quality by pursuing age (phenological)
gradients of the same plant species or taking ad-
vantage of a succession of different plants in the
same location. Most ungulate growth occurs during
this period when plants are at their highest quality.
It is unlikely that the ungulate growth rate could be
greatly enhanced during this time (that is, on sup-
plemental higher quality resources ungulate growth
rates cannot be greatly improved). I am speaking
here about extant high quality, ungulate popula-
tions; there are some populations in every species
for which growth rates are suboptimal even during
this peak resource season because range is poor, or
competition is great.

In summary, there are species in which a higher
peak in food quality increases body size, but most
ungulates benefit only modestly from supplemental
higher quality food during the peak of the growth

season. Higher quality food for ungulates and the
accompanying increased growth rates do not seem
to be the only, or even the major factor in accounting
for the general phenomenon of Pleistocene gigan-
tism. I do not want to discount the importance of
nutrient quality during the growth season in the de-
termination of body size. It is of obvious impor-
tance and is a major element in my model. The
question at hand, however, is whether the peak of
nutrient quality was higher in the Pleistocene. There
are several lines of evidence which suggest that a
higher nutrient peak was not the only or even the
major cause of Pleistocene gigantism. This evidence
is discussed in more detail later.

Larger Body Size Achieved in Additional
Annual Growing Seasons

Most species of vertebrates do use several annual
growth seasons to reach full size; a single growth
season is often not sufficient to attain optimum body
size for the ecological position they occupy. Gen-

1984

GUTHRIE- PLEISTOCENE GIGANTISM

485

erally speaking, a reduction in juvenile and adult
mortality (increased life expectancy) selects for pro-
longed growth. Stated another way, a more conser-
vative, elongated life history usually results in al-
locations of resources into additional growth over
a succession of years. The increased time invest-
ments in larger body size pay off in competitive
ability, in greater stature and reproductive advan-
tages, and in reduced predation. Because individual
productivity can be spread over a longer period of
time, litter size is normally reduced and greater care
is given each offspring.

This phenomenon can be seen for example, among
the elephants (proboscidians), which grow slowly
over many years and in general have a conservative
life history. The same is also true for other major
groups which attain a large body size (for example,
rhinos and cetaceans).

When growth is sustained in additional annual
growth seasons an evolutionary shift to a reduced
rate of growth almost always occurs. There are sev-
eral ways of evaluating whether this was or was not
the case for ungulates in the Pleistocene. First, there
should be a striking change in the survivorship and
growth curves, but (at least among the ruminants)
such change is not apparent. There is no indication
that mortality rates were lower in Pleistocene artio-
dactyls. In fact, several authors (for example, Geist,
1971) have argued that many extant large artiodac-
tyls are shorter-lived than their smaller counterparts
(large Marco Polo sheep, Ovis ammon versus the
smaller bighorn sheep, Ovis canadensis). In com-
parisons of many closely related extant artiodactyl
species, larger species or subspecies body size seems
to be associated with higher adult mortality— quite
contrary to what this second explanation of gigan-
tism would predict. Large forms seem to be larger
at the same individaul age as their smaller counter-
parts (Klein, 1965) suggesting that their larger size
is not due to extending growth over more annual
growth seasons. Also, larger species within genera
or larger subspecies within species mature earlier
(or as early as the smaller forms), suggesting that
their large size is not a more conservative trait. If
we could imagine many of the Pleistocene forms as
chronospecies or chronosubspecies of extant ungu-
late populations, this same principle seems to have
temporal as well as spatial dimension.

In summary, many species do attain large body
size by continuing to grow for additional years and

this has been an important element in some gigan-
tism in the past (proboscidians, for example), but it
does not seem to account for Pleistocene gigantism
or Holocene dwarfing.

Large Body Size Via Longer Growth
Seasons Each Year

Species can become much larger in body size with-
out greatly increasing growth rates or greatly chang-
ing survivorship or mortality curves. They simply
need to be supplied with growth nutrients for a lon-
ger period during each annual growth season. The
growth peak can be extended. Although supple-
menting ungulate diet during their normal growth
peak seldom results in major growth changes, sup-
plemental feeding before and after the normal growth
peak greatly increases annual growth gains without
affecting the growth rate. This indicates that ungu-
lates have the developmental potential for pro-
longed seasonal growth if nutrients are available.

If this extension of the growth season occurred
naturally it would theoretically: (1) allow animals
to reach their genetic maximum potential body size
and (2) select for animals with greater genetic size
potentials.

I should emphasize that it is not just the length
of season which is important but the conjunction of
high nutrient levels maintained over a longer period
of time that is required to maximize body size.

Additionally, heavy competition for survival re-
sources through interspecific or intraspecific com-
petition (outside the growth season) is required to
hold population numbers low enough to insure range
quality (and hence, ungulate quality) during subse-
quent growth seasons. This addendum is a necessary
part of any alternative model of events; it is likely
to have occurred during much of the Pleistocene.
The intensifying winter or dry season which accom-
panied the cooling process, and perhaps also, inter-
specific competition during the non-growth seasons
within the context of the greater species diversity,
would tend to reduce population numbers.

I will focus my discussion of Pleistocene gigan-
tism on Alaskan mountain sheep (Ovis) and, then
to a lesser extent, on bison (Bison), elk (Cervus),
caribou (Rangifer), moose (Alces), and muskoxen
(Ovibos) and discuss why these same principles do
not necessarily apply to such animals as mammoth
(Mammuthus), horses (Equus), and ground squirrels
(Spermophilus).

Length (mm)

NO. 8

486 SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

16 17 18 19 20 21 22 23 24

M3 Length (mm)

Fig. 2. — A comparison of body size in recent Dali sheep (Ovis dalli) using length of molars, illustrating the larger size of the well
preserved Dry Creek specimens (dated at around 1 1,000 B.P.). All of the measurements taken on recent sheep are from mature “trophy
class’’ rams from different areas in Alaska.

FOSSIL ALASKAN DALL SHEEP: IMPLICATIONS FOR
PLEISTOCENE SEASONALITY

My first real awareness that Alaskan fossil sheep
could be larger than their living counterparts came
from studying an archaeological site. Fossil sheep
teeth from the Dry Creek archaeological site (Powers
et al., in press) have been plotted against those of
some trophy ram skulls available to the author. Even
with these few samples, it is apparent that the Dry
Creek sheep were probably larger (Fig. 2).

Additionally, in Figs. 3 and 4 comparisons have
been made of sheep skull characters and horn core
dimensions from other fossil sheep collected in cen-
tral Alaska. They also differ significantly from extant
rams. These fossils came from the Fairbanks mining

district and were removed over the years of peak
mining activity earlier in this century; they now re-
side in the American Museum of Natural History.
The silt sediments from which the fossils were col-
lected are mainly Rancholabrean in age, predomi-
nantly Late Wisconsin Glacial (Guthrie, 1968). The
recent sheep skulls were taken from all over the
state — with a particular bias for large trophy rams —
so the differences between fossil and recent sizes
may even be greater than appear here.

Because the modern treeline in the north is cor-
related with the botanical length of the summer
growing season (number of degree-days), we have

1984

GUTHRIE- PLEISTOCENE GIGANTISM

487

E

E

CD

o

o

c

*_

o

X

CD

o3

E

o

Q

CL

c

CD

E

CD

v_

Z>

CO

O

CD

20

0

100-

90

Dali Sheep Ovis da Hi

A Fossil
• Recent
O Young

▲ ▲
▲

▲ A

A A A *

▲ | A

.** •**. A

#>® A.

* ®.

• A
• — ,

J _l J l

80

60 70 80 90 100

Measurement 12, M-L Diameter of Horncore (mm)

Fig. 3.— A comparison of horn size in fossil and recent Dali sheep {Ovis dalli) from Alaska. The fossils are mainly from the Fairbanks,
Alaska, area, whereas the recent sheep were adult rams collected from numerous places throughout Alaska.

assumed that a predominantly herbaceous flora in
the north and boreal forests in temperate latitudes
could only be achieved by shortening the present
growth season. Treeline in central Alaska now av-
erages about 1,000 m above sea level. Pollen and
macrofossil evidence indicates the absence, or vir-
tual absence, of spruce ( Picea ) during the late Wis-
consin Glacial in the same area (Matthews, 1974;
Ager, 1975). This means a shift in treeline of at least
some 870 m (Fairbanks, Alaska, is 1 33 m above sea
level).

In fact, the reduction of trees in Alaska and the
Yukon Territory during the Wisconsin Glacial (as
in the case of the Holocene reduction of the boreal
forests over what is now the Great Plains) is prob-
ably not a simple matter of length of summer grow-
ing season. It may be more closely controlled by

additional factors such as seasonal moisture changes,
like the aridity which now controls the southern
border of the boreal forest (La Roi, 1982).

If Alaskan Pleistocene plant communities were
known in specific detail we could approach the mat-
ter of climate and seasonality directly. However pol-
len is difficult to identify to lower taxonomic levels
for the non-woody plants— the very ones, I contend,
we would have to know about. Macrofossil plant
identifications have yet to be done at a scale which
would be of much assistance. I also contend that
physical indicators in the fossil record of mean tem-
perature, that is, permafrost and ice wedge forma-
tion, may also be relatively uninformative about
seasonality. The problem has to be approached more
obliquely.

In the north most large mammals grow only dur-

488

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

E

E

I70r

Dali Sheep Ovis dalli

60

50-

140

c

o

O

v_

O

v_

a>

c

<

GO

c

CD

£ 130

a>

V_

ZJ

if)

o

(D

20

A Fossil
• Recent
O Young

•°?

X

▲

▲

▲ •

• A

A ▲ A • •

• A A

• A

A

▲

▲

1 I L

J I L

J I L

74 76 78 80 82 84 86 88 90 92

Measurement 9, Posterior Cranial Width (mm)

94

Fig. 4. — A comparison of body size in recent Dali sheep using skull measurements as an index. The recent sheep are adult rams collected
from all over Alaska. The fossil sheep came mainly from the Fairbanks, Alaska, area.

ing a relatively short time of the year and experience
growth dormancy during the remaining portion
(Norden et al., 1968). During this latter period they
do not have the potential for growth; even when
given high quality food they continue to lose weight.
Because of this winter dormancy, actual body size
can therefore serve as a rough index to summer
range — the duration and quality of summer growth
resources— given, of course, that one does not use
body size out-of-context of ecological settings, or
compare forms which are not closely related.

To avoid as many problems as possible with the
latter two issues my main analysis will use North
American mountain sheep. There are two species
in North America Ovis canadensis and Ovis dalli.
Considerable information exists about their biology,
although the biology of these species is far from
being totally understood. The most important ad-
vantage of using sheep, however, is that we have
fossils from many different restricted localities where
there are still existing sheep populations. Sheep, un-
like bison and wapiti, are not good colonizers (Geist,

1984

GUTHRIE- PLEISTOCENE GIGANTISM

489

1971), nor are they given to erratic movements or
long migrations, so it is likely tnat the fossil sheep
lived in areas near where their remains were found
and that they are related to those sheep now living
in the same area.

There are a number of variables which affect size
and some of these are species-specific, such as the
physical constraints on an insectivorous hibernating
bat or a weasel which has to follow a mouse’s tunnel.
There are no obvious outside physical constraints
of this nature on the American sheep. They have a
broad thermal-neutral zone, so it is not likely that
temperature is an important metabolic size regu-
lator. Likewise, they do not depend directly on size
to fend off or escape predators. However, size of
body and horns are, as with many ungulates, very
important socially (Geist, 1971). Size is also directly
affected by nutrition, both ontogenetically and phy-
logenetically. Generally, one can argue convincingly
for a balance between the social advantages of large
size versus the metabolic expenses of rapid growth
and early reproductive involvement (Geist, 1 966);
and that these forces determine the optimum bal-
ance of size “strategy” for any point in time.

The resource or nutritional determinant has sev-
eral faces. One is sheer quantity of herbage available,
the other is its quality, and still another is the span
of time it is available. In conjunction one must con-
sider also the severity of the winter bottleneck. I
would like to propose the following hypotheses as
part of a general theory accounting for Pleistocene
gigantism in American mountain sheep:

1 . Competition during the peak of the growth sea-
son is not the ultimate force determining body
or horn size.

2. During the peak of the growth season, body and
horn growth are not limited by quality of suitable
herbage.

3. Body and horn size are also not limited during
tie major part of the growth season by quantity
of herbage.

4. The physiological-nutritional debts accumulated
during the winter are not the major determinants
of body size.

5. The colonization phenomenon (Geist, 1971) does
not of itself explain body size trends in fossil
North American sheep.

6. The major variable in body size, both past and
present, in North American sheep is the duration
of the peak of high quality resources — the length

of season during which the sheep can grow at a
maximum rate.

As with any life history feature there are neces-
sarily a number of interlacing variables. All of the
six items referred to in the above hypotheses prob-
ably do affect body size, some more than others, in
different situations and for different reasons. What
I have tried to do conceptually is show that the sixth
hypothesis is the principal component to the geo-
graphic and historical (chronoclinal) variations in
body and horn size in North American sheep. I will
next discuss the items individually.

1. Competition during the peak of the growing
season is not the ultimate force determining body or
horn size. Sheep numbers and body quality seem to
be density dependent (Heimer and Smith, 1975), so
intraspecific competition is certainly important in
sheep biology. Competition is particularly signifi-
cant for northern sheep on winter ranges which usu-
ally are quite discrete and restricted and it seems
here to be the fundamental density regulator (Geist,
1971). Also first green-up often occurs in limited
areas and it may be that on very poor early spring
or autumn ranges intraspecific competition does play
a role. Once into the full flush of spring and summer,
however, it is unlikely that even a highly selective
grazer like sheep meets with intraspecific competi-
tion (Whitten, 1975). The summertime dearth of
feeding displacement behavior which is common
among Dali sheep in winter also points to reduced
competition during the growth season peak.

2. During the peak of the growth season, body
and horn growth are not limited by quality of suit-
able herbage. Sheep select early seasonal growth
stages of plants when available. These are more pal-
atable and more digestable than later growth stages
(Oleberg, 1956; Stoddart and Smith, 1955). Klein
(1970) has reviewed the adaptations of alpine plants
which enable them to grow rapidly and exhibit an
unusually high nutrient content. Whitten (1975)
showed that most plants in the spring and summer
diets of Mt. McKinley Park sheep had protein per-
centages (dry weight) of 25-30%. Compare this to
high quality domestic alfalfa at only 1 6% or rolled
oats at only 1 1% protein (Church, 1972). There are
upper limits to the amount of protein sheep can
digest. As domestic sheep are about the same size
as wild sheep and have about the same intrinsic
growth rate we can make some very general com-
parisons. Protein percentages of over 1 1% are con-
sidered adequate for optimum domestic ram growth

490

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

(NRC, 1971; Church, 1972). Wallach (1972) rec-
ommends a maximum ration of 12 to 19% protein
for wild sheep in captivity. It thus appears that sheep
ranges provide more than adequate quality during
the peak of the growing season. The extremely high
protein quality of sheep spring and summer forage
is matched by other nutrients (Whitten, 1 97 5). Win-
ter’s (1980) comparisons of the nutrient content of
plants available to a high quality sheep population
with one of lower quality sheep population during
the growth peak found no significant differences.
Given that wild sheep are getting food of more than
ample quality, can we be sure that they are getting
it in sufficient quantity?

3. Body and horn size are also not limited during
the major part of the growth season by quantity of
herbage. This assertion is not easy to confirm di-
rectly as it is difficult to analyze the amount of forage
consumed in the wild. I think there are ways, how-
ever, to indirectly illustrate that ample quantities of
food are available.

Palmer (1944) reported that Dali sheep require 6
lbs forage dry weight per day. Hoefs (1974) calcu-
lating numbers of bites per day, and estimating
quantity per bite, found wild sheep to exceed this
amount. Arnold (1964) found that the maximum
rate of biting is 100 bites per minute when vegeta-
tion is small and thinly distributed necessitating
many bites. Horejsi (1976) found wild sheep on
summer range to be below this amount (55 bites per
minute in June). Hoefs (1974) observed a feeding
rate of 45 to 55 bites per minute in dense emerging
vegetation in May and June.

The length of time it takes to fill or “top-up” the
rumen in summer could also be used as a crude
index of quantity of food available. Viereck (1963)
found that summer feeding periods averaged only
1.1 hrs; Whitten’s (1975) average for the summer
was 1.6 hr. On winter range Viereck (1963) found
the feeding time averaged much longer (2.43 hr).
Studies of domestic sheep have shown that the rate
of food intake increases and the rate of grazing time
decreases with improved pasture quality (Arnold,
1960a, 19606, 1962). Also, Blaxter et al. (1961) ob-
served that sheep appear to feed to a given point of
distention of their digestive tracts. Thus the brevity
of feeding periods is an indicator of pasture quantity,
assuming at least a moderate rate of passage. Be-
cause the ruminant system is geared to increase
transit time with increased food quality (rates of
digestion), one must also temper the evaluation of
feeding time with some estimate of time of passage.

The higher the food quality the more quickly it moves
through the gut, allowing increased rates of con-
sumption—so in many ways quality and quantity
are potentially related. To be assured that the short
feeding periods in the summer are indicative of am-
ple food quantity, rather than a rumen which is filled
with poor quality, slowly digesting food, one can
look at comparative faecal pellet counts. Summer
food has a rapid transit rate (Hoefs, 1974) as can be
seen by increased faecal pellet production during
spring and early summer (41 defaecations per 24
hrs in spring as opposed to 30 per 24 hrs in Octo-
ber—pellet number per defaecation being similar).
This is true despite the fact the percentage of un-
digestible fiber consumed (the bulk of faecal mate-
rial) is much higher in autumn.

Hoefs’ observation that sheep have a high number
of bites per hour in the first feeding periods of the
day but decline later also suggests that quantity is
not limiting during most of the summer.

4. The accumulated physiological- nutritional
debts during the winter are not the major determi-
nant of body size. The best way to test this hypoth-
esis is to capture sheep from a low quality popula-
tion in early spring when they are at their lowest
physiological ebb, then provide them with ample
high quality food, and see if they respond rapidly
to the increase in growth resources. I did this with
seven yearling Dali sheep (three male, four female)
from Healy, Alaska. They were caught in late March
and fed in captivity on a moderate to high quality
dietary plane. They responded rapidly in growth
that first spring, one of the males grew a 28 cm horn
segment, whereas the average in the wild is about
18 (Nichols, 1978). The length of horn in the captive
sheep was longer than any I have found in wild Dali
sheep. The first autumn in captivity the yearling
rams inseminated the yearling ewes and the ewes
raised lambs the following spring. This again is evi-
dence of their high quality, as yearling ewes do not
commonly rear lambs in the wild.

It may be that in an extremely poor winter there
could be some ungulate disability and lag-time in
spring recovery. However, the traditional view of
winter stress being the sole limiting factor, as in the
management of deer quality, has been discredited.
Klein (1965) showed that summer resources were
the major factor.

Two of my captive sheep became severely ema-
ciated when ill. They began to gain weight imme-
diately after being cured. One two-year-old male
who recovered in midwinter caught up with his peers

1984

GUTHRIE- PLEISTOCENE GIGANTISM

491

in horn growth and body weight in early spring. This
also is evidence for no or very little lag in recovery
from winter debilitation.

Winter mortality is important in reducing spring
competition for many animals— it leaves fewer in-
dividuals among which to divide the next summer’s
resources. I suspect winter severity is the causative
agent in correlations between general population
quality and density (Woodgerd, 1 964), both between
populations and within single populations through
time (Heimer and Smith, 1975). I will argue later
that in the production of large sheep, length of sum-
mer growing season is often secondarily associated
with severe winters.

5. The colonization phenomenon does not itself
explain body size trends in fossil North American
sheep. Geist (1971) proposed that colonization re-
sults in larger animals and that intraspecific size
variability is due to this phenomenon. McDonald
(1981) expands on this theory in his interpretation
of large body size in bison. The expansion into new
habitat usually does result in a size increase (prob-
ably both developmentally and genetically) for most
species. As populations reach or exceed carrying ca-
pacity, however, body size is again reduced. There
are numerous examples of this delayed size reduc-
tion in introduced or colonizing species, for exam-
ple, an introduced bighorn sheep population
(Woodgerd, 1964).

Geist proposed that as the glaciers receeded, the
sheep population colonizing the newly exposed hab-
itat became larger. One would expect then a chrono-
clinal gradient from small to large in the fossil record
across the glacial-postglacial boundary. However,
just the reverse is true as I have shown in a later
section, which suggests some other cause.

As Geist observed, sheep species do exhibit a geo-
graphic gradient of increased size toward glaciated
areas. However, Klein (personal communication)
has argued that this is a secondary correlate of phys-
iographic-phenological features which promote a
long, nutrient rich growing season, being more like-
ly to occur in areas susceptible to glaciation (that is,
well-watered, high mountain pastures).

6. The major variable of sheep body size, both
past and present in North America is the duration
of the high quality resource peak— the length of the
growing season. As I have already mentioned, there
is a portion of the year during which wild sheep
cannot grow because of an inherent growth dor-
mancy. The other part of the year they may grow if
resources beyond maintenance demands are avail-

able. For wild sheep populations, however, re-
sources are such that the sheep can only grow in a
portion of this annual available potential growth
period (Fig. 1). From studies with Dali sheep reared
in captivity, I have determined that the potential
growth period begins early in the spring— well before
new green growth is currently available to wild sheep,
and that growth potential continues well after plants
on the sheep’s natural range have senesced. Al-
though I have not been able to match the quality of
resources available to wild sheep in summer alpine
pastures, I have been able to extend the period of
actual growth in the captive sheep by providing
quality feed to support an early onset of growth in
the spring and to sustain growth in the autumn and
early winter. Because most of the captive sheep are
still living I have no osteological measurements to
compare with other populations. Hoefs (1974),
however, has already shown a much more rapid
growth rate and absolute weight in captive Dali sheep
kept on nutritionally supplmented rations. I do have
horn growth measurements; these seem to be as
equally good an indicator of dietary quality as body
size (Geist, 1971). The horns of my captive sheep
are much larger than the horns of wild sheep of the
same age and area (Guthrie, in preparation).

It seems reasonable to assume that if high quality
resources are generally available in ample quantities
during the peak growth season, then the remaining
variable to body size is the duration of this nutri-
tional peak.

One can visualize the “quality growth season”
hypothesis in experimental terms. We know that
northern ungulates can only grow rapidly during a
short segment of the year. We also know that if
resources are kept at a minimum maintenance level
during a large portion of the potential growth season,
ungulates will mature stunted in size. Now, we can
imagine an experiment in which we gave a group of
sheep unlimited quantities of the very highest qual-
ity food for two weeks during that growth season,
but keep them on maintenance rations for the re-
mainder. Because there are intrinsic limits to con-
sumption, digestion, assimilation of food, and to
growth rates, the two weeks’ access to optimum re-
sources would still not allow them to reach maxi-
mum genetic potential. However, if one continues
to expand the period of free access to quality re-
sources, adding a week at a time, the animals will
eventually reach a point at which they will reach
maximum genetic potential. From observing wild
sheep reared in captivity we can see that maximum

492

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

growth comes only when high quality resources are
supplied throughout the entire annual period of po-
tential growth. All of the hypotheses discussed seem
to point to the support of this last hypothesis that
the major variable in body and horn size develop-
ment is the seasonal duration of high quality, high
quantity resources.

Winter on northern sheep ranges is usually severe.
Sheep exploit high windblown slopes that are rela-
tively free of snow (Summerfield, 1974), but winter
vegetation is sparse and of low quality. The thick
fat deposits that sheep develop on energy rich forbs
and grass in the autumn decline steadily on meager
winter resources. Rams lose a considerable fraction
of these resources during the rut in late November-
early December (Geist, 1966). Ewes must carry the
new lambs during late winter when their fat reserves
are depleted and range resources are at the very
lowest level.

As a consequence of this winter bottleneck, Amer-
ican mountain sheep have a relatively high mortal-
ity rate; the majority of lambs are dead by the end
of their second winter (Murie, 1944; Murphy, 1974).
However, once past this second winter, sheep have
an unusually rectangular survivorship curve which
drops off only in old age. Most two-year-old sheep
will live for another six to eight years. This high
adult survivorship in the face of severe winters is
due primarily to two things: (1) the use of escape
terrain — mountainous slopes where sheep can usu-
ally out-run predators uphill, and (2) the highly nu-
tritious summer growth resource that alpine pas-
tures provide for growth, maintenance, and fat
reserves.

Klein (1970) has reviewed the factors responsible
for the extraordinary quality of arctic and alpine
plants. These cool-season plants are adapted to grow
very rapidly in a short summer. This rapid growth
produces leafy tissue with high levels of nitrogen
and phosphorous and little undigestible fiber. The
lack of shading by an upper story of trees means
that more light reaches these low growth forms and
the long days increase the length of the photosyn-
thetic period, which in turn means less nighttime
catabolic activity. Also Chapin (1977) has shown
that northern herbs have the ability to extract phos-
phorous from the soil more rapidly than their south-
ern counterparts. During their early seasonal growth
stages, these plants are relatively undefended by anti-
herbivore compounds. In addition, biomass pro-
duction on a 24-hr basis can be as high in the arctic

as in many temperate climate communities (Bliss,
1962).

Klein ( 1 965, 1 970) presented a model of ungulate
strategy, which capitalized on this brief, high quality
plant growth by physically following its wave of
appearance over the spring and summer landscape.
Because of topographic and altitudinal variations,
plant phenology will vary considerably throughout
the summer. This local variation is further exag-
gerated by the relatively low sun angle in the north.
On some low southern exposures new plant growth
occurs shortly after the snow and may not disappear
until late summer. Northern ungulates exploit these
variations in their summer range and are able to
find young plant growth for a considerable portion
of the summer season. Whitten (1975) and Winters
( 1 980) have documented this process with Dali sheep
movements. Dali sheep tend to range higher and
higher as the summer progresses, following the ap-
pearance of new plant growth up the slopes.

Large-bodied and large-homed sheep occur today
on ranges with wide altitudinal spans where they
can exploit a long season of high quality alpine
growth. Other factors are also important in deter-
mining the length of the ungulate growth season in
addition to sun angle, altitude, and aspect; most of
these features can be summed as climate. Jet stream
and cyclonic patterns can change the air tempera-
ture, cloud cover, and moisture regime to greatly
lengthen or shorten the growth season for any one
locality. Decreased snow, for example, can accel-
erate the onset of spring greenery. Aridity may be
an important factor limiting plant growth on well
drained, windy slopes with shallow soils; early sum-
mer showers may contribute significantly to plant
quality. Cloud cover may act to retard or prolong
the growth season depending on its occurrence in
spring or autumn. Certainly air temperature is of
primary importance in determining length of growth
season.

In Alaska, Dali sheep horn size increases from
west to east in both the Brooks Range and the Alaska
Range (Heimer and Smith, 1975). The fact that the
summer snowline altitidues also follow the same
gradients (Pewe, 1975) suggests more equable tem-
peratures (and hence plant growth) at higher alti-
tudes from west to east. In fact Pewe’s mapped es-
timates of high altitude summer snowlines
correspond roughly to the gradients of large-bodied
sheep populations throughout Alaska.

It is likely that a number of factors can potentially

1984

GUTHRIE— PLEISTOCENE GIGANTISM

493

lengthen the peak of available growth resources and
that these may vary in importance among living
sheep populations. Any one factor may not neces-
sarily be used to explain the causes of greater body
size in a given population without empirically ex-
amining the circumstances in question.

With the above information as background I will
now turn to the problem posed by the large size of
Pleistocene sheep, how they differ from the living
sheep and the possible causes of this change.

Today, the narrow altitudinal confinement of
sheep pasture between the talus of the mountain
crests and treeline was stretched in the Pleistocene.
The absence of a woodland meant access to the early
plant emergence of the lowlands, which would ex-
tend the length of the sheep’s peak growth season.
Sheep fossils, in now wooded hills around Fair-
banks, many kilometers from their current distri-
bution, testifies to the greater access of grassy foot-
hills in the Pleistocene. The largest bodied sheep
species today, Ovis ammon, makes considerable use
of foothills (Skogland and Petocz, 1975) as part of
pastures with tremendous altitudinal expanse. In
these habitats, Marco Polo sheep (Ovis ammon poll)
face severe winters, and few individuals have ever
been known to live past nine years (Skogland and
Petocz, 1975). The winters were undoubtedly harsh
for Alaskan Pleistocene sheep as well, with winds
(Guthrie, 1982) and exposure providing access to
winter forage, while at the same time leaching the
available nutrients. This balance of tight winter bot-
tleneck keeping populations well below competitive
margins for growth resources and a long growing
peak would select for animals which allocate nu-
trients to rapid body growth and violent early par-
ticipation in rut (Geist, 1966). This same selection
balance-point seems to have existed throughout the
Pleistocene geographic range of North American
sheep.

Although sheep occurred only in the western part
of the North American continent, their fossil record
is comparatively complete. Alaska and the Yukon
Territory have produced a number of specimens and
there are some fossils from British Columbia. There
are also Wisconsin-aged sheep fossils from Jaguar

Cave in Idaho; Little Box Elder Cave, Catlaw Cave,
and Winnemucca in Nevada; Rampart Cave and
Stanton Cave in Arizona; Glendale, California; and
near Lake Bonneville, Utah. These localities are re-
viewed by Harington (1978).

Stock and Stokes ( 1 969) discussed the Wisconsin-
age sheep which some authors give separate specific
status— Ovis catclawensis because of their much
larger body size (an almost American version of the
giant Ovis ammon of Asia). Stock and Stokes pro-
vided a date of 15,030 ±210 years B.P. on asso-
ciated organic material from Dry Cave in Nevada.
A review of these more southern sheep fossils re-
veals that they do average much larger than O. can-
adensis, but appear to be quite closely related, and
in fact, are now considered by many to be a chron-
ological subspecies, O. candensis catclawensis. Har-
ris and Mundel (1974) found that these large sheep
were widespread in the western United States and
Mexico during the late Wisconsin. They propose
that “environmental deterioration” near the end of
the Wisconsin and into the Holcene resulted in se-
lection for smaller sized sheep. Thus extant bighorn
sheep, O. canadensis, are reduced or dwarfed rep-
resentatives of the late Pleistocene sheep from the
same areas.

This also seems to be the case with O. dalli and
its fossil counterparts. Harington (1978) has con-
cluded that the Wisconsin-aged sheep from the Yu-
kon Territory are much larger than sheep living there
today; measurments on fossil skeletal material av-
erage about 10 to 11% larger than the same mea-
surements on extant sheep from the Yukon Terri-
tory. Harington suggests that the larger size of these
fossil sheep can be better appreciated when one notes
that hind leg length between living rams and ewes
(exhibiting considerable sexual dimorphism) is only
in the order of 6%. As pointed out earlier, I have
recorded this same phenomenon in fossil and living
sheep in the Yukon-Tanana Uplands in central
Alaska. This additional evidence for postglacial size
reduction in Ovis now allows one to conclude that
such dwarfing was a general phenomenon through-
out North America.

HOLOCENE SIZE REDUCTIONS IN BISON AND CERVUS

In addition to sheep, there are two other North Cervus— both are present in Alaska well into the

American megafaunal grazers which did not become Holocene.

extinct at the end of the late Pleistocene, Bison and The decline in horn and body size of Bison sp. in

494

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

the early Holocene is probably the best documented
case of rapid evolution in mammals. Harington’s
(1978) most recent date for large-horned superbison
(B. priscus ) in the Yukon Territory is just prior to
12,000 B.P. We have a date 10,600 B.P. on a very
large bison mandible from the Dry Creek site (Pow-
ers et al., in press). After this time ( 1 1 ,000 to 1 0,000
B.P.) bison begin to decline rapidly in size toward
the living species sizes in Great Plains archaeolog-
ical sites beginning at 1 1,000 B.P. As is the case in
Alaska, these Great Plains bison become progres-
sively smaller until the present. Vereshchagin (1967)
has discussed a parallel phenomenon in central Eur-
asia and Sher (1971) has reported the same decline
in Siberia.

The mid-Pleistocene increase and subsequent re-
duction in body size which occurred in bison evo-
lution characterizes many ungulate size trends
throughout the Quaternary (Guthrie, 1980). The
peak in bison size was reached in the mid-Pleisto-
cene, though the chronology varied somewhat in
different areas. Bison remained large bodied through
to the early part of the late Pleistocene (Illinoian in
North America) and then began a decline in size.
This decline plateaued somewhat during the last
glaciation and then accelerated in a rapid descent
in body and horn size during the postglacial. The
timing of this last decline also varied with geograph-
ic location.

The major postglacial decline in bison body size
was a widespread event, occurring in northern and
southern Europe, in northern Asia, and much of
North America. I contend that the major cause of
this decline in body size was an abbreviation of the
annual optimum growth season. In the case of bison,
the effects of this shorter growing season must have
been exacerbated by increasing postglacial plant zo-
nation (Guthrie, 1984) which reduced local vege-
tational diversity.

In the New World the postglacial winter bottle-
neck was opened by the spreading of bison’s optimal
short-grass, Bouteloua-Buchloe, habitat, so bison
numbers increased concurrent with, and despite the
declining body size.

There must have been a somewhat cyclical rela-
tionship initiated by climatic change — an abbrevi-
ated season for plant growth resulted in plant zo-
nation; such zonation decreased local herbaceous
diversity; which in turn, further shortened the grow-
ing season for most ungulates (Guthrie, 1 984). Thus,
the reductions in “ungulate-growth-days” were not
a simple matter of a subtraction of the actual de-
crease in “degree days.”

The effects of the shortened season were inten-
sified for bison in the New World by the postglacial
disappearance of bison’s competitors, which further
exaggerated the differences between good and poor
bison habitat; due to the absence of regrowth from
the tall grass species eaten by the more fiber-tolerant
large caecalids, and the decline in whole-plant nu-
trient quality because the low quality parts (stems)
were not removed by the (now extinct) mammoths
and equids. The short-grass plains however became
dominated by bison, but the simplicity of vegetation
probably resulted in a decreased phenological
succession of local plant maturation — shortening the
peak availability of high quality grasses. These C4
grasses tend to be lower in peak quality than their
C3 counterparts (Mattson, 1980).

The fossil record of elk (or wapiti), Cervus, is not
as complete as that of Bison , but it is clear in the
far north at least that Cervus underwent a postglacial
decline in body size (Guthrie, 1966; Harington,
1978). This phenomenon is best seen in Europe
where the extant Cervus (red deer) is much smaller
than the Pleistocene forms (Altuna, 1972).

The large wapiti Cervus elaphus described by
Shackleton and Hills (1977) dating at 9,670 ± 160
years B.P. illustrates the same principle seen in Bi-
son and Ovis. Even during the glacial recession into
the early Holocene this megafaunal grazer had not
yet reached the reduced size of its modern repre-
sentatives. For example, the Cervus specimens from
the Star Carr Site dated at around 9,000 B.P. in
England are larger than those from any of the current
populations in the British Isles (Clark, 1954).

As with Ovis and Bison, the Pleistocene Cervus
have disproportionally larger horns or antlers than
their Holocene descendants. This is significant be-
cause social organs are an even more sensitive in-
dicators of range quality than body size. In times of
stress these are organs of “low growth priority” and
are reduced first developmentally, the evolution-
arily. The same processes operate for exceptionally
high quality ranges, except in the reverse — the social
organs grow larger.

The antlers of Cervus are informative in this re-
gard, because they give us additional information
about the Pleistocene growth season length. Antlers
grow somewhat like woody plants, having their
“meristemative” tissue on the ends of the tines. The
calcification process takes place in a proximal-distal
gradient. Elk or wapiti raised in captivity, where
they are fed a highly nutritious diet from early spring
to the time the velvet is shed, develop antlers shaped
differently than antlers of wild wapiti; the tines are

1984

GUTHRIE — PLEISTOCENE GIGANTISM

495

unusually long. This seems to be due to the main-
tenance of a high dietary plane for a longer time
than in the wild. Although antlers from captive wa-
piti tend to be somewhat larger than their wild coun-
terparts due to a better quality diet, it is these long
tines that give them their unusual form.

This form is unusual in extant wild Cervus, but
it is common in the large Pleistocene Cervus antlers,
suggesting two things — that the range quality was
excellent and that the high quality existed for a rel-
atively long season each year.

Kurten (1968) has discussed this post-Pleistocene
dwarfing for a number of species, including brown
bears, Ursus arctos. Garutt (1964) has shown that
it was characteristic of late glacial Siberian mam-
moth, Mammuthus primigenius, just prior to ex-
tinction.

Unfortunately, we know all too little about why

some extant populations, counterparts of species
discussed above, are of higher quality than other
populations. We can only say that dietary resources
are accessible and are usually of high quality because
of various combinations of situations. All ungulate
populations usually have access to high quality food
resources for some part of the growth season; how-
ever, it is the duration of the quality peak which
seems to be important.

Bison, for example, use high quality cool-season
grasses in the early spring, but then shift to more
fibrous, but moderate quality warm-season grasses
throughout their growth season, but the duration of
the peak relates to numerous seasonal variables in-
cluding plant phenology, moisture, and fire. Al-
though the proximate variables may differ some-
what, much the same is also true for these other
species.

HOLOCENE BODY SIZE REDUCTION IN RANGIFER . OVIBOS, AND ALCES

One would expect to find an increase in body size
in reindeer ( Rangifer ), muskoxen ( Ovibos ), and
moose ( Alces ), when they expanded in the Holocene
to dominate the northern ungulate faunas. Presum-
ably, caribou, muskoxen, and moose become more
abundant with the rise of more moist conditions
and the accompanying increase in shrubs and tundra
vegetation. I found that, in fact, the Alaska Pleis-
tocene specimens of these three species did average
larger in body size (Figs. 5 to 1 1 ) than living Alaskan
species. (The Alaskan subspecies are the largest forms
of the extant species.)

In Alaska, summer forage for caribou, muskoxen,
and moose is profuse and nutritious — yet, their
Pleistocene ancestors were even larger than those of
today. How can this be? We know of no floristic
sources that would have greatly increased the quality
of summer forage during the Pleistocene. The only
obvious route to larger body size is to extend the
period in which quality forage is available. The larg-
est moose today are found on the Alaska Peninsula
where summers are long and cool. The caribou on
the Alaska Peninsula and the Aleutian Islands, with
a similar seasonal pattern, are also among the largest
of the subspecies.

Alaskan moose, muskoxen, and caribou, like bi-
son on the Great Plains (Guthrie, 1980), experi-
enced a postglacial increase in numbers. Interspe-
cific competetion declined with the demise of most
of the other megafauna as forage for moose, mus-

koxen, and caribou expanded. This widespread
vegetational change resulted in a greater quantity of
winter forage (willow shrubs for moose and mus-
koxen, and lichen for caribou, Guthrie, 1982). It
thus allowed greater numbers of animals through
the winter bottleneck, but new climatic regime did
not provide more growth resources per animal in
the summer.

Lengths and widths of M1, M2, and Ms are com-
pared between fossil moose from interior Alaska
and several extant populations from the same area.
The differences are all highly significant (Figs. 8, 9
and 10). The data are presented in Table 1 and 2.
I have not as yet located sufficient moose metapo-
dials, but Dr. David R. Klein of the Cooperative
Wildlife Unit, University of Alaska, graciously let
me use his data for caribou metatarsals. The fossils
are significantly longer. There were two caribou
metatarsals from Adak Island which exceeded the
fossil metatarsals in length. The Adak caribou were
transplanted from the central Alaskan Nelchina herd,
and have become enormous on this virgin range
with a long cool maritime growing season. One adult
bull was weighed at 700 lbs (320 kg), twice the weight
of healthy bulls from most other populations. Mea-
surements were also taken on fossil Alaskan musk-
oxen ( Ovibos ) skulls to compare with wild animals
living today in the Canadian High Arctic. The fossils
are significantly larger (Table 2 and Fig. 1 1).

The moose, muskoxen, and caribou fossils used

Length (mm)

496

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Width (mm)

Fig. 5. — A comparison of fossil and recent caribou ( Rangifer ) from interior Alaska using the first lower molars. The fossil animals
(triangles) and the recent specimens (circled dots) are from the same range of hills near Fairbanks, Alaska. The uncircled black dots
represent recent caribou taken from all over Alaska.

Length (mm)

1984

GUTHRIE- PLEISTOCENE GIGANTISM

497

Fig. 6.— A comparison of fossil and recent caribou ( Rangifer ) from interior Alaska using second lower molars. All details the same as
Fig. 5.

498

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 1. — Comparison of dental and leg bone measurements between fossil and Recent caribou (Rangifer tarandusj.

Tooth or bone

Length (mm)

Width (mm)

Significance

M, (Fossil)

mean = 20.50 ±

1.63

mean = 10.79 ± 0.75

n = 41

n = 39

Student-t

L = 12.51

v. sig.

W = 10.47

v. sig.

M, (Recent)

mean = 1 7.66 ± 0.94

mean = 9.58 ± 0.52

n = 88

n = 87

M: (Fossil)

mean = 21.21 ±

1.41

mean = 1 1.62 ± 0.86

n = 33

n = 33

Student-t

L = 10.92

v. sig.

W = 8.5 1

v. sig.

M, (Recent)

mean = 18.81 ±

0.89

mean = 10.45 ± 0.59

n = 88

n = 88

M3 (Fossil)

mean = 23.93 ±

1.93

mean = 1 1.02 ± 1.11

n = 23

n = 24

Student-t

L = 7.23

V. sig.

W = 8.38

V. sig.

M3 (Recent)

mean = 21.56 ±

1.23

mean = 9.66 ± 0.55

n = 88

n = 88

Metacarpals (Fossil)

mean = 1 9.6 1 ±

1.12

mean = 43.86 ± 2.78

n = 71

n = 70

Metatarsals (Fossil)

mean = 27.5 1 ±

1.36

mean = 43.13 ± 2.26

n = 53

n = 52

Student-t

L = 16.86

v. sig.

Metatarsals (Recent)

mean = 23.68 ±

1.48

Extant Rangifer from Alaska

n = 163

(various populations)

for comparison came from Fairbanks area and were
collected in conjunction with the placer gold mining
earlier in the century. They date mainly from late
Pleistocene (late Rancholabrean). Pewe (1975) and

I (Guthrie, 1968) discussed the collection in pre-
vious papers. The collections are now in the Amer-
ican Museum of Natural History.

DWARFED MONOGASTRICS OF THE MAMMOTH STEPPE

Although present during the last glacial, horses
and hemionids ( Equus ), wooly mammoths ( Mam -
muthus), and camelids ( Camelops ) became extinct
at the beginning of the Holocene. Their extinction
is informative because these species are somewhat
of an exception to the Pleistocene gigantism trend.
Although wooly mammoths are large by anyone’s
standards, they are not large for proboscidians. Rel-
atives in the central and southern part of the con-
tinent were much larger (for example, Mammuthus
columbi), as are the living elephants today. The
hemionids and horses were also quite small through-
out most of the late Pleistocene in Alaska — not much
larger than Shetland ponies. The few bones and teeth
of camels ( Camelops ) also do not seem to be as large

as some of their relatives further south in North
America, and they did not approach the size of the
old world Camelus.

My assessment of this unusual reduced size among
the mammoth steppe non-ruminant fauna (I will
include camels in this category [Guthrie, 1984] with
qualifications) is that they suffered most from the
shorter growing season in the north.

Ruminants are capable of taking advantage of very
high nutrient planes, these caecal-colon digesters are
not. They are adapted to an extremely long season
of medium-to-low quality nutrient levels (Guthrie,
1 984). They fit in the second category in Fig. 1 . That
is, they are large because they prolong their growth
over a number of years. A long growing season on

Length (mm)

1984

GUTHRIE- PLEISTOCENE GIGANTISM

499

8 9 10 I ! 12 13

Width (mm)

Fig. 7. — A comparison of fossil and recent caribou ( Rangifer ) from interior Alaska using third lower molars. All details the same as
Fig. 5.

500

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Table 2. — Comparison between dental and leg bone measurements of fossil and Recent moose (Alces alcesj and cranial measurements

of fossil and Recent muskoxen (Ovibos moschatusj.

Tooth, leg bone, or
cranial measurement

Length (mm)

Width (mm)

Significance

A Ices

M, (Fossil)

mean

= 29.42

±

1.16

mean = 20.48 ± 0.85

Student-t

n

= 33

n = 37

L = 9.69

v. sig.

W = 5.44

V. sig.

M,(Recent)

mean

= 26.26

±

1.22

mean = 19.07 ± 1.13

n

= 22

n = 22

M, (Fossil)

mean

= 31.75

±

1.09

mean = 22. 1 1 ± 1.14

Student-t

n

= 42

n = 43

L = 1 1.75

v. sig.

W = 4.59

v. sig.

M, (Recent)

mean

= 28.36

±

1.19

mean = 20.68 ± 1 .36

n

= 24

n = 24

M3 (Fossil)

mean

= 43.31

±

1.69

mean = 22.6 1 ± 1 .26

Student-t

n

= 35

n = 36

L = 7.74

v. sig.

W = 4.05

v. sig.

M, (Recent)

mean

= 40.03

±

1.39

mean = 2 1 .06 ± 1 .64

n

= 23

n = 22

Metacarpals (Fossil)

mean

= 34.32

±

1.44

mean = 68.43 ± 3.77

n

= 25

n = 25

Metatarsals (Fossil)

mean

= 40.17

±

1.62

mean = 66.95 ± 3.21

n

= 21

n = 21

Ovibos

Postorbital constric-

mean = 14.7 ± .70

Student-t

tion (Fossil)

n = 32

POS = 6.40

v. sig.

Postorbital constric-

mean = 12.3 ± .30

tion (Recent)

n = 20

Occipital width

mean = 18.30 ± .62

(Fossil)

n = 32

Occipital width

mean = 1 7.05 ± .49

Student-t

(Recent)

n = 20

OC = 4.22

v. sig.

the high nutrient flushes in the north was still a short
growing season for the large mongastric caecalids.

As I have pointed out elsewhere (Guthrie, 1984),
these non-ruminant ungulates employ a very con-
servative growth strategy: 1) low percentages of milk
protein; 2) long gestation period; 3) long suckling
periods; 4) slow growth rates; 5) infrequent repro-
duction; 6) forage is mainly high-fiber low-protein;
7) slow maturation time; 8) long life expectancy; 9)
small litter size.

Because of this conservatism, a long growing sea-
son (low peak of nutrient levels) is of fundamental
importance. Young have to have time to grow, the
female must have time to both nurse her young and

recover from winter debilitation. The response to a
moderate growth season plus a long off-season of
imposed growth dormancy would be quite apparent
in these groups as a reduction in body size. A rapid
reduction in growth season length would mean their
demise. Alaskan mammoth from Colorado Creek
dating at 1 5,000 B.P. and at Teklanika dating 1 2,300
B.P. already show considerable dwarfing. Judging
by the Dry Creek fauna (around 1 1,000 B.P.), ex-
tinction among the caecalids had already taken place
but the vegetational change had not yet produced
the major dwarfing among the ruminants (mountain
sheep, bison, and wapiti).

Although it is a truism that the larger bodied un-

1984

GUTHRIE — PLEISTOCENE GIGANTISM

501

32

30

29

28

27

cn

£ 26

Moose A Ices a Ices

M

▲ A

A A

A A A

X-

A A A

• A

25

24

23

• _ •

A Fossil
• Recent

17 18

19 20 2

Width (mm)

22 23

Fig. 8. — A comparison of fossil and recent moose ( Alces ) from interior Alaska using the first lower molars. The fossil moose teeth are
taken from sediments of late Pleistocene age in the Fairbanks, Alaska, area.

gulates today are the ones which are adapted to
eating the lower quality forage (more fibrous, fewer
nutrients), that correlation can cloud the issue of

Pleistocene gigantism. The concept behind this body
size-dietary fiber association is that a larger caecum
or rumen is required to efficiently incubate roughage

502

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

35

34

33

32

E

E

cr>

c

CD

30

29

28

27

26

Moose A/ces a/ces
Mo

* Fossil
® Recent

±A *

\ *

▲ A

i

V

A A

A A

• • A

• • •

• •

18 19 20 21 22 23 24

Width (mm)

Fig. 9. — A comparison of fossil and recent moose ( Alces ) from interior Alaska using second lower molars. Specimens as Fig.

and hence would select for large bodied animals
(Janis, 1975). Ungulates thus fall on a gradient of
two opposing foraging strategies (Bell, 1971; Wes-

toby, 1974, 1978; Hanley, 1980). High quality for-
age is sparsely distributed, and to take full advantage
of it one has to be quite selective. Being selective

1984

GUTHRIE- PLEISTOCENE GIGANTISM

503

E

E

cr>

c

a>

47

46

45

44

43

42

41

40

39

38.

Moose A Ices a Ices

M

A Fossil
• Recent

16

A A
A

▲ A

AA A

. aa

T

20 21 22 23 24 25

Width (mm)

Fig. 10. — A comparison of fossil and recent moose (Alces) from interior Alaska using third lower molars. Specimens as Fig. 8.

requires much greater time and results in less quan-
tity per given effort. These quantity limitations limit
body size, so that more selective browsers such as
American deer ( Odocoileus ) or roe deer ( Capreolus )
are relatively small bodied. The lower quality for-
age, however, is normally abundant and quickly ob-
tained, thus less selective grazers can get adequate
quantities with which to grow large, but seem to be,
seasonally, on the margins of obtaining sufficient

quality. The wapiti or red deer ( Cervus ) are sym-
patric with the above small deer and are contrast-
ingly larger bodied and forage less selectively.

Each strategy is caught by the opposite limits. For
a red deer to grow larger, it must increase the forage
quality by being more selective, but this would limit
quantity. Roe deer in order to grow larger must
increase forage quantity, by being less selective, but
this would decrease quality. Limited in opposite

NO. 8

504

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

e

o

c

o

o

CO

c

o

CJ>

-Q

O

CO

o

Ql

8

17

14

12

14

Muskoxen Ovibos moschatus

A A
▲ A

▲

▲

A A ^
^ A* A

A

• •

# %•

A

A

• •

o •

* •

A Fossil
• Recent

16

18

20

22

Occ. Width (cm)

Fig. 1 1 . — A comparison of fossil and recent muskoxen. The muskoxen fossils are all from Alaska. The measurements were from recent
skulls taken on individuals from various places in the Canadian High Arctic.

ways, each species, occupying a different balance
point in these strategies, can live together without
much competition.

I would argue, however, that the train of this
thinking provides very few clues to Pleistocene gi-
gantism. The reason is that both quantity and qual-
ity can be increased. For example, roe deer and red
deer are both quite small on the Iberian Peninsula
(Altuna, 1972), yet both species are gigantic (over

twice the size) in southern Siberia. Forage quantity
and quality can vary absolutely up and down the
scale for both large and small ungulates as they did
for, say, small saiga and large bison in the Alaskan
Pleistocene which were larger than their living rel-
atives to the south or for, say, large mammoths or
small horses in the Alaskan Pleistocene which were
both smaller than their relatives to the south.

1984

GUTHRIE- PLEISTOCENE GIGANTISM

505

MINI-GROUND SQUIRRELS AMONG THE GIANT MEGAFAUNA

Cursory examination of Alaskan Pleistocene
ground squirrels showed them to be smaller than
their extant relatives for the treeless arctic. A de-
tailed comparison of several osteoloical measure-
ments confirmed that assessment. The mean length
of adult male and female skulls from arctic Alaska
is 61 mm and 58 mm, respectively, whereas the
Pleistocene specimens from the Fairbanks area av-
eraged 53 mm. All other measurements showed this
same size difference. The fossils were more com-
parable to subspecies further south. Why with all
this gigantism around them would not the Pleisto-
cene ground squirrels also have been large?

One obvious difference is that they are a hiber-
nator and hibemators seem to be governed by a little
different balance of life history forces than non-hi-
bemators. A long cold season of hibernation re-
quires a greater lipid reserve (Mrosovsky, 1976;
Wrazen and Wrazen, 1 982). Greater body size means
a greater absolute amount of reserve energy and
greater likelihood of surviving the winter (Kiell and
Millar, 1978).

There is, consequently, a body size gradient among
ground squirrels from south to north with increasing
length of hibernation, just as there is in analogous
fashion with increasing altitude. In fact, that alti-
tudinal analogy can help us unravel the puzzle of
Pleistocene body size. For with increasing latitude
and altitude, ground squirrels are not only larger
(Levenson, 1979), they have lower reproductive rates

in the form of smaller litters (Bronson, 1980; Murie
et al., 1980; Murie and Harris, 1982). Having to
reach near adult body size before the first winter and
living at high altitudes or latitudes, (facing a shorter
growing season) these squirrels must grow faster than
their temperate latitude or lowland counterparts.
Thus, there are altitudinal and latitudinal gradients
of intrinsic growth rates (Kiell and Millar, 1978;
Koeppe and Hoffman, 1981; Murie and Harris,
1982). Although forage quality in northern and
mountainous areas is known to be rich (Klein, 1970)
the ground squirrels obviously grow large, not sim-
ply because of a greater growth season peak or length
but because they allocate fewer resources to repro-
duction in the form of reduced litters.

Thus body size among ground squirrels is a better
reflection of the length of hibernation period (in
context with other life history features) than it is a
reflection of their growing season resources. In-
creased length of peak growing season seems to be
translated into increased litter size (at the expense
of body size), a quite different phenomenon than
occurs in larger nonhibernators.

Thus the small Alaskan Pleistocene ground squir-
rels, rather than indicating a shorter growing season
suggest a longer one with a shorter hibernation pe-
riod than occurs in those same areas today. This is
consistent with our generalizations from the mega-
fauna.

PLEISTOCENE GIGANTISM FROM AN ALASKAN PERSPECTIVE

I have tried to present a theory accounting for the
reductions in body size of Alaskan ungulates from
the peak of the glacial phase to and through the
postglacial. This conceptual model of seasonal shifts
and its impact on evolution of life history strategies
may be used to better understand Pleistocene gi-
gantism in general. Although there were large mam-
mals throughout the Tertiary, it is during the Qua-
ternary (particularly the middle portions) that many
groups experience an increased body size in many

of their phylogenetic lines. Some of the equids be-
came immense (for example, Equus giganteus) and
so did the bovids (for example, Bison /atifrons), cer-
vids (for example, Cervalces, Megaceros), probos-
cidians (for example, Mammuthus columbi) and
many other groups. Additionally many had enor-
mous tusks, horns, and antlers unparalleled in the
Tertiary. Geist (1971) and Edwards (1967) discuss
these in detail. The question is why? Why were large
body size and large social organs selected during the

506

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Pleistocene when we normally think of the cooler
climate in terms of climatic deterioration? Gener-
alizing from the Beringian situation I propose that
the effect was global due to several interrelated fea-
tures, all a product of that climatic shift:

1 . A tightening of the winter or dry season bot-
tleneck increased off-season mortality and
hence decreased competition during the fol-
lowing growth season.

2. A more stressful season of plant growth fostered
less conservative plants which experienced
marked seasonal pulses of nutrients and as a
consequence were more digestible to ungu-
lates—grasses, forbs, and shrubs (all in com-
plex mosaic communities).

3. The shortening growth seasons in the early and
mid-Pleistocene still were not limiting to body
size and were probably compensated for by an
accompanying reduction in large mammal
species diversity. However, at least in the north,
the further abbreviated growth seasons in the
late Pleistocene did become limiting. Thus,
many groups, particularly the conservative
large caecalids (proboscidians, equids, and
rhinocerids) underwent body size reductions,
and then became extinct at the beginning of
the Holocene. Among the large mammals that
did survive this revolutionary time, it is ap-
parent that most experienced at least some body
size reduction. From an ungulate’s eye view
the late Pleistocene afforded a longer growing
season than did the Holocene.

I have argued elsewhere (Guthrie, 1974, 1980,
1982, 1984) that where there is intraspecific com-
petition for growth resources (and there usually is
among ungulates) heavy mortality in winter in-
creases the quality and quantity of available growth
herbage per individual during the following growth
season (particularly during the critical beginning and
end of that time). This is a well-known principle
among wildlife managers who increase the hunting
bag during the non-growth season to increase in-
dividual deer quality in the herd (Klein, 1965).

Increasing resources by decreasing competition
not only increases the quality of available growth
nutrients, it can also increase their duration over a
longer growth season by increasing the quantity of
the rarer nutritious early plants in the spring and
late ones in the autumn — per individual animal.

The increased duration of growth resources pushes
the individuals nearer their maximum size growth

potential. If there are inherent advantages to being
larger when growth nutrients are available (and there
usually are these advantages for most ungulates),
the selection balance shifts toward a larger body size
optimum.

Larger individuals gain social stature sooner and
hence experience reproductive privilege sooner, thus
reproduce earlier. When the bottleneck of winter
survival is tightened, large size becomes a decided
fitness advantage as one cannot gamble on too con-
servative a growth, or reproductive strategy when
the chances of dying during the winter have in-
creased. Geist (1971) pointed out that the rapidly
growing, large “daring,” species and subspecies of
mountain sheep (Ovis) are the ones with very short
life expectancies. This is a general pattern in other
artiodactyl species as well where the factors outlined
above seem to be in play.

This selection regime affects social organs in the
same manner as it does body size— except its effects
are even more exaggerated. The relative size of an
individual’s social organs (antlers, horns, tusks,
manes, and similar structures) are often the major
clue to social rank and hence reproductive privilege.
The resources that are devoted to fighting and dis-
play paraphernalia become optimal evolutionary' in-
vestments when nutrients for growth abound and
when the likelihood of dying before the next repro-
ductive season has increased.

This combination of a long peak growth season
in an environment which is kept understocked be-
cause of high competition during the winter bottle-
neck seems to have reached its zenith during the
middle Pleistocene. The more open winter bottle-
neck of the Tertiary and early Pleistocene environ-
ment changed to a tighter bottleneck and a long peak
ungulate growth season during the mid-Pleistocene.
In the late Pleistocene the growth season shortened
and this change continued through the Holocene.

The postglacial dwarfing can thus be seen as a
continuation (albeit accelerated) of the general trend
toward reduced body size in ungulates since the mid-
Pleistocene. This was of course exaggerated in the
north where shortened ungulate growth seasons had
a more profound effect, but it can be seen occurring
at mid-latitudes as well ( Bison latifrons into B. an-
tiquus and B. antiquus into B. bison).

It is possible that an increased seasonality during
the mid-Pleistocene tended to favor plants higher
in available nutrients (that is, which grow more rap-
idly and are less well defended by antiherbivory
compounds and fiber and contain high percentages

1984

GUTHRIE — PLEISTOCENE GIGANTISM

507

of nitrogen and phosphorus). These are grasses,
forbs, and shrubs — the growth forms which ungu-
lates broadly seem to have evolved to use as their
dietary staples.

If the sequence of my argument is sound thus far,
several questions implied by this train of thought
now require investigation. The most prominent of
these is the need to look beyond such generalizations
as annual mean temperatures or July maximums,
and their correlations with modern biota, in our
work toward reconstruction of the past. Several au-
thors (for example, Guilday et al., 1977; Whithead,
1973) have shown that there was no simple climatic
displacement of the biomes we recognize as integral
today. Changes more profound than mean annual
shifts occurred; there had to be major changes in
seasonal relationships (Guthrie, 1984). A number
of paleoclimatologists have mentioned this likeli-
hood. Hare (1976), for example, proposes that there
was “probably less seasonal variation in the general
pattern of circulation than today.” Circumpolar wave
number two today undergoes a large shift of lon-
gitude between summer and winter (Hare, 1976).
This would not have happened in Wisconsin times,
because it is “forced by large thermal differences
between land and sea that develop today during the
low-albedo summer and high-albedo late winter and
spring seasons.”

Whatever the origins of these different climatic
states, we can conclude from the ungulate size strat-
egies that the ecological norms of today were dif-
ferent then. Along with Guilday et al. (1977), we
must conclude that there may be no good analogues
in the biomes of today with those of the Pleistocene.

So for optimum large northern ungulate body
growth, there would be an early spring moisture
peak and the summers would be dry (somewhat arid,
to keep the woody plants down) but with spotty
showers to moisten the herbaceous growth. The
strategies of these “mammoth steppe” megafaunas
were mostly nomadic, judging by their plains coun-
terparts today, so that they could migrate to the
areas of high quality plant growth.

Autumns may be less critical to body size because
in most northern ungulates somatic growth declines
before fall and autumn is a time body metabolism
switches to depositing fat for winter. Autumn so-
matic growth could be prolonged by clear skies,
which result in warm days without moisture, but
also clear skies would increase nighttime radiation
loss (creating nighttime frosts). This may not be a
detrimental element however as many northern

herbs are moderately frost resistant. Early snows,
however, would shorten the growth season, so it is
possible that they came later in the year than they
do today.

The Pleistocene regulator of densities was un-
doubtably severe winters. A severe winter for an
ungulate can involve deep snows which require in-
creased energy expenditure (we know there were no
deep snows, Guthrie, 1982, in Beringia) or poor
over-wintering plant resources (which probably was
the case). Standing dead cool-season plants are no-
toriously poor winter food (Peden et al., 1974). They
are made even worse by exposure (lack of protective
snow-cover). Winters must have been times of great
reliance on body fat. Ruminants on such a diet can
recycle nitrogen to break down carbohydrates, but
do so inefficiently on forage at very low nutrient
levels (Church, 1972). The high mortality (partic-
ularly of the early age classes) is ironically a boon
to ungulate quality as it greatly decreases competi-
tion for limited spring growth resources.

There are anatomical indicators that the winters
were in fact difficult for ungulates. The cementum
layers on the teeth are laid down in an annual cycle
of deposition, like a tree ring. In teeth from the large
ungulates of the mammoth steppes the winter an-
nulus is so constructed that it is strikingly apparent
to the naked eye— without having to make histo-
logical sections as one has to do for more southern
species. This annulus is produced by growth ces-
sation in response to winter debilitation.

A weather summary for the glacial conditions
which promoted Alaskan gigantism would seem to
involve a long dry windy summer, hazy clear skies
with little precipitation, early springs and late au-
tumns, with little winter snow cover.

Isdo (1976) proposed the climatic “thermal blan-
ket” theory for desert or extremely dusty areas. The
atmospheric particles serve to maintain ground level
temperatures higher than usual by reducing radia-
tion loss. The implications of this theory of thermal
blanketing may apply to glacial atmosphere. Glacial
outwash fans produced prodigious quantities of rock
flour in the silt-clay size range — a newly milled caf-
eteria of mineral nutrients. These fine particles were
lifted by winds and distributed across the country-
side as loess (Pewe, 1975). It is difficult to fully
reconstruct the atmosphere of the mammoth steppe,
but the thick loess deposits across northern Eurasia
and Alaska, and south of the North American con-
tinental ice sheets which produced an enriched min-
eral mantle suggest a dense “dust bowl” haze

508

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

throughout the growing season. In the far north the
“thermal blanket” of dust would probably have its
greatest effect during late summer and autumn be-
cause radiation loss to the cold night sky increases
as the days get shorter.

Surface dust also serves to decrease late winter
and spring albedos. In several windy regions of Alas-
ka today surface areas stripped free of snow con-
tribute dust which covers the remaining snow with
a thin coat of silt; this in turn hastens early subli-
mation or melting and produces early greenery.

The Asian people who moved into Alaska at the
end of the Pleistocene must have caught the last
edge of these colorful sunsets and hazy summers; a
major shift had begun toward the recent climate.
They were among the last to see the remaining giants
of the Beringian Pleistocene and experience the cli-
matic vestiges responsible for those animals. It was
a revolutionary time.

ACKNOWLEDGMENTS

I wish to thank the National Park Service and the National
Geographic Society who jointly funded the North Alaska Range
Project. Under that project, we excavated the Dry Creek Site and,
while doing so, I was stimulated to ponder the reasons underlying
the large body size of the Pleistocene Alaskan faunas and have
the time to research those ideas. Also I wish to thank Gary

Selinger who measured the fossil and recent ground squirrel bones
for me as an extra credit course. Colleagues provide an essential
sounding board and coffee-room critique for one’s ideas and I
would like to acknowledge their contribution. Dave Klein, Roger
Powers, Bob White, John Bryant, and others have significantly
affected my thinking on ecology and gigantism.

LITERATURE CITED

Ager, T. A. 1975. Late Quaternary environment history in the
Tanana Valley, Alaska. Inst. Polar Studies, Ohio State Univ.,
Columbus, 54:1-230.

Altuna, J. 1972. Fauna de Mamiferos de los Yacimientos.
Prehistoricos de Guipuzcoa Munibe, Ano 24, Fasc. 1-4.

Arnold, G. W. 1 960a. Selective grazing by sheep of two forage
species at different stages of growth. Australian J. Agric. Res.,
1 1:1062-1 133.

. 1960ft. The effect of quantity and quality of pasture

available to sheep and their grazing behavior. Australian J.
Agric. Res., 1 1 : 1034-1043.

. 1962. The influence of several factors in determining

the grazing behavior of border Leichester x merino sheep.
J. British Grassland Soc., 17:41-51.

. 1964. Factors within plant associations affecting the

behavior and performance of grazing animals. Pp. 10-38, in
Grazing in terrestrial and marine environments (D. J. Crisp,
ed.), Blackwell, Oxford, 430 pp.

Bell, R. H. V. 1971. A grazing system in the Serengeti. Sci.
Amer., 225:86-93.

Blaxter, K. L., F. M. Wainman, and R. S. Wilson. 1961. The
regulation of food intake by sheep. Anim. Prod., 3:51-61.

Bliss, L. C. 1962. Adaptations of arctic and alpine plants to
environmental conditions. Arctic, 15:117-144.

Bronson, M. 1980. Altitudinal variation in emergence time of
golden-mantled ground squirrels ( Spermophilus columbi-
anus). J. Mamm., 61:124-126.

Chapin, F. S., III. 1977. Temperature compensation in phos-
phate absorption occurring over diverse time scales. Arctic
and Alpine Res., 9:139-148.

Church, D. C. 1972. Digestive physiology and nutrition of
ruminants. Published by D. C. Church, Oregon State Univ.,
Corvalis, 3: 1-35 1 .

Clark, J. D. G. 1954. Excavations at Star Carr. Cambridge
Univ. Press, Cambridge, 200 pp.

Edwards, W. E. 1967. The late-Pleistocene extinction and dim-
inution in size of many mammalian species. Pp. 141-154,
in Pleistocene extinctions (P. S. Martin and H. E. Wright,
Jr., ed.), Yale Univ. Press, New Haven, 453 pp.

Garutt, V. E. 1964. Das Mammut (Mammuthus primigenius
Blusmenb.). Neue Brehm-Buherei, Wittenberg-Lutherstadt,
Siemsen, 283 pp.

Geist, V. 1966. The evolutionary significance of mountain sheep
horns. Evolution, 20:558-566.

. 1971. Mountain sheep — A study of behavior and evo-
lution. Univ. Chicago Press, Chicago, 383 pp.

Guilday, J. E., P. W. Parmalee, and H. W. Hamilton. 1977.
The Clark’s Cave bone deposit and the late Pleistocene pa-
leoecology of the central Appalachian Mountains of Virginia.
Bull. Carnegie Mus. Nat. Hist., 2:1-88.

Guthrie, R. D. 1966. The extinct wapiti of Alaska and Yukon
Territory. Canadian J. Zook, 44:47-57.

. 1968. Paleoecology of the large mammal community

in interior Alaska during the late Pleistocene. Amer. Midland
Nat., 79:346-363.

. 1974. Environmental influences on body size, social

organs, population parameters, and extinction of Beringian
mammals. Khabarovsk Conference on Beringia. Soviet
Academy of Science, Vladivostock, 80 pp.

. 1980. Bison and man in North America. Canadian J.

Anthro., 1:55-75.

. 1982. Mammals of the mammoth steppe as paleoen-

vironmental indicators. Pp. 307-326, in Paleoecology of
Beringia (D. M. Hopkins, J. V. Matthews, C. E. Schweger,
S. B. Young, eds.), Academic Press, New York, 489 pp.

. 1984. Mosaics, allelochemics and nutrients: an ecolog-
ical theory of late Pleistocene megafaunal extinctions. In
Pleistocene extinctions (P. S. Martin and R. G. Klein, eds.),
Univ. Arizona Press, Tuscon, second edition.

Hanley, T. A. 1980. Nutritional constraints on food and hab-

1984

GUTHRIE— PLEISTOCENE GIGANTISM

509

itat selection by sympatric ungulates. Unpublished Ph.D.
dissert., Univ. Washington, 176 pp.

Hare, F. K. 1976. Late Pleistocene and Holocene climates:
some persistent problems. Quat. Res., 6:507-517.

Harington, C. R. 1978. Quaternary vertebrate faunas of Can-
ada and Alaska and their suggested chronological sequence.
Syllogeus Series, 15:1-105.

Harris, A. H., and P. Mundel. 1974. Size reduction in big
horn sheep (Ovis canadensis) at the close of the Pleistocene
J. Mamm., 55:678-680.

Heimer, W. E. and A. C. Smith. 1975. Ram horn growth and
population quality— Their significance to Dali sheep man-
agement in Alaska. Alaska Dept. Fish and Game, Wildlife
Tech. Bull. 5:1-41.

Hoefs, M. E. G. 1 974. Food selection by Dali’s sheep Ovis dalli
dalli. Pp. 759-786, in Behavior of ungulates and its relation
to management (V. Geist and F. Walther, eds.), Intemat.
Union Conserv. ofNat. Res., Switzerland, n.s., 24:759-786.

Hopkins, D. M., J. V. Matthews, Jr., C. E. Schweger, and S.
B. Young. 1982. Paleoecology of Beringia. Academic Press,
New York, 489 pp.

Horejsi, B. 1976. Suckling and feeding behavior in relation to
lamb survival in bighorn sheep. (Ovis canadensis canadensis
Shaw). Unpublished Ph.D. thesis, Univ. Calgary, Calgary,

201 pp.

Isdo, S. B. 1976. A perspective on the role of climatic dust in
climatic change theory abstracts. Amer. Quat. Assoc.
(AMQUA), 4th biennial meeting, Tempe, Arizona, pp. 59-
60.

Janis, C. 1975. The evolutionary strategy of the Equidae and
the origins of the rumen and caecal digestion. Evolution, 30:
757-774.

Kjell, D. J., and J. S. Millar. 1978. Growth of juvenile arctic
ground squirrels (Spermophilus parryi) at McConnell River
N.W.T. Canadian J. Zool., 56:1475-1478.

Klein, D. R. 1965. Ecology of deer range in Alaska. Ecol.
Monog., 35:259-284.

. 1970. Tundra ranges of the boreal forests. J. Range.

Mgmt., 23:8-14.

Koeppl, J. W., and R. S. Hoffmann. 1981. Comparative growth
rates (postuatae) of four species of ground squirrels. J. Mamm.,
61:41-57.

Kurten, B. J. 1968. Pleistocene mammals of Europe. Wei-
denfield and Nicolson, London, 317 pp.

La Roi, G. H. 1982. From aspen parkland to the Arctic coast:
A vegetation transect. (Abstracts) Annual Meeting, Society
of Range Management, Calgary, Alberta, 1982.

Levenson, H. 1979. Sciuvid growth rates: some corrections and
additions. J. Mamm., 60:232-234.

Matthews, J. V. 1974. Wisconsin environment of interior
Alaska; pollen and macrofossil analysis of a 27-meter core
from Isabella Basin (Fairbanks, Alaska). Canadian J. Earth
Sci., 1 1:828-841.

Mattson, W. J. 1980. Herbivory in relation to plant nitrogen
content. Ann. Rev. Ecol. Sysf, 11:1 19-162.

McDonald, J. N. 1981. The Northern American Bison: Their
classification and evolution. Univ. of California Press,
Berkeley, 316 pp.

Mrosovsky, N. 1976. Lipid programmes and life strategies in
hibemators. Amer. Zool., 16:685-697.

Murie, A. 1944. The wolves of Mount McKinley. Fauna Ser.,
U.S. Fauna National Parks, 5:xix + 1-238.

Murie, J. O., D. A. Boag, and V. K. Kievett. 1980. Litter size
in Columbian ground squirrels (Spermophilus columbianus).
J. Mamm., 61:237-244.

Murie, J. O., and M. A. Harris. 1982. Annual variation of
spring emergence and breeding in Columbian ground squir-
rels ( Spermophilus columbianus). J. Mamm., 63:431-439.

Murphy, E. C. 1974. An age structure and a reevaluation of
the population dynamics of Dali sheep (Ovis dalli dalli).
Unpublished M.S. thesis, Univ. Alaska, Fairbanks, 1 13 pp.

Nichols, L., Jr. 1978. Dali’s sheep. Pp. 179-190, in Big game
in North America (J. L. Schmidt and D. L. Gilbert, eds.),
Stackpole Books, Harrisburg, 494 pp.

Norden, H. C., I. M. Cowan, and A. T. Wood. 1968. Nutri-
tional requirements of blacktailed deer in captivity. Pp. 89-
96, in Comparative nutrition of wild animals (M. A. Craw-
ford, ed.), Symp. Zool. Soc. London, 21.

NRC. 1971. Recommended nutrient allowances for sheep. Nat.
Academy Sci., Washington, D.C., 107 pp.

Oleberg, K. 1956. Factors affecting the nutritive value of range
forage. J. Range. Mgmt., 9:220-225.

Palmer, L. J. 1944. Food requirements of some Alaskan game
mammals. J. Mamm., 25:49-54.

Peden, D. G., G. M. Van Dyne, R. W. Rice, and R. M. Hansen.
1974. The trophic ecology of Bison bison in short grass
prairie. J. App. Ecol., 1 1:489-498.

Pewe, T. L. 1975. Quaternary geology of Alaska. U.S. Geol.
Soc. Prof. Paper, 835:1-212.

Powers, W. R., R. D. Guthrie, and J. F. Hoffecker. In press.
Dry Creek: archeology and paleoecology of a late Pleistocene
Alaskan hunting camp.

Schmidt, J. L., and D. L. Gilbert. 1978. Big game of North
America: ecology and management. Stackpole Books, Har-
risburg, 494 pp.

Shackelton, D. M., and L. V. Hills. 1977. Post-glacial un-
gulates (Cervus and Bison) from Three Hills, Alberta. Ca-
nadian J. Earth Sci., 14:963-986.

Sher, A. V. 1971. Pleistocene mammals and stratigraphy of
the far Northeast USSR and North America. Geol. Inst.,
USSR Academy of Sciences, Moscow, 310 pp. (Trans. Ameri.
Geol. Inst., Published 1974, in International Geol. Review,
16:1-284).

Skogland, T., and R. G. Petocz. 1975. Ecology and behavior
of Marco Polo sheep (Ovis ammon poll) in Pamir during
winter. Report from FAO/UNPD Project: Conservation and
Utilization of Wildlife Resources, 23 pp.

Stock, A. D., and W. L. Stokes. 1969. A re-evaluation of
Pleistocene bighorn sheep from the Great Basin and their
relationship to living members of the genus Ovis. J. Mamm.,
54:805-807.

Stoddart, L. A., and A. D. Smith. 1955. Range Management.
McGraw-Hill, New York, second ed., 433 pp.

Summerfield, B. C. 1974. Population dynamics and seasonal
movement patterns of Dali sheep in the Atigun Canyon area
Brooks Range Alaska. Unpublished Master Sci. thesis, LJniv.
Alaska, College, 212 pp.

Vereschagin, N. K. 1967. The mammals of the Caucasus.
Izadtel’stvo Akademu Nauk. (SSSR English Translation) Is-
rael Program for Scientific Translations, Jerusalem. 89 pp.

Viereck, L. A. 1963. Sheep and goat investigations: range sur-
vey. Unpublished report on project W-6-R-3, E, 2-a. Alaska
Dept. Fish and Game, Juneau.

Wallach, J. D. 1972. Nutrition and feeding of captive rumi-

510

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

nants in zoos. Pp. 292-307, in Digestive physiology and
nutrition of ruminants (D. C. Church, ed.), Corvalis, Oregon,
780 pp.

Westoby, M. 1974. An analysis of diet selection by large gen-
eralist herbivores. Amer. Nat., 108:290-304.

. 1978. What are the biological bases of varied diets?

Amer. Nat., 112:627-631.

Whithead, D. R. 1973. Late- Wisconsin vegetational changes
in unglaciated eastern North America. Quaternary Res., 3:
621-631.

Whitten, K. 1975. Habitat relationships and population dy-
namics of Dali sheep (Ovis dalli dalli ) in Mt. McKinley Na-

tional Park. Unpublished M.S. thesis, Univ. Alaska, Fair-
banks, 1 17 pp.

Winters, J. 1980. Summer habitat and food utilization by Dali
sheep and its relation to body and horn size. Unpublished
M.S. thesis, Univ. Alaska, Fairbanks, 109 pp.

Woodgerd, W. 1964. Population dynamics of bighorn sheep
on Wild-horse Island. J. Wildlife Mgmt., 28:381-391.

Wrazen, J. A., and L. A. Wrazen. 1982. Hoarding, body mass
dynamics, and torpor as components of the survival strategy
of the eastern chipmunk. J. Mamm., 63:63-72.

Zeldin, T. 1983. The French. Collins, London, 542 pp.

Address: Institute of Arctic Biology, University of Alaska, Fairbanks, Alaska 99701.

QUATERNARY MARINE AND LAND MAMMALS AND THEIR
PALEOEN VIRONMENT AL IMPLICATIONS - SOME
EXAMPLES FROM NORTHERN NORTH AMERICA

C. R. Harington

ABSTRACT

Workers on Quaternary mammals should try to obtain as much
information as possible from fossil evidence available— espe-
cially from well-preserved specimens. Examples discussed here
are drawn from marine (seal and walrus) and land mammal re-
mains (mammoth, stag-moose, saiga antelope, and arctic ground
squirrel) from northern North America.

It is particularly important that: specimens are correctly iden-
tified; stratigraphic and sedimentological situations of skeletons
found in place are carefully assessed; associated plant and animal
remains are evaluated for paleoenvironmental clues; direct and
indirect geochronological data applying to the specimens are
gathered— where possible testing the validity of such data by
cross-checking; bone surfaces are closely examined so that signs

of predator-prey relationships, scavenging, and disease may be
detected; an effort is made to establish individual ages of the
specimens (for example, by tooth-sectioning); in the case of un-
usually well-preserved specimens, stomach contents, droppings
and soft parts are analyzed in order to gain information on feeding
habits and other ecological parameters.

Above all. Quaternary paleobiologists should view these long-
dead animals as realistically as possible, trying to see their place
in the web of life— as mammalogists view living species. Finally,
we must try to pass on significant paleoenvironmental infor-
mation in an appealing way to the public — they should be aware
of the great changes that can occur in landscapes and their biota
over even relatively short periods of geological time.

INTRODUCTION

As a vertebrate paleontologist working on Qua-
ternary mammals, I am concerned that as much
evidence as possible on the habitat of these mam-
mals is collected, and that it is properly analyzed.
The closer we can come to discovering the natural
habits and habitats of these mammals the better.
Both marine and land mammals deserve study in
this respect. It is particularly important to gather
such information on articulated skeletons found in
primary position, and from associated fossils in en-
closing sediments. Indeed, the nature of the sedi-
ments themselves may help to explain the environ-
ment of deposition. Although whatever is discovered
about the environment of deposition of a species is
not necessarily typical of that species’ habitat, it is
important to accumulate such paleoenvironmental
evidence in the hope that habitat preferences of cer-
tain Pleistocene mammals may be more precisely
defined.

Sometimes, fossils of mammalian species with
well-defined adaptations and habitat needs provide
clues to the nature of past environments in an area —
the basic assumption being that species represented
by fossils had ecological requirements similar to
those of the same or closely allied living species.
Among the cases mentioned here, perhaps the ringed

seal and saiga antelope best exemplify this kind of
evidence (Fig. 1).

Occasionally, other fossils, such as plant macro-
fossils, pollen, insect remains, and mollusc shells
directly associated with ancient skeletons shed light
on habitats of extinct mammals. Examples used here
are the Babine Lake mammoth and the stag-moose
( Cervalces scotti ) from Columbia, New Jersey.

Rarely, nearly complete animals, including stom-
ach contents, droppings and other soft parts of ice
age mammals are preserved, providing direct evi-
dence of their environmental adaptations and eating
habits. The Berezovka mammoth from frozen
ground in Siberia is an outstanding international
example (Augusta and Burian, 1963). The potential
of this kind of evidence is seen in a 12,000-year-old
arctic ground squirrel skeleton found in its nest in
frozen ground near Dawson, Yukon.

I wish to emphasize that significant paleoenvi-
ronmental information should not be restricted to
scientists. We must be concerned with passing on
our information in an appealing way to the public:
they should be aware of the great changes that can
occur in landscapes and their biota over even rel-
atively short periods of geological time.

511

512

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

NORTH AMERICA

No. 2

Fig. 1 . — Localities of specimens mentioned in text. Key: 1) Columbia, New Jersey (stag-moose, Cervalces scotti)\ 2) Plattsburgh, New
York (harbour seal, Phoca vitulina)\ 3) Hull, Quebec (ringed seal, Phoca hispida)\ 4) Baillie Islands, Northwest Territories (saiga antelope.
Saiga tatarica ); 5) Glacier Creek, Yukon Territory (ground squirrel carcass, Spermophilus parryi)\ 6) Dominion Creek, Yukon Territory
(ground squirrel nest and skeleton, Spermophilus parryi)\ 7) Babine Lake, British Columbia (Columbian mammoth, Mammuthus cf.
columbi)\ 8) Qualicum Beach, British Columbia (walrus, Odobenus rosmarus)\ 9) San Francisco Bay, California (walrus, Odobenus
rosmarus).

1984

HARINGTON-QUATERNARY PALEOENVIRONMENTS

513

Fig. 2. — Left side of a ringed seal (Phoca hispida; NMC 6830) skeleton from Champlain Sea deposits at Hull, Quebec.

EXAMPLES

Marine Mammals

First, I want to focus on marine mammals of the
Champlain Sea (an inland sea that covered the St.
Lawrence Lowlands from about 12,000 to 10,000
years ago). Although many species of marine mam-
mals that once lived in this sea now occupy the Gulf
of St. Lawrence, at least three, the bowhead whale
(Balaena mysticetus), ringed seal ( Phoca hispida ),
and bearded seal ( Erignathus barbatus) are closely
adapted to arctic waters (Mansfield, 1967).

A skeleton of a ringed seal (Fig. 2) found in place
with Hiatella arctica shells in marine clay at Hull,
Quebec, was important in determining the coldness

of the Champlain Sea’s early stages. Ironically, be-
cause the specimen was incorrectly identified [as
probably a young harp seal ( Phoca groen/andica ),
then as a young harbour seal ( Phoca vitulina ), even-
tually being displayed in the National Museum of
Canada as a young harp seal], it was not until some
80 years after collection that its paleoenvironmental
implications were realized! A detailed study of the
specimen (Harington and Sergeant, 1972) showed
that it was a ringed seal, and, despite its small size,
an adult about seven years old. Ringed seals are now
the commonest and most widespread arctic seals
(Fig. 3). They are adapted to keep breathing holes

514

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY NO. 8

Fig. 3. — Ringed seal (Phoca hispida). Courtesy of New York Zoological Society, Carleton Ray.

open in sea ice during winter and to pupping on fast
ice (sea ice attached to the coast) in spring. So, it is
reasonable to conclude that fast ice clung to the
northern coast of the sea. Perhaps the individual
was so small for its age because of poor nourishment
caused by early separation from its mother. Such
cases occur on relatively straight coasts with unsta-
ble fast ice, and these conditions may have prevailed
along the northern margin of the Champlain Sea.
Finally, it is worth noting that corroboration of an
arctic phase of the sea has since come from studies
of marine ostracodes (Cronin, 1981). According to
these studies, conditions during the Hiatella arctica
phase (from 1 1,600 to about 1 1,000) were frigid to
subfrigid, with bottom temperatures of 0 to 1 2°C,
and normal marine salinities (18-35 PPT). There-
fore, if I were asked to predict the geological age of
the Hull specimen, I would be inclined to place it
between 1 1,600 and approximately 1 1,000 B.P.

Again, the importance of correct identification of
specimens is shown in the case of another Cham-
plain Sea seal fossil found in 1901 at Plattsburgh,
New York (Bishop, 1921; Hartnagel and Bishop,
1921). A tibial shaft of a juvenile, considered for
many years to represent a hooded seal (Cystophora
cristata ), has been relocated and reidentified as part
of a harbour seal ( Phoca vitulina ; NYSM 892-D/
8569; Fig. 4) (Ray, 1983). T. M. Cronin identified
the following benthic foraminifera from a clay sam-
ple with the bone— Haynesia obiculare, and El-

phidium cf. albiumbiculatum, suggesting that the clay
was deposited during the Hiatella arctica or Mya
arenaria (approximately 1 1,000-10,000 B.P.) phase
of the Champlain Sea. Evidently, the harbour seal
(Fig. 5) occupied shallow water along the western
margin of the sea’s southern arm. Such an environ-
ment corresponds to the known habitat preference
of harbour seals today— and is quite different from
that of the hooded seal (Mansfield, 1967).

As a museum scientist, I feel it is particularly
important to transmit significant paleoenvironmen-
tal information in an appealing way to the public.
If someone asked me what the Champlain Sea would
have looked like about 1 1 ,000 years ago, I would
try to convey the following impression based on
fossil finds made during the past 130 years (Har-
ington, 1 980a; Mott et al., 1981), and using modern
observations of the species concerned to achieve a
more vivid result:

“Under the bright summer sun, blue-green water scattered with
pans of ice extends in a broad sweep eastward. To the north,
partly bounded by freshly exposed, rounded granitic bluffs lies
the dazzling Laurentide ice sheet, streaked and smudged along
its edge by gray heaps of morainic debris. Some fast ice still clings
to the north shore, where several ringed seals bob in the cold sea
off the floe-edge. They twist and turn, feeding on abundant shrimp-
like crustaceans. Two small flocks of Common Eiders whistle
over the wavelets stirred up by cold winds draining off the edge
of the ice sheet. They curve toward a low, rocky island where
females are nesting. Other eiders in the colony dive in shallow
waters nearby for clam-like molluscs.

1984

H ARINGTON - QUATERNARY PALEOENVI RONMENTS

515

Fig. 4. — Localities of seal remains from Champlain Sea deposits. Stippling indicates the approximate maximum limit of Champlain
Sea. The cluster of symbols at the centre marks Montreal, and at the left marks Ottawa. Key: ringed seal, Phoca hispida (black star);
harp seal, Phoca ( Pagophilus ) groenlandica (white dot); bearded seal, Erignathus barbatus (white square); harbour seal, Phoca vitulina
(white star); unidentified seal remains (black dot).

In deeper water several kilometres off the sea’s northern coast,
a massive black shape cruises at the surface, periodically blowing
a v-shaped mist into the air. The migrating bowhead whale moves
steadily westward toward the broad sloping sands of the Peta-
wawa Delta, the sun making rainbows in its moist breath. The
first white whales have arrived from the Gulf of Saint Lawrence.
A pod of 20 appear from above like a corps de ballet. They slide
smoothly through the cool greenish water, their gleaming white
backs periodically breaking the surface sending chevron-like wakes
behind. A few new-born calves, like small shadows, press close
to the females. Their heart-shaped flukes beat rapidly to keep up.
Hundreds of belugas already loll in shallow, river-warmed waters
in the southern embayment of the sea, where Lake Champlain
now lies.

On the sandy beaches of the south shore, small birds pecking
fitfully with their sharp bills scuttle back and forth in harmony
with the advancing and retreating waves. Dead starfish and stray
patches of kelp are partly exposed on the beach. Just off-shore.

the sea is alive with spawning capelin. Their silvery bodies, in
hundreds, rise and fall with the waves. Multitudes lie rotting in
rows at the level of the last high tide, their stench overriding the
salt tang of the onshore breeze. Through the glare of sun on the
water, silhouettes of about a dozen harp seals can be seen leaping
and cavorting. A solitary bearded seal, whiskers glistening white
against its rotund gray body, hauls out on a nearby sandbar to
rest.

On the northeastern margin of the sea, several of the Monter-
egian Islands (tops of the volcanic mounds of Mount Royal, Mont
St. Hilaire, etc.) appear darkly below. Waves roll in strongly on
their shores, churning up and scattering chalky, quadrangular
valves of Hiatella arctica. Farther east near what is now Quebec
City, tundra stretches northward from a large coastal lagoon to
the ice front some 13 kilometres away. The taller grasses and
sedges near the edge of the lagoon quiver in the cold breeze.
Dense mats of mouse-ear chickweed almost cover the gravel
there. Patches of creamy mountain avens contrast with the waxy

516

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 5.— Harbour seal ( Phoca vitulina). Courtesy of New York Zoological Society, Carleton Ray.

green of shrub willows and the duller green of crimson-flowered
mountain sorrel. A hare, sheltering behind a granite boulder,
nibbles on succulent willow shoots, its ears delicately attuned to
each new sound. Frightened, it zig-zags toward the end of the
lagoon and pauses. There, brilliant flashes of yellow mark the
tall flowering heads of groundsel, which dominates the surround-
ing moist, peaty ground. Southward, across the 60-kilometre-
wide threshold to the Champlain Sea, lies more tundra. Only in
the valleys far to the south do stunted spruce and birch trees
mark the pioneering northern fringe of the Boreal Forest, in whose
shade stalks a beady-eyed marten.”

Many problems remain concerning this inland sea
and its environment. Why have no walrus fossils
been reported from Champlain Sea deposits when
specimens are known from 1 5 localities (see for ex-
ample Harington, 1977: tig. 3) in the eastern ap-
proaches to the sea? Important habitat require-
ments, such as accessible banks of abundant molluscs
(on which walruses are known to feed), and suitable
hauling-out grounds (particularly near the north
shore of the sea), seem to have been present. Fur-
thermore, walrus bones tend to preserve well, es-
pecially skulls.

Continuing on the topic of ice age walruses, but
shifting to the Pacific coast, I wish to comment on
a remarkable walrus skeleton (Fig. 6) found near
Qualicum Beach, Vancouver Island. The skeleton
was found in place in Dashwood glaciomarine clay-
ey silt of early Wisconsin age (S. Hicock, personal
communication, 1 98 1 ). A radiocarbon date on bone

collagen of >40,000 B.P. (1-11617) supported the
relatively early age of this specimen based on its
stratigraphic position. Further, finding the specimen
in the basal clay is in accord with features of this

Fig. 6. — Right side of walrus (Odobenus rosmarus\ NMC 38490)
cranium. This is part of an individual skeleton found in early
Wisconsin glaciomarine clay near Qualicum Beach, British Co-
lumbia.

1984

HAR1NGTON — QUATERNARY PALEOENVIRONMENTS

517

unit as described by Fyles (1963) (that is, an epi-
neritic to littoral marine clay containing dropstones
probably derived from floating sea ice, and abun-
dant well-preserved pelecypod shells). Mollusc as-
semblages collected from the unit are living today
in the southern part of Bering Sea (Wagner, 1959),
which approximates the southern limit of Pacific
walrus ( Odobenus rosmarus divergens) range. In
other words, the stratigraphic, geochronological, and
paleontological evidence seems to hang together. In
the hope that more paleoenvironmental (particu-
larly paleotemperature) data could be gained, I ex-
cavated a block of clay from directly beneath the
skeleton, and passed it on to Sigrid Lichti-Federo-
vich for analysis of marine microorganisms. Un-
fortunately only three species of marine diatoms
(Plagiogramma staurophorum, Amphora sp., and
Grammatophora sp.) were noted, and the paucity
of specimens precluded a meaningful paleoenviron-
mental interpretation based on this evidence. Suffice
to say that walruses lived far south of their present
range during late Pleistocene glacial advances (for
example, records from San Francisco Bay, Fig. 7;
Kittyhawk, North Carolina; Montrouge near Paris;
and Tokyo, Harington, 1975, and this paper).

Land Mammals

Paleoenvironmental evidence derived from Qua-
ternary land mammals is also of interest, and I wish
to mention a few pertinent cases.

Remains of a partly articulated mammoth skel-
eton were exposed during stripping at a mining site
on Babine Lake, central British Columbia. The bones
lay in silty pond deposits in a bedrock depression,
and were overlain by a thin layer of gravel and a
thick layer of glacial till. Although no molar teeth
were found, limb proportions indicated that the
specimen was a large mammoth, like the Columbian
mammoth ( Mammuthus cf. columbi). Two radio-
carbon dates of 42,900 ± 1860 B.P. (GSC-1657)
and 43,800 ± 1830 B.P. (GSC- 1 687) on wood from
the silty layer, and another of 34,000 ± 690 B.P.
(GSC- 1 754) on mammoth bone suggested that, dur-
ing this part of the Olympia Interglaciation, the
vegetation near Babine Lake was similar to present
shrub tundra just beyond the treeline in northern
Canada (Fig. 8). Probably birch and willow shrubs
bordered the pond and were scattered over the land-
scape. Open ground supported abundant grasses,
Artemisia and other composites, various members
of the rose family, pinks, willow-herb, and several
other herbs. Although the exact contemporaneity of

the plant and mammoth remains is questionable,
both types of fossils suggest the presence of grassland
in the Babine Lake area from about 34,000 to 43,000
B.P. (Harington et al., 1974).

I am presently studying, in collaboration with Gary
J. Sawyer of the Natural Science Museum, Blairs-
town, New Jersey, a virtually complete skeleton of
an extinct stag-moose ( Cervalces scotti\ NSM 264)
from a bog near Columbia, New Jersey (Figs. 9-10).
This case is noteworthy because it is providing the
first clear evidence on the habitat of this moose-like
animal. The specimen was excavated from the base
of a marl layer about 1 m thick, that was overlain
by approximately 1 m of consolidated peat and about

O. 3 m of rich black loam. A “leaf layer” containing
abundant remains of Potamogeton and Najas, com-
monly found on pond or lake margins (J. V. Mat-
thews, Jr., personal communication, 1979), occur-
ring just below the bones yielded a radiocarbon date
of 1 1,600 ± 170 B.P. (1-1 1335). A bone sample gave
a date of 1 1,230 ± 160 B.P. (1-1 1286). Dominant
species of gastropod molluscs in the marl surround-
ing the Cervalces bones are Gyraulus parvus and
Valvata tricarinata, suggesting a large cold lake with
abundant vegetation and a muddy bottom with some
stones (M. F. I. Smith, personal communication,
1979), which seems to fit the paleobotanical evi-
dence. Among the sphaeriids, Pisidium variabile and

P. nitidum were most common, whereas P. ferru-
gineum, P. casertanum, Musculium securis,
Sphaerium rhomboideum, and S. simile were rarer
(G. L. Mackie, personal communication, 1979). The
dominant species ( Pisidium variabile) presently oc-
curs in a variety of habitats, but Sphaerium rhom-
boideum may define the type of habitat more closely
(an environment of deposition in a eutrophic lake
less than 5 m deep, pH 7-9, dissolved oxygen >6
mg/1, total alkalinity 60-250 mg CaC03/l, free
co2 <15 mg/1). Looking at a map showing stages
in the retreat of the Laurentide ice sheet (Prest, 1 969),
I calculate that some of these moose-like animals
lived about 600 km south of the southern margin
of the ice approximately 1 1,000 years ago. Cervalces
scotti, according to the above radiocarbon dates, was
contemporaneous with the American mastodon
(Mammut americanum) in this part of New England
(Parris, 1983).

Turning from eastern North America to the ex-
treme northwestern part of this continent, I wish to
mention the apparent intrinsic value of saiga ante-
lope ( Saiga tatarica\ Fig. 11) bones as paleoenvi-
ronmental indicators. Nine saiga specimens are

518

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 7.- Front view of an anterior cranial fragment of a walrus (Odobenus rosmarus\ California Academy of Sciences 16677) dredged
from the bottom of San Francisco Bay, California. A radiocarbon date of 27,200 ± 950 B.P. (1-9994) was obtained on the tusk. A
check date using the accelerator technique may be advisable.

1984

H ARINGTON - QUATERNARY PALEOEN VI RON MENTS

519

Fig. 8. — Restoration of the Babine Lake mammoth ( Mammuthus cf. columbi) as it may have appeared in its natural shrub tundra
environment during the Olympia Interglaciation. Ink sketch by Charles Douglas.

Fig. 9. — Restoration of the extinct stag-moose (Cervalces scotti). Pictured under winter conditions in this sketch by R. Bruce Horsfall.
Copyright 1937 by the American Philosophical Society.

520

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 10. — Top view of cranium with partial antlers of a stag-moose (Cen’alces scottv, NSM 264) excavated with most of the remainder
of the skeleton from a bog near Columbia, New Jersey. Lighter areas are restored. A bone sample from the specimen yielded a radiocarbon
date of 11,230 ± 160 B.P. (1-1 1335).

known from late Pleistocene deposits in Eastern
Beringia (six from central Alaska, two from northern
Alaska— one dated at 37,000 ± 900 B.P. (GSC-
3050)— and one from Baillie Islands, Northwest
Territories; Fig. 12). Probably most are of Wiscon-
sin age. Because living saigas of central Eurasia are
particularly adapted to dry steppe-grasslands, I sug-
gest that they crossed broad, steppe-like plains of
the northern Bering Isthmus during glacial phases
of the late Pleistocene (Fig. 1 3). Presumably the kind
of northern steppe to which they had adapted once
extended eastward, up the Yukon valley to central
Alaska, and along the Arctic Coastal Plain to Baillie
Islands in Canada. Saiga antelope remains appear

to be useful indicators suggesting the presence of
steppe-like vegetation, generally low, flattish terrain,
rather arid climatic conditions, and shallow snow
cover in winter (Harington, 19806, 1981).

Of potential paleoenvironmental interest are sev-
eral nests of late Pleistocene arctic ground squirrels
( Spermophilus parryi) that have been collected from
frozen silts near Dawson, Yukon, during the past
1 5 years. One example from Dominion Creek con-
tained nesting grasses, part of a seed cache, fecal
pellets, and skeletal fragments of an individual
ground squirrel. The nesting grasses gave a date of
12,200 ± 100 B.P. (GSC-2641). Part of the nest
debris was analyzed for fossil pollen and revealed

1984

H ARINGTON - QUATERNARY PALEOEN VIRONMENTS

521

Fig. 11.— Saiga antelope ( Saiga tatarica). This lightly built animal with its lyre-shaped horns and expanded nasal region evidently
occupied steppe-like terrain in northwestern North America during the late Pleistocene.

abundant fungal elements and spores as well as nu-
merous slime mold and bryophyte spores. The an-
giosperm pollen, which composed only about 20%
of the total assemblage, represented several grasses
(Gramineae), cotton grass ( Eriophorum sp.), willow
( Salix sp.), Chenopodiaceae/Amaranthaceae, sev-
eral species of Compositae, and much sage ( Arte-
misia sp.) (D. M. Jarzen, personal communication,
1975). Pollen analysis of another nest from Hunker
Creek near Dawson yielded 73.5% grasses (Gra-
mineae), 14.6% sage ( Artemisia sp.), 7.4% Cheno-
podiaceae, 2.6% willow ( Salix sp.), 1.0% Compos-
itae, 0.3% birch ( Betula sp.), traces of Lycopodium
sp., Botrychium sp., Sphagnum sp., and a great

abundance of fungal spores (C. McAtee, personal
communication, 1979). As more work is completed
on these nests (keeping in mind modern analogues,
for example, Krog, 1954; Bee and Hall, 1956; Ban-
field, 1974), probably we will be able to tell exactly
what grasses were chosen for nests, what plants and
fungi were eaten, and what kinds of seeds were
cached. It may be possible to recreate, on this basis,
a microcosm of the paleoenvironment of this part
of Eastern Beringia toward the close of the last gla-
ciation. Indeed, these ground squirrels could be con-
sidered as small, furry botanists, industriously sam-
pling vegetation within a limited range of their nests
and storing it away in underground herbaria.

522

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 12. — Left side of left homcore with attached portions of
frontal and parietal bones of a saiga antelope ( Saiga tatarica\
NMC 12090) from Baillie Islands, Northwest Territories. It is
the first recorded saiga specimen from Canada.

In 1981, M. Peschke, a placer miner on Glacier
Creek, Yukon Territory, picked up a ball of silt with
bits of hair projecting from it. I later found that it
was a carcass of a ground squirrel ( Spermophilus
parryr. Fig. 14). An x-ray showed that a virtually
complete skeleton lay inside the skin and hair. The
animal had been curled up as if it had died during

hibernation — perhaps more than 10,000 years ago
(Fig. 1 5). A detailed study of this specimen may also
yield valuable paleoenvironmental evidence.

CONCLUSION

Workers on Quaternary mammals should try to
wring a maximum of information from the fossil
evidence available — particularly from significant,
well-preserved specimens. I suggest that valid goals
are: (a) correct identification of the mammal— even
fragments of some species, such as homcores of the
saiga antelopes can have intrinsic value as paleoen-
vironmental indicators; (b) careful assessment of the
stratigraphic and sedimentological situation of skel-
etons found in place — sedimentology often yields
data on environment of deposition; (c) collection of
associated plant and animal macrofossil remains, in
addition to substantial samples of matrix enclosing
the bones, in case they may yield valuable paleoen-
vironmental clues; (d) gathering direct (for example,
radiocarbon dates on bone) and indirect geochrono-
logical data (for example, radiocarbon dates on as-
sociated mollusc shells or plants, dates on enclosing
tephra, and other material), where possible making
an effort to cross-check such data; (e) close exami-
nation of bone surfaces so that relationships be-
tween prey and predators or scavengers (including
humans), or possible effects of disease may be de-
tected; (f) to establish the age of the individual rep-
resented by noting the degree of tooth wear and
suture fusion, and by sectioning teeth and other
methods; (g) analyses of fecal droppings, stomach
contents and ancient “tooth-jam”1 in unusually well-
preserved specimens, in order to gather data on eat-
ing habits of the species in question.

The most reliable paleoenvironmental evidence
is that which is clearly associated with a specimen,
is readily identifiable, abundant, undisturbed, based
on various sources, and which, ultimately, is com-
patible. Above all, it is important for Quaternary
paleontologists to view these long-dead animals as
realistically as possible, trying to see their place in
the web of life— as mammalogists would view living
species.

1 Remains of vegetation eaten by an animal, which may be preserved in “lakes” of
ungulate teeth.

1984

HARINGTON -QUATERNARY PALEOENVIRONMENTS

523

Fig. 13 — Generalized map of Beringia during the peak of the last glaciation showing localities of saiga (Saiga tatarica) fossils (black
dots) (Harington, 1980/?). Arrows indicate possible lowland routes used by saigas moving from western to eastern Beringia via the broad
exposed plains of the Bering Isthmus (right of center). Glaciated areas are shown in black, ocean shorelines are stippled and land is
white. Localities 1 to 4 are in northeastern Siberia, 5 to 9 are in Alaska, and 10 is in the Northwest Territories.

ACKNOWLEDGMENTS

Paleoenvironmental work usually requires a great deal of in-
terdisciplinary cooperation. In this case, I would particularly like
to thank: D. E. Sergeant, Graham Beard, S. Hicock, S. Lichti-
Federovich, H. Tipper, R. J. Mott, J. G. Fyles, Gary J. Sawyer

(who collected and is owner of the Cervalces specimen from near
Columbia, New Jersey), J. V. Matthews, Jr., M. F. I. Smith, G.

L. Mackie, the late Harold Schmidt, D. M. Jarzen, C. McAtee,

M. Peschke, and D. Drummond.

Finally, I would like to pay tribute to the late John Guilday,
who provided me with cheerful advice, help, and encouragement
since I began work in the field of Quaternary mammals.

524

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 14. — Carcass of a ground squirrel ( Spermophilus parryi\ NMC 37557)— perhaps more than 10,000 years old— from organic silt
deposits on Glacier Creek, Yukon Territory. Note fur projecting through silt, and right side of back part of cranium exposed centrally.

LITERATURE CITED

Augusta, J., and Z. Burian. 1963. A book of mammoths. Paul
Hamlyn, London, 50 pp.

Banfield, A. W. F. 1974. The mammals of Canada. Univ.

Toronto Press, Toronto, 438 pp.

Bee, J. W., and E. R. Hall. 1956. Mammals of northern Alaska
on the Arctic Slope. Univ. Kansas Mus. Nat. Hist., Misc.
Publ., 8:1-309.

Bishop, S. C. 1921. Remains of a fossil phocid from Plattsburg,
New York. J. Mamm., 2:170.

Cronin, T. M. 1981. Paleoclimatic implications of late Pleis-
tocene marine ostracodes from the St. Lawrence Lowlands.
Micropaleontology, 27:384-418.

Fig. 15. — X-ray image of ground squirrel (Spermophilus parryi\
NMC 37557) from Glacier Creek, Yukon Territory, suggesting
that the animal had died curled up nose to tail in hibernating
position.

1984

HARINGTON-QUATERNARY PALEOENVIRONMENTS

525

Fyles, J. G. 1963. Surficial geology of Home Lake and Parks-
ville map-areas, Vancouver Island, British Columbia. Mem.
Geol. Surv. Canada, 318:1-142.

Harington, C. R. 1975. A postglacial walrus from Bathurst
Island, Northwest Territories. Canadian Field-Nat., 89:249-
261.

. 1977. Marine mammals in the Champlain Sea and the

Great Lakes. Ann. New York Acad. Sci., 288:508-537.

. 1 980a. Whales and seals of the Champlain Sea. Trail &

Landscape, 15(1): 3 2-4 7.

. 19806. Radiocarbon dates on some Quaternary mam-
mals and artifacts from northern North America. Arctic, 33:
815-832.

. 1981. Pleistocene saiga antelopes in North America

and their paleoenvironmental implications. Pp. 193-225, in
Quaternary paleoclimate (W. C. Mahaney, ed.. Geo Books,
Univ. East Anglia, Norwich, 464 pp.

Harington, C. R., and D. E. Sergeant. 1972. Pleistocene
ringed seal skeleton from Champlain Sea deposits near Hull,
Quebec— a reidentification. Canadian J. Earth Sci., 9:1039-
1051.

Harington, C. R., H. W. Tipper, and R. J. Mott. 1974. Mam-
moth from Babine Lake, British Columbia. Canadian J. Earth
Sci., 1 1:285-303.

Hartnagel, C. A., and S. C. Bishop. 1921. The mastodons,
mammoths and other Pleistocene mammals of New York
State. Bull. New York State Mus., 241-242:1-1 10.

Krog, J. 1954. Storing of food items in the winter nest of the
Alaskan ground squirrel, Citellus undulatus. J. Mamm., 35:
586.

Mansfield, A. W. 1967. Seals of arctic and eastern Canada.
(2nd edition, revised). Bull. Fish. Res. Board Canada, 137:
1-35.

Mott, R. J., T. W. Anderson, and J. V. Matthews, Jr. 1981.
Late-glacial paleoenvironments of sites bordering the Cham-
plain Sea based on pollen and macrofossil evidence. Pp. 1 29-
171, Quaternary paleoclimate (W. C. Mahaney, ed.). Geo
Books, Univ. East Anglia, Norwich, 464 pp.

Parris, D. C. 1983. New and revised records of Pleistocene
mammals of New Jersey. The Mosasaur, 1:1-21.

Prest, V. K. 1969. Retreat of Wisconsin and Recent ice in
North America. Geol. Surv. Canada Map, 1257A.

Ray, C. E. 1983. Hooded seal, Cystophora cristata : supposed
fossil records in North America. J. Mamm., 64:509-512.

Wagner, F. J. E. 1959. Palaeoecology of the marine Pleistocene
faunas of southwestern British Columbia. Bull. Geol. Surv.
Canada, 52:1-67.

Address: National Museum of Natural Sciences, National Museums of Canada, Ottawa K1 A 0M8, Canada.

PHYLETIC TRENDS AND EVOLUTIONARY RATES

Larry D. Martin

ABSTRACT

Single lineage phylogenies are compared to branching phylog-
enies. Such phylogenies are as theoretically sound and testable
as are comparable branching phylogenies. Single lineage phylog-
enies provide unique data concerning evolutionary rates and pro-
cesses. They have direct bearing on models of “punctuated equi-
librium” and the importance of natural selection in evolution.
Single lineage phylogenies also yield our most exact biostratigra-

phies and might be utilized to form a new scheme of absolute
dating.

Examples of single lineage phylogenies are drawn from the
saber-toothed cats and arvicolid rodents. These examples show
long-term evolutionary trends which vary dramatically in their
rates of change.

INTRODUCTION

Many workers have accepted two assumptions
about the fossil record that seem to be mutually
exclusive. One of these assumptions is that the fossil
record is so good that major evolutionary changes,
without gradual intermediaries, must be the result
of unusual evolutionary processes (that is, macro-
evolution) rather than simply the result of missing
sediments and their correlated geological time. The
other assumption is that the fossil record is so bad
that it is impossible to draw evolutionary lineages
and hence the fossil record can give no useful clues
to evolutionary history that cannot be better ob-
tained from the study of modern biology (Patterson,
1981). The truth probably lies between these two
viewpoints, and I will present evidence for phylo-
genetic lineages and argue that they do provide in-
sight into evolutionary processes. On the other hand

these lineages seem to require no special evolution-
ary models and are gradual and overlapping when
they can be examined in detail. Changes without
intermediate forms in these examples seem to be
the result of missing geological time. Many of the
objections to lineage reconstruction are based on the
idea that the recovery of a member of an ancestral
population is an unlikely event. This objection may
be valid for poorly known groups whose samples
are separated by many millions of years, and lineage
hypotheses in such situations are usually stated
within the framework of higher taxonomic units.
These hypotheses are useful as long as their real
nature is understood. Highly refined lineages may
sometimes be proposed. The recognition and use of
such lineages is a promising area of vertebrate pa-
leontology.

LINEAGE PHYLOGENIES

Lineage phylogenies imply that the direct ances-
tors of a taxon are contained within earlier taxa in
the phylogeny. They do not imply that all species
within that taxon (for instance a genus) must be
ancestral to a subsequent taxon (another genus) any-
more than that all subspecies within a species are
in themselves directly ancestral to a later species.
This is a general statement of ancestry, and the direct
ancestor at the species level does not have to be
identified. Statements of general ancestry may be
falsified by showing that the common ancestor of
two taxa (one of which is thought to be ancestral)
must have preceded in time the development of the
characters that are thought to unite them.

Special phylogenies may be proposed on a species

to species basis for segments of evolutionary lines.
These phylogenies can be rigorously tested and de-
fended. Objections to ancestor recognition are nu-
merous in the cladistic literature, but it seems that
many of these objections are due to an unnatural
fear of paraphyletic groups (taxa that do not include
all descendents of the common ancestor). This is a
difficult policy to understand as all taxa must at
some point have ancestors that are not included in
the same taxon and whatever taxa those ancestors
belong to must then be paraphyletic by definition.
It would thus seem to be a waste of time to argue
whether paraphyletic taxa are “natural” or not, as
they exist in any classification containing ancestral
forms. However, it would seem to be a fear of cre-

526

1984

MARTIN- PH YLETIC TRENDS

527

ating paraphyletic taxa that has been the motivation
for arguments that phylogenies should always be
shown on branching diagrams. Branching diagrams
have also been defended as more “probable” and
as testable while single lineages have often been
treated as imaginary constructs without scientific
merit.

Hennig demonstrated that only the presence of
shared homologies at the level of their origin (syn-
apomorphies) can unite taxa phylogenetically by
testing between alternate hypotheses of the relative
recency of common ancestry. These hypotheses must
take the form of a three taxon statement of the sort:
“taxa A and B share a common ancestor not shared
by taxon C.” The presence of complex and identical
structures shared by “C” and “B,” but absent in
“A” and in all more remotely related groups would
render this hypothesis unparsimonious under the
criteria outlined below. This process of “outgroup”
comparisons relies upon higher-level phylogenies
previously accepted on the basis of other characters.
It is by these comparisons that the “level of origin”

of homologies is determined. This procedure is truly
reciprocally illuminative (Hull, 1967), although it
may appear outwardly circular. Shared primitive
characters (symplesiomorphies) and characters
unique to only one taxon in the analysis (autapo-
morphies) have no relevance to three taxon tests
(see further discussion by Wiley, 1976). A theory of
relationships is considered corroborated only so long
as it remains the most workable (parsimonious) in-
terpretation of the biology of the organisms consid-
ered. When it does not meet these standards, it is
said to be falsified.

Because primitive character states must tempo-
rally precede derived states, establishing the level
of origin of homologies provides a relative temporal
sequencing of the appearance of monophyletic groups
(Fig. 1, A, II), as well as groupings of taxa that share
more immediate common ancestors than other taxa
in the analysis (Fig. 1, A, I). The statement of re-
lationships is always a comparative hypothesis, and
it is not possible to say that one taxon is the sister
group of another.

TWO TAXON HYPOTHESES

The cladistic analysis of Hennig cannot provide
a complete statement of phylogeny because it cannot
test some types of phylogenetic relationships. Its
application always results in a statement of com-
parative relationship that must be expressed in terms
of three groups. Such an analysis cannot distinguish
among three geometric relationships (Fig. 1, B: I,
II, III) for any two taxa (B and C of Fig. 1, B). By
convention, some workers have accepted type II in
which common ancestors are hypothetical and evo-
lution is assumed to be branching (see, for example,
Tattersall and Eldredge, 1977).

I distinguish between simple phylogenies (I) and
complex phylogenies (“cladograms,” II, and III). A
simple phylogeny has, as its basis a one-to-one re-
lationship between ancestor and daughter taxa and
thereby implies direct ancestry. A complex phylog-
eny is a combination of two or more simple phy-
logenies which share a common lineage until a point
of cladogenesis (branching) occurs (Fig. 1, C). The
point of cladogenesis is unique to a complex phy-
logeny and marks that point where parts of an evolv-
ing population become members of separate simple
phylogenies. The time transected at a point of clado-
genesis (insofar as it is a unique point) is only a
single generation in length and is pragmatically un-

recognizable. Taxa are, in fact, defined on the basis
of recognized evolutionary change along a segment
of a simple phylogeny even when cladogenesis is
known to have occurred. Evolution presents us with
a continuum of simple phylogenies resolving them-
selves into complex phylogenies. Systematic anal-
ysis is applied to segments of this continuum. Be-
cause taxonomy involves the partitioning of a
continuum, we must eventually separate parents
from their offspring as we extend our analysis from
the present into the past.

A hypothesis of an ancestor-descendant relation-
ship between two taxa follows from a tested hy-
pothesis of the relative recency of common ancestry
(a three taxon hypothesis) based on shared derived
character suites. A two taxon model may therefore
be falsified by refutation of the common ancestor
hypothesis upon which it is based. If, for example,
it is demonstrated that “B” of Fig. 1, A shared a
unique ancestor with “A,” instead of with “C,” none
of the models of Fig. 1 , B are correct.

The type II phylogeny of Fig. 1, B implies not
only: 1) a cladogenetic event, and 2) a hypothetical
taxon, but also that 3) this taxon has not been sam-
pled and probably exists in some geographic region
outside of the study area. Thus the local sequence

528

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 1.— A) phylogenetic nesting arrangement (I) for three taxa (A, B, C) is provided by (II) temporal sequence of appearance of
monophyletic groups; B) Alternate phylograms for taxa B and C of (A.), the black wedge implies the presence of a physical barrier and
the open circle a hypothetical common ancestor. C) The cladogram as a combination of simple phylogenies which share a common
lineage until branching occurs; D) A local sequence of taxa, (A, B, C) as explained by a branching sequence with two hypothetical
common ancestors and two hypothetical barriers.

(Fig. 1, D) must be accounted for by 4) local ex-
tinction of B followed or caused by 5) immigration
of descendants of the hypothetical taxon into the
region of study. None of these assumptions are tested
at the three taxon level. If taxa A, B, and C are
related by derived characters as in Fig. 1, A and
stratigraphically as in Fig. 1, D, a hypothesis of di-
rect ancestry without a cladistic (branching) event
is the most parsimonious interpetation of their ge-
ometry and should be maintained until branching
is demonstrated.

The most basic assumption of a complex phylo-
gram is the existence of a cladogenetic (vicariance)
event. This is some aspect of the geographic or eco-
logical environment that effectively divides up the
population. This aspect of the environment I call a
vicariator and symbolize by a wedge (Fig. 1, B-D).
A cladogenetic event implies the existence of a vi-

cariator, even if its nature is not known. If two taxa
(B and C, for example) are found to exist at the same
time, and form a holophyletic assemblage, they must
represent sister clades which shared a common
ancestor until a cladogenetic event occurred. Their
geometric relationships must be as in either model
II or III of Fig. 1 , B. A hypothesis of phyletic descent
whereby taxon B evolves into taxon C without a
cladistic event (Model 1) has thereby been refuted.

A cladogenetic event may also be demonstrated
by the distribution of derived character states. Ex-
cept for reversals in daughter taxa (which are as-
sumed to be rare), an ancestral population must be
primitive in all character states where it differs from
a daughter taxon. Again, this is a two taxon state-
ment, easily falsifiable in terms of relative parsi-
mony because it is applicable to all morphological
states by which the taxa differ. If taxon B of Fig. 1,

1984

MARTIN -PHYLETIC TRENDS

529

B is found to have derived characters unique to
itself, type I and type II phylogenies are rejected.
Both of these geometric arrangements are also re-
jected if taxon C can be shown to predate the evo-
lutionary appearance of taxon B.

Falsification of a type I (simple) phylogeny cor-
roborates a complex phylogeny and its assumptions.
If a species is defined as a single evolutionary lin-
eage, as was done by Wiley (1978), then the test of
a type I phylogram is also a test of the hypothesis
that a vertically stacked sequence of populations
represents such a lineage. If a type I phylogeny has
been refuted, type II and type III phylograms may
be similarly tested until some geometric relationship
is considered sufficiently corroborated.

Direct species-to-species lineages require close
stratigraphic control, but they are clearly possible.
Any time that we accept the continuance of one
species through several geological horizons we have
accepted all of the assumptions and probabilities
that are required for species-to-species lineages. Such
examples are actually common and have been used
to support the idea of evolutionary “stasis” in punc-
tuated equilibrium models. The American badger,
Taxidea taxus, seems to be a good example, as the
living species appears to extend back into the late
Pliocene some 3.5 million years ago (Hibbard, 1 970).
Of course the complete anatomy of the Pliocene
badger isn’t known, and there could have been con-
siderable anatomical change without our detecting
it. In fact, we are always dealing with a subset of
the total suite of characters, and the evolution we
observe (or lack of it) must largely be a result of the
particular set of characters we choose to observe.
This is an important consideration when we talk
about stasis versus change. We must realize that the
great majority of characters in all organisms have
been in stasis for tens of millions of years. For in-
stance, red blood corpuscles in mammals are nearly
uniform throughout the order and their basic fea-
tures must have been established some time in the
Mesozoic.

EXAMPLES

I have previously applied Kurten’s terminology
for saber-toothed cats (dirk-toothed and scimitar-
toothed) to all cats with the proper morphology
(Martin, 1980), rather than to just the Smilodontini
and Homotherini as he first applied them (Kurten,
1968). I also added a third category, the conical-

Size reduction in lineages of large mammals that
survived the late Pleistocene extinction provide
another set of examples of species lineages. There
are many such examples, and in some the ice-age
form has been placed in a separate species, whereas
in others it is kept in the same species as its living
descendants. The best documented lineage is that
of the plains bison. Bison bison bison, whose descent
from B. antiquus has been described by a number
of workers (Schultz and Frankforter, 1946; Schultz
and Martin, 1970a; Wilson, 1974). In this lineage
the size reduction is gradual with very little change
in proportions.

One of the important questions in any phyloge-
netic analysis is the direction of evolutionary change
(polarity) in the characters used. Cladists tend to
favor outgroup comparisons (comparison with char-
acter states in related groups) to determine polarity,
but two other methods can also be applied to pro-
posed lineages. The first of these is stratigraphic
position. The youngest member of a lineage should
be derived in all ways that it differs from the oldest
member. Obviously stratigraphy in itself cannot be
used to establish a lineage, as it only gives a hy-
pothesis of polarity. This hypothesis can then be
applied to all the available samples from strati-
graphic positions lying between the two age ex-
tremes. In every case each successive stratigraphic
stage should become more like the youngest sample
in features that are changing. There is no reason to
expect this result if a lineage is not present and if
many successive stages are available; alternative ex-
planations soon become absurdly complicated.

Functional analysis can also be used to formulate
a hypothesis of polarity. If the function of a structure
is well understood, then it should be possible to
identify changes that improve the function; we would
expect such changes to be more derived. It should
also be possible to evaluate the importance of the
structure to the organism and how likely a reversal
in polarity might be.

OF LINEAGES

toothed cats, which includes all the living felids. The
dirk-toothed cats have the longest upper canines and
seem in most respects to be the most highly mod-
ified.

The dirk-toothed adaptive type has developed in-
dependently at least four times (Fig. 2). The pattern

530

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 2. — Suggested phylogeny of the cats (from Martin, 1980). The
toothed cats.

seems to be one of extinction followed by redevel-
opment of the adaptive type. In North America these
redevelopments seem to be spaced at fairly regular
intervals. The two best known lineages involve the
Barbourofelini, a group of nimravid cats, and the
Smilodontini, a group of felid cats.

The dirk-toothed cats of the tribe Barbourofelini
are known from a number of sites whose relative
stratigraphic positions are known. This makes it an
excellent group to test some of the hypothesis of
lineage relationships. The Barbourofelini may be
united by the following derived features which sep-
arate them from all other cats: 1 ) upper canine flat-
tened with prominent internal and external grooves;
2) enormous enlargement of the origin of the su-
perficial masseter muscle; 3) serration of the cutting
edges of all of the deciduous and permanent teeth,
and 4) displacement of the lower cheek laterally on
a distinct buttress of the lower jaw.

Smilodontini, Barbourofelini, Eusmilini, and Hoplphonini are dirk-

The oldest members of the tribe are found in Eu-
rope and the Barbourofelini first appeared in North
America about eleven million years ago, with Bar-
bourofelis whitfordi in the Clarendonian. The last
form in the lineage is B. fricki which became extinct
about seven million years ago. When we compare
the earliest European form, where both the skull and
lower jaws are known ( Sansanosmilus palmidens),
we find that there has been— enlargement of the
canines and camassials; reduction in the size of the
P33; change from an inclined to a vertical occipit;
lowering of the glenoid fossae on the skull until they
extend to the lower edge of the upper carnassials;
development of a distinct area of overlap between
P4 and Mh and reduction in the height of the cor-
onoid process on the ramus. The direction of the
polarity of all these characters is also confirmed by
outgroup comparison with living felid cats. There
is also a remarkable increase in size (Fig. 3).

1984

MARTIN -PHYLETIC TRENDS

531

Fig. 3. — Stratigraphic sequence from oldest (A) to youngest for skull and jaws of barbourofelin cats. Skulls with lower jaws: A)
Sansanosmilus palmidens (postorbital bar not developed yet); B) Barbourofelis whitfordi (earliest North American form); C) B. morrisi,

D) B. lover, E) B. fricki. Right rami: A) Syrtosmilus syrtensis\ B) Sansanosmilus palmidens\ C) Barbourofelis whitfordi, D) B. morrisi,

E) B. lover, F) Barbourofelis sp. from the Jack Swayze Quarry, Clark Co., Kansas; G) B. fricki.

The polarity of characters can also be tested func-
tionally. Many of the specializations found in the
skulls and lower jaws of saber-toothed cats are re-
lated to the elongation of the upper canines. The
functional bite of the animal is the arc that lies be-

tween the tips of the upper and lower canines. The
lower jaw has to rotate through a considerable dis-
tance before there is any separation of the canines.
In the most specialized saber-toothed cat, Barbour-
ofelis fricki, the total rotation must be slightly more

532

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

than 1 1 5°. Features that facilitate this rotation in-
clude: 1) the vertical occipital region; 2) the lowering
of the glenoid fossa on the skull; 3) the lowering of
the coronoid process on the ramus, and 4) the en-
largement of the paramastoid processes. Features
that facilitate the camassial bite include: 1) the en-
largement of the upper camassial; 2) the reduction
of P33 and 3) the overlapping of P4 on M,. We can
thus see that the functional interpretation of these
characters is in agreement with their stratigraphic
distribution and with outgroup comparisons with
other cats. When we look at the proposed lineage
for Barbourofelis (Fig. 3), we can see that the changes
in these features form a progressive series. We can
also see that not all features change at the same rate,
and that changes in a mechanical system stop after
the system is perfected. For instance, a postorbital
bar is achieved sometime between Sansanosmilus
and Barbourofelis, and that region remains fairly
constant after the development of the postorbital
bar. The relationship of the P4 to the M, is also
fairly constant after an initial overlap is achieved.

We can make an additional check on our under-
standing of the evolutionary process in some other
group of similar saber-toothed cats.

The Barbourofelini were very widespread during
the Miocene, and are now known from France, Ger-
many, Spain, Turkey, Mongolia, and a number of
sites in the United States. They also appear to have
occurred in the Miocene of Africa (Kurten, 1976;
Ginsburg, 1978). The African material is not very
good, and I accept it as belonging to the Barbour-
ofelini on the basis of comparisons made by its de-
scribes. One of the African forms is Burdigalian in
age and would probably be one of the oldest and
most primitive barbourofelin known (Fig. 3), if it is
correctly assigned. It would certainly be the smallest
known species.

In order for the Barbourofelini to enter North
America they must have had a northern Asiatic pop-
ulation around eleven million years ago. This pop-
ulation was probably continuous with the European
population. The American populations seem to have
been well separated from the European one soon
after the dispersal into North America. This is a
classical cladogenic event through dispersal with
separate lineages of Barbourofelini evolving con-
temporaneously in North America and Eurasia. The
European lineage shows very little increase in size
(Fig. 4), but it does show the progressive reduction
of P3 and the overlap of P4 and M, found in North
American Barbourofelis lineage. It is clear that the

Fig. 4. — Stratigraphic sequence from oldest (A) to youngest (D)
of barbourofelin cats in Eurasia. A) Prosansanosmilus peregrinus,
Langenau 1 (Heizmann et al., 1980); B) Sansanosmilus palmi-
dens, Sansan (Ginsburg, 1961); C) Sansanosmilus jourdani val-
lesiensis (Beaumont and Crusafont-Pairo, 1982); D) “ Sansano-
smilus” piveteaui (Ozansoy, 1965). Size does not increase very
much but the P3 becomes relatively smaller and loses a root, the
camassial becomes larger and it and P4 tilt posteriorly.

overall rate of known evolutionary change was less
in the European branch, and ultimately this resulted
in less modification at the time that the European
population became extinct.

We can make an additional check on our under-
standing of the evolutionary processes in the Bar-

1984

MARTIN- PHYLETIC TRENDS

533

bourofelini by looking at the same sort of series in
some other group of saber-toothed cats. If the func-
tional progression has been correctly interpreted,
and if the changes are the result of natural selection
to improve those functions, we might expect to see
the same pattern of changes in any group of animals
that become functional dirk-toothed cats. This seems
to actually be the case, and we can draw a parallel
example from the dirk-toothed felids of the late Plio-
cene and Pleistocene, the Smilodontini. The genus
Megantereon shares with Smilodon a short tail and
shortened distal limb segments. The Smilodontini
are felid cats and have the divided auditory bulla
characteristic of that family. The Barbourofelini are
nimravids (Fig. 2) and some workers consider their
relationship to the true cats (felids) to be remote.
The oldest known Megantereon seems to be the
Pliocene, M. hesperus (Schultz and Martin, 19706).
The American lineage terminated in Smilodon flor-
idanus about 10,000 years ago. The total duration
of the lineage in North America is something on the
order of 5,000,000 years. This lineage in North
America is well represented by lower jaws (Fig. 5)
but not by skulls of Megantereon. If we compare
the skull of the European M. megantereon to Smi-
lodon floridanus, we find that they differ in the youn-
ger form having enlarged the carnassial, reduced P3\
lowered the glenoid fossa, enlarged the paramastoid
process, lowered the coronoid process on the ramus,
overlapped M, with P4 and increased in size dra-
matically. These changes are the same as those ob-
served in Barbour of elis, and when we observe the
stacked series of lower jaws (Fig. 5), we see that they
develop in the same ordered, gradual progression
seen in Barbourofelis (Fig. 3). There is one notable
exception. The Barbourofelini have a progressive
increase in the size of the dependent flange on the
ramus. The Smilodontini have a progressive de-
crease in the relative size of the dependent flange,
and in the Smilodontin lineage the occipit region
actually becomes more inclined as the flange de-
creases in size.

The arvicolid rodents (voles and lemmings) pro-
vide a number of detailed species lineages (Martin,
1979). The arvicolids are grass-eating rodents, and
show progressive hypsodonty in the evolution of
their molars until most forms have ever-growing
teeth. As the crowns of their molars become higher,
the roots become lower and appear later ontoge-
netically. The function of the roots to anchor the
tooth in the jaw is at the same time taken over by
grooves in the enamel of the tooth crown that expose

the dentine covered by root cementum. The peri-
dontal ligaments attach to the root cementum within
these grooves, which are called dentine tracts. The
dentine tracts take over the functions of the dimin-
ishing roots and the measurement of the height of
the dentine tracts above the base of the tooth crown
is a good measure of the hypsodonty of the molar.
In evergrowing teeth the dentine tracts form con-
tinuous bands from the top to the bottom of the
tooth crown.

Besides increasing the crown height of the molars,
arvicolids increase the lifespan of their molars by
increasing the resistance of their molars to abrasion.
They do this by changing the arrangement of the
apatite prisms in their molar enamel (Koenigswald,
1 980), and the functional significance and directions
of these changes have been worked out by Koenigs-
wald.

The functional significance of the different ar-
rangements of apatite prisms in the enamel of ar-
vicolid molars seems to be related to the direction
of chewing forces (Koenigswald, 1980). The prim-
itive arrangement for arvicolids (the only type found
in most of the early members of the arvicolid ra-
diation and in some cricetids) is radial enamel. This
arrangement when stressed by the directional forces
of chewing permits fractures to transect the entire
thickness of the enamel. Arvicolids with only radial
enamel usually compensate for this by having very
thick enamel. Lamellar or tangential enamel inhibit
fractures from passing through the enamel and
thereby preserve the enamel edges, extending the
lifespan of the tooth.

The Mimomys lineage published by Koenigswald
(1980) shows the general pattern of evolutionary
change in enamel histology for an arvicolid rodent,
and these changes are gradual and regular. They
parallel very closely the other changes which are
related to the rate of tooth wear and subsequent
increases in hypsodonty. Such patterns argue very
strongly for the effect of natural selection in these
evolutionary patterns and provide little support for
either stochastic or macroevolutionary models.

In North America, the muskrats (Ondatrini) have
been the most extensively studied arvicolid rodents
(Nelson and Semken, 1970; Martin, 1979; Za-
krzewski, 1974), and the following species lineage
is well accepted: Pliopotamys minor , P. meadensis,
Ondatra idahoensis, O. annectens, O. nebracensis,
O. zibet hicus. Within this lineage we find the fol-
lowing evolutionary trends: 1) increase in size; 2)
increase in hypsodonty and in dentine tract height;

534

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

Fig. 5. — Stratigraphic sequence from oldest (A) to youngest (D) of lower jaws of Smilodontin cats in North America: A) Megantereon
hesperus ; B) M. gracilis\ C) Smilodon fatalis\ D) Smilodon floridanus.

1984

MARTIN- PH YLETIC TRENDS

535

O)

cr

LU

X

I-

q

$

-

-

□

ARVICOLINE

ZONE

n

-

+

arvicoline

ZONE

m

:

X

ARVICOLINE

ZONE

••• •

-

O

ARVICOLINE

ZONE

3zr

•

• • • •

-

•

arvicoline

ZONE

3zn

• •

••o°*** « •••

-

►3

• .°oV*s*

••

• •• • X •

-

•o • l • • •

• ••••• •

-

\ oO 0*0** •

• • • • •

_

O o • • °0 ••

• • •

-

O

X O oo oo° • *

•

-

° x oO o O

-

x o °

-

B ° +D+

+ □ □

X

X X

:

□

□

X

X

► 2

□

□ °

^

-

□

□ □

□

□

-

□ □

LENGTH IN MILLIMETERS

D o D

4

□

5

6 7

8

i i —

■ i i ■

i

i i i i i i i

I .i ,-L

1 1 1

■ i A i i i i i i i i i A i i i

j i i i i i A i

Fig. 6. — Measurements of lower first molars of muskrats showing increase in size from older (Arvicoline Zone II) to younger (Zone
VII) arvicoline zones in North America (from Martin, 1979). Note that the size increase is overlapping except for that area where no
sample was available (Zone IV).

3) increase in crown complexity on Ml5 and 4) ad-
dition of cement to the re-entrant angles of the molar
crowns. In Figs. 6 and 7 we can see that the increases
of size and crown complexity are gradual and over-
lapping.

The arvicolid molars consist of a series of gen-
erally alternating triangles. In the M, of very prim-
itive forms like Prosomys, there is an anterior loop
followed by three alternating triangles. These cor-
respond to the anteroconid, metaconid, protoconid,
and hypoconid, plus the posterior cingulum in the
cricetine M,. Most of the evolutionary modification
of the grinding surface of M, achieved by increasing
the length and complexity of the anteroconid, and
the elaboration of the crown pattern results from
selection acting on crenulations of the anteroconid
(anterior loop). In early arvicolids these crenulations
occupy the position of alternating triangles in later
forms. The new triangles are formed by enlargement
of these crenulations coupled with enlargement of
their re-entrants. The oldest known muskrat, Pliop-
tamys minor, has a highly crenulated anterior loop
followed by five distinct alternating triangles and a

posterior loop. The fourth and fifth triangles become
better separated from the anterior loop in P. mead-
ensis and Ondatra idahoensis. In Ondatra annec-
tens, O. nebracensis, and O. zibethicus, the posterior
portion of the anterior loop opens broadly into sixth
and seventh triangles. Cement first appears as tiny
deposits in the bases of the re-entrant angles of the
molars of O. idahoensis, and eventually fills re-en-
trants in later forms. The muskrat chronocline pub-
lished by Nelson and Semken (1970) includes four
samples that are reasonably well dated: Hagerman,
3.5 m.y.; Borchers, 1.9 m.y.; Cudahy, 0.6 m.y.; and
the Wisconsin samples, all of which should date less
than 100,000 years before present. Borchers is
lumped with Grandview which seems to be a slightly
older fauna, and the curve could probably be im-
proved in that area. Dentine tracts are extremely
low or absent in the Hagerman sample, and I start
the curve at zero at that point. There is a fairly slow
steady increase in both size of M, and dentine tract
height from the Hagerman to the Cudahy faunas, a
time of nearly two million years. After the Cudahy
faunas the last 600,000 years has been a time of very

536

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

a

Fig. 7. — Graph a showing change in dentine tract height on M, (lower curve) and length of M, (upper curve) for muskrats (measurements
adapted from Nelson and Semken, 1970); (A) Hagerman Local Fauna, (B) Borchers and Grandview local faunas, (C) Cudahy local
faunas, (D) Wisconsinan faunas, and (E) modern muskrats. Part b\ Left M,’s of muskrats arranged in stratigraphic sequence with oldest
at the bottom and North American arvicoline zone listed on the left. From bottom to top: Pliopotcunys minor, P. meadensis. Ondatra
idahoensis, O. cf. annectens, O. annectens, O. nebracensis, O. zibet hecus (from Martin, 1979).

1984

MARTIN -PHYLETIC TRENDS

537

M. Y. B. P.

Fig. 8.— Graph showing increase in lower jaw diastema length in two lineages of saber-toothed cats. A) Sansanosmilus palmidens ; B)
Barbourofelis whitfordr, C) B. lover, D) B. frickv, E-F) Megantereon hesperus ; G) M. gracilis-, H) Smilodon floridanus. Ages of A-D
based on age estimations in Savage and Russell (1983).

rapid increase in both tooth size and dentine tract
height.

If we calibrate the curve for increase in size of the
diastema in the Smilodontini (Fig. 8) against the
muskrat curve, we see a curve of very similar shape
with little change in size until the last 600,000 years.
Muskrats and saber-toothed cats have little in com-
mon either phylogenetically or ecologically, and any
set of factors that would tend to synchronize the
rates of evolutionary change in these two groups
would have to be profound and must have affected
many other animals at the same time. I suspect that
important climatic shifts were beginning in North
America some 600,000 years ago.

The rate of evolutionary change in dentine tract
height in muskrats seems regular enough, and the
curve well enough calibrated by radiometric dates
that it would be possible to estimate the absolute
age of faunas containing muskrats using the dentine

tract curve alone. In fact, it might be possible to
develop a system of biometric dating using a variety
of organisms and structures to obtain absolute age
estimates for geologic horizons comparable in qual-
ity with those from radiometric dating. Such a sys-
tem would have the advantage of multiple testing
through the use of different organisms. The dates
obtained would be on the actual age of the fauna,
and dates could be obtained from a wide variety of
sedimentary rocks whose absolute ages cannot now
be easily estimated.

The accuracy of such a system would be largely
dependent on the availability of reasonable strati-
graphic samples of the organism involved, the pres-
ence of structures that are changing rapidly along a
noncontroversial polarity and the direct association
of a sequence of faunas containing the organisms
with radiometric dates.

CONCLUSIONS

The recognition of phylogenetic lineages is an area
of paleontological research that will increase in im-
portance as our knowledge of the biostratigraphic
record improves. Stratigraphic data are a necessary
prerequisite for studies of this type (Gingerich, 1979).
But if such data are available, we can propose and

test hypotheses of both polarity and ancestry. We
can also study rates of change and speculate on the
types of evolutionary models that best explain the
known fossil record.

Saber-toothed cats and arvicolid rodents with
known stratigraphic records show patterns of change

538

SPECIAL PUBLICATION CARNEGIE MUSEUM OF NATURAL HISTORY

NO. 8

that are gradual and overlapping. The rate of change
is not constant in either group but varies with time.
Both the Smilodontini and the Ondatrini show a
marked increase in the rate of evolutionary change
some 600,000 years ago. This suggests some overall
controlling factor that affects the strength of selec-
tion on structure in both groups. This overall factor
may be a change in climate. The Barbourofelini show
a similar pattern of change at an earlier time (Fig.
8).

The Barbourofelini provide interesting examples
of both cladogenesis and anagenesis with differing
rates of evolutionary change. The general similarity
in the evolutionary patterns of the Barbourofelini
and the Smilodontini suggest that there are not very
many avenues open to success in an adaptive zone
and animals with similar ways of life are strongly
routed by natural selection into similar morpholo-
gies.

ACKNOWLEDGMENTS

I have had very stimulating conversations with D. Rasmussen, with me. The figures are by M. Tanner, M. A. Klotz, and J.

C. C. Black, C. B. Schultz, J. Chaline, W. von Koenigswald, and Martin. J. D. Stewart and R. W. Wilson critically read the text

especially R. W. Wilson on various aspects of this study. K. and I am grateful for their many helpful suggestions.

Whetstone has shared many useful insights into cladistic analysis

LITERATURE CITED

Beaumont, G. de, and M. Crusafont-Pairo. 1982. Les Fe-
lides (Mammiferes, Carnivores) du Vallesien du Valles, Ca-
talegne, Espagne. Arch. Sc. Geneve, 35:41-64.

Gingerich, P. D. 1979. The stratophenetic approach to phy-
logeny reconstruction in vertebrate paleontology. Pp. 41-77,
in Phylogenetic analysis and paleontology (J. Cracraft and
N. Eldredge, eds.), Columbia Univ. Press, New York, 233 pp.

Ginsburg, L. 1961. La faune des Carnivores Miocenes de San-
san (Gers). Mem. Mus. Nat. Hist. Nat., ser. c, 9:1-190.

. 1978. Syrtosmilus syrtensis, n. gen., n. sp., Felide ma-

chairodontiforme du Bardigzlien de Lybie. C. R. Somm. Soc.
Geol. Fr„ 1978:273-74.

Heizmann, E. P. J., L. Ginsburg, and Ch. Bulot. 1980. Pro-
sansanosmilus peregrinus ein neuer machairodontider felids
aus dem Miocan Deutschlands and Frankreichs. Stuttgarter
Beitr. Naturk., 58(27): 1-27.

Hibbard, C. W. 1970. Pleistocene mammalian local faunas
from the Great Plains and Central Lowland Provinces of the
United States. Pp. 395-433, in Pleistocene and Recent en-
vironments of the central Great Plains. (W. Dort, Jr., and J.
K. Jones, Jr., eds.), Univ. Press Kansas, Lawrence, 433 pp.

Hull, D. L. 1967. Certainty and circularity in evolutionary
taxonomy. Evolution, 21:174-189.

Koenigswald, W. v. 1980. Schmelzstruktur und Morphologie
in den Molaren der Arvicolidae (Rodentia). Abh. Senckenb.
Naturforsh. Ges., 539:1-129.

Kurten, B. J. 1968. Pleistocene mammals of Europe. Aldine
Publ. Co., Chicago, 317 pp.

. 1976. Fossil Carnivora from the late Tertiary of Bled

Douarah and Cherichira, Tunisia. Notes Service Geol. Tuni-
sie, 42:177-214.

Martin, L. D. 1979. The biostratigraphy of arvicoline rodents
in North America. Trans. Nebraska Acad. Sci., 7:91-100.

. 1980. Functional morphology and the evolution of cats.

Trans. Nebraska Acad. Sci., 8:141-154.

Nelson, G., and H. Semken. 1970. Paleoecological and strati-
graphic significance of the muskrat in Pleistocene deposits.
Bull. Geol. Soc. Amer., 81:3733-3738.

Ozansoy, F. 1965. Etude des giesments continentaux et des
Mammiferes du Cenozoique de Turquie. Mem. Soc. Geol.
France, n.s. 44(1), Mem. ( 1 02): 1 —92.

Patterson, C. 1981. Significance of fossils in determining evo-
lutionary relationships. Ann. Rev. Ecol. Syst., 12:195-223.

Savage, D. E., and D. E. Russell. 1983. Mammalian paleo-
faunas of the World. Addison-Wesley Publ. Co., Reading,
Massachusetts, 432 pp.

Schultz, C. B., and W. D. Frankeorter. 1946. The geologic
history of the bison in the Great Plains, a preliminary report.
Bull. Univ. Nebraska State Mus., 3:1-10.

, and L. D. Martin. 1970a. Quaternary mammalian

sequence in the central Great Plains. Pp. 341-353, in Pleis-
tocene and Recent environments of the central Great Plains
(W. Dort, Jr., and J. K. Jones, Jr., eds.), Univ. Press Kansas,
Lawrence, 433 pp.

. 1970A Machairodont cats from the early Pleistocene

Broadwater and Lisco local faunas. Bull. Univ. Nebraska
State Mus., 9:33-38.

Tattersall, I., and N. Eldredge. 1977. Fact, theory, and
fantasy in human paleontology. Amer. Scientist, 65:204-
211.

Wiley, E. O. 1976. The phylogeny and biogeography of fossil
and Recent gars (Actionopterygii:Lepisosteidae). Univ. Kan-
sas Mus. Nat. Hist. Misc. Publ., 64: 1-111.

. 1978. The evolutionary species concept reconsidered.

Syst. Zool., 28:17-26.

Wilson, M. 1 974. The Casper Local Fauna and its fossil bison.
Pp. 125-171, in The Casper Site (G. C. Frison, ed.). Aca-
demic Press, New York, 266 pp.

Zakrzewskj, R. J. 1974. Fossil Ondatrini from western North
America. J. Mamm., 55:284-292.

Address: Museum of Natural History and Department of Systematics and Ecology, University of Kansas,

Lawrence, Kansas 66045.