74

HARVARD UNIVERSITY

Library of the

Museum of

Comparative Zoology

NORTH AMERICAN QUATERNARY CANIS

Frontispiece

Three species of Canis. Top, Canis lupus (the gray wolf, photo by L.
David Mech). Middle, Canis rufus (the red wolf, photo by Curtis
Carley). Bottom, Canis latrans (the coyote, photo by Tom Smylie).
All photographs courtesy of the Fish and Wildlife Service, Depart-
ment of the Interior.

NORTH AMERICAN QUATERNARY CAMS

RONALD M. NOWAK

Staff Specialist

Office of Endangered Species

U.S. Fish and Wildlife Service

Department of the Interior

Washington, D.C. 20240

MONOGRAPH

OF THE
MUSEUM OF NATURAL HISTORY, UNIVERSITY OF KANSAS

NUMRER 6

1979

NORTH AMERICAN QUATERNARY CAN1S

MONOGRAPH OF THE MUSEUM OF NATURAL HISTORY

THE UNIVERSITY OF KANSAS

Number 6, pages 1-154, text figures 1-55
September 1, 1979

MUS. COMP. ZOOL"
LIBRARY

JUN 1 8 1980

HARVARD

UNIVERSITY

Editor: E. O. Wiley

Copyrighted

By

Museum of Natural History

The University of Kansas

Lawrence, Kansas 66045

U.S.A.

ISBN: 0-89338-007-5

Printed

By

The University of Kansas Printing Service

Lawrence, Kansas

U.S.A.

CONTENTS

INTRODUCTION _ 1

Acknowledgements 2

Methods .— 4

Age and Secondary Sexual Variation 6

HISTORY AND STATISTICAL ANALYSIS OF RECENT POPULATIONS 7

Comparison of Known Series of Dogs, Wolves, and Coyotes 7

Systematic Problems in the Northeast 12

Systematic Problems in the Southeast _.. 24

SYSTEMATIC DESCRIPTIONS 66

Genus Canis Linnaeus 66

Canis cedazoensis Mooser and Dalquest 68

Canis Jcpophagus Johnston 68

Canis Jatrans Say 73

Canis edwardii Gazin 82

Canis rufns Audubon and Bachman 85

Canis armbrusteri Gidley 90

Canis lupus Linnaeus 93

Canis familiaris Linnaeus 102

Canis dims Leidy - 106

SUMMARY 118

LITERATURE CITED 121

ADDENDUM 136

APPENDIX A 138

APPENDIX B 144

APPENDIX C 150

INTRODUCTION

According to the revisionary work of An-
derson (1943), Goldman (1937, 1944), and
Jackson ( 1951 ) , three living species of wild
Cards occur in North America: Cards la-
trans, the coyote; C. rufits, the red wolf; and
C. lupus, the gray wolf. Although this taxo-
nomic arrangement has been generally ac-
cepted, some questions have arisen concern-
ing matters not fully explained by the
revisions, and certain newly recognized phe-
nomena within canid populations.

Much attention, especially since 1960, has
been directed toward the systematics of Cards
in the eastern half of North America. There
has been controversy regarding the taxonomic
status of the wolves (C. lupus hjcaon and
subspecies of C. rufus) originally found there,
and of the populations of Canis presently
inhabiting the region (roughly, east of 100°W
and south of 50°N). The production of fully
fertile hybrids, of common occurrence among
captive Canis (Gray, 1972), has been sug-
gested as having affected wild Canis in the
eastern part of the continent.

In addition to the questions concerning
living Canis, there are problems involving the
paleontological history of the genus in North
America. Although many fossil specimens
have been described, not all of them have
been assigned to particular lineages ancestral
to living populations. Of special interest, be-
cause it is the only fossil kind represented by
what a modern mammalogist would call a
good series, is the dire wolf, Canis dims.

My aim in studying Canis was to examine
large series of specimens from throughout
North America, in order to obtain a clearer
understanding of the systematic relationships
between the species represented. I hoped to
get an idea of the extent of variation within
the Recent wolves found in the western and
northern parts of the continent, and to deter-
mine the relative positions of C. dints, C.
rufus, and C. lupus hjcaon. I wanted also to
ascertain, as well as possible, the origins and

relationships of the presently existing popu-
lations of Canis in the east. Partly from
study of populations, I hoped to distinguish
and more accurately delineate (morphologi-
cally, geologically, geographically) the living
and extinct species of North American Canis.

Because of their recognized taxonomic
value, abundance in museum collections, pa-
leontological preservation, and relative ease
of handling, I used skulls as the primary ma-
terial of my study. Approximately 5,000 spec-
imens were examined.

The first main part of the paper consists
of an historical sketch and a statistical analy-
sis based on those populations represented by
large series of complete skulls. This analysis
serves to delineate special groups and to as-
sess the probable origin and relationship of
questionable populations. The BMD07M
program of multivariate analysis was a pri-
mary method employed in this study. The
second main part of the paper consists of
descriptions of each recognized species of
North American Canis. Some of the speci-
mens discussed, including many of the fossils,
could not be used in multivariate analysis, but
the descriptions are supported in part by uni-
variate and bivariate statistics.

Collections cited in this paper are repre-
sented by the following abbreviations:
AMNH, American Museum of Natural His-
tory; ANSP, Academy of Natural Sciences,
Philadelphia; CM, Carnegie Museum; CNM,
National Museum of Canada; FGS, Florida
Geological Survey; FM, Field Museum; ISM,
Illinois State Museum; KU, University of
Kansas Museum of Natural History; LACM,
Los Angeles County Museum of Natural His-
tory; LPI, Louisiana Polytechnic Institute
Department of Zoology; LSUMZ, Louisiana
State University Museum of Zoology; MCZ,
Harvard University Museum of Comparative
Zoology; MSU, Michigan State University
Museum; NYEC, New York Department of
Environmental Conservation; PPM, Panhan-

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

die Plains Museum; PUWL, Purdue Univer-
sity Wildlife Laboratory; QWS, Quebec Wild-
life Service; ROM, Royal Ontario Museum;
SD, San Diego Natural History Museum;
SMUMP, Southern Methodist University Mu-
seum of Paleontology; SR, Sul Ross State Uni-
versity Department of Biology; TM, Texas
Memorial Museum; UAlb, University of Al-
berta Department of Zoology; UAriz, Univer-
sity of Arizona Department of Biological
Sciences and Laboratory of Paleontology;
UArk, University of Arkansas Department of
Zoology; UCMP, University of California
Museum of Paleontology; UCMVZ, University
of California Museum of Vertebrate Zoology;
UColo, University of Colorado Museum; UF,
University of Florida State Museum; UI, Uni-
versity of Illinois Museum of Natural History;
UMMP, University of Michigan Museum of
Paleontology; UMMZ, University of Michi-
gan Museum of Zoology; UMinn, University
of Minnesota Museum of Natural History;
UN, University of Nebraska State Museum;
UO, University of Oklahoma Museum;
USFWS, United States Fish and Wildlife
Service field collections; USNM, United
States National Museum of Natural History;
VFG, Vermont Fish and Game Department.
A few other collections are spelled out in the
text. Other common abbreviations in this
paper include "C." for Canis and "A." for
Aenocyon.

This paper is a slightly modified version
of a Ph.D. dissertation (Nowak, 1973) sub-
mitted to the University of Kansas in 1973.
Since that year substantial new information
has become available, as for example through
Kurten's ( 1974 ) study of fossil coyotes, Kole-
nosky and Standfield's (1975) analysis of
wolves in Ontario, and Mooser and Dal-
quest's ( 1975 ) description of a new species
of North American Pleistocene Canis. In
addition, a number of specimens have been
collected recently in southeastern Texas, and
this material allows an updating of the status
of the red wolf in that area. Although I have
devoted some space to the newly available

views and data, they have not, in all cases,
received the same degree of attention shown
the earlier material. Other differences be-
tween my dissertation and this paper include
the dropping in the latter of several figures
and tables of measurements, the relegation
of the statistical analysis of the dire wolf to
the section entitled "Systematic Descriptions,"
and the correction of several errors. I have
not cited my dissertation as a reference for
this paper, except in a few instances in which
mention of the contrast between the two
seemed warranted.

Acknowledgements

It is impossible to express my full appre-
ciation to all who assisted me. I must, how-
ever, single out Professor E. Raymond Hall,
Museum of Natural History, University of
Kansas, who was my major advisor until his
retirement in May 1972, and who then vol-
untarily continued to act in this capacity.
Professor Hall initially suggested that I do
graduate work at the University of Kansas,
and it was his idea for me to make a study
of the relationships between species of North
American Quaternary Canis. On countless
occasions he provided me with assistance and
advice regarding my dissertation and my
general program of work at the University.
I consider it a rare honor to have been among
his students.

Practically all of the other instructors, and
many of the students, with whom I have been
associated at the University of Kansas, have
at one time or another given me some help
that eventually contributed to this paper. I
thank them all, but specifically want to men-
tion Professors Robert S. Hoffmann, Robert
M. Mengel. and Craig C. Black (now of the
Carnegie Museum). I also am grateful to
Professor Peter M. Neely, Associate Director
of the University Computer Center, who took
much of his time to explain the processes and
results of the BMD07M computer program.
Although I could not have effectivelv used

1979

NOWAK: NORTH AMERICAN QUATERNARY CAMS

BMD07M without Professor Neely's help, I
take full responsibility for the application
and interpretation of this program with re-
gard to the problems of my study. Additional
valuable assistance on the use and under-
standing of computers was provided by my
fellow student, Alberto Cadena.

The research required in preparation of
this paper necessitated considerable travel to
museums and other localities throughout
North America. I therefore am especially
grateful to those organizations that aided me
in this regard. The Theodore Roosevelt Me-
morial Fund of the American Museum of
Natural History, and the National Science
Foundation each made a direct grant for
travel and related expenses. The Committee
on Systematics and Evolutionary Biology,
University of Kansas, provided travel grants
in 1970 and 1971, a research assistantship in
the summer of 1972, and also a traineeship for
the academic year 1971-1972.

No progress could have been made in my
research had it not been for the cooperation
of numerous persons who generously assisted
me in the examination of specimens and asso-
ciated materials in their care. I want to
especially thank John L. Paradiso, Bird and
Mammal Laboratories, United States National
Museum of Natural History (now of the
Office of Endangered Species, U.S. Fish and
Wildlife Service). During the four months
that my wife and I worked at the National
Museum, he aided us in every way possible
and spent a great deal of his own time to see
that we were well provided for both in and
out of the Museum. John and I actually have
been in close communication regarding Canis
since 1965. Many of the views expressed in
this paper were developed jointly with him
in the course of years of pleasant study, con-
versation, and correspondence.

I am also grateful to the following per-
sons who either sent me specimens on loan
or assisted me when I visited their areas:
Sydney Anderson, American Museum of Nat-
ural History; Rollin H. Baker, The Museum,

Michigan State University; Troy L. Best, Mu-
seum of Zoology, University of Oklahoma;
Elmer C. Birney, Museum of Natural History,
University of Minnesota; Ben Day, Vermont
Fish and Game Department; Diana Van El-
sacker. University of Colorado Museum;
David E. Fortsch, Los Angeles County Mu-
seum of Natural History; Philip S. Gipson,
Department of Zoology, University of Arkan-
sas; John W. Goertz, Department of Zoology,
Louisiana Polytechnic Institute; John E. Guil-
day, Carnegie Museum; C. R. Harington,
National Museum of Canada; Billy R. Harri-
son, Panhandle Plains Museum; Claude W.
Hibbard, Museum of Paleontology, Univer-
sity of Michigan; Donald F. Hoffmeister,
Museum of Natural History, University of
Illinois; Emmet T. Hooper, Museum of Zool-
ogy, University of Michigan; J. H. Hutchison,
Museum of Paleontology, University of Cali-
fornia; Frederick F. Knowlton, U.S. Fish and
Wildlife Service; Barbara Lawrence, Museum
of Comparative Zoology, Harvard University;
Everett H. Lindsay, Laboratoiy of Paleontol-
ogy, University of Arizona; George H. Low-
ery, Jr., Museum of Zoology, Louisiana State
University; Ernest L. Lundelius, Jr., Texas
Memorial Museum; Larry D. Martin, Museum
of Natural History, University of Kansas;
John D. Newsom, Louisiana State Univer-
sity Cooperative Wildlife Research Unit;
Robert T. Orr, California Academy of Sci-
ences; N. Panter, Department of Zoology,
University of Alberta; Paul W. Parmalee,
Illinois State Museum; Oliver P. Pearson,
Museum of Vertebrate Zoology, University
of California; Randolph L. Peterson, Royal
Ontario Museum; Charles Pichette, Quebec
Wildlife Service; Douglas H. Pimlott, Univer-
sity of Toronto; Clayton E. Ray, U.S. Na-
tional Museum of Natural History; Richard
L. Reynolds, Los Angeles County Museum
of Natural History; Horace G. Richards,
Academy of Natural Sciences, Philadelphia;
Glynn Riley, Jr., U.S. Fish and Wildlife
Service; C. B. Bobbins, Department of Bio-
logical Sciences, University of Arizona; Den-

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

nis N. Russell, Texas Parks and Wildlife
Department; Donald Schierbaum, New York
Department of Environmental Conservation;
James F. Scudday, Department of Riology,
Sul Ross State University; Beryl E. Taylor
and Richard H. Tedford, American Museum
of Natural History; Gilmer Voss, San Diego
Natural History Museum; S. David Webb,
University of Florida State Museum; J. Wil-
liam Yon, Florida Geological Survey; Phillip
M. Youngman, National Museum of Canada;
and Curtis J. Carley, U.S. Fish and Wildlife
Service.

These acknowledgements would not be
complete without the names of my parents,
Jacob and Esther Nowak, New Orleans, Lou-
isiana. Throughout the course of my research
they were always ready and willing to pro-
vide any assistance, whether requested or not.

I finally wish to express my gratitude to
my wife, Thu. Although having only re-
cently arrived in the United States, and with
an incomplete command of the English lan-
guage, she served as an indispensable full
time assistant, especially in the recording of
data. Subsequently, she prepared the base
maps and parts of the figures herein.

Methods

As a primary statistical tool I employed
the Biomedical computer program, number
07M, stepwise discriminant analysis (Dixon,
1970). This method is a modified version of
multivariate discriminant function analysis,
as used previously in the study of Canis and
explained in detail by Jolicoeur (1959), Giles
( 1960 ) , Lawrence and Bossert ( 1967 ) , and
Gipson (1972). The BMD07M program in-
volves a procedure known as canonical analy-
sis, as discussed by Rao ( 1952 ) and Seal
(1964).

In multivariate analysis a series of varia-
bles from an individual specimen are consid-
ered together to determine the position of
that specimen relative to other specimens. In
its simplest form this procedure resembles

that of a scatter diagram in which the loca-
tion of a specimen on a two dimensional
graph is determined by its position along
both a vertical and horizontal axis, each
representing a single variable. The multi-
variate analysis, through a process of matrix
inversion, can consider numerous variables,
but plots the results in the same form of a
two dimensional graph.

The BMD07M program requires that at
least two designated groups of individuals be
entered into the analysis. The variables are
tested one at a time for their ability to dis-
tinguish between the groups. If any variable
is found to have too low a discriminatory
power, that variable is rejected and not con-
sidered in the analysis. The effects of cor-
relation among the variables are eliminated
in this program by a process of eigenvalue
extraction.

On the basis of the variables selected, the
designated groups are separated as well as is
possible. The statistical distance between
groups (D2 of Mahalonobis), calculated from
the combined variables, may be printed out
if desired. In addition, each individual speci-
men is given a D2 distance from each group,
and is assigned canonical coordinates to plot
its position relative to all other specimens.
If the variables employed have effectively
distinguished the groups, the specimens
within a particular group will be nearer to
each other than to the specimens of other
groups. Once definite groups have been es-
tablished, specimens of questionable identity
may be individually entered into the analysis
to determine their position relative to the
groups and hence their possible taxonomic
affinity.

For use in multivariate analysis, the 15
measurements listed in appendix B were se-
lected. These measurements were considered
to represent all of the main dimensions of the
skull plus those of three of the more diag-
nostic teeth. Additional measurements, es-
pecially of the teeth, which are individually
of diagnostic value, could have been added.

1979

NOWAK: NORTH AMERICAN QUATERNARY CAMS

But large series of specimens were desirable,
and so measurements of parts too often
missing or defective were excluded. Also it
was reasoned that the 15 utilized measure-
ments would adequately express the major
functions of the skull. Because the entire
mandible was occasionally missing from
specimens, measurements of the lower jaw
and teeth were omitted in multivariate analy-
sis. Several test runs of the program, involv-
ing as many as 35 measurements, including
those of the mandible, did not seem to pro-
duce results different from those that follow,
nor to noticeably increase the discriminatory
ability of the analysis.

Lawrence and Bossert (1967) divided
each of 15 measurements by greatest length
of skull and entered their analysis with the
resulting series of fractions, intending thereby
to eliminate size as a discriminating factor.
Actually there are various expressions of size
of a skull, and dividing by any one of them
may produce different results. Furthermore,
it is questionable whether any attempt should
be made to eliminate the size factor, because
it appears to be a definite biological factor,
at least in distinguishing the wild species of
North American Canis. Certain skulls, repre-
senting two kinds of Canis that would not
ordinarily be confused because of size differ-
ences, may have similar proportions of great-
est length to other measurements.

Therefore raw measurements were used
in most of the following calculations. This
procedure considers the size of each meas-
urement simultaneously as a factor in classi-
fying a specimen. Since the sizes of the
various measurements may vary at different
rates between different species, proportion is
also a factor in the analysis. It is true that an
unusually large or small specimen may be
assigned to a group other than that which
its proportions indicate, but such occurrences
are rare. In any case, my tests of this par-
ticular computer program, using the selected
15 measurements, revealed that in most in-
stances variables based on raw measurements

and fractions of greatest length of skull pro-
duced similar depictments of relationship,
but that the raw measurements gave a wider
separation between groups.

With one major exception (dogs, see be-
low), different analyses were used for males
and females, and it was found that such a
procedure usually produced wider separa-
tion between groups than was achieved by
combining sexes. This wider separation oc-
curred regardless of whether raw measure-
ments or fractions of greatest length were
used as variables.

The sex of some of the skulls utilized in
statistical analysis was unknown, and these
skulls were assigned to male or female cate-
gories on the basis of size and the other
factors explained below. Fortunately, ex-
cepting domestic dogs, each of the major
standard groups, against which other material
was tested, consisted predominantly of speci-
mens of known sex. In the subsequent pages,
when a sample size of one sex is listed, it is
followed by the number (in parentheses) of
specimens in the series (if any) for which
sex had not been recorded, but which were
judged to belong to that sex.

In the statistical analyses, specimens of
domestic dogs (C. familiaris) were not sepa-
rated by sex. Dog skulls are poorly repre-
sented in museum collections, compared with
skulls of wild Canis, and less than half of the
50 specimens of C. familiaris used in my
analyses were of known sex. Individual vari-
ation in this species is so great that it tends
to obscure sexual differences in the morphol-
ogy of the skull. Consequently, dogs of male,
female, and unknown sex were combined in
one group.

In addition to multivariate analyses, tables
of measurements, with means, extremes,
standard deviations, and coefficients of varia-
tion, are provided in appendix B. In some
cases I also have drawn ratio diagrams that
depict differences in size and proportion be-
tween the specimens of various groups.
Whereas multivariate analysis demonstrates

6

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

the collective results of such differences, the
ratio diagram permits visualization of how
each group differs in individual measure-
ments. Simpson ( 1941 ) explained this method
in detail in the course of his account of
Pleistocene felines. Briefly, raw statistics (in-
dividual measurements, means of a series,
etc.), taken on two or more specimens or
series, are converted to their logarithms. One
of the specimens or series is taken as a stand-
ard, and the difference is found between the
logs of its individual measurements or means
and the respective logs of the other speci-
mens or series. In diagramming, the standard
values are all plotted in a vertical line repre-
senting the zero point, and the respective
values of the other specimens or series are
plotted at a horizontal distance from the
standard values, representing the difference
between the two values.

Many skulls were examined for which the
complete set of 15 measurements, required in
multivariate analysis, could not be obtained
because of damage, wear, or missing parts.
Except for fossil material, data from such
specimens were not incorporated in the ratio
diagrams or statistical tables (appendix B).
Therefore, the groups represented in the mul-
tivariate analyses, ratio diagrams, and tables
are all of identical composition. Specimens
not used for the calculation of statistics did
not appear to differ from the main series.

I finally want to make it clear that I used
multivariate analysis in a supporting and
demonstrative role, rather than as a problem
solver in itself. The analysis did not provide
any major conclusions that were not apparent
from more conventional methods of examina-
tion, but it did allow the efficient evaluation
of many data, and the objective, graphical
portrayal of a complex situation.

Age and Secondary Sexual Variation

The aging process in Canis was described
by Goldman (1944:400-401), Jackson (1951:
250-251), Miller, Christensen and Evans

(1965:652-653). and Mech (1970:139-143).
Gier (196S:54-55) showed how to estimate
the age of C. latrans by examination of wear
on the incisor and canine teeth. Linhart and
Knowlton ( 1967) demonstrated a method of
aging coyotes through evaluation of cemen-
tum layers in the canine teeth.

By the age of six months in Canis, the
permanent dentition, except for the canine
teeth, is fully in place, and the skull has
reached approximately 90 percent of its even-
tual total length. Complete emergence of
the canines, and maximum dimensions of the
skull, however, are not attained until about
12 months in coyotes and 15 months in larger
gray wolves. Therefore, for the calculation of
statistics in the following sections of this
paper, I used only skulls of animals estimated
to be at least 12 months old, and did not use
some wolves that were under 15 months old.
The males of Canis average larger than
the females in every measurable dimension
of the skull, but there is extensive overlap
between the two sexes. Males have propor-
tionally broader rostra and higher sagittal
crests. In many female coyotes the sagittal
crest is flattened, and the temporal ridges
that usually coalesce in males are in some
females Urate.

Statistical comparison was made of meas-
urements of skulls of 97 male and 61 female
C. latrans testes from Colorado and Idaho,
and of skulls of 51 male and 35 female C.
lupus mogollonensis, youngi and irremotus
from the mountainous region of the western
United States. In this particular test, only
specimens of known sex were used. Each
species was examined separately and a large
overlap of the two sexes was found in all 15
of the measurements considered. Males av-
eraged larger in each measurement, however,
and analysis of variance and STP tests
showed a significant difference (p less than
.05) between the males and females of each
species in all measurements except postorbi-
tal constriction of braincase.

HISTORY AND STATISTICAL ANALYSIS
OF RECENT POPULATIONS

Comparison' of Known Series of Dogs, Wolves, and Coyotes

The questions to be considered in this
paper concern primarily eastern North Amer-
ica and fossil history. Is the red wolf of the
southeastern United States a full species, a
subspecies of the gray wolf, a subspecies of
the coyote, or a hybrid between C. lupus
and C. latrans? What is the origin and affin-
ity of the populations of Canis that recently
have become established in much of the
eastern half of the continent? Is hybridiza-
tion a major factor in the situation? Is the
Pleistocene dire wolf completely distinguish-
able from the modern gray wolf, and is it
possibly ancestral to the latter?

Before attempting to answer these and
other questions, it would be advisable to
delineate the perimeters of those populations
that seem best to represent recognizable spe-
cies. Throughout most of that part of North
America in which the coyote, gray wolf, and
domestic dog are found together, they are
easily distinguishable and usually behave to-
ward one another as species. The gray wolf
once occurred in all of North America except
for parts of the southeastern United States,
most of the state of California, Baja Califor-
nia and the coastal lowlands of Mexico, and
the region south of central Mexico, ( Goldman,
1944:414). A record of C. lupus baileyi from
Tequisistlan, Oaxaca, southern Mexico
(Goodwin, 1969:224) seems to have been
based on questionable evidence. The coyote
was originally found throughout most of the
western half of the continent, and its range
in the northeast extended as far as the upper
Great Lakes (Young, 1951:29). The domes-
tic dog, C. familiaris, has long occurred in
all parts of the continent, almost always in
association with man. All dogs may have
descended from a small southwest Asian sub-
species of C. lupus that was domesticated 10

to 12 thousand years ago (Scott, 196S). The
dogs of the American Indians were appar-
ently introduced into the New World by man,
and do not seem to have been influenced by
interbreeding with native species of North
American Canis (Allen. 1920; Haag, 1948).
Specimens of the earliest known domestic
dogs on the continent were described from a
site in Lemhi County, Idaho dated at 10,400-
11,500 B.P. They reportedly already possess
the typical characters of C. familiaris (Law-
rence, 1966, 1968).

Only in the eastern part of North America
do hybridization and modification of the orig-
inal populations appear to be of possible
significance. A few isolated instances in
which C. familiaris hybridized with either
C. lupus or C. latrans in other regions, have
been reported (Young, 1944:180-210; Men-
gel, 1971; Gray, 1972), but such cases do not
seem to have had lasting effect on popula-
tions. No instances of interbreeding between
C. lupus and C. latrans in the western half
of the continent have yet been reported.
Therefore, it is reasonable to consider C.
lupus and C. latrans of the western and
northern parts of the continent as consisting
of natural, unmodified populations that may
confidently be used as a basis on which to
test more questionable populations.

For an initial analysis of known groups, I
decided upon using skulls of C. latrans and
C. lupus that had been collected not later
than 1925 in the mountainous region of the
west. The gray wolf sample included 57(6)
males and 37(2) females (parentheses con-
tain numbers of specimens in the series for
which sex had not been recorded, but which
were judged to belong to the particular sex
indicated; see p. 5). This group consisted
of all skulls of adult C. lupus mogollonensis,

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

ijoungi, and irrcmotus in the U.S. National
Museum of Natural History, upon which the
15 necessary measurements could be made,
except for two specimens taken after 1925
(see appendix A, part 1).

Of coyotes, 97 male and 61 female skulls,
all of the subspecies C. latrans lestes, were
utilized in the initial test (see appendix A,
part 2). These specimens comprised the en-
tire National Museum collection of Idaho and
Colorado adult lestes of known sex, taken
prior to 1926, except for skulls upon which
all of the needed 15 measurements could not
be made.

The selection of these particular speci-
mens as standard comparative material had
the following advantages : ( 1 ) the wolf and
coyote had long been sympatric in the region,
and thus theoretically would have evolved
the maximum amount of differential charac-
ters reflecting their separate ecological niches;
(2) the region had a minimum human (and
presumably domestic dog) population; (3)
the time period was one in which both the
wolf and coyote were common (most of the
specimens represent the first few years of
Federal predator control work which began
in 1915); and (4) the subspecies of both C.
lupus and C. latrans do not exhibit extremes
of size or other characters within their re-
spective species.

For a sample of domestic dogs, only those
skulls were selected which, while known to
be C. familiaris, were superficially nearest to
those of C. lupus or C. latrans in appearance.
The extremes of domestication represented
by broad-skulled dogs (as bulldogs), narrow-
skulled dogs (as Russian wolfhounds), and
dogs having greatly reduced rostra (as pugs)
were avoided. Extremely small dogs, those
in which the skull was less than 150 milli-
meters in greatest length, also were not used.
Specimens utilized in the sample included 1
Eskimo dog, 5 Irish wolfhounds, 3 German
shepherds, 2 sheep dogs, 2 Newfoundlands.
1 doberman pinscher, 1 greyhound, 1 great
Dane, 1 mastiff, 1 Irish setter, 1 beagle, 1

- LATRANS

LUPUS

FAMILIARIS

— I 1 r-

-4 -3 -2

-i 1 1-

2 3 4

5
4
3
2
I

0
-I
-2
-3
-4
-5
-6

LATRANS

LUPUS

FAMILIARIS

■3 -2

3 4 5 6

Fig. 1. — Graphical results of multivariate analy-
ses comparing samples of C. lupus and C. latrans
from the mountainous region of western North Amer-
ica, and C. familiaris. Only the margins of the range
of variation of each species are shown. In this and
in all subsequent portrayals of multivariate analyses,
the numbers along the vertical and horizontal axes
are canonical coordinates. These coordinates are used
to indicate relative position, and do not represent
any material values. In this figure and subsequent
portrayals of analyses, males are shown above and
females are shown below.

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

Fig. 2. — Range map of C. lupus in North America showing localities (black dots) of specimens used in the
statistical analyses of this paper. The numbers on the map represent recognized subspecies, as follows:

1. C. I. alces

2. C. I. arctos

3. C. Z. baileyi

4. C. I. beothucus

5. C. I. bernardi

6. C. /. columbianus

7. C. I. crassodon

8. C. I. fuscus

9. C. /. hudsonicus

10. C. /. griseoalbus

11. C. ?. irremotus

12. C. /. labradorius

13. C. /. Zigoni

14. C. /. lycaon

15. C. Z. mackenzii

16. C. /. manningi

17. C. /. mogollorwnsis

18. C. /. monstrabilis

19. C. Z. nubilus

20. C. Z. occidentalis

21. C. Z. orion

22. C. Z. pamba-sileus

23. C. Z. tundrarum

24. C. Z. youngi

The solid lines indicate subspecific boundaries. The dashed line in southeastern Ontario shows Standfield's
(1970) division between his "Ontario type" and "Algonquin type" of C. lupus lycaon. Because of the scale
of the map, it was not possible to plot all localities in crowded areas.

10

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

scale of miles

Fig. 3. — Range map of C. latrans showing localities of specimens used in the statistical analyses of this
paper. The numbers on the map represent recognized subspecies, as follows:

1. C. I. cagottis 7. C. /. impavidus 13. C. /. microdon

2. C. I. clepticus 8. C. /. incolatus 14. C. I. ochropus

3. C. /. dickeyi 9. C. I. jamesi 15. C. /. pcninsulae

4. C. Z. frustror 10. C. /. latrans 16. C. /. f<?xen.rfs

5. C. /. goldmani 11. C. /. /c.sres 17. C. /. thamnos

6. C. /. hondurensis 12. C. /. mearnsi 18. C. /. umpquensis

The solid lines indicate subspecies boundaries. Because of the scale of the map, it was not possible to plot all
localities in crowded areas. Certain additional localities of C. latrans are shown in Figs. 14, 25, and 31.

19. C. /. vigilis

20. C. I. "var." (Law-
rence and Bossert,
1969)

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

11

basset hound, and 30 dogs of unknown or
mixed breed (see appendix A, part 3). Of
these 50 specimens, only 11 males and 9 fe-
males had been previously identified as to
sex. For the reasons explained above, do-
mestic dogs of male, female, and unknown
sex were combined into a single sample.

The graphical results of multivariate anal-
yses of males and females of the initial sam-
ples are depicted in figure 1. For both sexes
there is complete separation between all
three species. Such an arrangement could
be expected, and it may serve as a sound
basis on which to evaluate other specimens.

The next step was to compare skulls taken
elsewhere in northern and western North
America to the above series of reliably dis-
tinguished specimens. Skulls of 176(27)
males and 114(33) females, previously identi-
fied as C. lupus, and of 69(2) males and
50(4) females, identified as C. latrans, were
tested individually against the three known
groups (see appendix A, parts 4 and 5). The
maps in figures 2 and 3 show localities of all
specimens of gray wolves and coyotes. Fig-
ure 4 shows the results of multivariate analy-
ses. Nearly all of the newly added material
falls within the range of variation of the ap-
propriate original sample, or at least is closer
to this range than to that of other species.

Only five of these specimens (see appen-
dix A, part 6) seem confusing as to identity.
Three skulls, previously identified as C. lupus
baileyi, are statistically and morphologically
intermediate to known samples of female gray
wolves and coyotes. Canis lupus baileyi, the
smallest subspecies of North American gray
wolf, shared its entire range with C. latrans,
and interbreeding between the two might
have been possible under certain conditions.
I thus consider these three specimens as prob-
able hybrids of C. lupus and C. latrans, and
henceforth have not used them in the forma-
tion of samples of either parent species.

Two females from the Sacramento Moun-
tains of New Mexico, previously identified as
C. lupus monstrabilis, appear both visually

Fig. 4. — Statistical positions of individual native,
\vi!d-caught specimens from throughout northern and
western North America, relative to ranges of variation
of the series of C. lupus, C. latrans, and C. familiaris
shown in Fig. 1. The black dots represent individuals
of C. lupus and C. latrans, all of which fall close to
the appropriate range of variation. The letter B
represents specimens throught to be hybrids between
C. lupus baileyi and C. latrans; the letter M repre-
sents specimens thought to be hybrids between C.
lupus monstrabilis and C. familiaris. Males are above,
females below; numbers along axes are canonical co-
ordinates.

12

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

LUPUS

4
3
2
I

0
-I
-2
-3
-A
-5
-6
-7
-8

FAMILIARIS

-5-4-3-2-IOI2345

Fie. 5. — Multivariate comparison of all specimens
of C. lupus and C. latrans from northern and western
North America, and C. familiaris. Only the margins
of the range of variation of each species are shown.
Males are above, females below.

and statistically (Fig. 4) to be intermediate
to the gray wolf and domestic dog. At least
two other skulls collected in New Mexico,

but not complete enough for inclusion in the
multivariate analysis, also suggest the oc-
currence of hybridization between the two
species. Nonetheless, the preponderance of
material that can be clearly identified as
either C. latrans or C. lupus, indicates that in
those regions hitherto discussed the relation-
ships of these two canids to one another and
to C. familiaris were those of normal species.
All of the northern and western speci-
mens, except the five considered to be hy-
brids, were incorporated with the appropriate
standard samples of C. lupus or C. latrans,
and these two groups along with the sample
of 50 domestic dogs were tested in new multi-
variate analyses. The graphical results de-
picted in figure 5 once again indicate clear
separation between the three species. Meas-
urements for the total series of dogs, and of
western and northern wolves and coyotes are
listed in appendix B (parts 1 and 2). The
means of these measurements (of males
only for C. lupus and C. latrans) are com-
pared in the ratio diagram in figure 6.

Systematic Problems in
the Northeast

Decline of the Gray Wolf

According to Goldman ( 1944 ) , a single
subspecies of gray wolf, C. lupus lycaon, orig-
inally occupied the region from eastern Min-
nesota to the Atlantic, and from northern
Ontario to parts of the southeastern United
States. Considering the enormity of this
range, however, and the problems associated
with the systematics of Canis in eastern North
America, Goldman used relatively few speci-
mens for describing the situation. From the
entire region south of Lakes Michigan and
Erie he assigned to lycaon only four complete
skulls and one mandibular ramus.

Standfield ( 1970 ) reported the presence
of two distinct kinds of lycaon in Ontario,
which he designated the "Ontario type" and
the "Algonquin type." The former was said
to occur mainly in the boreal forests north

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

13

-20 -.16 -.12 - 08 -.04 0
1 1 — i 1

-i 1 1 1 1 r

LATRANS

LUPUS

FAMILIARIS

.70

80

90

100

Fig. 6. — Ratio diagram comparing means of total
samples of C. lupus and C. latrans from northern and
western North America (males only are shown), and
C. familiaris (dashed line). Vertically arranged num-
bers represent the measurements so numbered in Ap-
pendix B. A log difference scale is provided above,
and a ratio scale below the diagram.

and northeast of Lake Superior, to be larger,
and to vary in pelage from pure white to jet
black. The latter was said to dccur in the
deciduous forests east and southeast of Lakes
Superior and Huron, to be smaller, and to be
invariably gray-brown. Kolenosky and Stand-
field (1975) made a multiple discriminant
analysis comparing 105 skulls of the "Algon-
quin type" and 122 skulls of the "Ontario
type" (now referred to as the "Boreal type"),
and concluded that there were significant
differences between the two. These authors
(p. 71) reported that "the ranges of the two
types overlap throughout a broad band
across eastcentral Ontario, but there is no
conclusive evidence of their interbreeding."

Mech and Frenzel ( 1971 ) suggested that
the present population of wolves in north-
eastern Minnesota consists at least in part of
the subspecies C. lupus nubilus. This idea
was based on their observations of black
wolves and white wolves in the area. Indi-
viduals of these colors were common among
nubilus, but were not reported in cited ob-
servations of lycaon in eastern Ontario.

Whatever their original systematic status,
northeastern wolves have suffered a drastic
loss in numbers and range because of perse-
cution by Caucasian man. The species prac-
tically disappeared south of the St. Lawrence
River between 1850 and 1900 (Peterson,
1966:200; Goodwin, 1936), and was gone
from southern Michigan and Wisconsin early
in the twentieth century (Jackson, 1961:293;
Arnold, 1952). Wolves were reported to be
present in moderate numbers in the upper
peninsula of Michigan by Stebler ( 1944 ) ,
and in northern Wisconsin by Schorger
(1942). In the 1950"s, however, the wolf
populations of these two states declined
sharply. Jackson (1961:293) estimated 50
wolves to be present in northern Wisconsin,
and Keener (1970) reported that none still
survived in the area. For the upper penin-
sula of Michigan, recent numerical estimates
have been 20 (Smits, 1963), less than 12
(Douglass, 1970), and about six (Hendrick-
son and Robinson, 1975). In March 1974
four wolves from northern Minnesota were
released in upper Michigan, but by Septem-
ber all had been killed through human
agency (Weise, et o/., 1975). A viable group
of about 20 to 30 individuals, however, has
maintained itself on Isle Royale in Lake
Superior since the late 1940's (Mech, 1966;
Wolfe and Allen, 1973).

The only major population of C. lupus
to be found anywhere in the United States
south of Canada is that in northern Minne-
sota. An estimated 1,000 to 1,200 individuals
are reported to exist in the area and they are
said to be in no immediate danger of extir-
pation (Mech, 1977).

14

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

Despite intensive control measures, wolves
reportedly still occupy all of the forested
parts of Ontario, even areas close to Toronto
and Ottawa. They do not, however, enter
the settled agricultural sections of that prov-
ince (Clarke, 1970). Here and in other parts
of the northeast, there evidently has been a
correlation between the decline of the gray
wolf and the intensity of human population
and agricultural development.

Rise of the Coyote

At the time the first white settlers arrived
in North America, coyotes were apparently
confined to open plains and more arid re-
gions, mainly in the western half of the con-
tinent (Young, 1951; Seton, 1929). The
original range of the species did, however,
follow the prairie peninsula through the mid-
western states, at least as far as northeastern
Indiana (Mumford, 1969:85). According to
Jackson (1961:285) coyotes were undoubt-
edly present in southern Wisconsin when the
early explorers arrived, and a few may have
inhabited the northern part of the state. The
name C. latrans thamnos was applied by
Jackson ( 1949 ) to the coyote of the north-
eastern portion of the range of the species.

Sometime after the middle of the nine-
teenth century, coyotes began appearing to
the north, east, and south of the prairies, and
by the mid-twentieth century they existed in
large numbers beyond their original range
(De Vos, 1964; Mech, 1959, 1961; Young,
1951). Many coyotes had escaped from cap-
tivity, or had been deliberately released by
sport or bounty hunters. Some of these ani-
mals formed local breeding populations that
maintained themselves over a period of time.
Records that seem attributable to such intro-
ductions are as follows.

Florida.— Palm Beach County, 1925; Col-
lier, Monroe, and Marion counties (Young,
1951:15); DeSoto County, 1933 (Sherman,
1937); Polk County, 1962 (Cunningham and
Dunford, 1970).

Alabama.— Barbour County, 1924-1929
(Young, 1951:16); Bazemore, Fayette Coun-
ty, 1956; Huntsville Arsenal, Madison County,
1961 (Holliman, 1963).

Georgia.— North Georgia, 1929; Haber-
sham County, 1930; Ware County (Young,
1951:15).

South Carolina. — Aiken County; Edisto
Island, Colleton County, 1924 (Golley, 1966:
141).

North Carolina.— Gaston County, 1938
(Young, 1951:14); Swain County, 1947 (Lin-
zey and Linzey, 1968).

Tennessee. — Grand Junction, Hardeman
County; Maury County, 1930; McCains,
Maury County, 1931 (Kellogg, 1939:267);
Hickman and Maury counties, early 1930's
(Young, 1951:15); Benton, Hickman, and Se-
quatchie counties (Schultz, et ah, 1954:205);
Tennessee National Wildlife Refuge, Henry
County, 1951 (Schultz, 1955).

Kentucky. — Near Fayette-Clark county
line, 1953 (Gale and Pierce, 1954).

Virginia. — Rockingham, Highland, and
Grayson counties (Handley and Patton, 1947:
140); Tazewell and Lee counties, 1968-1969;
near Mossy Creek, Augusta County, 1970
(Carpenter, 1971).

West Virginia. — Tucker County ( Hand-
Ian, 1946).

Maryland. — 5 mi. NW Poolsville, Mont-
gomery County, 1921 (Jackson, 1922); out-
skirts of Baltimore, 1931 (Redington, 1931:
27); Cecil County, 1961 (Paradiso, 1969:
134).

New Jersey. — Near Ringoes, Hunterdon
County, 1938 (Young, 1951:16); near Fishing
Creek, Cape May County, 1948 (Ulmer,
1949).

Pennsylvania. — Clinton County, 1915-1916
(Shoemaker, 1917:11); near Flowing Spring,
Blair County, 1907 (Gifford and Whitebread,
1951:46); Sheshequin Township, Bradford
County, 1939; Chestnut Hill section, Phila-
delphia, 1942 (Ulmer, 1949); Beaver and
Forest counties, 1946-1947 (Richmond and
Rosland, 1949:34).

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

15

New York. — Near Ithaca, Tompkins Coun-
ty, 1920 (Seton, 1929:369); Ontario County,
about 192S (Rump, 1941:415).

Massachusetts. — Vicinity of Amherst,
Hampshire County, 1936 (Warfel, 1937).

Maine.— Near Portland, 1932 (Aldous,
1939).

Some other occurrences in New England
and New York probably represent range ex-
tensions and are discussed below. But it is
difficult to determine whether some records
should be designated as introductions or
range extensions.

The actual eastward extension of the coy-
ote's range has seemingly been slowest in the
region between Michigan and the Ohio Val-
ley, possibly because of the intensity of hu-
man population there. Some sources, such as
Seton (1929:368) even suggest that the coy-
ote had been exterminated in part of this
region, although Mumford (1969:85) doubt-
ed that the species had ever been extirpated
in Indiana. Jackson (1961:285) noted a re-
versal of the original situation in Wisconsin,
in that coyotes had become much more com-
mon in the northern half of the state than in
the southern part. Burt ( 1946 ) observed that
the species was rare in the southern part of
the lower peninsula of Michigan. Hoffmeister
and Mohr (1957:119) stated: "The coyote
occurs in much of Illinois, but it is not com-
mon anywhere in the state." And Mumford
( 1969 : 84 ) reported coyotes to - be present
throughout Indiana, but not to be common.
Hamilton (1943:178) indicated that coyotes
were of sporadic occurrence in most of west-
ern Ohio, and that they had been established
in Logan County for 12 years. Other occur-
rences in the western counties of Ohio were
reported by Negus (1948), Whitacre (1948),
Young (1951:15), and Goodpaster and Hoff-
meister (1968). Wilson (1976) mentioned
the presence of a sparse coyote population in
western Kentucky.

The coyote's range seems to have ex-
panded mainly to the northeast of the orig-

inal distribution in the prairie peninsula. At
the base of the lower peninsula of Michigan,
individuals were reported in Rerrien County
in 1900-1901 (Wood and Dice, 1924) and in
Washtenaw County in 1905 and 1910 (Wood,
1922). A specimen was taken in Genesee
County, in the east-central part of the state
in 1917 ( Wood and Dice, 1924 ) , and an indi-
vidual was reported in Charlevoix County,
at the northern tip of the lower peninsula, in
1919-1921 (Dice, 1925).

Coyotes had also entered the upper penin-
sula of Michigan by the early twentieth cen-
tury. Shiras (1921:166) reported: "In the
past fifteen years the coyote unexpectedly ap-
peared in northern Wisconsin and Michigan,
coming from Minnesota. It has since become
very numerous." Wood and Dice (1924)
listed occurrences in five upper peninsula
counties between 1912 and 1915.

Even earlier, coyotes had begun to move
north in Minnesota. According to Bailey
( 1929 ) they first appeared in Sherburne
County in 1875, and Calm (1921) reported
them to be more common than wolves in
Itasca County.

Snyder (1938) thought that coyotes were
present in the western Rainy River District,
Ontario in 1S90. Peterson (1966:197) noted
that in Ontario before 1900 the species was
restricted to Rainy River and western Kenora
districts. Krefting (1969) suggested that coy-
otes moved to Isle Royalc in Lake Superior
sometime prior to 1912-1913, from the Sibley
Peninsula area of Ontario, where they had
arrived about 1900. Coyotes continued to in-
habit Isle Royale until gray wolves, in a re-
versal of the general trend in the northeast,
occupied the island in the 1940's.

From western Ontario, coyotes spread
eastward above Lake Superior, and northward
toward Hudson Ray. They also apparently
crossed the St. Clair River into southeastern
Ontario where the first specimen was taken,
north of Thedford. Lambton County, in 1919.
In 1943 specimens were collected in Essex,
Peterborough, and Carleton counties, Ontario

16

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

(Anderson, 1946), and by 1956 the species
was reported to occur nearly throughout the
province (Peterson, 1957). The first record
for Quebec was a specimen taken near Lusk-
ville, Gatineau County in 1944 (Rand, 1945).
According to Wolfram (1964) the first coyote
to be found in New Brunswick was killed
near Sussex in 195S. Subsequent expansion of
the species in Quebec, as far as the Gaspe
Peninsula, and in New Brunswick, was re-
ported by Georges ( 1976 ) .

Bromley (1956) summarized the history
of the coyote in northern New York. Indi-
viduals were shot in the St. Lawrence River
area in 1925 and in Franklin County in the
mid-1930's. The species was said to have
achieved a good foothold in the early 1940's
when its range included several areas of the
Adirondack Mountains. By the late 1940's
coyotes had spread throughout the Adiron-
dacks, and in the early 1950's they occupied
the entire northern third of the state. Subse-
quently, according to Severinghaus ( 1974a,
1974b), there have been records from much
of southern New York. An estimate of 5,000-
15,000 coyotes for the state was published by
Marvinney (1976).

Coyotes apparently are continuing to
move down the Appalachian Mountains from
the northeast. They are currently reported to
be rare, but widely distributed in Pennsyl-
vania (John L. George, Department of Wild-
life Management, Pennsylvania State Univer-
sity, pers. comm. ), and their "sporadic pres-
ence" in West Virginia was discussed by
Taylor, Counts, and Mills (1976). In May
1976 a specimen was taken near Nestorville.
Barbour County in northeastern West Vir-
ginia, and I found the skull to closely resem-
ble those of some New England coyotes.

In 1936 a coyote reportedly was killed in
Argyle Township, Penobscot County, Maine.
Over the next two years 11 other individuals
were taken in that vicinity, most of which
were considered to be hybrids between C.
latrans and C. familiaris (Aldous, 1939). A
specimen identified as C. latrans thamnos

was taken in Lower Enchanted Township,
Somerset County, Maine in 1961 (Carson,
1962). The subsequent occupation of nearly
the entire state by the species has been docu-
mented by Richens and Hugie (1974) and
Teer (1975). The Maine Department of In-
land Fisheries and Wildlife (1976) has pub-
lished an estimate of from 1,500 to 5,500
coyotes in the state.

Coyotes were first reported in Vermont in
1942 and have since been taken in all 14
counties of the state. The first record in New
Hampshire was one shot near Holdemess,
Grafton County in 1944, and the first in
Massachusetts was shot near Otis, Berkshire
County in 1957. Three more were trapped in
Massachusetts on the Prescott Peninsula of
Quabbin Reservoir in 1958, and one was shot
near Grafton, Worcester County in 1959. Four
individuals, believed to be coyotes or hybrids
between C. latrans and C. familiaris ("coy-
dogs"), were taken in western Connecticut
from 1957 to 1963 (Pringle, 1960, 1963).
Silver and Silver (1969:Fig. 30) depicted oc-
currences throughout Vermont. New Hamp-
shire, and Massachusetts.

Difficulties in Identifying Northeastern Canis

Although the recent presence of wild
Canis in the northeast does appear to repre-
sent primarily an extension of the range of
the species C. latrans, the exact identity of
certain individuals and populations has been
open to question. Perhaps because coyotes
had never before occurred in the region, their
initial appearance was something of a mys-
tery and the cause of excitement among the
public and local wildlife officials. Some per-
sons believed that wolves were returning to
areas in which they had been exterminated
long before, while others considered the new
canids to be wild dogs or coy-dogs. And it
is likely that each of these three kinds of ani-
mals contributed in some part to the mystery.
An actual specimen of C. lupus was killed in
Fulton County, New York in 1968 (Paradiso
and Schierbaum, 1969). Another wolf, of

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

17

unsuaUy large size and probably not a native
animal, was killed north of Kemptville, Carle-
ton County, Ontario, about 30 miles from the
New York border, in 1962. According to
Bump (1941) an undetermined number of
gray wolves escaped from captivity in south-
ern Franklin County, New York about 1930.
Silver and Silver (1969) said that four gray
wolves were imported to the vicinity of Croy-
don, Sullivan County, New Hampshire, and
that after the last died in 1914 there were
regular reports of wolflike animals in that
area.

Completely feral dogs seem to be uncom-
mon, but do occur on occasion, and are some-
times the cause, directly or indirectly, of re-
ported wolves and coyotes. Carson (1962)
referred to several cases of C. familiaris living
and breeding in the wild in Maine. McKnight
(1964:48) reported that a pack of wild dogs
inhabited a den near Hopkinton, New Hamp-
shire for six years prior to 1959. And Nesbitt
( 1975 ) made a five-year study of a feral pack
on Crab Orchard National Wildlife Refuge,
Illinois.

Aldous (1939) and Carson (1962) re-
ferred to the taking of numerous coy-dogs in
Maine. Wetzel and Penner (1962) reported
the collection of two specimens of coy-dogs
in Litchfield County, Connecticut. Cook
(1952) discussed the presence of such ani-
mals in New York. Wolfram (1964) cited a
report that 20 percent of the coyotes in On-
tario were actually coy-dogs. Paul (1970)
stated that 20 percent of the carcasses of wild
Cards found in Illinois were identified as coy-
dogs. In tracing the spread of coyotelike
animals in New York, Severinghaus ( 1974a,
1974b) noted that the first litters in a newly
occupied area were almost always obvious
coy-dogs.

Such records apply to wild-caught animals
that were presumed to be hybrids between
C. latrans and C. familiaris on the basis of
morphological characters. Cases in which
known hybrids were born in captivity under
controlled conditions were discussed by Dice

(1942), Young (1951:123), Kennelly and
Roberts (1969), Gier (1968), and Mengel
( 1971 ) . The latter three of these authors
reported such hybrids to be fertile.

Mengel (1971) reviewed the subject of
hybridization between C. latrans and C. fa-
miliaris, and presented information on his
own experiments. He noted that whereas
coyotes normally mate from late January to
March, and usually give birth in the spring,
coy-dogs have been observed to mate from
October to December, and to give birth in
the winter. This phase shift in the breeding
cycle was held to be a barrier restricting the
interbreeding of coy-dogs with C. latrans,
and hence preventing the introgression of
domestic dog genes into the wild coyote pop-
ulation. Furthermore, Mengel pointed out
that since the offspring of coy-dogs would be
born under harsh winter conditions, and since
male coy-dogs do not demonstrate the same
tendency to parental care as male coyotes,
the hybrid pups would be unlikely to survive.
These factors taken together seemingly would
prevent the establishment of a population of
canids of mixed coyote and dog ancestry.
Gier (1968) and Kennelly and Roberts (1969)
also reported the shift in breeding time
among coy-dogs. Iljin (1941) found a parallel
situation in captive hybrids of C. lupus and
C. familiaris.

Gipson (1972), Gipson, Sealander, and
Dunn (1974), and Gipson, Gipson, and Sea-
lander ( 1975 ) concluded that in Arkansas
introgression of dog genes into the coyote
population could and did occur. This view
was based in part on the fact that 38 of 284
skulls of wild canids recently collected in
Arkansas, were shown by multivariate analy-
sis to be morphologically intermediate to dog
and coyote populations, and thus were desig-
nated coy-dogs. There was no evidence, how-
ever, that any of these individuals were other
than first generation offspring. It also was
found that some male coyotes in Arkansas
were physiologically in breeding condition in
late November and December. On the basis

18

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

of examination of female reproductive tracts,
however, the earliest reported actual mating
was 17 February.

Freeman (1976:40-42) pointed out that
in Oklahoma, at least, winter birth would not
seriously affect the offspring of coy-dogs,
since weather conditions were not so severe
in that area. He also suggested that lack of
care by the male parent would not necessarily
result in loss of the litter, since female coy-
otes often successfully rear young after losing
their mate.

Silver and Silver ( 1969 ) raised and stud-
ied a litter of five coyotelike canids dug from
a den in Croydon, Sullivan County, New
Hampshire, and 50 of their descendents. Two
hybrid litters were born to one of the orig-
inal females crossed with domestic dogs. The
hybrids were fertile and produced two litters,
but once again it was found that breeding
occurred three to four months earlier than in
wild Cams, and that males did not assist in
rearing the young. Considering these factors,
and also the disadvantages faced by pups
born in midwinter, Silver and Silver thought
it unlikely that eoy-dogs could establish or
merge with a wild population. On the basis
of behavioral and physical studies of these
various animals, it was concluded that the
present population of New England Canis is
of predominantly coyote ancestry, but that
some dog and/ or wolf genes had been intro-
duced in the past. Such genes, however, had
become well integrated so that the population
was now stabilized and breeding true to
type. Therefore, Silver and Silver thought
that wild Canis in New England should not
be considered as hybrids, but as a kind of
coyote.

Lawrence and Bossert ( 1969 ) subjected
31 skulls of New England Canis to linear dis-
criminant analysis. Of these specimens, 16
were offspring of the captive canids studied
by Silver and Silver, and 15 were killed wild
in New Hampshire, Vermont, and Massachu-
setts. In the analysis, most of the specimens
fell between known samples of C. lupus and

C. latrans, a few overlapped with C. latrans,
and a few also approached the range of C.
familiaris. The positions of these wild indi-
viduals were, however, substantially different
from those of 21 known, captive raised Fi
and FL» coyote-dog hybrids. Lawrence and
Bossert, like Silver and Silver, concluded that
the New England population was predomi-
nantly coyote, probably with some dog/ wolf
ancestry. A series of 32 skulls of C. latrans
thamnos from Minnesota also demonstrated
a shift away from typical C. latrans toward
both C. familiaris and C. lupus, but not to the
extent found in New England Canis. Hence
the New England population was considered
to represent the development of a trend that
had begun at an earlier time through the
introduction of wolf or dog genes into the
coyote population. The designations "Canis
latrans var." and "eastern coyote" were used
for the wild population of New England
Canis.

Chambers, et al. (1974) used the same
technique as Lawrence and Bossert (1969)
on nearly 150 recently collected specimens
from New York, and came to the same con-
clusion regarding the systematic position of
the population. Still another statistical analy-
sis of skulls was done on Maine material by
Hilton (1977) who found that, except for
five dogs, all specimens received from 1968
to 1975 could be identified as eastern coyotes.

As previously mentioned, Mengel ( 1971 )
thought that the introduction of dog genes
into a population of wild Canis would be un-
likely. He therefore suggested that the pres-
ent population of New England Canis was
the result of the introgression of genes only
from the wolf ( C. lupus hjcaon ) into the wild
coyote population. This view was supported
in part by a discriminant function analysis in
which skulls of 55 coyotes taken in north-
eastern Kansas, 12 known captive raised coy-
ote-dog hybrids, and 13 presumptive wild
coy-dogs were compared to the same samples
of Canis used by Lawrence and Bossert
( 1969 ) . Since domestic dogs are abundant

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

19

in northeastern Kansas, and since coy-dogs
arc regularly reported in the area, Mengel
argued that the local coyote population
should demonstrate the same variability as
that of New England, if introgression from
C. familiaris were possible. Because in the
analysis the 55 coyotes fell close to the orig-
inal coyote sample, and because the positions
of the coyote-dog hybrids substantially dif-
fered from those of the Kansas coyotes, Men-
gel concluded that introgression from C.
familiaris to C. latrans was not occurring.
Canis lupus, long extinct in Kansas, but still
present in parts of the northeast, was thought
to be the only source of the genes that had
modified the northeastern coyote population.
Known cases of hybridization between C.
lupus and C. latrans are rarer than those
between other pairs of species of Canis
(Gray, 1972). Indeed, until recently the only
suggestion that such a cross was possible was
the statement by Young (1951:124) that two
specimens of supposed coyote-wolf hybrids,
born in captivity, were in the Royal Ontario
Museum. Kolenosky (1971) reported that in
May 1969 and May 1970, litters of five healthy
pups were born to a female C. lupus lycaon,
captured in Algonquin Provincial Park, and a
male C. latrans, taken in York County, On-
tario. Kolenosky and Standfield ( 1975 ) added
that a subsequent mating of Fi siblings pro-
duced a litter of four pups. They also ob-
served that the members of the Fj generation
were similar in appearance to many speci-
mens of a small kind of wolf, designated the
"Tweed type," that had been collected along
the southern limits of the range of their "Al-
gonquin type" of C. lupus lycaon in south-
eastern Ontario. They stated that evidence
was mounting that the "Tweed type" had
originated from hybridization between C. la-
trans and C. lupus lycaon in the wild.

Examination of Specimens

With the above background in mind, I
examined all skulls of wild Canis from the
northeastern United States and southeastern

LUPUS

LATRANS

LUPUS

FAMILIARIS

-5 -4 -3 -2 -I 0

2 3

Fie. 7. — Multivariate comparison of individual
specimens of C. lupus lycaon to the ranges of varia-
tion of the total series shown in Fig. 5. Black dots,
statistical positions of lycaon from Michigan, Minne-
sota, Wisconsin, and western Ontario; triangles, posi-
tions of lycaon from southeastern Ontario and south-
ern Quebec; squares, positions of two specimens
thought to be hybrids between C. lupus lycaon and
C. familiaris. Males are above, females below.

20

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

Canada that were available to me. The pre-
viously established total series of northern
and western wolves and coyotes, and the
series of 50 domestic dogs, were used as
standard groups against which the northeast-
ern material was tested (see pp. 11-12).

Each skull previously identified by others
as C. lupus tycoon, except the six to be dis-
cussed later, were compared individually to
the three standard groups by multivariate
analysis. Most of these skulls fall within the
range of variation of C. lupus, but, as could
be expected considering their relatively small
over-all size and narrow proportions, they
demonstrate a shift toward C. latrans (Fig.
7). Two males from the upper peninsula of
Michigan (see appendix A, part 6) have a
statistical position intermediate to those of
C. lupus and C. familiaris, and also appear
by eye to be of mixed blood. These two
skulls, therefore, are considered to represent
hybrids, and henceforth are not used in the
analyses.

Of the remaining specimens represented
in figure 7, a group of 72 were collected in
the upper peninsula of Michigan, northern
Wisconsin, northern Minnesota, and that part
of Ontario to the north and west of Stand-
field's (1970) line separating the "Algonquin
type" and "Ontario type" of C. lupus hjcaon;
and a group of 31 were obtained in south-
eastern Ontario and southern Quebec (see
appendix A, part 7. and Fig. 2). The multi-
variate distribution of these two groups shows
overlap and does not suggest any sharp dis-
tinction. Wolves from the more westerly re-
gion are larger, and have relatively broader
rostra and frontal shields, and smaller second
upper molars (see appendix R, part 3). In
these and other characters the western hjcaon
do appear to fall between the eastern hjcaon
and C. lupus nubilus from the Great Plains.
Thus Mech and Frenzel ( 1971 ) could have a
case in suggesting the survival of nubilus in
Minnesota. Throughout the entire region in
question, however, we seem to be dealing

4 r

LATRANS

LUPUS

4HATRANS

3

2

I

0
-I
-2
-3
-4
-5
-6
-7

-5 -4 -3 -2-10 I

3 4

Fig. 8. — Multivariate comparison of all specimens
of C. lupus (including hjcaon), specimens of C. la-
trans from northern and western North America, and
C. familiaris. Only the margins of the range of varia-
tion of each species are shown. Males are above,
females below.

only with minor and gradual variation, when
considering the species C. lupus as a whole.

1979

NOWAK: NORTH AMERICAN QUATERNARY CAMS

21

These findings are not intended to contra-
diet the conclusion of Kolenosky and Stand-
field (1975) that there is a significant statisti-
cal difference between specimens of C. lupus
hjcaon from various parts of Ontario. Those
authors dealt primarily with a question of
intraspecific variation. My studies tend to
show only that material referred to lycaon
is not more noticeably distinctive from C.
lupus in general than is material referred to
most other named subspecies of the gray
wolf.

Satisfied that the 103 skulls of lycaon
discussed above were representative of the
species C. lupus, I incorporated them into
my gray wolf samples. Graphical results of
multivariate analyses comparing these total
samples with the standard samples of do-
mestic dogs and coyotes are shown in figure 8.

All available skulls, previously identified
in collections as coyotes (with one exception
to be discussed later), and that had been
taken in the range delineated by Jackson
(1951:266-267) for C. latrans thamnos (plus
extreme southern Quebec), were individually
tested against the three standard groups of
C. lupus (including lycaon), C. latrans, and
C. familiaris. The specimens included 80(22)
males and 50(13) females from Manitoba,
North Dakota, Minnesota, Iowa, Illinois, In-
diana, Michigan, Wisconsin, Ontario, and
Quebec (see appendix A, part 8). Relative
multivariate positions are depicted in figure
9. Nearly all specimens fall within or near
the total range of variation of C. latrans, but
are concentrated closer to C. lupus than is the
standard coyote sample. One skull, obtained
in March Township, Ontario, appears to rep-
resent a coyote-dog hybrid. The overwhelm-
ing statistical affinity of thamnos to the stand-
ard sample of C. latrans suggests that the
coyote has established itself in essentially
unmodified form in much of the northeast.

In addition to the above, I was able to
examine a number of specimens from the
extreme northeastern United States where
the wild canid population was designated

LUPUS

LATRANS

LUPUS

-i 1 1 1 r~

-6 -5 -4 -3 -2

-i r-

-i 0

-i 1 1 1-

12 3 4

Fig. 9. — Multivariate comparison of individual
specimens of C. latrans thamnos to the ranges of
variation of the total series of C. lupus, C. latrans,
and C. familiaris shown in Fig. 8. Black dots, sta-
tistical positions of thamnos; square, position of prob-
able hybrid between C. latrans thamnos and C. fa-
miliaris. Males are above, females below.

22

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

"C. latrans var." by Lawrence and Bossert
(1969). Wild-caught animals were repre-
sented by skulls of 25(5) males and 20(2)
females (see appendix A, part 9). Multivari-
ate positions, relative to the three standard
groups, are depicted in figure 10.

Figure 10 also shows the statistical posi-
tions of nine known coyote-dog hybrids
(some of those discussed by Mengel, 1971;
all in KU), and the positions of 15 wild-
caught individuals (all of those in USNM
and KU that could be utilized) previously
identified by others as suspected coy-dogs.

The specimens of wild northeastern Canis
are statistically proximal to the standard sam-
ple of C. latrans, but demonstrate a pro-
nounced shift toward C. lupus. By eye, most
of these skulls appear coyotelike, though on
the average they are larger than C. latrans
and have relatively broader rostra and frontal
shields. Both visually and statistically, five
individuals stand out from the others and
have multivariate positions approaching the
limits of the sample of C. familiaris. My data
thus support the view of Lawrence and Bos-
sert (1969) that both domestic dog and gray
wolf genes have influenced Canis in the
northeast. Nonetheless, the amount of graph-
ical separation between the main clusters of
northeastern canids and the few that are
scattered in the direction of C. familiaris,
suggests to me that two separate phenomena
are involved. First, a limited amount of hy-
bridization between wild Canis and domestic
dogs has occurred, but there has been no
large-scale shift in characters, among the
wild population as a whole, in the direction
of C. familiaris. Secondly, the predominantly
coyotelike population of northeastern Canis
evidently has sustained the introgression of
genes from the gray wolf. There are not yet
enough data to determine exactly when and
where this process began, but apparently
wolf genes are now incorporated in the north-
eastern coyote.

The relationships among the populations
of Canis in the northeast and in other regions

LUPUS

4
3
2
I

0
-I
-2
-3
-4
-5
-6
-7

-5 -4 -3 -2

LATRANS

-6 -5 -4

-1 r-

-3 -2

-I

I

2 3

Fie. 10. — Multivariate comparison of various in-
dividual specimens to the ranges of variation of the
total series of C. lupus, C. latrans, and C. familiaris
shown in Fig. 8. Black dots, statistical positions of
specimens of wild-caught individuals from New Eng-
land, New York, and Pennsylvania; open circles,
wild-caught animals previously reported as coy-dogs;
H, captive born coy-dogs; triangles, specimens from
southeastern Canada thought to be hybrids between
C. lupus and C. latrans; M, specimens in ROM
thought to be from captive wolf-coyote hybrids.
Males are above, females below.

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

23

TARLE 1

Statistical distance, D2, between populations of Canis
in the northeast and elsewhere.

C. familiaris

A. Males

C. lupus 68.2

C. lupus lycaon 61.7

C. latrans _ 55.8

C. latrans thamnos ... 53.3

Northeast U.S. Canis 48.4

B. Females

C. lupus 51.6

C. lupus hjcaon 46.0

C. latrans 56.6

C. latrans thamnos 55.8

Northeast U.S. Canis 47.4

C. lupus

C. lupus
hjcaon

C. latr

C. latrans
thamnos

4.1

68.1

48.3

63.2

42.2

2.0

49.2

30.2

5.4

3.2

4.6

56.8

39.3

53.6

34.9

2.2

41.3

25.1

3.5

2.1

hitherto discussed, are summarized in table 1.
Data from those specimens considered to
represent hybrids between wild Canis and
C. familiaris were not used in the calculation
of the statistics shown. Northern and western
wolves demonstrate close relationship with
the subspecies C. lupus hjcaon of the north-
east; and there is affinity between northern
and western coyotes, C. latrans thamnos, and
the populations of Canis in the extreme north-
eastern United States. This latter group and
C. lupus hjcaon are separated by less than
half the statistical distance found between
the main groups of northern and western
wolves and coyotes.

I above postponed discussion of six speci-
mens previously identified in collections as
C. lupus lycaon, and one identified as C.
latrans tliamnos (see appendix A, part 6).
Each of these skulls appeared to be of un-
usual size or to have other characters sug-
gesting hybridization between C. lupus and
C. latrans. Multivariate positions are shown
in figure 10.

In addition to the above, I examined two
skulls (ROM 31-9-15-1 and 31-9-15-2) re-
ported to represent the hybrid offspring of a
captive wolf and coyote (Young, 1951:124;

Mengel, 1971:333). The labels of these speci-
mens read in part: "From Ontario stock."
Multivariate positions of each, relative to the
three standard series, are plotted in figure 10
(since ROM 31-9-15-2 lacked both PI, it
would not ordinarily have been included in
a multivariate analysis, but because of its
special interest I used an estimate of alveolar
length of upper toothrow).

If considered together, the positions of
these nine specimens bridge the gap be-
tween the statistical limits of the standard
sample of C. lupus and the positions of the
individual specimens from the northeastern
United States. It is thus reasonable to sup-
pose that hybridization between wolf and
coyote in southeastern Canada has, and prob-
ably still is permitting the flow of genes from
one species to the other. Evidently wolf
genes have spread through much of the coy-
ote population, and resulting wolflike char-
acters have been phenotypically expressed in
Canis of the extreme northeastern United
States. This introgression may have assisted
the coyote population of the region to adapt
to and flourish in an environment far from
the original prairie habitat of C. latrans.

24

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

Systematic Problems in
the Southeast

Background

When Europeans first entered the south-
eastern part of North America, they found
animals that appeared to be closely related to
the wolves of the Old World. For example,
Catesby (1771:xxvi), writing of Florida and
the Carolinas, noted: "The Wolves in Amer-
ica are like those of Europe, in shape and
colour, but are somewhat smaller."

Bartram (1791:199), the first author to
apply a binomial to Canis in the southeastern
region, referred to the black wolves that he
saw in Florida as "lupus niger." Harlan
(1825:82) used the name Canis lycaon for
these same animals, but also for black wolves
inhabiting mountainous areas of North Amer-
ica and Europe. The designation Canis ly-
caon originally had been given to a black
wolf from the vicinity of Quebec by von
Schreber in 1775 (Goldman, 1944:437-440),
and was restricted by Miller (1912b) to the
wolves of eastern Canada and the north-
eastern United States. Richardson (1829:70)
termed the black wolves that he saw on the
banks of the Mackenzie and Saskatchewan
rivers of western Canada as "Canis lupus
occidentalis var. E. Lupus ater." He said,
however, that the same animals occurred
throughout North America, and he included
Bartram's Florida wolves under this name.
It thus seems that early naturalists named
kinds of wolves largely on the basis of color,
and considered the names to apply where-
ever the particular colors were found. Since
it was eventually demonstrated that colora-
tion of wolves in nearly all parts of North
America is highly variable and of minimal
taxonomic value (Young, 1944:59-66; Gold-
man, 1944:401), these early writings are use-
less to an understanding of the problems
presently under review.

Apparently Audubon and Bachman (1851:
126, 240) were the first authors to set definite
bounds to the ranges of named kinds of North

American wolves, and to suggest that in the
southern United States there existed wolves
structurally different from those in other re-
gions. They kept Richardson's designation in
their description of Canis lupus var. Arer, the
"Black American Wolf," but recorded this
kind only from Florida, South Carolina,
North Carolina, Kentucky, southern Indiana,
southern Missouri, Louisiana, and northern
Texas. They also discussed Canis lupus, var.
Rufus, the "Red Texan Wolf," which they
said ranged from northern Arkansas, through
Texas, and into Mexico. In their description
of this animal Audubon and Bachman men-
tioned the long legs, pointed nose, and slen-
der proportions. They noted that although
the reddish shade predominated in Texas,
other colors also were represented there, and
that the wolves of different colors freely in-
terbred. They stated that except for Canis
latrans, the coyote, all of the wolves that they
described were only varieties of one species.

Audubon and Bachman's delineation of
Canis in the southern United States was gen-
erally accepted by biologists. But toward the
end of the nineteenth century finer taxonomic
splitting became more fashionable, and Bangs
(1898) designated the Florida wolf as a full
species, Canis ater. Bailey (1905) question-
abb referred to the large, dark wolf of east
Texas as C. ater, and expressed hope that
specimens could be obtained to confirm its
status. He also recognized Audubon and
Bachman's red wolf as a full species with
the name Canis rufus, and assigned it a range
in southern and central Texas. Bailey dis-
tinguished the larger gray wolf (C. griseus=
C. lupus) of western Texas from C. rufus,
and referred to the latter as "a large coyote
or small wolf."

Miller (1912a) explained that the name
ater was technically unavailable, and he des-
ignated the Florida wolf as C. floridanus.
This name then became generally accepted
for wolves in the forested areas of the south-
eastern United States, while C. rufus con-
tinued to be recognized in central and south-

1979

NOWAK: NORTH AMERICAN QUATERNARY CAMS

25

em Texas. At this time, however, there had
not yet been an attempt to associate C. rufus
and C. floriclanus into one group that demon-
strated characters different from those of all
other North American wolves.

Goldman (1937) combined the wolves of
the south-central and southeastern United
States into a single species, C. rufas, that he
considered distinct from all other North
American wolves. He combined the latter
into the species C. lupus. Goldman said that
C. rufus "exhibits a departure from the true
wolves, and in cranial and dental characters
approaches the coyotes." He listed the names
C. rufus rufus for the Texas subspecies, and
C. r. floridanus for the eastern race, and he
also described C. r. gregoryi, a new sub-
species in the lower Mississippi Valley.
Shortly thereafter, Harper (1942) pointed
out that Bartram's ( 1791 ) term for the Flor-
ida wolf was actually the earliest name to
have been applied in the southeast, and the
specific designation of the wolves in the
region then technically became C. niger. But
the International Commission on Zoological
Nomenclature (1957, opinion 447) rejected
Bartram's technical terms, and Hall (1965)
listed C. rufus as the proper name. There-
fore, the trinomials applied by Goldman
( 19.37 ) are presently considered valid and
are used in subsequent discussion.

Goldman (1944:481) observed that the
two eastern subspecies of C. rufus exhibited
a remarkable approach in size and general
proportions to the eastern gray wolf, C. lupus
hjcaon, but that there existed several specific
cranial differences. On the other hand, he
said that the subspecies C. r. rufus in central
Texas and Oklahoma was so small and in
general characters agreed so closely with C.
latrans, that some specimens were difficult to
distinguish. He suggested the possibility of
hybridization between C. rufus and C. latrans
in some localities in Texas.

Goldman (1944) used the vernacular "red
wolf" for the species C. rufus, presumably on
tlie basis of Audubon and Bachman's descrip-

tion of the "Red Texan Wolf." This term,
however, is not found in any of the early
literature discussing wolves in states east of
Texas. Kellogg (1915:41) said that trappers
in Cherokee County, southeastern Kansas, did
refer to "Red Wolves" in the area. But other-
wise this popular term seems to have been
restricted to parts of Texas until Goldman
introduced its use throughout the range of
C. rufus. This appellation may be unfortu-
nate, because although the rufous element in
the fur sometimes stand? out, the "red barn-
roof paint" color mentioned by Young (1946:
36) seldom shows up. Early records indicate
that a dark-colored or entirely black phase
was locally common in the eastern forests.
But most available specimens of "red wolves"
actually exhibit a typical wild canid color
pattern, consisting of an agouti gray or
brown, interspersed with black hairs, espe-
cially on the back, and with the muzzle, ears,
and outer surfaces of the limbs tending to-
ward a tawny color.

Lawrence and Bossert (1967) thought
that separation of C. rufus as a distinct spe-
cies rested too heavily on the small red wolves
of central Texas where hybridization with the
coyote may have been a factor. They said
that if the study of wolves in the south had
been based on adequate series of specimens
from Florida, separation of C. rufus from C.
lupus would have been highly unlikely. In a
multivariate analysis they compared all avail-
able skulls of wolves collected before 1920 in
Louisiana, Alabama, and Florida, including
one Florida specimen assigned by Goldman
to C. lupus hjcaon (according to Lawrence
and Bossert, 1975:81, a total of 12 specimens
were in this sample), with series of 20 adult
skulls each of C. familiaris, C. lupus (various
North American subspecies), C. lupus hjcaon,
and C. latrans. They found C. lupus hjcaon
and C. rufus (the early southeastern ma-
terial) both to overlap with C. lupus, but to
be distinct from each other. All three of
these populations formed a cluster distinct
from C. latrans, with C. rufus being the

26

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

farthest removed. Lawrence and Bossert in-
terpreted this analysis as demonstrating that
early populations described as C. rufus, "east
of the range of Canis latrans, are a local form
of Canis lupus, not a distinct species of wolf."

Lawrence and Bossert ( 1975 ) repeated
the above interpretation. They again ob-
served that their 12 early specimens from
Louisiana and eastward, which they now des-
ignated "floridanus," and their series of C.
lupus lycaon, were both less coyotelike than
their general sample of C. lupus. Lawrence
and Bossert ( 1975 ) also evaluated a series
of 30 specimens (designated as "gregoryi")
taken in the 1920's in Arkansas, south of the
Arkansas River. This group was found to be
nearer in D2 distance to the general sample
of C. lupus and to C. lupus lycaon, than to
"floridanus."

The suggestions by Lawrence and Bossert
(1967, 1975) — (1) that pre-1920 eastern C.
rufus is not more than subspecifically distinct
from C. lupus, (2) that C. rufus and C. lupus
lycaon are less coyotelike than C. lupus in
general, and ( 3 ) that Arkansas red wolves are
closer to C. luptis than to Louisiana red
wolves — are not in agreement with my own
findings. These suggestions, at least in part,
also seem not to correspond well with results
obtained by some other workers.

Paradiso ( 1968 ) argued that the relatively
small samples used by Lawrence and Bossert
( 1967 ) did not adequately represent the vari-
ability shown by Canis, and thus that con-
specificity of the red and gray wolves had
not been demonstrated. He also pointed out
several cranial and dental characters in which
C. rufus resembled C. latrans more than it did
C. lupus. Paradiso and Nowak (1972a) com-
pared data on 213 skulls of C. rufus, 214 of
C. lupus, and 336 of C. latrans, and concluded
that the red wolf was a distinct species. In
size and proportion C. rufus fell between C.
latrans and C. lupus, but was nearer to the
latter. Development of certain dental char-
acters in C. rufus suggested affinity to the
coyote. Atkins and Dillon (1971), on the

basis of a morphological study of the cerebel-
lum, considered the red and gray wolves to
be in the same group, distinct from other
Canis, but that C. rufus was a valid species.
Shaw (1975) evaluated data on the behavior,
ecology, vocalizations, allelic frequency, and
morphology of a living population of C. rufus
gregoryi in southeastern Texas, and con-
cluded that the population represented a spe-
cies distinct from both C. lupus and C. la-
trans.

Three recent studies, aimed primarily at
identifying newly collected specimens of wild
Canis in the south-central states, and each
employing a different method of multivariate
analysis, also have tended to uphold the spe-
cific status of C. rufus. Gipson, Sealander,
and Dunn (1974), who used a single meas-
urement and five separate ratios as variables,
compared as groups the skulls of 40 C. la-
trans, 34 C. latrans x C. familiaris, 31 C. fa-
miliaris, 37 C. lupus lycaon, 40 C. lupus from
the Great Plains region, and 40 C. rufus taken
in Arkansas prior to 1925. They found the
red wolf sample to be statistically intermedi-
ate to the coyote and gray wolf samples, and
reported their results to support recognition
of the red wolf as a distinct species. Freeman
( 1976 ) , who used 15 raw measurements as
variables, compared as groups the skulls of
40 C. latrans, 12 C. latrans x C. familiaris, 24
C. familiaris, 43 C. lupus from the Great
Plains region, and 44 C. rufus taken in Arkan-
sas prior to 1925. The graphical results of his
analysis indicate that the red wolf, again, is
statistically between the coyote and gray wolf.
Unlike Gipson, Sealander, and Dunn (1974),
however, who found some overlap between
samples of C. lupus and C. rufus. Freeman
obtained complete statistical separation of the
two species. Elder and Hayden (1977),
whose variables were proportions obtained by
dividing 14 measurements by greatest length
of skull ( in accordance with Lawrence and
Bossert, 1967), compared as groups the skulls
of 29 C. latrans, 30 C. familiaris, and IS C.
lupus from widely scattered regions. Then,

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

27

27 skulls of C. rufus taken in Missouri in the
1920's and early 1930's were compared as
individuals to these three groups. None of
the red wolves fell within the limits of the
groups, but they were distributed statistically
between the coyote and gray wolf samples.

Mech (1970:25, 285, 351) suggested that
the differing views on the status of C. rufus
could be reconciled by recognition of the red
wolf as a hybrid population that could be
properly known as C. lupus x C. latrans. Re-
gardless of the taxonomic disputes, it is now
generally agreed that the original populations
of Canis inhabiting the southeastern quarter
of the United States have all but disappeared,
and have been replaced in large part by
another kind of Canis (McCarley, 1962; Para-
diso and Nowak, 1972a; Pimlott and Joslin,
1968).

Examination of Earliest Available
Eastern Material

In my own study, I wished to consider all
available specimens from south of Lakes
Michigan and Erie, and east of the Mississippi
River. Goldman ( 1944 ) examined only four
skulls (including one damaged and one im-
mature) and one mandibular ramus, assigned
to C. lupus Tycoon, from this region. He also
had only two skulls of C. rufus gregonji (both
subadults), from Indiana and Illinois, and
two of C. r. floridanus (one damaged), from
Alabama and Florida. Some other workers
have considered this small collection to be
insufficient for an assessment of the relation-
ships among the original populations in the
region, particularly for settling the question
of whether C. rufus and C. lupus intergraded
in the Ohio Valley and along the central
Atlantic coast. Unfortunately, there are
scarcely any additional skulls available, and
there doubtless never will be, because the
native wolves of the region were exterminated
long ago. Excavation of fossil and aboriginal
sites offers some hope, but most material from
such sources that I have examined is frag-

LUPUS

LATRANS

-6 -5 -4

- r~
0

2 3

Fie. 11. — Multivariate comparison of individual
specimens of C. rufus to the ranges of variation of
the total series of C. lupus, C. latrans, and C. famil-
iaris shown in Fig. 8. L, pre-1920 specimens from
Louisiana; A, specimen taken in 1917 in Alabama;
C, specimen found in Alabama cave deposit; F,
Florida specimen; K, specimen from Garvin County,
Oklahoma; P, Pennsylvania specimen; T, southeastern
Texas specimen. Males are above, females below.
The Florida, Oklahoma, and Pennsylvania specimens
are compared both with male and female series.

28

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

mentary and of limited value. This material
is discussed later in the paper.

Lawrence and Bossert ( 1967) limited their
sample of C. rafu.s to specimens collected
before 1920, believing that such material was
the least likely to show influence from C.
latrans. As an initial step, this paper also
considers material taken prior to 1920. But
from this period, and from east of the Missis-
sippi River, there are available only four
skulls that arc adult and unbroken, and hence
suitable for inclusion in multivariate analysis.
Two of these specimens are unknown as to
sex, but since the smallest of the four is a
known male, I compared each of them indi-
vidually with the previously compiled sam-
ples of 50 C. familiaris, 166 male C. latrans,
and 294 male C. lupus ( including 61 C. lupus
hjcaon ) . Their relative multivariate positions
are plotted in figure 11 (the two skulls of
unknown sex are also compared to appro-
priate female samples). The positions of the
four are beyond the range of variation of any
of the known samples, but are distributed
between C. lupus and C. latrans. Measure-

ments of these skulls are listed in appendix
B (part 4), and other details are given in
table 2.

Although these four specimens are indica-
tive of the former presence in the eastern
United States of a kind of wolf different from
any subspecies of C. lupus, their value is lim-
ited by missing data. A better collection of
pre-1920 material was taken in Louisiana,
and was identified by Goldman as C. rufus
gregoryi. The Louisiana material suitable for
inclusion in my analysis consists of 7(2)
males and 2(1) females (parentheses contain
numbers of specimens in the series for which
sex had not been recorded, but which were
judged to belong to the particular sex in-
dicated; see p. 5). A single skull of a
female C. rufus gregoryi, taken in extreme
southeastern Texas in 1906, also is available.

As depicted in figure 11, the nine Louisi-
ana specimens, and the one from southeastern
Texas, have multivariate positions interme-
diate to the graphical limits of C. lupus and
C. latrans. Seven of these skulls have indi-
vidual D- distances farther from C. lupus

TABLE 2
Data on early specimens of C. rufus.

Collection D" from D= from

and number Sex Locality Date C. lupus C. latrans

ANSP 2261 male? Pennsylvania _ pre-1859 27.2 23.0

MCZ 11179 male? Florida" 1854? 24.5 37.8

USNM 223936 male 12 mi. S Cherokee, Colbert County, Alabama 1917 14.5 30.6

USNM 348063 male Fern Cave, Jackson County, Alabama00 .... 43.8 26.4

USNM 132229 male Mer Rouge, Morehouse Parish, Louisiana 1904 25.6 24.9

USNM 136834 male 23 mi. SW Tallulah, Madison Parish, Louisiana __ 1905 18.1 28.1

USNM 137125 male 20 mi. SW Vidalia, Concordia Parish, Louisiana .... 1905 37.9 27.2

MCZ 9114 male Mer Rouge, Morehouse Parish, Louisiana 1898 23.2 22.8

USNM 136731 male 18 mi. SW Tallulah, Madison Parish, Louisiana .... 1905 46.1 52.3

USNM 133687 male? 15 mi. NW Tallulah, Madison Parish, Louisiana .... 1904 38.4 38.1

USNM 133688 male? 10 mi. SW Floyd, West Carroll Parish, Louisiana 1904 29.7 37.5

USNM 136105 female 15 mi. SW Tallulah, Madison Parish, Louisiana .... 1904 22.4 19.4

USNM 234227 female? 12 mi. N Avery Island, Iberia Parish, Louisiana .... 1919 31.7 19.1

USNM 147701 female Kountze, Hardin County, Texas 1906 38.0 25.2

"Apparently with regard to this specimen, Barbour (1944:142) wrote: "we have in the Agassiz Museum in
Cambridge a skull of a fine old adult animal which was said to have been killed in 1S54 in the region where
the city of Miami now stands."

" This specimen was found in a cave; its state of preservation suggested that it lived in Recent times (Para-
diso and Nowak, 1973).

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

29

-5 -4 -3 -2

Fig. 12. — Multivariate positions (black dots) of
individual pre-1920 males of C. rufus from Louisi-
ana, relative to series of males of C. lupus, C. latrans,
and C. familiaris. The variables used in this par-
ticular analysis were derived by dividing the meas-
urements by greatest length of skull.

than from C. latrans (table 2). A separate
analysis of males only, in which the employed
variables were fractions derived by dividing
each of the other measurements by greatest
length (in accordance with Lawrence and
Bossert, 1967), produced much the same pic-
ture (Fig. 12).

Measurements of the pre-1920 Louisiana
material are listed in appendix B (part 4),
and other details are given in table 2. In
figure 13, specimens of males are compared
with the series of C. lupus and C. latrans in
a ratio diagram. The Louisiana skulls ap-
proach those of C. lupus in over-all length,
but are comparatively small in most other
dimensions, a notable exception being their
large M2. In general the proportions of C.
rufus gregoryi seem intermediate to those of
C. lupus and C. latrans, but the frontal shield
in the Louisiana material is even narrower,
relatively, than in the coyote, and the post-
orbital constriction is narrower, relatively,
than in the gray wolf. The results of these
analyses indicate that C. rufus morphologi-

cally resembles C. lupus lycaon more than it
does any other subspecies of gray wolf.

We are still left with what Lawrence and
Bossert (1967) called "the biologically diffi-
cult problem of reconciling the existence of
two similarly-sized forms of wolf in one con-
tinuous habitat." Actually, because of the
limited number of specimens, there may
never be indisputable proof that the red and
gray wolves did not undergo intergradation
in the eastern forests. The existence, how-
ever, of 14 complete skulls collected before
1920 in this region, that do not overlap in

-.16

.12 -.08 -.04 0

LATRANS

RUFUS LUPUS

70

80

.90

1.00

Fig. 13. — Ratio diagram comparing means of
total series of C. lupus, series of C. latrans from
northern and western North America, and two series
of C. rufus (males only are shown for all series). The
solid line under rufus represents the mean values of
pre-1920 Louisiana specimens; the dashed line rep-
resents the mean values of specimens taken from
1919 to 1929 in the south-central states. The verti-
cally arranged numbers correspond to the measure-
ments so numbered in appendix B. A log difference
scale is provided above, and a ratio scale below the
diagram.

30

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

statistical position with series of 482 speci-
mens of C. lupus, including 103 tycoon, sug-
gests to me that the population represented
by these skulls is more than subspecifically
different from the gray wolf. Furthermore,
the slender proportions of the red wolf indi-
cate that its prey averaged smaller than that
of the gray wolf, and that its ecological niche
may have approached that of the coyote
which did not exist in the eastern forests
when white settlers first arrived. Possibly,
when C. lupus entered the northeastern for-
ests it underwent a degree of parallel evo-
lution with C. rufus, just to the south. The
subspecies C. lupus lycaon may even have
been in the process of replacing C. rufus,
when the white man interfered. And it is
reasonable to think, considering the ease with
which interbreeding occurs in Canis, that
hybridization between C. lupus hjcaon and
C. rufus did occur, and that a zone of intro-
gression may have developed that tended to
modify one or both populations. The advent
of the white man would probably have stim-
ulated the spread of such hybridization, just
as it seems to have encouraged interbreeding
between C. rufus and C. latrans in Texas
(Paradiso and Nowak, 1972a). Be that as it
may, available early specimens from the
southeastern United States can all be sepa-
rated from known series of C. lupus, and
seem to represent a different species, C. rufus.

The exact distribution of C. lupus and C.
rufus, and the extent to which their ranges
overlapped in eastern North America, will
probably never be known. But if one Florida
specimen, called C. lupus lycaon by Goldman
(see table 2), is in fact C. rufus as I have
tried to show, then there is no longer any
confirmation of the original presence of the
gray wolf in Florida. And if correct, identifi-
cation of a Pennsylvania specimen as C. rufus
rather than lycaon means that the range of
the red wolf once extended farther to the
northeast than was formerly thought.

As explained by Paradiso and Nowak
(1972a:7-8), Goldman was also incorrect in

his reasons for stating that the range of the
red wolf in the Mississippi Valley once ex-
tended as far north as Warsaw, Hancock
County, Illinois; and Wabash County, Indi-
ana. The specimen supposedly from the for-
mer locality had been in possession of an
animal dealer, and there is no evidence that
it represents a native wolf of the area. The
other specimen was actually collected in the
Wabash River area of southwestern Indiana,
not in Wabash County father north. Both
skulls appear to be referable to C. rufus, but
they are subadults and hence not suitable for
use in multivariate comparisons made on the
basis of adults.

Examination of Material Collected from

1919 to 1929 in the South-central

United States

In the course of Federal predator control
work in the south-central United States, from
1919 to 1929, a large number of specimens of
Canis were taken. Most of these were re-
ferred by Goldman (1944) to C. rufus gre-
goryi, the "Mississippi Valley Red Wolf."
Lawrence and Bossert (1967) did not use
this material in their attempt to define the
red wolf, because of the possibility that some
of the specimens represented hybridization
between red wolf and coyote. Paradiso and
Nowak ( 1972a ) , however, noted that al-
though a few of these specimens appeared to
be hybrids, the remainder demonstrated the
continued survival of the species C. rufus.

Since many of the specimens assigned to
C. r. grcgoryi may indeed represent the red
wolf in unmodified form, they should be con-
sidered in any effort to ascertain the relation-
ship between C. rufus and C. lupus. But
since some of the specimens may actually be
the result of hybridization between C. rufus
and C. latrans, their inclusion with unmodi-
fied C. rufus in statistical compilation could
result in a misunderstanding of the situation.
There is no completely objective method of
separating red wolves from red wolf-coyote
hybrids, but earlier work (McCarley, 1962;

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

31

Fig. 14. — Map showing localities of specimens taken prior to 1930 in the south-central United States.
Triangles, C. rufus; black dots, specimens originally identified as C. latrans; squares, specimens originally iden-
tified as C. rujus, but considered in this paper to represent hybridization between red wolf and coyote.
See figures 18-23 for more details. The dotted line shows the western limits of oak forest in central Texas.
The dashed line divides localities of specimens previously identified as C. rujus gregoryi (east) from those
previously identified as C. rufus rujus (west). Note: because of the scale of the map it was not possible
to plot all localities in crowded areas.

Paradiso, 1968) indicated that skulls of C. r.
gregoryi almost invariably have a greater
maximum length than those of C. latrans. In
my preliminary examination of the series
identified as C. rufus gregoryi, I noticed 15
adult skulls from Arkansas and eastern Okla-
homa that seemed comparatively short (less
than 215 millimeters in greatest length for
males, and less than 210 millimeters for fe-
males), and which were within or near the
size range of my standard series of western
coyotes. These 15 specimens were considered
the most likely to represent hybridization,

and hence were not directly compared to the
standard series of C. lupus. They are dealt
with later in this account.

Of the remaining specimens collected
from 1919 to 1929 that Goldman assigned to
gregoryi, those that could be subjected to
multivariate analysis were 63(1) males and
52(1) females (see Fig. 14; and appendix A,
part 10). Of these 115 skulls, 64 were from
Arkansas, 19 were from southern Missouri,
29 were from southeastern Oklahoma, and
three were taken in northern or western
Louisiana in 1928 (a specimen collected in

32

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

" LATRANS

LUPUS

4
3
2

I

0
-I
-2
-3
-4
-5
-6
-7

LATRANS

-6 -5-4-3-2-1 0 I 2 3 4

Fig. 15. — Multivariate positions (black dots) of
individual specimens of C. rufus taken from 1919 to
1929 in the south-central United States, relative to
the ranges of variation of the total series of C. lupus,
C. latrans, and C. familiaris shown in Fig. 8. Males
above, females below.

extreme southern Louisiana in 1919 was in-
cluded with the early eastern material cov-
ered above; see pp. 2S-30).

TABLE 3

Statistical distance, D2, between C. lupus,
C. latrans, and C. rufus.

C. lupus

C

latrans

A. Males

C. latrans

C. rufus .

71.9

18.6

22.2

B. Females

C. latrans

C. rufus

63.4

16.9

18.8

These 115 skulls were compared individ-
ually, by multivariate analysis, to the standard
series of C. lupus, C. latrans, and C. familiaris.
Their distribution (Fig. 15), like that of the
earlier material from the southeast (Fig. 11),
is intermediate to the graphical limits of C.
latrans and C. lupus. There is no overlap
with the coyote sample, but three male and
three female specimens, all from the Ozark
region of Arkansas and Missouri, fall within
the range of variation of the gray wolf. There-
fore, on the basis of a multivariate analysis
involving 15 cranial measurements, complete
separation of C. lupus and C. rufus is not
possible. The amount of overlap between the
two, however, is much less than that found
among the various subspecies of C. lupus.

Measurements of these 115 gregoryi are
listed in appendix B (part 5). Means of males
are compared with those of other series in a
ratio diagram (Fig. 13). Size and proportion
in both the early red wolf and the later series
of gregoryi are seen to match closely, and to
differ from C. lupus and C. latrans. There is
no indication that the later and more nor-
therly concentrated series of gregoryi is sub-
stantially different or any more coyotelike
than the early Louisiana material. These two
series thus were combined into a single group
that was considered as a standard sample of
C. rufus. The D- statistics comparing the
three total standard samples of C. lupus, C.
latrans, and C. rufus, were calculated and
are shown in table 3. The intermediate val-

1979

NOWAK: NORTH AMERICAN QUATERNARY CAMS

33

5-
4-
3-

2-

-2-
-3-
-4-

-5-

-6
-7

-8
-9

-4 -3 -2 -I

— i—
0

-i 1 1 1 r-

2 3 4 5 6

Fig. 16. — Multivariate comparison of 886 individ-
ual specimens of Canis from the south-central United
States, to the ranges of variation of the total series
of pre-1930 C. rufus gregoryi, the total series of C.
latrans from northern and western North America,
and C. familiaris. The dashed lines show the margins

ues of the red wolf are almost as near to
those of the coyote as to those of the gray
wolf.

Influence of the Domestic Dog on Wild Canis
of the Southeast

Since it has been concluded that C. rufus
is specifically distinct from C. lupus, and
since there is no evidence that C. lupus
existed in the southeastern United States in
the twentieth century, the gray wolf need not
be considered further in discussion of the
systematic problems of this region. Hence-
forth, this section of the paper deals primarily
with the relationships between C. rufus and
C. latrans, and with their comparative status

-3

■I

- r—

0

I

— 1 1 1-

2 3 4

of the statistical distribution of 879 of the individual
specimens. The seven exceptions are from Arkansas
( A ) ; Louisiana ( L ) ; Missouri ( M ) ; Lavaca County,
Texas (T); and Van Zandt County, Texas (Z). See
appendix A (part 11) for additional details. Males
are right, females left.

in the southeast. Before proceeding, however,
the question of interbreeding between C.
familiaris and wild Canis should be covered.
In order to evaluate the influence of the
domestic dog, the three previously estab-
lished samples of C. familiaris, C. latrans, and
C. rufus were compared as groups with 886
other individual skulls collected in the south-
eastern and south-central states. All of these
specimens had been previously identified in
collections as wild Canis, and they include
the members of nearly all of the samples
referred to in the following account of the
southeastern situation. The graphical results
of the analyses are shown in figure 16. There
is no statistical blending of wild southeastern

34

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

Canis and C. familiaris, and no suggestion of
massive introgression from the domestic dog
into any of the wild populations. Only seven
specimens have statistical positions or D2
values indicating hybridization involving C.
familiaris (see appendix A, part 11).

Pre-1930 Relationship of Red Wolf and

Coyote in Arkansas, Missouri, and

Oklahoma

With the gray wolf and domestic dog
eliminated from the picture, we now are
dealing exclusively with the relationship of
the red wolf and coyote. As an initial step,
standard series of both these species were
compared with each other by multivariate
analysis. For the red wolf, these series con-
sisted of the previously used 70(3) males
and 55(2) females collected before 1930 and
identified as C. rufus gregoryi. For the coy-
ote, the previously established standard sam-
ple of western material was used, except that
one male specimen was withheld for consid-
eration at a later point. The comparative
series of C. latrans thus consisted of 165(2)
males and 111(4) females. The results of
these analyses, demonstrating clear separa-
tion between the two species, are depicted in
figure 17.

All remaining group comparisons involv-
ing problems in the southeast are based on
the statistical distribution of C. rufus and C.
latrans shown in figure 17. All other speci-
mens, mostly those collected farther to the
west or later in time than the standard sam-
ple of C. rufus, are evaluated on the basis of
their relative distance from this sample and
from the standard sample of C. latrans. There
is no objective manner of setting definite
limits to the ranges of variation shown by
red wolves, coyotes, and hybrids between the
two. It would therefore be meaningless to
attempt to assign each and every specimen
to one of these three categories. Therefore I
decided to examine each group of specimens,
collected in a given area and over a certain

RUFUS

3

2

I

0
-I
-2
-3 -

-3

-I

i
0

I

3 4 5 6

Fie. 17. — Multivariate comparison of the series of
pre-1930 C. rufus gregoryi and the series of C. latrans
from northern and western North America. Only
the margins of the range of variation of each species
are shown. Males are above, females below; C. rufus
to the right, C. latrans to the left. All subsequent
graphs of multivariate analyses that involve the
southeastern or south-central United States are based
on the illustrations shown here. The numbers along
the vertical and horizontal axes are canonical coordi-
nates. The position of a single female C. rufus is not
included within the range of variation shown for
that species. The coordinates of this specimen fall
beyond the coverage of the computer plot, though
not in the direction of the range of variation of C.
latrans. The specimen is large and might possibly
represent a mislabeled male. The position of this
specimen is among those plotted for females in Fig.
15.

period, and to try to interpret relative affinity
on the basis of the total group position.

We have already seen that a large group
of specimens identified as C. rufus gregoryi
may be clearly distinguished from a large
series of western coyotes. The great majority
of these specimens referred to gregoryi were
collected between 1919 and 1929 in the
Ozark-Ouachita uplands of Arkansas, south-

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

35

em Missouri, and eastern Oklahoma. Coyotes
also were found along the western and north-
ern edges of this region, but the original
limits of their range are unknown. According
to Audubon and Bachman (1851:152), C.
latrans was "well known throughout the west-
ern part of the States of Arkansas and Mis-
souri."

The coyotes of this region are called C.
latrans frustror. This name was given by
Woodhouse ( 1851 ) to a canid from the
Cimarron River, about 100 miles west of Fort
Gibson, Oklahoma. For a while, there was
uncertainty regarding whether frustror ap-
plied to the coyote or to the larger wolf
group. Merriam (1897) considered the holo-
type to be a coyote, but Bailey (1905:175)
subsequently noted: "A series of topotypes
of frustror secured since at Red Fork, Ind. T.,
shows it to be a widely different species,
more nearly related to Canis rufus." Bailey
seems not to have considered the possibility
that both red wolf and coyote could have oc-
curred together in the same vicinity. Jackson
(1951:271) considered frustror a subspecies
of C. latrans with a range from eastern Kan-
sas and Missouri south to the Texas Gulf
coast.

Also, according to Jackson (1951:274):
"A very few specimens from the Ozark re-
gion of Arkansas and Missouri superficially
hint that there may be possible hybridization,
but probably not intergradation, with the
Mississippi Valley red wolf, Canisniger gre-
goryi Goldman, in that region." And Law-
rence and Bossert ( 1967 ) found a series of
eight skulls, collected in 1921 at Fallsville,
Newton County, Arkansas, and identified as
C. rufus gregonji, to "span the whole range
of variation from coyote to wolf."

As stated above, I did not include 15
skulls, identified as gregonji, in my standard
sample of C. rufus, because they seemed un-
usually small in greatest length. These 3(1)
males and 12(1) females were all obtained
in Arkansas and eastern Oklahoma between
1919 and 1929 (see appendix A, part 12).

They were compared individually to the
standard red wolf and coyote series, and their
relative statistical positions are plotted in
figure IS. Twelve of the specimens cluster
within or near the range of variation of C.
rufus. It is impossible to say whether these
specimens are small gregonji or represent ge-
netic influence from C. latrans. But the fact
that their positions are concentrated close to
C. rufus, and do not form an evenly distrib-
uted bridge between red wolf and coyote,
suggests the former possibility. Although
these 12 skulls are short, compared to most
C. rufus, multivariate evaluation of all meas-
urements indicates affinity to this species. Of
the other three specimens, one taken at Falls-
ville, Newton County, Arkansas, occupies

I- 3

LATRANS

RUFUS

□

i

~\ 1 — r-

2 3 4

-i r-

5 6

Fig. 18. — Multivariate positions of certain indi-
vidual specimens relative to the ranges of variation
of the series of C. rufus and C. latrans shown in Fig.
17. Black dots, C. latrans frustror taken prior to
1930 in Missouri; triangles and squares, specimens
originally identified as C. rufus gregoryi that have
unusually short greatest lengths. The squares indi-
cate specimens that probably represent C. latrans or
red wolf-coyote hybrids.

36

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

such an intermediate position (D2 from coy-
ote 23.4, from red wolf 24.7) that it almost
certainly represents a hybrid. The two re-
maining specimens appear to be coyotes.

All of the questionable specimens were
obtained in Arkansas and eastern Oklahoma.
In southeastern Texas, Louisiana, and states
farther east, the only specimens of wild Cards
taken prior to 1930 are clearly referable to
C. rufus. In Missouri, however, in addition
to the 19 specimens that were identified as
C. rufus gregoryi and used in my standard
sample of red wolves, nine skulls were col-
lected from 1923 to 1925 and originally iden-
tified as C. latrans frustror (see appendix A,
part 13 ) . They were taken in the same period
and in approximately the same area of the
southeastern part of the state (see Fig. 14)
as the series of gregoryi. The multivariate
positions of these specimens all fall within
the range of variation of the standard coyote
sample (Fig. 18). Therefore, as first pointed
out by Paradiso and Nowak (1972a), there
is direct evidence that C. rufus and C. latrans
occurred sympatrically in this area, without
intergradation or hybridization.

Until now, this section of the paper has
dealt only with red wolves identified as C. r.
gregoryi, and not with C. r. rufus. Most of
the pre- 1930 specimens that Goldman as-
signed to the latter named subspecies were
collected in Texas, but a few were taken in
Oklahoma and northwestern Arkansas. In
addition, Hall and Kelson (1952:340-341)
assigned two specimens taken in 1923 and
1924 at Reeds Spring, Stone County, south-
western Missouri, to C. r. rufus, although
Goldman had listed them as C. r. gregoryi.
McCarley (1962) suggested that the taxon
C. r. rufus might actually represent the result
of hybridization between C. r. gregoryi and
C. latrans, but Paradiso and Nowak (1972a)
continued to recognize its validity.

Since specimens assigned to C. r. rufus
are on the whole smaller and more narrowly
proportioned, and hence more coyotelike than
most specimens of gregoryi, there was no

need herein to depict a direct comparison of
C. r. rufus with the standard sample of C.
lupus. But one skull (USNM 8098), obtained
sometime in the nineteenth century at Chero-
kee Town, in what is now Garvin County,
central Oklahoma, within the designated
range of C. r. rufus, is of special interest. It
was placed in the National Museum's collec-
tion of C. lupus nuhilus, and was not in-
cluded by Goldman (1944) in his list of
specimens examined. Nevertheless he did
handwrite "C. rufus" on the specimen tag.
My comparison of this skull with standard
series of C. lupus, C. latrans, and C. familiaris
shows that its position is close to that of other
old southeastern material identified as C.
rufus (Fig. 11).

s

LATRANS

RUFUS

□

i

3 4

Fig. 19. — Multivariate positions of certain indi-
vidual specimens relative to the ranges of variation of
the series of C. rufus and C. latrans shown in Fig. 17.
Squares and triangles, specimens from eastern Okla-
homa and Arkansas originally identified as C. rufus
rufus; S, specimens from Reeds Spring, Stone County,
Missouri; K, specimen from Garvin County, Okla-
homa; black dots, C. latrans from western and central
Oklahoma; E, C. latrans from eastern Oklahoma.

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

37

Figure 19 depicts the multivariate posi-
tions of eight males identified as C. r. rufus
(including USNM 8098). Only males of ru-
fus were available from Arkansas, Oklahoma,
and Missouri (see appendix A, part 14). Six
fall well within the limits of C. rufus, whereas
two of the Oklahoma skulls have intermediate
positions (D- from C. rufus 28.S and 31.4,
from C. latrans 26.8 and 31.3). Although
available data are thus inconclusive regarding
the status of C. r. rufus in this region, some
of the animals obtained here are indistin-
guishable from red wolves taken farther east.
A number of skulls identified as C. latrans
also were taken prior to 1930 in Oklahoma,
mostly in the western and central parts of the
state, and in an adjoining county of Texas
(see Fig. 14; appendix A, part 15). Figure
19 shows the relative statistical positions of
15(7) males and 12(6) females to fall mainly
within the limits of the standard sample of
western C. latrans. Two females from Creek
and Tulsa counties, northeastern Oklahoma,
are also included within these limits. Inter-
estingly, of a series of three specimens taken
at Red Fork, Tulsa County, Oklahoma, in
1904-1905, one is statistically identical to the
coyote, one is identical to the red wolf, and
one is intermediate in characters.

Although there is evidence of early hy-
bridization between C. rufus and C. latrans
at certain localities in Arkansas and eastern
Oklahoma, genetic exchange appears to have
remained very limited before 1930. Nearly
all specimens taken until then in the lower
Mississippi Valley, and identified as C. rufus,
are statistically separate from western coyotes
and from coyotes taken in the south-central
United States. Hence the sum of available
evidence indicates that the species C. rufus.
in essentially unmodified form, survived
through the 1920's in this region. Rut by that
period the species was under heavy pressure
from man and already was making its last
stand in the Ozark-Ouachita uplands. Refore
proceeding with a discussion of the red wolf's
decline, however, it is necessary first to de-

scribe the complex pre-1930 situation in
Texas.

History of Texas Canis

As we have seen, Audubon and Rachman
(1851) wrote that the "Red Texan Wolf"
occurred from northern Arkansas through
Texas and into Mexico, although the only
specific place they mentioned was 15 miles
west of Austin, where Goldman (1944:488)
fixed the type locality for C. r. rufus. Other
early naturalists reported on the presence of
wolves in this same general area, and, like
Audubon and Bachman, observed consider-
able local variation in color. For example,
Roemer (1849:80), referring to an 1846 visit
to a plantation in what is now Colorado
County, Texas, wrote that "the owner of the
farm offered several wolf skins for sale. He
had taken them from wolves recently caught
in steel traps. The pelts were of various col-
ors, one black, the other yellow, and still
another greyish brown. The farmer informed
us that such variation in color was quite com-
mon among the larger wolves. They were
very plentiful in the forest surrounding his
house and a number of hogs had been killed
by them."

References are few as to how far east in
Texas the coyote originally occurred, but
there is no evidence of its former presence
beyond the prairies. One early source
(Fisher, 1841:33) discussing Rrazoria County
in 1840, stated: "The large black wolf
abounds in the country; but the small prairie
wolf of the western states, I think, is seldom,
if ever, found so far south."

Bailey's detailed biological survey of Texas
(1905:171-177) included the following refer-
ences to Canis in the state.

C. lupus. — Still common over most of the plains
and mountain country of western Texas, mainly west
of the one hundredth meridian.

C. latrans. — More or less common over at least
middle and southern Texas and apparently eastward
on strips of prairie as far as Gainesville [Cooke
County] and Richmond [Fort Bend County]. There
are vague reports of a small wolf occurring farther
east on the coast prairie even to the border of Louisi-

38

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

ana, but specimens are needed before these reports
can be associated with definite species. East of the
semiarid mesquite region coyotes are rare and prob-
ably mere stragglers. True to their name of prairie
wolf, they do not enter the timbered country to any
extent, although at home in the scrub oak, juniper,
mesquite, and chaparral, as well as over the open
prairies of the southern part of the state.

C. rufus. — A definite range can be assigned the
species, covering the whole of southern Texas north
to the mouth of the Pecos and the mouth of the
Colorado, and still farther north along the strip of
mesquite country east of the plains, approximately
covering the semiarid part of the Lower Sonoran
zone. As yet there are no specimens to show whether
these wolves extend into the more arid region west
of the Pecos. While apparently nowhere overlapping
the range of the larger, lighter-colored 'lobo' or
'loafer' [C. lupus] of the plains, they take its place
to the south and east as soon as the plains break
down and the scrub oak and mesquite country begins,
but their whole range is shared with the coyote. The
ranchmen invariably distinguish between them and
coyotes, and with good reason, for the wolves kill
young cattle, goats, and colts with as much regularity
as the coyotes kill sheep. While paying a bounty of
SI or S2 for coyotes, the ranchmen usually pay S10
or $20 for red wolves.

Although Bailey considered C. rufus to
occupy all of southern Texas, and to share its
entire range with C. Jatrans, some modifica-
tion of this delineation is suggested by the
Bureau of Biological Survey field reports (on
file, National Fish and Wildlife Laboratories,
U.S. National Museum of Natural History)
upon which his published work was based.
These reports (made by Bailey, J. H. Gaut.
A. H. Howell, W. Lloyd, and H. C. Ober-
holser from 1891 to 1905) were prepared at
a time when little was known about the sys-
tematics of Canis in southern Texas. They
refer to C. rufus as either a large species of
coyote or a wolf. But more important than
the names applied is the fact that the pres-
ence of both this larger species of Canis and
a distinct smaller species (C. latrans) was
generally recognized. The reports suggest
that the larger species, or the animal now
referred to as the red wolf, was originally
common only along the Texas coast above
Nueces Bay, and that C. latrans was rare or
entirely absent from this area. Farther in-
land, in the more arid parts of southern
Texas, the red wolf progressively diminished

in numbers, whereas the coyote became
abundant. The reports do not provide any
insight on the original situation farther north
in the Edwards Plateau area.

Examination of Pre-1930 Material
from South Texas

The main problem presented by the above
information is the relationship between the
red wolf and coyote where their ranges met
along the Texas coast, and farther inland
toward the Rio Grande Valley. Unfortu-
nately, the number and distribution of avail-
able pre-1930 specimens are not sufficient for
thorough analysis of the situation. Although
a useful series of coyotes was collected in
south Texas, few skulls were saved from the
critical areas immediately to the north of
Nueces Bay and inland toward San Antonio.

A group of eight skulls, previously identi-
fied as C. rufus rufus, was taken in 1900 and

3
2
I

0
-I
-2
-3

3

LATRANS

RUFUS

A

-3 -2

-I

0

I

Fie. 20. — Multivariate positions of pre-1930 speci-
mens from southern Texas relative to series of C.
rufus and C. latrans. Triangles, C. rufus rufus; black
dots, C. latrans tcxcnsis.

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

39

1904 near the Texas coast in Calhoun, Colo-
rado, and Liberty counties (see appendix A,
part 16). The multivariate positions of these
specimens are shown in figure 20. Clearly,
these specimens have close affinity with those
of the standard red wolf sample, which were
taken farther east and identified as C. rufus
gregoryi, but they are smaller and more nar-
rowly proportioned. In other words, they fit
Goldman's (1944:487) description of the sub-
species C. r. rufus. A specimen taken in 1906
in Hardin County, extreme southeastern
Texas, and identified as C. r. gregoryi, was
incorporated above in my standard red wolf
sample. There are no other usable, pre-1930
specimens of red wolves obtained in southern
and coastal Texas.

Skulls of 31(5) males and 26(8) females
labeled as C. latrans texensis were taken from
1891 to 1918 in south Texas (see map, Fig.
14; appendix A, part 17), more than half
coming from Nueces County. Most of these
specimens fall within the range of variation
of western C. latrans (Fig. 20). Although
three Nueces County specimens have statisti-
cal positions that are removed in the direction
of the standard red wolf sample, their D2
values confirm closer affinity to C. latrans.
Therefore, although there is suggestion of
limited hybridization, early specimens from
south Texas may be separated into two dis-
tinct groups, representing the species C. la-
trans and C. rufus.

Examination of Pre-1930 Material from
Central Texas

Early specimens from the Edwards Pla-
teau area of central Texas are not so easily
divisible into recognized species, and have
been the cause of much confusion. A report
by Allen (1896:75-76) is pertinent, since it
contains both the earliest account of possible
hybridization in the area, and an early refer-
ence to the presence there of three kinds of
Canis. Allen's informant, H. P. Attwater, said
that Canis lupus was "formerly common in
Bexar County, but I have not heard of their

occurrence here for several years. They are
still found in the broken, hilly country north-
west of San Antonio, particularly in Edwards
County." And, regarding C. latrans, he stated:
"In Kerr County and adjoining counties they
are the 'thorn in the side,' of the sheepmen.
Mr. Lacey [Howard Lacey, a rancher said to
be a careful and reliable observer] says that
the Coyotes of that region are different from
the Coyotes of the prairies, being much
larger. They are believed by the ranchmen
to be a cross between the 'Lobo' (Wolf) and
the Coyote. Two years ago, when the bounty
act was in force, the regular 'Lobo' price was
allowed for the large Coyotes of the rocky
region to the northwestward of San Antonio."

With regard to the red wolf and coyote
in central Texas, Goldman (1944:481) re-
ported: "Specimens collected in the vicinity
of Llano, Tex., include typical examples of
both species and individuals not sharply dis-
tinctive of either. Close approach in essential
details and the apparent absence of any in-
variable unit character suggests the possibil-
ity of hybridism in some localities in Texas."
McCarley (1962) considered a series of eight
skulls from the vicinity of Llano to suggest
the occurrence of hybridization. Paradiso and
Nowak (1972a) concluded that specimens
taken throughout the Edwards Plateau area
of Texas between 1890 and 1918 represented
a hybrid swarm of C. rufus x C. latrans. For
purposes of this paper, all available early
material from central Texas, as well as that
collected in western Texas in the same period,
was re-examined and subjected to more criti-
cal statistical evaluation.

To facilitate an interpretation of the situa-
tion, the central and west Texas material was
divided on an arbitraiy geographic basis. Fig-
ure 21 shows multivariate positions of speci-
mens collected in Texas counties to the west
and just north of Tom Green County; figure
22 does the same for skulls taken in Tom
Green County; and figure 23 shows positions
of specimens taken in the following counties
of central Texas: Blanco, Burnet, Coleman,

40

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

-3 -

8

LATRANS

-3 -2

-I

- r-
0

I

H 1-

3 4

RUFUS

western Texas, and into Tom Green County.
Goldman (1944) assigned to C. r. riifus
a male from 22 miles north of Sheffield, Pecos
County, which is about 100 miles farther west
than the closest other record of C. rufus
listed by him. Paradiso and Nowak ( 1972a )
considered the skull to represent a coyote
and to hardly be distinguishable from some
specimens of C. latrans lestes. While making
a preliminary examination of my standard
coyote sample, I noticed one large skull of
lestes from Bountiful, Conejos County, south-
ern Colorado, that closely resembled the
Pecos County specimen. The Colorado skull
was withheld from the coyote series in the
multivariate analysis comparing the standard
samples of C. rufus and C. latrans. Its rela-
tive position is plotted in figure 21 and dem-
onstrates that a few western coyotes are sta-
tistically well removed from normal C. latrans
in the direction of C. rufus. The Pecos

Fig. 21. — Multivariate positions of certain indi-
vidual specimens relative to series of C. rufus and
C. latrans. Black dots, C. latrans texcnsis from west-
ern Texas; P, specimen from Pecos County, Texas
originally identified by Goldman (1944) as C. rufus
rufus, but considered in this paper to probably rep-
resent C. latrans texensis; C, specimen of C. latrans
lestes from Conejos County, Colorado.

Concho, Edwards, Gillespie, Kerr, Llano, Mc-
Culloch, Menard, San Saba, and Sutton (see
also Fig. 14; appendix A, parts 18-20). All
material was collected from 1915 to 1918,
except for the specimens from Edwards, Gil-
lespie, and Kerr counties, which were taken
from 1899 to 1906.

Some 30 males and 15 females from west-
ern Texas, each previously identified as C.
latrans texensis, fall within or near the range
of variation of the standard coyote sample.
The skulls of 36 males and 14 females taken
farther east in Tom Green County, each also
previously identified as texensis, show the
same kind of statistical distribution. Hence
there appears to be no clinal shift, in the
direction of C. rufus, as the range of C. la-
trans passes from the western states, across

3
2
I

0
-I
-2
-3

3
2
I

0
-I
-2
-3

I- S

LATRANS

— i r

-3 -2

9

-3 -2 -I

0

RUFUS

Fig. 22. — Multivariate positions (black dots) of
individual pre-1930 specimens of C. latrans texensis
from Tom Green County, Texas, relative to series of
C. rufus and C. latrans.

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

41

County specimen, the position of which is
close to that of the Colorado skull, could be
such a coyote. The D2 values of both skulls
indicate closer affinity to C. latrans than to
C. rufus, but there is no way of definitely
determining whether the Pecos County speci-
men represents some genetic influence from
the red wolf. The skin of this specimen was
examined, but could not be distinguished
from those of other coyotes taken in central
and western Texas.

Central Texas is a land of major natural
transition. Here the high plains break down
into the rough country of the Edwards Pla-
teau; the arid Sonoran area of the Lower
Austral Life-zone passes into the humid Aus-
troriparian area; and typical prairie and des-
ert vegetation merge with eastern deciduous
forest. Regarding the transition of the vege-
tation pattern on the Edwards Plateau, Bray
(1904:14) noted that "conditions vary enough
to give in some places, as in well-watered and
sheltered canyons, a relatively luxuriant
growth, while in other situations, as upon
stony arid slopes, there is the scantiest vege-
tation."

At the turn of the century, central Texas
also was undergoing pronounced changes be-
cause of the influence of man. Overgrazing
by cattle and sheep, and the frequent occur-
rence of fire, had suppressed the native prai-
ries and permitted the spread of mesquite.
Simultaneously, clearing and cultivation along
the river bottoms had deprived the area of
its richest forest growth (Bray, 1904). And,
of course, the hand of man was turned heav-
ily against predatory animals such as wolves
and coyotes.

The specimens of Cards collected in cen-
tral Texas were originally identified as C.
rufus rufus, C. latrans texensis, and C. latrans
frustror. Jackson (1951:271, 279) considered
this last named subspecies to range north of
Nueces Bay, and to the eastern edge of the
Edwards Plateau where he recorded speci-
mens from Blanco, Burnet, Llano, and San
Saba counties. Jackson wrote that texensis

occurred to the south of Nueces Bay, but that
its range extended northwestward into cen-
tral Texas (he listed specimens there from
Coleman, Concho. Gillespie, Kerr, McCul-
loch, Menard, and Sutton counties). Gold-
man (1944:488-489) thought that C. rufus
rufus occurred throughout central Texas, and
he recorded pre-1930 specimens from Burnet,
Edwards, Kerr. Llano, and McCulloch coun-
ties.

Those specimens from central Texas that
could be subjected to multivariate analysis in-
cluded animals identified as the following:
C. latrans texensis, male — 13, female — 12(2);
C. latrans frustror, male — 9, female — 15; C.
rufus rufus, male — 21(1), female — 5. Locali-
ties are plotted on the map in figure 14, and
statistical positions are shown in figure 23.
Also shown are positions of the 14 skulls of
C. lupus monstrabilis collected in central and
western Texas, each of which was used above
in my standard gray wolf sample (most were
collected from 1900 to 1920). Their positions
are plotted here to emphasize that, although
C. lupus was present in the area, it was dis-
tinct from the animals being called red
wolves, and was not a factor in the problems
under discussion.

The multivariate positions of the central
Texas skulls identified as texensis, frustror,
and rufus form a statistical bridge between
the ranges of variation of standard C. latrans
and C. rufus. There is no meaningful place
to draw a line separating these skulls into
coyotes and red wolves, and we apparently
are dealing with some sort of blending be-
tween the two. Specimens collected in cer-
tain restricted localities, such as those from
the vicinity of Llano and Burnet, by them-
selves bridge the statistical gap between red
wolves and coyotes. After an evaluation of
these statistics, a visual re-examination of the
skulls, and a check of available skins from
central Texas, it must be concluded that the
original identification of these specimens was
arbitrary and based on an incomplete under-
standing of the situation. In this regard it is

42

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

3
2
I

0
-I
-2
-3

. LATRANS

RUFUS

m

m
m

m

m

m

m

-4

~i 1 r

2 3 4

5

6

—r-
7

— 1-
9

0

8

10

3

2

I

0-
-I-

-2"
-3-

LATRANS

RUFUS

m

m

m
m

m

m

n 1 r

2 3 4

5

6

7

~i r-

8 9

-4 -3 -2

0 I

10

Fig. 23. — Multivariate positions of certain individual pre-1930 specimens from Texas, relative to series of
C. rufus and C. latrans. Black dots, specimens from central Texas originally identified as C. latrans texensis
or C. latrans frustror; squares, specimens from central Texas originally identified as C. rufus rufus; M, speci-
mens of C. lupus monstrabilis from central and western Texas. Males are above, females below.

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

43

interesting to note an unusual sex ratio among
the available specimens of rufus and frustror
from the area. For specimens originally iden-
tified as frustror, females outnumber males
15 to 9; but for skulls previously referred to
rufus, males outnumber females 20 to 5. This
suggests to me that the original identification
of the specimens was based partly on the bias
of assigning larger skulls, predominantly
males, to rufus, and smaller skulls, predomi-
nantly females, to frustror.

No other group of specimens, at least
among those collected before 1930, falls to
such an extent between the statistical limits
of C. latrans and C. rufus. This condition
could have resulted from one of the following
factors: (1) intergradation in central Texas
between the small western coyote and the
larger eastern red wolf; (2) long-term inter-
breeding, and the production of a hybrid
zone serving as a bridge for the flow of genes
from one species to the other; (3) short-term
interbreeding caused possibly by drastic al-
teration of the environment.

I do not think that the statistical distribu-
tion of specimens is indicative of intergrada-
tion and hence the conspecificity of red wolf
and coyote. Recognized subspecies of C.
latrans intergrade throughout western North
America and show no morphological overlap
with red wolves. There is no evidence of a
cline of characters approaching those of
standard C. rufus, even among specimens
taken as far east as Tom Green County,
Texas. The sudden breakdown found in the
limited area of central Texas is highly atypi-
cal of North American Canis, and does not
represent normal subspecific intergradation.
Furthermore, as we have seen, there is no
suggestion of original intergradation in any
other area where the ranges of C. rufus and
C. latrans met, and there is direct evidence
of the sympatric occurrence of the two spe-
cies in Missouri (see pp. 35-36).

While I therefore consider hybridization
responsible for the situation in central Texas,
there is no reason to believe that the process

had been occurring over a long period. Had
interbreeding been going on for many years
or centuries, its effects seemingly would have
spread beyond central Texas.

Available evidence favors the theory that
hybridization between C. rufus and C. latrans
was of relatively short-term occurrence, and
probably had begun in the latter part of the
nineteenth century in response to man's dis-
ruption of the habitat and his persecution of
native wild canid populations. This is in
accord with earlier statements by McCarley
(1962) and Paradiso and Nowak (1972a).
Since central Texas was an area of natural
transition, where the eastern red wolf and
western coyote would have overlapped in
range, if not in habitat, opportunities for
interbreeding undoubtedly occurred. Once
man altered the environment, and at the same
time attacked the wolves and coyotes, ecologi-
cal and behavioral isolation might have brok-
en down, and large-scale hybridization be-
gun.

According to Mayr (1963:128): "By far
the most frequent cause of hybridization in
animals is the breakdown of habitat barriers,
mostly as a result of human interference."
Mayr (1963:118-121) used the term "hybrid
swarm" to describe populations, in such areas
of breakdown and interbreeding, that form
a continuous bridge between two parent
species. This term seems applicable to the
population of Canis in central Texas.

The exact original limits of the ranges
of C. rufus and C. latrans in central Texas are
unknown, partly because the hybridization
factor has obscured the picture. The red
wolf, however, usually is considered a species
of the eastern forests. In this regard it is in-
teresting to note that the area of central Texas
under discussion falls just within the extreme
western boundary of eastern forest elements
(see Fig. 14). Indeed, oak trees first appear
immediately to the east and south of Tom
Green County.

At this point I would like to summarize
briefly the early status of Canis in the south-

44

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

east, as evaluated on the basis of specimens
collected prior to 1930. The red wolf, a spe-
cies distinct from the gray wolf of northern
and western North America, was found along
the Atlantic coast and westward probably to
the edge of the prairies. Large series of spe-
cimens show that before 1930 the red wolf
in the lower Mississippi Valley was easily
separable from the western coyote. The two
species occurred sympatrically, or at least
showed no tendency to intergrade in most
areas where their ranges approached. A small
percentage of specimens taken in eastern
Oklahoma and Arkansas suggest that limited
hybridization had occurred at certain locali-
ties in those states. In central Texas, how-
ever, more extensive interbreeding had re-
sulted in the formation of a hybrid swarm
between the two species.

Survival of the Red Wolf from the
1930's to 1950's

Although the red wolf, in unmodified
form, seems to have survived in the lower
Mississippi Valley through the 1920's, the
species had been under heavy human pressure
throughout most of its range. The course of
its decline was discussed by me ( 1967, 1970,
1972, 1974), Russell and Shaw (1972), and
Young ( 1944 ) . Wolves disappeared from
Pennsylvania, Maryland, West Virginia, Vir-
ginia, North Carolina, South Carolina, Ten-
nessee, Kentucky, Ohio, Indiana, and Illinois
without specimens having been saved which
would have enabled us to determine their
systematic status.

Farther to the south, wolves, probably
only C. rufus, did not survive much longer.
Harper (1927:315-317) said that the last
known kill of C. r. floridanus in the Okefino-
kee Swamp of Georgia took place about 1908,
although there were later reports. There
seem to be no other definite records of wolves
in Georgia after 1900. According to Chap-
man (1894:345) wolves in Florida were al-
ready "on the verge of extinction." The last
reported occurrence in the state, listed by

Young (1944:25), was in 1903 near the Ever-
glades. H. H. Bailey (1930) wrote that a
wolf was killed about 1918 or 1920 on the
Osceola side of the Kissimee River.

Howell (1921:30) reported that wolves in
Alabama were "on the verge of extinction.
Their last stronghold appears to be the rough,
hilly country stretching from Walker County
northwestward to Colbert County." The last
recorded kill took place south of Cherokee,
Colbert County in 1917. Apparently, how-
ever, wolves held out for a while longer in
the state. In 1937 and 1938 reports by the
Predator and Rodent Control branch of the
U.S. Bureau of Biological Survey (on file at
the offices of the Division of Animal Damage
Control, U.S. Fish and Wildlife Service,
Washington, D.C. ) it is said that in Alabama
nine wolves were killed in 1937 and three in
1938. Holliman (1963:242) reported a speci-
men of C. rufus collected at Livingston, Sum-
ter County, western Alabama in 1944. The
multivariate position of its skull falls within
the statistical limits of the standard red wolf
sample ( Fig. 24 ) .

Wolves apparently disappeared from most
of the higher country of Mississippi at an
early date, but survived until comparatively
recently along the Mississippi River and Gulf
Coast. Jenkins (1933:125-127) reported "but
few wolves in Mississippi from 1892 to the
present time." He added that during the
great flood of 1927 many wolves crossed into
southwestern Mississippi from Louisiana. Ac-
cording to Young (1944:29), these wolves
were still present in 1932 and were preying
on livestock. Goldman (1944:485) reported,
but did not examine, a specimen of C. rufus
gregoryi taken at Biloxi in 1931. I found this
skull to fall squarely within the range of vain
ation of standard C. rufus (Fig. 24).

Very little information is available con-
cerning the possible former occurrence of
the red wolf in Kansas. Lantz (1905) re-
ferred to "Canis ater" as "Once abundant.
Still found in a few scattered sections of the
state." Cockrum (1952:229) cited reports of

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

45

RUFUS

Fig. 24. — Multivariate positions of specimens of
C. rufus, and two hybrids, taken from the 1930's to
the 1950's. A, Alabama gregoryi; K, Arkansas gre-
goryi; L, Louisiana gregoryi; M, Mississippi gregoryi;
O, Oklahoma gregoryi; T, Texas gregoryi; R, Texas
rufus; B, specimen of C. r. rufus taken in 1954 in
Brazoria County, Texas; D, specimens of C. r. rufus
taken in Madison County, Texas (originally identified
as C. latrans, but considered in this paper to repre-
sent C. rufus); S, specimen of C. r. rufus taken in
southern Harris County, Texas (originally identified
as C. latrans, but considered in this paper to repre-
sent C. rufus); N, specimens of probable red wolf-
coyote hybrids taken in northern Harris County,
Texas.

a red wolf killed in Cherokee County, ex-
treme southeastern Kansas, in 1908 or 1909,
and of one heard howling there in 1915.

The situation in the Ozark-Ouachita up-
lands, and adjacent areas, is most confusing
in the 1930's and 1940's, and is discussed in
detail in the next subsection of this paper. I
will first deal with those areas in which the
red wolf seems to have survived past 1930 in
unmodified form.

Six skulls obtained in northern Louisiana
from 1935 to 1940 are indicative of the con-
tinued presence of C. rufus in that state (Fig.

24). No other specimens of wild Canis taken
between 1930 and 1950 arc available from
Louisiana. All evidence suggests that up
until the last few years of this period the red
wolf maintained moderate numbers over
much of Louisiana. Ry the early 1950's, how-
ever, the species had been decimated in
most areas, except for the eastern bottom
lands and southern marshes of the state
(Nowak, 1967). McCarley (1962) reported
specimens of C. rufus obtained in 1956 and
1957 in Terrebonne and Madison parishes,
respectively. The skull taken in Terrebonne
Parish, southern Louisiana, falls within the
statistical limits of the red wolf (Fig. 24).
The skull from Madison Parish, northeastern
Louisiana, was not suitable for multivariate
analysis, but does represent C. rufus.

According to Gipson (1972:4), red wolves
were present in southern Arkansas as late as
the 1950's. A specimen taken in 1942 on the
Union-Columbia county line, just north of the
Louisiana border, has a multivariate position
within the limits of C. rufus (Fig. 24).

A male and a female specimen taken in
1936 near Rattiest, McCurtain County, ex-
treme southeastern Oklahoma were reported
by McCarley (1962) to be red wolves. These
two skulls are statistically well removed from
my standard coyote sample (Fig. 24), and
probably indicate continued survival of C.
rufus in the area.

The southeastern part of Texas, especially
the area known as the Rig Thicket, seems to
have been one of the last major refuges for
the red wolf. Goldman ( 1944:4S6-489) re-
corded post-1930 specimens of C. rufus gre-
goryi from Hardin, Newton, and Polk coun-
ties; and C. r. rufus from Rrazoria, Rrazos,
Liberty, Montgomery, and Walker counties.
In addition, Jackson (1951:275) listed speci-
mens of C. latrans frustror from Harris
County, in the same area. Most of this ma-
terial, plus two other skulls in the National
Museum from Madison County that had been
mistakenly labeled as C. latrans texensis, were
subjected to multivariate analysis. The re-

46

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

suits (Fig. 24) confirm the survival of the
red wolf in southeastern Texas. As in other
cases, specimens identified as C. r. rufus tend
to show some statistical approach to C. la-
trans. The two Madison County specimens
apparently represent C. r. rufus. One of the
Harris County skulls was collected at Genoa
on the south side of Houston. Although orig-
inally identified as C. latrans frustror, it falls
within the range of variation of the standard
red wolf sample, and I think it actually repre-
sents C. rufus. The other two Harris County
skulls were taken at Humble, on the north
side of Houston, and their intermediate sta-

tistical positions suggest genetic influence of
the coyote at that locality. All of the above
listed Texas specimens were collected be-
tween 1930 and 1943, and I examined very
little material taken over the next 20 years in
the state. The skull of a female taken in
eastern Brazoria County in 1954 has a multi-
variate position suggesting the survival of
C. r. rufus in the area.

In summary, it appears that the red wolf
continued to exist in parts of its former range
from the 1930's to the 1950's. No usable speci-
mens of wild Canis, other than those men-
tioned in this subsection of the paper, were

100

90

39

!

\

•
[

\ ' 1
3 • \C^

39

_%_B_5__^ «J '

|*«

•

•

3
•

•

A A

<

f

•□n5

A )

A

A)

30

I A /

N. •

A A A
A AA.

aaaa

a

4 ~~k ^Xj^)/*^0-

30

0 200

i /

1

scale of miles

100

90

Fig. 25. — Map showing localities of specimens taken from the 1930's to the 1950's in the south-central
United States. Triangles, C. rufus; black dots, specimens originally identified as C. latrans; squares, speci-
mens originally identified as C. rufus, but considered in this paper to represent hybridization between red
wolf and coyote (also two specimens taken in northern Harris County, Texas originally identified as C. la-
trans). See Figs. 24, 26, 27, 28, 29, and 30 for more details. Note: because of the scale of the map it was not
possible to plot all localities in crowded areas.

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

47

collected in the areas under discussion. Al-
though the distribution of the red wolf was
certainly not stable through the entire period,
it seems safe to say that about 1940 the range
of unmodified C. rufus extended from extreme
southeastern Oklahoma and southeastern
Texas, across Louisiana, southern Arkansas,
and southern Mississippi, and into western
Alabama . Localities of the available speci-
mens from this region are listed in appendix
A (part 21), and are plotted in figure 25.

Increase of Hybridization from the
19.30's to 1950's

Specimens collected prior to 1930 indicate
the effects of hybridization between C. riifus
and C. latrans only in central Texas, and at
a few localities in Arkansas and eastern Okla-
homa. In the following decades, specimens of
apparent mixed genetic origin were taken
over larger areas. Information from the 1930's
and early 1940's seems important to a full
understanding of the phenomena then engulf-
ing populations of southern Canis, but, un-
fortunately, comparatively few specimens
were saved from the areas of greatest interest.

As explained earlier, specimens taken
in 1900 in Calhoun County, on the middle
Gulf Coast of Texas, have multivariate posi-
tions within or near the limits of standard C.
rufus. No other specimens of wild Canis were
collected in this area until after 1930. Jack-
son (1951:275) listed specimens of C. latrans
frustror from Aransas, Refugio, and Victoria
counties. Goldman (1944:488-489) stated
that specimens of C. rufus rufus had been
collected in the same area. Skulls of 10 males
obtained from 1936 to 1942 in these three
adjoining counties, were suitable for analysis
(see Fig. 25; appendix A, part 22). Five had
been referred originally to C. rufus and five
to C. latrans. As shown by the positions
plotted in figure 26 these previous designa-
tions are questionable and some genetic ex-
change seems to have occurred.

RUFUS

Fig. 26. — Multivariate positions of individuals
taken from 1936 to 1942 in Aransas, Refugio, and
Victoria counties on the central coast of Texas.
Squares, specimens originally identified as C. rufus;
black dots, specimens originally identified as C.
latrans.

Few specimens from farther inland in
southern and central Texas are available from
this period. No skulls at all were saved from
the central Texas counties that were appar-
ently occupied by a hybrid swarm of C. rufus
x C. latrans prior to 1920 (see pp. 41-43).
Probably, as is shown in the next subsection
of the paper, all wild Canis had been extermi-
nated in this area by the 1930's. Three males
of C. /. texensis, taken in 1942 in Bexar Coun-
ty, have multivariate positions within the
range of C. latrans (Fig. 27).

Farther north, a series of specimens was
taken from 1930 to 1942 in the area between
the Colorado and Red rivers (see Fig. 25;
appendix A, part 23). Records of both C.
rufus and C. latrans were listed from various
localities in this area by Goldman (1944) and
Jackson ( 1951 ) , but I can see no basis for
separating these specimens into two species.
I subjected 18 of the skulls to multivariate
analysis and found most to fall within or near
the range of variation of standard C. latrans
(Fig. 27). The more intermediate positions
of two females (D2 from coyote 36.7 and
22.8; from red wolf 39.3 and 25.6) suggest
genetic influence from C. rufus.

48

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

LATRANS

RUFUS

~i i i i i i i i i i i i

-4-3-2-10 1 2 3 4 5 6 7

□

I

-| 1 1-

2 3 4

Fig. 27. — Multivariate positions of individuals
taken in Bexar County, Texas (t), and in north-
central Texas from 1930 to 1942. Black dots, speci-
mens originally identified as C. latrans; squares,
specimens originally identified as C. rufus.

From 1933 to 1942 a large series of skulls
was collected on the Wichita Mountains Na-
tional Wildlife Refuge, Comanche County,
southwestern Oklahoma (see Fig. 25; appen-
dix A, part 24). Nearly all of these speci-
mens were originally identified as C. latrans
frustror, and this designation is supported by
statistical positions of 22 males and 25 fe-
males (Fig. 28). Goldman (1944:488) listed
a single specimen of a male from this group
as C. rufus rufus. I, however, was not able
to distinguish this skull from those of some
large coyotes, and its multivariate position
(Fig. 28) suggests affinity with C. latrans.
Therefore, if the red wolf ever did occur as
far west in Oklahoma as Comanche County,
it apparently disappeared before the 1930's.
Nonetheless, the presence of coyotes, some of

which had perhaps received an introgression
of genes from C. rufus, caused persons to
think that wolves still inhabited the area.
According to Halloran and Glass (1959:363):
"The Texas red wolf reaches the western edge
of its range in the Wichita Mountains. Most
of the canids present on the refuge are coy-
otes, but the situation is clouded by the local
custom of calling everything larger than a
small coyote a wolf." And Duck and Fletcher
(1945:128) reported that C. rufus rufus was
being taken over most of the state.

Several other specimens, identified as C. I.
frustror, were taken in central and north-
eastern Oklahoma in 1932. Those usable in

3
2

I

0
-I
-2
-3

3
2

I

0

-I
-2
-3 -

\- s
LATRANS

RUFUS

~l 1-

5 6

-4

-i 1 r

-3 -2 -I

-T-
0

3 4

Fig. 28. — Multivariate positions of individuals
taken in Oklahoma from 1932 to 1942, relative to
series of C. rufus and C. latrans. Black dots, C.
latrans frustror from Comanche County; square,
specimen from Comanche County identified by Gold-
man (1944) as C. rufus, but considered in this paper
to represent C. latrans; v, specimens from Cleveland
County; E, coyotes from eastern Oklahoma; triangle,
specimen from Atoka County apparently representing
C. r. rufus.

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

49

multivariate analysis include three males from
Cleveland County, one male from Cherokee
County, and two females from Osage County
(see appendix A, part 25). These six speci-
mens demonstrate statistical approach to C.
ntfus (Fig. 28). A single female, taken the
same year in Atoka County, southeastern
Oklahoma, and identified originally as C. r.
rufus, seems to be within the range of varia-
tion of other specimens assigned to that sub-
species (Fig. 28). Two other skulls (in
USNM) collected in 1932 in eastern Okla-
homa are a female from Cherokee County
identified as C. r. rufus, and a male from
LeFlore County assigned by Goldman to
C. r. gregoryi. Neither one is suitable for
multivariate analysis, but both appear to be
intermediate in characters between the stand-
ard samples of C. rufus and C. htrans. As
mentioned in the last subsection of this paper,
two skulls of red wolves were collected in
1936 in McCurtain County, southeastern
Oklahoma.

The distribution of specimens thus sug-
gests that by the 1930's the red wolf had
become restricted to parts of southeastern
Oklahoma. A few skulls taken in the north-
eastern and central areas of the state, how-
ever, may represent the results of interbreed-
ing between C. rufus and C. latrans (see Fig.
25).

Specimens discussed earlier in the paper
demonstrate the sympatric occurrence of C.
rufus and C. latrans in southern Missouri
prior to 1930. Ry that year, however, the red
wolf had been decimated by government and
private hunting (Nowak, 1970; Sampson,
1961). In contrast, the coyote seems to have
maintained its numbers and increased its
range. One possible indication of subsequent
hybridization in the area was provided by
Rennitt and Nagel (1937:168) who wrote:
"It is difficult to outline exactly the range of
these two species in Missouri, since so many
observers cannot tell them apart."

According to Sampson (1961), the last
pure specimen of a Missouri red wolf was
collected in 1932 in Dade County. This speci-
men, however, was not listed by either Gold-
man (1944) or Jackson (1951), and its multi-
variate position (Fig. 29) is well within the
range of variation of C. latrans. A specimen
collected the same year in Iron County was
listed by Goldman (1944:486) as gregoryi,
and was considered by Paradiso and Nowak
(1972a: 11) possibly to be a red wolf. At
present, however, I think the skull represents
hybridization involving C. familiaris (see Fig.
16).

Sampson ( 1961 ) reported that in 1941
and 1942 a series of 171 specimens was taken
throughout Missouri, partly in order to de-

3
2
I

0
-I

-2
-31-

8

LATRANS

RUFUS

□

"i r

-3 -2

~i 1 1

5 6 7

3
2

I

0

-I
-2
-3 -

9

-3 -2

0

Fig. 29. — Multivariate positions of individuals
taken in Missouri from 1932 to 1942, relative to series
of C. rufus and C. latrans. Black dots, specimens
originally identified as C. latrans; d, specimen from
Dade County listed as C. rufus by Sampson (1961);
squares, specimens from Ozark and Oregon counties
identified as C. r. rufus by Leopold and Hall (1945).

50

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

termine whether the red wolf was still present
in the state. He wrote that 44 of the skulls
were sent to the U.S. National Museum, but
suggested that these included all specimens
thought most likely to represent C. rufus.
Hence the sample sent in probably already
was biased in favor of larger and more wolf-
like individuals. In any case, the Museum
reportedly identified the specimens as 39 coy-
otes, one dog, one coy-dog, and one red wolf-
coyote hybrid. Jackson (1951:274) referred
most of these specimens to C. latrans frustror,
and listed records all across southern Mis-
souri. Of the specimens taken in 1941 and
1942 in the southern half of Missouri, eight
males and seven females were suitable for
multivariate analysis (see Fig. 25; appendix
A, part 26). Their relative positions are
shown in figure 29, and indicate close affinity
with C. latrans, but possibly some genetic
influence from C. rufus.

Elder and Hayden (1977) evaluated 20
specimens of Canis from Missouri, deposited
in the collection of the University of Missouri,
and reported one to be C. familiaris and two
to be coy-dogs. Of the others, four taken
between 1945 and 1950 in the Ozarks were
identified as C. rufus, nine taken mostly from
1940 to 1955 were identified as C. latrans,
and four taken between 1946 and 1949 were
considered to probably be red wolf-coyote
hybrids. The graphical results of this analysis
suggest to me that these last 17 specimens
represent a single interbreeding population,
and not that the red wolf survived as a dis-
tinct entity in Missouri until 1950. The fact
that Elder and Hayden also identified the
two skulls mentioned in the next sentence as
C. rufus, suggests that their accepted limits
for this species were less restrictive than my
own.

Leopold and Hall (1945) referred to C. r.
rufus two specimens of males, taken in 1941
and 1942, respectively, in Ozark and Oregon
counties, extreme southern Missouri. In a
1952 letter cited by Sampson (1961), Leopold
wrote: "In all respects these two specimens

are intermediate between typical coyotes and
typical red wolf." Multivariate analysis of
the skulls (Fig. 29) supports this statement.
Leopold and Hall had assigned the specimens
to C. r. rufus on the basis of their small size,
even though they were obtained within the
range designated for C. r. gregoryi by Gold-
man (1944:484-486). Since Goldman had
assigned two male specimens from Reeds
Spring, Stone County, Missouri to gregoryi,
an awkward situation developed in which
material of rufus was reported from farther
to the east. Hall and Kelson (1952:340-341)
technically solved this problem by stating
that the two Stone County skulls resembled
C. r. rufus in small size and cranial charac-
ters, and by referring them to this subspecies.
The issue, however, is a temporal as well as
a geographic one. The Stone County material
was taken in 1923 and 1924, and as we have
seen, falls right in the statistical center of a
group of 70 skulls of C. r. gregoryi (Fig. 19).
The Ozark and Oregon County specimens
were taken almost 20 years later at a time
when the red wolf had been nearly or com-
pletely exterminated in Missouri. They are
statistically beyond standard C. rufus, and
should, I think, be looked upon as represent-
ing hybridization between C. rufus and C.
latrans, rather than a particular subspecies of
the red wolf.

A related problem seems to exist just to
the south in Arkansas. Locality records from
this state listed by Goldman (1944:485-488)
include 42 for C. rufus gregoryi and 6 for
C. r. rufus. If the adjacent marginal records
for each subspecies are plotted on a map, as
they were by Hall and Kelson (1959:852),
an awkward picture emerges. Several of the
localities from which rufus was reported are
seen to be nearly surrounded by the desig-
nated range of gregoryi. There are no geo-
graphic barriers that might account for such
a distribution, but once again the time ele-
ment seems to be a factor. As best as I can
determine, all but one of the specimens as-

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

51

signed by Goldman to gregoryi were col-
lected prior to 1930. In contrast, there is only
one locality from which specimens assigned
to rufus were taken before 1930. This locality
is Boxley, Newton County, which is not one
of those surrounded by the range of gregoryi.
A skull taken in 1922 at this site falls within
the standard range of variation of C. rufus
(Fig. 19). The localities that are encircled
by the range of gregoryi are Raspberry and
Hector, both in Pope County. Of the four
specimens recorded by Goldman from these
places, two are not usable in my analyses, one
has been identified by me as a dog hybrid
(Fig. 16), and the statistical position of one
is shown in figure 30. Of the other post- 1930
specimens listed by Goldman, only one could
be subjected to analysis, and its position is
also depicted in figure 30. The specimens
evaluated in the analysis, as well as the

RUFUS

Fig. 30. — Multivariate positions of individuals
taken in Arkansas from 1932 to 1951, relative to
series of C. rufus and C. latrans. Black dots, speci-
mens originally identified as C. latrans; squares,
specimens originally identified as C. rufus.

others, suggest a more pronounced influence
by C. latrans after 1930.

Jackson (1951:274) recorded C. latrans
frustror from a number of localities in western
and northern Arkansas. Most of the speci-
mens involved were collected between 1932
and 1942, but not all are suitable for analysis.
The positions of those that were statistically
evaluated are plotted in figure 30. A few
other skulls, taken as late as 1951 in the same
areas, are also depicted. Arkansas specimens
taken in this period are listed in appendix A
(part 27), and localities are shown in figure
25. The number of specimens is hardly
enough to allow a full understanding of the
situation, but their distribution suggests that
after 1930 C. rufus was no longer prevalent
in Arkansas, and that animals of more inter-
mediate or coyotelike characters were begin-
ning to predominate.

The Empty Zone and Expansion of
Modified Canis

The maps in figures 14 and 25 show many
areas in which no records of Canis are plot-
ted. Of course, material from states west of
Texas, and north of Oklahoma and Missouri,
was available, but was not considered appli-
cable to the problems under discussion. The
records that are shown seem to form a rough-
ly circular pattern around a blank area cen-
tered in northeastern Texas. It could be
argued that this area is not represented only
because no one ever collected there, or be-
cause no material from there was preserved.
But it should be understood that almost all
specimens discussed in this section of the
paper were taken in the course of Federal
predator control efforts. This control work
was done in areas where complaints existed,
and thus is indicative of the presence and
abundance of wild Canis. To be sure, only
a small percentage of the animals killed were
sent as specimens to the National Museum,
and collecting was not continued on a regular
geographic or temporal basis. Nonetheless,
the available material may serve as one relia-

52

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

ble means of evaluating the situation over a
lengthy period. Therefore, the constant lack
of specimens from the area in question leads
me to think that no significant population of
wild Canis had survived there into the twen-
tieth century.

Bailey (1905:172) stated: "The black wolf
is reported from a few localities in the tim-
bered region of eastern Texas, but in most
cases as 'common years ago, now very rare
or extinct.' " But in another publication, per-
haps because it was written for different
purposes, Bailey (1907:13) stated: "In the
timbered region of eastern Texas, especially
in the extensive swamps and bottoms, the
black wolf is still abundant and very de-
structive to cattle and hogs, while it renders
sheep raising practically impossible." Bailey's
own field reports from 1904 (on file, National
Fish and Wildlife Laboratory, U.S. National
Museum of Natural History) do indicate that
in the area of the Big Thicket in Hardin and
Liberty counties, wolves were still common.
Strecker (1926:16) reported wolves in ex-
treme eastern Texas to be "now almost or
quite extinct." Apparently then, by the early
twentieth century wolves had been wiped out
in most of the higher country of eastern Texas,
but continued to live in the bottom land
swamps, especially in the area of the Big
Thicket.

To the north of the Big Thicket, it seems
as though there was actually a zone nearly or
completely empty of wild Canis. This idea
is supported by the existence of certain maps
showing the distribution of Canis in Texas at
various times. One of these, prepared by
Bailey (1907:7), depicts a large area of north-
central Texas as being free of the presence of
wolves. The status of coyotes in the area,
however, is not indicated.

In the annual report for fiscal 1931 of the
Texas District of the U.S. Bureau of Biologi-
cal Survey (on file, offices of U.S. Fish and
Wildlife Service, San Antonio), there is a
map showing the "Present Infestation of
Predatory Animals." The term "predatory

animals" is vague, but since coyotes and
wolves were usually the main target of Fed-
eral control programs, especially in Texas,
it is safe to assume that canids are being re-
ferred to. The map indicates that predators
were absent in the eastern part of Texas,
except along the coast, in the vicinity of the
Big Thicket, and immediately south of the
Red River. In a large area of central Texas,
including all of those counties thought occu-
pied by a hybrid swarm of Canis prior to
1920, the map shows predatory animals to be
"under control" in 1931. This term may be
taken to mean nearly, if not completely, ex-
terminated. There is evidence that all wild
Canis had been decimated in central Texas
by this time. According to Gabrielson (1936:
64), sheep raisers in this area sustained losses
of ten percent to predators in 1915, but, be-
cause of Federal control programs, had losses
of only a fraction of one percent in 1935.
Several of the reports cited by Russell and
Shaw (1971a) indicated that by the late
1930's coyotes and wolves had been com-
pletely exterminated in some central Texas
counties.

Another map, drawn by Russell and Shaw
(1971a), was based on reports made in 1940
by field personnel of the Texas Game, Fish
and Oyster Commission. The range of the
red wolf at this time was shown to extend all
along the Gulf Coast north of Nueces Bay,
to extend inland in southeastern Texas up the
Trinity and Neches river basins, and to in-
clude strips along the Red, Sabine, and Sul-
phur rivers in extreme northeastern Texas.
Most of the reports cited from northeastern
and inland Texas, however, state that wolves
were very rare in these areas. The only major
area of abundance indicated by the reports
cited was the southeastern corner of the state.
In this regard there is agreement with the
1931 Biological Survey map and the available
specimens (see pp. 45-46).

Although Russell and Shaw dealt pri-
marily with C. rufas, reports by Texas game
officials regarding the status of coyotes were

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

53

also made between 1940 and 1942 (on file,
offices of Texas Parks and Wildlife Depart-
ment, Austin). These reports indicate the
presence of coyotes only in those areas south
of San Antonio, west of the Pecos River, and
northwest of Fort Worth.

The sum of evidence from three different
sources (specimens. Federal reports, and
Texas government reports) suggests that the
following situation existed in Texas in the
1930's and early 1940's. Red wolves were
present in moderate numbers in the south-
eastern corner of the state and along the
central Gulf Coast. In the latter area hybridi-
zation with coyotes was probably occurring.
Coyotes themselves were common in the
southern, western, and northern parts of
Texas. In a large area centered in north-
eastern Texas, however, all kinds of wild
Canis seem to have been rare, if not totally
absent. I think that this empty zone may
have been significant for two reasons : ( 1 )
it served as an actual barrier that for a time
limited contact, and probably interbreeding,
between C. latrans to the north and west, and
C. rufus in southeast Texas, and hence may
have assisted in maintaining an unmodified
red wolf population in the latter area; and
(2) it formed an empty niche for predators
that awaited reoccupation by suitable canids.

This void apparently began to fill by the
1940s. As discussed in the previous subsec-
tion of this paper, coyotes and animals inter-
mediate to C. latrans and C. rufus seemed by
that time to be increasing in Oklahoma,
southern Missouri, and Arkansas. In this
period there are several references to the
presence of wild Canis in Arkansas. Black
( 1936 ) wrote that the "timber wolf" was rare
in the northwestern part of the state, and he
did not mention the presence of coyotes.
Dellinger and Black (1940) listed several
records of coyotes, and also remarked:
'Wolves are becoming rather common in the
Ozarks." The Arkansas Game and Fish Com-
mission (1951:96-99) reported that coyotes
had been extending their range and could be

found as far east as the central part of the
state. Wolves, which also had reportedly in-
creased, were said to be most common in the
Ozark area, and second most common in
southwestern Arkansas. But in this last named
account, wolves and coyotes were discussed
together and perhaps were being confused.
Furthermore, it was stated that the Missis-
sippi Valley wolf, C. r. gregoryi, had become
rare and should not be totally exterminated.
This implies recognition that the canids re-
ported to be common were actually some-
thing other than the original native wolves
of most of Arkansas. In contrast to this view-
point, Sealander (1956:279) reported: "The
race gregoryi occurs throughout the State and
is quite numerous in some counties. It evi-
dently has largely replaced the race rufus
over its former range in Arkansas." Sea-
lander, however, informed me (pers. comm.)
that this statement had been based partly on
Goldman's (1944:487) report that rufus had
become restricted to parts of central and
southern Texas. Since rufus was thus not sup-
posed to exist in Arkansas, but since canids
thought to be red wolves were certainly pres-
ent, Sealander had to refer the latter to
gregoryi.

These references suggest an increase in
wild Canis in Arkansas between the 1930's
and 1950's, with the animals involved be-
lieved to be wolves. None of these reports,
however, was based on a thorough examina-
tion of specimens. The few specimens that
are available from this period suggest that
the original wolf population of Arkansas was
being replaced by more coyotelike animals
(Fig. 30). Later, Paradiso (1966) reported
that three specimens taken in 1964 in Chicot
County, extreme southeastern Arkansas, rep-
resented a range extension of C. latrans frus-
tror. The first study of large series of Arkan-
sas specimens collected since 1930 was
described by Gipson (1972), and Gipson,
Sealander, and Dunn (1974). On the basis
of multivariate analysis of 284 adult skulls
taken from 1968 to 1971, they concluded that

54

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

the existing population of Cams in the state
was predominantly coyote, though with some
red wolf influence. The essentially coyotelike
character of this population also is recognized
in the next subsection of my paper. There-
fore, the canids that were said to be common
in Arkansas through the 1950's could not have
been red wolves.

While true red wolves were being deci-
mated in the Ozark-Ouachita uplands in the
1920's, coyotes probably moved in and in-
creased in numbers, partially filling the va-
cant ecological niche. Surviving red wolves
probably interbred with coyotes and pro-
duced some of the intermediate specimens
discussed in the previous subsection. These
hybrids served as a bridge for the introgres-
sion of genes from C. rufus into C. latrans.
Much of the coyote population was thus
modified, probably in a manner that favored
its continued expansion into a woodland habi-
tat. Of course, lumbering and agricultural
practices may also have assisted the spread
of C. latrans into the south-central states
(McCarley, 1962; Gipson, 1972).

The increase of coyotes in Arkansas, and
undoubtedly also in Oklahoma, was followed
by their large-scale build up in the previously
empty zone centered in northeastern Texas.
Halloran (1959, 1960) provided records that
almost completely block in northeastern
Texas. Although he thought that these rec-
ords depicted the occurrence of black-colored
red wolves, the animals in question were cer-
tainly members of the expanding population
of modified coyotes.

As Halloran explained, the existence of
black animals was thought to indicate the
presence of red wolves, rather than coyotes.
Whereas C. rufus had a locally common black
phase, only one record of a black coyote was
known to Young (1951:52). Halloran (1958)
also used records of black canids in an effort
to plot the distribution of C. rufus in Okla-
homa. In recent years, however, there has
been widespread recognition of the existence
of black coyotes in the south-central United

States. Halloran himself (1963) reported one
from Comanche County, Oklahoma. Dal-
quest ( 1968 ) wrote that a specimen of C.
latrans tcxensis from Wilbarger County,
Texas was in the black color phase. Pimlott
and Joslin (1968) attempted to locate black
animals in Arkansas, hoping that these would
be C. rufus, but the four that they found
were coyotes of medium size. Of the 284
Arkansas skulls analyzed by Gipson (1976),
24 were from black animals. Of these, 12
were identified as coyotes, six as coyote-dog
hybrids, five as coyote-red wolf hybrids, and
one as a domestic dog. Freeman (1976:14)
reported that 12 of 121 Oklahoma specimens
were black or very dark, and that the skulls
of eight of these were identified as coyotes,
two as coyote-dog hybrids, and two as coyote-
red wolf hybrids. Elder and Hayden (1977)
stated that of the seven Missouri specimens
they considered to be red wolves or coyote-
red wolf hybrids, five were black or had been
associated with black animals.

All recent records of black coyotes have
come from the former range of C. rufus (ex-
cept for one reported in Michigan by Ozoga
and Harger, 1966). Although it could be
argued that this phenomenon is the result of
preservation of a newly favorable mutation,
I consider it further evidence of the recent
introgrcssion of genes from C. rufus into
C. latrans.

The coyotelike population apparently in-
creased rapidly in the 1950's. McCarley
( 1959 ) reported C. latrans to be common in
most areas of east Texas. He said also that
the red wolf was either extirpated or ex-
tremely rare in the area. McCarley ( 1962 )
reported data on 110 skulls collected since
1948 in east Texas, eastern and central Okla-
homa, and Arkansas. Observing that over-all
skull size, as represented by greatest length
and zygomatic width, was the only consistent
character separating red wolves and coyotes,
he referred all but one of the specimens to
C. latrans frustror (the exception was a skull
from an unknown Arkansas locality, assigned

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

55

Fig. 31. — Map showing localities of specimens taken from 1961 to 1971 in the south-central United
States. Triangles, C. rufus; squares and open circles, specimens apparently representing hybridization between
C. rufus and C. latrans (see Figs. 37-39); black dots, specimens of the modified population of C. latrans (see
Figs. 32-36). Note: because of the scale of the map, it was not possible to plot all localities in crowded
areas, especially along the Texas coast.

to C. r. rufus). McCarley concluded that coy-
otes had replaced red wolves in Oklahoma,
Arkansas, and east Texas, and that hybridiza-
tion between the two may have occurred.
He thought that C. rufus was extant only in
a few isolated parts of eastern and southern
Louisiana, on the basis of two skulls obtained
there in 1956 and 1957 (see p. 45).

Evidently red wolves were present over
much of Louisiana through the 1940's (Low-
ery, 1943; Nowak, 1967; St. Amant, 1959).
By the early 1950's, however, their numbers
had been greatly reduced by government
trapping and private hunting. Their last
major concentrations were in the bottom lands

along the Mississippi River in the eastern
part of the state, and in the southern coastal
marshes and prairies. For a brief period, the
northwestern and north-central parts of Lou-
isiana were apparently left without a sig-
nificant population of wild Canis, and, in
effect, these areas became an extension of the
"empty zone" in adjacent northeastern Texas.
But just after the heaviest period of wolf
trapping, from 1947 to 1952, coyotelike ani-
mals were reported to be moving into the
northwestern part of Louisiana. These canids
subsequently spread over most of northern
and central Louisiana, and continued to in-

56

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

crease in numbers each year (Nowak, 1967,
1970; Paradiso, 1966; Wilson, 1967).

McCarley's (1959, 1962) reports were the
first published indications that the species
C. nifus might be in serious trouble through-
out its entire range. Until then, there had
been a general belief that red wolves were
common over large areas. Actually, Goldman
(1937) had written that both the subspecies
floridanus and rufiis might be extinct, but
subsequently ( 1944 ) he reported the latter to
be present in parts of central and southern
Texas. Young (1946:43) thought that red
wolves could survive indefinitely in large
areas of Texas and Louisiana where they
would be in no direct conflict with man.
Later Federal reports told of large and even
increasing numbers of red wolves in Texas,
Arkansas, and Oklahoma. Obviously such re-
ports were based on the movement of coyotes
into areas formerly occupied by C. rufus, but
as late as 1964 some authorities considered
the species to be in no danger (see Nowak,
1967, 1970, 1972). By that year, however,
the precarious status of the red wolf was
generally recognized, and efforts were under-
way to evaluate the threats both from man
and from interbreeding.

Examination of Post- 1960 Material from

Oklahoma, Arkansas, Louisiana, and

Inland East Texas

The disappearance of C. rufus from most
of its former range, and its partial replace-
ment by other canids, motivated the collect-
ing of many new specimens. Nearly all of
this material was taken in the normal course
of predator control work by Federal and state
employees. For convenience, the following
discussion is arranged on a partly arbitrary
geographic basis. Localities of all specimens
discussed in this subsection (that were ex-
amined by me) are plotted on the map in
figure 31, and listed in appendix A (parts
28-32).

Oklahoma. — Skulls of 12(4) males and
13(3) females, taken in 1965 in southeastern

Oklahoma, all fall within or near the statisti-
cal limits of standard C. latrans (Fig. 32).
These specimens were obtained in the same
area from which a large series of C. rufus
had been taken prior to 1930. After a multi-
variate analysis of 138 skulls taken in 1975
and 1976, and 114 collected prior to 1975,
mostly from 1953 to 1970, Freeman (1976:
13-14, 28, 33, 47-48) concluded that the cur-
rent Oklahoma population of wild Canis was
essentially coyotelike. Of his total of 252
specimens, 203 were identified as C. latrans,
one as C. familiaris, 33 as coy-dogs, and 15
as intermediate to C. latrans and C. rufus.
There was no significant difference between
the older and newer groups, except in the
southeastern part of the state where skulls
taken in 1975 and 1976 were found to be
smaller. Freeman suggested, and I agree,
that the genetic influence of the red wolf in
eastern Oklahoma had declined to the point

I- 8

LATRANS

RUFUS

-3 -

-1 1 1 1~-

2 3 4 5

-4 -3 -2

0

Fie. 32. — Multivariate positions (black dots) of
individual post- 1960 specimens from southeastern
Oklahoma, relative to the ranges of variation of the
series of C. rufus and C. latrans shown in Fig. 17.

1979

NOWAK: NORTH AMERICAN QUATERNARY CAMS

57

at which only an occasional individual ex-
hibited any characters of C. rufus.

Northern Arkansas. — Statistical positions
of 20 males taken in 1969 and 1970 in the
Ozark area of northern Arkansas (Conway,
Franklin, Newton, Pope, and Van Buren
counties ) , also were found to lie mainly with-
in the range of variation of C. latrans (Fig.
33). Two specimens, however, have D2 val-
ues slightly closer to C. rufus, thus indicating
the continued genetic influence of this spe-
cies. No female specimens from northern
Arkansas were tested by multivariate analysis,
but 20 skulls from the area were examined,
and most were indistinguishable from those
of western coyotes. None exceeded 200 milli-
meters in greatest length or 100 millimeters
in zygomatic width. Gipson, Sealander, and
Dunn (1974) found small "pockets of red
wolf influence" in the Ozark and Ouachita
mountains, but did not refer any particular
specimens from those areas to C. rufus. Pirn-

il- 8

LATRANS

RUFUS

"l 1 1 1 r~

2 3 4 5 6

Fig. 33. — Multivariate positions (black dots) of
individual post-1960 specimens from northern Ar-
kansas, relative to series of C. rufus and C. latrans.

8

LATRANS

RUFUS

~i 1 r

2 3 4

Fig. 34. — Multivariate positions (black dots) of
individual post-1960 specimens from southern Ar-
kansas, relative to series of C. rufus and C. latrans.

lott and Joslin (1968) heard what they con-
sidered probably to be wolves in the Ozark
National Forest in 1964, but found only coy-
otes there in 1965. Subsequently, a group of
investigators from Arkansas Polytechnic Col-
lege continued attempts to locate red wolves
in the Ozarks, both through examination of
skulls and elicitation of howling responses
(see Nowak, 1970). Although some wolf like
calls were heard, no conclusive evidence of
the presence of C. rufus was obtained (Henri
D. Crawley, pers. comm. ). There is little
doubt that true red wolves have now disap-
peared throughout the Ozark-Ouachita up-
lands.

Southern Arkansas. — The multivariate po-
sitions of skulls of 52(6) males and 11(3)
females, taken from 1964 to 1970 in southern
Arkansas, are depicted in figure 34. Most of
these fall within or near the range of varia-
tion of C. latrans, but there is a pronounced
over-all shift toward the standard red wolf

58

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

sample. Four individuals have D2 values
nearer to C. rufus than to C. latrans. An ad-
ditional 20 skulls of females from the area,
examined, but not tested by multivariate anal-
ysis, appear to follow the same pattern. Gip-
son, Sealander, and Dunn (1974) reported
that two specimens from southern Arkansas
were not significantly different from C. rufus,
and that there was a strong genetic influence
from the red wolf in the area. They added,
however, that the red wolf, in pure form,
probably no longer existed in Arkansas, and
that its genes had been incorporated into a
predominantly coyote population. Of the 284
skulls they subjected to multivariate analysis,
only these two were classified as red wolves,
208 were identified as coyotes, 27 were said
to be intermediate to C. rufus and C. latrans,
and the remainder were considered to repre-
sent wild dogs or dog hybrids. The fact that
Arkansas Canis is more wolflike in the south-
ern part of the state may reflect longer sur-
vival of C. rufus there, and hence more recent
introgression from individuals of that species.
Gipson (1972:50-51) suggested that a "zone
of red wolf influence" in the south-central
part of the state might have resulted from the
establishment of an unofficial refuge in that
area, in which red wolves were released in
the late 1950's.

Louisiana.— Skulls of 22(17) males and
19(17) females, taken from 1963 to 1969 in
Louisiana, demonstrate similar statistical dis-
tributions to those formed by the southern
Arkansas material (Fig. 35). Goertz, Fitz-
gerald, and Nowak (1975) concluded that
155 skulls collected in Louisiana from 1963
to 1973 represented an essentially coyotelike
population that had been slightly modified
through introduction of genes from C. rufus.
Elements of this same population apparently
now have spread across Mississippi and into
Alabama (Cahalane, 1964; Paradiso, 1966;
Linzey, 1971; Wolfe, 1972). Recent informa-
tion indicates that coyotelike animals are now
also established in southwestern and south-
central Tennessee ( D. W. Yambert, Tennessee

V 8

LATRANS •

RUFUS

0

-I
-2

-3 -

$

-4 -3 -2

0

3 4

n r-

5 6

Fig. 35. — Multivariate positions (black dots) of
individual post-1960 specimens from Louisiana, rela-
tive to series of C. rufus and C. latrans. The triangle
shows the position of a specimen of C. rufus gregoryi
collected in 1964 in St. Landry Parish.

Wildlife Resources Agency, pers. comm.),
and in southwestern Georgia (Daniel W.
Speake, Cooperative Wildlife Research Unit,
Auburn University, pers. comm.). Red wolves
seem to have been extirpated throughout this
region, except possibly in southern Louisiana.
In 1964 Pimlott and Joslin (1968) heard
small groups of wolves in the northeastern
part of the state and in adjoining sections of
Mississippi, but there have been no subse-
quent records from that area. Persistent re-
ports of red wolves have come from the
coastal marshes in Cameron and Vermilion
parishes, but I have examined no specimens
from this area. In 1976, personnel of the
U.S. Fish and Wildlife Service carried out
field studies in Cameron Parish at the ex-
treme southwestern corner of Louisiana, and
live-captured several animals that appeared
to be red wolves. Coyotelike individuals,

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

59

however, also were reported from this area.
A skull, obtained in 1964 near Washington,
St. Landry Parish, south-central Louisiana,
appears distinct from other recent material
taken in the state. According to Pimlott and
Joslin (1968:383), this specimen was identi-
fied as C. rufus gregoryi by Barbara Law-
rence. Although the skull has several dental
anomalies, its multivariate position (Fig. 35)
supports this identification. Red wolves have
received legal protection in Louisiana since
July 1970 (Nowak, 1971).

Inland East Texas. — Figure 36 shows the
statistical positions of 77(6) males and 42(6)
females, collected from 1964 to 1971 in coun-
ties of east Texas more than 100 miles inland.
Paradiso ( 1968 ) examined most of this ma-
terial earlier, and reported that it represented
an interbreeding population bridging the size
gap between red wolves and coyotes. Para-
diso, however, had statistically combined the
inland material with a number of specimens

-3 -

s

LATRANS

RUFUS

-3 -2

0

3 4

Fig. 36. — Multivariate positions (black dots) of
individual post-1960 specimens from inland east
Texas, relative to series of C. rufus and C. latrans.

taken along the Gulf Coast. Later, Paradiso
and Nowak ( 1972a ) , also having combined
inland and some coastal samples, came to
about the same conclusion. They thought
that the hybrid swarm, established at the turn
of the century in central Texas, had expanded
eastward, and by the 1960's had engulfed
most of east Texas and adjacent parts of
Louisiana and Arkansas. This hypothesis no
longer seems fully tenable. It was pointed
out above (pp. 52-53) that wild Canis in
central Texas had been nearly extirpated by
the 1930's, a period prior to the build up of
Canis in east Texas. Thus the population of
the former area could hardly have been the
source for that of the latter. Furthermore,
when intensively analyzed, recent east Texas
material does not show exactly the same kind
of statistical distribution as that of the earlier
specimens. The present population is pre-
dominantly coyotelike, and does not form an
even distribution bridging the gap between
C. rufus and C. latrans (compare Figs. 23
and 36). Two specimens do fall within the
range of variation of standard C. rufus, a male
from Lamar County and a female from Ham-
ilton County. There is no apparent correla-
tion between the geographical and statistical
distributions of the inland east Texas material.
More wolflike and more coyotelike specimens
occur throughout the area, sometimes to-
gether at the same locality (see also appendix
B, part 6).

The large samples from inland east Texas,
Louisiana, and southern Arkansas all have
similar multivariate distributions, and appar-
ently represent a single population with a
common origin. The region occupied by this
population corresponds in large part to the
zone in which Canis was rare or absent over
much of the first half of the century. Its geo-
graphical distribution also lies mainly along
the southern fringe of the area in which speci-
mens of hybrid characters appeared in the
1930's and 1940's. Such hybridization led to
the introgression of C. latrans, and probably

60

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

also to the genetic swamping of any red
wolves in the area that had survived man's
onslaught. After having been modified
through introgression from C. rufus, the coy-
ote population expanded southward and
eastward across the south-central states. Ad-
ditional hybridization probably occurred as
remnant pockets of red wolves were encoun-
tered. I do not think, however, that the cur-
rent population of wild Canis in this region
should be spoken of as a "hybrid swarm."
The specimens demonstrate a much greater
affinity with C. latrans than with C. rufus or
any other species, and may be properly re-
ferred to as coyotes. The subspecies consid-
ered by Jackson (1951:271) to inhabit the
most proximal geographic area was C. latrans
frustror. His description of this subspecies
as the "largest coyote" was based in part on
the examination of specimens that probably
represented introgression from C. rufus.

Examination of Post-1960 Material
from Coastal Texas

The situation within 100 miles of the Texas
Gulf Coast is markedly different from that
found farther inland. Red wolves survived
in the southeastern corner of Texas, east of
the Rrazos River, through the 1930's and
1940's (pp. 45-46). Some skulls taken farther
south along the coast also seem to resemble
those of C. rufus. Halloran (1961) reported
that five supposed red wolves were collected
on the Aransas National Wildlife Refuge in
1956, but the listed weights were much less
than those of specimens that he said had
been taken in 1939-1940. Davis (1966:112-
113) depicted the recent range of C. rufus
as extending all along the Gulf Coast from
the Louisiana border to Raffin Bay, but he
observed that the species was "on the verge
of extinction." Beezley (1967) wrote that
red wolves were present on the coastal prai-
ries from Jefferson to San Patricio counties,
and farther inland in Harris, Wharton, Colo-
rado, Lavaca, and Victoria counties.

Once again, my analysis of material is

divided on a partly arbitrary basis. The lo-
calities, many known only to county, are
plotted in figure 31 and listed in appendix A
(parts 33-38).

Kenedy County. — One of the most inter-
esting of the coastal specimens was taken in
December 1961 near Armstrong, Kenedy
County, not far north of the Mexican border.
The skin and skeleton were reported as C.
rufus by Paradiso ( 1965), but with no further
comment. The statistical position of the large
skull falls well within the limits of the stand-
ard red wolf sample (Fig. 37). And yet the
specimen was taken in an area beyond the
southern edge of the range assigned by Gold-
man (1944:487) to C. r. rufus. Other ma-
terial, taken south of Nueces Bay in an earlier
period, does not approach the Kenedy County
skull in size or other characters, and appears

Fie. 37. — Multivariate positions (squares) of in-
dividual post- 1960 specimens from the central Gulf
Coast of Texas, relative to series of C. rufus and
C. latrans. The triangle shows the position of an
apparent specimen of C. rufus collected in 1961 in
Kenedy County, extreme southern Texas.

1979

NOWAK: NORTH AMERICAN QUATERNARY CAMS

61

to represent only C. latrans (see Fig. 20).
No additional recent specimens from the area
are available, but further investigation may
yet reveal evidence of the presence of a
small, isolated red wolf population.

Central Gulf Coast. — Figure 37 shows the
relative positions of skulls of 24(7) males and
18(5) females, taken from 1962 to 1969 in an
area including Calhoun, Victoria, Matagorda,
Lavaca, Colorado, Austin, and Fort Rend
counties. No character clines are apparent
in this area, and several counties were the
source of both wolflike and coyotelike speci-
mens. Most fall between the extreme limits
of C. rufus and C. latrans, and their statistical
distribution is similar to that shown by the
former hybrid swarm of central Texas (Fig.
23). Apparently, significant numbers of red
wolves survived on the central coast of Texas
longer than farther inland. Coyotes have now
occupied the same area and hybridized with
the wolves, and there is no way of separating
the two parent species. Specimens taken in
the 1930's and 1940's (Fig. 26) indicate that
coyotes were then already beginning to pre-
dominate, but evidently red wolf influence
persisted. Possibly, small groups of relatively
unmodified C. rufus still are present in the
area, but it would be difficult to pinpoint
their locations.

Harris County. — The vicinity of the Ad-
dicks Reservoir, just west of Houston, was the

RUFUS

RUFUS

Fig. 38. — Multivariate positions (squares) of in-
dividual post-1960 specimens from Harris County,
Texas, relative to series of C. rufus and C. latrans.

Fig. 39. — Multivariate positions (squares) of in-
dividual post-1960 specimens from western Brazoria
County, Texas, relative to series of C. rufus and C.
latrans.

source of 12(2) males taken from 1964 to
1971. Seven of these have a smaller D2 sepa-
ration from C. rufus than from C. latrans.
The over-all statistical distribution (Fig. 38)
suggests that the coastal hybrid population is
established in the area, but that red wolf in-
fluence still is strong. On the basis of investi-
gation of the ecology, behavior, morphology,
and allelic frequency of canids taken mainly
in southeastern Harris County, Shaw (1975:
95) suggested that the population there rep-
resented hybridization between C. rufus and
C. latrans.

Western Brazoria County. — From 1968 to
1971, 17(13) males and 14(10) females were
collected in the vicinity of the Clemens Prison
Farm, south of the town of Rrazoria and west
of the Brazos River. Their statistical distribu-
tion (Fig. 39) is indicative of a hybrid popu-
lation bridging the gap between C. rufus and
C. latrans.

62

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

Eastern Brazoria County. — Paradiso and
Nowak ( 1972a ) reported that typical C. rufus
rufus still lived in Brazoria County east of the
Brazos River. Russell and Shaw (1971b) lo-
cated red wolves in the same area, through
elicitation of howling responses. A total of
11(5) male and 9(4) female adult skulls,
some of them not available for Paradiso and
Nowak's earlier work, were tested by multi-
variate analysis. All of this material was ob-
tained between 1966 and 1971 in the east-
central part of the county, in the vicinity of
Angleton, Liverpool, and Hopkins Mound.
Statistical positions (Fig. 40) are concen-
trated along the lower limits of the standard
red wolf sample, not unlike those of older
material assigned to the subspecies C. r.
rufus. Four of the specimens, however, ac-
tually have a greater D2 separation from C.
rufus than from C. latrans. While it thus is
evident that gene flow from coyotes has had
an effect in this area, the influence of C. rufus

RUFUS

Fig. 40. — Multivariate positions (triangles) of in-
dividual specimens taken from 1966 to 1971 in east-
ern Brazoria County, Texas, relative to series of C.
rufus and C. latrans.

seems stronger here than farther west on the
Gulf Coast. Several other skulls from eastern
Brazoria County, not suitable for multivariate
analysis, were examined and found to possess
typical red wolf characters. One of these
skulls was 241.0 millimeters in greatest length.
Recent reports from field personnel, however,
suggest that red wolf influence in this area
has declined, and conservation efforts there
have been largely discontinued.

Jasper, Tyler, Liberty, and ivestern Cham-
bers counties. — Red wolves held out for many
years in the area of the Big Thicket to the
northeast of Houston. Currently, however, it
is not likely that the species, in unmodified
form, is present in the area. According to
recent reports, it seems that the expanding
coyote population has occupied the wood-
lands of southeastern Texas in moderate
numbers (Russell and Shaw, 1971b; Glynn
Riley, U.S. Fish and Wildlife Service, pers.
comm.). The status of wild Canis here is not
well understood, and I have examined no
specimens from within the borders of the Big
Thicket itself. The positions of 3(1) males
taken in 1970 on the eastern side of the
Thicket, two near Fred, Tyler County, and
one near New Blox, Jasper County, indicate
that the influence of C. rufus is still strong
in the area (Fig. 41). Skulls of 11 males and
10 females were taken from 1965 to 1971
along the lower edge of the Big Thicket in
southern Liberty and western Chambers
counties. This area contains a mixture of
coastal prairie and woodland habitat. Some
of the specimens are indistinguishable from
red wolves collected many years ago in the
south-central states, but others are coyotelike
or intermediate in characters (Fig. 41). Ap-
parently then, coyotes have spread through
east Texas, and have come into contact with
surviving red wolves on the coastal prairies
where hybridization probably has occurred.

Jefferson and eastern Chambers counties.
— Even while the disappearance of red wolves
over most of their former range was being

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

63

RUFUS

Fig. 41. — Multivariate positions of individual
specimens taken from 1965 to 1971 in Jasper and
Tyler counties, Texas (open circles), and in southern
Liberty and western Chambers counties (squares),
relative to series of C. rufus and C. latrans.

generally acknowledged, there were continu-
ous reports of their survival on the coastal
prairies and marshes of extreme southeastern
Texas between Galveston Bay and Sabine
Lake. Paradiso (1965) reported that seven
large skulls, collected in 1963 or early 1964
on the Anahuac National Wildlife Refuge,
southeastern Chambers County, were unques-
tionably those of C. rufus. Pimlott and Joslin
(1968) wrote that in 1965 they located three
packs of red wolves in the area east of Gal-
veston Bay between Highway 73 and the
Gulf Coast. Paradiso and Nowak (1972a)
considered that specimens collected between
1963 and 1969 demonstrated the survival of
the subspecies C. rufus gregoryi in the area.
For the purposes of this paper, skulls of
15(3) males and four females taken from
1963 to 1970 were tested by multivariate
analysis. All skulls are from animals killed

in eastern Chambers and southern Jefferson
counties, the area between Galveston Bay and
Sabine Lake, and south of the line formed by
U.S. Interstate Highway 10 and Texas High-
way 73. The relative statistical positions of
these specimens (Fig. 42) fall predominantly
within the range of variation of standard C.
rufus, and confirm the presence of the species,
in apparently unmodified form, in extreme
southeastern Texas (see also measurements
in appendix B, part 7).

The suggestion that the red wolf survived
in southeastern Texas until about 1970 was
supported by field studies carried out in that
area. Glynn Riley, an agent of the U.S. Fish
and Wildlife Service, who was locally respon-
sible for red wolf management from 1969 to
1973, said (pers. comm.) that the species was
present in moderate to high numbers in
Chambers and Jefferson counties. In Septem-

A RUFUS

n r-

2 3

-3 -

-3 -2 -I

0

I

3 4

Fig. 42. — Multivariate positions (triangles) of in-
dividual specimens taken from 1963 to 1970 in Jef-
ferson and eastern Chambers counties, extreme south-
eastern Texas, relative to series of C. rufus and C.
latrans.

64

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

ber 1972 he estimated the population south
of Interstate Highway 10 to consist of about
150 individuals. Russell and Shaw (1971b)
wrote that in 1970 they located a dense red
wolf population in southern Chambers and
Jefferson counties, through elicitation of
howling responses. Shaw (1975) reported a
relatively high population density of C. rufus
gregonji in the area during his field work from
June 1971 to September 1972, and found the
animals so identified to have gross morpho-
logical and behavioral characters distinct from
those of C. latrans.

Probably an important factor in the sur-
vival of these wolves was their geographic
location. Far to the south along the Gulf
Coast, they were well removed from the orig-
inal and subsequently expanding coyote pop-
ulation. The temporary existence in eastern
Texas of a large zone in which wild Cam's
was rare or absent (see pp. 52-53), helped
to prevent genetic exchange between coyotes
to the north and the surviving wolves on the
coast. To the west, Galveston Bay and the
Houston metropolitan area limited the spread
of the hybridization process that engulfed
most of the coast.

In my dissertation (Nowak, 1973:146) I
expressed hope that conservation efforts
might help to maintain the unmodified popu-
lation of C. rufus in southeastern Texas. Even
before I wrote those words, however, the
situation probably had deteriorated beyond
the point of practical control. Indeed, there
is good evidence that canids with coyotelike
characteristics were already in Jefferson and
eastern Chambers counties in the 1960's, and
that the skulls of such animals may never
have been submitted for examination, be-
cause persons obtaining specimens thought
that only "wolves" were wanted (Carley and
McCarley, 1976). This bias probably was
not an overriding factor, however, since both
wolflike and coyotelike skulls were received
from many other parts of Texas (see Figs.
36-41). While it is likely that a portion of
the population of wild Canis in Jefferson and

eastern Chambers counties in the 1960's did
not represent unmodified C. rufus, I continue
to think that this portion was small relative
to other areas of coastal Texas.

In any event, the situation does appear to
have worsened in the late 1960's and early
1970's. Conservation efforts by the U.S. Fish
and Wildlife Service, including some trapping
of non-wolflike canids in the vicinity, did not
prevent the continued decline of the red wolf.
By 1974, if not sooner, field personnel recog-
nized that a substantial number of the wild
canids south of Interstate Highway 10 were
not wolves (Curtis J. Carley, Project Leader,
Red Wolf Recovery Program; pers. comm.).
Skulls of 13 males and 12 females, collected
from 1973 to 1975 in the same area as those
represented in figure 42, were evaluated by
multivariate analysis. The statistical positions
of these new specimens are shown in figure

RUFUS

Fig. 4.3. — Multivariate positions (squares) of in-
dividual specimens taken from 1973 to 1975 in Jef-
ferson and eastern Chambers counties, extreme south-
eastern Texas, relative to series of C. rufus and
C. latrans.

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

65

43, and the general pattern resembles that of
material taken just to the north from 1965
to 1971 (Fig. 41). Thus it is evident that
eoyotelike animals have now completely oc-
cupied the former range of C, ruftts in Texas,
and that no red wolf populations (though
probably some individuals) remain free from
significant genetic influence by C. latrans.

When it was recognized that the last red
wolf population was being swamped through
hybridization, the U.S. Fish and Wildlife
Service accelerated a program to live-capture
some individuals that did appear to represent
pure C. nifus, and to place these animals in
a breeding pool at the Point Defiance Zoo at
Tacoma, Washington. In November 1977
this facility contained 29 adults, all wild-
caught (including three from Cameron Par-
ish, Louisiana), and 13 young produced dur-

ing the spring of 1977 (Norman R. Winnick,
Director, Red Wolf Captive Rreeding Pro-
gram, Point Defiance Zoo; pers. comm.). On
13 December 1976 a pair from this group had
been released on Rulls Island, Cape Romain
National Wildlife Refuge, South Carolina.
Both animals were recaptured on 22 Decem-
ber 1976 after one had crossed to the main-
land, and the female subsequently died of
natural causes. Another release was planned
for January 1978. Although Bulls Island is
too small (5,000 acres) to support a viable
pack of wolves, it is hoped that procedures
can be developed there that will aid possible
future reintroductions in other parts of the
original range of C. rufus. At present, how-
ever, the wild wolf populations that once in-
habited the southeastern quarter of North
America are not in existence.

SYSTEMATIC DESCRIPTIONS

ORDER CARNIVORA
FAMILY CANIDAE
Genus Canis Linnaeus

1758. Canis Linnaeus, Systema Naturae, 10th
ed., p. 38. Type, Canis familiaris Lin-
naeus.

1816. Thos Oken, Oken's Lehrbuch der Na-
turgeschichte, pt. 3 (Zoologie), sect.
2, p. 1037. Type, Thos vulgaris Oken
(=Canis aureus Linnaeus).

1839. Lyciscus Hamilton-Smith, in The nat-
uralist's library (edit. Jardine), 25:160.
Type, Canis latrans Say.

1918. Aenocyon Merriam, Univ. California
Publ. Bull. Dept. Geol.. 10:532. 20
April. Type, Canis dims Leidy, subse-
quent designation by Hay, Second bib-
liography and catalogue of the fossil
Vertebrata of North America, 2:501,
27 January 1930.

Geological distribution. — Middle Pliocene
to Recent.

Geographical distribution. — In historical
time, the species of Canis have been distrib-
uted as follows:

Canis familiaris Linnaeus, 1758 (including
C. dingo Blumenbach, 1780, of Australia), the
domestic dog, throughout the world, usually
in association with man;

Canis lupus Linneaus, 1758, the gray or
timber wolf, in most of North America and
Eurasia, and on associated northern islands;

Canis rufus Audubon and Bachman, 1851,
the red or southern wolf, in southeastern
North America;

Canis latrans Say, 1823, the coyote or prai-
rie wolf, primarily in western North America;

Canis aureus Linnaeus, 1758 (including
C. lupaster Hcmprich and Ehrenberg, 1832,
of northern Africa), the golden jackal, from
central Africa to India and southeastern Eu-
rope;

Canis adustus Sundevall, 1846, the side-

striped jackal, in eastern, southern, and west-
ern Africa;

Canis mesomelas Schreber, 1778, the
black-backed jackal, in eastern and southern
Africa; and

Canis simensis Ruppell, 1835, the Abys-
sinian wolf, in the mountains of Ethiopia.

If most species assigned to the genus
Dusicyon are actually referable to Canis, as
suggested by Langguth ( 1975 ) , the Recent
geographical distribution of Canis would in-
clude much of South America. Pleistocene
fossils of Canis have been reported from
South America, as well as from North Amer-
ica, Eurasia, and Africa.

Relationship with other Recent genera. —
The Recent Canidae often have been divided
into three subfamilies: Caninae, with the
genera Canis, Alopcx, Fennecus, Vulpes,
Urocyon, Dusicyon. Chrysocyon, Atelocynus,
Cerdocyon, and Nyctereutes; Simocyoninae,
with the genera Cuon, Lycaon, and Speothos;
and Otocyoninae, with the genus Otocyon.
Van Valen (1964) suggested that Otocyon
does not warrant subfamilial distinction from
the Caninae, and most workers seem to be
following this viewpoint. Clutton-Brock, Cor-
bett, and Hills (1976) proposed abandoning
the use of all subfamily names for the Recent
Canidae. These authors also placed Urocyon
and Fennecus within the genus Vulpes, and,
like Osgood (1934) and Simpson (1945:109),
included Cerdocyon and Atelocynus in the
genus Dusicyon. Langguth (1975), however,
retained the genera Cerdocyon and Atelocy-
nus, put all but one of the species of Du-
sicyon in the genus Canis, and placed the
remaining species (Dusicyon vetulus) in the
genus Lycalopcx.

Osgood (1934) prepared a systematic key
in an attempt to demonstrate the common
affinity of all South American Caninae and
their collective distinction from Canis. The
South American genera (including Urocyon)

66

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

67

were said to be characterized by a reduced
upper carnassial, with a length always less
than the combined length of the two upper
molars; an occipital shield depressed at the
apex; a usually flattened or lyrate sagittal
crest (except in Chrysocyon); relatively small
upper incisors, the inner pair imperfectly or
not trifid; and long, slender canine teeth.
These characters, however, may be only rela-
tive expressions of trends found throughout
the Caninae. For example, in some North
American coyotes the saggital crest is flat-
tened and lyrate, and P4 is shorter than the
combined length of Ml and M2.

Study of canid karyotypes thus far seems
to support the composition of the genus Canis
as set forth above under "geographical dis-
tribution." Each species of Canis that has
been analyzed (C. familiaris, C. dingo, C.
lupus, C. rufus, C. latrans, and C. aureus) has
a diploid chromosome number of 78. Species
of other genera within the Caninae, including
Vulpes, Fennecus, Alopex, Urocyon, Chryso-
cyon, Nyctereutes, and Otocyon, have been
found to have diploid numbers different from
that of Canis. The species Dusicyon vetulus
(placed in the genus Lycalopex by Lang-
guth, 1975) has a diploid number of 76
(Wurster and Benirschke, 1968; Chiarelli,
1975).

The production of hybrids provides fur-
ther evidence of the validity of the presently
accepted structure of the genus Canis. Viable
hybrids have been reported between C. fa-
miliaris and C. lupus, C. familiaris and C.
latrans, C. familiaris and C. aureus, C. famili-
aris and C. dingo, C. latrans and C. aureus,
C. latrans and C. lupus, C. lupus and C.
dingo (Gray, 1972:45-51), and C. latrans and
C. rufus (this paper). There are few data on
interbreeding between members of Canis and
other recognized genera, and no evidence that
viable hybrids ever have been produced (see
Chiarelli, 1975:50).

Remarks. — Determination of the actual
origin of the genus Canis is not within the
scope of this paper. One widely held view

is that the genus Tomarctus of the Miocene
and early Pliocene is near the ancestral line
of Canis and other modern Caninae. Simpson
(1945:109, 222) placed Tomarctus within the
Caninae, but noted that it might represent
the ancestry of another subfamily. Matthew
(1930:131) and Vanderhoof and Gregory
(1940:145) considered Tomarctus to occupy
a central position from which both the various
borophagine dogs and the modern canines
arose. Green (1948) concluded that certain
species of Tomarctus were in the direct line
of the borophagines, whereas other species
probably gave rise to Canis and possibly to
Vulpes. Williams (1962) referred Tomarctus
to the Borophaginae and placed all but one
of its named species in the genus Aelurodon.
He suggested that some lesser known canine
of the late Miocene or early Pliocene, such as
Leptocyon vafer (Leidy), was the ancestor of
the Quaternary Caninae. Presumably then,
according to this interpretation, the lineage
leading to the living Caninae has been dis-
tinct from other subfamilial groupings at least
since the late Miocene, and Canis has de-
scended from small, foxlike ancestors. At pres-
ent, the number of reported Pliocene canines
is small, and the particular point at which
Canis arose has not been established.

It is true that many North American speci-
mens from the Pliocene and even earlier
times, some of them representing animals
larger than a modern wolf, were named as
species of Canis. Nearly all of these names,
however, now have been referred to various
other genera, and they seem to have no bear-
ing on the lineage of Canis (Hay, 1902:769-
776, and 1929-1930:488-512 may be consulted
for these named kinds and their places of
description). Most recently, the species Canis
texanus Troxell, 1915 was referred to the
genus Protocyon by J. L. Kraglievich (1952:
621).

Only three Pliocene species of Canis are
now known from North America. These in-
clude C. davisii from southwestern Nevada
(Merriam, 1911:242-243) and southeastern

68

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

Oregon (Shotwell, 1970:70), and C. condoni
from northeastern Oregon (Shotwell, 1956:
733). Shotwell (1970) noted that additional
material might prove these two species to be
synonymous. Canis lepophagus, discussed in
detail below, is known from a number of late
Pliocene and early Pleistocene sites.

Of the various specific names applied to
North American Pleistocene fossils of Canis,
the following 17 have not hitherto been
placed in other genera or formally reduced
to synonyms or subspecies:

C. cedazoensis Mooser and Dalquest,

1975*
C. lepophagus Johnston, 1938°
C. latrans Say, 1823°
C. andersoni Merriam, 1910
C. riviveronis Hay, 1917
C. caneloensis Skinner, 1942
C. irvingtonensis Savage, 1951
C. edwardii Gazin, 1942°
C. rufus Audubon and Bachman, 1851°
C. priscolatrans Cope, 1899
C. armbrusteri Gidley, 1913°
C. lupus Linnaeus, 1758°
C. milled Merriam, 1912
C. familiaris Linnaeus, 1758°
C. petrolei Stock, 1938
C. dims Leidy, 1858°
C. ayersi Sellards, 1916.

The status of all of these names is dis-
cussed in the following pages. Each of those
marked above with an asterisk ( ° ) is main-
tained as a specific name in this paper.

Canis cedazoensis Mooser and Dalquest

1975. Canis cedazoensis Mooser and Dal-
quest, Jour. Mamm., 56:787.

Holotype. — Right maxillary fragment con-
taining P3 to Ml and the alveolus of M2; no.
9780, Midwestern State University Depart-
ment of Biology; Arroyo Cedazo, 3 kilometers
SE City of Aguascalientes, Aguascalientes,
Mexico.

Geological distribution. — Early Ranchola-
brean.

Geographical distribution. — Known only
from the type locality.

Description. — A small canid, larger than
any North American fox, but smaller than C.
latrans; Ml relatively small, with pronounced
buccal cingulum (for more detail, see Mooser
and Dalquest, 1975:787-788).

Comparison with C. latrans. — Smaller; Ml
with relatively smaller medial section.

Remarks. — The single available specimen
is too fragmentary, and the teeth too heavily
worn, to allow the same kind of account given
to other species in this paper. In general I
am in accord with the description and com-
parisons provided by Mooser and Dalquest
(1975:787-788), and with their conclusion
that the specimen does not represent C. la-
trans.

Record of occurrences. — Type locality;
early Rancholabrean (probably Illinoian);
specimen examined: holotype, Midwestern
State University Department of Biology 9780
(formerly no. FC 634 in collection of O.
Mooser); measurement in appendix B (part
8).

Evolutionary position. — On the basis of
the one available specimen, the relationships
of C. cedazoensis can not be carefully as-
sessed. The species might represent an aber-
rant line that separated from C. latrans in the
Irvingtonian, or it could be a surviving ele-
ment of one of the smaller-sized populations
of the Blancan C. lepophagus.

Canis lepophagus Johnston

1938. Canis lepophagus Johnston, Amer.
Jour. Sci., ser. 5, 35:383.

Holotype. — Skull without mandibles; no.
W. T. 881, Panhandle Plains Museum;
stratum no. 2, Harold Ranch, North Cita Can-
yon, center of west half of sec. 164, block 6,
Randall County, Texas.

Geological distribution. — Blancan.

Geographical distribution. — Known from
California, Florida, Idaho, Kansas, Nebraska,
and Texas.

Description. — A small coyotelike canid;

1979

NOWAK: NORTH AMERICAN QUATERNARY CAN1S

skull small with mostly narrow proportions;
rostrum elongated and narrow; braincase rela-
tively small and little inflated dorsoposte-
riorly; frontal and supraoccipital shields rela-
tively broad; sagittal crest prominent; mandi-
ble long and narrow; premolars and molars
with trenchant, laterally compressed cusps;
P4 with prominent deuterocone; Ml with
pronounced buccal cingulum; p4 with promi-
nent second cusp.

Comparison with C. latrans. — Skull aver-
aging smaller, but overlapping in most meas-
ured dimensions; braincase relatively smaller
and less inflated dorsoposteriorly; frontal and
supraoccipital shields relatively broader; sa-
gittal crest more prominent; some mandibles
relatively deeper; dentition usually more
crowded and with more trenchant, laterally
compressed cusps; metaconule on Ml less
separated from protocone; p4 usually with
more prominent second cusp, and with third
cusp and posteromedial cingulum more re-
duced (if they are present); anterior margin
of ml usually more nearly vertical. Bjork
(1970:14) and Kurten (1974) discussed a
number of postcranial differences between
the two species.

Other comparison. — See account of C.
edwardii.

Remarks.— Johnston (1938) wrote that C.
lepophagus differed from C. latrans in having
more prominent sagittal and lambdoidal
crests, less expanded braincase, deeper man-
dible, and lesser distance between premolars.
Subsequently, other workers assigned addi-
tional specimens to C. lepophagus, but not
all of this material shared the characters cited
by Johnston. Hibbard (1938:243) wrote that
a specimen from the Rexroad fauna, which
he later ( 1941a: 268-269) referred to C. le-
pophagus, represented a "small dog with a
light tapering ramus." Hibbard (1941a)
stated that C. lepophagus from this fauna
was characterized by small size and relatively
narrow teeth. Fine (1964) wrote that a man-
dibular fragment, assigned later by Bjork
(1970:14) to C. lepophagus, was "from a

lightly built coyote having a jaw quite long
for its depth as compared with those of Re-
cent specimens." Bjork (1970) described C.
lepophagus from the Hagerman local fauna
as having a long and slender lower jaw, and
dentition more slender than that of C. latrans.
Giles (1960) subjected six mandibles of C.
lepophagus from Cita Canyon to a multi-
variate analysis involving five measurements.
The results suggested to him that C. lepoph-
agus was subspecifically, but not specifically,
distinct from Recent C. latrans.

According to Kurten (1974:27-28), "Canis
lepophagus of the Blancan differs from the
Recent and Rancholabrean C. latrans in its
shorter distal limb segments, more tapering
snout, relatively larger M2 (this character is
however retained in some Recent popula-
tions), narrower p2-p4, and shorter ml (rela-
tive to its own width, and also to the length
of p4)." He considered all coyotelike ma-
terial from the Blancan, except from the latest
part of this age, to represent C. lepophagus,
and that this species formed a stage in the
evolutionary line leading to C. latrans. Ma-
terial from most of the localities listed below
under "record of occurrences," was referred
by Kurten to C. lepophagus.

I recognize most material previously as-
signed by others to C. lepophagus as repre-
senting a single variable species, separate
from and ancestral to C. latrans. Hibbard
(1941b) thought that C. lepophagus had as
large a geographic range as Recent C. latrans.
Considering this likelihood, plus the long
period of time through which material of C.
lepophagus is distributed, considerable varia-
tion would be expected. The three skulls
from Cita Canyon reported by Johnston
(1938), plus one other obtained at the same
site, are the only specimens of C. lepophagus
upon which meaningful measurements can be
made of elements other than the teeth and
lower jaws. And, as listed above, there are
differences between these skulls and those of
C. latrans. The mandibles of C. lepophagus
from Cita Canyon overlap in size and pro-

70

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

portion those of Recent C. latrans (Fig. 44),
but average deeper. Most specimens from
other sites are mandibles, some of them rela-
tively shallow. Differentiating characters of
the lower dentition, as listed in the above

88

-

o

87

-

•

o o

86

• o o o
• • o

85

•
• o s O

84

o.o o0 o
• ,oo • s •

83

o • o c
• ° oc» -° o °

82

O O C o

81

80

# b„ •o«c» o

79

••

78

•• •• • *o

•

• • • c

77

•

••

•

•• •

76

•
•

c • •

75

• • •

74

• .. c

73

.•

•

72

•
s

71

s s •

i

14

15 16 17 18 19 20 21 22

Fig. 44. — Scatter diagram comparing length and
depth of mandible in certain specimens of C. le-
pophagus and C. latrans. The vertical axis indicates
measurement of the distance from the anterior edge
of the alveolus of pi to the posterior edge of the
alveolus of m3. The horizontal a\is indicates meas-
urement of the minimum depth from the dorsal sur-
face of the mandible between p3 and p4 to the
ventral surface of the mandible. Black dots, Recent
C. latrans testes; open circles, Pleistocene C. latrans
orcutti; c, C. lepophagus from Cita Canyon; s, C.
lepophagus from other sites; v, C. latrans irvingtonen-
sis.

"comparison with C. latrans," are present in
most of the Cita Canyon mandibles and in
most of the other specimens that can be eval-
uated. These characters are matched in
some individuals of C. latrans, but their pres-
ence in specimens of C. lepophagus from
several major sites suggests affinity among
the populations represented.

The type locality of C. lepophagus, Cita
Canyon, now is recognized to contain an
Aftonian (Hibbard, 1970:414) or late Blancan
(Kurten, 1974:5) fauna, and thus is among
the youngest sites from which the species
has been reported. Since no specimens from
contemporary sites are complete enough for
definite assignment to C. latrans, but since
some do have characters typical of C. lepoph-
agus, all coyotelike material from the Blancan
is referred below to C. lepophagus.

Record of occurrences. — The following
list is arranged alphabetically by state, and
geographically (north to south, west to east)
within states. Specimens examined by me
are identified by element, museum number,
or both; and selected measurements are found
in appendix B (part 9) and appendix C (part
1). Occurrences also are shown on the map
in figure 45.

CALIFORNIA.— Sacramento Valley, Sacramento
River, Tehama County; late Pliocene; as Canis (Van-
derhoof, 1933:383); ml, UCMP 29828.

FLORIDA.— Santa Fe River IB, Gilchrist County;
Blancan (Nebraskan); as C. cf. lepophagus (Webb,
1974b: 17); five mandibular fragments, UF 10423,
10424, 10836, 10837, 10858. Two of these speci-
mens are comparatively large and deep, and sug-
gestive of the Cita Canyon material.

IDAHO. — Grand View, Owyhee County; late
Blancan (Hibbard, et al., 1965); as "Canis n. sp."
(Schultz, 1938a:297); two mandibular fragments,
LACM 118-1246, 1343. The specimens are com-
paratively large and deep; the p4 has a prominent
second cusp, but no third cusp or posteromedial
cingulum.

Hagerman local fauna, Twin Falls County; early
Blancan (late Pliocene); as C. lepophagus (Bjork,
1970:13-16); four maxillary fragments, USNM
25136, UMMP V52280, V54995, V56401; seven man-
dibular fragments, USNM 25131, UMMP V45222,
V50249, V53519, V53817, V53910, V56282; two
Ml, UMMP V50000, V56034; Ml fragment, UMMP
V57016; p2, UMMP V56809; two p3, UMMP
V50008, V51052; p4 fragment, UMMP V53817; ml,

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

71

Fig. 45. — Map showing localities of C. lepophagus (black dots) and C. armbrusteri (triangles). Be-
cause of the scale of the map, it was not possible to plot all localities in crowded areas.

USNM 25132; ml fragment, UMMP V52757; post-
cranial elements, UMMP. As explained by Bjork
(1970), the Hagerman material indicates an animal
smaller than the average Recent C. latrans.

KANSAS. — Rexroad fauna, Meade County State
Park, Meade County; early Blancan (late Pliocene);

as C. lepophagus and Canis sp. ( Hibbard, 1941a:
268-269); three mandibular fragments, KU 3914,
4602, 4603; Ml, m2, KU 3915; P4, Ml, UMMP
37132. The mandibles are relatively small and shal-
low, and most of the teeth are small with trenchant
cusps. Hibbard wrote that large size distinguished

72

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

an Ml and m2, found at the site, from C. lepoph-
agus. Although these teeth are larger than the others
in the Rexroad fauna, they are within the size range
of C. lepophagus from other sites.

Rexroad fauna, Keefe Canyon, Meade County;
early Blancan (late Pliocene); as C. lepophagus
(Hibbard and Riggs, 1949:838); ml, KU 7692.

Deer Park local fauna, Meade County; late
Blancan (Aftonian); as Canis sp. (Hibbard, 1956:
172); two ml, UMMP 31945.

NEBRASKA.— Sand Draw, 6 mi. N Ainsworth,
Brown County; late Blancan (Aftonian: Hibbard,
et ah, 1965); as C. cf. latrans ( McGrew, 1944:53),
as C. lepophagus (Skinner and Hibbard, 1972:107).

Broadwater quarry, north side of North Platte
River, Morrill County; late Blancan (Aftonian); as
"Canis sp. (near C. latrans Say)" (Barbour and
Schultz, 1937:4), as "Canis sp. (Coyote)" (Schultz
and Stout, 1948:563); two skulls, UN 26111, 26112;
mandibular fragment, UN 26116; P4, Ml, M2, UN
26113. Assignment herein to C. lepophagus is un-
certain on the basis of morphology alone, as the
cranial parts are in too poor a condition for evalua-
tion of specific characters. A measurement of depth
on the mandible is not possible, but it is relatively
thick lateromedially, and has the massive appearance
attributed to C. lepophagus by Johnston (1938). The
teeth are approximately the same size and have
the same structure as those of the Cita Canyon speci-
mens.

Lisco quarry, north side of North Platte River,
Garden County; late Blancan (Aftonian)-, as "Canis
sp. (Coyote)" (Schultz and Stout, 1948:563); skull,
UN 26107; mandible, UN 26114. This material offers
some evidence for recognizing that skulls of the Cita
Canyon kind, as well as small mandibles from other
sites, represent the same species. The Lisco mandi-
ble is comparatively small and shallow, with trench-
ant, laterally compressed cusps on the teeth, and
with a high second cusp on p4. The skull also is
small, but has a prominent sagittal crest, as do the
Cita Canyon skulls.

TEXAS.— Red Coral, Proctor Ranch, Oldham
County; late Blancan; "within the variation range
of C. lepophagus" (Kurten, 1974:5).

Cita Canyon, Randall County; late Blancan (Af-
tonian: Hibbard, 1970:414); as C. lepophagus
(Johnston, 1938); four skulls, maxillary fragment,
32 mandibles and fragments thereof, various isolated
teeth and postcranial elements, PPM.

Blanco local fauna, near Mount Blanco, Crosby
County; Blancan (latest Pliocene or earliest Pleisto-
cene); as C. lepophagus (Dalquest, 1975:22, 47).

Beck Ranch, Scurry County; early Blancan; as
C. lepophagus (Kurten, 1974:5).

Red Light Bolson, southeastern Hudspeth County;
Blancan (Nebraskan); as C. cf. lepophagus (Aker-
sten, 1970:86; 1972:19-24).

Evolutionary position. — The small size
and the dental characters of some specimens
of C. lepophagus suggest affinity to the Old

World jackals and to the foxes (Vulpes). The
other two recognized species of Pliocene
Canis, C. davisii and C. condoni, also were
described as being foxlike in size and certain
other details. Possibly then, the lineage of
Canis separated from that of Vulpes in the
early or middle Pliocene. Older specimens
of C. lepophagus are small and especially
indicative of descent from foxlike ancestors.
The species C. latrans probably evolved from
within the complex assigned above to C.
lepophagus. Perhaps the shallow-jawed ma-
terial represents the ancestral element of C.
latrans, whereas the more massive Cita Can-
yon specimens represent another line.

In my dissertation (Nowak, 1973:164), I
agreed with Johnston ( 1938 ) in recognizing
C. lepophagus as an ancestral coyote, but
unlike him I did not consider it to be in the
direct phylogenetic line of the modern wolves.
My opinion was based in part on acceptance
of a late Blancan age for the type of C.
edwardii and several other early specimens
of the wolf line. Since the most wolflike
specimens of C. lepophagus, especially those
from Cita Canyon, also were late Blancan in
age, but were morphologically distinct from
the supposedly contemporary wolf material,
they could hardly have been ancestral to C.
edwardii. New information (Johnson, Op-
dyke, and Lindsay, 1975; Kurten, 1974) sug-
gests that the type of C. edwardii, as well as
most other early North American specimens
that apparently represent the wolf line, ac-
tually are Irvingtonian in age. The one ex-
ception is a Blancan specimen from Mifiaca
Mesa, Chihuahua, which consists of a mandi-
ble about the size of that of a small wolf,
but which has certain dental characters re-
sembling those of some specimens of C. le-
pophagus (see account of C. edwardii be-
low). There thus is a possibility that this
mandible represents a transitional phase
through which the wolf line did evolve from
a population of C. lepophagus.

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

73

Canis latrans Say

1823. Canis latrans Say, in Long, Account of
an expedition from Pittsburgh to the
Rocky Mountains performed in the
years 1819 and 1820, 1:168. Type lo-
cality. Engineer Cantonment, about 12
mi. SE present town of Blair, Wash-
ington County, Nebraska, on west bank
of Missouri River.
1910. Canis orcutti Merriam, Univ. Califor-
nia Publ. Bull. Dept. Geol., 5:391.
Type from Rancho La Brea, Los An-
geles County, California. Valid as a
subspecies of C. latrans.
1910. Canis andersoni Merriam, Univ. Cali-
fornia Publ. Bull. Dept. Geol., 5:393.
Type from Rancho La Brea, Los An-
geles County, California. A synonym
of C. latrans orcutti.
1917. Canis riviveronis Hay, Florida State
Geol. Surv. Rept., 9:59. Type from
Vero, Indian River County, Florida.
Valid as a subspecies of C. latrans.
1942. Canis caneloensis Skinner, Bull. Amer.
Mus. Nat. Hist., 80:163. Type from
Papago Springs Cave, Santa Cruz
County, Arizona. Valid as a subspecies
of C. latrans.
1951. Canis irvingtonensis Savage, Univ. Cal-
ifornia Publ. Bull. Dept. Geol. Sci.,
28:231. Type from Irvington, Alameda
County, California. Valid as a sub-
species of C. latrans.
1961. Canis latrans harriscrooki Slaughter,
Jour. Mamm., 42:505. Type from Lew-
isville site, Denton County, Texas.
Valid as a subspecies of C. latrans.

In addition to those listed above, 22 names
based on Recent specimens are available for
use at the subspecific level. These are to be
found in the systematic revision by Jackson
( 1951 ) , and also are listed by Hall and Kel-
son (1959:843-846).

Type. — None designated.

Geological distribution. — Irvingtonian to
Recent.

Geographical distribution. — Pleistocene
and early Recent records from Alberta, Alas-
ka, Arizona, Arkansas, California, Colorado,
Florida, Idaho, Illinois, Indiana, Iowa, Kan-
sas, Maryland, Mississippi, Missouri, Ne-
braska, Nevada, New Mexico, Oklahoma,
Oregon, Pennsylvania, Texas, Utah, Wiscon-
sin, Wyoming, Aguascalientes, Estado de
Mexico, Nuevo Leon, Oaxaca, and Puebla;
original range in historical time included
most of western half of North America and
the plains region as far east as southern
Wisconsin, northwestern Indiana, western Ar-
kansas, and central Texas; see also "'remarks"
below.

Description. — Size small to medium for
the genus; skull usually small with mostly
narrow proportions; rostrum elongated, nar-
row, and shallow; braincase relatively large
and well inflated dorsoposteriorly, often
broader at level of parietotemporal sutures
than at base; postorbital constriction broad
lateromedially and short anteroposteriorly;
zygomata usually slender and not widely
spreading; orbits large; frontals usually only
slightly elevated above rostrum, not promi-
nently convex, and forming relatively narrow
and flat shield; temporal ridges usually not
sharp, seldom obscuring frontal suture, and
sometimes forming lyrate pattern over brain-
case; sagittal crest seldom prominent, often
thickened lateromedially, sometimes flat-
tened; supraoccipital shield comparatively
small; external side of occipital usually with
thin-walled projection just above foramen
magnum; tympanic bullae well inflated; man-
dible long, narrow, and shallow, ventral edge
often rounded; incisors relatively small, up-
per canines prominent, thin anteroposteriorly,
their alveoli set low in premaxillae, their ven-
tral tips normally extending below level of
anterior mental foramina when jaws are
closed; premolars with trenchant, laterally
compressed cusps; P4 with prominent deu-
terocone and lingual cingulum; Ml with rela-
tively large, deeply sculptured medial sec-
tion, prominent metaconule well separated

74

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

from protocone, usually with pronounced
buccal and anterior cingula; M2 relatively
large, cusps well developed; p2 usually lack-
ing posterior cusp: p.3 sometimes with second
and third cusp; p4 with second cusp, and usu-
ally with pronounced third cusp and postero-
medial cingulum extending behind third cusp;
m2 and talonid of ml relatively large, with
trenchant cusps. For details on pelage and
postcranial skeleton, see Jackson ( 1951 ) ;
Grinnell, Dixon, and Linsdale (1937); Hilde-
brand (1952a, 1952b, 1954); and Merriam
(1912).

Comparisons. — See accounts of C. ceda-
zoensis, C. lepophagus, C. rufus, C. lupus,
and C. familiaris.

Remarks. — When the white man first ar-
rived in North America, C. latrans had a wide
distribution, primarily in the western half of
the continent. The exact southern, northern,
and eastern limits of its former range are not
known. Young (1951:29) thought that the
species was originally found only as far as
central Mexico, and that movement farther
to the south occurred after the introduction
of livestock in the region. But Jackson ( 1951)
recognized the presence of three separate
subspecies in Central America, and it is ques-
tionable whether all of these could have come
into existence in only 400 years. Also, accord-
ing to Young ( 1951 ) , the coyote did not be-
come established in northwestern Canada and
Alaska until the nineteenth and twentieth
centuries. But Jackson (1951:265) suspected
that the subspecies C. latrans incolatus was
a very long time resident of Alaska.

Information provided in the first main
part of this paper indicated that the original
range of the coyote extended at least as far
east as southern Wisconsin, northwestern In-
diana, western Arkansas, and central Texas.
Skeletal remains, identified as C. latrans,
have been reported from archeological ex-
cavations in Crawford County, Illinois (Par-
malee and Stephens, 1972); Will County,
Illinois (Parmalee, 1962b); Madison County,
Illinois (Parmalee, 1959a); Phelps County,

Missouri (Parmalee, 1965); and Washington
County, Arkansas (Morrison, 1970). Indian
sites from farther east have apparently not
yielded specimens of C. latrans, but in the
late Pleistocene the species ranged as far east
as Pennsylvania and Florida. Man's extermi-
nation of the larger wolves, and disruption of
the environment, contributed to an expansion
of the coyote's range since the mid-nineteenth
century. The species is now established in
Ontario, southern Quebec, New England,
New York, the Ohio Valley, and the lower
Mississippi Valley; and introduced individuals
have been reported from all states of the
southeast.

Specimens, especially skulls, of C. latrans
can almost always be easily separated from
those of wolves. Hybridization under com-
pletely natural conditions seems to have oc-
curred rarely, if ever. The three specimens
described on page 11 are the only ones
known from the western half of the continent
that appear to represent hybridization be-
tween C. lupus and C. latrans. More exten-
sive interbreeding, probably caused by man's
influence, has taken place in southeastern
Canada. This process has permitted the in-
trogression of wolf genes into the expanding
coyote population of the region, and has
resulted in a modification of that population
(see pp. 21-23). Lawrence and Bossert
( 1969 ) referred to the existing coyotes of the
northeastern United States as "C. latrans var."
Specimens taken there in recent years are
variable, but, in comparison with coyotes
from farther west, are often characterized by
larger size, more massive skull, broader and
more elevated frontal shield, more prominent
sagittal and lamboidal crests, and teeth with
less trenchant cusps.

A related process in the south-central
states has allowed introgression of genes from
C. rufus into the coyote population that occu-
pied the region in the twentieth century. In-
dividuals of this population generally closely
resemble coyotes from farther north and west
in proportions and dentition, but are charac-

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

75

terized by larger skull, more prominent sagit-
tal and lambdoidal crests, and more elongate
postorbital constriction.

Kurten (1974) grouped all coyotelike ma-
terial from the Irvingtonian and latest Blan-
can of North America under the name Canis
priscolatrans, which he considered to repre-
sent an intermediate stage in the evolutionary
transition from C. lepophagus to C. latrans.
He acknowledged that there were difficulties
in drawing specific boundaries in this se-
quence, but thought that differences in limb
proportions, between Rancholabrean C. la-
trans and available Irvingtonian specimens,
supported a separation of species at that
point. In addition, C. priscolatrans was re-
ported to be considerably larger, on the av-
erage, than C. latrans or C. lepophagus, and
to differ from those species in certain dental
proportions.

It is important to note that Kurten's
(1974) delineation of C. priscolatrans was
based in part on his inclusion within that
species of several large specimens which had
been assigned by me (Nowak, 1973) or
others to the wolf line. Among this material
is the type of C. priscolatrans, which I con-
sidered not more than subspecifically distinct
from C. rufus; the type of C. edwardii, which
Gazin (1942) and I considered to be closely
related to C. rufus; and specimens from the
Inglis IA fauna in Florida which were re-
ferred to C. rufus by Webb (1974b: 17). I
continue to think that this material repre-
sents the wolf line, and that several other
(but certainly not all) Irvingtonian or late
Blancan specimens, assigned by Kurten to
the coyote line, probably are referable to C.
rufus or C. edwardii. If such specimens are
removed from consideration, we find that the
Irvingtonian coyotes are not notably larger
than those of the Rancholabrean. Although
I did not make a detailed analysis of limb
bones, as did Kurten, he apparently evaluated
only three specimens designated by him as
C. priscolatrans, one of which was part of the
Inglis material referred to C. rufus by Webb

(1974b: 17), and another of which was from
a partial skeleton, recovered at Arkalon gravel
pit in Kansas, which did not include cranial
portions. I do not consider that this sample,
or the analysis based thereon, can substan-
tially support a specific division of the coy-
ote line at the Irvingtonian-Rancholabrean
boundary, and I continue to think that this
line was represented in the Irvingtonian only
by the species C. latrans.

Kurten (1974:12, 27) also stated that the
species Canis arnensis of the Villafranchian
of Europe closely resembled the North Amer-
ican material he referred to C. priscolatrans,
but that the former averaged smaller. He
observed that future studies might demon-
strate that these two species formed part of
a single Holarctic coyote population, in which
case C. arnensis would have to be regarded as
a synonym or subspecies of C. priscolatrans.
I would not agree with the use of the name
C. priscolatrans for such a population, but
since I have not examined any specimens of
C. arnensis I can not comment on the question
of their relationship to North American Canis.

Fossil record. — The following list is ar-
ranged alphabetically by state and province,
and geographically (north to south, west to
east) within states, except that Alberta is
listed first and Mexican states last. Speci-
mens examined by me are identified by ele-
ment, museum number, or both; and selected
measurements are found in appendix B (part
10) and appendix C (part 2). Occurrences
also are shown on the map in figure 46.

ALBERTA. — Medicine Hat; Kansan, Sangamon,
Wisconsin; as C. cf. latrans (Churcher, 1969b:180;
Kurten, 1974:7, 10).

ALASKA. — Cripple Creek Mine, near Fairbanks;
Wisconsin; as C. latrans (Guthrie, 1968:352).

ARIZONA. — Anita, Coconino County; early Irv-
ingtonian ( Richard H. Tedford, American Museum
of Natural History, pers. comm.); as C. latrans (Hay,
1921:633); CI, USNM.

Ventana Cave, Papago Indian Reservation, Pima
County; Wisconsin (Hibbard, 1958); as C. latrans
(Colbert, 1950:132).

Papago Springs Cave, southeast of Sonoita, Santa
Cruz County; Wisconsin (Hibbard, 1958); as C.
caneloensis (Skinner, 1942:163), as C. latrans cane-

76

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

Fig. 46. — Map showing localities (black dots) of fossil C. latrans. Because of the scale of the map, it
was not possible to plot all localities in crowded areas.

loensis (Slaughter, 1966b:480), as C. latrans (An-
derson, 1968:22); skull without mandibles, AMNH
42800. Skinner originally distinguished caneloensis
as a species, on the basis of its following characters,
as compared to the living C. estor ( =C. latrans
mearnsi ) : proportionally wider face, larger bullae,

wider P4, less prominent hypocone on Ml, and
wider M2 with less developed hypocone and proto-
conule. Now, however, C. estor and all other named
kinds of Recent coyotes have been arranged as sub-
species of C. latrans (Jackson, 1951), and it seems
unlikely that these differences presently would be

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

77

considered more than subspecific. Slaughter (1966b:
480) demonstrated that the relative facial width of
caneloensis fell within the range of variation of Recent
C. latrans. The additional characters listed by Skin-
ner are matched in large series of Recent coyotes.
The interruption of the ridge extending anteriorly
from the hypocone of Ml is unusual, but does occur
rarely in modern coyotes. In the structure of the
occipital, braincase, frontals, and sagittal crest, and
in all other features that can be evaluated, the skull
is not different from typical western C. latrans. I
thus concur with Slaughter in reducing caneloensis
to subspecific rank.

Murray Springs, 1 mi. W Lewis Spring on San
Pedro River, Cochise County; late Pleistocene; man-
dibular fragment, UAriz 2406.

Double Adobe, Cochise County; late Wisconsin;
as C. latrans (Kurten, 1974:8).

ARKANSAS.— Eddy Bluff shelter, near Spring-
dale, Washington County; early Recent; as C. latrans
(Morrison, 1970).

Conard fissure, 15 mi. S Harrison, Newton County;
Illinoian (Kurten, 1963:100); as C. latrans (James
H. Quinn, Department of Geology, University of
Arkansas, pers. comm.).

Peccary Cave, eastern Newton County; early
Recent; as "coyote" (Quinn, 1972:93).

CALIFORNIA.— Samwel Cave, Shasta Lake,
Shasta County; Wisconsin ( Hibbard, 1958); as C.
latrans (Graham, 1959).

Hawver Cave, 5 mi. E Auburn, El Dorado Coun-
ty; Wisconsin (Hibbard, 1958); as C. ochropus
(Stock, 1918:479), as C. latrans (Anderson, 1968:
22); cranial fragment, UCMP 11041.

Teichart gravel pit, Sacramento County; late
Pleistocene; mandibular fragment, UCMP 85379.

Murphys, Calaveras County; Pleistocene; as C.
latrans (Whitney, 1879:246), as "C. latrans?" ( Mer-
riam, 1903:290).

Irvington, Alameda County; Irvingtonian; as C.
irvingtonensis (Savage, 1951:231); two mandibular
fragments, UCMP 38748, 38805; cranial fragment,
UCMP 56090; radius, UCMP 38804. According to
Savage, irvingtonensis has a relatively deeper hori-
zontal ramus, and relatively wider and more closely
spaced premolars than C. latrans. He compared the
two mandibles to 79 specimens of Recent western
coyotes, and to 30 jaws of Pleistocene C. latrans
from Rancho La Brea. He listed eleven dental
measurements of irvingtonensis, but none on the
comparative material, and also none to demonstrate
the proportional depth of the Irvington jaws. Al-
though I agree that these mandibles have a more
massive appearance than those of most other coy-
otes, I find all measurements to fall within the
range of variation of Recent and Pleistocene C.
latrans from western North America. The relative
depth of the horizontal ramus and width of the teeth
also overlap the corresponding dimensions in C.
latrans (Figs. 44, 47). The premolars are unusually
close together, but their spacing varies in Cams. In
development of the posterior cusps and cingulum
on p4, and in other features that can be evaluated,

10
9
8h

20 21 22 23 24 25 26

Fig. 47. — Scatter diagram comparing crown
length of ml (horizontal axis) and crown width of
ml (vertical axis), in 40 specimens of Pleistocene
C. latrans orcutti from Rancho La Brea (range of
variation indicated by solid line) and two specimens
of C. latrans irvingtonensis (v).

the Irvington material matches series of C. latrans.
Consequently I do not consider this material to rep-
resent a separate species. Savage's name may be
tentatively maintained as a subspecific designation,
C. latrans irvingtonensis, until sufficient material is
available to more fully evaluate the situation.

Tranquility, Fresno County; late Wisconsin; as
C. latrans (Kurten, 1974:8).

McKittrick tar seeps, Kern County; Wisconsin;
as C. latrans orcutti (Schultz, 1938b: 165; Giles,
1960:385); 27 skulls and cranial fragments, 16 man-
dibles and fragments thereof, LACM. Although he
recognized considerable variation in cranial and den-
tal characters, Schultz referred all McKittrick coyotes
to the subspecies orcutti. He observed that some of
the skulls were characterized by comparatively more
massive dentition, larger size, and a broader muzzle
than is to be seen in the living California subspecies,
C. latrans ochropus.

Maricopa Brea, near Maricopa, Kern County;
Wisconsin; 16 skulls and cranial fragments, 23 man-
dibles and fragments thereof, various isolated teeth,
LACM. This hitherto unreported collection closely
resembles that of C. latrans orcutti from Rancho La
Brea and McKittrick.

Carpinteria asphalt, Santa Barbara County; Wis-
consin (Hibbard, 1958); as "resembling C. latrans"
(Wilson, 1933:68), as C. latrans (Anderson, 1968:
22).

Rancho La Brea, Los Angeles, Los Angeles
County; Wisconsin; as C. ochropus orcutti (Mer-
riam, 1912:255), as C. latrans orcutti (Giles, 1960);
60 skulls and 50 mandibles, LACM. Compared to
those of Recent coyotes, Merriam stated that skulls
of orcutti averaged larger, and had broader palates
and zygomata, deeper and thicker mandibles, and
thicker carnassials. Giles (1960) found no significant
difference between samples of orcutti from Rancho
La Brea and McKittrick. He also reported that the
statistical separation among the Recent subspecies,
testes, mearnsi, and ochropus, was considerably less
than that between orcutti and any one of them. My
own findings concur with those of Merriam and
Giles. The skulls of orcutti average larger in all meas-
urable dimensions than those of Recent subspecies,

78

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

04

ORCUTTI

I
2

3

4

5

6

7

8

9

10

II

12

13

14

15

..00 1.10

Fig. 48. — Ratio diagram comparing means of C.
lot ram orcutti from Rancho La Brea and the total
series of Recent C. latrans from northern and western
North America (vertical line). Vertically arranged
numbers represent the measurements so numbered
in appendix B. A log difference scale is provided
above, and a ratio scale below the diagram. The
Recent series consisted of the combined male and
female samples (n=277). Sample sizes for the vari-
ous measurements of Rancho La Brea coyotes were
as indicated in appendix B (part 10).

and several individuals from Rancho La Brea are
larger than the largest Recent specimens that I
examined. Comparisons of proportion are provided
in figures 44, 48, and 49. In addition, orcutti differs
from Recent C. latrans in usually having more convex
frontals, a more prominent sagittal crest (but not
matching that of C. lepophagus), a broader supra-
occipital shield, and temporal ridges that join an-
terior, rather than posterior, to the coronal suture.
In size and all other characters there is overlap with
series of living coyotes, and I agree with earlier
authors in recognizing orcutti as a local late Pleisto-
cene population of C. latrans. The fact that the
coyotes in this group were large may be correlated
with the sympatric presence of a much larger kind
of wolf (C. dims) than the kind (C. lupus) that
shared the historical range of C. latrans. Hence the
morphology and ecological niche of the late Pleisto-
cene coyotes may have been shifted toward those of
wolves. Merriam (1912:258) noted that a few
cranial fragments from Rancho La Brea suggested
the presence of a smaller coyote, more like that found

presently in the area, in addition to orcutti. Schultz
( 1938b: 164- 168) also considered that two subspecies
might be represented in the McKittrick deposits. But
each of these authors decided that it would be best
to refer all specimens of C. latrans from the respec-
tive sites to a single subspecies. I found one adult
skull from Rancho La Brea, which was strikingly
smaller than the others, but considering the variation
that can be expected in a local population of Recent
Canis, there is no valid reason why this single skull
should be taxonomically distinguished from C. /.
orcutti. And we would expect variation in the tar
pit fauna to be especially great since the deposits
were formed over thousands of years. Both in state
of preservation and in abundance, the collections
from Rancho La Brea contain the most useful series
of Pleistocene C. latrans. The number of individual
animals represented in the Los Angeles County Mu-
seum was counted at 200 by Stock (1929) and at
239 by Marcus (1960).

Rancho La Brea, Los Angeles, Los Angeles
County; Wisconsin; as C. andersoni (Merriam, 1910:
393; 1912:260); skull without mandibles, UCMP
12249. Merriam specifically distinguished this speci-
men from C. latrans primarily because of its rela-
tively short and broad rostrum. lackson (1951:232)
thought it possible that additional material might
show andersoni to be synonymous with the living C.
latrans clcpticus of Baja and southern California.
Giles (1960) also questioned the specific status of
andersoni, and found statistical similarity between
andersoni and four skulls of clepticus from San Diego
County, California. Slaughter ( 1966a:479-48I )
found the rostral proportions of andersoni to be com-
parable to those of a Pleistocene skull from the Lau-
bach Cave, Williamson County, Texas, which he
thought might be referable to C. latrans harriscrooki.
Anderson (1968:24) suggested that andersoni was
probably a subspecies of C. latrans. The name C.
andersoni here is synonymized under C. latrans, and
is referred to the Rancho La Brea population of C. /.
orctitti. The single skull upon which Merriam based
his description, is from a juvenile, perhaps five
months old at time of death. Although the rostrum
is indeed relatively broader than in nearly all other
available specimens of C. latrans (Fig. 49), this
condition may be accounted for in part by age or by
retarded development lengthwise. There are several
other visible abnormalities that might contribute to
the aberrant appearance of the skull. The alveolus
for the left PI is missing, and the alveolus for the
right PI crowds that for the right P2. The whole
skull seems to be slightly twisted out of line, and
there is an unusual inflation of the dorsal surface of
the rostrum immediately anterior to the orbits. The
characters of the occipital, braincase, frontals, and
sagittal crest are those of normal C. latrans, and
confirm that the specimen is a coyote, not a "coyote-
like wolf" as stated by Merriam (1912:260).

Harbor freeway, Los Angeles, Los Angeles
County; Wisconsin; as C. cf. latrans (Miller, 1971:
54).

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

79

68
66
64
62
60
58
56
54
52
50^

O O

O o
ho o

OOo •

o<

>o

ad •

' c • o • • • •

• •

•o

• • •

• •

o

62 64 66 68 70 72 74 76 78

Fig. 49. — Scatter diagram comparing alveolar
length of maxillary toothrow (horizontal axis) to
maximum crown width across upper cheek teeth
(vertical axis) in certain specimens of C. latrans
(see appendix B for full description of measure-
ments). Black dots, Recent C. latrans lestes; open
circles, Pleistocene C. latrans orcutti from Rancho
La Brea; c, C. latrans canelocnsis; d, C. latrans hon-
durensis; h, C. latrans harriscrooki; a, C. andersoni
( = C. latrans orcutti).

La Mirada, Los Angeles County; Wisconsin; as
C. cf. latrans (Miller, 1971:49).

Costeau pit, 2 mi. S El Toro, Orange County;
Wisconsin; as C. cf. latrans (Miller, 1971:17)- Ml
LACM 1S220.

Vallecito Creek, San Diego County; Irvingtonian
(Hibbard, et al, 1965); six mandibular, fragments,
various postcranial fragments, LACM.

COLORADO. — Chimney Rock animal trap, Lari-
mer County; late Pleistocene or early Recent; as C.
latrans ( Hager, 1972:65).

FLORIDA.— Ichetucknee River, Columbia Coun-
ty; Wisconsin; as C. latrans (Webb, 1974b: 17). Ac-
cording to Kurten (1974:10) a left mandible (UF
1151) from this site represents a "medium to large"
C. latrans. Martin and Webb (1974:128), however,
stated that the same specimen "belongs to a smaller
individual of the same general character" as material
from Devil's Den, which they referred to C. familiaris
(see below). In my own opinion UF 1151 represents
C. familiaris, but Webb (1974b: 17) may have based
his report of C. latrans on other material. In my
dissertation ( Nowak, 1973:204) I associated another
mandible ( UF 11517) with the Ichetucknee River
site, but that specimen actually is from Devil's Den.

Haile XIIB, Alachua County; Rancholabrean;
mandibular fragment, UF.

Devil's Den, near Williston, Levy County; late
Wisconsin or early Recent (7,000-8,000 B.P.); three
mandibular fragments, UF 11514, 11515, 11517.
Subsequent to my assignment of these specimens to
C. latrans (Nowak, 1973:204-205), they, along with
additional material from the site, were referred to
C. familiaris by Martin and Webb (1974:127-128).
The suggestion by these same authors, that canid
specimens from the Ichetucknee River, Seminole
Field, and Melbourne localities represent C. famili-
aris, rather than C. latrans, has raised questions
about the over-all status of the coyote line in the
Wisconsin and early Recent of Florida.

Reddick IA, Marion County; Sangamon (Webb,
1974b:13); as C. latrans (Gut and Ray, 1964:325)!
Withlacoochee River VIIA, Citrus County; Sanga-
mon; as C. latrans (Webb, 1974b: 13, 17). '

Seminole Field, near St. Petersburg, Pinellas
County; Wisconsin (Hibbard, et al, 1965); as C.
cf. riviveronis (Simpson, 1929a:573), as C. familiaris
(Martin and Webb, 1974:128).

Lake Cutaline, Pinellas County; late Pleistocene;
mandibular fragment, UF.

Phillipi Creek-Fruitville Ditch, 7 mi. E Sarasota,
Sarasota County; Wisconsin; as C. cf. riviveronis
(Simpson, 1929b: 275).

Melbourne, Brevard County; Wisconsin (Hibbard,
et al, 1965); as C. riviveronis (Gazin, 1950:12;
Simpson, 1929b:268), as C. cf. latrans (Ray, 1958:
433), as C. familiaris (Martin and Webb, 1974:
128); rostral fragment, MCZ 5909; mandibular frag-
ment, isolated teeth, USNM.

Vero (stratum 3), Indian River County; late
Wisconsin (Webb, 1974b:13); as C. cf. latrans
(Sellards, 1916:157), as C. riviveronis (Hay, 1917a:
59), as C. latrans (Weigel, 1962:38); maxilla with
P4, FGS 7036. In his description of C. riviveronis,
Hay wrote that the specimen differed from those of
C. latrans in having a relatively shorter anterior lobe
of P4, and a relatively greater transverse extent of
the sockets of Ml and M2. The measurements he
provided, however, were only slightly different from
those of four comparative specimens of C. latrans.
Ray (1958:433) considered that there was no basis
upon which C. riviveronis could be distinguished
from C. latrans. I agree, but the Vero specimen and
others of C. latrans from Florida average smaller
than western coyotes, possibly because of the sym-
patic presence of a small wolf (C. rufus).

IDAHO. — Jaguar Cave, Beaverhead Mountains,
Lemhi County; late Wisconsin (C-14 dates: 10,370
±350 and 11,580±250 B.P.); as C. latrans (Kurten
and Anderson, 1972:24).

Moonshiner Cave, Bingham County; late Wiscon-
sin or early Recent; as C. latrans (Kurten and An-
derson, 1972:37).

Middle Butte Cave, Bingham County; early Re-
cent; as C. latrans (Kurten, 1974:9).

American Falls, Power County; Rancholabrean
(Hibbard, et al, 1965), Illinoian (Kurten, 1974:7);

80

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

as C. latrans (Kurten, 1974:7). According to Kur-
ten, two mandibles from this locality are "large."
With respect to material from the same site, Gazin
(1935:298) stated: "An incomplete humerus, a tibia
and a third metatarsal are recognized as belonging
to a dog somewhat smaller than Canis occidentalis
but larger than a coyote." The possibility that all
of this material represents the lineage of C. rujus,
should not be overlooked.

Rainbow Beach local fauna, American Falls Res-
ervoir, Power County; Wisconsin (C-14 dates:
21,500±700 and 31,300±2,300 B.P.); as C. latrans
(McDonald and Anderson, 1975:26).

Twin Falls, Twin Falls County; Wisconsin; as
C. latrans (Kurten, 1974:9).

ILLINOIS.— Galena, Jo Daviess County; Wis-
consin (Kurten, 1974:10); as C. latrans (Hay, 1923:
337).

Polecat Creek gravel pits, 1 mi. S Ashmore, Coles
County; late Wisconsin ( Hibbard, et ah, 1965); as
C. latrans (Galbreath, 1938:306, 311).

INDIANA. — Boone County; Wisconsin (Kurten,
1974:10); as C. latrans (Cope and Wortman, 1884:
7; Lyon, 1936:150).

IOWA. — Dubuque, Dubuque County; Wiscon-
sin; as C. latrans (Kurten, 1974:9).

KANSAS. — Borchers local fauna, sec. 21, T33S
R28W, Meade County; Yarmouthian; as C. cf. la-
trans (Getz, 1960:363).

Adams local fauna, north of Cimarron River,
Meade County; early Illinoian (Hibbard, 1970); as
C. latrans ( Schultz, 1969:30).

Butler Spring local fauna, 15 mi. SSW Meade,
Meade County; late Illinoian; as C. cf. latrans (Hib-
bard and Taylor, 1960:178).

Cragin Quarry local fauna, north of Cimarron
River, Meade County; Sangamon; as C. cf. latrans
(Hibbard and Taylor, 1960:178).

MARYLAND.— Cumberland Cave, 4 mi. NW
Cumberland, Allegany County; Illinoian ( Kurten,
1963:100); as C. cf. priscolatrans (Gidley and
Gazin, 1938:23); cranial fragment, USNM 7660.
The specimen is in poor condition, and represents
an animal that may not have been a year old. The
skull is the size of that of a large Recent C. latrans,
and the braincase and postorbital constriction are
coyotelike in shape. The frontals are more convex
than in most Recent C. latrans.

MISSISSIPPI.— Vicksburg (south of), Warren
County; "Wisconsin?"; as C. latrans (Kurten, 1974:
10).

MISSOURI.— Brynjulfson Caves, 6 mi. SSE Co-
lumbia, Boone County; late Wisconsin (about 10,000
B.P.); as C. latrans (Parmalee and Oesch, 1972:29).

Younger's Cave, St. Clair County; early Recent;
as C. latrans (Kurten, 1974:10); two mandibles, KU
5952, 7072.

Herculaneum (near), Jefferson County; Wiscon-
sin (Hibbard, et ah, 1965); as C. latrans (Olson,
1940:42).

Bat Cave, 8 km. NW Waynesville, Pulaski Coun-

ty; late Wisconsin (10,000-16,000 B.P.); as C. la-
trans (Hawksley, Reynolds, and Foley, 1973:77).

Zoo Cave, 1 mi. ENE Hilda, Taney County; early
Recent (less than 9,000 B.P.); as C. latrans (Hood
and Hawksley, 1975:24; Saunders, 1977:14).

NEBRASKA. — Hay Springs quarry, Sheridan
County; probably Illinoian (Hibbard, 1958); as C.
cf. latrans (Matthew, 1902; Schultz, 1934:369).

Rushville fossil quarry, Sheridan County; Yar-
mouthian (Schultz and Martin, 1970); as "Canis
sp. — Coyote" (Schultz and Tanner, 1957:71); man-
dibular fragment, UN 2913.

Mullen, Cherry County; late Irvingtonian (Kur-
ten, 1974:7); as C. latrans (Martin, 1972:174);
mandible, UN 26115.

Schmidt gravel pit, West Point, Cuming County;
late Pleistocene; skull without mandibles, UN 2909.

Allen site, Frontier County; late Wisconsin; as
C. latrans (Schultz, Martin, and Tanner, 1970:120).

Angus fossil quarry, Nuckolls County; Yar-
mouthian; as C. latrans (Schultz and Tanner, 1957:
67).

NEVADA.— Tule Springs, Clark County; late
Wisconsin; as C. latrans (Mawby, 1967).

NEW MEXICO.— Isleta Caves, 8 mi. W Isleta,
Bernalillo County; late Wisconsin; as C. latrans
(Harris and Findley, 1964:115).

Blackwater Draw, near Clovis, Curry County;
Wisconsin (Lundelius, 1967:301); CI, Ml, TM 937-
896, 937-897.

Shelter Cave, near Las Cruces, Dona Ana County;
late Pleistocene; skull without mandibles, LACM.

Conkling Cavern, near Las Cruces, Dona Ana
County; late Pleistocene; skull, LACM 1634; man-
dible, LACM 1631.

Burnet Cave, 50 mi. W Carlsbad, Eddy County;
Wisconsin (Hibbard, 1958); as C. latrans lestes and
C. microdon (Schultz and Howard, 1935:284); man-
dibular fragment, CI, P4, UN 13454, 13455.

Dark Canyon Cave, Eddy County; late Pleisto-
cene; mandible, LACM.

Dry Cave, 15 mi. W Carlsbad, Eddy County; Wis-
consin (Kurten, 1974:8); as C. latrans (Harris, 1970:
14).

OKLAHOMA. — Berends local fauna, near Gate,
Beaver County; probably Illinoian ( Hibbard and
Taylor, 1960:57); as "a canid the size of a coyote"
(Rinker and Hibbard, 1952:101), as C. latrans
(Starrett, 1956:1187); mandibular fragment, UMMP
33319.

Alton, Ottawa County; Wisconsin (Kurten, 1974:
9); as C. latrans (Hay, 1920:129); axis, USNM
9131.

OREGON. — Fossil Lake, Lake County; early or
middle Wisconsin (Allison, 1966:32); as C. lestes
(Elftman, 1931:7); three mandibular fragments,
AMNH 8584, 8585, 8586.

PENNSYLVANIA.— Frankstown Cave, Blair
County; Wisconsin (Hibbard, 1958); as C. priscola-
trans (Peterson, 1926:283); two mandibular frag-
ments (from same individual), CI, CM 11027.
Peterson thought that the material indicated an ani-

1979

NOWAK: NORTH AMERICAN QUATERNARY CAMS

81

mal about the same size as that represented by the
type of C. priscolatrans Cope, 1899 (which in this
paper is referred to a subspecies of C. rufus). The
mandibles and teeth are large, but fall within the
size range of Recent and Pleistocene C. latrans from
western North America. As noted by Peterson, there
is a prominent posterior cusp on the p2 of each
mandible. In this respect, the Frankstown mandibles
resemble those of C. latrans harriscrooki Slaughter,
1961 from the late Pleistocene of Texas.

TEXAS. — Rock Creek, Briscoe County; Kansan
(Hibbard, 1970); as C. priscolatrans (Troxell, 1915:
628, 634). The description and illustration provided
by Troxell indicate that an Ml from this site is best
referred to C. latrans, not C. rufus priscolatrans.

Quitaque local fauna, Motley County; Wisconsin
(Kurten, 1974:9); as C. latrans ( Dalquest, 1964:
501).

Howard Ranch local fauna, Groesbeck Creek,
northwest of Quanah, Hardeman County; Wisconsin
(C-14 date: 16,775+565 B.P.); as C. latrans (Dal-
quest, 1965:71).

Wichita Falls, Wichita County; early Recent; as
C. latrans (Dalquest, 1961:75).

Lubbock Reservoir, Lubbock County; Wisconsin
(Lundelius, 1967:302); as C. latrans (Kurten, 1974:
9); mandibular fragment, TM.

Slaton quarry, 5 mi. N Slaton, Lubbock County;
Sangamon (Hibbard, 1970); as near C. latrans (Dal-
quest, 1967:9).

Gilliland local fauna, Knox County; Irvingtonian
(late Kansan: Hibbard, et al., 1965); as C. cf.
latrans (Hibbard and Dalquest, 1966:20).

Benjamin Franklin local fauna, North Sulphur
River, Delta County; late Wisconsin (11,000 B.P.);
as C. latrans (Slaughter and Hoover, 1963:141).
Two mandibular fragments were reported not to
have a posterior cusp on p2, as typical of C. latrans
harriscrooki ( see following accounts ) . The material
was thought to represent a more modern coyote that
was either a replacement for or a descendent of the
older harriscrooki.

Clear Creek local fauna, north of Denton, Den-
ton County; Wisconsin (C-14 date: 28,840±4,740
B.P.; Hibbard, et al, 1965); as C. latrans cf. harris-
crooki (Slaughter and Ritchie, 1963:125).

Lewisville site, Denton County; Sangamon or
Wisconsin interstadial; as C. latrans harriscrooki
(Slaughter, 1961); cast of type, mandible, SMUMP
60315. This subspecies originally was called "wolf
like," and was distinguished from other coyotes by
its well developed posterior cusp on p2, longer tooth-
row relative to depth of mandible, reduced distance
between premolars, and more vertical ascending
ramus. Slaughter examined 52 mandibles of Recent
C. latrans from the United States, and found none
with a posterior cusp on p2. He said, however, that
this cusp was present on two specimens from San
Luis Potosi, Mexico, and on one from Archaga, Hon-
duras. The same condition existed in late Pleistocene
specimens from Brazos County and Ingleside, Texas,
which Slaughter referred to harriscrooki. Another

mandible, collected at a Pleistocene site in Hender-
son County, Texas, also was assigned to this sub-
species, on the basis of its more vertical ascending
ramus. Slaughter speculated that harriscrooki might
be a southern kind of coyote that could have in-
habited Texas only in an interglacial or interstadial.
Possible affinity to Recent C. I. hondurensis of Hon-
duras was implied, and a jaw of that subspecies
reportedly had the angle of ascending ramus about
the same as in harriscrooki. I examined six speci-
mens of C. I. hondurensis ( 1 in AMNH, 1 in KU,
2 in MCZ, 2 in USNM), of which five had a promi-
nent posterior cusp on p2, and one had the cusp
slightly developed. In contrast, only six of 250
Recent specimens from the western United States,
and only one of 40 Pleistocene specimens from
Rancho La Brea, which could be checked for this
character, had any trace of the cusp. Hondurensis
originally was reported to have a relatively broad
palate, and I found such a condition to exist in most
skulls of that subspecies that I examined (Fig. 49).
Slaughter (1966b) reported that a Pleistocene skull
from Laubach Cave, Williamson County, Texas,
which he said might be referable to harriscrooki,
had a relatively broader palate than Recent C. la-
trans. Therefore, it seems a reasonable hypothesis
that a warmth-adapted coyote, with certain more
wolflike characters than typical Recent C. latrans,
was found in Texas during part of the Pleistocene,
and might still be represented by the living coyote
of Honduras.

Moore Pit local fauna, Dallas, Dallas County;
Sangamon; as C. latrans cf. harriscrooki (Slaughter,
1966a:481; 1966b:79).

Trinity River terraces, 2.5 mi. NW Trinidad,
Henderson County; Wisconsin; as "Canis sp. — Coy-
ote" (Stovall and McAnulty, 1950:233), as C. latrans
harriscrooki (Slaughter, 1961:509).

Clamp Cave, San Saba County; early Recent; as
C. latrans (Lundelius, 1967:293).

Carson Holloway Ranch, San Saba County; Wis-
consin; as C. latrans (Kurten, 1974:9).

Miller's Cave (Travertine unit), Llano County;
early Recent; as C. latrans (Lundelius, 1967:293).

Longhorn Cavern, 8.5 mi. S Burnet, Burnet
County; late Wisconsin (Hibbard, et al, 1965); as
Canis (Semken, 1961), as C. latrans (Lundelius,
1967:293).

Laubach Cave, Georgetown, Williamson County;
Wisconsin (Kurten, 1974:9); as C. latrans, possibly
C. /. harriscrooki (Slaughter, 1966a:479-481 ); skull
without mandibles, SMUMP 61269. As Slaughter
reported, the specimen has a relatively broad facial
width as compared to most C. latrans. Some Recent
and Pleistocene specimens, however, especially those
assigned to C. latrans hondurensis, approach the
Laubach Cave specimen in relative broadness (Fig.
49). Nonetheless, assignment to harriscrooki seems
a reasonable procedure (see account of Lewisville
site, above ) .

Cage gravel pit, 5 mi. N Cameron, Milam Coun-

82

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

ty; late Pleistocene; "closely related to C. latrans"
(Hay, 1927:291).

Brazos County; late Pleistocene; as "a coyote,
Canis sp." (Peterson, 1946:166), as C. latrans harris-
crooki (Slaughter, 1961:509).

Levi shelter, Travis County; early Recent; as C.
latrans ( Lundelius, 1967:293).

Schulze Cave, 28 mi. NE Rock Springs, Edwards
County; Wisconsin or early Recent; "probably re-
ferable to C. latrans harriscrooki or . . . intermediate
between that extinct race and the modern coyote"
(Dalquest, Roth, and Judd, 1969:255-257).

Klein Cave, 12 mi. WSW Mountain Home, Kerr
County; late Wisconsin; as C. latrans (Roth, 1972:
78).

Cave Without a Name, Kendall County; late
Wisconsin (C-14 date: 10,900+190 B.P.); as C.
latrans (Lundelius, 1967:293).

Wunderlich site, Comal County; early Recent;
as C. latrans (Lundelius, 1967:293).

Friesenhahn Cave, near Bulverde, Bexar County;
Wisconsin; as C. latrans (Hay, 1920:141; Lundelius,
1960:38); four mandibles, TM 933-670, 933-1622,
933-2454, 933-3398. No posterior cusp was present
on the p2 of the one specimen that could be checked
for this character.

Ingleside gravel pit, San Patricio County; Wis-
consin; as C. latrans harriscrooki (Slaughter, 1961:
509); as C. latrans (Lundelius, 1972:20).

UTAH. — Silver Creek local fauna, 5 mi. N Park
City, Summit County; late Sangamon to early Wis-
consin; as "Canis? latrans" (Miller, 1976:401).

WISCONSIN.— Blue Mounds, Dane County (or
"Iowa lead region"); late Pleistocene; as C. latrans.
O. P. Hay apparently wrote of this same material
three different times, stating ( 1 ) that it was prob-
ably from Iowa (1914:491); (2) that it was prob-
ably not from Blue Mounds, but from another crevice
(1918:347); and (3) that it was found at Blue
Mounds (1923:341).

WYOMING.— Little Box Elder Cave, west of
Douglas, Converse County; Wisconsin; as C. latrans
(Anderson, 1968, 24).

Bell Cave, Albany County; Wisconsin to early
Recent; as C. latrans (Anderson, 1974:81).

AGUASCALIENTES.— Cedazo local fauna, near
City of Aguascalientes; early Rancholabrean (prob-
ably Illinoian); as C. latrans ( Mooser and Dalquest,
1975:786).

ESTADO DE MEXICO.— Tequixquiac (near);
late Pleistocene; as "C. cf. ocropus" (Furlong, 1925:
139, 152), as "C. ocrupus" (Alvarez, 1965:27).
Both of these authors apparently were referring to
C. ochropus, a name for the living coyote of Cali-
fornia, that has been arranged as a subspecies of
C. latrans.

NUEVO LEON. — San Josecito Cave, near Aram-
berri; Wisconsin; as C. latrans ( Kurten, 1974:7); 6
cranial fragments and 22 mandibles, LACM. The
upper cranial elements can not be fully evaluated
because of their poor condition, but seem not to
differ from typical C. latrans. Measurements of the

mandibles approach those of specimens from Rancho
La Brea. None of the mandibles have a p2 with a
posterior cusp. Age has not yet been reliably deter-
mined for the mammalian fauna of this site, but
Jakway (1958:326) suggested that it was approxi-
mately as old as the fauna of Papago Springs Cave,
Arizona (Wisconsin), and older than that of Rancho
La Brea.

OAXACA.— Monte Flor Cave, 2 km. NE Valle
Nacional; early Recent; as C. latrans (Alvarez, 1963).

PUEBLA. — Valsequillo, near Puebla; late Pleisto-
cene; as C. latrans (Kurten, 1967:173).

Evolutionary position. — The species C.
latrans apparently arose from certain popula-
tions within the species C. lepophagus in the
Blancan. Subsequently, there appears to have
been relatively little change in the coyote
line, at least with respect to the skull, but
some Pleistocene populations became larger
and more massive than most, if not all Recent
subspecies.

Canis edwardii Gazin
1942. Canis edwardii Gazin, Proc. U.S. Natl.

Mus., 92:499.
1954. Canis lupus baileyi, Hoffmeister and

Goodpaster, Illinois Biol. Monogr.,

24:34.

Holotype. — Skull with mandibles; no.
12S62, U.S. Natl. Mus.; about two miles north-
east by east of Curtis Ranch House, San
Pedro Valley, Cochise County, Arizona.

Geological distribution. — Late Blancan to
early Irvingtonian.

Geographical distribution. — Known from
Arizona, Kansas, Oregon, Texas, and Chihua-
hua.

Description. — A medium-sized canid re-
sembling C. rufus in most observable char-
acters; skull medium-sized with mostly nar-
row proportions; rostrum elongated and
narrow; braincase relatively small and mod-
erately inflated dorsoposteriorly; postorbital
constriction elongated, broad lateromedially;
zygomata slender, not deep; frontals moder-
ately elevated above rostrum, not prominently
convex; sagittal crest prominent; mandible
long, slender, and shallow, with ascending
ramus set at comparatively high angle to the

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

83

vertical; teeth relatively large and set closely
together in jaws; upper canines prominent,
thin anteroposteriorly; P4 with moderately
developed deuterocone; Ml with relatively
large, deeply sculptured medial section, and
pronounced buccal cingulum; p2 lacking pos-
terior cusp; p3 usually with second cusp; p4
usually with second and third cusp, and pro-
nounced posteromedial cingulum extending
behind third cusp.

Comparison with C. lepophagus. — Larger
and relatively broader in all measurable di-
mensions; teeth usually with less trenchant
cusps.

Comparison with C. rtifus. — Close resem-
blance in size and proportions of skull; post-
orbital constriction of braincase relatively
broader; ascending ramus of mandible set at
more acute angle to the vertical; Ml with
broader medial section and more prominent
buccal cingulum.

Remarks. — Hoffmeister and Goodpaster
(1954:34) considered the name C. edwardii
a synonym of C. lupus baileyi, the small gray
wolf found in southern Arizona in historic
time. Actually, this Pleistocene wolf can not
be referred to a living subspecies. As is dis-
cussed later in this paper, C. lupus seems
not even to have entered North America
until the Illinoian, and the Curtis Ranch
fauna is early Irvingtonian in age. Although,
as explained by Hoffmeister and Goodpaster,
there is little difference in over-all size be-
tween the type of C. edwardii and some speci-
mens of baileyi, most measurable dimensions
of C. edwardii are much smaller than the
means of those of baileyi. Moreover, the ros-
trum of C. edivardii is narrower and the man-
dible shallower, than those of any skull of
baileyi examined by me. The Ml of C. ed-
wardii differs greatly from that of any modern
C. lupus, in having a pronounced buccal
cingulum and a relatively large, deeply sculp-
tured medial section. In these and other fea-
tures, the skull of C. edwardii approaches
that of C. rufus, and, as realized by Gazin,
that species is the only one to which the fossil

need be critically compared. Indeed, the red
wolf probably is a direct descendent or im-
mediate relative of C. edwardii, and the latter
eventually may be shown to be only a syno-
nym or subspecies of C. rufus.

In my dissertation (Nowak, 1973:208, 229-
230), I had indicated that C. edwardii was
known only from the type locality, but that
several other early wolf specimens might be
referable to this species, rather than to C.
rufus. Additional material, subsequently
made available to me through the kindness of
Richard H. Tedford and Beryl E. Taylor at
the American Museum of Natural History,
suggests a greater range for C. edwardii and
also provides a basis for assigning some pre-
viously examined specimens to this species.
Of particular importance in this regard is a
skull from the Rome Beds, Oregon, which
allows direct comparison between its own
parts and those of material from the type
locality and from the Gilliland local fauna,
Texas. Cranial elements from the latter site
appear to have about the same size and shape
as those of the Rome Beds specimen. The
type of C. edwardii. as well as a maxillary
fragment from the same locality, not men-
tioned in Gazin's (1942) original description,
have an Ml with a remarkably pronounced
buccal cingulum, and the Rome Beds speci-
men shares this character. This cingulum is
never so prominent in C. rufus as in these
three specimens, and this factor is one reason
for not now synonymizing C. eduardii with
the red wolf.

Kurten (1974) considered the type of C.
edwardii to be one of the large Irvingtonian
and late Blancan coyotes which he grouped
under the name C. priscolatrans ( see p.
75). Although Kurten correctly associ-
ated C. edwardii with C. priscolatrans, I
think that both names represent the lineage
of the wolf, rather than the coyote.

Record of occurrences. — The following
list is arranged alphabetically by state, except
that Chihuahua is placed last. Specimens ex-
amined by me are identified by element,

84

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

museum number, or both; and selected meas-
urements are found in appendix B (part 11)
and appendix C (part 3). Occurrences also
are shown on the map in figure 50.

ARIZONA. — Anita, Coconino County; early Irv-
ingtonian ( Richard H. Tedford, American Museum
of Natural History, pers. comm.); as "C. nubilus?"
(Hay, 1921:632); two mandibular fragments, USNM
10210 R and C. As reported by Hay, the most com-
plete ramus, in comparison with that of Recent C.
lupus, is lower and thinner, and has thinner teeth.
The proportion of length to depth actually is unlike
that found in any specimen of C. lupus, but is close
to that in some specimens of Recent C. rufus. The
other ramus also is shallow, but is not complete
enough for full evaluation. Still another mandibular
fragment was found at the site, and was question-
ably referred to C. lupus by Hay. This specimen is
much larger and deeper than the other two, and
probably represents C. armbrusteri. Its presence at
Anita, together with the other specimens, suggests
that by the early Irvingtonian there already had
been a divergence between the line of small primi-
tive wolves (C. edwardii and C. rufus) and the line
leading to the larger wolves of the late Quaternary.
Assignment of the Anita material to C. edwardii,
rather than to C. rufus, is arbitrary because ade-
quate samples are unavailable.

Curtis Ranch, San Pedro Valley, Cochise County;
early Irvingtonian (Johnson, Opdyke, and Lindsay,
1975); as C. edwardii (Gazin, 1942:499); skull
with mandibles, USNM 12862; maxillary fragment
with P3-M2, USNM 12864.

KANSAS. — Arkalon gravel pit, south side of
Cimarron River, Seward County; Irvingtonian (late
Kansan ) ; "The humerus, femur and other elements
are the length of those of a large Canis latrans Say.
The bones are heavy and nearly as large in diameter
as those of Canis lupus Linne (Hibbard, 1953:115).

Cudahy fauna, Rig Springs Ranch, Meade Coun-
ty; Irvingtonian (late Kansan: Hibbard, et al., 1965);
as Canis sp. (Getz, 1960:361). An astragalus was
reported to be from a canid the size of a small wolf
and larger than a coyote.

OREGON. — Rome Reds, Malheur County; Irving-
tonian; as C. priscolatrans (Kurten, 1974:6); par-
tial skull without mandibles, USNM 23898 (in U.S.
Geological Survey collections, Menlo Park, Cali-
fornia). The specimen is approximately the same
size as the type of C. edwardii. The teeth are larger
than those of nearly all coyotes examined.

TEXAS. — Gilliland local fauna, Knox County;
Irvingtonian (late Kansan: Hibbard, et al., 1965);
as C. cf. lupus (Hibbard and Dalquest, 1966:20);
rostral fragment, UMMP 46483; parietal, UMMP
46460. Hibbard and Dalquest thought these frag-
ments to represent "a canid the size of the gray
wolf," but actually the specimens are smaller than
the corresponding parts of any skull of C. lupus
examined by me. They clearly are not referable to

C. latrans or C. lepophagus, and in features that
can be evaluated they closely resemble the specimen
of C. edwardii from Rome Reds, and also fall well
within the range of variation of C. rufus. Hibbard
(pers. comm.) came to consider the Gilliland fauna
to be pre-Kansan, and, if so, these specimens repre-
sent one of the earliest known occurrences of a wolf
in North America.

CHIHUAHUA.— Mina Erupcion, 90 mi. SSE
Juarez; Pleistocene; as Canis sp. (Eaton, 1923:233).
Eaton wrote that six vertebrae from an adult animal
were smaller than those of C. lupus, but larger than
those of C. latrans. He implied affinity to C. prisco-
latrans Cope, 1899.

Miiiaca Mesa, approximately 100 mi. W, 10 mi.
S City of Chihuahua; Rlancan (Kurten, 1974:6);
mandibular fragment, LACM 105/149. Although
Kurten associated this specimen with the coyote line,
it is larger than any mandible of C. lepophagus or
C. latrans examined by me. The specimen does,
however, resemble most available material of C.
lepophagus in the pronounced development of the
second cusp of p4, and the reduced development of
the third cusp. In addition, unlike most wolves, p3
(as well as p2) lacks posterior cusps. While referral
to C. edwardii still seems most appropriate, the
possibility remains that this specimen represents a
transitional phase through which the wolf line evolved
from C. lepophagus.

Evolutionary position. — Canis edwardii
may represent the first unquestionable ap-
pearance of a wolf in North America. Ma-
terial referable to this species is on the aver-
age older and more primitive in characters
than that assigned to other Pleistocene species
of wolves. Perhaps C. edwardii descended
from a late Blancan population of C. lepopha-
gus, but we also can not rule out the alterna-
tive that the wolf and coyote lines had been
distinct at an earlier time. The presence of
a large wolf (C. armbrusteri) in the early
Irvingtonian, as well as C. edwardii and C.
rufus, suggests that radiation of the wolf
group had been in progress for a considerable
period. This radiation may have been asso-
ciated with the initial glacial advances of the
Pleistocene, and also with the simultaneous
extinction of the large borophagine dogs.

The early history of the wolves, and their
exact relationships with the coyotes, can not
now be assessed because of the scarcity of
fossil Canis in the Blancan and Hemphillian
(middle Pliocene). A single mandibular frag-
ment (UN 2908) from the Hemphillian

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

85

Mitchell Creek Ash Hollow formation. Fron-
tier County, Nebraska, has measurable di-
mensions close to those of C. edwardii, and
appears to represent a wolf. Because of the
poor condition of this specimen, and its re-
moval in time from the scope of this paper,
I do not now refer it to a particular species.

Canis etruscus, a wolf resembling C. ed-
wardii, was present in the early Pleistocene
of Europe (Kurten, 1968:109). Thus, a group
of small, relatively unspecialized wolves, re-
taining some coyotelike characters, seems to
have become widespread at this time. This
group apparently formed the basic stock from
which the larger wolves of the late Quater-
nary descended. Kurten (1968:108-109)
thought that C. etruscus probably gave rise
to C. lupus mosbachensis of the middle Pleis-
tocene, from which in turn modern C. lupus
developed. In the New World, C. edwardii
seems to have been close to the line from
which arose the larger C. armbrusteri and C.
dirus.

The geographical distribution of C. ed-
wardii (Fig. 50) appears to have been con-
centrated in the southwestern quarter of the
continent, while the closely related C. rufus
occupied the southeast. The latter was able
to survive, but C. edwardii eventually disap-
peared, perhaps because of changing habitat
conditions or competition with C. latrans
and/or C. lupus.

Canis rufus Audubon and Bachman

1791. Lupus niger Bartram, Travels, p. 199.
Not available because Bartram was not
consistently binomial (according to
Int. Comm. Zool. Nomen., 1957, opin-
ion 447 ) .

1851. Canis lupus var. Rufus Audubon and
Bachman, Quadrupeds of North Amer-
ica, 2:240. Type locality, 15 mi. W
Austin, Texas (Goldman, Jour. Mamm.,
18:38, 1937).

1899. Canis priscolatrans Cope, Jour. Acad.
Nat. Sci., Philadelphia, ser. 2, 9:227.
Type from Port Kennedy deposit, Up-

per Merion Township, Montgomery
County, Pennsylvania. Valid as a sub-
species of C. rufus.

1905. Canis rufus, Bailey, N. Amer. Fauna,
25.174.

1912. Canis floridanus Miller, Proc. Biol. Soc.
Washington, 25:95. Type from Horse
Landing, about 12 mi. S Palatka, Put-
nam County, Florida. Valid as a sub-
species of C. rufus.

1937. Canis rufus gregonji Goldman, Jour.
Mamm., 18:44. Type from Macks
Bayou, 3 mi. E Tensas River, 18 mi.
SW Tallulah, Madison Parish, Louisi-
ana.

1942. Canis niger, Harper, Jour. Mamm., 23:
339.

1965. Canis rufus, Hall, Univ. Kansas Mus.
Nat. Hist. Misc. Publ., no. 43, p. 13.

Recent subspecies revised by Goldman
(1944); subspecies listed and distribution
mapped by Hall and Kelson (1959:851-852),
and by this paper (Fig. 50).

Type. — None designated.

Geological distribution. — Early Irvington-
ian to Recent.

Geographical distribution. — Pleistocene
and early Recent records from Arkansas, Flor-
ida, Pennsylvania, Texas, and Estado de Mex-
ico; historical range confined to southeastern
quarter of North America, from central Texas
to Atlantic, and from Gulf of Mexico north
to southern Pennsylvania, Ohio Valley, and
southeastern Kansas; presently found only in
extreme southeastern Texas and southern
Louisiana.

Description. — Medium-sized for the genus;
skull medium-sized with narrow proportions;
rostrum elongated and narrow; braincase rel-
atively small and not much inflated dorso-
posteriorly; postorbital constriction elongated,
narrow lateromedially, lateral margins often
appearing parallel when viewed from above;
zygomata usually slender and not widely
spreading; orbits usually large; frontals usu-
ally moderately elevated above rostrum, not

86

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

prominently convex, and forming a relatively
narrow and flat shield; temporal ridges often
sharp, often obscuring frontal suture, and
usually joining anterior to coronal suture;
sagittal crest prominent and sharp dorsally;
supraoccipital shield moderately large; ex-
ternal side of occipital often well ossified;
tympanic bullae usually well inflated; mandi-
ble long, narrow, and shallow, ventral edge
usually not notably convex when viewed from
side; incisors often relatively small; upper
canines prominent, thin anteroposteriorly,
ventral tips usually extending below level of
anterior mental foramina when jaws are
closed; premolars with trenchant, laterally
compressed cusps; P4 usually with prominent
deuterocone and lingual cingulum; Ml often
having relatively large, deeply sculptured
medial section, the metaconule usually promi-
nent and well separated from protocone, buc-
cal and anterior cingula usually pronounced;
M2 relatively large, cusps well developed; p2
occasionally with a posterior cusp; p3 some-
times with second and third cusp; p4 with
second cusp, usually with a moderately de-
veloped third cusp and posteromedial cingu-
lum extending behind third cusp; m2 and
talonid of ml relatively large, with moder-
ately trenchant cusps. For details on pelage
and postcranial skeleton, see Goldman
(1944), Young (1946:36), and Paradiso and
Nowak (1972b).

Comparison with C. latrans. — Skull larger
and relatively broader in most dimensions;
rostrum usually relatively broader and deep-
er; braincase relatively smaller and not so
much inflated, never broader at level of pa-
rietotemporal sutures than at base; postor-
bital constriction narrower and more elon-
gated; zygomata deeper and more widely
spreading; jugal more deeply inserted in max-
illa; frontals often more elevated above ros-
trum; temporal ridges usually sharper, more
often obscuring frontal suture, and usually
joining anterior, rather than posterior, to
coronal suture; sagittal crest more prominent;
supraoccipital shield broader, projecting far-

ther posteriorly; external side of occipital
more ossified, more often lacking projection
dorsal to foramen magnum; occipital con-
dyles usually extending farther transversely;
mandible usually relatively thicker and deep-
er; premolars set more closely together in
jaws; upper canines thicker anteroposteriorly,
usually not extending so far ventrally; pre-
molars usually relatively broader with less
trenchant cusps; deuterocone and lingual
cingulum of P4 usually less prominent; Ml
usually with relatively smaller medial section,
and less prominent hypocone, metaconule,
and buccal cingulum; p2 more often with
posterior cusp; posteromedial cingulum on
p4 usually less prominent; metaconid of ml
less pronounced, not projecting so far me-
dially; m2 and talonid of ml with less trench-
ant cusps. Atkins and Dillon ( 1971 ) listed
differences between C. rufus and C. latrans
in the morphology of the cerebellum; Russell
and Shaw (1972) and Jackson (1951:240)
discussed distinguishing characters in external
appearance.

Comparison with C. lupus. — Cranial dif-
ferences usually or more often apparent are
as follows: skull smaller and relatively nar-
rower in most dimensions; rostrum narrower;
braincase relatively deeper; lateral margins
of postorbital constriction appearing more
nearly parallel when viewed from above, not
rising so steeply into frontal region; zygomata
more slender, not so deep, not so widely
spreading; orbits relatively larger; frontals
less elevated above rostrum, less convex, and
forming a flatter and relatively narrower
shield; tympanic bullae more inflated; mandi-
ble shallower; incisors smaller; upper canines
thinner anteroposteriorly, extending more
ventrally; premolars narrower; P4 with more
prominent deuterocone, its root appearing to
pass more vertically into palate; Ml with
relatively larger, more deeply sculptured me-
dial section, and more prominent metaconule
and buccal cingulum; M2 relatively larger;
p4 more often with third cusp and postero-
medial cingulum extending behind third cusp;

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

87

talonid of ml relatively larger with more
trenchant cusps. Atkins and Dillon (1971)
listed differences between C. rufus and C.
lupus in the morphology of the cerebellum;
Goldman (1944), and Paradiso and Nowak
(1972b) discussed differences in pelage and
external appearance.

Other comparisons. — See accounts of C.
edwardii and C. armbrusteri.

Remarks. — Two critical problems concern-
ing the systematics of C. rufus are: (1) its
original relationship with C. lupus, particu-
larly the question of whether the two inter-
graded in the forests of the eastern United
States; and (2) the relationship in historic
time between C. rufus and C. latrans in the
south-central United States.

Although it is sometimes difficult to sepa-
rate specimens of red and gray wolves, the
previous part of this paper showed that multi-
variate analysis could distinguish nearly all
skulls of C. rufus, including all taken before
1920 in the eastern United States, from large
series of C. lupus. In addition to the meas-
urements used in multivariate analysis, the
characters listed in the above "comparison
with C. lupus" usually serve to distinguish
the two species. Problems may still arise,
especially if complete skulls are not available.

In nearly all measurements and other fea-
tures in which C. rufus differs from C. lupus,
the former approaches C. latrans. Indeed,
available specimens of the red wolf almost
bridge the morphological gap between the
proximal extremes of the other two species.
Hybrid origin for C. rufus thus seems to be
one possibility, but there are other solutions
to the problem. The most reasonable expla-
nation is that C. rufus represents a primitive
line of wolves that has undergone less change
than C. lupus, and has thus retained more
characters found in the ancestral stock from
which both wolves and coyotes arose.

The frontal shield and postorbital con-
striction are the only parts of the reo1 wolf's
skull that often do not have a form interme-
diate between that of typical C. latrans and

C. lupus. The postorbital constriction in C.
rufus is sometimes relatively narrower than
in both the gray wolf and coyote, and the
elongated lateral margins often appear paral-
lel when viewed from above, unlike the nor-
mal condition in either of the other two
species. The frontal shield of C. rufus is also
relatively narrower, and in some specimens
has a more flattened aspect than in either
C. lupus or C. latrans. These characters, to-
gether with a prominently rising sagittal crest,
give a unique appearance to certain speci-
mens of C. rufus, including both some pre-
1920 and some post- 1960 individuals.

Lawrence and Bossert ( 1967 ) considered
that if initial study of the red wolf had been
based on adequate series from the southeast-
ern United States (rather than from Texas),
C. rufus and C. lupus probably would not
have been taxonomically separated. As we
have seen, however (pp. 25-2S), complete
skulls taken prior to 1920 in Louisiana and
eastward are rare, and can all be distin-
guished from those of C. lupus. In addition
to the 14 skulls listed in table 2, Goldman
(1944) assigned two other early specimens
to C. rufus. One of these, a subadult female
taken in 1832 on the Wabash River, Indiana
(in AMNH), was assigned to C. rufus gre-
goryi. I agree with this designation, as the
specimen is comparatively small and narrow-
proportioned, and has the dental characters
normally associated with the red wolf. The
other skull (in USNM), the type of C. rufus
floridamis, was taken in 1890 on the St. Johns
River, Putnam County, Florida. This speci-
men is difficult to evaluate because the pos-
terior part is missing, and there is also a
dental anomaly in that M2 on both sides is
missing. Nonetheless, the specimen seems to
be within the morphological range of other
skulls of C. rufus.

Archeological sites and other Recent de-
posits in the eastern United States have
yielded various specimens of wolves, but few
in good condition. Such specimens examined

88

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

by me, which seem best referred to C. rufus,
include the following.

Banks site, 1.5 mi. N Clarksdale, Crittenden
County, Arkansas; as C. lupus (Parmalee, 1959c:6);
mandible, ISM.

Blain site, west bank Scioto River, south of Chilli-
cothe, Ross County, Ohio; as C. lupus (Parmalee
and Shane, 1970:198); maxillary fragment, ISM.

New Paris Sinkhole No. 2, Bedford County,
Pennsylvania; as C. lupus hjcaon (Guilday and Ben-
der, 1958:134); incomplete skull and pair of mandi-
bles, CM.

Eschelman site, 3 mi. S Columbia, Lancaster
County, Pennsylvania; as C. lupus (Guilday, Parma-
lee, and Tanner, 1962:64); cranial fragment and pair
of mandibles, CM.

Buffalo Village site, Putnam County, West Vir-
ginia; as C. lupus (Guilday, 1971:9); three mandib-
ular fragments and isolated ml, CM.

Lauderdale Indian mound, Washington County,
Virginia; isolated Ml (collection of Ronald M.
Nowak).

Crow Island Indian midden, Jackson County, Ala-
bama; mandible, UMMZ. Recently, Barkalow (1976:
25-26) reported this specimen, and material from
two other sites in the \icinity, to be either C. rufus
or C. lupus.

Jungerman site, Indian River, just south of south-
ern tip of Merritt Island, Brevard County, Florida;
as C. cf. niger (Wing, 1963:52); ml, m2, UF.

Nichol's Hammock, .7 mi. NE Princeton, Dade
County, Florida; sinkhole with contemporary fauna;
as C. niger (Hirschfeld, 1968:180); mandible, UF.

Bullen and Benson (1967) reported the
discovery of three cut and perforated canid
jaws on Tick Island, Florida. Although refer-
ral to C. rufus is a possibility, the fragmentary
nature of the material (as illustrated by Bul-
len and Benson) would make assignment to
any particular species of Canis difficult.

Webb and Baby (1957:61-71) described
three specimens of wolves from the Wright
Mounds, Montgomery County, Kentucky;
near New Liberty, Owen County, Kentucky;
and the Wolford Mounds, Pickaway County,
Ohio. The most complete of these specimens,
from the Wright Mounds, was a spatula-
shaped artifact cut from the upper jaws and
palate of a wolf. The specimen now has ap-
parently been lost, but a published photo-
graph suggests that the skull represented had
a narrow rostrum, and may have belonged to
a red wolf.

Fossil record. — The following list is ar-

ranged alphabetically by state, except that
Estado de Mexico is placed last. Specimens
examined by me are identified by element,
museum number, or both; and selected meas-
urements are found in appendix B (part 12)
and appendix C (part 4). Occurrences also
are shown on the map in figure 50.

ARKANSAS— Eddy Bluff shelter, near Spring-
dale, Washington County; early Recent; as C. rufus
(Morrison, 1970); maxillary fragment, UArk. The
fossil closely matches series of modern C. rufus in
size and other characters.

FLORIDA.— Haile VILA, Alachua County; Sanga-
mon (Webb, 1974b: 13); cranial fragment, UF. The
specimen is larger than comparative material of C.
latrans and is smaller than C. lupus or C. dirus. In
size, and in characters of the frontal region and
dentition, the specimen is well within the range of
variation of Recent C. rufus. Martin (1974:77)
compared measurements of the P4 (incorrectly la-
beled as p4 in his figure 3.13) of nine specimens
from this site, to those of other wolves, and stated
that either C. lupus or C. rufus was represented.
According to his scatter diagram, the measurements
of the Haile material are substantially closer to those
of C. rufus than to those of C. lupus.

Devil's Den, near Williston, Levy County; late
Wisconsin or early Recent (7,000-8,000 B.P.); as
C. rufus (Martin and Webb, 1974:126).

Inglis IA, Citrus County; early Irvingtonian; as
C. cf. niger (Klein, 1971:17), as C. rufus (Webb,
1974b: 17), as C. lupus (Martin, 1974:72), as C.
priscolatrans ( Kurten, 1974:6); right and left maxil-
lary fragments, UF 18046; P4, UF 18049; cl, UF
18052; Ml, UF 19406; ml, UF 19404; two mandib-
ular fragments, UF 19323, 19324. This material
indicates the presence of a canid close in size and
dental characters to the wolf (C. rufus) that inhab-
ited the southeast in historical time. Several of the
specimens are larger than those of any Recent or
late Pleistocene coyote examined by me. The first
upper molars in this series have deeply sculptured
medial sections, as does the type of C. priscolatrans,
but neither the Inglis nor Port Kennedy specimens
have the buccal cingulum on Ml as strongly devel-
oped as in C. edwardii. Klein (1971:17-18) ob-
served that the measurements of the Inglis speci-
mens approached those of C. edwardii from Curtis
Ranch, and that these specimens indicated a wolf
very close to, if not conspecific with, C. rufus.

Crystal River Power Plant, Citrus County; San-
gamon; maxillary fragment, UF 17074. Kurten
(1974:10) assigned another maxilla from this site
to C. latrans, but noted that it was "large." Possibly
that specimen should be referred to C. rufus.

Melbourne, Brevard County; Wisconsin (Hib-
bard, et al., 1965); as C. cf. lupus (Ray, 1958:434),
as C. rufus (Webb, 1974b: 17); mandibular frag-
ment, MCZ 17789. As explained by Ray, C. lupus
and C. rufus can not always be distinguished on the

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

1 '

scale of m.les

Fig. 50. — Map showing localities of C. rufus from archeological sites (triangles), fossil C. rufus (black
dots), and C. edwardii (squares). The solid lines show the distribution of Recent subspecies: C. rufus rufus
(R), C. rufus gregoryi (G), and C. rufus floridanus (F). Recause of the scale of the map, it was not possi-
ble to plot all localities in crowded areas.

basis of the mandible and lower teeth. I tentatively it has a posteromedial cingulum extending behind

refer the Melbourne specimen to C. rufus because it the third cusp of p4.

is within the size range of that species, and because Vero, Indian River Country; late Wisconsin

90

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

(Webb, 1974b:13); as C. cf. niger (Weigel, 1962:
37). Only a PI and the anterior half of a P4 were
so referred, and Weigel considered the material too
meager for positive identification. Webb (1974b:17),
however, listed C. rufus for this site.

PENNSYLVANIA.— Port Kennedy deposit, Upper
Merion Township, Montgomery County; Irvingtonian
(probably Yarmouthian: Hihbard, 1958); as C.
priscolatrans (Cope, 1899:227); P4, Ml, M2, p4,
ANSP 57-58. Cope regarded the upper teeth as
"the type of a distinct species, having important
points of resemblance to the coyote" ( although his
publication listed the premolar as a PI, the tooth
actually is a P4). Cope also noted: "The forms of
the cusps and cingula in this species are like those
of the corresponding teeth of the coyote, except as
to the conules. The size is that of the large, but not
largest wolves." This description agrees well with
that of C. rufus by Goldman (1944), and Paradiso
and Nowak (1972a), who considered the red wolf
to have coyotelike teeth, but to approach C. lupus in
size. A very few specimens of Recent and late
Pleistocene C. latrans have teeth as large as those
found at Port Kennedy, but I disagree with Kurten's
(1974) suggestion that the type of C. priscolatrans
represents a coyote ancestral to modern C. latrans.
In contrast, the Port Kennedy teeth fall well within
the range of variation of Recent C. rufus, and are
nearly equal in size to those of C. edwardii.

TEXAS. — Miller's Cave, Llano County; late Wis-
consin; as Cains sp. (Patton, 1963:31). According
to Patton, a single m2 from the site was larger than
those of coyotes, and slightly larger than that of one
available specimen of C. rufus.

Buffalo Bayou, Houston, Harris County; late
Pleistocene; as "Canis sp. cf. lupins [sic] baileyi"
( Du Bar and Clopine, 1961:99). Since the gray
wolf is not known to have occurred near Houston
in Recent time, since the nearest geographical sub-
species to Houston is not baileyi, and since the
specific name was not correctly spelled, there is
reason to suspect that identification of the pertinent
material was not made carefully. The specimen
apparently has been lost, but I think that any late
Pleistocene remains of small wolves in the area would
be referable to C. rufus.

ESTADO DE MEXICO.— Upper Becerra forma-
tion, northwest of Puente del Gallo, Valley of Te-
quixquiac; Sangamon or Wisconsin; as Canis sp.
(Hibbard, 1955:52). According to Hibbard, a man-
dible from the site "is smaller than Canis lupus
Linnaeus and appears closely related to Canis niger
( Bartram ) ."

Evolutionary position. — Modern C. rufus
apparently represents a comparatively un-
modified surviving line of the primitive stock
of small wolves that had developed by the
early Pleistocene. The red wolf evolved from
C. edwardii, or a close relative, and then
remained in North America through the mid-

dle and late Quaternary. The gray wolf prob-
ably evolved from a branch of the same stock,
but one that had entered the Old World and
become isolated there through factors asso-
ciated with glaciation. While C. lupus devel-
oped in Eurasia and eventually became the
only species of wolf throughout most of the
Northern Hemisphere, and while C. dims
underwent its sudden rise and fall in the
New World, the smaller C. rufus held on to
its niche in the southern forests and marshes.

Goldman (1944:399) wrote that certain
Pleistocene remains from Rancho La Rrea
suggested the presence there of a species with
relationship to C. rufus. All of the specimens
of wild Canis from Rancho La Brea that I
examined, however, could be referred to C.
dints, C. lupus, or C. latrans.

There has been a suggestion that C. rufus
evolved from a coyotelike ancestor that had
become isolated by glaciation in a Florida
refugium (Nowak, 1970:84). This hypothesis
no longer is tenable in the light of the above
outlined evolutionary sequence of the red
wolf. Furthermore, the subspecies of coyote
(C. latrans riviveronis) that inhabited Florida
was small, and survived into early Recent
time, and hence could not have given rise to
the much larger C. rufus which already was
present in Florida by the Irvingtonian.

Canis armbrusteri Gidley

1913. Canis armbrusteri Gidley, Proc. U.S.
Natl. Mus., 46:98.

Type. — Portion of a left lower jaw con-
taining p4 to m2; no. 7662, U.S. Natl. Mus.;
Cumberland Cave, about 4 mi. NW Cumber-
land, Allegany County, Maryland.

Geological distribution. — Early (?) Irv-
ingtonian to early Rancholabrean.

Geographical distribution. — Known from
Maryland and Florida, with possible records
from Arizona, California, Nebraska, Pennsyl-
vania, South Carolina, and Texas.

Description. — Size large for the genus;
skull usually large and relatively narrow in

1979

NOWAK: NORTH AxMERICAN QUATERNARY CANIS

91

most proportions; rostrum elongated and nar-
row; braincase moderately inflated dorsally;
zygomata usually deep and broadly spread-
ing; frontals moderately elevated above ros-
trum, not prominently convex, and forming
relatively narrow shield; sagittal crest promi-
nent and sharp dorsally; supraoccipital shield
large; tympanic bullae notably large and well
inflated; mandible long, moderate in depth;
teeth comparatively large; P4 usually with
prominent deuterocone; Ml having relatively
large, deeply sculptured medial section, and
pronounced buccal cingulum; M2 relatively
large; p2 and p3 lacking posterior cusps in
available specimens; p4 having second and
third cusps, and pronounced posteromedial
cingulum extending well behind third cusp:
talonid of ml relatively large.

Comparison with C. rufus. — Usually much
larger; postorbital constriction with lateral
margins not parallel; zygomata usually deep-
er and more broadly flaring; mandible rela-
tively deeper; teeth, especially carnassials,
sometimes relatively larger; p4 having more
pronounced posteromedial cingulum.

Comparison with C. lupus. — Skull usually
narrower in most proportions; rostrum rela-
tively longer and narrower; braincase usually
more inflated dorsally; frontals less convex
and usually forming relatively narrower
shield; tympanic bullae larger and more in-
flated; P4 usually having more prominent
deuterocone; Ml having relatively larger,
more deeply sculptured medial section, and
pronounced buccal cingulum; M2 relatively
larger; p4 with third cusp, and pronounced
posteromedial cingulum extending behind
third cusp; ml with relatively larger talonid.

Comparison with C. dims. — Usually
smaller; skull narrower in most proportions;
rostrum relatively longer and much narrower;
braincase more inflated dorsally; postorbital
constriction not rising so steeply into frontal
region; frontal shield much narrower; sagittal
crest usually less prominent; supraoccipital
shield broader and not projecting so far pos-
teriorly; tympanic bullae larger and more in-

flated; postpalatine foramina more anteriorly
placed ( arrangement of the optic and anterior
lacerated foramina can not be evaluated in
available specimens of C. armbrusteri); an-
terior parts of vertical plates of palatines
flaring less broadly (placement of the vomer
can not be evaluated); mandible usually
shallower; P4 usually relatively smaller with
more prominent deuterocone; Ml having rel-
atively larger, more deeply sculptured medial
section, more reduced paracone and meta-
cone, more prominent hypoeone with its an-
terior ridge extending around protocone, and
more pronounced buccal cingulum; M2 rela-
tively larger; p4 usually similar; ml smaller
with relatively larger talonid.

Remarks. — Gidley's (1913) original de-
scription of C. armbrusteri was based on three
lower jaws from Cumberland Cave, which
reportedly differed from those of C. lupus in
having relatively greater depth, smaller ca-
nines, p2 and p3 without posterior cusps, p4
with a third cusp and posterior cingulum,
and ml with a larger heel. Not all of these
characters can now be considered diagnostic,
but on the whole Gidley's distinction of C.
armbrusteri was borne out by the discovery
of additional material. The upper teeth of
the species were first described by Patterson
(1932), who, like Gidley, noted certain coy-
otelike characters. Gidley and Gazin (1938:
15-23) discussed a number of skulls and man-
dibles from Cumberland Cave, which sup-
ported the continued recognition of C. arm-
brusteri.

Martin (1974:76) suggested that C. arm-
brusteri is synonymous with C. lupus, and
that specimens from Cumberland Cave are
closely matched by skulls of large, northern
gray wolves. I disagree with this interpreta-
tion for reasons provided in the above com-
parison of the two species. Probably of great-
est value in distinguishing C. armbrusteri
from Recent C. lupus is the presence in the
former of a pronounced buccal cingulum on
the Ml, and a posteromedial cingulum ex-
tending well behind the third cusp on p4.

92

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

These characters are conspicuous in speci-
mens both from Cumberland Cave and Flor-
ida. Also, the bullae of C. armbrusteri are
more inflated than those of C. lupus, and, in
fact, are larger than those of any other species
of Cards,

Goldman (1944:399) thought that the
Cumberland Cave wolf appeared closely al-
lied to the red wolf. In most proportions and
dental characters, C. armbrusteri does ap-
proach Recent C. rufus, and there would be
a basis for considering it a giant Pleistocene
red wolf. The two species are easily distin-
guished by size and other characters, how-
ever, and there is some evidence that they
occurred together at certain Pleistocene lo-
calities.

Whereas some specimens of C. armbru-
steri seem not very different from C. rufus,
others are nearly as large as C. dims. Al-
though the range of variation shown by C.
armbrusteri is not unusually great for a spe-
cies of Canis, the available material does span
much of the moi-phological gap between C.
rufus and C. dims, and may represent a part
of the evolutionary sequence through which
the dire wolf developed from more primitive
stock. Martin (1974:75) reported that one of
the Cumberland Cave skulls (USNM 11886)
has an inion projection as pronounced and
hooked as in C. dims, and might represent
a population beginning to develop into that
species. In my own opinion this specimen
does match some skulls of C. dirus in size
and height of sagittal crest, but not in projec-
tion of inion. Furthermore, in all distinguish-
ing characters of the dentition that can be
evaluated, the specimen is unlike C. dirus.

Record of occurrences. — Only the remains
from Cumberland Cave and two Florida sites
are complete enough for reliable assignment
to C. armbrusteri. Various other fragments of
large wolves have been reported from pre-
Illinoian sites, and are not definitely referable
to C. lupus or C. dirus on a moiphological or
chronological basis (specimens with the typi-
cal characters of these two species do not

appear in North America until the Illinoian).
The earlier material may represent the line-
age of C. armbrusteri and is listed at this
point. The following list is arranged alpha-
betically by state; specimens examined by me
are identified by element, museum number,
or both; and selected measurements are found
in appendix R (part 13) and appendix C
(part 5). Occurrences also are shown on the
map in figure 45.

ARIZONA. — Anita, Coconino County; early Irv-
ingtonian ( Richard H. Tedford. American Museum
of Natural History, pers. comm.); as "C. nuhilus?"
(Hay, 1921:632); mandibular fragment, USNM
10210 A. Of three such specimens from this site,
two are referred above to C. edwardii. The third
mandible is much larger and deeper, and is almost
identical in size and proportion with certain speci-
mens from Cumberland Cave.

CALIFORNIA. — Irvington, Alameda County;
Irvingtonian; as C. cf. dirus (Savage, 1951:230).
The few pertinent fragments from this site do not
seem adequate for identification, but do represent a
large canid and may be referable to C. armbrusteri.

FLORIDA.— McCleod lime rock mine, 2.5 mi. N
Williston, Levy County; Irvingtonian; cranial frag-
ment, AMNH 67286; two maxillary fragments (prob-
ably from same individual), AMNH 67287-67288;
two mandibular fragments ( probably from same in-
dividual), AMNH 67289-67290; mandibular frag-
ment, AMNH 67291. The specimens are large, but
do not match C. dims in size or other critical char-
acters. Referral to C. armbrusteri is supported by
the presence of a pronounced buccal cingulum on
Ml, a prominent deuterocone on P4, a pronounced
posteromedial cingulum on p4, and a relatively
shallow mandible.

Coleman IIA local fauna, Sumter County; Irving-
tonian; as C. lupus (Martin, 1974:76); skull without
mandibles, UF 11519; cranial fragment, maxillary
fragment, and three mandibular fragments, UF
11520; mandibular fragment, two ml, UF 12121;
mandibular fragment, UF 11518; two P4, UF 12114;
various teeth and postcranial elements, UF. Al-
though Martin (1974:75) considered one skull of
C. armbrusteri from Cumberland Cave (USNM
7994 ) to be "essentially identical" to the Coleman
skull UF 11519, he assigned these and other speci-
mens from both sites to C. lupus. For reasons stated
above I recognize C. armbrusteri as a distinct spe-
cies, probably most closely related to C. rufus, and
the Coleman material seems best referred to C.
armbrusteri. Some of the specimens are compara-
tively small, but most dimensions fall within the size
range for the Cumberland material. Other char-
acters in which the Coleman wolves resemble C.
armbrusteri, rather than C. lupus, include: the pro-
nounced buccal cingulum, and relatively large, deeply
sculptured medial section on Ml; the prominent

1979

NOWAK: NORTH AMERICAN QUATERNARY CAMS

93

deuterocone on P4; the pronounced posteromedial
cingulum extending well behind the third cusp on
p4; and the large, well inflated bullae of the one
specimen on which they could he evaluated.

MARYLAND.— Cumberland Cave, 4 mi. NW
Cumberland, Allegany County; Illinoian (Kurten,
1963:100); as C. armbrusteri (Gidley, 1913:98; Gid-
ley and Gazin, 1938:15), as C. lupus (Martin, 1974:
76); eight skulls, USNM 7994, 8144, 11881, 11883,
11885, 11886, 11887, 12288; 13 mandibular fragments,
USNM 7482, 7661, 8144, 8168, 8169, 8172, 11881,
11882, 11887, 11888, 12290, 12293, 12295; P4,
USNM 12289; Ml, M2, FM P14790.

NEBRASKA. — Rushville fossil quarry, Sheridan
County; Yarmouthian ( Schultz and Martin, 1970);
as C. dirus nebrascensis Frick ( Schultz and Tanner,
1957:71); maxillary fragment, UN 25691. The fea-
tures of the Ml in this specimen are not those of
C. dims, but fall within the range of variation shown
by C. armbrusteri.

Angus fossil quarry, Nuckolls County; Yarmouth-
ian; as C. dirus nebrascensis Frick ( Schultz and
Martin, 1970:347). According to Larry D. Martin
( Department of Systematies and Ecology, University
of Kansas, pers. comm. ), this record is based on an
ulna, but Merriam (1912:236) reported that in C.
dirus this element shows no sharp distinguishing char-
acters.

PENNSYLVANIA.— Port Kennedy deposit, Up-
per Merion Township, Montgomery County; Irving-
tonian (probably Yarmouthian: Hibbard, 1958); as
possibly C. indianensis (Cope, 1899:227). Cope
reported three postcranial elements to be larger than
those of any wolf known to him. Such material is
not reliable in the identification of the dire wolf,
and the record herein is listed under C. armbrusteri.

SOUTH CAROLINA.— Ashley River, Charleston
County; Pleistocene; as C. occidentalis (Hay, 1923:
365). A mandibular fragment with p4 was compared
by Hay to C. dirus and C. lupus, and was found to
be closer in size to the latter. Unfortunately, the
single specimen of C. lupus used by Hay (USNM
9001 ) is the largest skull of that species that I have
examined, and thus is hardly typical. It -seems un-
likely that gray wolves of this size ever occurred as
far to the southeast as Charleston. But the measure-
ments of depth of jaw ( 28.0 millimeters ) and length
of p4 (18.5 millimeters), listed by Hay for the
Ashley River specimen, are almost identical to those
of several specimens of C. armbrusteri from Cum-
berland Cave and Florida.

TEXAS. — Rock Creek, Briscoe County; Kansan
(Hibbard, 1970); as C. dirus (Troxe.ll, 1915:633).
Troxell referred a tibia and several other postcranial
elements to C. dirus, solely on the basis of size.
According to Stock and Lance (1948), however, the
body of C. dirus was small relative to its skull. Thus
size would not be a reliable character in distinguish-
ing the postcranial skeleton of the dire wolf from
that of other large species.

Evolutionary position. — Cards armbrusteri
is one of several large species that arose from
the basal stock of primitive wolves repre-
sented by C. edwardii and C. rufus. Descent
could have been directly from either of these
latter two species. The presence of speci-
mens of large wolves at several early Irving-
tonian sites suggests that divergence between
the lineages of C. armbrusteri and C. rufus
occurred early in the Pleistocene, and is evi-
dence for a lengthy independent evolution of
the wolf group. Unfortunately, this early
material is so fragmentary that it is impossible
to determine how many species of wolves are
represented. Specimens clearly showing typi-
cal characters of C. lupus and C. dirus do
not appear in North America until the Illi-
noian, at which time the less specialized C.
armbrusteri was still present. Therefore, it is
reasonable to suppose that before then C.
armbrusteri was the only large wolf in North
America, and that it may have occurred over
much of the continent. In the Illinoian, C.
armbrusteri might have become restricted to
the east by the initial movement of circum-
polar C. lupus into the plains and western
mountains. Canis armbrusteri disappeared
by the end of the Illinoian, but we do not
know if its lineage ended then or if it gave
rise to C. dirus, as suggested by Martin
(1974:76). This latter hypothesis is not ade-
quately supported by available morphologi-
cal evidence, and perhaps C. dirus was a
replacement for, rather than a descendent of,
C. armbrusteri.

Canis lupus Linnaeus

1758. Canis lupus Linnaeus, Systema Na-

utrae, 10th ed., p. 39. Type locality,

Sweden.
1910. Canis occidentalis furlongi Merriam,

Univ. California Publ. Bull. Dept.

Geol., 5:393. Type from Rancho La

Brea, Los Angeles County, California.

Valid as a subspecies of C. lupus.
1912. Canis milled Merriam, Mem. Univ.

California, 1:247. Type from Rancho

94

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

La Brea, Los Angeles County, Califor-
nia. A synonym of C. lupus furlongi.
1918. Aenocyon milleri, Merriam, Univ. Cali-
fornia Publ. Bull. Dept. Geol., 10:533.

In addition to those listed above, 24 names
based on North American Recent specimens
are available for use at the subspecific level.
These are to be found in the systematic re-
vision by Goldman (1944), and also are
listed by Hall and Kelson (1959:847-851).

Type. — None designated.

Geological distribution. — Late Irvington-
ian to Recent in North America.

Geographical distribution. — Pleistocene
and early Recent records from Alberta, Sas-
katchewan, Yukon, Alaska, Arizona, Arkan-
sas, California, Colorado, Georgia, Idaho, Illi-
nois, Kansas, Michigan, Minnesota, Nebraska,
Nevada, New Mexico, Oklahoma, Oregon,
Pennsylvania, Texas, Virginia, Wisconsin,
Wyoming, Nuevo Leon, and many localities
in Eurasia. Historical range throughout Eu-
rasia, except tropical forests of southeastern
corner; throughout North America, except
parts of southeastern quarter, southern and
coastal Mexico, Central America, Baja Cali-
fornia, and most of California; and on most
adjacent continental islands. Presently extir-
pated in many areas settled by man, includ-
ing most of Europe and the 48 southern con-
tinental states of the United States.

Description. — Size large for the genus;
skull usually large with mostly broad propor-
tions; rostrum elongated, usually relatively
broad and deep; braincase relatively small,
not much inflated dorsoposteriorly; postor-
bital constriction elongated, narrow latero-
medially; zygomata thick, deep, broadly flar-
ing; orbits relatively small; frontals usually
well elevated above rostrum, prominendy
convex, forming relatively broad shield; tem-
poral ridges sharp, often obscuring frontal
suture, usually joining anterior to coronal su-
ture; sagittal crest prominent, sharp dorsally;
supraoccipital shield large; external side of
occipital well ossified; tympanic bullae usu-

ally moderate in size, not much inflated; man-
dible thick and deep, ventral margin not con-
vex when viewed from side, toothrow bowed
outward in center; incisors relatively large;
upper canines prominent, thick anteropos-
teriorly, alveoli set relatively high in premax-
illae, ventral tips usually not extending to
level of anterior mental foramina when jaws
are closed; premolars relatively broad; P4
usually lacking prominent deuterocone and
lingual cingulum; Ml having relatively large
paracone and metacone, relatively small me-
dial section without trenchant cusps, the
metaconule reduced and not well separated
from protocone; Ml lacking pronounced buc-
cal cingulum; M2 usually relatively small;
p2 often with posterior cusp; p3 usually with
second and third cusp; p4 with second cusp,
sometimes lacking third cusp, usually with-
out posteromedial cingulum extending be-
hind third cusp; ml relatively broad, usually
having relatively small talonid. For details
on pelage and postcranial skeleton see Gold-
man (1944), Iljin (1941), Mech (1970), and
Hildebrand (1952a, 1952b, 1954).

Comparison with C. latrans. — Usually
much larger; skull larger and relatively broad-
er in most dimensions; rostrum relatively
broader and deeper, especially in posterior
half, flaring out more anterolaterally; brain-
case relatively smaller, less inflated dorsally,
never broader at level of parietotemporal su-
tures than at base; postorbital constriction
narrower, more elongated, rising more steeply
into frontal region; zygomata deeper, thicker,
more broadly flaring; orbits relatively smaller;
frontals more elevated above rostrum, more
depressed medially, more prominendy con-
vex, forming broader shield; temporal ridges
sharper, more often obscuring frontal suture,
and joining anterior, rather than posterior, to
coronal suture; sagittal crest more prominent;
supraoccipital shield broader, projecting far-
ther posteriorly; external side of occipital
more ossified, seldom with any trace of thin-
walled projection dorsal to foramen mag-
num; occipital condyles extending farther

1979

NOWAK: NORTH AMERICAN QUATERNARY CAMS

95

transversely; tympanic bullae usually less in-
flated; mandible thicker and deeper, ventral
edge less convex when viewed from side,
toothrow more bowed outward in center;
incisors larger, extending farther transverse-
ly; upper canines thicker anteroposteriorly,
their alveoli set more dorsally in premaxillae,
not extending so far ventrally; premolars
broader with less trenchant cusps, usually
more closely set in jaws; P4 with deuterocone
and lingual cingulum much less prominent or
absent; Ml having relatively larger paracone
and metacone, relatively smaller medial sec-
tion with less trenchant cusps, hypocone less
prominent, metaconule smaller and less dis-
tinct from protocone, protoconule often less
distinct from protocone, anterior cingulum
less pronounced, buccal cingulum less pro-
nounced or absent; M2 usually relatively
(and occasionally absolutely) smaller with
less trenchant cusps; p2 more often with pos-
terior cusp; p4 with posterior cusps more
reduced, more often lacking well developed
third cusp and posteromedial cingulum; ml
relatively broader, metaconid less prominent
and not projecting so far medially, heel rela-
tively smaller; m2 and talonid of ml with
less trenchant cusps. Hildebrand (1952a,
1954) discussed differences between the post-
cranial skeletons of C. lupus and C. latrans;
Atkins and Dillon ( 1971 ) listed distinguish-
ing features of the cerebellum.

Other comparisons. — See accounts of C.
rufus, C. armbrusteri, C. familiaris, and C.
dims.

Remarks. — The gray wolf is probably the
most widely distributed and most naturally
successful species of Canis ever to exist. Its
size, intelligence, and social nature are singu-
larly adapted for its role as the major preda-
tor of northern ungulates. The systematics of
the species have long been a source of con-
fusion, and are still not completely under-
stood. Especially difficult problems involve
the status of the small Recent subspecies or
species of wolves that existed all along the
southern margins of the range of C. lupus.

Imaizumi (1970a, 1970b) recently raised the
extinct Japanese wolf, hodophilax, back to the
level of a full species. The wolf of China and
central Asia (chanco), and of India and the
Near East (pallipes) are probably not more
than subspecifically distinct from C. lupus,
but adequate series of specimens from these
vast regions never have been studied in de-
tail. On the basis of cranial measurements
provided by Pocock (1935:671), and the few
specimens that I have seen, pallipes seems to
be a highly variable entity with a cranial size
range bridging the gap between North Amer-
ican C. lupus and C. latrans. Even more in-
teresting in this regard is arabs of southern
Arabia, of which the cranial measurements
listed by Harrison (1968:203) indicate an ani-
mal averaging not much larger than C. la-
trans. Lawrence (1966:57) suggested that
arabs may have been influenced by hybridi-
zation with C. familiaris. Harrison (1973:
190), however, reported that all available
skulls of arabs could be distinguished by the
relatively greater size and inflation of their
bullae. Two other southern subspecies listed
by Ellerman and Morrison-Scott (1951:218-
220), C. lupus italicus of Italy, and C. I. sig-
natus of Spain, had been synonymized under
C. /. lupus of most of Eurasia by Pocock
(1935). One more named subspecies, C. I.
deitanus, was based only on two live animals
from southeastern Spain. Miller (1912c:315)
noted that they had a "general appearance
much as in C. aureus." On the basis of this
description, Pocock (1935:653) suggested the
possibility that deitanus was a representative
of the North African jackal. The question ap-
parently never has been resolved.

For North America no attempt has been
made to go beyond previous studies in assess-
ing the intraspecific relationships within C.
lupus. All of the names and their areas of
application, summarized by Hall and Kelson
(1959:847-851), are maintained in this paper.
Some comment, however, is necessary regard-
ing a confusing situation on the Arctic is-
lands. Anderson (1943) and Goldman (1944)

96

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

considered that four subspecies inhabited this
region (see map, Fig. 2). Manning and Mac-
pherson (1958). following extensive statistical
analysis, concluded that the kind of wolf
represented by a series of eight skulls (in-
cluding only two adults) collected in 1914-
1916 on Banks Island, and described as C.
lupus bernardi by Anderson (1943), had been
replaced by a different kind of wolf, repre-
sented by 16 specimens collected in 1953-
1955, that seemed closest to C. I. arctos of
Prince Patrick and Ellesmere islands. I did
not measure the series of specimens taken on
Banks Island in 1914-1916, but skulls of six
males and two females collected there in
1953-1955 were suitable for inclusion in multi-
variate analyses. The series of males demon-
strates a consistently high statistical distance
from each subspecies of gray wolf, including
arctos. The most striking character of the
recently collected Banks Island skulls is their
great maximum width across the upper cheek
teeth. In all but one of these specimens this
width actually exceeds the alveolar length
from PI to M2. In most other skulls of C.
lupus, including all but one of the 21 arctos
that I measured, the length was greater than
the width. The Banks Island skulls also differ
from arctos, and most other subspecies of C.
lupus, in their greater width of frontal shield.
Although I agree with Manning and Mac-
pherson (1958:43) that the more recently
collected skulls from Banks Island differ from
Anderson's description of C. /. bernardi, I am
not so certain that these specimens may be
"assigned to C. I. arctos with confidence."

The most critical problem that concerned
the Recent wolves of the New World was the
relationship of C. lupus with C. rufus of the
southeast. Information provided in the pre-
vious part of this paper, and in the account
of C. rufus in this part, has to me confirmed
the specific status of the red wolf. Nonethe-
less, the paucity of available material from
the eastern United States gives an incomplete
picture of the original situation in that region.
In the above account of C. rufus, I discuss a

number of eastern specimens which seem
best referred to that species. Various other
fragments from the east, including many from
archeological sites and not listed by Gold-
man (1944) or Hall and Kelson (1959), prob-
ably represent C. lupus hjcaon. Such speci-
mens examined by me (indicated by ele-
ment), or reported by others, include the
following.

Tick Creek Cave site, 12 mi. W Rolla, Phelps
County, Missouri; as C. lupus (Parmalee, 1965:19).
Parmalee also reported that some remains from this
site may represent C. rufus.

Bell site, 5 mi. W. Oshgosh, Winnebago County,
Wisconsin; as C. lupus (Parmalee, 1963:61).

Raddatz rock shelter, centra] Sauk County, Wis-
consin; as C. lupus (Parmalee, 1959b:85); maxillary
fragment, mandibular fragment, ISM.

Moccasin Bluff site, west of Buchanan, Berrien
County, Michigan; as "wolf" (Cleland, 1966:205).

Anker site, Cook County, Illinois; as C. lupus
(Parmalee, 1959a:91); two maxillary fragments, pre-
maxillary fragment, two mandibular fragments (prob-
ably all from same individual), ISM.

Fisher site, south bank Des Plaines River, Will
County, Illinois; as C. lupus (Parmalee, 1962b:402);
mandibular fragment, ISM.

Kingston Lake site, 15 mi. SW Peoria, Peoria
County, Illinois; as C. lupus (Parmalee, 1962a:10).

Hummel Camp site, 1 mi. S London Mills, Fulton
County, Illinois; as C. nubilus ( Cole and Deuel,
1937:265).

Weaver site, Fulton County, Illinois; as C. lupus
(Parmalee, 1959a: 91).

Clear Lake site, Tazewell County, Illinois; as C.
lupus (Parmalee, 1959a:91); maxillary fragment,
mandibular fragment, ISM.

Busch Estate site. Pike County, Illinois; as C.
lupus (Parmalee, 1959a:91).

Knight site, Calhoun Countv, Illinois; as C. lupus
(Parmalee, 1959a:91).

Snyders site, Calhoun County, Illinois; as C.
lupus (Parmalee, 1959a: 91).

Apple Creek site, Greene County, Illinois; man-
dibular fragment, ISM.

Cahokia site, near East St. Louis, Madison Coun-
ty, Illinois; as C. lupus (Parmalee, 1957:239).

Palestine site, Palestine, Crawford County, Illi-
nois; as C. lupus (Parmalee and Stephens, 1972:71);
mandibular fragment, ISM.

Sugar Camp Hill site, Williamson County, Illi-
nois; as C. lupus (Parmalee, 1959a:91).

Fifield site, Porter County, Indiana; as "gray
wolf? Canis lupus" (Parmalee, 1972:205).

Breck Smith Cave, 8 mi. W Lexington, Fayette
County, Kentucky; as "wolf" (Miller, 1922).

Cirico Mound, near Citico Creek, Hamilton
County, Tennessee; as C. rufus floridanus (Kellogg,

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

97

1939:267), as C. lupus lycaon (Goldman, 1944:
441); mandibular fragment, USNM.

Madisonville ancient cemetery, Cincinnati vicin-
ity, Hamilton County, Ohio; as C. lupus (Langdon,
1881:299).

Hobson site, near Middleport, Meigs County,
Ohio; as C. lupus (Murphy, 1968:12).

Fairchance Mound, Moundsville, Marshall Coun-
tv, West Virginia; as C. cf. lupus (Guilday and
Tanner, 1966:42).

Mount Carbon site, 3.5 mi. SW Montgomery,
Fayette County, West Virginia; as C. lupus (Guil-
day and Tanner, 1965:2); ml, CM.

Doepkin's Farm site, U.S. Hwy. 50, Anne Arun-
del County, Maryland; maxillary fragment, Doep-
kin's Farm collection.

Quaker State Rockshelter, 3 mi. SE Franklin,
Venango County, Pennsylvania; as C. lupus (Guil-
day and Tanner, 1962:134).

Sheep Rock shelter, west bank Raystown branch
Juanita River, Huntingdon County, Pennsylvania; as
C. lupus (Guilday and Parmalee, 1965:38).

Johnston site, Indiana County, Pennsylvania;
maxillary fragment, CM.

Hartley site, Greene County, Pennsylvania; man-
dibular fragment, CM.

Eschelman site, 3 mi. S Columbia, Lancaster
County, Pennsylvania; as C. lupus (Guilday, Parma-
lee, and Tanner, 1962:64); three mandibular frag-
ments, CM. The red wolf, C. rufus, is also repre-
sented by material from this site.

Lewiston Mound, Lewiston, Niagara County,
New York; as "wolf" (Ritchie, 1969:218).

Garoga site, Fulton County, New York; maxillary
fragment, two mandibular fragments, CM.

Frontenac Island, Cayuga Lake, Cayuga County,
New York; as C. lupus (Ritchie, 1969:106); maxil-
lary fragment, CM.

Lamoka Lake site, near Tyrone, Schuyler County,
New York; as C. lupus (Guilday, 1969:55).

Sawyer's Island, near Boothbay, Lincoln County,
Maine; as C. occidentalis ( Loomis and Young, 1912:
27).

In addition to the above records, Man-
ville and Sturtevant ( 1966) reported the pres-
ence of two Indian artifacts, containing parts
of wolf skulls, in the collection of the Skok-
Ioster Castle Museum in Sweden. The speci-
mens had probably been obtained from In-
dians near the Swedish colony on the Dela-
ware River, or the Dutch colony on the Hud-
son River. The material was identified as C.
lupus lycaon, and the measurements provided
indicate that the gray wolf, rather than the
red wolf, is represented.

Whereas there is sometimes difficulty in
distinguishing specimens of C. lupus and C.

rufus, cranial material of C. lupus and C. la-
trans can always be separated. The clear dis-
tinction of the two species was recognized by
American taxonomists at least as early as
Audubon and Bachman (1851). Baird (1857:
104) adequately described some of the major
cranial differences between the gray wolf and
coyote. Cope (1879:184) was the first to
point out the discriminating features of the
cusps on the medial section of Ml. Gidley
(1913:98-102) listed what he considered to
be diagnostic characters of the lower denti-
tion, but, as explained by Jackson (1951:242),
these characters are not always reliable. The
most thorough discussion of the differences
in proportions and other characters, between
the skulls of C. lupus and C. latrans, was that
provided by Lawrence and Bossert ( 1967 ) .

The coyote and gray wolf shared a large
part of their respective ranges in North Amer-
ica, but hybridization under completely nat-
ural conditions occurred rarely, if ever. In-
terbreeding in eastern Canada, caused largely
by recent human environmental disruption,
has resulted in the production of some speci-
mens with intermediate characters.

Fossils of wolves, other than C. dims, are
comparatively rare in North America, and it
sometimes is difficult to determine what spe-
cies are represented. Martin (1974:76) con-
sidered C. armbrusteri of Maryland and Flor-
ida to be synonymous with C. lupus, but, as
explained previously, the two are distinct.
One named Pleistocene species that now can
be synonymized with a subspecies of C. lupus
is C. (Aenocyon) milleri Merriam from
Rancho La Brea. The single specimen on
which the species was based falls within the
morphological range of C. lupus, and seems
best referred to C. lupus furlongi (see ac-
count of Rancho La Brea, below).

Fossil record. — Most fossil wolf material
is so fragmentary that determination as to
species is difficult. The first specimens that
show the specific characters of C. lupus ap-
pear in Illinoian deposits. Several pre-Illi-
noian fragments that had been referred by

98

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

Fig. 51. — Map showing localities (black dots) of fossil C. lupus. Because of the scale of the map, it was
not possible to plot all localities in crowded areas.

others to this species, are discussed above in
the accounts of C. edwardii, C. rufus, and C.
armbrusteri. The following list contains ad-
ditional literary references to fossil C. lupus.

The list is arranged alphabetically by state
and province, and geographically (north to
south, west to east) within states and prov-
inces, except that Canadian provinces are

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

99

listed first and Mexican states last. Specimens
examined by me are identified by element,
museum number, or both; and selected meas-
urements are found in appendix B (part 14)
and appendix C (part 6). Occurrences also
are shown on the map in figure 51.

ALBERTA— Medicine Hat (north of); early Re-
cent; as C. lupus (Churcher, 1969b: 181).

Medicine Hat; Sangamon; as C. lupus (Churcher,
1970:62).

Island Bluff, near Medicine Hat; Sangamon; as
C. lupus (Churcher, 1969b: 181).

Mitchell Bluff, near Medicine Hat; late Pleisto-
cene; as C. lupus (Churcher, 1969a :2).

SASKATCHEWAN— Fort Qu'Apelle; late Pleis-
tocene; tibia, CNM 12178 (see also Khan, 1970:13).

YUKON. — Old Crow area; late Pleistocene; as
"wolf" (Geist, 1955:1702); mandibular fragment,
CNM 17311.

Hunker Creek vicinity, Klondike River; late Pleis-
tocene; as "Canis (wolf)" (Quackenbush, 1909:127);
skull without mandibles, CNM 9929. The specimen
shares certain characters with Recent wolves of the
Arctic islands, including a broad rostrum, crowded
toothrow, and relatively large carnassials.

Gold Run Creek, 30 mi. SE Dawson; Wisconsin
(C-14 dates: 22,200 and 32,250 B.P.); as C. cf.
lupus (Harington and Clulow, 1973:699).

Quartz Creek; late Pleistocene; cranial fragment,
CNM 17311.

ALASKA. — Historic Bluff, entrance to Esch-
scholtz Bay; late Pleistocene; as "Canis (wolf)"
(Quackenbush, 1909:97); maxillary fragment and
other cranial elements, AMNH 13753.

Buckland River, southeast of Eschscholtz Bay;
late Pleistocene; as "Canis (wolf)" (Quackenbush,
1909:120).

Fairbanks (near); Illinoian; as "Canis sp. (wolf)"
(Pewe and Hopkins, 1967:267).

Fairbanks Creek Mine, near Fairbanks; Wiscon-
sin; as C. lupus (Guthrie, 1968:352).

Engineer Creek Mine, near Fairbanks; Wiscon-
sin; as C. lupus (Guthrie, 1968:352).

Gold Hill Mine, near Fairbanks; Wisconsin; as
C. lupus (Guthrie, 1968:352).

Cripple Creek Mine, near Fairbanks; Wisconsin;
as C. lupus (Guthrie, 1968:352).

ARIZONA. — Ventana Cave, Papago Indian Res-
ervation, Pima County; Wisconsin (Hibbard, 1958);
as C. lupus (Colbert, 1950:132).

Papago Springs Cave, southeast of Sonoita, Santa
Cruz County; Wisconsin (Hibbard, 1958); as C.
nubilus (Skinner, 1942:164).

ARKANSAS. — Conard fissure, 15 mi. S Harrison,
Newton County; Illinoian ( Kurten, 1963:100); as
"C. occidentalis?" (Brown, 1908:182); cranial frag-
ment, isolated teeth, postcranial fragments, AMNH
11762; mandibular fragment, AMNH 11761.

CALIFORNIA.— Samw el Cave, Shasta Lake,
Shasta County; Wisconsin (Hibbard, 1958); "the

specimen appears to resemble the northern wolves;
for example C. /. pambasileus, rather than specimens
from the southern part of the range of C. lupus"
(Graham, 1959:58).

Potter Creek Cave, 1 mi. SE Baird, Shasta Coun-
ty; Wisconsin (Hibbard, 1958); as C. lupus (Kurten
and Anderson, 1972:37); mandibular fragment,
UCMP 5018.

McKittrick tar seeps, Kern County; Wisconsin;
as "gray or timber wolf" (Sternberg, 1928:226), as
Aenocijon near mitteri ( Schultz, 1938b:169). Stern-
berg reported that he had collected ten specimens
of gray wolves at McKittrick, but Schultz did not
mention the presence of C. lupus or C. furlongi at
the site. Schultz did note that two Ml seemed best
referred to Aenocyon milleri, which in this paper is
considered a synonym of C. lupus furlongi.

Maricopa Brea, near Maricopa, Kern County;
Wisconsin; skull and mandible, LACM 18419; skull
without mandibles, LACM 21921; maxillary and
mandibular fragments, LACM 18798; three maxillary
fragments, LACM 20531 and two unnumbered;
cranial fragment, LACM; two mandibular fragments
LACM 17890, 22288; Ml, LACM. These specimens
were identified by me from among a larger number
of specimens of C. dirus in the hitherto unreported
Maricopa Brea collection.

Rancho La Brea, Los Angeles, Los Angeles Coun-
ty; Wisconsin; as C. occidentalis furlongi (Merriam
1910:393; 1912:251; Schultz, 1938b:163; Goldman!
1944:399; Anderson, 1968:26), as C. furlongi (Hay,
1927:184; Stock, 1956:33), as C. lupus (Hibbard,
1958:18); eight skulls without mandibles, LACM
2300-44, 2300-56, 2300-353, 2300-384, 2600-1,
2600-5, 236(315), one unnumbered; cranial frag-
ment, LACM; two mandibles, LACM 2301-L476,
2301-L495; incomplete skull, UCMP 19792; maxil-
lary fragments, UCMP 10733; maxillary fragment,
mandibular fragment, UCMP 11283. Merriam (1910)
considered that certain specimens from Rancho La
Brea represented an animal closely related to the
modern gray wolf. His name for this animal, C.
occidentalis furlongi, indicated his recognition of it
as a subspecies of the North American gray wolf
which was then (1910) often referred to as C. occi-
dentalis. Miller (1912b), however, restricted the
name occidentalis to the interior forests of northern
Canada, and Hay (1927:184) considered it improb-
able that a subspecies of occidentalis ever would
have been present in southern California. Subse-
quently, various authors either followed Hay in list-
ing furlongi as a full species, or continued to use the
trinomial C. occidentalis furlongi. Since all of the
Recent gray wolves of North America, including
occidentalis, were arranged as subspecies of C. lupus
by Goldman (1944), and since Merriam's original
intention was obviously to recognize furlongi as a
subspecies of gray wolf, the proper name for the
animal in question would be C. lupus furlongi. Mer-
riam (1910, 1912) based his descriptions of furlongi
on three fragmentary specimens in the University of
California Museum. The material was said to be

100

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

smaller than that of the much more abundant C.
dims from Rancho La Brea, and to be characterized
by a more prominent hypocone on Ml. Subsequently,
Stock (1929:20) reported that 20 individuals of
furlongi from Rancho La Brea were represented in
the collections of the Los Angeles County Museum,
but later (1956:33) he wrote that "only eight speci-
mens have been recognized in the Museum collec-
tions." Considering that taxonomists of the early
twentieth century sometimes named species on the
basis of less critical analysis than is usual today, it
might be tempting to write off the few specimens of
furlongi as small, aberrant examples of C. dims (of
which 1,646 individuals were reported to be repre-
sented in the LACM collections from Rancho La
Brea by Marcus, 1960:5). There is no question,
however, that C. lupus is also present. The speci-
mens of C. dints from Rancho La Brea are remark-
ably consistent in certain critical characters, and
while going through the unlabeled collection of
wolves at the Los Angeles County Museum, I found
the eight above listed skulls of C. lupus to stand out
clearly from the others. I think that Stock (1956:
33) must have been referring to the same eight
specimens. Each of these skulls is from an adult
animal, and has each of the following characters for
which it can be evaluated: f rentals depressed me-
dially; temporal ridges sharp; orbital angle under
47°; supraoccipital shield broad, not projecting far
posteriorly; postpalatine foramina placed well anterior
to posterior edges of P4; optic foramen and anterior
lacerated foramen well separated; vertical plates of
palatines not broadly flaring anteriorly; posterior end
of vomer extending well behind posterior nasal open-
ing; Ml with large hypocone, its ridge extending
completely or almost completely around anterior
base of protocone. These characters, along with
over-all moderate size, distinguish the LACM speci-
mens, as well as those in the UCMP, from C. dirus.
In addition to these skulls, I found a cranial frag-
ment having the characters of C. lupus, and two
mandibles characterized by small size, a prominent
posterior cusp on p2, no posterior cingulum on p4,
and a relatively high set heel on ml. Most of the
specimens of furlongi have comparatively large teeth,
especially carnassials, a broad rostrum, and a rela-
tively broad frontal shield. These characters are
shared by C. dirus, and might suggest that furlongi
represents an evolutionary transition between C.
dims and C. lupus. But there is no chronological evi-
dence to support this view; material of both species
was found together in the same pits at Rancho La
Brea. Furthermore, in the great majority of char-
acters, the material from this site shows no tendency
toward blending; each specimen can be unquestion-
ably referred to either C. lupus or C. dims. Large
carnassials, and a broad rostrum and frontal shield
are also present in specimens of modern gray wolves
of the Arctic islands (see "remarks," above). Mac-
pherson (1965:164) hypothesized that wolves with
such characters had been isolated by late Pleistocene
glaciation in a Pearyland refugium, and had subse-

quently spread back across the Arctic. Possibly this
population had once occupied a large northern area,
and had then been driven by the Wisconsin glacia-
tion both northeastward into Pearyland, and south-
westward as far as Rancho La Brea. Not all speci-
mens of C. lupus from this site are as massive as
Recent Arctic wolves, and, as noted by Merriam
(1912:253) there is considerable variation in size
of teeth.

Rancho La Brea, Los Angeles, Los Angeles Coun-
ty; Wisconsin; as C. milleri (Merriam, 1912:24), as
Aenocyon milleri (Merriam, 1918:533), as C. lupus
(Martin, 1974:76); skull with mandibles, UCMP
11257. According to Merriam's original description,
the single known specimen from Rancho La Brea is
intermediate in characters between C. lupus and
C. dims. He compared the specimen most critically
with C. dirus, from which it was said to differ in
having lesser size, a smaller frontal shield, lower
sagittal crest, less overhang of inion, more anteriorly
placed postpalatine foramina, and more prominent
hypocone on Ml. Merriam reported milleri to have
a much broader palate and much more massive
dentition than C. lupus. Later (1918) he observed
that the characters of milleri justified placing it to-
gether with dims in the new genus Aenocyon. Sub-
sequently, according to Stock, Lance, and Nigra
(1946:109), the validity of milleri as a species was
questioned, but they did not indicate whether refer-
ral to C. dirus or C. lupus was being considered.
Martin (1974:76) recognized C. milleri as a synonym
of C. lupus. Had Merriam been able to examine the
eight skulls of C. lupus furlongi from Rancho La
Brea in the Los Angeles County Museum, and a
series of Recent C. lupus from the Arctic islands, he
might not have established milleri as a separate
species. As he himself observed (1912:247), the
combination of characters found in milleri is ap-
proached most closely in C. I. furlongi, and several
of the LACM specimens are almost identical to
milleri. In its relatively large carnassials and unusu-
ally broad rostrum, milleri resembles some of the
living wolves of the Arctic. Merriam's (1918) asso-
ciation of the specimen with the genus Aenocyon
was a mistake, as the specimen differs from the
dire wolf in all of the critical characters mentioned
above in the discussion of C. lupus furlongi.

Schuiling Cave, 2 mi. SE Newberry, San Ber-
nardino County; late Pleistocene; as C. cf. lupus
(Downs, ct al., 1959:9).

COLORADO.— Chimney Rock animal trap, Lari-
mer County; late Pleistocene or early Recent; as
C. lupus (Hager, 1972:65).

GEORGIA.— Ladds, near Cartersville, Bartow
Countv; late Pleistocene; as C. cf. lupus ( Rav, 1967:
133); Ml, USNM 23698.

IDAHO. — Jaguar Cave, Beaverhead Mountains,
Lemhi Countv; late Wisconsin (C-14 dates: 10,370
±350 and 11,580±250 B.P.); as C. lupus (Kurten
and Anderson, 1972:24).

Moonshiner Cave, Bingham County; late Wiscon-
sin or early Recent; as C. lupus ( Kurten and Ander-
son, 1972:37).

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

101

ILLINOIS.— Polecat Creek gravel pits, 1 mi. S
Ashmore, Coles County; late Wisconsin ( Hibbard,
et al, 1965); as C. lupus (Galbreath, 1938:309).

KANSAS. — Goodland, Sherman County; late
Pleistocene or early Recent; as C. lupus (Williston,
1898:93; Hibbard, Frye, and Leonard, 1944:10);
skull with mandibles, KU 2851. The specimen re-
sembles skulls of C. lupus baileyi from the south-
west, and is smaller than skulls of C. I. nubilus col-
lected on the Great Plains in historical time.

MICHIGAN.— Millington, Tuscola County; late
Wisconsin; as C. lupus (Wilson, 1967:211); pair of
mandibles, upper incisors, UMMP 33770.

MINNESOTA.— Itasca bison site, Clearwater
County; late Pleistocene; as "wolf" (Shay, 1963:48).

MISSOURI.— Brynjulfson Caves, 6 mi. SSE Co-
lumbia, Boone County; late Wisconsin (about 10,000
B.P.); as C. lupus (Parmalee and Oesch, 1972:29).

NEBRASKA.— Hay Springs quarry, Sheridan
County; probably Illinoian (Hibbard, 1958); as C.
cf. Occident alis (Matthew, 1918; Schultz, 1934:369);
rostral fragment, AMNH (Frick Collection) 25511;
mandibular fragment, UN 2912.

Mullen, Cherry County; late Irvingtonian (Kur-
ten, 1974:7); as Aenocyon dims (Martin, 1972:174);
maxillary fragment, UN 39337; mandibular frag-
ment, UN 26117. In size and other characters, these
specimens resemble C. lupus, not C. dims.

Freedom (near), Frontier County; late Pleisto-
cene; mandible, UN 2911.

Republican River, 1 mi. S Guide Rock, Webster
County; late Pleistocene or early Recent; cranial
fragments, USNM 18749.

NEVADA. — Lake Lahontan, near Fallon, Church-
ill County; early Recent; as C. lupus (Morrison,
1964:73).

Smith Creek Cave, Baker, White Pine County;
late Pleistocene; cranial fragment, LACM 7190.

NEW MEXICO.— Isleta Caves, 8 mi. W Isleta,
Bernalillo County; late Wisconsin; as C. cf. lupus
(Harris and Findley, 1964:115), as C. lupus (An-
derson, 1968:22).

Blackwater Draw, near Clovis, Curry County;
Wisconsin ( Lundelius, 1967:301); crushed skull,
TM 937-521; mandible, TM 937-895; two Ml, TM
937-885, 937-905.

Burnet Cave, 50 mi. W Carlsbad, Eddy County;
Wisconsin (Hibbard, 1958); as C. nubilus (Schultz
and Howard, 1935:284); mandible, UN 14004.

Hermit's Cave, east slope of Guadalupe Moun-
tains, Eddy County; late Wisconsin (C-14 dates:
11,850±350 and 12,900±350 B.P.; Schultz, Martin,
and Tanner, 1970:119); two maxillary fragments and
two mandibular fragments (probably all from same
individual), UN 19211, 19217, 19218, 19220; pair
of mandibles, UN 19216. These specimens were
associated with a man-made hearth from which the
C-14 dates were obtained. At least one of the frag-
ments is charred, as from fire, and the single skull,
apparently represented by the first four fragments
listed above, may have been deliberately broken
apart. Schultz, Martin, and Tanner (1970) re-

corded only C. dims from Hermit's Cave, and that
species is present, but the five specimens listed above
unquestionably represent C. lupus.

Dark Canyon Cave, Eddy County; late Pleisto-
cene; mandible, LACM 1644.

OKLAHOMA.— Selman Cave system, 7 mi. SW
Freedom, Woodward County; Recent, as C. lupus
(Black and Best, 1972); mandible, collection of
Troy L. Best.

Afton, Ottawa County; Wisconsin (Kurten, 1974:
9); as C. nubilus (Hay, 1920:129); skull with man-
dibles, USNM 196943; three mandibles, USNM
196946, 196947, 196948; P4, CI, USNM 9128; eight
canine teeth, USNM 9129; premaxillary fragment,
P4, Ml, p4, USNM 9130.

Arkansas River, Le Flore County; Pleistocene;
cranial fragment, AMNH 32669.

OREGON.— Bend (near), Deschutes County;
late Pleistocene; tibia, CNM 12178.

Fossil Lake, Lake County; early or middle Wis-
consin (Allison, 1966:32); as C. cf. occidentalis
(Elftman, 1931:7).

PENNSYLVANIA.— Crystal Hill Cave, 3 mi. W
Stroudsburg, Monroe County; late Pleistocene or
early Recent; as C. lupus (Leidy, 1889; Hay, 1923-
310).

TEXAS. — Lubbock Reservoir, Lubbock County;
Wisconsin (Lundelius, 1967:302); mandible, TM
892-255.

Schulze Cave, 28 mi. NE Rock Springs, Edwards
County; Wisconsin or early Recent; as C. cf. lupus
(Dalquest, Roth, and Judd, 1969:256).

VIRGINIA.— Natural Chimneys, 1 mi. N Mt.
Solon, Augusta County; late Wisconsin (ca. 10,000-
15,000 B.P.); as C. cf. lupus (Guilday, 1962:94).

WISCONSIN.— Blue Mounds, Dane County; late
Pleistocene; as C. occidentalis (Hay, 1918:347- 1923-
341).

WYOMING.— Little Box Elder Cave, west of
Douglas, Converse County; Wisconsin; as C. lupus
(Anderson, 1968:25); two mandibles from subadult
individual, UColo 22287, 24683.

Bell Cave, Albany County; Wisconsin to early
Recent; as C. lupus (Anderson, 1974:81).

NUEVO LEON. — San Josecito Cave, near Aram-
berri; Wisconsin; skull without mandibles, LACM
192-3017; mandibular fragment, LACM 192-28338.
In the large collection of canid material from San
Josecito Cave, the gray wolf is represented only by
these two specimens which probably belonged to
the same individual. The skull is much smaller than
those of C. dims from the same site, and, indeed, is
the smallest skull of an adult C. lupus that I have
examined. I am indebted to Richard L. Reynolds
of the Los Angeles County Museum for recognizing
the presence of these specimens in the collection
from San Josecito Cave, and for loaning them to me.

Evolutionary position. — The North Amer-
ican gray wolf, like many of our other larger
mammals, appears to be a comparatively late
immigrant from the Old World. Kurten

102

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

(1968:109-110) traced the probable evolution
of the species in Europe from a relatively
small ancestor. The primitive stock that gave
rise to C. lupus in the Old World was in all
likelihood the same that is represented in
North America by C. edwardii and early spec-
imens of C. rufus. At various times in the
Pleistocene, factors associated with glaciation
presumably divided this widespread basic
stock, and permitted development of several
species. While C. rufus, C. armbrusteri, and
C. dims evolved in the New World, C. lupus
arose in Eurasia and apparently became the
species of wolf most suited for the challeng-
ing environment of the late Quaternary. We
do not know at what point the gray wolf
completed its occupation of Eurasia and be-
gan to move across the Bering Strait, but con-
ceivably this could have been as early as the
Kansan glaciation. A few pre-Illinpian frag-
ments had been questionably assigned to C.
lupus by previous authors, but at present
these specimens seem best referred to C. ru-
fus or C. armbrusteri. The earliest material
that clearly displays the specific characters of
C. lupus is that from the Illinoian deposits at
Hay Springs and Mullen in Nebraska. Illi-
noian specimens also have been reported from
the Conard fissure, Arkansas, and from near
Fairbanks, Alaska. It can be reasonably as-
sumed that the species was able to cross the
Bering Land Bridge in the Illinoian, and
that it eventually established itself in some
parts of North America. Glacial movement
may have been responsible for the initial ap-
pearance of C. lupus in the central United
States.

The only known Sangamon records of the
species in North America are based on a few
fragments collected near Medicine Hat, Al-
berta (Churcher, 1969b:lSl; 1970:62). Pos-
sibly the cold-adapted C. lupus had with-
drawn from more southerly regions during
that interglacial period. The number and
distribution of Wisconsin records is much
greater, but the fossil history of the species is
comparatively poor, and few occurrences are

represented by well preserved cranial ma-
terial.

So fragmentary the material, so variable
the existing gray wolf, and so incomplete our
understanding of Pleistocene chronology, that
it is difficult to assess the factors leading to
the present situation. A number of the speci-
mens discussed in the above list do not ap-
pear to differ significantly from specimens of
C. lupus taken in the same areas in historic
time. The most interesting Wisconsin speci-
mens are the massive skulls from Rancho La
Brea and the Yukon, that resemble the skulls
of modern Arctic wolves. A population of
wolves with such skulls may have been wide-
spread in the north at the beginning of the
Wisconsin, or, more likely, at the beginning
of the last major stadial. The glacial move-
ment may then have split the population, one
element moving into the Pearyland refugium
where it survived, and the other element be-
ing driven southward where it either became
extinct or was eventually absorbed into other
populations of C. lupus. A few of the speci-
mens from Rancho La Brea are small, and
there are also unusually small skulls of C.
lupus from Goodland, Kansas and San Jose-
cito Cave, Nuevo Leon. These specimens may
represent the result of character displacement
following an initial late Pleistocene or early
Recent movement of C. lupus into areas
where C. dims still predominated. Eventu-
ally, C. lupus prevailed over the dire wolf,
either through competition or because of ex-
ternal factors, and established itself as the
major large predator of most of North Amer-
ica.

Canis familiaris Linnaeus

175S. Canis familiaris Linnaeus, Systema
Naturae, 10th ed., p. 38.

1938. Canis petrolei Stock, Bull. S. California
Acad. Sci., 37:50. Type from Rancho
La Brea, Los Angeles County, Califor-
nia. The type and only known speci-
men, originally described by Stock as
"a coyote-like wolf jaw," probably rep-

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

103

resents C. familiaris. A study on this
and other specimens of early domestic
dogs from Rancho La Brea is being
made by Richard L. Reynolds of the
Los Angeles County Museum.

Type. — None designated.

Geological distribution. — Late Ranchola-
brean to Recent.

Geographical distribution. — World-wide in
association with man.

Description. — Exceptionally variable in
size and other characters because of the in-
fluence of domestication. Less specialized
breeds characterized as follows: moderate
size; skull medium-sized, relatively broad in
most proportions; rostrum usually relatively
short, broad, and deep; braincase relatively
small, not much inflated dorosposteriorly,
broadly based, set low relative to other parts
of skull; postorbital constriction usually elon-
gated, broad lateromedially, rising very
steeply into frontal region; zygomata rela-
tively thick, deep, moderately flaring; orbits
relatively small; frontals rising steeply above
rostrum, prominently convex, forming broad
shield with bulging postorbital processes;
temporal ridges usually sharp, often obscur-
ing frontal suture, often joining anterior to
coronal suture; sagittal crest sometimes prom-
inent; supraoccipital shield small, not project-
ing far posteriorly; external side of occipital
well ossified; tympanic bullae usually small,
not much inflated; medial part of posterior
margin of palate often extending well behind
toothrow; mandible thick, deep, ventral edge
often convex when viewed from side, tooth-
row bowed out prominently in center; ascend-
ing ramus sometimes curving dorsoposterior-
ly; teeth relatively small, usually lacking
trenchant cusps, usually widely spaced in
jaws; upper canines thick anteroposteriorly,
short dorsoventrally; P4 usually lacking prom-
inent deuterocone and lingual cingulum; Ml
having relatively large paracone and meta-
cone, relatively small medial section, usually
lacking buccal cingulum; M2 usually small;

p2 and p3 sometimes with posterior cusps;
p4 with second cusp, sometimes lacking third
cusp, usually without posteromedial cingulum
extending behind third cusp; ml relatively
broad, usually with relatively small talonid;
m3 occasionally absent. For details on pelage
and postcranial skeleton see Haag (1948);
Iljin (1941); Miller, Christensen, and Evans
(1965); and Scott and Fuller (1965).

Comparison with C. latrans. — Often close
in size; skull usually relatively broader in
most dimensions; rostrum relatively broader
and deeper ( a few specialized breeds, such as
collies and Russian wolfhounds, may have
relatively longer and narrower rostra than
coyotes); braincase relatively smaller, less
inflated dorsoposteriorly, broader ventrally,
set lower relative to other parts of skull ( some
breeds with smaller skulls than coyotes, may
have relatively well inflated braincases,
broader at level of parietotemporal sutures
than at base); postorbital constriction nar-
rower, more elongate, rising much more
steeply into frontal region; zygomata thicker,
deeper, usually more broadly flaring; orbits
usually relatively smaller; frontals rising much
more steeply above rostrum, more depressed
medially, more prominently convex, and
forming broader, more bulging shield; orbital
angle greater; temporal ridges usually sharp-
er; more often obscuring frontal suture, more
often joining anterior, rather than posterior
to coronal suture; sagittal crest often more
prominent (may be absent in some smaller
breeds), usually sloping more posteroven-
trally; supraoccipital shield often not project-
ing so far posteriorly; external side of occipi-
tal more ossified, more often lacking thin-
walled projection dorsal to foramen magnum
(present in some small individuals); tympanic
bullae relatively smaller, usually more rugose
and much less inflated; palate relatively
broader, central part of posterior margin
sometimes extending well behind toothrow;
distance between toothrow and bulla relative-
ly longer; mandible relatively thicker and
deeper, ventral margin sometimes more con-

104

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

vex when viewed from side, toothrow more
bowed outward in center; teeth usually rela-
tively smaller with less trenchant cusps, more
widely spaced in jaws (some specimens have
crowded teeth); upper canines relatively
thicker anteroposteriorly and much shorter
dorsoventrally; P4 more often lacking promi-
nent deuterocone and lingual cingulum; Ml
having relatively larger paracone and meta-
cone, relatively smaller medial section, more
often lacking buccal cingulum; M2 relatively
much smaller; p2 more often having posterior
cusp; p4 having posterior cusps more reduced,
more often lacking well developed third cusp
and posteromedial cingulum; ml relatively
broader, metaconid less prominent and not
projecting so far medially; m2 and talonid of
ml usually relatively smaller with less trench-
ant cusps; dental and other anomalies more
common. Atkins and Dillon (1971) listed
differences between C. familiaris and C. la-
trans in the morphology of the cerebellum.

Comparison wit!} C. lupus. — Usually much
smaller; skull usually much smaller ( some
breeds, as Irish wolfhounds and great Danes,
may have skulls larger than those of most
wolves ) ; rostrum usually relatively shorter;
braincase usually broader based, set relative-
ly lower; postorbital constriction usually rela-
tively broader and rising more steeply into
frontal region; zygomata usually shallower,
not so broadly flaring; frontals usually rising
more steeply above rostrum, more depressed
medially, more prominently convex, and form-
ing relatively higher, broader, and more bulg-
ing shield; orbital angle usually greater; tem-
poral ridges more often joining posterior,
rather than anterior to coronal suture; sagit-
tal crest usually less prominent (may be
higher in some large breeds); supraoccipital
shield much smaller, not projecting so far
posteriorly; tympanic bullae usually smaller,
less inflated; central part of posterior margin
of palate more often extending behind tooth-
row (especially in wolf-sized individuals);
distance between bulla and toothrow usually
relatively larger; mandible usually relatively

thicker, ventral edge more often convex when
viewed from side (especially in wolf -sized in-
dividuals), toothrow usually more bowed out-
ward in center; teeth, including incisors, of
larger individuals relatively much smaller and
more widely spaced in jaws; dental and other
anomalies more common. Atkins and Dillon
( 1971 ) listed differences between C. famili-
aris and C. lupus in the morphology of the
cerebellum; Hildebrand (1952b), Iljin (1941:
377-379), and Young (1944:179) discussed
distinguishing features of external appear-
ance.

Remarks.- — An account of C. familiaris is
herein included primarily for comparative
purposes. The origin of the domestic dog,
and its fossil and archeological history in
North America, are not within the scope of
this paper. Most authorities now think that
the domestic dog was derived from one of the
small Eurasian subspecies of C. lupus (Deg-
erb0l, 1961; Lawrence, 1966; Olsen and Ol-
sen, 1977; Reed, 1961; Scott, 1968; Trouessart,
1911). Skaggs (1946:345) suggested that In-
dian dogs found at a site in Kentucky may
have had a coyotelike ancestor, but Allen
(1920:440) thought that all American aborig-
inal dogs had been introduced from the Old
World. The oldest known remains of C. fa-
miliaris in North America, dated at about
10,400-11,500 B.P., were obtained from Jaguar
Cave in Lemhi County, Idaho (Lawrence,
1966, 196S).

Attempts to describe the cranial charac-
ters that distinguish C. familiaris and C. lu-
pus have been common in the literature of
several nations for many years. One of the
earliest and most detailed of these efforts was
that of Series (1835), who listed the follow-
ing characters of the dog, as compared to
those of the wolf: braincase more broadly
based; frontals rising more steeply, with
broader postorbital processes; sagittal crest
less prominent; supraoccipital shield smaller;
canine teeth shorter dorsoventrally; carnas-
sials shorter anteroposteriorly; mandible
thicker with more convex ventral surface; m3

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

105

TABLE 4

Orbital angle (in degrees) of C. latrans, C. lupus,
C. familiaris, and C. dims.

lower
extreme

upper
extreme

sample
size

C. latrans, western U.S. .._

C. lupus, western U.S

C. familiaris — _

C. dirus, Rancho La Brea

42.8

36

50

53

42.8

38

49

76

52.9

40

64

58

53.1

48

60

75

occasionally lacking. A few other features
mentioned by Serres do not now seem as re-
liable as those listed here.

Reynolds (1909:22-23) compiled a list of
cranial characters, mostly those cited earlier
by Serres, that had been used by various
authors to separate the dog and wolf. He
reported the most important character to be
that the plane of the eye socket is more
obliquely inclined to the brow (the orbito-
frontal or orbital angle is less) in the wolf.
The method of measuring this angle was
illustrated by Iljin (1941:387) and Mech
(1970:27). I took the measurement on many
specimens and found it to be reliable in sepa-
rating skulls of dogs from those of both
wolves and coyotes (see table 4).

Other characters suggested as useful in
distinguishing skulls of dogs, as compared to
those of wolves, are: smaller average size,
relatively smaller teeth, less inflated bullae
(Miller, 1912c:313); projection of palate be-
hind M2 (Allen, 1920:436); less acute angle
of zygomatic process of maxillary, smaller
bullae (Iljin, 1941:390); braincase more ossi-
fied and joining rostrum at greater angle
(lower set relative to other parts of skull),
interorbital region more elongated (Lawrence
and Bossert, 1967:225); mandible relatively
thicker lateromedially (Lawrence, 1968).

The cranial differences between C. latrans
and C. familiaris received less attention until
relatively recently. Coues (1873) and Pack-
ard (1885) were both struck by what they
considered to be close resemblance in external
appearance of the coyote and the American

Indian dog. But Allen (1920:434-435, 450)
pointed out that no one had yet made a care-
ful comparison of dog and coyote skulls. He
reported that dogs had a smaller heel of ml,
less trenchant cusps on the molars and pre-
molars, and a more prominent outward bend
of the lower toothrow at the junction of the
molar and premolar series.

Other characters proposed as useful in
distinguishing skulls of domestic dogs, as
compared to those of coyotes, are: more in-
flated frontal sinuses, greater orbital angle,
more obtuse angle in maxillo-jugal suture,
smaller tympanic bullae, more widely spaced
teeth, smaller medial section of Ml (Hall,
1943 ) ; broader rostrum and more widely
spaced incisors (Burt, 1946:61-62); smaller
ratio obtained by dividing the distance be-
tween the inner margins of the alveoli of PI
by the distance from the anterior margin of
the alveolus of PI to the posterior margin of
the alveolus of M2 (Howard, 1949); canine
teeth shorter dorsoventrally and thicker an-
teroposteriorly (Jackson, 1951:242); relatively
broader base of braincase and deeper mandi-
ble ( Bee and Hall, 1951 ) ; deeper and thicker
mandible (Lawrence, 1968).

A problem associated with attempts to dis-
tinguish between C. lupus and C. familiaris
is the role of domestication in modifying the
phenotype of the skull. The most critical
question is whether nutritional or other fac-
tors involved in captivity cause the develop-
ing skull of a wolf to take on characters nor-
mally found in the dog. Studies have demon-
strated that wolves raised in captivity some-

106

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

times show such changes as a shortening of
the jaws, overlapping of the teeth, more
steeply raised forehead, and general decrease
in size. This subject was discussed in detail
by Degerbol (1961), Iljin (1941:390-392),
Lawrence (1966). Scott (1968:246-247), and
various other authors cited therein. Thus far
all of the investigations have concerned what
happens to wolves in captivity, but it has not
yet been determined whether C. familiaris
could take on wolflike characters under feral
conditions.

Interbreeding of domestic dogs with
wolves or coyotes has occasionally occurred,
and was discussed in the first main part of
this paper. Hybridization between C. lupus
and C. familiaris has long been accepted (see
reviews by Allen, 1920:433-434; Iljin, 1941:
360-361; and Young, 1944:180-210). Some
early naturalists, such as Coues (1873)
thought hybridization between dog and coy-
ote to be common in the west, but Allen
( 1920 ) refuted this view on the basis of
cranial evidence. Later, however, specimens
began appearing in the eastern states, that
were considered to represent hybrids between
the two species (coy-dogs). Mengel (1971)
has reviewed this subject in detail.

Canis dirus Leidy

1854. Canis primaevus Leidy, Proc. Acad.
Nat. Sci., Philadelphia, 7:200. Not C.
primaevus of Hodgson, 1833.

1855. Canis clirus Leidy, Proc. Acad. Nat.
Sci., Philadelphia, 1858, p. 21. Type
from banks of Ohio River below Evans-
ville, Vanderburgh County, Indiana.

1869. Canis indianensis Leidy, Jour. Acad.
Nat. Sci., Philadelphia, 7:368. An in-
advertant renaming of C. dirus.

1876. Canis mississippiensis J. A. Allen,
Amer. Jour. Sci., ser. 3, 11:49. Type
from Lead Region of Upper Missis-
sippi.

1884. Canis lupus, Cope and Wortman, Ann.
Rept. State Geol. Indiana, 14:9.

1912. Canis dirus, Merriam, Mem. Univ. Cal-
ifornia, 1:218.

1916. Canis ayersi Sellards, Ann. Rept. Flor-
ida Geol. Surv., 8:152. Type from
Vero, Indian River County, Florida.

1918. Aenocyon dims, Merriam, Univ. Cali-
fornia Publ. Bull. Dept. Geol., 10:533.

1918. Aenocyon ayersi, Merriam, Univ. Cali-
fornia Publ. Bull. Dept. Geol., 10:533.

1929. Canis ( Aenocyon ) ayersi, Simpson,
Bull. Amer. Mus. Nat. Hist., 56:572.

1946. Canis (Aenocyon) dirus, Stock, Lance,
and Nigra, Bull. S. California Acad.
Sci., 45:109.

1962. Canis ayersi, Weigel, Florida Geol.
Surv. Spec. Publ., no. 10, p. 37.

1972. Canis dirus, Knrten and Anderson,
Tebiwa, 15:37.

1974. Canis dirus, R. A. Martin, in Pleisto-
cene mammals of Florida (edit.
Webb), p. 73.

Type. — Left maxillary fragment with P2-
M2; no. 11614, Acad. Nat. Sci., Philadelphia;
banks of Ohio River below Evansville, Van-
derburgh County, Indiana.

Geological distrilmtion. — Rancholabrean
to early Recent.

Geographical distribution. — Known from
Alberta, Arizona, Arkansas. California, Flor-
ida, Idaho, Illinois, Indiana, Kansas, Ken-
tucky, Louisiana, Missouri, Nebraska, Ne-
vada, New Mexico, Oklahoma, Oregon, Penn-
sylvania, Tennessee, Texas, Utah, Virginia,
West Virginia, Wisconsin, Aguascalientes, Ja-
lisco (possibly), Estado de Mexico, Nuevo
Leon, Puebla, and northern Peru.

Description. — Size large for the genus;
skull averaging largest in genus, with mostly
broad proportions; rostrum elongated, rela-
tively broad and deep; braincase relatively
small, not much inflated dorsoposteriorly;
postorbital constriction elongated, narrow
lateromedially, rising steeply into frontal re-
gion; zygomata thick, deep, broadly flaring;
orbits relatively small; frontals usually well
elevated above rostrum, moderately convex.

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

107

forming broad shield; temporal ridges usu-
ally not sharp, seldom obscuring frontal su-
ture, usually joining at or anterior to coronal
suture; sagittal crest prominent, sharp dor-
sally; supraoccipital shield narrow, pro-
jecting far posteriorly; external side of occipi-
tal well ossified; tympanic bullae usually
moderate in size, not well inflated; vertical
plates of palatines flaring broadly anteriorly:
posterior end of vomer usually extending only
slightly behind posterior nasal opening; post-
palatine foramina usually opposite posterior
ends of P4; optic foramen and anterior lacer-
ated foramen normally close together in com-
mon pit; mandible thick and deep, ventral
margin not convex when viewed from side,
toothrow bowed outward in center; incisors
relatively large; upper canines prominent,
thick anteroposteriorly; premolars relatively
broad; P4 relatively large, usually lacking
prominent deuterocone and lingual cingulum;
Ml having relatively large paracone and met-
acone, relatively small medial section with-
out trenchant cusps, reduced hypocone with
ridge seldom extending anteriorly around
base of protocone, and sometimes with pro-
nounced buccal cingulum; M2 relatively
small; p2 usually lacking posterior cusp; p3
often with second and third cusps; p4 usually
having second and third cusps, and pro-
nounced posteromedial cingulum extending
behind third cusp; ml relatively large, usu-
ally having relatively small, low-set talonid.
For additional details, and description of
postcranial skeleton, see Merriam (1912);
Stock, Lance, and Nigra (1946); Nigra and
Lance (1947); Stock and Lance (1948); and
Galbreath (1964).

Comparison with C. lupus. — Skull usually
larger; posterior part of rostrum usually not
so deep; postorbital constriction rising more
steeply into frontal region; frontals less con-
vex, less depressed medially, forming rela-
tively broader and flatter shield; orbital angle
greater; temporal ridges smoother, less often
obscuring frontal suture; sagittal crest usu-
ally more prominent; supraoccipital shield

narrower, projecting farther posteriorly, more
often having posteroventral hook; vertical
plates of palatines flaring more broadly an-
teriorly; posterior end of vomer not extending
so far behind posterior nasal opening; post-
palatine foramina usually set more posterior-
ly; optic foramen and anterior lacerated
foramen closer together; mandible usually
relatively thicker and deeper; upper canines
usually relatively smaller; Ml having more
reduced hypocone, its anterior ridge much
less often extending anteriorly around base of
protocone, and more often with buccal cingu-
lum; p2 more often lacking posterior cusp;
p4 with posterior margin of second cusp usu-
ally sloping anteroventrally rather than pos-
teroventrally, more often with third cusp and
posteromedial cingulum extending behind
third cusp; ml usually relatively larger, usu-
ally with relatively smaller, narrower, lower
set talonid. For additional details, and com-
parison of postcranial elements of C. dims
and C. lupus, see Merriam (1912); Stock,
Lance, and Nigra ( 1946 ) ; and Stock and
Lance (1948).

Other comparison. — See account of C.
armbrusteri.

Remarks. — The nomenclatural confusion
regarding the type specimen, and other early
reported material of dire wolves, was resolved
by Merriam (1912:218-221). He correctly
determined that Cams dims was the proper
name for all specimens that had been re-
ported up to the time of his study.

In his original description of Canis primae-
vus (=Canis dims), Leidy (1854:200) noted:
"Certain naturalists may regard the fossil as
an indication of a variety only of the Canis
lupus, and of the correctness of such a view
I shall not attempt to decide." Subsequently,
Cope and Wortman (1884:10) wrote that "it
is impossible to admit this fossil to the rank
of a distinct and well defined species, but it
appears, in our judgement, to be but a variety
which has a living representative in the
mountains of Oregon, today." Nonetheless,
Leidv's original recognition of the distinctness

108

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

of the dire wolf was borne out when a large
number of specimens became available.

Rancho La Brea has yielded by far the
greatest amount of dire wolf material. In
fact, there are few complete skulls from all
other localities combined. Thus, for purposes

10

9-

8-

7 -

6

5-

4 -

3

2

I

0
-I
-2
-3

D

D

D

D

D

D

D

D

D

N

Y
I

B

0

I sxinp .
kayrra°tNkl

S KHT dQt n

A R FHBST H I

N F LO , P F LA
SB,AlYMs

0

A

T

0

My

F

N

F,

M

-1 1 r-

-4 -3 -2

■I

— r
0

Fig. 52. — Graphical results of multivariate analy-
sis comparing a group of 62 skulls of C. dims from
Rancho La Brea (D), with a total of 467 Recent
C. lupus divided into the following subspecific
groups: C. /. arctos (A), C. /. baileyi (B), C. /.
fuscus (F), C. /. hudsonicus (H), C. /. irremotus
(I), C. /. mackenzii (K), C. I. ligoni (L), C. /. mo-
goUonensis (M), C. /. nubilus (N), C. /. occidentalis
(O), C. /. pambasileus (P), C. /. lycaon (R), C. I.
monstrabilis (S), C. /. tundrarum (T), C. /. Ijcrnardi
(X), C. /. youngi (Y). As in previous multivariate
analyses in this paper, 15 measurements of the skull
were used, but in this case sexes were combined in
each group. Only marginal positions of individuals
of each group are plotted.

of statistical comparison, I have restricted my
sample of C. dims to 62 specimens from
Rancho La Brea that were complete enough
for use in multivariate analysis. These were
all unknown as to sex, and so were tested
against the combined male and female sam-
ples of each subspecies of C. lupus for which
more than five specimens were available (a
total of 467 specimens). Figure 52 shows the
resulting ranges of variation of each group,
and the complete separation of the dire wolf
and the gray wolf. Measurements of the
series of C. dims are listed in appendix B
(part 15), and their means are compared
with those of C. lupus in figure 53. The skull
of the dire wolf is seen to be much larger
than that of C. lupus, and to be proportion-
ally broader at the canines and frontal
shield, but not so deep, relatively, between
the toothrow and orbit. The P4 is propor-
tionally much longer, but the upper canine
and M2 have relatively smaller diameters.

The lack of multivariate overlap between
C. dims and C. lupus, and the striking differ-
ences in size and proportion, lead me to no
other conclusion than that the two must be
treated as distinct species. There also is no
statistical evidence to suggest that the dire
wolf is ancestral to the living gray wolf.

Although the material from Rancho La
Brea is by far the most abundant from any
one locality, well preserved upper cranial ele-
ments from other sites in California, Mexico,
Texas, Missouri, and Kentucky has confirmed
that the same species, with the same well
marked characters, was broadly distributed
in the late Pleistocene. Among the most re-
liable cranial characters that distinguish C.
dims from other wild species of Canis are:
large over-all size, relatively broad frontal
shield, large orbital angle (see table 4), nar-
row supraoccipital shield projecting far pos-
teriorly, vertical plates of palatines flaring
broadly anteriorly, postpalatine foramina set
relatively far posteriorly, optic foramen and
anterior lacerated foramen close together in
common pit, large carnassial teeth, and re-

1979

NOWAK: NORTH AMERICAN QUATERNARY CAMS

109

0

.04 .08

LUPUS DIRUS

1.00 1.10 1.20 1.30

Fig. 53. — Ratio diagram comparing means of C.
dirus (n=62) and the total sample of Recent C.
lupus (combined sexes, n=482). Vertically arranged
numbers represent the measurements so numbered
in appendix B. A log difference scale is provided
above, and a ratio scale below the diagram.

duced hypocone on Ml. I examined more
than 500 skulls from Rancho La Brea in the

Los Angeles County Museum, and found few
that lacked more than one of these characters.
There was no difficulty in assigning most of
this material to C. dims, and picking out the
eight skulls in the LACM collection that are
referable to C. lupus (see pp. 99-100). The
specimens from Rancho La Brea represent a
population that was geographically local, but
the deposits were laid down over many thou-
sands of years, and thus considerable intra-
specific variation would be expected at the
site.

Sellards (1916:152-157) described Canis
ayersi from Vero, Florida as belonging to the
same group as C. dims, but being specifically
distinct. The single skull on which he based
his description was thought to have a rela-
tively narrower rostrum than the skulls found
at Rancho La Brea. But the rostrum of
ayersi had apparently been damaged prior to
its recovery, and since Sellards' study some
restorative work has been carried out. A full
evaluation is now impossible, but the skull is
clearly narrower than most specimens from
Rancho La Brea. In addition, some of the
mandibles from Florida that I examined are
longer than any from the southwest. Dire
wolves in Florida may therefore have tended
to have relatively long and narrow jaws, but
the differences are slight and I do not con-
sider them to be of specific value. In all
other characters, as acknowledged by Sel-
lards, the Florida material resembles C. dims
from Rancho La Brea.

In describing C. ayersi, Sellards (1916:
156) noted that there was no effective barrier
between Florida and Evansville, Indiana, the
type locality of C. dims, and that the type of
C. dirus, a maxillary fragment, might even-
tually be shown to represent the same species
as found in Florida. Olson (1940:44) sus-
spected that the Indiana and Florida speci-
mens were probably referable to the same
species, C. dirus on the basis of nomencla-
tural priority, and that a new name might be
needed for material from California. Re-
cently, a very broad skull of a dire wolf was

110

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

obtained at Welsh Cave, Woodford County,
Kentucky, only 150 miles from Evansville,
and was assigned to C. dims by Guilday,
Hamilton, and McCrady (1971:274). In ros-
tral proportions this specimen resembles
skulls from Rancho La Brea more than it does
the type of C. ayersi, and it thus supports
recognition of the type of C. dims and speci-
mens from Rancho La Brea as representing
the same species.

Martin (1974:73) reported that a skull
from Reddick, Marion County, Florida was
similar in proportion to specimens from
Rancho La Brea, and he thus implied that
there was morphological overlap between the
two populations represented. Martin sug-
gested that the type of C. ayersi, as well as
all other dire wolf material from Florida, was
referable to C. dims, and I have followed the
same course in this paper. I have not re-
tained ayersi as a subspecies, and presently
prefer to recognize C. dims as a widespread
monotypic species.

In a popular article, Frick (1930:79) listed
two new subspecific names, Aenocyon dims
alaskensis from the pre-tundra fauna of
Alaska-Yukon, and A. dims nebrascensis from
the Sheridan fauna of Nebraska. No type
specimens or exact localities were desig-
nated, and no descriptions were provided.
Nonetheless, several authors referred to sup-
posed dire wolves from Nebraska as nebra-
scensis Frick (Schultz and Stout, 1948:565;
Schultz and Tanner, 1957:71; Schultz and
Martin, 1970:347). According to Schultz
(1934:369), there were specimens of A. d.
nebrascensis Frick from Hay Springs, Sheri-
dan County, Nebraska in the Frick collection
of the American Museum of Natural History.
No one, however, discussed particular speci-
mens, and the original faunal list from this
site had included only the gray wolf (Mat-
thew, 1918). The specimens of wolves from
Hay Springs that I examined are referable to
C. lupus, and are discussed above in the ac-
count of that species. The only specimens
of wolves reported from specific localities in

Alaska also have been referred to C. lupus,
and it is unlikely that the dire wolf ever oc-
curred so far north. I therefore consider
Frick's names as nornena nuda and have not
employed them in this paper.

Merriam (1918) created the genus Aeno-
cyon to include what he thought were three
species: A. dims, A. ayersi, and A. milleri.
Goldman (1944:400) observed that the cra-
nial and dental details of the dire wolf justi-
fied recognition of Aenocyon as a genus or
subgenus. Stock, Lance, and Nigra (1946:
109) formally used Aenocyon as a subgenus
for C. ( A. ) dims. Various other authors have
either followed Merriam, employed Aenocyon
as a subgenus, or ignored the term (see "Rec-
ord of occurrences," below).

As pointed out above, milleri and ayersi
are synonyms of C. lupus furlongi and C.
dims, respectively, and thus there is only one
species of dire wolf, C. dims. Although this
species has the most pronounced specific
characters of any North American member of
the genus Canis, these characters do not
seem to be of a higher order than those dis-
tinguishing other species. Furthermore, there
is no evidence that the lineage of the dire
wolf was long separate from that of other
Canis. Indeed, C. dims seemingly was the
last wild species of Canis to appear in North
America, and it probably descended from the
same basic stock that gave rise to other
wolves. Thus I have not employed Aenocyon
as a genus or subgenus, and have synony-
mized it under Canis.

Record of occurrences. — The dire wolf is
represented by more fossil material than any
other species of Canis, and yet its known
geological range is relatively short, being re-
stricted to the Rancholabrean and early Re-
cent. A number of large canids from older
sites were assigned to C. dims by previous
authors, but in such cases identification was
incorrect or material was inadequate to allow
careful evaluation. Irvingtonian records from
Irvington, California (Savage, 1951:230);
Rushville fossil quarry (Schultz and Tanner,

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

111

Fig. 54. — Map showing localities (black dots) of C. dints. Because of the scale of the map, it was
not possible to plot all localities in crowded areas.

1957:71) and Angus fossil quarry (Schultz
and Martin, 1970:347), Nebraska; Port Ken-
nedy, Pennsylvania (Cope, 1899:227); and
Rock Creek, Texas (Troxell, 1915:633) are

discussed above in the account of C. arm-
brusteri. And records from Hay Springs
(Schultz, 1934:369; Schultz and Stout, 1948:
563; Schultz and Tanner, 1957:71) and Mul-

112

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

len (Martin, 1972:174), Nebraska are dis-
cussed above in the account of C. lupus. In
addition, the dire wolf was associated with
the Blancan Broadwater quarry site, Ne-
braska (Barbour and Schultz, 1937:4), and
Blanco fauna, Texas (Vanderhoof, 1937). No
particular specimens, however, were dis-
cussed, and subsequent papers on the Broad-
water quarry (Schultz and Stout, 1945:234;
1948:563; Hibbard, 1970:414) and Blanco
fauna (Meade, 1945; Johnston and Savage,
1955:36-37; Dalquest, 1975) did not mention
the presence of C. dims. Perhaps the initial
reports of C. dims had been based mistakenly
on remains of Borophagus.

The following list is arranged alphabeti-
cally by state and province, and geographi-
cally (north to south, west to east) within
states and provinces, except that Latin Amer-
ican areas are placed last. Specimens exam-
ined by me are identified by element, mu-
seum number, or both; and selected measure-
ments are found in appendix B (part 15) and
appendix C (part 7). Occurrences also are
shown on the map in figure 54.

ALBERTA.— Castleguard icefield, Banff National
Park; late Pleistocene; as C. dirus (Cowan, 1954:44).
This record is based on a single lower canine tooth
found lying on the surface.

Medicine Hat; Sangamon; as C. dirus (Churcher,
1970:63).

ARIZONA. — Ventana Cave, Papago Indian Res-
ervation, Pima County; Wisconsin (Hibbard, 1958);
as C. dims (Colbert, 1950:132).

Murray Springs, 1 mi. W Lewis Spring on San
Pedro River, Cochise County; late Pleistocene; two
mandibular fragments, UAriz 4394, 4395.

Whitewater Draw, near Douglas, Cochise Coun-
ty; early Recent; as C. dirus (Hester, 1960:69).

ARKANSAS. — Peccary Cave, eastern Newton
County; early Recent; as C. dirus (Davis, 1969:164;
Quinn, 1972:92).

CALIFORNIA.— Samwel Cave, Shasta Lake,
Shasta County; Wisconsin (Hibbard, 1958); as C.
dirus (Kurten and Anderson, 1972:37); mandibular
fragment, UCMP 9566.

Potter Creek Cave, 1 mi. SE Baird, Shasta Coun-
ty; Wisconsin (Hibbard, 1958); as C. indianensis
(Sinclair, 1904:17), as C. dirus (Anderson, 1968:
22).

Hawver Cave, 5 mi. E Auburn, El Dorado Coun-
ty; Wisconsin (Hibbard, 1958); as C. near dims
(Stock, 1918:478); as C. dirus (Anderson, 1968:22).

Cool quarry. El Dorado County; late Pleistocene;
mandibular fragment, UCMP 38328.

Teichart gravel pit, Sacramento County; late
Pleistocene; mandibular fragment, UCMP 85380.

Arroyo Las Positas, Alameda County; Pleistocene;
mandible, FM PM664.

Livermore Valley, near San Leandro, Alameda
County; Pleistocene; as C. indianensis ( Leidy, 1873:
230), as C. dirus (Merriam, 1912:244).

Oil Springs, Tulare County; Pleistocene; as C.
indianensis (Merriam, 1903:288), as C. dirus (Mer-
riam, 1912:244).

McKittrick tar seeps, Kern County; Wisconsin; as
A. dims (Merriam and Stock, 1921; Schultz, 1938b:
169); two complete skulls, LACM; six maxillary
fragments, LACM; 13 mandibles, LACM; three man-
dibles, UCMP.

Maricopa Brea, near Maricopa, Kern County;
Wisconsin (C-14 date: 13,860 B.P.); as C. dirus
(Shakespear, 1975); three complete skulls with man-
dibles, six other mostly complete skulls; 10 cranial
and maxillary fragments; 20 mostly complete mandi-
bles; many other fragments and isolated teeth, all
in LACM.

Carpinteria asphalt, Santa Barbara County; Wis-
consin (Hibbard, 1958); as A. near dirus (Wilson,
1933:69).

Rancho La Brea, Los Angeles, Los Angeles Coun-
ty; Wisconsin; as "C. indianensis (?)" (Merriam,
1906), as C. dirus (Merriam, 1912:218; Marcus,
1960:2), as A. dims (Merriam, 1918:533; Stock,
1929:286), as C. (A.) dirus (Stock, Lance, and
Nigra, 1946:109; Nigra and Lance, 1947:26; Stock
and Lance, 1948:79); 520 complete or mostly com-
plete skulls without associated mandibles, LACM;
one skull, AMNH; two skulls, KU; five skulls, USNM;
ten skulls, UCMP; 90 complete mandibles, LACM;
three mandibles, AMNH; three mandibles, USNM;
five mandibles, UCMP; numerous fragments and iso-
lated teeth, LACM. The dire wolf material from
Rancho La Brea is the most abundant of any large
mammal from any fossil site in North America. The
number of individual wolves represented in the col-
lection of the Los Angeles County Museum was
estimated at 2,000 by Stock, and was counted at
1,646 by Marcus (1960). The amount of well pre-
served material offers an unparalleled opportunity for
studies of variation in a local population of canids,
but as yet little has been done in this regard. Nigra
and Lance (1947) found the average size of meta-
podials of C. dirus to differ between the five major
tar pits from which remains were recovered. Popula-
tion studies of this kind are qualified by the prob-
ability that the different pits were active at different
times, and that each was active over a lengthy
period in which chronological mixing of elements
occurred. Radiocarbon dates based directly on speci-
mens of C. dirus from Rancho La Brea were given
as 9,860±550 and 10,710±320 B.P. by Miller ( 1968:
14). It is likely, however, that remains of wolves
were deposited at the site over thousands of years. I
found the skulls of C. dirus to be remarkably con-

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

113

sistent in the critical characters that distinguish the
species, and on the whole I agree with the detailed
description provided by Merriam in 1912. Studies
of postcranial elements from Rancho La Brea have
indicated that the dire wolf had a relatively stockier
body, lighter limbs, and shorter feet than modern
C. lupus (Merriam, 1912:236; Stock, Lance, and
Nigra, 1946; Stock and Lance, 1948:79).

Harold Beds, 5 mi. SE Palmdale, Los Angeles
County; Pleistocene; three metapodials, USNM 13085.

Harbor freeway, Los Angeles, Los Angeles Coun-
ty; Wisconsin; as C. cf. dirus (Miller, 1971:54).

San Pedro, Los Angeles County; Wisconsin; as
C. cf. dirus (Miller, 1971:45).

La Mirada, Los Angeles County; Wisconsin; as
C. cf. dirus (Miller, 1971:49).

Newport Bay Mesa, Orange County; Wisconsin;
as C. cf. dirus (Miller, 1971:34).

Costeau pit, 2 mi. S El Toro, Orange County;
Wisconsin; as C. cf. dirus (Miller, 1971:17).

FLORIDA.— Aucilla River IA, Jefferson County;
Wisconsin; as C. dirus (Webb, 1974b: 17).

Ichetucknee River, Columbia County; Wisconsin;
as C. dirus (Webb, 1974b: 17); maxillary fragment,
UF 8006; three mandibular fragments, UF 8005,
12899, 17717; three canine teeth, UF 1995, 8214,
8215; postcranial fragments, MCZ 18347-18349. The
mandibles resemble those of dire wolves from Rancho
La Brea, but are larger.

Santa Fe River IIA, Gilchrist County; Rancho-
labrean (Webb, 1974b:31; apparently incorrectly
designated as late Irvingtonian on p. 13); as C. dirus
(Webb, 1974b: 17).

Hornsby Springs, near High Springs, Alachua
County; Wisconsin; as A. mjersi (Bader, 1957:71);
maxillarv fragment, UF 3988; mandibular fragment,
UF 3987.

Haile VIIIA, Alachua County; Sangamon; as C.
dirus (Webb, 1974b: 17).

Arredondo IB, 4 mi. SW Gainesville, Alachua
County; Sangamon; as A. ayersi (Bader, 1957:54),
as C. dirus (Webb, 1974b: 17).

Devil's Den, near Williston, Levy County; late
Wisconsin or early Recent (7,000-8,000 B.P.); as
C. dims (Martin and Webb, 1974:126); incomplete
sk-ull, UF 7996.

Wekiva River, Levy County; late Pleistocene;
mandibular fragment, UF 14204.

Reddick IA, Marion County; Sangamon (Webb,
1974b:13); as C. ayersi (Gut, 1939), as C. (A.)
ayersi (Gut and Ray, 1964), as C. dirus (Martin,
1974:73); crushed skull with mandibles, UF 2923;
two crushed skulls without mandibles, UF 3081, one
unnumbered; mandibular fragment, UF; two isolated
Ml, isolated M2, P4, ml, UF; P4, MCZ.

Eichelberger Cave, 2 mi. SW Belleview, Marion
County; late Pleistocene; cast of pair of mandibles,
MCZ 7349; two mandibular fragments ( probably
from same individual), UF 1622, 1623; isolated
teeth, UF. The mandibles and lower carnassials are
the largest that I examined.

Sabertooth Cave, 1 mi. NW Lecanto, Citrus

County; Wisconsin; as C. ayersi (Simpson, 1928:9),
as C. dirus (Webb, 1974b:17).

Rock Springs, Orange County; Sangamon; as C.
dims (Webb, 1974b: 17).

Seminole Field, near St. Petersburg, Pinellas
County; Wisconsin (Hibbard, et al, 1965); as C.
(A.) ayersi (Simpson, 1929a:572), as C. dirus
(Webb, 1974b: 17); mandibular fragment, AMNH
23568; Ml, AMNH 23582; M2, AMNH 23569; two
ml, AMNH 23565, 23567; various other fragmentary
teeth, AMNH. Simpson reported that a large and
a small kind of canid were represented both at Semi-
nole Field and Sabertooth Cave. All of the material
that I examined appears to be within the range of
variation of C. dirus. A small P4 listed by Simpson
may possibly have belonged to a large red wolf.
Webb (1974b:17), and Martin and Webb (1974:
128) reported C. familiaris from Seminole Field. A
C-14 date of only 2,040±90 B.P. for this site was
questioned by Hester (1960).

Melbourne, Brevard County; Wisconsin ( Hib-
bard, et al., 1965); as C. (A.) cf. ayersi (Gazin,
1950:400), as A. cf. ayersi (Ray, 1958:433), as
C. dirus (Webb, 1974b: 17); mandible, USNM
12946; two isolated P4, two Ml, four ml, USNM.

Sebastian Canal, Brevard County; Wisconsin; as
C. dirus (Webb, 1974b: 17).

Vero (stratum 2), Indian River County; late
Wisconsin (Webb, 1974b:13); as C. ayersi (Sellards,
1916:152; Weigel, 1962:37), as C. dirus (Martin,
1974:73); skull without mandibles, FGS 7166. The
status of C. ayersi, the type of which was obtained
at Vero, is discussed in the above "remarks."

Bradenton, Manatee County; Sangamon (Webb,
1974b: 13); maxillary fragment, UF 3276; mandibular
fragment, UF 2259.

Phillipi Creek-Fruitville Ditch, 7 mi. E Sarasota,
Sarasota County; Wisconsin; as C. ayersi (Simpson,
1929b:275).

IDAHO. — Jaguar Cave, Beaverhead Mountains,
Lemhi County; late Wisconsin (C-14 dates: 10,370
±350 and 11,580±250 B.P.); as C. cf. dirus (Kur-
ten and Anderson, 1972:24).

American Falls, Power County; Rancholabrean
(Hibbard, et al, 1965), Illinoian (Kurten, 1974:7);
as ef. A. dirus (Gazin, 1935:298), as C. (A.) dirus
(Hopkins, Bonnichsen, and Fortsch, 1969:3). Ga-
zin's original faunal list for this site stated only that
distal portions of two humeri and an abraded phal-
ange could not be distinguished from corresponding
parts of dire wolves from Rancho La Brea. This
material is not reliable in the identification of C.
dirus, and might represent some other large canid.

Rainbow Beach local fauna, American Falls Res-
ervoir, Power County; Wisconsin (C-14 dates:
21,500±700 and 31,300±2,300 B.P.); as C. dirus
(McDonald and Anderson, 1975:26).

ILLINOIS. — Galena, Jo Daviess County; Wiscon-
sin (Kurten, 1974:10); as C. or A. mississippiensis
(Hay, 1923:337).

INDIANA. — Ohio River, below Evansville, Van-
derburgh County; late Pleistocene; as C. primaevus

114

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

(Leidy, 1854:200; 1856:167), as C. dims (Leidy,
1858:21; Merriam, 1912:240), as C. indianensis
(Leidy, 1869:368), as C. lupus (Cope and Wort-
man, 1884:9), as A. dims (Hay, 1923:204); maxil-
lary fragment, ANSP 11614. The nomenclatural his-
tory of the type specimen of C. dirus was reviewed
by Merriam (1912:218-221). As he noted (1912:
240-241), the type resembles specimens from Rancho
La Brea in the reduction of the hypocone of Ml,
as well as in other features that can be evaluated.
There is a pronounced buccal cingulum on the Ml,
as found in some specimens of C. dints, and the
teeth are within the size range of those from Cali-
fornia.

KANSAS. — Twelve Mile Creek, Logan County;
Pleistocene; as C. occidcntalis (Hay, 1924:143, 165);
P4, M2, p4, KU 392.

Pendennis, Lane County; Pleistocene; as C. occi-
dentalis (Hay, 1924:71); ml, KU 393.

Cragin Quarry local fauna, north of Cimarron
River, Meade Countv; Sangamon; as "C. occiden-
talis?" (Hay, 1917b:48), as A. dims ( Hibbard, 1939:
464; 1949:84; Hibbard and Taylor, 1960:178), as
C. dirus (Schultz, 1969:53); mandible, KU 4613.

KENTUCKY.— Welsh Cave, 3.5 mi. SW Troy,
Woodford County; late Wisconsin (ca. 13,000 B.P.);
as C. dirus (Guilday, Hamilton, and McCrady, 1971:
274); cast of skull without mandibles, CM 12625;
cast of mandible from different individual, CM
12625a. The specimens are among the most complete
that have been collected at sites outside of the
southwest, and in all characters they resemble speci-
mens from Rancho La Brea.

LOUISIANA.— Avery Island, Iberia Parish; late
Pleistocene; as C. dirus (Gagliano, 1967:40).

MISSOURI.— Brynjulfson Caves, 6 mi. SSE Co-
lumbia, Boone County; late Wisconsin (about 10,000
B.P.); as C. dirus (Parmalee and Oesch, 1972:31);
isolated teeth, ISM. On the basis of radii, Parmalee
and Oesch reported that the wolves from this site
were slightly larger than the huge individual found
at Powder Mill Creek Cave (see below).

Cherokee Cave, St. Louis, St. Louis County; late
Pleistocene (Webster, 1964); as Canis (Simpson,
1949:16). Simpson reported that eight metapodials
were larger than those of C. dirus, but he did not
refer them to a species. The measurements he listed
are much greater than the means given by Nigra and
Lance (1947) for the same elements of C. dirus
from Rancho La Brea; but the size of more recently
collected metapodials of Missouri C. dirus, found in
association with cranial material, is close to that of
the Cherokee Cave specimens (Galbreath, 1964;
Hawksley, 1963). Galbreath 's specimen (see ac-
count of Powder Mill Creek Cave, below) reportedly
represented a female, and a male of the same popu-
lation probably would have had metapodials as large
as those found in St. Louis.

Herculaneum (near), Jefferson County; Wiscon-
sin (Hibbard, et al, 1965); as C. dirus (Olson,
1940:42); P4, M2, FM WC1736.

Carroll Cave, Camden County; Wisconsin; as

A. dirus (Hawksley, 1963), as C. dints (Hawksley,
1965:79).

Perkins Cave, Camden County; Wisconsin; as
C. cf. dirus (Hawksley, 1965:82).

Bat Cave, 8 km. NW Waynesville, Pulaski Coun-
ty; late Wisconsin (10,000-16,000 B.P.); as A. dints
(Hawksley, 1963), as C. dirus (Hawksley, 1965:81;
Hawksley, Reynolds, and Foley, 1973:72-77). Ac-
cording to these last authors, data from Bat Cave
tended to bear out Galbreath's (1964) suggestion
that dire wolves from Missouri were larger than those
from Rancho La Brea (see account of Powder Mill
Creek Cave, below ) .

Cox Cave, Pulaski County; late Pleistocene; "pos-
sibly Canis dints" ( Mehl, 1962:44).

Bushwacker Cave, Pulaski County; Wisconsin;
as C. dirus (Hawksley, Reynolds, and Foley, 1973:
73).

Powder Mill Creek Cave, Shannon County; late
Wisconsin (C-14 date: 13,170±600 B.P.); as C.
(A.) dirus (Galbreath, 1964). Galbreath reported
the discovery of most of the skeleton, but not in-
cluding the upper parts of the skull, of a large
female dire wolf. Most of the postcranial measure-
ments were found to exceed those of even the largest
reported specimens of C. dims from Rancho La Brea
(as listed by Merriam, 1912; Nigra and Lance, 1947;
and Stock and Lance, 1948). Galbreath thus con-
sidered the limbs and feet of the Missouri individual
to be relatively larger than those of specimens from
the tar pits. The size of this individual, and of others
reported by Hawksley (1963), led Galbreath to
suggest the possibility that Missouri dire wolves
averaged larger than those of California.

Zoo Cave, 1 mi. ENE Hilda, Taney County; late
Wisconsin (9,000-13,000 B.P.); as C. dirus (Hood
and Hawksley, 1975:25, 28). According to these
authors the material from this site represented an
adult that was "quite small by Missouri standards,
closely approaching the size of Rancho La Brea
specimens."

NEBRASKA. — Heckendorf gravel pit, Stanton
County; late Pleistocene; cranial fragment, UN 2911.

NEVADA.— Gypsum Cave, 16 mi. E Las Vegas,
Clark County; late Wisconsin; as "Canis or Aenocyon
sp." (Harrington, 1933:192), as C. dints (Hester,
1960:69).

NEW MEXICO.— Conkling Cavern, near Las
Cruces, Dona Ana County; late Pleistocene; mandi-
ble, LACM.

Hermit's Cave, east slope of Guadalupe Moun-
tains, Eddy County; late Wisconsin (C-14 dates:
11,850±350 and 12,900±350 B.P.); as C. dirus
(Schultz, Martin, and Tanner, 1970); maxillary frag-
ment, UN 19212; cranial fragment, UN 19215; man-
dibular fragment, UN 19213.

OKLAHOMA. — Marlow, Stephens County; late
Pleistocene; skull with mandibles, USNM 10278.

OREGON.— Willamette Valley, near Woodburn,
Marion County; late Pleistocene; as A. dirus (Pack-
ard, 1950:89).

Fossil Lake, Lake County; early or middle Wis-

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

115

consin (Allison, 1966:32); as C. cf. dims (Elftman,
1931:5).

PENNSYLVANIA.— Frankstown Cave, Blair
County; Wisconsin (Hibbard, 1958); as C. dims
(Peterson, 1926:282); maxillary fragment, CM
11023; three mandibular fragments, CM 11022,
11024, 11026. Although the Frankstown material
is referable to C. dims, there is some approach in
dental characters to C. armbrusteri from Cumberland
Cave, only about 50 miles away. More accurate age
estimates of these two sites would be desirable, so
that we might evaluate the idea of the Frankstown
specimens representing a transition between C. arm-
brusteri and C. dims.

TENNESSEE. — Jewell Cave, near Ruskin, Dick-
son County; late Pleistocene; as "wolf" (Barr, 1961:
178), as C. dims (Corgan, 1976).

Robinson Cave, 8 mi. SW Livingston, Overton
County; Wisconsin; as C. dims (Guilday, Hamilton,
and McCrady, 1969:60).

Whitesburg, Hamblen County; late Pleistocene;
as "A. ayersi?" (Hay, 1921:95); isolated incisors and
premolars, USNM 8997.

TEXAS. — Tule Canyon, Briscoe County; Pleisto-
cene; as C. indianensis (Cope, 1895:453), as C. dims
(Merriam, 1912:242).

Slaton quarry, 5 mi. N Slaton, Lubbock County;
Sangamon (Hibbard, 1970); as Aenocyon sp. (Dal-
quest, 1967:10).

Pemberton Hill, Denton County; Sangamon; as
Aenocyon sp. (Slaughter, et al, 1962:17).

Moore Pit local fauna, Dallas, Dallas County;
Sangamon; as A. cf. dims (Slaughter, 1966b:79).

Williams Cave, southern end of Guadalupe Moun-
tains, Culberson County; late Pleistocene; as C. dims
(Ayer, 1936:608).

Scharbauer site, south of Midland, Midland
County; Wisconsin; as C. dims or ayersi (Wendorf,
Krieger, and Albritton, 1955:113).

Clamp Cave, San Saba County; early Recent;
as A. dims ( Lundelius, 1967:293).

Laubach Cave, Georgetown, Williamson County;
Wisconsin (Kurten, 1974:9); as C. cf. dims (Slaugh-
ter, 1966a:481).

Levi shelter, Travis County; late Wisconsin; as
A. dims (Lundelius, 1967:293).

Friesenhahn Cave, near Bulverde, Bexar County;
Wisconsin; as "A. dims?" (Hay, 1921:141), as A.
dims (Lundelius, 1960:38).

Kincaid shelter, Uvalde County; late Pleistocene;
as A. dims (Lundelius, 1967:293).

Blanco Creek, Bee County; late Pleistocene; as
A. ayersi (Sellards, 1940:1636).

Ingleside gravel pit, San Patricio County; Wiscon-
sin; as A. dims (Lundelius, 1962), as C. dims (Lun-
delius, 1972:12); skull without mandibles, mandibu-
lar fragment, Ml, various other fragments, TM. The
skull is crushed, but appears to be the largest of
C. dims that I examined. Lundelius (1972:12, 20)
observed that two skulls from the Ingleside fauna
resembled the type of C. ayersi more than they did
skulls from Rancho La Brea, but that there was no

basis for considering C. ayersi a separate species
from C. dims.

UTAH.— Silver Creek local fauna, 5 mi. N Park
City, Summit County; late Sangamon to early Wis-
consin; as C. cf. dims (Miller, 1976:401).

VIRGINIA.— Clark's Cave, 12 km. SW Williams-
ville, Bath County; late Wisconsin (less than 10,000
B.P.); as C. cf. dims (Guilday, 1977:69).

WEST VIRGINIA. — Rennick, Greenbrier County;
late Pleistocene; mandible, CM 24327.

WISCONSIN.— Blue Mounds, Dane County; late
Pleistocene; as C. mississippiensis (Allen, 1876:49;
Hay 1914:484; 1923:342), as C. dims (Merriam,
1912:221); four limb bones, MCZ 10988-10991.
Allen originally referred to this record as being from
only the "Lead Region of Upper Mississippi," but
Hay (1923:342) restricted the locality to Blue
Mounds. Allen thought that the great size of the
bones warranted their referral to a distinct species,
but he compared them only to a single small indi-
vidual of C. lupus. Hay's continued recognition of
C. mississippiensis also was based on scanty compara-
tive material. The measurements of length listed by
Allen actually fall within the size range of both
C. lupus and C. dims as given by Stock and Lance
(1948:82), but are closer to the means of the latter
species. It seems best for now to follow Merriam
(1912) in synonymizing C. mississippiensis under
C. dims.

AGUASCALIENTES.— Cedazo local fauna, near
City of Aguascalientes; early Rancholabrean (prob-
ably Illinoian); as C. dims (Mooser and Dalquest,
1975:788); four mandibular fragments, Midwestern
State University Department of Biology 9781-9784.
These specimens, clearly referable to C. dims, pro-
vide the only well supported record of a pre-Sanga-
mon dire wolf. Actually, however, there is some
question about the age of the fauna, as Mooser and
Dalquest (1975:783) had stated: "Early Ranchola-
brean age (Savage, 1951) is indicated. We think
the Cedazo local fauna could be as old as Yarmouth-
ian or as young as Sangamon, but favor Illinoian
Age."

JALISCO.— Lago de Chapala; late Pleistocene;
as "Canis sp., large wolf" (Downs, 1958).

ESTADO DE MEXICO.— Tequixquiac (near);
late Pleistocene; as C. dims (Merriam, 1912:243),
as A. dims (Furlong, 1925:152; Maldonado-Koerdell,
1955); cast of cranial fragment, UCMP 27615.

NUEVO LEON. — San Josecito Cave, near Aram-
berri; Wisconsin; as Aenocyon (Russell, 1960:541);
two skulls, LACM 3106, 9795; 27 cranial and maxil-
lary fragments, LACM; 30 mandibular fragments,
LACM. This large amount of material represents a
population not differing in characters from that of
Rancho La Brea.

PUEBLA. — Valsequillo, near Puebla; late Pleisto-
cene; as C. (A.) dims (Thenius, 1970:59).

PERU.— La Brea, 30 mi. SE Talara (northern
part of Peru, not mapped in Fig. 54); late Pleisto-
cene; as C. (A.) dims (Churcher, 1959).

116

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

Evolutionary position. — The species C.
dints is known only from the late Pleistocene
and early Recent, and is the most common
fossil wolf of that period. The dire wolf was
not an ancestral species, but rather a highly
specialized animal, well adapted for life in
the megafaunal community of its time. In
its large size, broad proportions, large teeth,
and other critical characters, it stood on the
opposite end of the evolutionary line from
the small species of Canis of the Pliocene and
early Pleistocene. Its initial appearance in
the southern part of the continent may be cor-
related with a northern withdrawal of C.
lupus. Hay (1927:192) speculated that C.
dims was restricted to a more southerly range
by the presence of the gray wolf in the north.

Both the geographic and phylogenetic
origin of the dire wolf are unknown. Kurten
(1968:109) suggested that C. falconeri, a
large wolf of the early Pleistocene of Europe,
might be related to C. dims of the New
World. But there is no chronological or geo-
graphic evidence to support recognition of a
connection between the two, and the meas-
urements listed by Del Campana (1913:220-
229) indicate that the skull of C. falconeri
did not closely approach that of C. dims in
size.

Martin (1974:76) suggested that a popu-
lation represented by C. armbrusteri of Cum-
berland Cave, Maryland may have given rise
to C. dims. This view is reasonable in that
the disappearance of C. armbrusteri in the
Illinoian coincided with the initial appear-
ance of C. dirus (see account of C. armbrus-
teri). Moreover, some specimens of C. arm-
brusteri approach those of C. dirus in size,
and the two species share other characters
such as in the morphology of the lower pre-
molars. As yet, however, there is no conclu-
sive evidence to indicate immediate relation-
ship between C. armbrusteri and the dire
wolf.

Apparently C. dims developed exclusively
in the Western Hemisphere, and its ancestry
probably lies in the basic stock of small

wolves represented by C. edwardii and C.
rufus. At some point in the Pleistocene, an
element of this stock, comprised of individ-
uals resembling either C. eduardii, C. rufus,
or C. armbrusteri, must have become isolated
and begun separate evolution. But how was
C. dirus able to appear suddenly all across
North America in the late Pleistocene, with
the most distinctive set of characters in the
genus Canis already fully developed?

One hypothesis that can not now be disre-
garded is that the dire wolf arose and devel-
oped in South America. There are several
pieces of evidence to support this idea. First,
C. dirus has been reported from South Amer-
ica, specifically from the La Brea tar pits near
Talara, Peru (Churcher, 1959). The species
also is known from several sites in Mexico,
and at one time probably was distributed
throughout that country and Central America.
Intriguingly, the earliest occurrence of the
species, that is supported by good evidence,
also is among the most southerly (see ac-
count of Cedazo local fauna, Aguascalientes,
above). The known range of the dire wolf
(Fig. 54) suggests a southern, warmth adapt-
ed species, in contrast to the boreal C. lupus.
The primitive stock of small wolves, repre-
sented by the living C. rufus, also seems to
have been warmth adapted to some degree.
Factors associated with one of the glaciations
may have driven an element of this stock
into South America where it eventually
evolved into C. dirus. Possibly the Sangamon
interglacial afforded the opportunity for re-
invasion of much of North America. Further
evidence is offered by available information
on large South American fossil Canis (L.
Kraglievich, 1928; J. L. Kraglievich, 1952:
63). One specimen in particular, C. nehringi
from the province of Buenos Aires in Argen-
tina, appears to have points of resemblance
to C. dims. L. Kraglievich's photographs and
measurements show that the skull of C.
nehringi shares at least the following char-
acters with C. dirus: large size and massive
proportions, broad frontal shield, prominent

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

117

sagittal crest, narrow supraoccipital shield
projecting far posteriorly, vertical plates of
palatines flaring broadly anteriorly, postpala-
tine foramina opposite posterior ends of P4,
and relatively large carnassial teeth. Of
course proof of relationship between C. dints
and C. nehringi would not in itself establish
South American origin for the dire wolf.

Although its origin remains a mystery, the
dire wolf was clearly a common mammal of
the North American late Pleistocene. The
nature of its fossil remains suggests that it
was found primarily in open lowlands, and
was a predator of its contemporary large
herbivores. The extinction of most of this
megafauna at the close of the Pleistocene, for
any or all of the reasons discussed by Martin
and Wright ( 1967 ) , probably also signaled
the end of the dire wolf. An additional factor
in the extinction of C. dims may have been a
renewed influx of gray wolves following the

withdrawal of the Wisconsin ice sheet. A
general consensus among authors who have
speculated on the behavior of C. dims is
that it was a powerful creature, but was
slower and possibly not so alert as C. lupus
(Matthew, 1916; Merriam, 1912:218; Scott,
1937:578; Stock, 1956:32; Stock and Lance,
1948). The dire wolf may not have been so
well adapted in the pursuit of the predomi-
nantly smaller, swifter herbivores that sur-
vived through the Recent, and it may have
lost in competition with the gray wolf. The
sympatric occurrence of the two species is
demonstrated by good cranial material from
Rancho La Rrea, the Maricoa Brea, San
Josecito Cave, and Hermit's Cave. Fossils of
both species also have been reported from
Fossil Lake, Jaguar Cave, Samwel Cave, Pot-
ter Creek Cave, Ventana Cave, Blue Mounds,
Brynjulfson Caves, and Medicine Hat.

SUMMARY

Systematic problems in the genus Cants
center on its paleontological history and on
Recent populations in eastern North America.
In order to investigate these matters, approxi-
mately 5,000 specimens were examined. Many
of these were complete skulls, 15 measure-
ments of which were utilized in the BMD07M
computer program of multivariate analysis.
For statistical purposes, material was sepa-
rated by sex, except for fossils and specimens
of domestic dogs.

The gray wolf (Canis lupus) and the coy-
ote (Canis latrans) are readily distinguished
from one another, and from the domestic dog
(Canis familiaris). By multivariate analysis,
only five skulls of wild Canis from northern
and western North America appeared to rep-
resent hybrids. The remaining 379 specimens
of C. lupus and 277 of C. latrans from these
regions, along with a series of 50 C. familiaris,
were used as standard groups with which to
compare individuals taken in the east. The
subspecies C. lupus hjcaon, which has been
nearly exterminated in the eastern United
States, still survives in the upper Great Lakes
region, as well as in southeastern Canada.
Nearly all specimens that had been previously
identified as hjcaon showed close statistical
affinity to the standard sample of C. lupus,
and thus were combined with that sample.

Available information indicates that by
1900 the coyote had begun to extend its range
to the east and north of the prairies. The
subspecies C. latrans thamnos, of the north-
central United States and southeastern Can-
ada, is statistically close to western C. latrans.
A few specimens, however, suggest that lim-
ited wolf-coyote hybridization has occurred
recently in southern Ontario and Quebec, and
has allowed introgression of genes from C.
lupus into C. latrans. As a result, the multi-
variate position of the coyote population now
expanding through the northeastern United
States is shifted in the direction of the wolf.

Hybridization of C. latrans and C. familiaris
also has taken place, but has not had sub-
stantial effect on wild species of Canis.

In historical time, the red wolf (C. rufus)
inhabited the region from central Texas to
the Atlantic, and from the Gulf Coast to the
Ohio Valley and Pennsylvania. The 14 earli-
est available specimens, from the part of this
region that was well separated from the orig-
inal range of the coyote, show no statistical
overlap with the standard samples of 482 C.
lupus (including 103 hjcaon), 277 C. latrans,
and 50 C. familiaris. An additional 115 skulls
collected from 1919 to 1929 in Arkansas, Lou-
isiana, southern Missouri, and eastern Okla-
homa, and previously identified as C. rufus
gregoryi, have almost the same multivariate
distribution. These skulls, plus most of the
older specimens, were combined to make a
standard red wolf sample of 125 individuals.
This sample and the standard coyote sample
were used to compare all other southeastern
material.

Series of specimens taken prior to 1930
indicate that hybridization between C. rufus
and C. latrans generally was uncommon
where their ranges approached or overlapped,
except in the Edwards Plateau area of central
Texas. Material from that area forms a sta-
tistical bridge between the ranges of varia-
tion of the two standard samples. Subse-
quently, as the red wolf became rare, hybridi-
zation increased along the Texas coast, and
in north-central Texas, eastern Oklahoma,
southern Missouri, and Arkansas. This inter-
breeding apparently allowed introgression of
red wolf genes into the expanding coyote
population, which by the 1960's had become
established in most inland areas of the south-
central states. This population is essentially
coyotelike, but is shifted statistically in the
direction of the red wolf, and contains a few
individuals that are phenotypically close to
C. rufus.

118

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

119

Material collected in the 1960's and 1970's
shows that the genetic influence of the red
wolf remained strong within 100 miles of the
Texas coast. Samples from most localities
there fall mainly between the statistical dis-
tributions of C. rufas and C. latrans. Until
about 1970, an unmodified population of the
red wolf survived in extreme southeastern

Texas and probably in adjacent parts of
Louisiana.

The genus Canis apparently arose by the
middle Pliocene (Hemphillian), and its sub-
sequent hypothetical evolution is shown in
figure 55. The relationships of C. cedazoen-
sis of Mexico are not well understood, but
otherwise the North American species can be

OLD
WORLD

NEW WORLD

RECENT

RANCHO-
LABREAN

SI ME AD AU

LATRANS

IRVING-
TONIAN

BLANCAN

CEDAZO-
ENSIS

--?--

RUFUS

ARMBRUSTERI

EDWARD I

I

OLD
WORLD

i

FAMIL-
LUPUS IARIS

DIRUS % __J

'LEPO-

Iphagus

HEMP-
HILLIAN

J

r

--?-■

Ietruscus

1

Fig. 55. — Hypothetical phylogenetic diagram of the evolution of Canis. Dashed lines indicate possible
lineages in which fossil evidence is lacking. Names of species are placed at the latest levels at which those
species are known. Only names of species recognized in this paper are shown. Horizontal and vertical dis-
tances are not necessarily to scale, and do not indicate degree of affinity. The following species are abbre-
viated: C. simensis (SI), C. mesomelas (ME), C. adustus (AD), C. aureus (AU). Question marks (?)
indicate alternative lineages.

120

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

separated into coyote and wolf groups. The
Blancan C. lepophagus was a variable entity,
with some individuals resembling small coy-
otes, and others having certain wolflike char-
acters. It is known from 15 localities, from
Florida to Idaho. The species C. latrans prob-
ably descended from certain populations of
C. lepophagus by the end of the Blancan, and
subsequently there appears to have been
little change in the coyote line. Some Pleisto-
cene coyotes, especially those of the late
Rancholabrean of California, became larger
than modern C. latrans, but others were about
the same size or smaller. Fossil C. latrans has
been reported from 109 localities across North
America, from Florida to Alaska, and from
Oaxaca to Pennsylvania.

It is not clear whether the wolf group de-
scended from some population of C. lepopha-
gus, or was already distinct from the coyote
line throughout the Blancan. Several species
of wolves appeared by the early Irvingtonian,
the first of which was probably C. edwardii.
This was a small species, known from a few
localities in the southwestern quarter of
North America. An immediate relative, C.
rufus of the southeast, continued to represent
the primitive stock of wolves until Recent
times. Still another Irvingtonian wolf, C.
armbrusteri, was much larger, but resembled
the red wolf in certain dental characters and
skull proportions. It is known by good ma-
terial only from Maryland and Florida,
though fragmentary remains from elsewhere
suggest that it may once have occurred all
across the continent.

During the early Pleistocene, an element

of the primitive stock of small wolves appar-
ently entered Eurasia where it gave rise to
C. lupus. In the course of the Illinoian gla-
ciation this species probably moved into
North America, where its fossils have been
reported from 58 localities. Some Ranchola-
brean gray wolves were remarkably small,
but others, particularly those of Rancho La
Brea, had massive skulls resembling those of
some modern Arctic wolves.

The extinct C. dims did not appear in
North America until the Rancholabrean, and
may have originated in South America or
descended from C. armbrusteri. This large,
highly specialized species was not ancestral
to modern wolves, and its skull is easily dis-
tinguished from that of C. lupus. By multi-
variate analysis, there was no overlap be-
tween 62 specimens of C. dims from Rancho
La Brea and 467 specimens of Recent C.
lupus. The dire wolf has been reported from
96 localities and apparently was common
throughout that part of North America to
the south of Canada.

The generic name Aenocijon, sometimes
applied to the dire wolf, is here synonymized
under Canis. The following names no longer
are considered to represent separate species:
C. caneloensis, C. irvingtonensis, and C. rivi-
veronis ( all are fossil subspecies of C. la-
trans); C. andersoni (a synonym of C. latrans
orcutti); C. petrolei (a synonym of C. famil-
iaris); C. priscolatrans (a fossil subspecies of
C. rufus); C. milleri (a synonym of C. lupus
furlongi); and C. ayersi (a synonym of C.
dims ) .

LITERATURE CITED

Aldous, C. M.

1939. Coyotes in Maine. Jour. Mamm., 20:104-
106.
Akersten, W. A.

1970. Interpretation of sediments and vertebrate
fossils in fill of Red Light Bolson, south-
eastern Hudspeth County, Texas. In Geol-
ogy of the southern Quitman Mountains
area, Trans-Pecos, Texas. Soc. Econ. Pa-
leontol. and Mineral., publ. 70-12, pp. 82-
87.
1972. Red Light local fauna (Blancan) of the
Love Formation, southeastern Hudspeth
County, Texas. Bull. Texas Mem. Mus.,
no. 20, 53 pp.
Allen, CM.

1920. Dogs of the American aborigines. Bull.
Mus. Comp. Zool., 63:431-517.

1942. Extinct and vanishing mammals of the West-
ern Hemisphere. Amer. Comm. Int. Wildl.
Prot., xv+620 pp.

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. Jour. Sci., ser. 3, 11:
47-51.

1896. On mammals collected in Bexar County and
vicinity, Texas, by Mr. H. P. Attwater, with
field notes by the collector. Bull. Amer.
Mus. Nat. Hist., 8:47-80.
Allison, I. S.

1966. Fossil Lake Oregon. Oregon State Univ.
Studies in Geol., no. 9, 48 pp.
Alvarez, T.

1963. Restos de mamiferos encontrados en una
cueva de Valle Nacional, Oaxaca, Mexico.
Rev. Biol. Trop., 11:57-61.

1965. Catalogo paleomastozoologico Mexicano.
Inst. Nac. Antro. Hist., Mexico, Dept. Pre-
hist., publ. 17, 70 pp.
Anderson, E.

1968. Fauna of the Little Box Elder Cave, Con-
verse County, Wyoming. Univ. Colorado
Studies, Earth Sci. Ser., no. 6, 59 pp.

1974. A survey of the late Pleistocene and Holo-
cene mammal fauna of Wyoming. Geol.
Surv. Wyoming Rept. Invest., 10:76-87.
Anderson, R. M.

1943. Summary of the large wolves of Canada,
with description of three new Arctic races.
Jour. Mamm., 24:386-393.

1946. Catalogue of Canadian Recent mammals.

Bull. Natl. Mus. Canada, 102:i-v+l-238.
Arkansas Game and Fish Commission

1951. A survey of Arkansas game. Little Rock,

iv+155 pp.

Arnold, D. A.

1952. About wolves. Michigan Conserv., 21(1):
23-25.
Atkins, D. L., and L. S. Dillon

1971. Evolution of the cerebellum in the genus
Canis. Jour. Mamm., 52:96-107.
Audubon, J. J., and J. Bachman

1851. The quadrupeds of North America. New
York, vol. 2, 334 pp.
Ayer, M. Y.

1936. The archaeological and faunal material from
Williams Cave, Guadalupe Mountains,
Texas. Proc. Acad. Nat. Sci., Philadelphia,
88:599-618.

Bader, R. S.

1957. Two Pleistocene mammalian faunas from
Alachua County, Florida. Bull. Florida
State Mus., Biol. Ser., 2:53-75.

Bailey, B.

1929. Mammals of Sherburne County, Minnesota.
Jour. Mamm., 10:153-164.

Bailey, H. H.

1930. Correcting inaccurate ranges of certain Flor-
ida mammals and others of Virginia and the
Carolinas. Bull. Bailey Mus. and Lib. Nat.
Hist., no. 5.

Bailey, V.

1905. Biological survey of Texas. N. Amer. Fauna,

no. 25, 222 pp.
1907. Wolves in relation to stock, game, and the
national forest reserves. U.S. Dept. Agr.
Forest Serv. Bull., no. 72, 31 pp.
Baird, S. F.

1857. General report upon the zoology of the sev-
eral Pacific railroad routes. Part I: Mam-
mals. Washington, D.C., xlviii-f757 pp.
Bangs, O.

1898. The land mammals of peninsular Florida
and the coast region of Georgia. Proc.
Boston Soc. Nat. Hist., 28:157-235.
Barbour, E. H., and C. B. Schultz

1937. An early Pleistocene fauna from Nebraska.
Amer. Mus. Novit., no. 942, 10 pp.

Barbour, T.

1944. That vanishing Eden. A naturalist's Florida.
Little, Brown and Co., Boston, 250 pp.
Barkalow, F. S., Jr.

1976. (Erroneously labeled as 1972.) Vertebrate
remains from archeological sites in the Ten-
nessee Valley of Alabama. Southern Indian
Studies, 24:3-53.
Barr, T. C.

1961. Caves of Tennessee. Bull. Dept. Conserv.
and Commerce, Tennessee, Div. Geol., no.
64, \ii+567 pp.
Bartram, W.

1791. Travels. Philadelphia, xxxiv+522 pp.
Bee, J. W., and E. R. Hall

1951. An instance of coyote-dog hybridization.
Trans. Kansas Acad. Sci., 54:73-77.

121

122

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

Beezley, C.

1967. Marsh fugitive. Texas Parks and Wildl., 25
(l):18-20.
Benxitt, B., and W. O. Nagel

1937. A survey of the resident game and furbear-
ers of Missouri. Univ. Missouri Studies, 12
(2):1-215.
Bjobk, P. B.

1970. The Carnivora of the Hagerman local fauna
(late Pliocene) of southwestern Idaho.
Trans. Amer. Phil. Soc, new ser., 60:3-54.

Black, J. D.

1936. Mammals of northwestern Arkansas. Jour.
Mamm., 17:29-35.
Black, J. H., and T. L. Best

1972. Remains of a gray wolf (Canis lupus) from
northwestern Oklahoma. Proc. Oklahoma
Acad. Sci., 52.120.
Bray, W. L.

1904. The timber of the Edwards Plateau of
Texas. U.S. Dept. Agr., Bur. Forestry
Bull., no. 49, 30 pp.

Bromley, A. W.

1956. Adirondack coyotes. New York State Con-
serv., 10(4) :8-9.
Brown, B.

1905. The Conard fissure. Mem. Amer. Mus. Nat.
Hist., 9:157-208.

Bullen, B. P., and C. A. Benson

1967. Cut wolf jaws from Tick Island, Florida.
Florida Anthropol., 20:175-177.
Bump, G.

1941. The introduction and transportation of game
birds and mammals into the state of New
York. Trans. N. Amer. Wildl. Conf., 5:409-
420.
Burt, W. H.

1946. The mammals of Michigan. Univ. Michi-
gan Press, xv-f-288 pp.
Cahalane, V. H.

1964. A preliminary study of distribution and
numbers of cougar, grizzly and wolf in
North America. New York Zool. Soc, 12
pp.
Cahn, A. R.

1921. The mammals of Itasca County, Minnesota.
Jour. Mamm., 2:68-74.
Carley, C. J., and H. McCarley

1976. An evaluation of red wolf (Canis rufus)
hybridization in southeast Texas. Paper No.
153 at 56th Ann. Mtg. Amer. Soc. Mamm.,
Texas Tech Univ., Lubbock.
Carpenter, M.

1971. Some recent coyote records in Virginia.
Virginia Wildl., 32(6): 14-15.

Carson, H. S.

1962. Covote, coy-dog, or dog. Maine Fish &
Game, 4(1) :4-7.

Catesby, M.

1771. The natural history of Carolina, Florida,
and the Bahama Islands. London, 2 vols.

Chambers, B. E., P. N. Gaskin, R. A. Post, and
S. A. Cameron

1974. The coyote. Conservationist, 29(2) :5-7.
Chapman, F. M.

1894. Remarks on certain land mammals from
Florida, with a list of the species known to
occur in the state. Bull. Amer. Mus. Nat.
Hist., 6:333-346.

Chiarelli, A. B.

1975. The chromosomes of the Canidae. In Fox
(1975), pp. 40-53.

Churcher, C. S.

1959. Fossil Canis from the tar pits of La Brea,
Peru. Science, 130:564-565.

1969a. The vertebrate fauna of Surprise, Mitchell
and Island bluffs, near Medicine Hat, Al-
berta. Mid-Western Friends of the Pleisto-
cene, 5 pp.

1969b. A fourth report on some Pleistocene locali-
ties in southern Alberta and Saskatchewan
and their vertebrate fossil faunas. Geol.
Surv. Canada, viii-|-410 pp.

1970. A fifth report on some Pleistocene locali-
ties in southern Alberta and Saskatchewan
and their vertebrate fossil faunas. Geol.
Surv. Canada, v-(-133 pp.
Clarke, C. H. D.

1970. Wolf management in Ontario. In Jorgen-
sen, Faulkner, and Mech (1970), pp. 19-23.
Cleland, C. E.

1966. The prehistoric animal ecology and ethno-
zoology of the upper Great Lakes region.
Univ. Michigan Anthropol. Papers, no. 29,
x+294 pp.
Clutton-Brock, J., G. B. Corbett, and M. Hills

1976. A review of the family Canidae, with a
classification by numerical methods. Bull.
British Mus. (Nat. Hist.), Zool., 29:119-
199.

Cockrum, E. L.

1952. Mammals of Kansas. Univ. Kansas Publ.
Mus. Nat. Hist., 7:1-303.
Colbert, E. H.

1950. The fossil vertebrates. In Haury, E. W.,
The stratigraphy and archaeology of Ven-
tana Cave Arizona, Univ. Arizona Press,
Tucson, pp. 126-148.
Cole, F., and T. Deuel

1937. Rediscovering Illinois/archeological explora-
tions in and around Fulton County. Univ.
Chicago Press, xvi-|-295 pp.
Cook, R.

1952. The coy-dog: hybrid with a future? Jour.
Hered., 43:71-73.
Cope, E. D.

1879. On the genera of Felidae and Canidae.
Proc. Acad. Nat. Sci., Philadelphia, 1879,
pp. 168-194.

1895. Extinct Bovidae, Canidae and Felidae from
the Pleistocene of the plains. Jour. Acad.
Nat. Sci., Philadelphia, ser. 2, 9:453-459.

1899. Vertebrate remains from Port Kennedy bone

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

123

deposit. Jour. Acad. Nat. Sci., Philadelphia,
ser. 2, 11:193-267.
Cope, E. D., and J. L. Wortman

1884. Post-Pliocene vertebrates of Indiana. Ann.
Rept. State Geol. Indiana, 14:1-62.

CORGAN, J.

1976. Vertebrate fossils of Tennessee. Tennessee
Div. Geol. Bull., no. 77, 100 pp.
Cory, C. B.

1912. The mammals of Illinois and Wisconsin.
Field Mus. Nat. Hist. Pub!., no. 153, Zool.
Ser., 11:1-505.

Coues, E.

1873. The prairie wolf or coyote: Canis Jatrans.
Amer. Nat., 7:385-389.
Cowan, I. M.

1954. The occurrence of the Pleistocene wolf,
Canis dims, in the Rocky Mountains of cen-
tral Alberta. Canadian Field-Nat., 68:44.
Cunningham, V. D., and R. D. Dunford

1970. Recent coyote record from Florida. Quart.
Jour. Florida Acad. Sci., 33:279-280.
Dalquest, W. W.

1961. Two species of bison contemporaneous in
early Recent deposits in Texas. South-
western Nat., 6:73-78.

1964. A new Pleistocene local fauna from Motley
County, Texas. Trans. Kansas Acad. Sci.,
67:499-505.

1965. New Pleistocene formation and local fauna
from Hardeman County, Texas. Jour. Pa-
leontol., 39:63-79.

1967. Mammals of the Pleistocene Slaton local
fauna of Texas. Southwestern Nat., 12:1-30.

1968. Mammals of north-central Texas. South-
western Nat., 13:13-21.

1975. Vertebrate fossils from the Blanco local
fauna of Texas. Occas. Papers Mus. Texas
Tech Univ., no. 30, 52 pp.
Dalquest, W. W., E. Roth, and F. Judd

1969. The mammal fauna of Schulze Cave, Ed-
wards County, Texas. Bull. Florida State
Mus., Biol. Sci., 13:205-276.

Davis, L. C.

1969. The biostratigraphy of Peccary Cave, New-
ton County, Arkansas. Proc. Arkansas Acad.
Sci., 23:192-196.

Davis, W. B.

1966. The mammals of Texas. Texas Game and
Fish Comm., Austin, 267 pp.

De Vos, A.

1964. Range changes of mammals in the Great
Lakes region. Amer. Midi. Nat., 71:210-
231.

DEGERB0L, M.

1961. On a find of preboreal domestic dog (Canis
familiaris L.) from Star Carr, Yorkshire,
with remarks on other Mesolithic dogs.
Proc. Prehist. Soc, 27:35-55.
Del Campana, D.

1913. I cani pliocenici di Toscana. Palaeontogr.
Italica, 19:189-254.

Dellinger, S. C, and J. D. Black

1940. Notes on Arkansas mammals. Jour. Mamm.,
21:187-191.
Dice, L. R.

1925. A survey of the mammals of Charlevoix
County, Michigan, and vicinity. Occas. Pa-
pers Univ. Michigan Mus. Zool., no. 159,
33 pp.
1942. A family of dog-coyote hybrids. Jour.
Mamm., 23:186-192.
Dixon, W. J.

1970. BMD biomedical computer programs. Univ.
California Press, Berkeley, x+600 pp.
Douglass, D. W.

1970. History and status of the wolf in Michigan.
In Jorgensen, Faulkner, and Mech (1970),
pp. 6-8.
Downs, T.

1958. Fossil vertebrates from Lago de Chapala,
Jalisco, Mexico. Cong. Geol. Int., Mexico
City, Sec. 7 — Paleontol., Taxon., and
Evol., pp. 75-77.

Downs, T., H. Howard, T. Clements, and G. A.
Smith

1959. Quaternary animals from Schuiling Cave in
the Mojave Desert, California. Los Angeles
Co. Mus. Contrib. Sci., no. 29, 21 pp.

Du Bar, J. R., and G. Clopine

1962. Late Pleistocene deposits in the vicinity of
Houston, Texas: a preliminary investiga-
tion. Trans. Gulf Coast Assoc. Geol. Socie-
ties, 11:83-108.
Duck, L. G., and J. B. Fletcher

1945. A survey of the game and furbearing ani-
mals of Oklahoma. Oklahoma Game and
Fish Comm., 144 pp.
Eaton, G. F.

1923. Vertebrate fossils from the Miiia Erupcion.
Amer. Jour. Sci., ser. 5, 6:229-238.
Elder, W. H., and C. M. Hayden

1977. Use of discriminant function in taxonomic
determination of canids from Missouri.
Jour. Mamm., 58:17-24.
Elftman, H. O.

1931. Pleistocene mammals of Fossil Lake, Ore-
gon. Amer. Mus. Novit., no. 481, 21 pp.
Ellerman, J. R., and T. C. S. Morrison-Scott

1951. Checklist of Palaearctic and Indian mam-
mals. British Mus., London, 810 pp.
Fine, M. D.

1964. An abnormal p2 in Canis cf. C. latrans
from the Hagerman fauna of Idaho. Jour.
Mamm., 45:483-485.
Fisher, O.

1841. Sketches of Texas in 1840. Walters and
Weber, Springfield, Illinois, vii-(-64 pp.
Fox, M. W. (ed.)

1975. The wild canids/their systematics, behav-
ioral ecology and evolution. Van Nostrand
Reinhold, New York, xvi4-508 pp.

Freeman, R. C.

1976. Coyote x dog hybridization and red wolf
influence in the wild Canis of Oklahoma.

124

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

Unpublished M. A. thesis, Northeastern
Oklahoma State Univ., vii+62 pp.
Frick, C.

1930. Alaska's frozen fauna. Nat. Hist., 30(1):
71-80.

FUNKHOUSER, W. D.

1925. Wildlife in Kentucky. Kentucky Geol. Surv.,
Frankfort, 385 pp.
Furlong, E. L.

1925. Notes on the occurrence of mammalian re-
mains in the Pleistocene of Mexico, with
description of a new species Capromeryx
mexicana. Univ. California Publ. Geol. Sci.,
15:137-152.
Gabrielson, I. N.

1936. Report of the chief of the Bureau of Bio-
logical Survey. Washington, D.C.
Gagliano, S. M.

1967. Occupation sequence at Avery Island. Lou-
isiana State Univ. Coastal Studies Ser., no.
22, xiii+110 pp.
Galbreath, E. C.

1938. Post-glacial fossil vertebrates from east-
central Illinois. Field Mus. Nat. Hist. Geol.
Ser., 6:303-313.
1964. A dire wolf skeleton and Powder Mill
Creek Cave, Missouri. Trans. Illinois State
Acad. Sci., 57:224-242.
Gale, L. R., and R. Pierce

1954. Occurrence of the coyote in Kentucky.
Jour. Mamm., 35:256-258.

Ganier, A. F.

1928. The wild life of Tennessee. Jour. Tennes-
see Acad. Sci., 3:10-22.
Cabman, H.

1894. A preliminary list of the vertebrate ani-
mals of Kentucky. Bull. Essex Inst., 26:
1-63.
Gazin, C. L.

1935. Annotated list of Pleistocene mammalia
from American Falls, Idaho. Jour. Washing-
ton Acad. Sci., 25:297-302.
1942. The late Cenozoic vertebrate faunas from
the San Pedro Valley, Arizona. Proc. U.S.
Natl. Mus., 92:475-518.
1950. Annotated list of fossil mammalia associated
with human remains at Melbourne, Florida.
Jour. Washington Acad. Sci., 40:397-404.
Geist, O. W.

1955. Vertebrate paleontological reconaissance of
the Old Crow River area, Yukon Territory,
Canada. Bull. Geol. Soc. Amer., 66:1702.

Georges, S.

1976. A range extension of the coyote in Quebec.
Canadian Field-Nat., 90:78-79.
Getz, L. L.

1960. Middle Pleistocene carnivores from south-
western Kansas. Jour. Mamm., 41:361-365.
Gidley, J. W.

1913. Preliminary report on a recently discovered
Pleistocene cave deposit near Cumberland,
Maryland. Proc. U. S. Natl. Mus., 46:93-
102.

Gidley, J. W., and C. L. Gazin

1938. The Pleistocene vertebrate fauna from Cum-
berland Cave, Maryland. Bull. U. S. Natl.
Mus., no. 171, vi+99 pp.
Gier, H. T.

1968. Coyotes in Kansas. Kansas State Univ. Agr.
Exp. Sta. Bull., no. 393, 118 pp.

GlFFORD, C. L., AND R. WhTTEBREAD

1951. Mammal survey of south central Pennsyl-
vania. Pennsylvania Game Coram., 75 pp.
Giles, E.

1960. Multivariate analysis of Pleistocene and Re-
cent coyotes (Canis latrans) from Cali-
fornia. Univ. California Publ. Geol. Sci.,
36:369-390.
Gipson, P. S.

1972. The taxonomy, reproductive biology, food
habits and range of wild Canis (Canidae)
in Arkansas. Unpublished Ph.D. disserta-
tion, Univ. Arkansas, viii-f-188 pp.
1976. Melanistic Canis in Arkansas. Southwestern
Nat., 21:123-139.
Gipson, P. S., I. K. Gipson, and J. A. Sealandeb
1975. Reproductive biology of wild Canis (Cani-
dae) in Arkansas. Jour. Mamm., 56:605-
612.
Gipson, P. S., J. A. Sealander, and J. E. Dunn

1974. The taxonomic status of wild Canis in
Arkansas. Syst. Zool., 23:1-11.

Goertz, J. W., L. V. Fitzgerald, and R. M. Nowak

1975. The status of wild Canis in Louisiana.
Amer. Midi. Nat., 93:215-218.

Goldman, E. A.

1937. The wolves of North America. Jour.

Mamm., 18:37-45.
1944. Classification of wolves. Part II in Young,
S. P., and E. A. Goldman, The wolves of
North America, Amer. Wildl. Inst., Wash-
ington, D.C, pp. 389-636.
Golley, F. B.

1966. South Carolina mammals. Charleston Mus.,
xiv-f 181 pp.
Goodpaster, W. W., and D. F. Hoffmeister

1968. Notes on Ohioan mammals. Ohio Jour.
Sci., 68:116-117.

Goodwin, G. C.

1936. Big game animals in the northeastern
United States. Jour. Mamm., 17:48-50.

1969. Mammals from the state of Oaxaca, Mex-
ico, in the American Museum of Natural
History. Bull. Amer. Mus. Nat. Hist., 141:
1-269.

Graham, R.

1959. Additions to the Pleistocene fauna of Sam-
wel Cave, California. I. Canis lupus and
Canis latrans. Cave Studies, Univ. Cali-
fornia Mus. Vert. Zool., 10:54-67.
Gray, A. P.

1972. Mammalian hybrids. Commonwealth Agr.
Bur., Slough, England, x+262 pp.
Green, M.

1948. A new species of dog from the lower Plio-

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

125

cene of California. Univ. California Publ.
Bull. Dept. Geol. Sci., 28:81-90.
Grinnell, J., J. S. Dixon, and J. M. Linsdale

1937. Fur-bearing mammals of California. Univ.
California Press, Berkeley, 2:i-xiv+377-777.

GUILDAY, J. E.

1962. The Pleistocene local fauna of the Natural
Chimneys, Augusta County, Virginia. Ann.
Carnegie Mus., 36:87-122.

1969. Bone refuse from the Lamoka Lake site.
In Ritchie (1969), pp. 54-59.

1971. Biological and archaeological analysis of
bones from a 17th century Indian village
(46 PU 31), Putnam County, West Vir-
ginia. West Virginia Geol. and Econ. Surv.,
Rept. Archeol. Invest., no. 4, vii+64 pp.

1977. The Clark's Cave bone deposit and the late
Pleistocene paleoecology of the central Ap-
palachian Mountains of Virginia. Bull. Car-
negie Mus. Nat. Hist., no. 2, 87 pp.

GUILDAY, J. E., AND M. S. BENDER

1958. A recent fissure deposit in Bedford County,
Pennsylvania. Ann. Carnegie Mus., 35:127-
138.
Guilday, J. E., H. W. Hamilton, and A. D. Mc-
Crady

1969. The Pleistocene vertebrate fauna of Robin-
son Cave, Overton County, Tennessee. Pa-
laeovertebrata, 2:25-75.

1971. The Welsh Cave peccaries (Platygonus)
and associated fauna, Kentucky Pleistocene.
Ann. Carnegie Mus., 43:249-320.
Guilday, J. E., and P. W. Parmalee

1965. Animal remains from the Sheep Rock shel-
ter (36 HU 1), Huntingdon County, Penn-
sylvania. Pennsylvania Archaeol., 35:34-49.
Guilday, J. E., P. W. Parmalee, and D. P. Tanner

1962. Aboriginal butchering techniques at the
Eschelman site (36 La 12), Lancaster
County, Pennsylvania. Pennsylvania Ar-
chaeol., 32:59-83.
Guilday, J. E., and D. P. Tanner

1962. Animal remains from the Quaker State
Rockshelter (36 Ve 27), Venango County,
Pennsylvania. Pennsylvania Archaeol., 32:
131-137.

1965. Vertebrate remains from the Mount Carbon
site (46-Fa-7), Fayette County, West Vir-
ginia. West Virginia Archeol., 18:1-14.

1966. Vertebrate remains from the Fairchance
Mound (46 Mr 13), Marshall County, West
Virginia. West Virginia Archeol., 21:41-54.

Gut, H. J.

1939. Additions to the recorded Pleistocene mam-
mals from Ocala, Florida. Proc. Florida
Acad. Sci., 3:54-55.
Gut, H. J., and C. E. Ray

1964. The Pleistocene vertebrate fauna of Red-
dick, Florida. Quart. Jour. Florida Acad.
Sci., 26:315-328.
Guthrie, R. D.

1968. Paleoecology of the large-mammal commu-

nity in interior Alaska during the late Pleis-
tocene. Amer. Midi. Nat., 79:346-363.

Haag, W. G.

1948. An osteometric analysis of some aboriginal
dogs. Univ. Kentucky Repts. Anthropol.,
7:107-264.

Hager, M. W.

1972. A late Wisconsin-Recent vertebrate fauna
from the Chimney Rock animal trap, Lari-
mer County, Colorado. Univ. Wyoming
Contrib. Geol., 11:63-71.

Hall, E. R.

1943. Cranial characters of a dog-coyote hybrid.

Amer. Midi. Nat., 29:371-374.
1965. Names of species of North American mam-
mals north of Mexico. Univ. Kansas Mus.
Nat. Hist. Misc. Publ., no. 43, 16 pp.
Hall, E. R., and K. R. Kelson

1952. Comments on the taxonomy and geographic
distribution of some North American mar-
supials, insectivores and carnivores. Univ.
Kansas Publ. Mus. Nat. Hist., 5:319-341.

1959. The mammals of North America. Ronald
Press, New York, 2:i-viii+547-1083+l-79.
Halloran, A. F.

1958. Black red wolves. Oklahoma Wildl., 14
(4):6-8.

1959. Shadow in the woods. Texas Game and
Fish, 17(12):10-11.

1960. Black red wolf map. Texas Game and Fish,
18(3):25.

1961. The carnivores and ungulates of the Aran-
sas National Wildlife Refuge, Texas. South-
western Nat., 6:21-26.

1963. A melanistic coyote from Oklahoma. South-
western Nat., 8:48-49.
Halloran, A. F., and B. P. Glass

1959. The carnivores and ungulates of the Wich-
ita Mountains Wildlife Refuge, Oklahoma.
Jour. Mamm., 40:360-370.
Hamilton, W. J., Jr.

1943. The mammals of eastern United States.
Comstock Publ. Co., Ithaca, New York, 432
pp.
Hamnett, W. L., and D. C. Thornton

1953. Tar heel wildlife. North Carolina Res.
Comm., Raleigh.

Handlan, J. W.

1946. Hunter slays Monroe "wolf." West Virginia
Conserv., 10(8): 13, 23.

Handley, C. O., Jr., and C. P. Patton

1947. Wild mammals of Virginia. Virginia Comm.
Game and Inland Fisheries, Richmond, vi-f-
220 pp.

Harincton, C. R., and F. V. Clulow

1973. Pleistocene mammals from Gold Run Creek,
Yukon Territory. Canadian Jour. Earth Sci.,
10:697-759.

Harlan, R.

1825. Fauna Americana. Philadelphia, x+318 pp.
Harper, F.

1927. The mammals of the Okefinokee Swamp

126

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

region of Georgia. Proc. Boston Soc. Nat.
Hist, 38:191-396.
1942. The name of the Florida wolf. Jour.
Mamm., 23:339.
Harrington', M. R.

1933. Gypsum Cave, Nevada. Southwest Mus.
Papers, no. 8, ix-f-197 pp.
Harris, A. H.

1970. The Dry Cave mammalian fauna and late
pluvial conditions in southeastern New Mex-
ico. Texas Jour. Sci., 22:3-27.
Harris, A. H., and J. J. Findley

1964. Pleistocene-Recent fauna of the Isleta Caves,
Bernalillo County, New Mexico. Amer.
Jour. Sci., 262:114-120.

Harrison, D. L.

1968. The mammals of Arabia. Ernest Benn, Ltd.,
London, 2 :i-xiv+ 193-381.

1973. Some comparative features of the skulls of
wolves (Canis lupus Linn.) and pariah dogs
(Canis familiaris Linn.) from the Arabian
Peninsula and neighbouring lands. Bonn
Zool. Beitr., 24:185-191.
Hawksley, O.

1963. The dire wolf in Missouri. Missouri Speleol.,
5:63-72.

1965. Short-faced bear (Arctodus) fossils from
Ozark caves. Bull. Natl. Speleol. Soc, 27:
77-92.

Hawksley, O, J. F. Reynolds, and R. L. Foley

1973. Pleistocene vertebrate fauna of Bat Cave,
Pulaski County, Missouri. Bull. Natl.
Speleol. Soc, 35:61-87.
Hay, O. P.

1902. Bibliography and catalogue of the fossil
vertebrata of North America. U. S. Geol.
Surv., 868 pp.

1914. The Pleistocene mammals of Iowa. Rept.
Iowa Geol. Surv., 23:1-662.

1917a. Vertebrata mostly from stratum number 3,
at Vera, Florida, together with descriptions
of new species. Ann. Rept. Florida Geol.
Surv., 9:43-68.

1917b. On a collection of fossil vertebrates made
by Dr. F. W. Cragin in the Equus beds of
Kansas. Kansas Univ. Sci. Bull., 10:39-51.

1918. Quaternary vertebrates in southeastern Wis-
consin. In Alden, W. C, The Quaternary-
geology of southeastern Wisconsin, U. S.
Geol. Surv. Prof. Paper, no. 106, pp. 346-
347.

1920. Descriptions of some Pleistocene vertebrates
found in the United States. Proc. U. S.
Natl. Mus., 58:83-146.

1921. Descriptions of species of Pleistocene verte-
brata, types or specimens of which are
preserved in the United States National
Museum. Proc. U. S. Natl. Mus., 59:599-
642.

1923. The Pleistocene of North America and its
vertebrated animals from the states east of
the Mississippi River and from the Cana-

dian provinces east of longitude 95°. Car-
negie Inst. Washington Publ., no. 322, viii
+ 499 pp.

1924. The Pleistocene of the middle region of
North America and its vertebrated animals.
Carnegie Inst. Washington Publ., no. 322A,
vii-f-385 pp.

1927. The Pleistocene of the western region of
North America and its vertebrated animals.
Carnegie Inst. Washington Publ., no. 322B,
vii+346 pp.

1929-1930. Second bibliography and catalogue of
the fossil vertebrata of North America. Car-
negie Inst. Washington, 2 vols.
Hendrickson, J., AiND W. L. Robinson

1975. Status of the wolf in Michigan, 1973.
Amer. Midi. Nat., 94:226-232.
Hester, J. J.

1960. Late Pleistocene extinction and radiocarbon
dating. Amer. Antiquity, 26:58-77.
Hibbard, C. W.

19.38. An upper Pliocene fauna from Meade Coun-
ty, Kansas. Trans. Kansas Acad. Sci., 40:
239-265.

1939. Notes on some mammals from the Pleisto-
cene of Kansas. Trans. Kansas Acad. Sci.,
42:463-479.

1941a. Mammals of the Rexroad fauna from the
upper Pliocene of southwestern Kansas.
Trans. Kansas Acad. Sci., 44:265-313.

1941b. Paleoecology and correlation of the Rexroad
fauna from the upper Pliocene of south-
western Kansas, as indicated by the mam-
mals. Univ. Kansas Sci. Bull., 27:79-104.

1949. Pleistocene stratigraphy and paleoecology
of Meade County, Kansas. Contrib. Mus.
Paleontol. Univ. Michigan, 7:63-90.

1953. Equus (Asinus) calobatus Troxell and asso-
ciated vertebrates from the Pleistocene of
Kansas. Trans. Kansas Acad. Sci., 56:111-
126.

1955. Pleistocene vertebrates from the upper Be-
cerra (Becerra Superior) formation, Valley
of Tequixquiac, Mexico, with notes on other
Pleistocene forms. Contrib. Mus. Paleontol.
Univ. Michigan, 12:47-96.

1956. Vertebrate fossils from the Meade forma-
tion of southwestern Kansas. Papers Mich-
igan Acad. Sci., Arts, and Letters, 41:145-
203.

1958. Summary of North American Pleistocene
mammalian local faunas. Papers Michigan
Acad. Sci., Arts, and Letters, 43:3-32.

1970. Pleistocene mammalian local faunas from
the Great Plains and central lowland prov-
inces of the United States. In Dort, W.,
Jr., and J. K. Jones, Jr., Pleistocene and
Recent environments of the central Great
Plains, Univ. Kansas Dept. Geol., Spec.
Publ. no. 3, pp. 395-433.
Hibbard, C. W., and W. W. Dalquest

1966. Fossils from the Seymour formation of
Knox and Baylor counties, Texas, and their

1979

NOWAK: NORTH AMERICAN QUATERNARY CAWS

127

bearing on the late Kansan climate of that
region. Contrib. Mus. Paleontol. Univ.
Michigan, 21:1-66.
Hibbard, C. W., J. C. Frye, and A. B. Leonard

1944. Reconaissance of Pleistocene deposits in
north-central Kansas. Kansas State Geol.
Surv. Bull., no. 52, 28 pp.
Hibbard, C. W., C. E. Ray, D. E. Savage, D. W.
Taylor, and J. E. Guilday

1965. Quaternary mammals of North America.
In Wright, H. E., and D. G. Frey (eds.),
The Quaternary of the United States,
Princeton Univ. Press, pp. 509-525.
Hibbard, C. W., and E. S. Riggs

1949. Upper Pliocene vertebrates from Keefe Can-
yon, Meade County, Kansas. Bull. Geol.
Soc. Amer., 60:829-860.
Hibbard, C. W., and D. W. Taylor

1960. Two late Pleistocene faunas from south-
western Kansas. Contrib. Mus. Paleontol.
Univ. Michigan, 16:1-223.
Hildebrand, M.

1952a. An analysis of body proportions in the

Canidae. Amer. Jour. Anat., 90:217-256.
1952b. The integument in Candidae. Jour. Mamm..

33:419-428.
1954. Comparative morphology of the body skele-
ton in Recent Canidae. Univ. California
Publ. Zool., 55:399-470.
Hilton, H.

1977. More answers to Maine coyote questions.
Maine Fish and Wildl., 19(l):2-4.

HlBSCHFELD, S. E.

1968. Vertebrate fauna of Nichol's Hammock, a
natural trap. Quart. Jour. Florida Acad.
Sci., 31:177-189.

HOFFMEISTER, D. F., AND W. W. GOODPASTER

1954. The mammals of the Huachuca Mountains,
southeastern Arizona. Illinois Biol. Monogr.,
Univ. Illinois Press, no. 24, v+152 pp.
Hoffmeister, D. F., and C. O. Mohr

1957. Fieldbook of Illinois mammals. Illinois Nat.
Hist. Sun-., xi+233 pp.

HOLLIMAN, D. C.

1963. The mammals of Alabama/ Unpublished
Ph.D. dissertation, Univ. Alabama, 504 pp.
Hood, C. H., and O. Hawksley

1975. A Pleistocene fauna from Zoo Cave, Taney
County, Missouri. Missouri Speleol., 15:
1-42.
Hopkins, M. L„ R. Bonnichsen, and D. Fortsch

1969. The stratigraphic position and fauna] asso-
ciates of Bison (Gigantobison) latifrons in
southeastern Idaho, a progress report. Teb-
iwa, 12:1-8.

Howard, W. E.

1949. A means to distinguish skulls of coyotes
and domestic dogs. Jour. Mamm., 30:169-
171.
Howell, A. H.

1921. A biological survey of Alabama. N. Amer.
Fauna, no. 45, 88 pp.

Iljin, N. A.

1941. Wolf-dog genetics. Jour. Genetics, 42:359-
414.
Imaizumi, Y.

1970a. Systematic status of the extinct Japanese
wolf, Canis hodophilax. Jour. Mamm. Soc.
Japan, 5:27-32.
1970b. Systematic status of the extinct Japanese
wolf, Canis hodophilax, 2. Similarity rela-
tionship of hodophilax among the species of
the genus Canis. Jour. Mamm. Soc. Japan,
5:62-66.
International Commission on Zoological No-
menclature

1957. Opinion 447. Opinions and Declarations,
15 (part 12):211-224.

Jackson, H. H. T.

1922. A coyote in Maryland. Jour. Mamm., 3:
186-187.

1949. Two new coyotes from the United States.
Proc. Biol. Soc. Washington, 62:31-32.

1951. Classification of the races of the coyote.
Part II in Young, S. P., and H. H. T. Jack-
son, The Clever Coyote, Wildl. Mgmt. Inst.,
Washington, D.C., pp. 227-441.

1961. Mammals of Wisconsin. Univ. Wisconsin
Press, Madison, xiv+504 pp.
Jakway, D. E.

1958. Pleistocene Lagomorpha and Rodentia from
the San Josecito Cave, Neuvo Leon, Mexico.
Trans. Kansas Acad. Sci., 61:313-327.

Jenkins, W.

1933. Wild life of Mississippi. Natchez, 155 pp.
Johnson, N. M., N. D. Opdyke, and E. H. Lindsay
1975. Magnetic polarity stratigraphy of Pliocene-
Pleistocene terrestrial deposits and verte-
brate faunas, San Pedro Valley, Arizona.
Bull. Geol. Soc. Amer., 86:5-12.
Johnston, C. S.

1938. Preliminary report on the vertebrate type
locality of Cita Canyon, and the description
of an ancestral coyote. Amer. Jour. Sci.,
ser. 5, 35:383-390.
Johnston, C. S., and D. E. Savage

1955. A survey of late Cenozoic vertebrate faunas
of the Panhandle of Texas. Part I. Intro-
duction, description of localities, preliminary
faunal lists. Univ. California Publ. Geol.
Sci., 31:27-50.
Jolicoeur, P.

1959. Multivariate geographical variation in the
wolf Canis lupus L. Evolution, 13:283-299.

Jorgensen, S. E., C. E. Faulkner, and L. D. Mech

( eds. )

1970. Proceedings of a symposium on wolf man-
agement in selected areas of North Amer-
ica. U. S. Bur. Sport Fisheries and Wildl.,
Twin Cities, Minnesota, iii+50 pp.

Keener, J. M.

1970. History of the wolf in Wisconsin. In Jor-
gensen, Faulkner, and Mech (1970), pp.
4-5.

128

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

Kellogg, R.

1915. The mammals of Kansas with notes on their
distribution, habits, life histories and eco-
nomic importance. Unpublished M.A. the-
sis, Univ. Kansas, 310 pp.

1937. Annotated list of West Virginia mammals.
Proc. U. S. Natl. Mus., 84:443-479.

1939. Annotated list of Tennessee mammals. Proc.
U. S. Natl. Mus., 86:245-303.
Kennelly, J. J., and J. D. Roberts

1969. Fertility of coyote-dog hybrids. Jour.
Mamm., 50:830-831.

Khan, E.

1970. Biostratigraphy and paleontology of a San-
gamon deposit at Fort Qu'Appelle, Sas-
katchewan. Natl. Mus. Canada Publ. Pa-
leontol., no. 5, viii+82 pp.

Klein, J.

1971. The ferungulates of the Inglis IA local
fauna, early Pleistocene of Florida. Un-
published M.S. thesis, Univ. Florida, 115 p.

Kolenosky, G. B.

1971. Hybridization between wolf and coyote.
Jour. Mamm., 52:446-449.

Kolenosky, G. B., and R. O. Standfteld

1975. Morphological and ecological variation
among gray wolves (Canis lupus) of On-
tario, Canada. In Fox (1975), pp. 62-72.
Kraglievich, J. L.

1952. Un canido del Eocuartario de Mar del
Plata y sus relaciones con otras formas Bra-
silenas y Norteamericanas. Rev. Mus. Mar
del Plata, 1:53-70.
Kraglievich, L.

1928. Contribucion al conocimiento de los grandes
canidos extinguidos de Sud America. An.
Soc. Cien. Argentina, 106:332-342.
Krefting, L. W.

1969. The rise and fall of the coyote on Isle Roy-
ale. Naturalist, 20(4) : 24-31.

KURTEN, B.

1963. Notes on some Pleistocene mammal migra-
tions from the Palaearctic to the Nearctic.
Eiszeitalter u. Gegenwart, 14:96-103.

1967. Prariewolf und sabelzahntiger aus dem
Pleistozan des Valsequillo, Mexiko. Quar-
tar, 18:173-178.

1968. Pleistocene mammals of Europe. Aldine,
Chicago, viii-(-317 pp.

1974. A history of coyote-like dogs (Canidae,
Mammalia). Acta Zool. Fennica, no. 140,
38 pp.

Kurten, B., and E. Anderson

1972. The sediments and fauna of Jaguar Cave.
II — The fauna. Tebiwa, 15:21-45.

Langdon, F. W.

1881. The mammalia of the vicinity of Cincinnati
— a list of species with notes. Jour. Cincin-
nati Soc. Nat. Hist., 1:297-313.

Lancguth, A.

1975. Ecology and evolution in the South Amer-
ican canids. In Fox (1975), pp. 192-206.

Lantz, D. E.

1905. A list of Kansas mammals. Trans. Kansas
Acad. Sci., 19:171-178.
Lawrence, B.

1966. Early domestic dogs. Zeit. f. Saugetier-
kunde, 32:44-59.

1968. Antiquity of large dogs in North America.
Tebiwa, 11:43-49.

Lawrence, B., and W. H. Bossert

1967. Multiple character analysis of Canis lupus,
latrans, and familiaris, with a discussion of
the relationships of Canis niger. Amer.
Zool., 7:223-232.

1969. The cranial evidence for hybridization in
New England Canis. Breviora, no. 330,
13 pp.

1975. Relationships of North American Canis
shown by a multiple character analysis of
selected populations. In Fox (1975), pp.
73-86.
Leidy, J.

1854. Note on 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.

1856. Description of some remains of extinct
mammalia. Jour. Acad. Nat. Sci., Phila-
delphia, ser. 2, 3:166-171.

1858. Notice of remains of extinct vertebrata,
from the Valley of the Niobrara River.
Proc. Acad. Nat. Sci., Philadelphia, 1858,
pp. 20-29.

1869. The extinct mammalian fauna of Dakota
and Nebraska. Jour. Acad. Nat. Sci., Phila-
delphia, ser. 2, 7:1-472.

1873. Contributions to the extinct vertebrate
fauna of the western territories. Rept. U. S.
Geol. Sun., 358 pp.

1889. Notice and description of fossils in caves
and crevices of the limestone rocks of
Pennsylvania. Ann. Rept. Geol. Surv. Penn-
sylvania, 1887, pp. 1-20.
Leopold, AS., and E. R. Hall

1945. Some mammals of Ozark County, Missouri.
Jour. Mamm., 26:142-145.
LlNHART, S. B., and F. F. Knowlton

1967. Determining age of coyotes by tooth ce-
mentum layers. Jour. Wildl. Mgmt, 31:
362-365.

LlNZEY, D. W.

1971. Animal harvested in south Alabama prob-
ably coyote-red wolf hybrid. Alabama
Conserv., 41(6):6-7.

LlNZEY, D. W., AND A. V. LlNZEY

1968. Mammals of the Great Smoky Mountains
National Park. Jour. Elisha Mitchell Sci.
Soc, 84:384-414.

Loomis, F. B., and D. B. Young

1912. On the shell heaps of Maine. Amer. Jour.
Sci., ser. 4, 34:17-42.
Lowery, G. H., Jr.

1943. Check- list of the mammals of Louisiana

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

129

and adjacent waters. Occas. Papers Louisi-
ana State Univ. Mus. Zool., 13:213-257.
Lundelius, E. L., Jr.

1960. Mylohyus nasutus long-nosed peccary of
the Texas Pleistocene. Bull. Texas Mem.
Mus., no. 1, 40 pp.

1962. Late Pleistocene vertebrate fauna from San
Patricio County, Texas. Geol. Soc. Amer.
Spec. Paper, no. 68, p. 222.

1967. Late-Pleistocene and Holocene faunal his-
tory of central Texas. In Martin and Wright
(1967), pp. 287-319.

1972. Fossil vertebrates from the late Pleistocene
Ingleside fauna, San Patricio County, Texas.
Univ. Texas, Bur. Econ. Geol. Rept. Invest.,
no. 77, 74 pp.
Lyon, M. W., Jr.

1936. Mammals of Indiana. Amer. Midi. Nat.,
17:1-384.
Macpherson, A. H.

1965. The origin of diversity in mammals of the
Canadian Arctic tundra. Syst. Zool., 14:
153-173.

Maine Department of Inland Fisheries and
Wildlife

1976. Coyote. Maine Fish and Wildl., 18(4):
S-19.
Maldonado-Koerdell, M.

1955. Sobre un craneo de Aenocyon dims (Leidy)
del Pleistoceno superior de Tequixquiac,
Mexico. An. Inst. Nac. Antro. Hist. Mexico,
7:51-58.
Manning, T. H., and A. H. Macpherson

1958. The mammals of Banks Island. Arctic Inst.
N. Amer. Tech. Paper, no. 2, 74 pp.
Manville, R. H., and W. C. Sturtevant

1966. Early specimens of the eastern wolf, Canis
lupus lycaon. Chesapeake Sci., 7:218-219.

Marcus, L. F.

1960. A census of the abundant large Pleistocene
mammals from Rancho La Brea. Los An-
geles Co. Mus. Contrib. Sci., no. 38, 11 pp.

Martin, L. D.

1972. The microtine rodents of the Mullen assem-
blage from the Pleistocene of north central
Nebraska. Bull. Univ. Nebraska State Mus.,
9:173-182.

Martin, P. S., and H. E. Wright (eds.)

1967. Pleistocene extinctions. Yale Univ. Press,
New Haven, x+453 pp.

Martin, R. A.

1974. Fossil mammals from the Coleman II A
fauna, Sumter County. In Webb (1974a),
pp. 35-99.
Martin, R. A., and S. D. Webb

1974. Late Pleistocene mammals from the Devil's
Den fauna, Levy County. In Webb
(1974a), pp. 114-145.
Marvinney, S.

1976. Will the predators return? Conservationist,
31(l):iv-v.

Matthew, W. D.

1902. List of the Pleistocene fauna from Hay
Springs, Nebraska. Bull. Amer. Mus. Nat.
Hist., 16:317-322.

1916. The grim wolf of the tar pits. Amer. Mus.
Jour., 16(l):45-47.

1918. Contributions to the Snake Creek fauna
with notes upon the Pleistocene of western
Nebraska American Museum expedition of
1916. Bull. Amer. Mus. Nat. Hist., 38:
183-229.

1930. The phylogeny of dogs. Jour. Mamm., 11:
117-138.
Mawby, J. E.

1967. Fossil vertebrates of the Tule Springs site,
Nevada. Nevada State Mus. Anthropol. Pa-
pers, 13:106-128.
Mayr, E.

1963. Animal species and evolution. Harvard
Univ. Press, Cambridge, xiv-f-797 pp.

McCarley, H.

1959. The mammals of eastern Texas. Texas Jour.

Sci., 11:385-426.
1962. The taxonomic status of wild Canis (Cani-
dae) in the south central United States.
Southwestern Nat., 7:227-235.
McDonald, H. C, and E. Anderson

1975. A late Pleistocene vertebrate fauna from
southeastern Idaho. Tebiwa, 18:19-37.
McGrew, P. O.

1944. An early Pleistocene (Blancan) fauna from
Nebraska. Field Mus. Nat. Hist. Geol. Ser.,
9:33-66.

McKnight, T.

1964. Feral livestock in Anglo-America. Univ.
California Publ. Geogr., 16:1-78.

Meade, G. E.

1945. The Blanco fauna. Univ. Texas Publ., 4401 :
509-556.

Mech, L. D.

1959. The coyote comes east. Frontiers, 23:117-
119, 126.

1961. Exit timber wolf, enter coyote. Animal
Kingdom, 64(3):89-92.

1966. The wolves of Isle Royale. Fauna Natl.
Parks, U.S., Fauna Ser., no. 7, xiv+210 pp.

1970. The wolf: the ecology and behavior of an
endangered species. Natural History Press,
Garden City, New York, xx+384 pp.

1977. A recovery plan for the eastern timber wolf.
Natl. Parks and Conserv. Mag., 51(1):
17-21.
Mech, L. D., and L. D. Frenzel, Jr.

1971. The possible occurrence of the Great Plains
wolf in northeastern Minnesota. In Mech,
L. D., and L. D. Frenzel, Jr. (eds.), Eco-
logical studies of the timber w-olf in north-
eastern Minnesota, N. Cent. Forest Exp.
Sta., St. Paul, pp. 60-62.

Mehl, M. G.

1962. Missouri's ice age mammals. Missouri Div.
Geol. and Water Res., Ed. Ser., no. 1, xi+
104 pp.

130

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

Mexcel, R. M.

1971. A study of dog-coyote hybrids and implica-
tions concerning hybridization in Canis.
Jour. Mamm., 52:316-336.
Merriam, C. H.

1897. Revision of the coyotes or prairie wolves,
with descriptions of new forms. Proc. Biol.
Soc. Washington, 11:19-33.
Merriam, J. C.

1903. The Pliocene and Quaternary Canidae of
the Great Valley of California. Univ. Cali-
fornia Publ. Bull. Dept. Geol., 3:277-290.

1906. Recent discoveries of Quaternary mammals
in southern California. Science, n.s., 24:
248-250.

1910. New Mammalia from Rancho La Brea.
Univ. California Publ. Bull. Dept. Geol.,
5:391-395.

1911. Tertiary mammal beds of Virgin Valley and
Thousand Creek in northwestern Nevada.
Univ. California Publ. Bull. Dept. Geol.,
6:199-304.

1912. The fauna of Rancho La Brea. Part II.
Canidae. Mem. Univ. California, 1:217-272.

1918. Note on the systematic position of the
wolves of the Canis dims group. Univ.
California Publ. Bull. Dept. Geol., 10:531-
533.
Merriam, J. C., and C. Stock

1921. Occurrence of Pleistocene vertebrates in an
asphalt deposit near McKittrick, California.
Science, n.s., 54:566-567.

Miller, A. M.

1922. Licks and caves of the lower Ohio Valley
as repositories of mammalian remains, in-
cluding those of man. Bull. Geol. Soc.
Amer., 33:156-159.

Miller, G. J.

1968. On the age distribution of Smilodon cali-
fornicus Bovard from Rancho La Brea. Los
Angeles Co. Mus. Contrib. Sci., no. 131,
17 pp.
Miller, G. S., Jr.

1899. Preliminary list of New York mammals.

Bull. New York State Mus., 6:271-390.
1912a. The names of two North American wolves.

Proc. Biol. Soc. Washington, 25:95.
1912b. The names of the large wolves of northern
and western North America. Smithsonian
Misc. Coll., 59(15): 1-5.
1912c. Catalogue of the mammals of western Eu-
rope (Europe exclusive of Russia). British
Mus., London, xv-|-1019 pp.
Miller, M. E., G. C. Christexsex, axd H. E. Evans
1964. Anatomy of the dog. W. B. Saunders Co.,
Philadelphia, xii-)-941 pp.
Miller, W. E.

1971. Pleistocene vertebrates of the Los Angeles
Basin and vicinity (exclusive of Rancho La
Brea). Bull. Los Angeles Co. Mus. Nat.
Hist., Sci.: no. 10, 124 pp.

1976. Late Pleistocene vertebrates of the Silver
Creek local fauna from north central Utah.
Great Basin Nat., 36:387-424.
Mivart, St. G.

1890. Monograph of the Canidae. London, 216

PP-
Mooser, O., and W. W. Dalquest

1975. Pleistocene mammals from Aguascalientes,
central Mexico. Jour. Mamm., 56:781-820.
Morris, R. F.

1948. The land mammals of New Brunswick.
Jour. Mamm., 29:165-176.
Morrison, J. D.

1970. The Eddy Bluff shelter of Beaver Reservoir
of northwest Arkansas. Proc. Arkansas
Acad. Sci., 24:85-91.
Morrison, R. B.

1964. Lake Lohontan: geology of southern Car-
son Desert, Nevada. U. S. Geol. Surv. Prof.
Paper, no. 401, v+156 pp.
Mumford, R. E.

1969. Distribution of the mammals of Indiana.
Indiana Acad. Sci. Monogr., no. 1, vii-f-
114 pp.

Murphy, J. L.

1968. The Hobson site: a Fort Ancient compo-
nent near Middleport, Meigs County, Ohio.
Kirtlandia, no. 4, 14 pp.
Negus, N. C.

1948. A coyote, Canis latrans. from Preble County,
Ohio. Jour. Mamm., 29:295.
Nesbitt, W. H.

1975. Ecology of a feral dog pack on a wildlife
refuge. In Fox (1975), pp. 391-396.
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. S. California Acad.
Sci., 46:26-34.
Nowak, R. M.

1967. The red wolf in Louisiana. Defenders of
Wildl. News, 42:60-70.

1970. Report on the red wolf. Defenders of
Wildl. News, 45:82-94.

1971. Louisiana protects wolf, cougar, and all
birds of prey. Defenders of Wildl. News,
46:278.

1972. The mysterious wolf of the south. Nat.
Hist., 81(l):50-53, 74-77.

1973. North American Quaternary Canis. Un-
published Ph.D. dissertation, Univ. Kansas,
380 pp.

1974. Red wolf: our most endangered mammal.
Natl. Parks and Conserv. Mag., 48(8):9-12.

Olsen, S. J., and J. W. Olsen

1977. The Chinese wolf, ancestor of New World
dogs. Science, 197:533-535.
Olsox, E. C.

1940. A late Pleistocene fauna from Herculaneum,
Missouri. Jour. Geol., 48:32-57.
Osgood, W. H.

1934. The genera and subgenera of South Amer-
ican canids. Jour. Mamm., 15:45-50.

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

131

Ozoca, J. J., AND E. M. Harger

1966. Occurrence of albino and melanistic coyotes
in Michigan. Jour. Mamm., 47:339-340.
Packard, A. S.

1885. Origin of the American varieties of the dog.
Zoologist, ser. 3, 9:367-372.
Packard, E. L.

1950. A large wolf from the Pleistocene of Wil-
lamette Valley, Oregon. Geol. News Letter,
Geol. Soc. Oregon Country, 16:89-90.
Paradiso, J. L.

1965. Recent records of red wolves from the Gulf
Coast of Texas. Southwestern Nat., 10:
318-319.

1966. Recent records of coyotes, Canis latrans,
from the southeastern United States. South-
western Nat., 11:500-501.

1968. Canids recently collected in east Texas,
with comments on the taxonomy of the red
wolf. Amer. Midi. Nat., 80:529-534.

1969. Mammals of Maryland. N. Amer. Fauna,
no. 66, iv+193 pp.

Paradiso, J. L., and R. M. Nowak

1972a. A report on the taxonomic status and dis-
tribution of the red wolf. U. S. Bur. Sport
Fisheries and Wildl. Spec. Sci. Rept. —
Wildl., no. 145, ii+36 pp.

1972b. Canis rufus. Mammalian Species, Amer.
Soc. Mamm., no. 22, 4 pp.

1973. New data on the red wolf in Alabama.
Jour. Mamm., 54:506-509.
Paradiso, J. L., and D. Schierbaum

1969. Recent wolf record from New York. Jour.
Mamm., 50:384-385.
Parmalee, P. W.

1957. Vertebrate remains from the Cahokia site,
Illinois. Trans. Illinois State Acad. Sci., 50:
235-242.

1959a. Use of mammalian skulls and mandibles by
prehistoric Indians of Illinois. Trans. Illi-
nois State Acad. Sci., 52:85-95.

1959b. Animal remains from the Raddatz rock-
shelter, Sk 5, Wisconsin. Wisconsin Ar-
cheol., 40:83-90.

1959c. Animal remains from the Banks site, Crit-
tenden County, Arkansas. Tennessee Ar-
chaeol. Soc. Misc. Paper, no. 5, 8 pp.

1962a. Additional fauna] records from the Kings-
ton Lake site, Illinois. Trans. Illinois State
Acad. Sci., 55:6-12.

1962b. The faunal complex of the Fisher site, Illi-
nois. Amer. Midi. Nat., 68:399-408.

1963. Vertebrate remains from the Bell site, Win-
nebago County, Wisconsin. Wisconsin Ar-
cheol., 44:58-69.

1965. The food economy of Archaic and Wood-
land peoples at the Tick Creek Cave site,
Missouri. Missouri Archaeol., 27:1-34.

1972. Vertebrate remains from the Fifield site,
Porter County, Indiana. Indiana Hist. Soc.
Prehist. Res. Ser., 5:202-205.
Parmalee, P. W., and R. D. Oesch

1972. Pleistocene and Recent faunas from the

Brynjulfson Caves, Missouri. Illinois State
Mus. Rept. Invest., no. 25, vii-(-52 pp.
Parmalee, P. W., and O. C. Shane, III

1970. The Blain site vertebrate fauna. In Prufer,
O. H., and O. C. Shane, III, Blain Village
and the Fort Ancient tradition in Ohio,
Kent State Univ. Press, pp. 185-206.
Parmalee, P. W., and D. Stephens

1972. A wolf mask and other carnivore skull arti-
facts from the Palestine site, Illinois. Penn-
sylvania Archaeol., 42:71-74.
Patterson, B.

1932. Upper molars of Canis annbrusteri Gidley
from Cumberland Cave, Maryland. Amer.
Jour. Sci., ser. 5, 23:334-336.
Patton, T. H.

1963. Fossil vertebrates from Miller's Cave, Llano
County, Texas. Bull. Texas Mem. Mus.,
7:1-41.
Paul, J. R.

1970. Coyotes and kin. Explorer, 12(l):23-25.
Peterson, O. A.

1926. The fossils of the Frankstown Cave, Blair
County, Pennsylvania. Ann. Carnegie Mus.,
16:249-315.
Peterson, R. L.

1946. Recent and Pleistocene mammalian fauna
of Brazos County, Texas. Jour. Mamm., 27:
162-169.
1957. Changes in the mammalian fauna of On-
tario. In Urquhart, F. A. (ed.), Changes
in the fauna of Ontario, Contrib. Royal
Ontario Mus., pp. 43-58.

1966. The mammals of eastern Canada. Oxford
Univ. Press, Toronto, xxxii-|-465 pp.

Pewe, T. L., and D. M. Hopkins

1967. Mammal remains of pre- Wisconsin age in
Alaska. In Hopkins, D. M. (ed.), The
Bering Land Bridge, Stanford Univ. Press,
pp. 266-270.

PlMLOTT, D. H., AND P. W. JOSLIN

1968. The status and distribution of the red wolf.
Trans. N. Amer. Wildl. and Nat. Res. Conf.,
33:373-389.

Pocock, R. I.

1935. The races of Canis lupus. Proc. Zool. Soc.
London, 1935, pt. 3, pp. 617-686.
Princle, L. P.

1960. Notes on coyotes in southern New England.

Jour. Mamm., 41:278.
1963. Covotes in New England. Massachusetts
Audubon, 48(2) : 52-54.

QUACKENBUSH, L. S.

1909. Notes on Alaskan mammoth expedition of

1907 and 1908. Bull. Amer. Mus. Nat.

Hist., 26:87-130.
Quinn, J. H.

1972. Extinct mammals in Arkansas and related

C14 dates circa 3000 years ago. Int. Geol.

Congr., 24(sect. 12):89-96.
Rand, A. L.

1945. Mammals of the Ottawa district. Canadian

Field-Nat., 59:111-132.

132

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

Rao, C. R.

1952. Advanced statistical methods in biometric
research. Wiley, New York, 390 pp.
Ray, C. E.

1958. Additions to the Pliestocene mammalian
fauna from Melbourne, Florida. Bull. Mus.
Comp. Zool., 119:421-451.
1967. Pleistocene mammals from Ladds, Bartow
County, Georgia. Georgia Acad. Sci. Bull.,
25:120-150.
Redington, P. G.

1931. Report of the chief of the Bureau of Bio-
logical Survey. Washington, D.C.
Reed, C. A.

1961. Osteological evidences for prehistoric do-
mestication in southwestern Asia. Sonder-
druck Zeit. Tierzuchtung und Zuchtungs-
biologie, 76:31-38.
Reynolds, H. S.

1909. A monograph of the British Pleistocene
mammalia, vol. 2, part 3, pp. 1-28, The
Canidae. London Palaeontographical Soc.
Rhoads, S. N.

1903. The mammals of Pennsylvania and New
Jersey. Philadelphia, 266 pp.
Richardson, J.

1829. Fauna Boreali-Americana. John Murray,
London, xvi-f-300 pp.
Richens, V. B., and R. D. Hugie

1974. Distribution, taxonomic status, and charac-
teristics of coyotes in Maine. Jour. Wildl.
Mgmt., 38:447-454.
Richmond, N. D., and H. R. Rosland

1949. Mammal survey of northwestern Pennsyl-
vania. Pennsylvania Game Comm., 87 pp.

RlNKER, G. C, AND C. W. HlBHARD

1952. A new beaver and associated vertebrates,
from the Pleistocene of Oklahoma. Jour.
Mamm., 33:98-101.
Ritchie, W. A.

1969. The archaeology of New York State. Nat-
ural History Press, Garden City, New York,
xxxiv-|-357 pp.

ROEMER, F.

1849. Texas. 1935 reprint by Standard Printing

Co., San Antonio, xii-f-301 pp.
Roth, E. L.

1972. Late Pleistocene mammals from Klein Cave,

Kerr County, Texas. Texas Jour. Sci., 24:

75-84.
Russell, D. N., and J. H. Shaw

1971a. Notes on the red wolf (Canis rufus) in the

coastal marshes and prairies of eastern

Texas. Texas Parks and Wildl. Dept., 5 pp.
1971b. Distribution and relative density of the red

wolf in Texas. Texas Parks and Wildl.

Dept., 11 pp.
1972. The red wolf — situation critical. Texas

Parks and Wildl., 30(3): 12-15.
Russell, R. J.

1960. Pleistocene pocket gophers from San Jose-

cito Cave, Nuevo Leon, Mexico. Univ.

Kansas Publ. Mus. Nat. Hist., 9:539-548.

St. Amant, L. S.

1959. Louisiana wildlife inventory and manage-
ment plan. Louisiana Wild Life and Fish-
eries Comm., xx+329 pp.
Sampson, F. W.

1961. Missouri's vanishing wolves. Missouri Con-
serve, 22(6) :5-7.
Saunders, J. J.

1977. Late Pleistocene vertebrates of the western
Ozark Highland, Missouri. Illinois State
Mus. Rept. Invest., no. 33, x4-118 pp.
Savage, D. E.

1951. Late Cenozoic vertebrates of the San Fran-
cisco Bay region. Univ. California Publ.
Bull. Dept. Geol. Sci., 28:215-314.

SCHORGER, A. W.

1942. Extinct and endangered mammals and birds
of the upper Great Lakes region. Trans.
Wisconsin Acad. Sci., Arts and Letters, 34:
23-44.
Schultz, C. B.

1934. The Pleistocene mammals of Nebraska.
Bull. Nebraska State Mus., 1:357-393.

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.

Schultz, C. B., and L. D. Martin

1970. Quaternary mammalian sequence in the
Great Plains. In Dort, W., Jr., and J. K.
Jones, Jr., Pleistocene and Recent environ-
ments of the central Great Plains, Univ.
Kansas Dept. Geol. Spec. Publ., no. 3, pp.
341-353.
Schultz, C. B., L. D. Martin, and L. G. Tanner

1970. Mammalian distribution in the Great Plains
and adjacent areas from 14,000 to 9,000
years ago. Abstr., Amer. Quat. Assoc. Mtg.,
1:119-120.
Schultz, C. B., and T. M. Stout

1945. Pleistocene loess deposits of Nebraska.
Amer. Jour. Sci., 243:231-244.

1948. Pleistocene mammals and terraces in the
Great Plains. Bull. Gaol. Soc. Amer., 59:
533-588.
Schultz, C. B., and L. G. Tanner

1957. Medial Pleistocene fossil vertebrate locali-
ties in Nebraska. Bull. Univ. Nebraska
State Mus., 4:59-81.
Schultz, G. E.

1969. Geology and paleontology of a late Pleisto-
cene basin in southwest Kansas. Geol. Soc.
Amer. Spec. Paper, no. 105, viii-|-85 pp.
Schultz, J. R.

1938a. Early Pleistocene mammal fauna from the
vicinity of Grand View, Ada and Owyhee
counties, Idaho. Proc. Geol. Soc. Amer.,
1937, p. 297.

1938b. A late Quaternary mammal fauna from the
tar seeps of McKittrick, California. Carne-
gie Inst. Washington Publ., 487:111-215.
Schultz, V.

1955. Status of the coyote and related forms in

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

133

Tennessee. Jour. Tennessee Acad. Sci., 30:
44-46.
Schultz, V., et al.

1954. Statewide wildlife survey of Tennessee.
Tennessee Game and Fish Comm., Nash-
ville, 506 pp.
Scott, J. P.

1968. Evolution and domestication of the dog.
Evol. Biol., 2:243-275.
Scott, J. P., and J. L. Fuller

1965. Genetics and the social behavior of the dog.
Univ. Chicago Press, xviii+468 pp.
Scott, W. B.

1937. A history of land mammals in the Western
Hemisphere. Amer. Phil. Soc, New York,
xiv+786 pp.
Seal, H.

1964. Multivariate statistical analysis for biolo-
gists. Wiley, New York, 207 pp
Sealander, J. A., Jr.

1956. A provisional check-list and key to the
mammals of Arkansas (with annotations).
Amer. Midi. Nat., 56:257-296.
Sellards, E. H.

1916. Human remains and associated fossils from
the Pleistocene of Florida. Ann. Bept. Flor-
ida Geol. Surv., 8:123-160.
1940. Pleistocene artifacts and associated fossils
from Bee County, Texas. Bull. Geol. Soc.
Amer., 51:1627-1658.
Semken, H. A.

1961. Fossil vertebrates from Longhorn Cavern,
Burnet County, Texas. Texas Jour. Sci.,
13:290-310.
Serres, M.

1835. On the distinctive characters of the dog,
the wolf, and the fox, as supplied by the
skeleton. Edinburgh New Philos. Jour., 19:
244-253.
Seton, E. T.

1929. Lives of game animals. Doubleday, Doran
& Co., Garden City, New York, l(part 2):
339-640.
Severinghaus, C. W.

1974a. Notes on the history of wild canids in New
York. New York Fish and Game Jour., 21:
117-125.
1974b. The coyote moves east. Conservationist,
29(2):8, 36.
Shakespear, S.

1975. An osteometric description of the Pleisto-
cene dire wolf, Canis dims, from the Mari-
copa Brea of California. Proc. Utah Acad.
Sci. Arts Letters, 52:77-78.
Shaw, J. H.

1975. Ecology, behavior, and systematics of the
red wolf (Canis rufus). Unpublished Ph.D.
dissertation, Yale Univ., v+99-f-xi pp.
Shay, C. T.

1963. A preliminary report on the Itasca bison
site. Proc. Minnesota Acad. Sci., 31:24-27.
Sherman, H. B.

1937. List of the Becent wild land mammals of

Florida. Proc. Florida Acad. Sci., 1:102-
128.
Shiras, G., Ill

1921. The wildlife of Lake Superior, past and
present. Natl. Geogr., 40:113-204.
Shoemaker, H. W.

1917. Extinct Pennsylvania animals. Altoona
Tribune, Altoona, Pennsylvania, 201 pp.
Shotwell, J. A.

1956. Hemphillian mammalian assemblage from
northeastern Oregon. Bull. Geol. Soc.
Amer., 67:717-738.

1970. Pliocene mammals of southeast Oregon and
adjacent Idaho. Bull. Mus. Nat. Hist. Univ.
Oregon, 17:1-103.
Silver, H.

1957. A history of New Hampshire game and fur-
bearers. New Hampshire Fish & Game
Dept. Surv. Bept., no. 6, xiv+466 pp.

Silver, H., and W. T. Silver

1969. Growth and behavior of the coyote-like
canid of northern New England with ob-
servations on canid hybrids. Wildl. Mon-
ogr., no. 17, 41 pp.
Simpson, G. G.

1928. Pleistocene mammals from a cave in Citrus
County, Florida. Amer. Mus. Novit., no.
328, 16 pp.

1929a. Pleistocene mammalian fauna of the Semi-
nole Field, Pinellas County, Florida. Bull.
Amer. Mus. Nat. Hist., 56:561-599.

1929b. The extinct land mammals of Florida. Ann.
Bept. Florida Geol. Surv., 20:229-279.

1941. Large Pleistocene felines of North America.
Amer. Mus. Novit., no. 1136, 27 pp.

194.5. The principles of classification and a classi-
fication of mammals. Bull. Amer. Mus. Nat.
Hist., 85:v-xvi+l-350.
1949. A fossil deposit in a cave in St. Louis.
Amer. Mus. Novit., no. 1408, 46 pp.
Sinclair, W. J.

1904. The exploration of the Potter Creek Cave.
Univ. California Publ. Amer. Archaeol. and
Ethnol., 2:1-27.
Skaccs, O.

1946. A study of the dog skeletons from Indian
Knoll with special reference to the coyote
as a progenitor. In Webb, W. S., Indian
Knoll, Univ. Kentucky Bept. Anthropol.
and Archaeol., 4:341-355.
Skinner, M. F.

1942. The fauna of Papago Springs Cave, Arizona.
Bull. Amer. Mus. Nat. Hist., 80:143-220.

Skinner, M. F., and C. W. Hibbard

1972. Early Pleistocene preglacial and glacial
rocks and faunas of north-central Nebraska.
Bull. Amer. Mus. Nat. Hist., 148:1-148.
Slaughter, B. H.

1961. A new coyote in the late Pleistocene of

Texas. Jour. Mamm., 42:503-509.
1966a. The Moore pit local fauna; Pleistocene of
Texas. Jour. Paleontol., 40:78-91.

134

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

1966b. Plaiyppnus compressus and associated fauna
from the Laubach Cave of Texas. Amer.
Midi. Nat., 75:475-494.
Slauchter, 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. Univ. Texas, Bur. Econ. Geol. Rept.
Invest., no. 48, viii+75 pp.

Slaughter, B. H., and B. R. Hoover

1963. Sulphur River formation and the Pleisto-
cene mammals of the Ben Franklin local
fauna. Jour. Grad. Res. Center, S. Meth.
Univ., 31:132-148.

Slauchter, B. H., and R. Ritchie

1963. Pleistocene mammals of the Clear Creek
local fauna, Denton County, Texas. Jour.
Grad. Res. Center, S. Meth. Univ., 31:117-
131.
Smits, L.

1963. King of the wild. Michigan Conserv., 32
(l):45-50.
Snyder, L. L.

1938. A faunal investigation of western Rainy
River District, Ontario. Trans. Royal Ca-
nadian Inst., Toronto, 22:157-180.
Standfield, R.

1970. Some considerations on the taxonomy of
wolves in Ontario. In Jorgensen, Faulkner,
and Mech ( 1970), pp. 32-38.
Starrett, A.

1956. Pleistocene mammals of the Berends fauna
of Oklahoma. Jour. Paleontol., 30:1187-
1192.
Stealer, A. M.

1944. The status of the wolf in Michigan. Jour.
Mamm., 25:37-43.
Sternberg, C. H.

1928. Extinct animals of California. Sci. Amer.,
139:225-227.

Stock, C.

1918. The Pleistocene fauna of Hawver Cave.
Univ. California Publ. Bull. Dept. Geol.,
10:461-515.

1929. A census of the Pleistocene mammals of
Rancho La Brea, based on the collections
of the Los Angeles Museum. Jour. Mamm.,
10:281-289.

1938. A coyote-like wolf jaw from the Rancho La

Brea Pleistocene. Bull. S. California Acad.

Sci., 37:49-51.
1956. Rancho La Brea, a record of Pleistocene

life in California. Los Angeles Co. Mus.

Nat. Hist., Sci. Ser., no. 20, 83 pp.
Stock, C, and J. F. Lance

1948. The relative length of limb elements in

Canis dims. Bull. S. California Acad. Sci.,

47:79-84.
Stock, C, J. F. Lance, and J. O. Nigra

1946. A newly mounted skeleton of the extinct

dire wolf from the Pleistocene of Rancho

La Brea. Bull. S. California Acad. Sci., 45:

108-110.

Stovall, J. W., and W. N. McAnulty

1950. The vertebrate fauna and geologic age of
Trinity River terraces in Henderson Countv,
Texas. Amer. Midi. Nat., 44:211-250.
Strecker, J. K.

1926. A check-list of the mammals of Texas. Bay-
lor Univ. Bull., 29(3):l-48.

SURBER, T.

1932. The mammals of Minnesota. Minnesota
Game and Fish Dept., St. Paul, 84 pp.

Taylor, R. W., C. I. Counts, III, and S. Mills
1976. Occurrence and distibution of the coyote,
Canis latrans. Say, in West Virginia. Proc.
West Virginia Acad. Sci., 48:3-4.
Teer, C. L.

1975. The eastern coyote. Maine Fish and Wildl.,
17(4):9-12.
Thentus, E.

1970. Einige jungpleistozane saugetiere (Plattj-
gonus, Arctodus und Canis dints) aus dem
Valsequillo, Mexico. Quartar, 21:57-66.
Trouessart, E. L.

1911. Le loup de Flnde (Canis pallipes Sykes),
souche ancestrale du chien domestique.
Compt. Rend. Acad. Sci. Paris, 152:909-
913.
Troxell, E. L.

1915. The vertebrate fossils of Rock Creek, Texas
Amer. Jour. Sci., ser. 4, 39:613-638.
Ulmer, F. A.

1949. Recent records of coyotes in Pennsylvania
and New Jersey. Jour. Mamm., 30:435-436.
Van Valen, L.

1964. Nature of the supernumary molars of Oto-
cyon. Jour. Mamm., 45:284-286.
Vanderhoof, V. L.

1933. Additions to the fauna of the Tehama upper
Pliocene of northern California. Amer. Jour.
Sci., ser. 5, 25:382-384.

1937. Critical observations on the Canidae in
Cope's original collection from the Blanco
of Texas. Proc. Geol. Soc. Amer., 1936.
p. 389.
Vanderhoof, V. L., and J. T. Gregory

1940. A review of the genus Aelurodon. Univ.
California Publ. Bull. Dept. Geol., 25:143-
164.
Warfel, H. E.

1937. A coyote in Hampshire County, Massachu-
setts. Jour. Mamm., 18:241.
Webb, S. D. (ed.)

1974a. Pleistocene mammals of Florida. University
Presses of Florida, Gainesville, x-f-270 pp.
Webb, S. D.

1974b. Chronology of Florida Pleistocene mam-
mals. In Webb (1974a), pp. 5-31.
Webb, W. S., and R. S. Baby

1957. The Adena People. No. 2. Ohio Hist. Soc,
Ohio State Univ. Press, xi-(-123 pp.
Webster, D.

1964. Cherokee Cave bone deposit. Missouri
Speleol., 16:79-86.

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

135

Weigel, R. D.

1962. Fossil vertebrates of Vero, Florida. Florida
Geol. Surv. Spec. Publ., no. 10, vii-f 59 pp.
Weise, T. F., W. L. Robinson, R. A. Hook, and
L. D. Mech

1975. An experimental translocation of the eastern
timber wolf. Audubon Conserv. Rept., no.
5, 28 pp.

Wendobf, F., A. D. Krieger, and C. C. Albbitton
1955. The Midland discovery. Univ. Texas Press,
viii+139 pp.
Wetzel, R. M., and L. R. Penner

1962. Coydog in Connecticut. Jour. Mamm., 43:
109-110.
Whit acre, D.

1948. The mysterious coyote pack of Ohio. Ohio
Conserv. Bull., 12(3):29.
Whitney, J. D.

1879. The animal remains, not human, of the
auriferous gravel series. Mem. Mus. Comp.
Zool., 6:239-258.
Williams, C. T.

1962. Classification of the Borophaginae (Cani-
dae). Unpublished M.A. thesis, Univ. Kan-
sas, 103 pp.

WlLLISTON, S. W.

1898. The Pleistocene of Kansas. Trans. Kansas
Acad. Sci., 15:90-94.
Wilson, J.

1976. Bobcats, bears and coyotes .... are they
here? Kentucky Happy Hunting Ground,
32(6) :6-8.

Wilson, B. L.

1967. The Pleistocene vertebrates of Michigan.
Papers Michigan Acad. Sci., Arts, and Let-
ters, 52:197-234.
Wilson, R. W.

1933. Pleistocene mammalian fauna from the Car-
penteria asphalt. Carnegie Inst. Washing-
ton Publ., 440:59-76.
Wilson, W. C.

1967. Food habits of the coyote, Canis latrans, in
Louisiana. Unpublished M.S. thesis, Louisi-
ana State Univ., ix+49 pp.

Winc, E. S.

1963. Vertebrates from the Jungerman and Good-
man sites near the east coast of Florida.
Contrib. Florida State Mus., Soc. Sci., 10:
51-60.

Wolfe, J. L.

1972. Wolves in Mississippi? Mississippi Game
and Fish, 35(2): 10-11.

Wolfe, M. L., and D. L. Allen

1973. Continued studies of the status, socializa-
tion, and relationships of Isle Royale wolves,
1967 to 1970. Jour. Mamm., 54:611-635.

Wolfram, G.

1964. Coyotes: the silent invaders. Canadian
Audubon, 26:112-115.

Wood, N. A.

1922. The mammals of Washtenaw County, Mich-
igan. Occas. Papers Mus. Zool. Univ. Mich-
igan, no. 123, 23 pp.
Wood, N. A., and L. R. Dice

1924. Records of the distribution of Michigan
mammals. Papers Michigan Acad. Sci.,
Arts, and Letters, 3:425-469.

WOODHOUSE, S. W.

1851. The North American jackal — Canis frustror.
Proc. Acad. Nat. Sci., Philadelphia, ser. 1,

5:147-148.

WURSTER, D. H., AND K. BeNTRSCHKE

1968. Comparative cytogenetic studies in the or-
der Carnivora. Chromosoma, 24:336-382.
Young, S. P.

1944. History, life habits, economic status, and
control. Part I in Young, S. P., and E. A.
Goldman, The wolves of North America,
Amer. Wildl. Inst., Washington, D.C., pp.
1-385.

1946. The wolf in North American history. Cax-
ton Printers, Ltd., Caldwell, Ohio, 149 pp.

1951. History, life habits, economic status, and
control. Part I in Young, S. P., and H. H. T.
Jackson, The clever coyote, Wildl. Mgmt.
Inst., Washington, D.C., pp. 1-226.

Addendum

Subsequent to preparation of the galley
proof of this paper, the following information
came to my attention.

Canis Jepophagus. — Bjork (1974) assigned
a newly discovered specimen from the Wen-
dell Fox pasture locality of the Rexroad fauna,
Meade County, Kansas, to this species. Dal-
quest (1978) listed "Canis cf. lepophagus"
from the Beck Ranch local fauna, Scurry
County, Texas.

Canis latrans. — Parmalee, Munson, and
Guilday (1978) reported specimens of "C.
latrans Say — Coyote?" from the Harrodsburg
Crevice, Monroe County, Indiana. Although
a radiocarbon analysis made on bones from
this site provided a date of 25,050 ±660 B.P.,
some of the fauna! components suggested
that a Sangamon age was more likely. Corner
(1977) reported this species from a Ranchola-
brean fauna, 4.5 mi. W McCook, Red Willow
County, Nebraska. Grayson (1977) reported
"Canis cf. latrans Coyote" from zones dated
6,500-9,500 B.P. in the Dirty Shame Rock-
shelter, Malheur County, Oregon. Martin,
Gilbert, and Adams ( 1977 ) listed this species
from the late Pleistocene Natural Trap Cave,
Big Horn County, Wyoming.

Canis rufus. — On 5 January 1978 the U.S.
Fish and Wildlife Service made a second re-
lease of a pair of wild-caught red wolves on
Bulls Island, Cape Romain National Wildlife
Refuge, South Carolina. The animals appar-
ently adapted well, and did not leave the
vicinity of Bulls Island and a small adjacent
island, until they were recaptured (un-
harmed ) on 19 October and 1 November 1978
to terminate the experiment. In November
1978, the breeding colony at Tacoma, Wash-

ington and a second facility at Winnie, Texas
contained a total of 31 wild-caught animals
thought to be red wolves, 5 surviving young
produced in the spring of 1977, and 15 young
produced in the spring of 1978.

Canis lupus. — Corner ( 1977 ) reported this
species from a Rancholabrean fauna, 4.5 mi.
W McCook, Red Willow County, Nebraska.
Martin, Gilbert, and Adams (1977) listed
"Canis sp. (wolf)" from the late Pleistocene
Natural Trap Cave, Big Horn County, Wyom-
ing.

Canis familiaris. — Arredondo and Varona
( 1974 ) described Cubacyon transversidens, a
new genus and species of canid from a Pleis-
tocene site in western Cuba. Based on the
published description and an examination of
specimens of domestic dogs, E. Raymond Hall
(Museum of Natural History, Univ. Kansas;
pers. comm.) considers Cubacyon transver-
sidens to be a synonym of Canis familiaris,
and I agree with this assessment. Beebe
(1978) reported a specimen of C. familiaris
from the Old Crow River Basin of the north-
ern Yukon, with a minimum age of 20,000
B.P., and observed: "The highly evolved
morphology of the specimen suggests a much
earlier time of domestication."

Canis dims. — Parmalee, Munson, and
Guilday ( 1978 ) reported specimens of "Canis
cf. dims" from the Harrodsburg Crevice,
Monroe County, Indiana (see above para-
graph on C. latrans), and from the Guy
Wilson Cave, Sullivan County, Tennessee.
According to Berta and Marshall (1978), fos-
sils referrable to C. dims have been reported
in South America from Talara, Peru; Tarija,
Bolivia; and Muaco, Venezuela.

Literature Cited

Arredondo, O., and L. S. Varona

1974. Nuevos genero y especie de mamifero (Car-
nivora: Canidae) del Cuaternario de Cuba.
Poeyana (Havana), no. 131, 12 pp.
Beebe, B. F.

1978. Two new Pleistocene mammal species from
Beringia. Amer. Quat. Assoc, Abstr. 5th
Bien. Mtg., p. 159.
Berta, A., and L. G. Marshall

1978. South American Carnivora. In Westphal,

F. (ed.), Fossilium catalogus, I: Animalia,
W. Junk, The Hague, part 125, ix + 48 pp.

Bjork, P. R.

1974. Additional carnivores from the Rexroad For-
mation (upper Pliocene) of southwestern
Kansas. Trans. Kansas Acad. Sci., 76:24-38.

Corner, R. G.

1977. A late Pleistocene-Holocene vertebrate fauna
from Red Willow County, Nebraska. Trans.
Nebraska Acad. Sci., 4:77-93.

136

1979

NOWAK: NORTH AMERICAN QUATERNARY CAN1S

137

Dalquest, W. W.

1978. Early Blancan mammals of the Beck Ranch
local fauna of Texas. Jour. Mamm., 59:269-
298.
Grayson, D. K.

1977. Paleoclimatic implications of the Dirty
Shame Rockshelter mammalian fauna. Teb-
iwa, no. 9, 26 pp.

Martin, L. D., B. M. Gilbert and D. B. Adams

1977. A cheetah-like cat in the North American
Pleistocene. Science, 195:981-982.

PARMALEE, P. W., P. J. MUNSON, AND J. E. GUDLDAY

1978. The Pleistocene mammalian fauna of Har-
rodsburg Crevice, Monroe County, Indiana.
Natl. Speleol. Soc. Bull., 40:64-75.

APPENDIX A

The following list provides details on samples
used in multivariate analyses. Specimens are in the
USNM unless otherwise indicated.

1. Canis lupus taken not later than 1925 in the
mountainous region of western North America.

C. I. irremotus

ALBERTA.— 25 mi. SE Lethbridge, 1.

IDAHO.— Bannock Co.: 10 mi. E Pocatello, 1;
Tyhee Basin, 1. Caribou Co.: Soda Springs, 2. Clark
Co.: Argora, 1. Lemhi Co.: Leadore, 1; 10 mi. S
Leadore, 1.

MONTANA. — No precise locality, 1. Beaverhead
Co.: Dillon, 1. Carbon Co.: Red Lodge, 1. Carter
Co.: Ridge, 1. Cascade Co.: Belt, 1. Powder River
Co.: Kruger, 1. Rosebud Co.: Ingomar, 1; Lame
Deer, 1.

WYOMING.— Campbell Co.: Gilette, 1. Con-
verse Co.: Glenrock, 1; Lost Springs, 1. Fremont
Co.: Lenore, 1; Split Rock, 1. Johnson Co.: Barber,
1. Sublette Co.: Cora, 1; Pinedale, 1. Sheridan Co.:
Arvada, 2. Teton Co.: Elk, 3; Kelly, 1. Yellowstone
National Park, 1.

C. I. mogollonensis

ARIZONA. — No precise locality, 1. Apache Co.:
Escudilla Mts., 3. Greenlee Co.: Clifton, 1; 15 mi.
SW Alma, New Mexico, 1. Maricopa Co.: Aguila, 1.
Navajo Co.: Cibecue, 1; Heber, 1.

NEW MEXICO.— Catron Co.: Datil Mts., 1; Gila
National Forest, 6; Luna, 1; 15 mi. SE Reserve, 1.
Grant Co.: head of Mimbres River, 1; Silver City, 1.
Sierra Co.: Fairview, 1; Chlorida, 5; Monticello, 1.
Socorro Co.: Magdalena, 1.

C. I. youngi

COLORADO. — Mesa Co.: no precise locality, 1;
Glade Park, 2; West Creek, 1. Pueblo Co.: 25 mi.
NW Pueblo, 1. Rio Blanco Co.: Piceance, 2; Sul-
phur, 1; Turman's Creek, 1.

NEW MEXICO.— Rio Arriba Co.: Abiquiu, 2;
Canjilon, 1; Dulce, 2; El Vado, 1; Hayes, 2. San
Juan Co.: La Plata, 1. Sandoval Co.: Cuba, 3;
Senorita, 1. Santa Fe Co.: Lamy, 1. Valencia Co.:
San Mateo, 1.

UTAH. — No precise locality, 1. Box Elder Co.:
Grouse Creek, 1. Duchesne Co.: Duchesne, 1. San
Juan Co.: 10 mi. NW Monticello, 2.

WYOMING.— No precise locality, 2. Laramie
Co.: Federal, 2. Sweetwater Co.: Rock Springs, 1.
Not located, Black Tail Creek, 1.

2. Canis latrans lestes taken not later than 1925
in the mountainous region of western North America.

COLORADO.— Conejos Co.: Bountiful, 1; Ceni-
cro, 4; La Jara, 4; Rio Grande, 1. Delta Co.: Cedar
Edge, 1; Grand Mesa, 2. Garfield Co.: Austin, 2;
East Salt Creek, 2; Salt Creek, 1. Grand Co.:
Kremmling, 4. Larimer Co.: Arkins, 3; Loveland,
1. Mesa Co.: Mesa, 1. Moffatt Co.: Craig, 2. Park
Co.: South Park, 2; Tarryall, 3. Rio Blanco Co.:

Piceance, 6. Rio Grande Co.: Monte Vista, 24.
Routt Co.: Battle Creek, 4; Russell Springs, 4;
Steamboat Springs, 2. Summit Co.: Gore Range, 1.
IDAHO.— Ada Co.: Boise, 1. Bannock Co.: no
precise locality, 1; Chesterfield, 1; McCammon, 1;
Pocatello, 3; Tyhee Basin, 2. Bingham Co.: Alridge,
2; Cerro Grande, 4; Ft. Hall, 2. Blaine Co.: Saw-
tooth National Forest, 3. Bonneville Co.: John Gray's
Lake, 1. Boundary Co.: Schnoors, 1. Canyon Co.:
Bowmont, 4. Caribou Co.: Preuss Mts., 2. Cassia
Co.: Almo, 1; Oakley, 6. Clark Co.: Dubois, 1;
Kilgore, 1; Medicine Lodge Creek, 4. Custer Co.:
Bigfoot River, 2. Elmore Co.: Arrow Rock, 1.
Gooding Co.: Gooding, 1. Goodnow Co.: Bliss, 1.
Idaho Co.: Orangeville, 1; Rice Creek, 2; West Lake,
7; White Bird, 1. Lemhi Co.: Leadore, 2; Leesburg,
1; Salmon, 3. Lewis Co.: Forest, 1; Salmon River,
2. Lincoln Co.: Shoshone, 1. Owyhee Co.: Grand
View, 1; Grassmere, 1; Hot Springs, 2; Three Creek,
8. Payette Co.: French, 1. Pegram Co.: Bear Lake,
3.

3. Canis familiaris.— 50 (10 in KU, 6 in MCZ,
1 in ROM, 5 in UArk, 3 in USFWS).

4. Canis lupus from northern and western North
America (other than as listed in 1 above).

C. I. alces

ALASKA. — Kachemak Bay, Kenai Peninsula, 2.

C. I. arctos

NORTHWEST TERRITORIES.— Ellesmere Is-
land: Bear Peninsula, 1 (CNM); Eureka Sound, 6
(CNM); Griese Fjord, 2 (CNM); Hare Fjord, 1
(CNM); Slidre Fjord, Foshien Peninsula, 8 (6 in
CNM). Graham Island: Norwegian Bay, 1 (CNM).
Prince Patrick Island: Cherie Bay, 1; Mould Bay,
1 (CNM).

C. I. baileyi

ARIZONA. — Cochise Co.: Huachuca Mts., 1.
Pima Co.: 5 mi. SE Arivaca, 1; Helvetia, 2.

NEW MEXICO.— Dona Ana Co.: Hatch, 1.
Grant Co.: Cloverdale, 2; Hatchita, 4. Hidalgo Co.:
Animas, 1; 30 mi. SE Animas, 2; 35 mi. SE Animas,
1; Animas Mts., 1; Animas Peak 1 (KU); San Luis
Valley, 1.

TEXAS.— Brewster Co.: 10 mi. S Alpine, 1
(SR). Jeff Davis Co.: Fort Davis, 1. Pecos Co.:
near Longfellow, 1 (SR).

CHIHUAHUA.— Colonia Garcia, 1; Colonia
Juarez, 1; near corner adjoining Sonora, Arizona,
and New Mexico, 3.

SONORA.— Sierra Pinto Mts., 1.

C. I. beothucus

NEWFOUNDLAND (Island).— No precise lo-
cality, 3 (2 in MCZ).

C. /. bernardi (including all specimens from Banks
Island)

138

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

139

NORTHWEST TERRITORIES (all in CNM).—
Banks Island: no precise locality, 1; North Adam
River, 3; Big River, 1; Egg River, 1; 25 mi. E Sachs
Harbor, 2.

C. I. crassodon

BRITISH COLUMBIA.— Vancouver Island:
Quatsino Sound, 2.

C. I. fuscus

BRITISH COLUMBIA.— No precise locality, 1.

OREGON. — Clackamas Co.: Clackamas Lake, 1
(SD). Curry Co.: Rogue River, 2. Douglas Co.:
Tiller, 2; Glide, 2. Jackson Co.: 25 mi. NE Ashland,
1; Peavine Mt., 3 (SD). Lake Co.: Sycan, 1. Lane
Co.: 20 mi. S Oakridge, 1 (SD). Linn Co.: Cas-
cadia, 2.

WASHINGTON.— Jefferson Co.: 22 mi. S Port
Angeles, 1.

C. I. hudsonicus

NORTHWEST TERRITORIES.— Aberdeen Lake,
5 (CNM); Beaver Hill Lake, 3 (CNM); Cape Ful-
lerton, 1 (AMNH); Hudson Bay, 1 (AMNH); Nuel-
tin Lake, 1 (CNM); Red River, 1 (KU); head of
Schultz Lake, 2; Simon's Lake, 1; Thelon River, 1;
Wajer River, 1 (AMNH).

C. I. labradorius

NEWFOUNDLAND ( LABRADOR ) .—Porcu-
pine, 1.

QUEBEC (UNGAVA).— No precise locality, 1.

C. I. ligoni

ALASKA. — Conclusion Island, 1; Ketchikan, 1;
Kuiu Island, 2; Kupreanof Island, 2; Prince of Wales
Island, 2; Revillagigedo Island, 1; Wrangell, 6.

C. I. mackenzii

NORTHWEST TERRITORIES.— Amundsen
Gulf, 1 (CNM); south side of Coronation Gulf, 1;
Port Epworth Harbor, 1 (CNM); head of Hood
River, 1 (CNM); Mackenzie Delta, 3 (CNM); Rae
River, 1 (CNM).

C. I. manningi

NORTHWEST TERRITORIES.— Baffin Island:
no pricise locality, 2 ( 1 in CNM, 1 in collection of
Douglas H. Pimlott); Pangnirtung Fjord, 1 (CNM).

C. I. monstrabilis

NEW MEXICO.— No precise locality, 1. Otero
Co.: Elk, 2; Sacramento Mts., 2; Mayhill, 1.

TEXAS.— Crockett Co.: Ozona, 2. Culbertson
Co.: Guadalupe Mts., 1. Jack Co.: Fort Richardson,
1. Kimble Co.: 1. Presidio Co.: 40 mi. SW Marfa,
1. Reagan Co.: Big Lake, 1. Upton Co.: Rankin,
6. Ward Co.: Monahans, 1.

C. I. nubilus

MANITOBA.— Southeast of Carberry, 1 (CNM);
Duck Mountain, 2; Riding Mountain National Park,
3 (CNM).

COLORADO.— Bent Co.: 3.

KANSAS.— Gove Co.: 3 mi. W Castle Rock, 1.
Trego Co.: near Castle Rock, 1 (KU).

MINNESOTA.— Becker Co.: 25 mi. N Detroit
Lakes, 1.

NEBRASKA.— Platte River, 3. Kearny Co.: Ft.
Kearny, 3.

NEW MEXICO.— Guadalupe Co.: Santa Rosa,
1. Lincoln Co.: 40 mi. SE Corona, 1. Socorro Co.:
Carthage, 3. Torrence Co.: Mountain Air, 1.

NORTH DAKOTA.— Billings Co.: Medora, 2.
Golden Valley Co.: near Beach, 1.

OKLAHOMA.— Panhandle area, 1 (AMNH).
Comanche Co.: Wichita Mountains National Wild-
life Refuge, 1.

SOUTH DAKOTA.— No precise locality, 1. Cus-
ter Co.: Folsom, 1. Harding Co.: 20 mi. NE Buffalo,
1. Meade Co.: Faith, 1. Pennington Co.: Imlay, 1.
Ziebach Co.: Red Elm, 1.

WYOMING.— Converse Co.: Douglas, 2. Na-
trona Co.: Natrona, 2.

C. I. occidentalis

ALBERTA. — Edmonton, 1; Simonette River, 1
(UAlb); 30 mi. N Whitecourt, 3 (UAlb); 50 mi. N
Whitecourt, 3 (UAlb); Wood Buffalo National Park,

4 (AMNH).

BRITISH COLUMBIA.— Barking Horse River, 2
(KU); upper Henry River, 1.

NORTHWEST TERRITORIES.— Artillery Lake,

5 (4 in CNM); Aylmer Lake, 1 (AMNH); Fort
Good Hope, 1; Fort Simpson, 1; Fort Smith, 1; Great
Bear Lake, 1 (AMNH); 52 mi. up Keele River, 1
(CNM); Nahanni Butte, 10 (CNM); mouth of
Netla River, 1 (CNM); Salt Plains, 5 (CNM); Slave
River, 4 (CNM).

YUKON. — 40 mi. SE Crow Base, 3; north fork
McMillan River, 1; Pelly Lakes, 4; White River, 4
(ROM).

C. I. orion

GREENLAND.— No precise locality, 2 (AMNH).

C. /. pambasileus

ALASKA. — No precise locality, 1. Anaktuvak
Pass, 7 mi. N Tolugak, 1; Big Delta River, 1; Cold
Bay, 1; Fairbanks, 1; 100 mi. N Fairbanks, 1; Fare-
well Mts., 1; Gold Creek (near head, above Curry),
1; Jarvis Creek, 1; upper John River, 7; Little Delta
River, 1; Mt. Hayes, 3; Nome, 1; Savage River, 1;
Sushana River, 2; Tanana River, 2; Teklanika River,
2; Teller, 1; Tolugak Lake, 1; Yukon River, 35 mi.
below Beaver, 1.

YUKON. — No precise locality, 1; Hoole Canyon,
1.

C. I. tundrarum

ALASKA. — No precise locality, 1; Noatak River,
2; Pitmega River, Cape Sabine, 1; Point Barrow, 1;
Umiat, 2; upper Meade River, 1 (UCMVZ); Wahoo
Lake, Brooks Range, 1 (KU).

C. I. youngi

CALIFORNIA. — San Bernardino Co.: 12 mi. W
Lanfair, 1 (UCMVZ).

5. Canis latrans from northern and western North
America (other than as listed in part 2).

140

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

C. I. incolatus

ALASKA.— Big Delta River, 12; Copper River
Flats, 1; Eagle River, 4; Fairbanks, 2; Mt. Hayes, 4;
Tanana, 1.

C. I. latrans

WYOMING.— Albany Co.: lelm, 1; Laramie, 22;
Red Mts., 1. Carbon Co.: Shirley, 1. Converse Co.:
Douglas, 4. Crook Co.: Manville, 4; Sundance, 4.
Laramie Co.: Federal, 4. Natrona Co.: Casper, 1.
Sheridan Co.: Arvada, 1.

C. I. mearnsi

ARIZONA. — Apache Co.: Marsh Lake, 1; Spring-
erville, 1. Coconino Co.: Anderson Mesa, 1; Bright
Angel Spring, 1; Flagstaff, 1; Fredonia, 3; Kaibab
National Forest, 1; Ryan, 6; Tuba, 4. Graham Co.:
Chiricahua Ranch, 1. Mojave Co.: Trumbull Mts.,
1. Navajo Co.: Antelope Springs, 1; Ft. Apache, 4.
Yuma Co.: Gila Mts., 1; Tinajas Altas, 1; Tule Tanks,
1.

C. I. texensis

NEW MEXICO.— Bernalillo Co.: Isleta, 1. Eddy
Co.: Salt Valley, 2. Lincoln Co.: Callo Canyon, 3.
Otero Co.: Cienega, 3; Lincoln National Forest, 3.
San Juan Co.: Fruitland, 1. San Miguel Co.: Pecos,
1. Santa Fe Co.: Lamy, 2. Socorro Co.: Carthage,
6; San Andres Mts., 1. Torrance Co.: Manzano Mts.,
1; Mesa Jiminez, 1.

6. Suspected hybrids.

Canis lupus x Canis latrans

ARIZONA.— Not located, Lanks, 1.

CHIHUAHUA.— Colonia Garcia, 1 (MCZ).

VERACRUZ.— Orizaba, 1 (MCZ).

ONTARIO.— Captives, 2 (ROM). Lanark Co.:
Sherbrooke, 1 ( ROM ) . Nipissing District: Preston,
1 (ROM).

QUEBEC. — Gatineau Co.: northern part, 1
(QWS); central part, 1 (QWS); Gracefield, 1
(CNM, originally identified as C. latrans thamnos) ,
Papineau Co.: Montebello, 1 (ROM). Pontiac Co.:
Head Lake, 1 (CNM).

Canis lupus x Canis familiaris

MICHIGAN.— Luce Co.: McMillan, 1 (UMMZ).
Schoolcraft Co.: Cusino, 1.

NEW MEXICO.— Otero Co.: Sacramento Mts.,
2.

7. Canis lupus lycaon.
Western Group

ONTARIO.— A/goma District: Batchwana Bay,
1 (UCMVZ); McMahon Tvvp., 1 (ROM). Cochrane
Dist.: Kapukasing, 1 (collection of Douglas H. Pim-
lott). Kenora Dist.: no precise locality, 1 (ROM);
Ball Lake, 1 (CNM); Eagle Lake, 1 (ROM); 100
mi. W Fort William, 1 (CNM); Kenora, 1 (ROM);
Whitefish Bay, 2 (ROM). Parry Sound Dist.: Bur-
ton Twp., 1 (ROM); Carling Twp., 1 (ROM).
Rainy River Dist.: Quetico, 3 (UI). Thunder Bay
Dist.: Hurkett, 1 (ROM); Killala Lake, 2 (ROM);

Lake Leopard, 1 (ROM); Lape Nipigon, 2 (CNM);
north shore of Lake Superior, 1 (CNM); Silver Islet

1 (ROM).

MICHIGAN.— Alger Co.: southern part, 1
(UMMZ); Grand Marais, 1 (UMMZ); 14 mi. SW
Grand Marais, 1 (UMMZ); 25 mi. NE Munising, 1
(UMMZ). Baraga Co.: Phcshika River, 1 (UMMZ);
Sec. 6, T50N, R31W, 1 (MSU). Chippewa Co.:
north shore of Whitefish Bay, 1 (UMMZ); Sec. 35,
T47N, R5W, 2 (MSU). Delta Co.: West Escanaba
River, 1. Dickinson Co.: no precise locality, 1;
Randville, 1; West Escanaba River, 1. Gogebic Co.:
Iron River, 1 (UMMZ); Marinesco, 1 (UMMZ);
Presque Island, 2 (UMMZ); 7 mi. N Watersmeet, 1
(MSU). Houghton Co.: Kenton, 1. Luce Co.: north
of Newberry, 1. Marquette Co.: 30 mi. NW Mar-
quette, 2. Ontonagon Co.: Calderwood, 1. School-
craft Co.: 1.

MINNESOTA.— Beltrami Co.: Red Lake Na-
tional Wildlife Refuge, 1 (UMinn). Cook Co.: no
precise locality, 3 (AMNH); near Dunn Lake, 2
(UMinn); Horland, 1 (UMinn). Koochiching Co.:

2 ( 1 in TM, 1 in UMinn). Lake Co.: Clearwater
Lake, 1; Eskwagama Lake, 1; Hart Lake, 1; Horse
River, 1; South Fowl Lake, 1. Lake of the Woods
Co.: Baudette, 1 (UMinn); 12 mi. S Williams, 1
(UMinn). St. Louis Co.: Duluth, 1 (AMNH); Ely,
6 ( UMinn ) ; Four Town Lake, 1 ( UI ) . Sherburne
Co.: Elk River, 1.

WISCONSIN.— Vilas Co.: Eagle River, 1.

Eastern Group

ONTARIO. — Nipissing Dist.: Algonquin Pro-
vincial Park, 7 (3 in CNM, 3 in ROM); Bishop
Twp., 1 (ROM); Clancy Twp., 1 (KU); Lake Ni-
pissing, 1 (ROM); Preston, 1 (ROM); Whitney,
1 (UCMVZ). Peterborough Co.: north of Apsley,

1 (CNM). Renfrew Co.: Dacre, 1 (ROM).

QUEBEC. — Southern part, no precise locality, 2
(QWS). Gatineau Co.: Aylwin, 1 ( CNM ); Lucerne,

2 (UCMVZ). Labelle Co.: Boyer, 1 (QWS); La-
coste, 1 (QWS); Mont Laurier, 1 (QWS); Nomin-
ingu, 1 (QWS); Ste. Veronique, 1 (QWS); Val-
Barrette, 1 (QWS). Papineau Co.: Montebello, 2
(ROM). Pontiac Co.: near Cabonga Reservoir, 2
(QWS); Jim's Lake, 1 (CNM). Temiscamingue
Co.: 40 mi. NE Mattawa, 1.

8. Canis latrans thamnos.

MANITOBA.— Carman, 4; Duck Mountain, 1.

ONTARIO.— Algoma Dist.: Dean Lake, 1
(ROM); Prince, 1 (ROM); Tarbutt, 4 (ROM);
Wolford, 1 (ROM). Greg Co.: Markdale, 1 (ROM).
Huron Dist.: Zurich, 1 (ROM). Kenora Dist.: Ox-
drift, 1 (ROM). Kent Co.: Chatham, 1 (ROM).
Lambton Co.: Thedford, 1 (CNM). Lanark Co.:
Sherbrooke, 1 (ROM). Nipissing Dist.: Algonquin
Provincial Park, 2 (CNM). Norfolk Co.: 1 (ROM).
Parry Sound Dist.: Monteith, 1 (ROM). Peter-
borough Co.: Lakefield, 1 (CNM). Rainy River
Dist.: Pinewood, 1 (ROM).

QUEBEC— No precise locality, 2 (QWS).
Bcauce Co.: Beauceville, 1 (QWS). Charlevoix

1979

NOWAK: NORTH AMERICAN QUATERNARY CAWS

141

Co.: Baie St. Paul, 1 (QWS). U Islet Co.: St. Au-
bert, 1 (QWS). Maskinonge Co.: St. Leon, 1
(QWS). Portneuf Co.: Valcartier, 1 (QWS).

ILLINOIS.— Lake Co.: Camp Logan, 1 (FM).
McClean Co.: LeRoy, 1. Marshall Co.: 9 mi. W
Henry, 1.

INDIANA.— Clinton Co.: Jefferson, 1 (PUWL).
Jasper Co.: McCoysburg, 1. Newton Co.: 5 mi. S
Roselawn, 1 (PUWL). Tippecanoe Co.: West
Point, 1 (PUWL).

IOWA.— Adair Co.: Richland, 2 (KU). Appa-
noose Co.: Moravia, 1 (KU). Monroe Co.: 2 mi.
N Avery, 2 (KU).

MICHIGAN.— Alcona Co.: Aldair, 1 (UMMZ).
Alger Co.: Miners River, 1. Baraga Co.: no precise
locality, 2 (MSU); Baraga, 1 (UMMZ). Barry Co.:
1 (UMMZ). Cheboygan Co.: Beaugrand, 1
(UMMZ). Chippewa Co.: Brimley, 1 (UMMZ);
7 mi. NW Pickford, 1 (UMMZ); Race, 1 (UMMZ).
Clinton Co.: St. Johns, 1 (MSU). Crawford Co.:
Hanson Game Refuge, 1 (UMMZ). Delta Co.: Bark
River, 1; Rapid River, 1. Dickinson Co.: Cedar
River, 1 (UMMZ). Gogebic Co.: Ironwood, 2
(UMMZ); Montreal River, 1 (UMMZ). Houghton
Co.: Isle Royale, 3 (UMMZ). Ingham Co.: 1
(MSU). Iron Co.: 3 (MSU). Jackson Co.: Liberty,
1 (UMMZ). Marquette Co.: Negaunee, 2; Yalmar,
1 (UMMZ). Menominee Co.: Cedar River, 2
(UMMZ); Dagett, 1; Ingalls, 4 (UMMZ); Michiga-
mee River, 1; Whitney, 2 (UMMZ); Wilson, 2
(UMMZ). Montcalm Co.: 1 (MSU). Ontonagon
Co.: no precise locality, 1 (UMMZ); Ewen, 1
(UMMZ). St. Clair Co.: 1 (MSU). Schoolcraft
Co.: no precise locality, 2 (MSU); Manistique, 1
(UMMZ). Washtenaw Co.: Dexter, 1 (UMMZ).
Not located, Warheim, 1 (UMMZ).

MINNESOTA. — Beltrami Co.: no precise local-
ity, 4 (UMinn); Red Lake National Wildlife Refuge,
1 (UMinn). Isanti Co.: 1 (UMinn). Lake Co.:
Fernberg, 1 (UMinn). Lake of the Woods Co.:
no precise locality, 5 (UMinn); Norris Camp, 2
(UMinn). Pennington Co.: 1 (UMinn). Pine Co.:
1 (UMinn). Sherburne Co.: Elk River, 5.

NORTH DAKOTA.— Benson Co.: Ft. Totten, 1;
Sully Hill National Park, 1.

WISCONSIN.— Ashland Co.: Basswood Island,
Apostle Islands, 1. Forest Co.: Crandon, 4 (MSU);
Wabeno, 1 (MSU). Iron Co.: Kenosa, 1 (FM).
Vilas Co.: Eagle River, 1. Walworth Co.: Delavan,
1.

9. Wild Canis from the Northeastern United
States.

MAINE. — Franklin Co.: Rangley, 1. Kennebec
Co.: Monmouth, 1.

MASSACHUSETTS.— Berkshire Co.: Otis, 1
(MCZ). Franklin Co.: Colrain, 1 (MCZ); Leyden,
1 (MCZ).

NEW HAMPSHIRE.— Coos Co.: Lancaster, 1
(MCZ); Stewartstown, 1 (MCZ). Hillsborough Co.:
Temple, 2 (MCZ). Merrimack Co.: Boscawen, 1
(MCZ). Sullivan Co.: Croydon, 1 (MCZ).

NEW YORK.— No precise locality, 1 (NYEC).
Franklin Co.: Faust, 1 (NYEC); Santa Clara, 2
(NYEC). Lewis Co.: 4 (NYEC). Oneida Co.:
Hawkinsville, 2 (NYEC); Woodgate, 1 (NYEC).
Oswego Co.: Fulton, 1 (NYEC). Schenectady Co.:
1 (NYEC). Yates Co.: 1.

PENNSYLVANIA.— Clearfield Co.: Clearfield, 1.
Potter Co.: 1.

VERMONT. — No precise locality, 1. Addison
Co.: Granville, 1 (VFG). Chittenden Co.: Shel-
burne, 1 (VFG). Orange Co.: Brookfield, 1 (VFG).
Orleans Co.: Barton, 2 ( 1 in VFG); Glover, 1
(VFG); Jay, 1 (VFG); Troy, 2 (VFG). Rutland
Co.: no precise locality, 1 (VFG); Middletown
Springs, 1 (VFG); Sudbury, 1 (VFG). Washington
Co.: Berlin, 1 (VFG); Montpelier, 1 (VFG). Wind-
ham Co.: Brookline, 2 (MCZ); Wardsboro, 1
(MCZ).

10. Canis rufus gregoryi, 1919-1929 (for data on
earlier material see table 2).

ARKANSAS.— Boone Co.: Bergman, 1. Cle-
burne Co.: Almond, 1. Dallas Co.: Carthage, 1.
Garland Co.: Crystal Springs, 1; Lonsdale, 2 ( 1 in
MCZ). Marion Co.: Mull, 1. Newton Co.: Falls-
ville, 7; Lurton, 2. Perry Co.: Ava, 1; Cedar, 1.
Polk Co.: Egger, 1; 12 mi. NE Egger, 2; 10 mi. W
Egger, 1; Mena, 1; Shady, 1. Pope Co.: Mill Creek,
2; Simpson, 5; Solo, 1. Pulaski Co.: Femsdale, 5;
Pinnacle, 2. Saline Co.: Isaac, 6. Scoff Co.: Blue
Ball, 4; Cardiff, 2; 4 mi. S Parks, 1. Yell Co.: 8 mi.
NW Aly, 1; Onyx, 8; Stillwater, 3.

LOUISIANA. — Beauregard Parish: near Sabine
River, 2. Madison Pa.: 1.

MISSOURI.— Carter Co.: Barren, 4. Crawford
Co.: Cook Station, 3. Howell Co.: West Plains, 2.
Iron Co.: Arcadia, 2. Ripley Co.: Gatewood, 3.
Stone Co.: 3. Texas Co.: Tyrone, 1. Wayne Co.:
Upalika, 1.

OKLAHOMA. — Le Flore Co.: Octavia, 1; Page,
3; Talihina, 1. McCurtain Co.: Bethel, 3; Broken
Bow, 7; Sherwood, 4; Smithville, 7. Pushmataha
Co.: Cedar Creek, 1; Fewell, 1; Nashoba, 1,

11. Southeastern specimens that suggest hybridi-
zation with Canis familiaris.

ARKANSAS.— Pope Co.: 4 mi. S Raspberry, 1.

LOUISIANA. — Northern part, no precise locality,
1 (LPI). Jackson Pa.: 1 (LPI). Winn Pa.: Sikes,
1.

MISSOURI.— Iron Co.: 4 mi. S Sabula, 1.

TEXAS.— Lavaca Co.: 20 mi. S Hallettsville, 1.
Van Zandt Co.: 1.

12. Specimens, originally identified as C. rufus
gregoryi, with short greatest lengths.

ARKANSAS.— Marion Co.: Mull, 1. Newton
Co.: Fallsville, 2. Perry Co.: Ava, 1; 8 mi. W Wye,
2. Pope Co.: Simpson, 1. Pulaski Co.: Fernsdale,
1. Sf. Francis Co.: Forrest, 1.

OKLAHOMA.— Le Flore Co.: Octavia, 1; Page,
1. McCurtain Co.: Bethel, 1; Broken Bow, 1; Smith-
ville, 2.

142

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

13. Canis latrans, pre-1930, southern Missouri. —
Carter Co.: Barren, 2. Phelps Co.: Rolla, 5. Saline
Co.: 4 mi. N Napton, 1 (MCZ). Texas Co.: Tyrone,
1.

14. Pre-1930 specimens originally identified as
Canis rufus rufus.

ARKANSAS.— Newton Co.: Boxley, 1.

OKLAHOMA.— Atoka Co.: near Atoka, 2. Gar-
vin Co.: Cherokee Town, 40 mi. N Ardmore, 1.
Tulsa Co.: Red Fork, 2.

MISSOURI.— Sfone Co.: Reeds Springs, 2.

15. Pre-1930 specimens originally identified as
Canis latrans.

OKLAHOMA.— Canadian Co.: Calumet, 5. Co-
manche Co.: Cache, 3. Creek Co.: Manford, 1.
Custer Co.: Anthon, 1; Butler, 5. Tillman Co.:
Frederick, 9. Tulsa Co.: Red Fork, 1.

TEXAS.— Hemphill Co.: 2.

16. Canis rufus rufus, 1900 and 1904, coastal
Texas. — Calhoun Co.: O'Connorsport, 4; 7 mi. SW
Port Lavaca, 2. Colorado Co.: Frelsburg, 1. Liberty
Co.: 6 mi. N Dayton, 1.

17. Canis latrans texensis, pre-1930, southern
Texas. — Frio Co.: Frio Town, 1; 11 mi. W Frio
Town, 1; 20 mi. W Frio Town, 2; 8 mi. SW Frio
Town, 4; 9 mi. S Moore, 1; Pearsall, 7; 5 mi. E
Pearsall, 1; 20 mi. W Pearsall, 3. Nueces Co.: Cor-
pus Christi, 27; 45 mi. SW Corpus Christi, 3; Nueces
Bay, 2; San Diego, 1. Uvalde Co.: Sabinal, 1; 10
mi. N Sabinal, 1; 5 mi. S Sabinal, 1. Zavala Co.:
12 mi. NE Batesville, 1.

18. Canis latrans texensis, pre-1930, western
Texas. — Brewster Co.: Alpine, 3. Coke Co.: 10 mi.
N Water Valley, 1. Crockett Co.: Ozona, 3; 9 mi.
W Ozona, 1; 12 mi. NW Ozona, 5. Pecos Co.: Shef-
field, 1. Reagan Co.: Big Lake, 4; 3 mi. N Big
Lake, 1; 25 mi. E Big Lake, 2; 12 mi. S Big Lake,
2. Sterling Co.: Broome, 1; Sterling City, 3; 30 mi.
S Sterling City, 1. Upton Co.: Rankin, 7; 10 mi. SW
Rankin, 9. Nolan Co.: Sweetwater, 1.

19. Canis latrans texensis, pre-1930, Tom Green
County, Texas. — Carlsbad, 1; 6 mi. NE Carlsbad, 1;
15 mi. NE Carlsbad, 2; Christoval, 4; 6 mi. NE
Christoval, 1; 10 mi. NE Christoval, 1; 15 mi. NE
Christoval, 2; 20 mi. NE Christoval, 1; Mereta, 1;
San Angelo, 25; 15 mi. W San Angelo, 2; Water
Valley, 9.

20. Specimens from central Texas, pre-1930. —
Blanco Co.: Blanco, 2; Round Mt., 1. Burnet Co.:
Burnet, 1; 5 mi. E Fairland, 1; Marble Falls, 6; 6
mi. S Marble Falls, 1. Coleman Co.: 16 mi. N Cole-
man, 1. Concho Co.: 5 mi. N Pasche, 1. Edwards
Co.: Nueces River, 1. Gillespie Co.: 2. Kerr Co.:
no precise locality, 4; Kerrville, 1. Llano Co.: no
precise locality, 4; Baby Head, 1; 22 mi. S Bird
Range, 1; Castell, 7; Click, 2; Llano, 7; 20 mi. N
Llano, 1; 15 mi. E Llano, 1; 20 mi. S Llano, 3;

7 mi. NW Llano, 2; Vallev Springs, 2. McCulloch
Co.: Brady, 3; 13 mi. SW Brady, 1; 5 mi. SE Doole,
1. Menard Co.: Callan, 1; Ft. McKavett, 1; Menard,
10. San Saba Co.: Cherokee, 2. Sutton Co.: So-
nora, 1; 25 mi. W Sonora, 2.

21. Canis rufus, 1930's-1950's.
C. r. gregonji

ALABAMA. — Sumter Co.: Livingston, 1.

ARKANSAS. — Union-Columbia county line, 1.

LOUISIANA— La Salle Pa.: Little River, 1
(LSUMZ). Madison Pa.: Tallulah Reservation, 2
(LUSMZ). Terrebonne Pa.: near Houma, 1
(LSUMZ). Winn Pa.: 3.

MISSISSIPPI.— Harrison Co.: Biloxi, 1 (AMNH).

OKLAHOMA.— McCurtain Co.: near Battiest, 2
(UArk).

TEXAS. — Hardin Co.: no precise locality, 1;
Honey Island, 1 (UAriz); Kountze, 1. Newton Co.:
1. Polk Co.: southern part, 1; Carmona, 1
(UCMVZ); near Wakefield, 2.

C. r. rufus

TEXAS.— Brazoria Co.: 12 mi. S, 4 mi. E Alvin,

I (KU); Angleton, 1; 9 mi. NE Angleton, 2; 5 mi.
E Angleton, 1; 12 mi. E Angleton, 1. Brazos Co.:
15 mi. S Bryan, 1. Harris Co.: Genoa, 1. Liberty
Co.: Cleveland, 1; 1.5 mi. N Rye, 2. Madison Co.:

II mi. SE Madisonville, 2. Montgomery Co.: Porter,
2; Security, 2. Walker Co.: New Waverly, 1.

22. Specimens from the central coast of Texas,

1936-1942.— Aransas Co.: Aransas National Wildlife
Refuge, 5. Refugio Co.: 22 mi. E Refugio, 1; 12 mi.

5 Tivoli, 1; 7 mi. S Woodsboro, 1. Victoria Co.:
Bloomington, 1; 6 mi. S Bloomington, 1.

23. Specimens from northern Texas, 1930-1942. —
Eastland Co.: Cisco, 1. Jack Co.: Henry Lewis
Ranch, 2; 25 mi. NW Jacksboro, 1. Palo Pinto Co.:
no precise locality, 1; 6 mi. NE Graford, 1. Parker
Co.: 1. Shackelford Co.: 2. Throckmorton Co.: 5.
Wilbarger Co.: 16 mi. SE Vernon, 1. Young Co.:

6 mi. NE Murray, 3 ( also examined, 3 specimens
from near San Antonio, Bexar Co.).

24. Canis latrans frustror, Wichita Mountains Na-
tional Wildlife Refuge, Comanche County, south-
western Oklahoma, 1933-1942.--47.

25. Specimens from central and northeastern
Oklahoma, 1932. — Cherokee Co.: 1. Cleveland Co.:
Noble, 3. Osage Co.: 2 (also examined, one speci-
men of C. r. rufus from Redden, Atoka Co.).

26. Specimens from southern Missouri, 1941-1942.
— Christian Co.: 1. Crawford Co.: 1. Taney Co.:
3. Texas Co.: 1. Vernon Co.: 9 (also examined,
one specimen from Dade Co., collected 1932; one
from 3 mi. N Thomasville, Oregon Co. (UCMVZ),
collected 1942; and one from 5 mi. N Gainesville,
Ozark Co. (UCMVZ), collected 1941).

27. Specimens from Arkansas, 1930-1951. — No
precise locality, 1 (UArk). Benton Co.: Cherokee

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

143

City, 3; Siloam Springs, 1; Springtown, 1. Howard
Co.: Umpire, 1, Lawrence Co.: 1. Stone Co.: State
Game Refuge, 1 (UArk). Washington Co.: Sum-
mers, 2; Devil's Den State Park, 2 (UArk).

28. Specimens from southeastern Oklahoma, post-
1960. — Bryan Co.: 4. Choctaw Co.: 8. McCurtain
Co.: 6. Pushmataha Co.: 7.

29. Specimens from northern Arkansas, post- 1960
(all in UArk).— Conway Co.: 8. Franklin Co.: 2.
Newton Co.: 2. Pope Co.: 1. Van Buren Co.: 7.

30. Specimens from southern Arkansas, post-1960.
—Calhoun Co.: 1 (UArk). Chicot Co.: 3. Clark
Co.: 1. Hempstead Co.: 26 (15 in UArk). Hot
Springs Co.: 4 (UArk). Howard Co.: 1 (UArk).
Little River Co.: 14. Miller Co.: 6. Nevada Co.:

4 (UArk). Sevier Co.: 3 (UArk).

31. Specimens from Louisiana, post-1960. — Beau-
regard Pa.: near Merryville, 8. Bienville Pa.: 2 ( 1
in LPI, 1 in LSUMZ). Bossier Pa.: 1 (LPI). Con-
cordia Pa.: Ferriday, 1. De Soto Pa.: 3 (LPI).
East Carroll Pa.: 4 mi. N Transvlvania, 1 (LPI).
Jackson Pa.: 7 (LPI). Natchitoches Pa.: 4 (LPI).
Red River Pa.: 1. St. Landry Pa.: Thistlewaite
Game Management Area, 3 (LSUMZ). Union Pa.:
3 mi. S Farmerville, 1 (LPI); 2 mi. N Farmerville,
1 (LPI). Webster Pa.: 1 (LPI). Wed Baton Rouge
Pa.: 2 mi. W Addis, 1 (LSUMZ). Winn Pa.: 2 mi.

5 Brewster's Mill, 5 (LPI); 4 mi. E Dodson, 1
(LPI) (also examined, one specimen, apparently C.
rufus gregoryi, from near Washington, St. Landry
Pa., collection of Douglas H. Pimlott).

32. Specimens from inland east Texas, post- 1960.
—Bell Co.: 2. Bosque Co.: 2. Bowie Co.: 3.
Cherokee Co.: 12. Collin Co.: 2. Delta Co.: 2.
Denton Co.: 16. Freestone Co.: 7. Grayson Co.:
14. Hamilton Co.: 3. Hopkins Co.: 4. Hunt Co.:
11. Johnson Co.: 8. Lamar Co.: 7. Leon Co.: 10.

Limestone Co.: 2. Milam Co.: 5. Morris Co.: 3.
Rusk Co.: 2. Smith Co.: 4.

33. Specimens from the central coast of Texas,

post-1960.— Austin Co.: 7 mi. NW Sealy, 2 (MSU).
Calhoun Co.: 7. Colorado Co.: 14 (MSU). Fort
Bend Co.: 3. Lavaca Co.: 7. Matagorda Co.: 8.
Victoria Co.: 1 (USFWS) (also examined, one speci-
men, apparently C. rufus rufus, from near Armstrong,
Kenedy Co., collection of Russell E. Mumford).

34. Specimens from the vicinity of the Addicks
Reservoir, Harris Co., Texas, post-1960. — 12 (7 in
USFWS).

35. Specimens from the vicinity of the Clemens
Prison Farm, western Brazoria Co., Texas, post-1960.
—31 (11 in USFWS).

36. Specimens from the eastern part of Brazoria
Co., Texas, post-1960.— 15 mi. S Alvin, 1 (USFWS);
5 mi. E Angleton, 5; 7 mi. N Angleton, 1 (USFWS);
Graham Ranch, 1; near Hoskins Mound, 10 (7 in
USFWS); Liverpool, 1; Stringfellow Ranch, 1.

37. Specimens from the vicinity of the Big
Thicket southeastern Texas, post-1960. — Chambers
Co.: 7 mi. E Baytown, 4 (USFWS). Jasper Co.:
near New Blox, 1 (USFWS). Liberty Co.: no pre-
cise locality, 3; 1 mi. S Ames, 1 (USFWS); 4 mi.
S Ames, 1 (USFWS); 5 mi. S Dayton, 2 (USFWS);
2 mi. E Devers, 2 (USFWS); 5 mi. N Liberty, 2
(USFWS); 15 mi. E Liberty, 2 (USFWS); 3 mi.
S Raywood, 1 (USFWS); 5 mi. S Raywood, 3
(USFWS). Tyler Co.: near Fred, 2 (USFWS).

38. Canis rufus gregoryi, southeastern Texas,
1963-1970.— Chambers Co.: Anahuac National Wild-
life Refuge and vicinity, 6 (2 in USFWS); Barrows
Ranch, 1 (UO); Canada Ranch, 2 (USFWS);
Double Bayou, 4; Logan Ranch, 2; Monroe City, 1;
Smith Point, 1 (USFWS). Jefferson Co.: near Port
Arthur, 2.

APPENDIX B

This appendix provides measurements for some
of the key series used in multivariate analysis, and
for some of the fossil specimens examined. The num-
bered parts of the appendix (left margin) are the
same as referred to in the text. The numbers along
the tops of the columns correspond to the numbers
of the 15 measurements described below. If no
sample size (n) is indicated, or if an asterisk (•)
follows the sample size, then the figures shown are
actual measurements of individuals. Otherwise, the
five horizontal rows under the designation and sam-
ple size (n) of the series are mean, lower extreme,
upper extreme, standard deviation, and coefficient
of variation. Figures in parentheses, following the
sample size of most series for which sex is designated,
represent the number of specimens in the sample
that were unknown as to sex, but which were
judged to belong to the category indicated.

Descriptions of Measurements

1. Greatest length. — Length from anterior tip of
premaxillae to posterior point of inion.

2. Zygomatic width. — Greatest distance across
zygomata.

3. Braincase width. — Maximum breadth of
braincase across level of parietotemporal sutures.

4. Alveolar length of maxillary toothrow. — Dis-
tance from anterior edge of alveolus of PI to pos-
terior edge of alveolus of M2.

5. Maximum crown width across upper cheek
teeth. — Greatest breadth between outer sides of most
widely separated upper teeth (P4 or Ml).

6. Palatal width at PI. — Minimum width be-
tween inner margins of alveoli of first upper pre-
molars.

7. Width at CI. — Greatest breadth across max-
illae at outer edges of alveoli of canines.

8. Width of frontal shield. — Maximum breadth
across postorbital processes of frontals.

9. Postorbital constriction. — Least width across
frontals at constriction behind postorbital processes.

10. Length from toothrow to bulla. — Minimum
distance from posterior edge of alveolus of M2 to de-
pression in front of bulla at base of muscular process.

11. Height from maxillary toothrow to orbit. —
Minimum distance from outer alveolar margin of Ml
to most ventral point of orbit.

12. Depth of jugal. — Minimum depth of jugal
anterior to postorbital process, at right angle to its
anteroposterior axis.

13. Diameter of CI. — Maximum anteroposterior
width of upper canine at base of enamel.

14. Crown length of P4. — Maximum anteropos-
terior length of crown measured on outer side.

15. Crou;n width of M2. — Maximum transverse
diameter from outermost point to innermost point of
crown.

MEASUREMENTS

1 2

3

4

5

6

7

8 9 10 11

12 13 14 15

C. familiaris, n:

=50

217.2 112.4

58.95

70.22

68.12

29.76

41.64

60.94 39.21 60.24 33.61

15.25 11.21 19.28 10.91

151.0 84.0

50.5

52.5

51.5

21.5

30.0

40.5 32.2 33.6 20.5

10.1 8.4 14.4 7.7

285.0 154.0

65.0

88.0

85.5

42.3

59.0

87.4 44.8 88.0 53.5

23.6 14.0 22.7 13.0

30.88 12.91 3.07 8.19 7.33 4.96 6.42 9.78 3.17 11.24 6.78 2.67 1.47 1.66 1.27
14.27 11.48 5.21 11.66 10.76 16.67 15.42 16.09 8.08 18.66 20.17 17.51 13.11 8.61 11.64

2. Total sample of northern and western C. lupus, male, n=233(33)

259.6

141.1

65.92

86.63

82.20

31.89

48.27

65.41 41.46

66.46

40.91

19.56

14.57

25.92

13.82

235.0

126.0

58.8

76.5

72.2

26.3

40.2

55.1 31.0

57.0

33.0

14.5

10.9

22.2

11.4

293.0

164.0

71.8

98.4

94.0

39.1

55.0

76.9 49.0

78.5

50.8

24.1

17.2

30.5

16.7

12.27

6.11

2.46

4.17

3.76

2.09

2.73

4.72 3.10

4.56

2.96

1.53

1.21

1.42

.99

4.73

4.33

3.73

4.81

4.57

6.55

5.66

7.22 7.45

6.86

7.24

7.82

8.30

5.48

7.16

Total sample of northern and western C. lupus,

female

n=146(33)

247.7

133.5

64.89

83.70

78.30

30.53

45.73

61.35 40.46

62.68

38.50

18.23

13.53

24.79

13.44

224.0

120.0

59.2

73.7

70.2

24.3

39.3

50.6 34.1

54.0

32.4

14.8

11.4

22.2

11.2

278.0

154.0

71.3

95.1

90.3

37.6

53.9

73.5 48.5

75.8

45.8

23.4

15.9

28.2

16.3

12.18

6.62

2.69

3.78

3.60

2.10

2.59

4.18 2.86

4.26

2.82

1.35

.97

1.26

.91

4.92

4.96

4.14

4.52

4.56

6.88

5.66

6.81 7.07

6.80

7.32

7.40

7.17

5.08

6.77

Total sample of northern and western C

'. latrans, male,

n=166(2)

197.1

99.4

57.79

70.04

56.53

20.06

30.91

46.77 34.10

46.96

26.06

12.15

9.35 20.38

11.81

178.0

88.0

52.0

61.4

49.7

17.5

27.1

36.1 29.1

39.5

21.2

9.1

7.7

17.6

9.7

213.0

109.0

63.6

78.4

62.4

29.7

35.5

54.6 39.7

52.8

29.6

14.3

11.0

22.8

13.8

7.28

3.96

2.00

2.85

2.40

1.14

1.40

3.13 2.13

2.61

1.63

.88

.62

.96

.64

3.69

3.98

3.46

4.07

4.24

5.68

4.53

6.69 6.25

5.56

6.25

7.24

6.63

4.71

5.42

144

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

145

1 234 5 6789

Total sample of northern and western C. latrans, female, n=rlll(4)
188.1 95.0 56.80 67.39 54.19 19.47 29.44 44.81 33.74
172.0 87.0 53.7 60.7 48.8 16.7 26.7 38.7 28.7
204.0 106.0 60.2 74.0 60.4 22.5 32.4 53.6 40.9
6.57 3.45 1.59 2.56 2.22 1.12 1.55 3.12 2.18

10

11

12 13

44.49 24.75

39.0 21.9

51.3 28.7

2.36 1.46

11.66

9.7

13.8

8.78
7.7
10.0
.56

14

19.60

17.7

21.6

15

11.52
10.4
13.0
.55

3.49 3.63 2.80 3.80 4.10 5.75 5.26 6.96 6.46 5.30 5.90 7.36 6.38 4.39 4.77

C. lupus hjcaon, western group, male, n=42(ll)

253.2
238.0
274.0
7.35
2.90

136.7
125.0
150.0
5.00
3.66

66.38
59.8
72.0
2.99
4.51

84.39
77.5
90.3
2.71
3.21

79.42
73.0
85.5
2.80
3.52

30.31
26.7
34.3
1.93
6.38

45.75
41.6
50.6
2.10
4.59

C. lupus hjcaon, western group, female, n=30(8)

241.2
224.0
268.0
10.03
4.16

129.0
121.0
142.0
5.36
4.15

64.34

59.0

69.3

2.67
4.16

81.07
72.4
87.5
3.34
4.12

75.14
70.5
81.0
2.86
3.81

28.81
26.0
32.5
1.64
5.71

C. lupus hjcaon, eastern group, male, n=19(4)

947 1 1 .14 1 fift 95 R9 fid 77 85 98 TO

247.1

237.0
255.0
5.96
2.41

134.1
128.0
140.0
3.59
2.68

63.25
58.3
68.0
2.49
3.93

82.69
78.5
87.7
2.48
2.99

77.85
74.2
84.3
2.71
3.48

28.59
26.0
32.0
1.47

5.14

42.47
39.5
47.3
1.79
4.22

44.26
40.8
47.6
1.69

3.81

C. lupus hjcaon, eastern group, female, n=12(2)

231.4
223.0
241.0
6.64
2.87

125.0
116.0
132.0
4.79

3.82

62.68
60.5
66.0
1.70
2.71

79.15
75.0
83.5
2.92
3.69

73.76
69.0
78.3
3.20
4.34

26.39
23.5
30.0
2.17
8.21

40.76
37.6
44.3
2.08
5.11

235.0
250.0
235.0
228.0
246.0
240.0
234.0
224.0
250.0
247.0
240.0

121.0
128.0
126.0
115.0
120.0
122.0
114.0
122.0
126.0
124.0
120.0

64.1
65.4
60.6
64.1
61.1
64.1
61.1
60.9
62.6
65.7
66.4

81.0
78.5
79.5
82.5
78.9
80.4
77.5
82.8
82.6
80.2

71.1
71.0
66.0
70.0
72.5
70.7
70.1
71.2
71.0
69.4

25.1
26.6
27.2
24.0
28.5
26.7
27.4
28.1
25.8
26.1
26.7

38.3
43.8
42.3
39.0
42.5
40.4
39.9
40.5
39.4
40.2
40.2

64.52
54.1
76.5
5.61
8.69

60.43
49.4
89.7
4.81
7.96

60.99
49.4
72.8
5.52
9.07

56.58
51.9
60.6
2.86
5.05

41.56
35.3
46.3
2.57
6.18

40.14

34.9

46.1

2.76
6.88

65.70 38.65

60.0 34.0

76.0 45.0

3.25 2.35

4.95 6.07

61.21
51.7
70.0
4.26
6.96

36.04
32.0
40.9
2.20
6.10

18.72
16.3
22.8
1.27
6.80

17.24
15.6
20.8
1.20
6.98

36.93
35.0
42.5
2.21
5.99

59.32 3.5.10

54.6 32.3

64.4 37.7

3.30 1.96

5.57 5.58

4. Early specimens of C. rufus, as listed in table 2, male, n

84.0 70.0

11(4)°
56.7 41.3

61.9
57.4
58.9
53.0
52.4
49.8
57.6
58.2
56.5
52.5

43.3
37.7
41.3
35.4
35.7
38.5
38.3
32.3
37.0
38.3

Early specimens of C. rufus, as listed in table 2, female, n=3( 1)°
222.0 114.0 60.5 75.0 66.0 24.4 37.3 48.4 31.5
230.0 117.0 60.7 79.5 65.0 25.5 38.8 54.0 38.3
222.0 115.0 64.1 77.3 64.6 25.5 37.0 52.3 37.3

59.9
63.8
57.5
57.5
60.3
62.7
61.7
56.3
60.6
62.7
62.0

55.7
58.4
54.5

5. C. rufus gregoryi, south-central United States, 1919-1929, male, n=63( 1)

232.6

121.2

61.93

78.98

69.37

26.21

40.00

218.0

110.0

58.3

72.6

63.6

22.3

35.7

261.0

138.0

68.0

86.8

75.3

32.0

47.2

8.76

5.93

2.10

2.79

2.73

1.95

2.24

3.77

4.89

3.39

3.53

3.93

7.45

5.61

55.52 37.33 58.32

47.2 32.0 51.8

62.1 42.5 65.4

3.73 2.40 3.45

6.71 6.43 5.91

34.5
37.6
35.3
34.3
35.2
34.2
34.0
33.2
32.2
35.0
32.5

33.6
32.3
31.3

33.64
29.1
38.0
2.24
6.65

16.3
18.3
18.4
18.3
15.8
15.8
16.8
16.0
15.4
16.2
17.5

15.2
13.7
14.5

15.78
13.3
18.5
1.26
7.96

13.61

11.8

15.5

.99
7.30

25.19
23.7
27.4
.93
3.70

39.78 63.34 37.78 18.15

36.0 59.0 34.3 16.3

44.9 69.5 42.2 20.0

2.77 2.86 1.89 1.08

6.95 4.51 5.00 5.93

12.46 23.56
10.4 21.1
13.9 26.0
.89 1.27

7.18 5.38

13.09 24.55
12.0 22.6
14.8 27.5
.80 1.20

6.12 4.88

16.54
14.0
18.5
1.49
9.01

11.98
11.3

12.7

.50
4.21

22.67

21.3

24.2

.93
4.10

11.1
12.6
13.7
11.2
13.6
12.2
12.4
11.4
11.1
11.2
13.5

24.5
25.2
23.8
23.5
24.8
23.6
23.8
24.6
23.8
24.8
24.5

14.02
12.8

15.7

.67
4.79

13.39
11.0
14.8
.83
6.18

14.25
13.4
15.7
.66
4.64

13.64

13.2

14.3

.35
2.61

13.1
14.2
13.5
14.4
13.9
13.7
14.4
13.8
14.6
15.6
13.1

10.5 22.1 12.0
10.1 22.3 13.0
11.3 22.2 13.4

11.93 23.66
10.5 21.4
13.2 26.0
.73 1.00
6.10 4.23

13.68

10.6

16.0

.82
5.99

146 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 6

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
C. rufus grcgoryi, south-central United States, 1919-1929, female, n=52(l)

220.9 115.4 61.14 75.15 66.78 25.32 37.99 52.93 37.93 53.89 31.66 14.84 11.12 22.31 1.3.29

210.0 108.0 57.6 68.5 61.6 21.2 33.4 42.7 30.8 50.4 27.3 12.0 9.6 20.0 11.7

245.0 130.0 64.8 80.5 74.7 29.7 45.0 63.0 41.5 66.1 36.1 17.3 12.9 24.4 14.7

5.58 4.40 1.89 2.58 2.85 1.98 2.23 4.20 2.44 2.66 1.72 1.02 .71 1.08 .72

2.53 3.81 3.11 3.45 4.27 7.62 5.87 7.94 6.42 4.93 5.33 6.88 6.36 4.84 5.42

6. Specimens from inland eastern Texas, post- 1960, male, n=77(6)

rtr>r> o l A A n p^o m ^l r\r\ Pfl r»n m C/D oo on AC\ ere" Off

206.8 104.2 58.61 71.90 59.22 21.56 33.32 49.55 35.57 50.74 27.91 13.15 10.05 21.04 12.29

192.0 96.0 54.6 65.5 53.3 18.8 29.5 43.5 29.6 44.0 24.3 10.8 8.8 19.2 10.3

221.0 112.0 63.7 76.8 64.8 24.6 36.3 61.5 40.2 56.8 30.5 15.1 11.5 23.0 13.6

5.62 3.45 2.11 2.14 1.91 1.25 1.42 3.59 2.12 2.64 1.44 .97 .56 .85 .64

2.72 3.31 3.60 2.98 3.22 5.80 4.26 7.24 5.96 5.21 5.17 7.35 5.56 4.02 5.23

Specimens from inland eastern Texas, post-1960, female, n=42(6)

198.0 99.52 58.01 69.81 57.39 20.59 31.75 47.36 35.66 47.86 26.60 12.41 9.51 20.43 12.31

180.0 91.0 54.2 64.7 53.6 18.6 29.1 42.1 30.1 42.2 22.3 10.7 8.4 18.5 10.5

214.0 109.0 62.0 75.5 62.8 23.1 37.8 56.0 39.8 54.8 30.8 15.5 11.5 23.0 13.6

7.51 3.78 1.61 2.77 1.94 1.08 1.65 3.23 2.01 3.20 1.77 1.01 .64 .96 .72

3.80 3.79 2.77 3.97 3.39 5.24 5.19 6.83 5.63 6.69 6.65 8.15 6.70 4.72 5.85

Specimens from Jefferson and eastern Chambers counties, Texas, 1963-1970, male, n=15(3)

233.7 119.1 60.63 76.73 68.05 26.59 39.59 50.99 35.79 60.23 33.31 15.07 11.31 22.27 13.86

218.0 105.0 56.0 67.0 61.5 24.0 33.8 41.3 27.5 52.5 28.3 11.8 10.0 21.1 12.0

247.0 130.0 64.7 82.6 73.6 29.2 43.0 58.0 40.0 65.0 36.8 16.7 12.5 23.3 14.8

6.98 5.95 2.39 3.76 3.06 1.71 2.31 4.14 3.44 3.23 2.30 1.23 .64 .79 .80

2.99 4.99 3.95 4.90 4.50 6.43 5.83 8.12 9.60 5.36 6.90 8.15 5.63 3.55 5.79

Specimens from Jefferson and eastern Chambers counties, Texas, 1963-1970, female, n=4°

222.0 112.0 58.7 75.0 65.5 27.7 38.5 47.7 33.5 55.0 33.6 15.5 11.0 21.1 12.3

224.0 111.0 61.8 77.1 63.0 24.0 35.8 52.5 39.1 56.7 33.7 14.7 10.6 21.1 13.2

220.0 111.0 58.8 76.1 66.8 25.4 35.4 48.5 35.7 58.0 30.8 14.0 10.1 21.1 13.9

225.0 110.0 59.6 74.0 63.8 24.1 38.3 45.8 33.8 55.5 31.6 15.3 10.8 21.2 12.5

8. C. cedazoensis, Cedazo, Aguascalientes

9. C. lepophagus, Hagerman, Idaho

C. lepophagus, Rexroad fauna, Kansas

C. lepophagus, Broadwater, Nebraska
C. lepophagus, Lisco, Nebraska

16.9

17.0 10.3
16.0 10.0

19.1 ....

16.9 _..

— 10.2

19.1 11.5
19.6 12.3

17.9 11.0

C. lepophagus, Cita Canyon, Texas

194.0 103.0 .... 68.6 ... 20.5 ... 54.0 .... 11.8 .... _

187.0 .... 54.3 64.7 _._ _.. .... .... 26.6 .... 8.7 19.0 11.3

- - -- _ 57.2 35.0 51.0 .... .... .... _.. _

190.0 99.0 50.0 70.0 _.. 20.0 „ 48.0 33.0 _ 25.0 12.0 20.0 12.0

74.5 __ .... _.. .... .... _.. .... .... 8.8 20.7 12.0

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

147

10.

12 3 4 5 6

C. latrans, Papago Springs Cave, Arizona
174.0 99.0 56.9 67.1 57.1 20.2

C. latrans, McKittrick, California

10

11

12

13 14

15

30.0 40.6 32.9 43.1 25.4 12.6

19.6 11.3

'98.6

58.3

74.5

59.5

47.0

38.2 _

._ 25.2

11.5 _. 22.0

12.6

62.4

76.5

65.0

24.5

_ 27.6

.... 22.3

12.3

210.0

77.0

64.4

22.5

35.5

55.0

36.0 .

_ 28.0

13.1 .... 22.5

12.6

219.0

110.0

58.4

77.5

64.5

21.5

53.1

35.5

... 28.8

14.0 .... 24.0

13.8

207.0

72.5

61.1

23.0

33.3

.... 10.3 21.3

12.3

206.0

61.5

75.5

61.8

20.0

33.5

56.0

36.5 -

._

.... 22.5

—

C. latrans, Maricopa, California
204.0 — 59.0 73.5 56.2

C. latrans, Rancho La Brea, California

21.8 32.8

20.1

n=44

n=36

n=42

n=49

n=21

n=44

n=38

n=44

205.5

106.7

60.54

72.54

61.20

22.18

34.29

51.21

185.0

90.0

55.7

65.5

50.0

17.3

29.8

44.0

222.0

116.0

65.1

78.0

67.4

25.5

38.3

61.5

9.03

5.58

1.76

3.23

4.31

1.70

2.00

4.14

4.39

5.23

2.91

4.45

7.05

7.68

5.83

8.0S

;47 n=40 n=40 n=36 n=8

1 36.73 48.16 27.05 13.07 9.88

32.9 44.0 22.0 10.2 9.0

42.0 52.8 32.3 15.2 10.9

2.12 2.55 2.33 1.10 .53

5.78 5.30 8.63 8.41 5.32

C. latrans, Rancho La Brea, California (unusually small specimen)

179.0 90.0 56.5 64.0 50.8 19.6 28.7 41.1 34.9 41.8 22.5

10.2

Type of C. andersoni

173.0 93.0 57.5 63.5 56.0

11. C. edwardii, Curtis Ranch, Arizona

__ 118.0 __ 77.0 65.4 22.2

C. edwardii, Rome Beds, Oregon

59.0 78.8 68.0 19.5

12. C. rufus, Eddy Bluff shelter, Arkansas

C. rufus, Haile VIIA, Florida
78.0

35.1 38.3 33.3 40.0 24.5 12.2 .._

.... . _ 29.1 13.7 12.4

50.8 38.8 .... _... __ 11.3

14.5

C. rufus, Inglis IA, Florida

C. rufus, Port Kennedy, Pennsylvania

13. C. armbrusteri, McCleod, Florida

C. armbrusteri, Coleman IIA, Florida

__ 133.0 66.4 85.6 78.2 24.0

63.0 45.2

35.0 18.5

n=43

n=33

21.05

11.84

18.2

10.2

23.5

13.6

1.29

.89

6.14

7.54

18.0

10.2

20.5

_._

24.0

12.6

24.0

13.7

23.0

13.1

20.9

13.3

22.1

22.8

—

24.0

12.3

26.4

14.9

26.0

15.5

26.1

13.9

28.8

26.5

148

MONOGRAPH MUSEUM OF NATURAL HISTORY

NO. 6

14.

12 3 4 5 6 7
C. armbrusteri, Cumberland Cave, Maryland

258.0 128.0 .... 96.0 82.2 27.0 43.5
64.8 .... 78.1

.... 150.0 .... 95.0 87.0 29.0 45.8

285.0 161.0 75.0 98.0 80.0 30.7 49.0

10

11 12 13 14

15

270.0

94.0

65.0

83.5

61.5 50.0 61.7 34.5 17.2

59.8 44.5 51.0 ....

68.0 44.0

70.0 43.0 77.0 45.0 21.0

.... 65.0 35.0 ....

59.0 39.8 60.0 35.6 19.0

C. armbrusteri, Rushville, Nebraska

C. lupus, Hunker Creek, Yukon
257.0 .... 61.0 88.7

C. lupus, Maricopa, California
258.0 144.0 69.0 87.5
255.0

.... 135.0
C. lupus, Rancho La Brea, California

35.8 52.0 57.5 40.0 66.3 42.7 20.0

33.8
32.5

48.2 ' 64.0 39.0

41.6 18.8

28.0

26.6

14.1

10.5

12.8

28.7

1577

27.9

15.5

28.9

15.2

28.0

17.0

29.5

17.1

28.9

15.2

—

30.0

15.5

—

28.5

13.7

26.6

13.7

26.6

12.5

26.2

28.5

23.3

13.1

28.1

12.2

240.0

138.0

62.7

81.5

79.7

33.6

48.1

74.3

46.0

62.7

38.0

25.0 .

- 26.8

14.1

140.0

82.7

80.3

28.5

44.0

64.0

44.0

39.0

18.3 .

... 24.8

14.2

123.0

66.9

78.7

72.3

29.0

58.9

42.2

56.0

35.6

16.0

_ 23.0

12.5

230.0

128.0

66.0

77.8

78.8

29.0

45.0

53.0

42.7

56.3

37.0

19.0

... 26.8

13.1

265.0

150.0

71.4

90.0

85.7

33.3

49.4

68.0

48.2

66.7

39.5

21.0 .

_ 29.0

14.3

269.0

147.0

69.2

89.9

36.1

52.0

75.0

41.3

70.5

44.5

22.1 .

... 26.8

13.2

252.0

62.7

82.5

79.9

33.5

44.6

63.8

39.5

19.7 .

... 25.1

13.8

148.0

69.0

94.0

94.5

38.0

68.3

43.7

70.6

44.3

23.5 .

... 29.2

12.8

248.0

140.0

64.5

84.0

83.2

31.0

46.8

57.5

41.5

62.0

_ 27.6

13.9

....

—

—

....

....

—

—

—

—

....

—

—

_ 23.5
_ 25.5

12.5

Type of C. milleri

247.0

137.0

66.7

83.0

89.6

33.5

49.8

63.9

44.3

58.6

39.0

18.8 _

_ 28.6

13.2

C. lupus, Goodland, Kansas
223.1 126.0 67.0 76.5

C. lupus, Hay Springs, Nebraska

73.0 30.0 43.6 59.0 34.0 54.5 34.0 15.0 13.3 23.1 14.0

C. lupus, Hermit's Cave, New Mexico

C. lupus, San Josecito Cave, Nuevo Leon
216.0 125.0 63.0 76.5 75.0 26.5

15. C. dims, McKittrick, California

309.0 170.0 72.5 106.5 99.0 40.0

_ _.. .... .... __ _ 27.4 13.6

.... 14.9 27.9 14.6

40.0 54.0 40.5 50.5 32.4 17.7 11.9 25.0 12.4

58.0 78.5 47.8 75.5 49.4 24.0 15.8 33.0 15.6

1979

NOWAK: NORTH AMERICAN QUATERNARY CAMS

149

12 3 4

C. dims, Maricopa, California

311.0
298.0
300.0

310.0
318.0
290.0

174.0 76.0

160.0
168.0

158.0

74.5

80.0

73.0

106.1
100.0
102.5
100.0
101.0
105.0
101.0

96.7

95.2
97.0

44.3
38.0
39.0
40.0
40.0
39.0
38.0

C. dims. Rancho La Brea, California, n=62

294.8
258.0
316.0
11.31
3.84

163.3
148.0
177.0
7.15

4.38

74.73
64.0
83.0
3.08
4.12

99.99
85.0
107.0
4.18
4.17

96.15
87.7
104.0
3.92
4.08

39.27
35.0
45.3
2.36
6.01

69.0
59.0
60.0

65.0
58.0

58.02
52.0
65.5
3.30
5.68

10

90.8 52.5 77.5
93.0 ....

82.5
95.0
87.0
89.0

47.9
54.0

47.0

75.0

11

47.0

45.2
43.8
43.0

45.0

12

27.0

24.5
22.5
21.5
22.0
21.5
22.8

13

14

17.2 32.3

31.8

83.45
73.4
100.0
5.63
6.75

49,33
43.5
54.4
2.13
4.32

72.43
63.5
77.5
3.54
4.89

42.39
36.6
48.5
2.66
6.38

C. dims, Hornsby Springs, Florida

C. dims, Reddick IA, Florida

C. dims, Melbourne, Florida

C. dims, Bradenton, Florida

C. dims, Ohio River, Indiana

C. dims, Twelve Mile Creek, Kansas

C. dims, Welsh Cave, Kentucky

309.0 180.0 78.0 104.3 100.7 39.0 59.1 104.0 57.3 82.0 45.5 25.0

C. dims, Herculaneum, Missouri

C. dims, Hermit's Cave, New Mexico

40.4 20.7

C. dims, Ingleside, Texas
333.0 179.0 79.0 110.0

44.0

53.0 .... 48.8 27.0

32.5
31.0

21.75
18.8
26.5
1.66

7.63

C. dims, Marlow, Oklahoma

310.0 170.0 78.5 111.5 102.5 37.0 60.0 100.0 50.0 79.0 49.0 24.0 18.0

C. dims, San Josecito Cave, Nuevo Leon

— 161.0 72.0 102.0 101.2 37.6 _ 49 2 45 5 21 4

297.0 169.0 76.0 104.3 103.0 37.2 58.8 94.0 54.2 74.0 44.9 21.0 14.!

15

14.9
16.0
15.4
14.4

15.66
13.5
17.5
1.15

7.34

31.75
28.7
35.3
1.38
4.35

15.15
13.1
17.0
.90

5.94

29.6 14.5
32.2 _._

30.7 16.5

30.8 15.5

32.0
30.0

30.1 15.1

35.5 14.0

30.0 15.4

30.5 15.7

32.5 15.6

33.0 15.4

31.0 .._

35.5 17.5

33.6 15.4

33.7 15.0

APPENDIX C

The numbers along the top of each of the following columns correspond to four measurements of the man-
dible and lower dentition: (1) distance from anterior edge of alveolus of pi to posterior edge of alveolus of
m3; (2) minimum depth from dorsal surface of mandible between p3 and p4 to ventral surface of mandible;
(3) crown length of p4; (4) crown length of ml. If a sample size (n) is listed for a series, then the five
horizontal rows under that figure are mean, lower extreme, upper extreme, standard deviation, and coefficient
of variation. If no sample size is given, the numbers shown are actual measurements of individuals, rather
than means, etc.

1. C. lepoplmgus

Santa Fe River, Florida

Grand View, Idaho
Hagerman, Idaho

Rexroad fauna, Kansas

Broadwater, Nebraska
Lisco, Nebraska

Cita Canyon, Texas

2. C. latrans

Recent, western U.S., male

Recent, western U.S., female

Irvington, California
McKittrick, California

4

78.0

15.9

12.9

22.5

83.5

19.4

12.3

22.6

....

20.0

82.5

19.1

20.8

—

18.7

13.5

22.4

19.0

17.1

13.2

70.5

17.5

12.3

20.8

70.0

14.0

11.5

18.5

78.7

16.0

—

—

....

18.1

75.5

—

20.5

70.0

14.5

11.1

17.9

n=14

n=14

n=16

n=13

78.61

18.16

13.08

21.21

73.2

15.1

12.1

19.5

82.0

21.3

14.0

23.0

2.63

1.48

.52

1.15

3.35

8.14

3.94

5.42

n=99

n=99

n=99

n=99

79.78

17.10

12.58

21.94

71.4

14.3

10.7

19.5

88.5

19.9

14.0

24.3

3.26

1.14

.70

1.01

4.10

6.66

5.58

4.60

n=99

n=99

n=99

n=99

76.35

15.98

12.08

21.10

69.1

12.7

10.8

18.6

82.3

19.0

13.9

23.4

2.71

1.14

.60

.90

3.55

7.12

4.96

4.26

79.9

19.3

13.8

23.9

78.9

16.9

13.4

21.7

n=10

n=10

n=14

n=16

82.24

18.94

12.78

22.78

77.0

17.4

11.7

21.0

88.5

20.4

14.1

24.6

3.82

1.14

.70

1.11

4.64

6.02

5.51

4.88

150

1979

NOWAK: NORTH AMERICAN QUATERNARY CANIS

151

l

Maricopa, California n=10

82.80
80.0
85.0
1.81
2.19

Rancho La Brea, California n=41

81.77
77.0
87.5
2.52
3.08

Vallecito Creek, California 85.5

Haile XIIB, Florida __ 78.1

Devil's Den, Florida _.. 72.1

Lake Cutaline, Florida

Rushville, Nebraska _

Mullen, Nebraska 73.6

Frankstown Cave, Pennsylvania 83.5

Lewisville, Texas _. 82.0

Friesenhahn Cave, Texas 78.6

75.5

76.5

San Josecito Cave, Nuevo Leon : n = 12

80.25

76.0

84.0

2.61
3.25

C. edwardii

Anita, Arizona

Curtis Ranch, Arizona ..

Minaca Mesa, Chihuahua

n=10

n=15

n=22

18.15

12.75

23.13

17.0

11.6

21.1

19.8

13.8

25.0

1.00

.55

.84

5.55

4.33

3.62

n=41

n=37

n=40

18.69

12.56

22.44

16.0

11.3

21.0

21.1

13.7

25.2

1.30

.62

1.00

6.97

4.90

4.45

18.9

13.1

22.0

16.8

....

....

15.2

20.2

17.5

21.5

96.5
87.2
91.5

18.0

17.5

19.6

17.8

18.0
16.0
15.8
15.5

n=15

18.01

16.1

19.9

1.24

6.90

19.5
20.0

11.7

11.6

13.7
13.8

12.3
12.1
11.4

n=10
12.96
12.0
14.1

.69
5.30

16.0
15.2
14.5

21.4
20.4
23.3

23.8

21.4
20.4

20.7

n=9

22.44
20.8
24.8
1.44
6.43

27.5
24.8
23.3

Hemphillian specimen from Ash Hollow formation, Nebraska

4. C. rufus

Recent, male .._

Recent, female

82.0

20.3

13.5

25.0

n = 64

n=64

n=64

n=64

89.66

21.60

14.40

25.68

83.2

19.2

13.2

23.4

99.5

24.9

16.0

28.1

3.37

1.31

.80

1.04

3.77

6.08

5.56

4.05

n=61

n=61

n=61

n=61

86.05

20.98

13.94

24.52

79.6

18.1

12.3

22.1

93.6

25.2

15.4

27.1

2.78

1.47

.72

1.11

3.24

7.00

5.16

4.54

152 MONOGRAPH MUSEUM OF NATURAL HISTORY NO. 6

Inglis IA, Florida

Melbourne, Florida

Port Kennedy, Pennsylvania

5. C. armbrusteri

Anita, Arizona

McCleod, Florida

Coleman IIA, Florida

Cumberland Cave, Maryland

6. C. lupus

Recent, western U.S., male

Recent, western U.S., female

Maricopa, California

Rancho La Brea, California

Type of C. milleri

Goodland, Kansas

Millington, Michigan

Hay Springs, Nebraska

Mullen, Nebraska _

Hermit's Cave, New Mexico

1

2

3

4

27.5

87.5

20.7

15.6

26.0

94.0

25.2

15.0

26.3

—

....

15.8

—

105.5

27.5

18.0

102.5

23.6

16.4

31.3

113.0

27.5

17.8

32.0

27.4

28.5

30.9

—

....

28.0
32.0

104.0

27.0

17.3

29.4

18.2

110.0

28.2

17.6

30.5

106.5

28.2

18.4

30.0

105.0

28.7

17.7

30.5

30.7

....

30.9

— -

—

—

30.9

n=62

n=62

n=62

n=62

95.31

26.62

15.68

28.53

86.7

23.0

13.6

26.0

104.0

31.0

17.0

31.5

3.44

1.66

.73

1.20

3.61

6.25

4.68

4.22

n=47

n=47

n=47

n=47

92.15

24.90

15.04

27.07

84.7

22.5

12.7

25.1

97.5

27.7

16.5

30.0

2.66

1.29

.74

1.00

2.88

5.17

4.95

3.70

17.5

30.2

27.5

16.3

29.1

15.5

28.4

97.7

27.5

16.1

30.5

99.9

29.1

18.2

31.2

95.5

26.8

16.8

29.0

95.0

25.8

18.0

32.0

84.3

22.1

15.0

26.3

101.0

28.0

15.0

30.0

98.0

29.0

17.4

29.9

100.0

30.0

17.3

29.3

96.0

30.6

16.6

30.2

95.0

24.6

15.8

29.9

San Josecito Cave, Nuevo Leon 84.0 23.0 14.3 26.5

1979 NOWAK: NORTH AMERICAN QUATERNARY CANIS 153

12 3 4
7. C. dims

Murray Springs, Arizona 110.0 .... 18.6 34.8

18.4 33.9

Cool quarry, California ... . 31.5 19.0

Teichart gravel pit, California 20.8

Arroyo Las Positas, California 110.0 31.9 19.0 34.3

McKittrick, California n=7 n=7 n=13 n=15

113.40 33.73 20.08 35.33

106.5 32.0 18.8 33.7

123.0 35.0 21.7 37.0

5.57 1.25 .89 .98

4.91 3.71 4.41 2.78

Maricopa, California n=10 n=12 n=18 n=17

113.2 34.01 19.87 35.01

108.0 29.8 18.0 33.5

119.5 36.9 21.5 37.5

3.82 2.38 1.04 1.47

3.36 7.01 5.23 4.19

Rancho La Brea, California n=73 n=73 n=73 n=73

110.64 31.81 19.48 34.25

102.0 25.5 17.9 31.8

117.5 36.5 20.6 38.5

3.31 1.96 .66 1.44

2.99 6.17 3.36 4.20

Ichetucknee River, Florida 117.0 32.5 20.3 35.5

117.0 31.1 19.8 36.0
116.5 33.2

Hornsby Springs, Florida _. . 116.0 29.0 18.3 35.5

Reddick IA, Florida .... .... 19.3 34.4

21.3 37.5

Eichelberger Cave, Florida .._ 125.0 32.1 20.8 37.7

34.2 20.9 37.6

Melbourne, Florida 112.5 31.6 19.3 35.3

33.5
34.8
35.9

Bradenton, Florida ..._ 32.5 18.4 36.2

Twelve Mile Creek, Kansas .... 18.0

Pendennis, Kansas ... . .— 34.0

Cragin Quarry, Kansas 108.5 30.5 18.3

Welsh Cave, Kentucky 113.5 29.6 19.6 36.3

Conkling Cavern, New Mexico 32.1 20.4 35.2

Hermit's Cave, New Mexico 118.0 35.5

Marlow, Oklahoma _ 120.0 39.0 20.2 36.0

Frankstown Cave, Pennsylvania _ 108.6 30.0 17.9 32.0

110.5 29.2 17.7 32.6

154 MONOGRAPH MUSEUM OF NATURAL HISTORY

1 2

Ingleside, Texas 118.0 34.0

Rennick. West Virginia ..... 107.0 28.2

Cedazo, Aguascalientes 112.0

33.0
33.5

San Josecito Cave. Nuevo Leon - - n=9 n=10

110.0 33.23

107.0 30.8

118.2 36.0

3.32 1.63

3.02 4.92

NO. 6

3

4

20.1

36.0

18.0

31.8

34.5

34.5

n=14

n=18

19.52

34.64

17.8

32.7

20.3

36.5

.71

1.00

3.64

2.89

Date Due

QE882.C15 N94 1979

North American Quitcmar) < mis

Harvard MCZ Library \ll>uri

3 2044 062 338 033