DETECTION OF HYBRIDIZATION EVENTS BETWEEN THE COYOTE, CANIS LATRANS, AND THE DOMESTIC DOG, CANIS FAMIL/ARJS, USING TWO POLYMORPHIC MICROSA TELLITE LOCI AND CRANIAL MORPHOMETRIC ANALYSIS
A Thesis
Presented to
the Faculty of the College of Science and Technology
Morehead State University
In Partial Fulfillment
of the Requirements for the Degree
Master of Science in Biology
by
John Jobe Cox
August I, I 997
CAMDEN CARROLL Lllik11tl'Y iv1Uhcrlb-1u, i\i 'IU,XJ 1 Arp-k''j Thes1'ei 5qq, 77:1.5 C 'B' 1'1,d..
Accepted by the Faculty of the College of Science and Technology, in partial fulfillment of the requirements for the Master of Science Degree.
ii DETECTION OF HYBRIDIZATION EVENTS BETWEEN THE COYOTE, CANIS LATRANS, AND THE DOMESTIC DOG, CANIS FAMILIARIS, USING TWO POLYMORPHIC MICROSATELLITE LOCI AND CRANIAL MORPHOMETRIC ANALYSIS
John Jobe Cox, M.S. Morehead State University, 1_997
Director of Thesis: --~---.::,,..~______Cranial morphometric and genetic DNA microsatellite analyses were utilized to determine the taxonomic status of the coyote in Kentucky, and to detect any potential hybridization events between coyotes, Canis /atrans, and domestic dogs,
Canis familiaris. Cranial morphometric analysis involved the employment of 19 linear cranial measurements, previously found to be discriminatory between wild and domestic canids, in a discriminant function analysis utilizing Mahalanobis D2 values.
One hundred and seventy-four canid skulls from the United States and Canada were analyzed and subsequently used to classify 65 unknown canids from Kentucky.
Discriminant function analysis indicated hybridization between coyotes and domestic dogs to be 7-11 %. However, only one of 28 (3 .5%) wild samples indicated hybridization, thus reflecting a possible overestimation of hybridization that may incurred by a potential bias of hybrid sample retention found in institutional collections.
Tissue samples from 55 Kentucky canids (31 coyotes and 24 domestic dogs) were obtained and DNA samples were isolated from canid tissues and amplified using the polymerase chain reaction. Genetic analysis involved the
iii examination of two microsatellite loci (263 and 377), previously determined to be polymorphic. Alleles were subsequently analyzed using polyacrylamide gel electrophoresis. Resultant genetic data indicate a high degree of polymorphism and interspecific overlap of alleles between the two canid species at locus 263, thus indicating a lack of utility of this locus for hybridization studies. Analysis of locus 377
revealed distinctive alleles occurring at high frequencies that show species
specificity, therefore, indicating this locus' potential utility for hybridization
assessment. At locus 377, four coyote-like canids shared allele L with domestic
dogs, however, hybridization was not confirmed by morphological data. Therefore,
based on morphological and genetic data, the Kentucky canids analyzed in this
study are best described as Canis /atrans, the coyote.
Accepted by:
iv ACKNOWLEDGMENTS
I wish to thank the following people and organizations for their respective contribution to my thesis project, of whom without, this project would not have been possible.
Faculty Committee Members:
Dr. Craig Tuerk: methodological guidance, advice and support, finances.
Dr. David Magrane: advice and support, finances.
Dr. Geoffrey Gearner: advice and support.
Dr. Les Meade: advice. support and the western Kentucky trip.
Faculty (Noncommittee Members):
Dr. Scott Rundell: dog tissue specimens.
Dr. Francis Osborne: statistics advice.
Dr. Brian Reeder: advice and support, olfactory tolerance.
Dr. Joe. Winstead: finances, advice and support.
Kentucky Fish and Wildlife Personnel:
David Yancy, Barth Johnson, Jim Delk, Jason Plaxico, Herbert Booth, and
Randy Joseph: coyote carcasses.
David Yancy, Mike Gilliam, Mark Cramer, Steve Bonney: literature.
David Yancy, Mike Gilliam, Mark Cramer, Steve Bonney, Barth Johnson, Jim
Delk, Jason Plaxico, Herbert Booth, Randy Joseph, Jeff Preston,
V ACKNOWLEDGMENTS (cont.)
Arnold Carter, Randy Hedges, Kenny Skaggs, Dale McKenzie, Bill Jolly, Dale
Gee, Paul Teague and all other KDFWR personnel who distributed
information concerning this project.
Taxidermists:
Nancy Logan, Dean Harney, Cobb and Son Taxidermy, Don Lee, Jerry May,
Tony Kamer: coyote carcasses.
Organizations:
Rowan County Animal Shelter: dog carcasses.
Museums. Universities. and Associated Personnel:
Canid skulls were examined from the following:
Christine Peterson, California Academy of Sciences.
Ted Daeschler, The Academy of Natural Sciences, Philadelphia.
Barbara Stein, University of California, Berkeley, Museum of Vertebrate
Zoology.
Steven D. Sroka, University of Illinois at Urbana-Champaign, Museum of
Natural History.
C.M. Dardia, Cornell University Mammalogy Collection.
Dave Snyder, Austin Peay State University.
Bob Martin and Leon Debenes-Gray, Murray State University.
Charles Elliot, Eastern Kentucky University.
vi ACKNOWLEDGMENTS (cont.)
David Westneat and Jim Krupa. University of Kentucky.
Les Meade, Morehead State University.
I would especially like to thank Robert Fischer of the National
Museum of Natural History for his generous hospitality during our scientific
visit to examine the Smithsonian's canid collection.
Graduate Students:
Les Meade and Mehrunissa Lambat for assistance with skull measurements
at the Smithsonian and at Murray State, and for advice and support.
Mike Fultz, Mike Spencer, Todd Jacobs, Ryan McCall, Chris Goshorn,
Teresa Caudill, Kevin Goetz, and Marlin Lawson for advice and
support.
Family:
For patience, financial assistance, olfactory tolerance, and advice and
support throughout the duration of my project.
vii TABLE OF CONTENTS
Title Page ...... Acceptance Page ...... ii Abstract ...... iii Acknowledgments ...... v Table of Contents ...... viii List of Tables ...... ix List of Figures ...... x
CHAPTER PAGE
I. Introduction ...... 1 The Eastern Coyote Phenotype ...... 2 Coyote Hybridization ...... 4 Methodological Approaches to Taxonomic Determination and Hybridization Analysis ...... 6 The Kentucky Coyote ...... 10 Research Objectives ...... 11 11. Materials and Methods ...... 12 Acquisition of Can id Tissues ...... 12 General Morphological Analysis ...... 12 Discriminant Function Analysis of Canid Skulls ...... 15 Isolation of Canid DNA ...... 19 PCR Reactions and Product Analysis ...... 22 111. Results ...... 26 Cranial Morphometric Analysis ...... 26 Microsatellite DNA Analysis ...... 33 IV. Discussion ...... 39 Cranial Morphometric Analysis ...... 39 Microsatellite DNA Analysis ...... 41 V. Conclusions ...... 43 Literature Cited ...... 45
viii LIST OF TABLES
TABLE PAGE
1. Sample localities of coyotes in Kentucky ...... 13
2. Sampling localities of canid skulls in the United States and Canada ...... 17
3. Canid DNA microsatellite oligonucleotide primers and associated dinucleotide repeat insert...... 22
4. Expected product size and polymorphism index of canid DNA microsatellite loci...... 23
5. Standardized canonical discriminant function coefficients using C. tatrans, C. familiaris, C. familiaris x C. tatrans, and C. /atrans x C. rufus as knowns and using Kentucky canids as unknowns...... 27
6. Discriminant classification of unknown Kentucky canids using known C. tatrans, C. familiaris, and C. ta/rans x C. fami/iaris. 29
7. Discriminant classification of unknown Kentucky canids using known C. Jatrans, C. fami/iaris, C. /atrans x C. familiaris and C. ta trans x C. rufus...... 32
8. Individual alleles of can id microsatellite DNA at locus 263...... 36
9. Allele frequencies of canid microsatellite DNA at locus 263 and 377 ...... 37
10. Individual alleles of canid microsatellite DNA at locus 377...... 38
ix LIST OF FIGURES
FIGURE PAGE
1. Presettlement geographical distribution and range expansion, of the coyote, Canis /atrans, in North America...... 1
2. Description of cranial characters...... 16
3. Map of canid skull sample localities in the United States and Canada...... 18
4. Map of coyote-like canid skull sampling localities in Kentucky. 20
5. Map of coyote-like canid DNA sampling localities in Kentucky. 21
6. A discriminant function plot comparing C. latrans, C. familiaris, and C. latrans x C. familiaris as knowns, and Kentucky canids as unknowns...... 28
7. A discriminant function plot comparing C. latrans, C. familiaris, C. /atrans x C. familiaris, and C. latrans x C. rufus as knowns, and Kentucky canids as unknowns...... 31
8. Denaturing, polyacrylamide gel electrophoresis of canid microsatellite DNA at locus 263, as revealed by silver staining. ... 34
9. Denaturing, polyacrylamide gel electrophoresis of canid microsatellite DNA at locus 377, as revealed by silver staining. ... 35
X CHAPTER 1
INTRODUCTION
The coyote, Canis latrans, is an 8-20 kg medium-sized canid of the order
Camivora whose presettlement geographical distribution included the central plains of the United States and southwestern Canada, as well as the arid regions of the
southcentral United States and northcentral Mexico (Young and Jackson, 1951).
Within the last 150 years, the coyote has rapidly expanded its range across North
America to include all 48 contiguous states, Alaska, subarctic Canada, Mexico, and
central Ame~ca into Panama (Figure 1; Parker, 1995). This expansion has
a
Figure 1. Presettlement geographical distribution (a) and range expansion (circa 1900) (b) of the coyote, Canis /atrans, in North America (Parker, 1995). 2
occurred despite repeated attempts by governmental and local coyote eradication programs to control or eliminate coyote populations and has been attributed to a number of factors, including the extirpation of interspecific competitors such as the gray wolf (Canis lupus) and the red wolf (Canis rufus), increased food availability resulting from deforestation and urbanization, flexibility in diet, and a high fecundity, dispersal ability, and sociality (Young and Jackson, 1951; Fox, 1975; Gipson,
1978). It has also been suggested that local releases may have facilitated the coyote's establishment in the east (Schultz, 1955; Galley, 1962; Barbour and Davis,
1974; Fischer, 1977; Gipson, 1978; Parker, 1995). The coyote is also the smallest and least morphologically specialized of the three wild canids in North America.
Nowak (1978) has suggested that this lack of specialized morphology has contributed to the coyote's adaptability and survivability.
The Eastern Coyote Phenotype
Associated with the coyote's movement east has been an apparent increase in body size as compared to western conspecifics, especially in southeastern
Canada and New England (Young and Jackson, 1951; Richens and Hugie, 1974;
Lawrence and Bossert, 1975; Nowak, 1978; Schmitz and Kolenosky, 1985; Schmitz
and Lavigne; 1987; Thurber and Peterson, 1991 ). This larger eastern coyote
phenotype is often characterized as having wolf and/or dog-like morphological
traits. Many taxonomists in the northeast have labeled this coyote phenotype as
Canis /atrans variant. In the southeastern U.S., the coyote phenotype shows only a
slight increase in body size, and is often described as having red wolf 3
characteristics. Despite these observations, the southeastern coyote is often classified as an extension of the southcentral coyote subspecies Canis Jatrans frustror. In general, however, there is little published data on the coyote in the southeast, especially east of the Appalachian Mountains.
Three main hypotheses have been proposed to explain the increase in size of eastern coyotes. These include (1) a genotypic selection to compensate for increases in prey size (Schmitz and Kolenosky, 1985; Schmitz and Lavinge, 1987;
Thurber and Peterson, 1991) and prey abundance (Rosenzweig, 1968), (2) a more immediate phenotypic response to prey size and abundance (Geist, 1986), and (3) hybridization with wolves, which includes gray wolves in the northeastern United
States and southeastern Canada (Silver and Silver, 1969; Mengel, 1971; Lawrence and Bossert, 1975; Hilton, 1978; Nowak, 1978; Lehman et al., 1991; Wayne et al.,
1991; Wayne, 1993; Roy et al., 1994), and red wolves in the southeastern United
States (Young and Jackson, 1951; Mccarley, 1962; Paradiso, 1968; Lawrence and
Bossert, 1967; Gipson et al., 1973; Goertz et al., 1975; Freeman, 1976; Gipson,
1976; French and Dusi, 1979; Phillips and Henry, 1992; Wayne, 1993; Wayne and
Gittleman, 1995). The phenotypic and genotypic response hypotheses suggest that eastern coyotes have adapted to increases in prey size and/or abundance. In general, there is a positive correlation between predator mass and prey mass in carnivores (Gittleman, 1985). The risk of injury or death caused by increased prey resistance may represent a selective pressure for increased body size in the eastern coyote (Lariviere and Crete, 1993), a phenomenon observed in other canids (Rausch, 1967; Nelson and Mech, 1985; Mech and Nelson, 1990). Coyotes 4
have been observed actively preying on white-tailed deer, a large prey species which may constitute the bulk of the coyote's winter diet in the eastern United
States (Messier et al., 1986; Parker, 1986; Parker and Maxwell, 1989; Dibello et al.,
1990). The eastern United States also includes a diverse mixture of agricultural and forest habitats as well as high evapotranspiration (Thurber and Peterson,
1991), a factor which increases primary productivity, and thus enhances the population of white-tailed deer and other coyote prey. Eastern coyotes of large size have been found in areas of high evapotranspiration and precipitation (Kennedy et al., 1986) and are known to occur more frequently around river systems (Silver and
Silver, 1969). However, sev1::ral generations of enhanced nutrition may be required for both a maximum body size to be obtained and a genotypic response to be sustained within a population (Geist, 1986). An increase in prey size is known to affect coyote social organization by inducing an increase in pack hunting among individuals (Fox, 1975; Beckoff and Wells, 1980; Bowen, 1980). This may contribute to a more immediate phenotypic increase in size if population densities are sufficient to form packs, but does not account for the phenomenon in areas where coyotes have recently invaded or are sparsely populated.
Coyote Hybridization
The rapid acquisition of wolf and/or dog-like characteristics by the eastern
coyote, and the lack of these modifications in other areas of expansion, has
suggested potential hybridization between coyotes and conspecific canids (Hilton,
1978). Hybridization among canids is well documented in the literature. Canid 5
hybrids of unknown (Bee and Hall, 1951; Cook, 1952; McCar1ey, 1959; Carson,
1962) and known origin (Seton, 1929; lljen, 1941; Dice, 1942; Hall, 1943; Young and Jackson, 1951; Gier, 1957; Silver and Silver, 1969; Kolenosky, 1971; Mengel,
1971) have been described. Coyotes have been reported to hybridize with dogs
(Young and Jackson, 1951; Kennedy and Roberts, 1969; Silver and Silver, 1969;
Mengel, 1971), gray wolves (Young and Jackson, 1951; Kolenosky, 1971), and red wolves (Paradiso, 1968; Gipson et al., 1974).
All three wild canid species in North America share the same diploid number of chromosomes (2N=78) and are thus able to freely hybridize in the absence of behavioral and morphological constraints (Benirschke, 1967; Wayne, 1993).
Wolves are phenotypically larger than coyotes and are often described as behaving
aggressively towards and even preying on the smaller coyote, thus making
hybridization improbable except under intense reproductive pressures (Mech, 1966;
Peterson, 1977; Carbyn, 1982). Kolenosky (1971) described a successful captive
coyote x gray wolf mating that produced offspring similar to eastern coyote
phenotypes. However, traditional isolating mechanisms that would normc.1lly be
present in the wild may have artificially deteriorated in captivity (Hilton, 1978).
Eastern coyotes have been reported to behave more aggressively and exhibit social
dominance more frequently than their western counterparts (Silver and Silver,
1969), but it is unknown whether this behavioral modification could potentially
contribute to more frequent hybridization.
Hybridization between coyotes and dogs appears to be more common in the
wild. Coyote x dog hybrids, or coydogs, are frequently found in areas of recent 6
coyote expansion (Andrews and Boggess, 1978; Weeks et al., 1990) or in areas where high populations of humans and dogs occur (Mahan et al., 1978). Feral dogs are often the same size as coyotes and their diets are similar (Gipson, 1974), suggesting that coyotes, feral dogs, and some free-ranging domestic dogs may occupy similar ecological niches (Gipson, 1978). However, establishment and successful reproduction by coydog populations may be limited due to a number of factors including (1) lack of parental assistance in pup care by coydog males, (2) a continual loss of wild canid behavioral characteristics over successive generations,
(3) the requirement of anthropogenic sources of food such as dumps or agricultural carrion to compensate for a decrease in killing efficiency, and (4) coydogs tend to breed in December, which results in the birth of coydog pups in mid-winter, thus reducing their survivability (Mengel, 1971; Gipson et al., 1973; Gipson, 1978).
However, a study of the reproductive organs of wild canids and hybrids by Gipson
(1975) revealed an overlap ifl the breeding season, thus indicating a potential for hybridization under natural conditions.
Methodological Approaches to Taxonomic Determination and Hybridization
Analysis
Cranial morphometric analysis has been used extensively in canid classification and hybridization detection (Young and Jackson, 1944; Howard, 1949;
Young and Jackson, 1951; Jolicoeur, 1959; McCarley, 1962; Lawrence and
Bossert, 1967, 1969; Gipson et al., 1974; Richens and Hugie, 1974; Goertz et al.,
1975; Elder and Hayden, 1976; Mahan et al., 1978; French and Dusi, 1979; 7
McGinnis, 1979; Schmitz and Kolenosky, 1985; Kennedy et al., 1986; Lydeard et al., 1986; Wayne, 1986; Lydeard and Kennedy, 1988; Wayne et al., 1989; Weeks et al., 1990). Skull size (Young and Jackson, 1944), index of ratio of palatal width to upper molar tooth row length (Howard, 1949), size of canines and spacing of premolars (Young and Jackson, 1951) have all been employed in cranial analyses.
However, groups of organisms may be distinct with respect to several combinations of characters jointly, yet overlap if these characters are considered separately
(Jolicoeur, 1959). Because coyotes are highly variable in phenotype (Young and
Jackson, 1951), canid classifications based on single morphological indices may be limited in their effectiveness.
More recently, discriminant analysi.s of combinations of several skull characteristics has been utilized in ~anid classification and hybridization detection
(Jolicoeur, 1959; Lawrence and Bossert, 1967, 1969; Gipson et al., 1974; Elder and
Hayden, 1976; Mahan et al., 1976; McGinnis, 1979; Schmitz and Kolenosky, 1985;
Kennedy et al., 1986; Lydeard et al., 1986; Wayne, 1986; Lydeard and Kennedy,
1988; Wayne et al., 1989; Weeks et al., 1990). Discriminant function analysis is a ' statistical technique involving the formation of a stepwise linear combination of the
independent variables which serves as a basis for classifying ungrouped cases into
a "known• group. In discriminant analysis, the weighted averages of the
independent variables are estimated to obtain the maximum separation between
groups and to summarize information contained in multiple independent variables
on a single index (Nei et al., 1975). Discriminant analysis is an excellent
multivariate technique for separating phenotypically similar organisms because 8
morphological differences among, rather than within similar groups are maximized and summarized on a few discriminant axes. Also, characters which contribute most to the discrimination are summarized on these same axes (Wayne, 1986).
Previous discriminant analyses have indicated potential hybridization among wild canids. Using discriminant analysis, Lawrence and Bossert (1967) were the first to demonstrate that canid hybrids were intermediate of their respective parental species. This same analysis also provided evidence suggesting that the red wolf was of hybrid origin. In New England and southeastern Canada, discriminant analyses have demonstrated that the eastern coyote phenotype is similar to that of coyote x gray wolf hybrids (Lawrence and Bossert 1967, 1969; Schmitz and
Kolenosky, 1985). It is speculated that New England coyotes may have descended from the northern subspecies Canis /atrans thamnos that migrated eastward and hybridized with the small Algonquin-type gray wolves (Hilton, 1978). In the southeast, discriminant analyses have shown the occurrence of coyote x red wolf hybridization (Gipson et al., 1973; Elder and Hayden, 1976) as well as coyote x dog hybridization (Gipson et al., 1973; Elder and Hayden, 1976; Freeman, 1976;
Lydeard et al., 1986).
Within the last two decades, molecular and biochemical techniques ha·,e been employed in canid classification, hybridization detection, and phylogenetic analysis (Hamilton and Kennedy, 1986; Wayne and O'Brien, 1987; McKenzie,
1988; Wayne et al., 1989; Kennedy et al., 1990; Lehman et al., 1991; Lehman and
Wayne, 1991; Wayne and Jenks, 1991; Wayne et al., 1991; Lehman et al., 1992;
Wayne et al., 1992; Gotelli, 1994; Roy et al., 1994; Wayne and Gittleman, 1995.). 9
Both nuclear and organelle DNA have become frequently utilized for inferring such phylogenetic and evolutionary relationships. In canids, mitochondrial DNA (mtDNA) analysis has shown evidence of hybridization between coyotes and gray wolves in
Minnesota, and southeastern Canada (Lehman et al., 1991). Mitochondrial DNA analysis has also demonstrated a lack of distinct red wolf mtDNA genotypes.
Instead, the red wolf had genotypes identical to either coyote or gray wolf genotypes, thus suggesting a potential hybrid origin of the red wolf, as well as the possibility of an extensive hybrid zone involving coyotes and gray and red wolves in the southeastern United States (Wayne and Jenks, 1991).
However, the mitochondrial genome is maternally inherited without recombination, and thus estimates of gene flow and hybridization may be biased
(Roy et al., 1994). Nuclear DNA analysis utilizing dinucleotide repeat units, or
"microsatellites", has been employed recently to support the red wolf mitochondrial
DNA findings (Wayne and Gittleman, 1995), and analyze both coyote x wolf hybridization in North America (Roy et al., 1994), and Ethiopian wolf x domestic dog hybridization in Ethiopia (Gotelli et al., 1994). Because microsatellite loci evolve through the gain or loss of repeat units rather than sequence substitutions, are highly polymorphic with frequent!y more than a dozen alleles per locus, have high mutation rates, and are widely dispersed in eukaryotic genomes, these loci can be
informative with regard to hybridization (Roy et al., 1994). The Kentucky Coyote
The coyote migrated into Kentucky approximately 25-30 years ago and is currently believed to inhabit every county. The geographic origin of this immigration is unknown, but may have occurred from several different invasionary fronts
(Parker, 1995). Road kill data reveal that coyote population densities are highest in the western and central regions of the state in comparison to the more heavily forested eastern Cumberland Plateau region (Cramer, 1995). This population density gradient may reflect an earlier establishment of the coyote population in the western part of the state, but is probably also a function of the geography, since coyote populations tend to be higher in areas with a lower density of foliage where they can more readily detect prey (Parker, 1995).
Reports of coyote hybrids within the state of Kentucky are common.
Historically, both the gray and red wolfs range extended into Kentucky (Young and
Goldman, 1944). However, both species have been extinct in Kentucky since early settlement (Young and Goldman, 1944; Hall and Kelson, 1959). Thus, if hybridization does exist, it is likely to originate from coyote x dog matings. Hybrids have been detected in states bordering Kentucky, but are described as infrequent, with numbers usually less than 5% (McGinnis, 1979; Kennedy et al., 1986; Lydeard et al., 1986; Weeks et al., 1990). Total coyote highway mortalities of 25 in 1978 and nearly 400 in 1995 indicate that the coyote population in Kentucky is increasing
(Cramer, 1995). This population increase, combined with the limited and largely unpublished knowledge base about the coyote in Kentucky, demonstrates a need for a better comprehension of the taxonomic status of the coyote in Kentucky. II
Research Objectives
In this study, both cranial morphometric and microsatellite DNA analysis will be employed in an attempt to detect potential hybridization events between the coyote, Canis /a trans, and the domestic dog, Canis familiaris within the state of
Kentucky. In addition, the effectiveness of each of the aforementioned techniques at hybridization detection will be subjectively evaluated based on their utility in this study.
It is the author's desire that this study will significantly contribute to a better understanding of the taxonomic,status of the coyote in Kentucky, as well as providing methodological insight and guidance for future hybridization assessment involving canids. 12
CHAPTER2
MATERIALS AND METHODS
Acquisition of Canid Tissues
Thirty-one coyote-like carcasses and/or tissues were obtained from
Kentucky Department of Fish and Wildlife Resources personnel, taxidermists, hunters, and road kills (Table 1 ). Ten domestic dog carcasses were obtained from the Rowan County Canine Shelter following euthanasia. Fourteen domestic dog tissue samples were obtained from the Morehead State Veterinary Technology
Clinic following surgical castration or ovariectomy of dogs.
General Morphological Analysis
Coyote-like canid and domestic dog carcasses were returned to Morehead
State University. The mass, total length, tail length, ear length, hindfoot length, and the pelage characteristics were recorded for unskinned coyote-like canids. Sex was recorded for both skinned and unskinned coyotes, as well as dogs. All canids were decapitated using a hacksaw. A 5-10 gram sample of muscle tissue was then removed from the neck area of each canid and frozen for future DNA analysis.
Skulls were cleaned by dermestid beetles (Carolina Biological Supply, Burlington,
North Carolina) for 2-3 weeks and were then boiled for 30-60 minutes to remove
any remaining flesh. 13
Table 1. Sample localities of coyotes in Kentucky.
County Age Sex Source Skull DNA
Ballard A F Murray yes no Ballard A M Murray yes no Ballard A M Murray yes no Bath A F carcass yes yes Boyd A carcass yes yes Calloway A F Murray yes no Carlisle A M Murray yes no Clark A F carcass yes yes Clinton A F carcass yes yes east KY J EKU yes no east KY (4) A MSU yes no Fayette roadkill no yes Fleming A UK yes no Fleming A F carcass yes yes Franklin roadkill no yes Franklin A carcass yes yes Fulton A F Murray yes no Grant A carcass yes yes Grant A carcass yes yes Gran: A carcass yes yes Grant carcass no yes Graves A F Murray yes no Harlan J F carcass yes yes Harrison A M Murray yes no Henderson A M Murray yes no Jessamine A M carcass yes yes Kenton A MSU yes no Kenton J M NMNH yes no Lewis A carcass yes yes Lincoln J F EKU yes no Livingstor. A M Murray yes no Livingston A M NMNH yes no Madison J NMNH yes no Madison A EKU yes no Marion A carcass yes yes Marion A M EKU yes no Mason A carcass yes yes McClean A M Murray yes no McCracken A F Murray yes no Menifee A carcass yes no Table 1. (continued).
County Age Sex Source Skull DNA
Muhlenburg A F Murray yes no Muhlenburg A M Murray yes no Nicholas A F carcass yes yes Owen A M carcass yes yes Owen A M carcass yes yes Owen A M NMNH yes no Pendleton A carcass yes yes Rowan A F carcass yes yes Rowan A F carcass yes yes Rowan A M carcass yes yes Rowan A M carcass yes yes Scott roadkill no yes Scott A F carcass yes yes Shelby A carcass yes yes Shelby A EKU yes no Shelby J EKU yes no Trigg A Murray yes no Washington A M NMNH yes no Wayne A carcass yes yes west KY (4) A Murray yes no Woodford A carcass yes yes Woodford A M carcass yes yes Woodford A F EKU yes no
(J) Juvenile; (A) Adult; (M) Male; (F) Female; (MSU) Morehead State University collection (Murray) Murray State University collection (EKU) Eastern Kentucky University collection (UK) University of Kentucky collection (NMNH) National Museum of Natural History collection 15
Discriminant Function Analysis of Canid Skulls
In order to detem,ine the taxonomic status of the coyote in Kentucky, and
detect any potential hybridization events between coyotes and domestic dogs, 19
linear cranial measurements (Figure 2), previously identified as being useful for
discriminating between wild and domestic canids (Kennedy et al., 1986; Wayne,
1986), were measured on 239 skulls from the United States and Canada (Table 2).
Canid sampling locations were selected to include a wide geographic separation of
coyote subspecies (Figure 3), and thus establish a size gradient of skull sizes in
order to compensate for exceptionally large or small coyotes sampled in Kentucky.
All measurements were recorded to the nearest 0.1 mm using vernier calipers.
Adult coyotes were distinguished from juveniles using the criteria of Nellis et al.,
(1978). Of the 239 canid skulls, 174 (122 coyotes, 22 coydogs, 21 dogs, and 9
coyote-red wolf hybrids), previously identified by scientists and/or museum
personnel, were employed in a linear stepwise discriminant function analysis, a
multivariate statistical method which was utilized to maximize the distance between
known canid groups, as well as to assess which cranial characteristics were most
important in canid group separation. Statistical analysis was performed using the
SPSSx subprogram DISCRIMINANT. Stepwise analysis was based on maximizing
the Mahalanobis distance between the two closest groups. Mahalanobis' distance,
d'-, is a generalized measure of the distance between two groups. The distance
2 between groups a and bis defined as D1Jb = (n - g) f ~ W11 * (X,.- x,b) ('>?1a- xjb) . i•I ~-I 16
TSL 7 7 -'>t;l"""- M L, M W 1 \r.~.l..,:::...;..;J ... M\, M W ... JIL ,..p\
Figure 2. Description of cranial characters.· TSL, total skull length; RW, rostrum width; PW, palatal width; ZH, zygomatic height; MH, mandible height; CD, cranial depth; PD, palatal depth; ZW, zygomatic width; MCW, maximum cranial width; LCW, least cranial width; WP', rostrum width at first premolar; M,L, length of first lower molar; MWP 4, mandible width at fourth lower premolar; P'L, length of third upper premolar; P4L, length of fourth upper premolar; M'L, length of first upper molar; M'W, width of first upper molar; M2L, length of second upper molar; M2W, width of second upper molar (modified from Wayne, 1986). Table 2. Sampling localities of canid skulls in the United States and Canada.
Species N Subspecies N LOC3les N
Canis /a/rans 122 thamnos 60 IL 10 IN 9 Ml 5 MO 6 NY 29 Ont 1 frustror 17 AL 5 AR 6 MO 6 incolatus 5 AK 1 BC 4 texensis 5 TX 5 /a/rans 5 NB 5 meamsi 5 AZ 5 /estes 8 CA 3 WY 5 ochropus 10 CA 10 Canis familiaris 21 KY 21 C. /a/rans x C. familiaris 22 AL 1 IL 1 IN 2 KY 1 . ME 5 MO 2 ND 1 NY 1 OK 2 SD 1 VA 1 WY 4 C. /a/rans x C. rufus 9 LA 5 MS 1 TX 3 Unknown Canids 65 KY 65
Total Canids 239 Figure 3. Map of canid skull sample localities in the United States and Canada. Single or multiple samples exist at several localities on the map (see Table 2). Symbols include: filled circles (•), Canis /atrans; open circles (o), Canis latrans x canis familiaris; and open squares (□), Canis /atrans x Canis rufus. 19
where p is the number of variables in the model, X1• is the mean for the ith variable in the group a, and Wq • is an element from the inverse of the within-groups covariance matrix (Nei et al., 1975). Discriminant functions were used to classify
65 unknown coyote-like canids (37 museum specimens and 28 carcasses) fror:n
Kentucky (Figure 4). All sexes were pooled due to a lack of information on sex for many of the canids measured. Analysis was performed with and without the coyote x red wolf hybrid data, due to the probability of potential bias incurred by the small sample size of this group (n=9).
Isolation of Canid DNA
DNA was extracted from thirty-one coyote-like canid (Figure 5) and twenty four domestic dog tissue samples from Kentucky, and subsequently isolated using the following protocol. Approximately 100mg of muscle or hair/hide tissue from each canid tissue sample was removed and placed in 1.2ml of digestion buffer
(100mM NaCl, 10mM Tris, 25mM EDTA, 0.5% SOS pH 8, 0.1% proteinase K,
50mM dithiothreitol) and incubated at 37°C for 3-5 days. An equal volume of phenol was then added to each digest, the solution inverted and mixed for five minutes, then centrifuged at 1 O,OOOg for 2 minutes to obtain phase separation. The aqueous phase from each tube was then decanted and placed into clean 1.5ml microcentrifuge tubes. Tubes with low volumes of aqueous extract were back extracted with an equal volume of phenol and recentrifuged as previously described. An equal volume of ether was then added to each tube, the tube inverted and mixed, and the ether discarded by pipetting to remove the remaining Figure 4. Map of coyote-like canid skull sampling localities in Kentucky. Filled circles(•) indicate approximate sampling locale.
"'0 Figure 5. Map of coyote-like canid DNA sampling localities in Kentucky. Filled circles (•) indicate approximate sampling locale. 22
aqueous solution. To each tube, a 1/10th volume of 3M sodium acetate and one half volume of isopropanol were added. The tubes were inverted, mixed and placed into a 10°c refrigerator for 24 hours to allow DNA precipitation. Samples were then microcentrifuged at 13,000g for 15 minutes, and the supernatant decanted. DNA pellets were then resuspended in 250µL of 70% ethanol and centrifuged at 13,000g for 10 minutes. The ethanol was then decanted and the pellet allowed to air dry,
before final resuspension in 75-1 00µL of distilled water for 24 hours. All DNA
samples were subsequently frozen at 4°C for storage purposes. DNA
concentrations for each sample were calculated by mixing 5µL of sample with 5µL of
Sµg/mL ethidium bromide and placing these drops on cellophane wrap stretched
over a Haake Buchler UVT transilliuminator. A similar ethidium-stained, calf thymus
DNA standard dilution series was placed next to the samples so that sample DNA
concentrations could be estimated visually.
PCR Reactions and Product Analysis
DNA amplification was performed using the polymerase chain reaction (PCR)
according to the following procedure (Saiki et al., 1988). Two master mixes ("hot"
and "cold") for each PCR reaction were created. The hot mixture contained Taq
DNA polymerase (Promega Corporation, Madison, Wisconsin), distilled water, and
5X reaction buffer (Promega). The "cold" mixture contained a set of oligonucleotide
primers obtained from Oligos Etc. Inc. (Wilsonville, Oregon) previously used to
characterize canid microsatellite dinucleotide repeats (Tables 3 and 4), MgCI, and Table 3. Canid DNA microsatellite oligonucleotide primers and associated dinucleotide repeat insert. (Ostrander et al., 1993)
Locus# Forward Primer Sequence Microsatellite Insert Reverse Primer Sequence
263- TCAATCATAGGTCGATTGTTG (TAho (CA)s N22 (AC),s AGTCACCCTCCAAGATGTG
377- ACGTGTTGATGTACATTCCTG (AC),2 CCACCCAGTCACACAATCAG
Table 4. Expected product size and polymorphism index of canid DNA microsatellite loci (Ostrander et al., 1993).
Locus# Polymorphism (PIG) Index Expected Product Size
263 0.81 50-150 bp
377 ? 150 - 250 bp 24
dNTPs (Ostrander, 1993). The hot mixture was created, and then dispersed in aliquots into 0.5ml microcentrifuge tubes each containing a separate canid DNA sample of approximately 200ng. 50µL of sterile mineral oil was added on top of the reaction mixture and the tubes placed in a MJ Research minicycler and heated to
95°C for 3 minutes, then rapidly cooled to 4°C until addition of the "cold" mixture.
The "cold" mixture was created on ice and aliquots added to each of the reaction tubes within the minicycler to yield a final reaction mixture concentration (0.8 U/µL
Taq, 1.25X of 1OX reaction buffer, 20ng/µL genomic DNA, 2mM Mg Cl, 5mM dNTPs,
12pmol/µL primers). The total volume for each reaction was 10µL. A pGEM®-
3Zf( +) control DNA and a pUC/M 13 forward primer using the same concentration of the aforementioned reaction mixture components in a 10µI total reaction volume was used as a positive control. All samples were immediately amplified in the minicycler for 30-40 cycles using the following program: denaturation at 93°C (60 sec), followed by primer annealing at 45-50°C (30 sec), and extension at 70°C (30 sec). A final extension step at 70°C (5 min) was performed after the final cycle. After PCR, reaction tubes were immediately placed into storage at 4 °C.
To determine PCR success, 5µL of bromophenol blue loading dye was mixed with 5µL of PCR reaction product and loaded onto a 6% nondenaturing
polyacrylamide gel and electrophoresed at 110-120V for 2-3 hours. An Msp 1 DNA
digest (New England Biolabs, Beverly, Massachusetts), heated to 95°C for 3
minutes, was loaded alongside the samples and used as an absolute DNA size
standard. Gels were stained for 10-15 minutes using 20% ethidium bromide and 25
subsequently viewed using a Haake Buchler UVT transilluminator. 3µL of the successful PCR reactions were then mixed with 3µL of bromophenol blue loading dye and heated to 95°C for 3 minutes before loading onto a 0.4mm 6% denaturing polyacrylamide gel. An Msp 1 DNA digest, heated to 95°C for 3 minutes, was again loaded alongside the samples to serve as an absolute DNA size standard. Samples were electrophoresed at 1200V for 3-4 hours. Samples were silver stained for visualization according to the Promega Silver Sequencing Kit protocol. Gels were allowed to dry overnight, then photoscanned and recorded on computer for allele comparison. Alleles were scored in relation to flanking and internal Msp 1 standards run simultaneously alongside the canid samples on the gel. 26
CHAPTER 111
RESULTS
Cranial Morphometric Analysis
Using combined sexes for coyotes, dogs, and coyote x dog hybrids as knowns and Kentucky canids as unknowns in a discriminant function analysis, 14 cranial characters were found to be useful discriminators (Test 1 of Table 5).
3 Characters P L, RW, and PW, had the highest discriminant functions for canonical discriminant function 1. The standardized discriminant canonical functions accounted for 100% of the variability. Functions 1 and 2 had high canonical correlations (r = 0.89 and 0.64, respectively) and were highly significant (P < 0.001).
A plot of the discriminant scores for each unknown in relation to known canid groups is given in Figure 6. These results indicate that among known canid groups, there is a distinct separation with little overlap when considering these 14 cranial characters in combination. Of all individuals in the three known groups,
93.4% were correctly classified (Table 6). One-hundred twenty-two coyotes
(100%), 20 domestic dogs (95.2%), and 13 coydogs (59.1%) were correctly classified (one known dog specimen was classified as a coyote x dog hybrid, and nine coydog specimens were classified as coyotes). Of the 65 unknown canids, 59
(86.8%) were classified as coyotes, 5 were classified as coydogs (7.7%) and_ 1 was 27
Table 5. Standardized canonical discriminant function coefficients using C. /atrans, C. familiaris, C. /atrans x C. familiaris, and C. /atrans x C. rufus as knowns and using Kentucky canids as unknowns.
Test 1• Test 2•
Cranial Cranial Character Function 1 Function 2 Character Function 1 Function 2
P3L -.48909 .24946 PW .50776 -.03976 RW .46079 .04477 RW .47492 .02727 PW .43986 -.17484 PJL -.46716 .25403 ZH ·-.33690 -.01790 ZH -.33074 .03727 2 M L .30553 .23234 MCW -.31795 .22567 p•L -.28154 · -.70666 M2L .28664 -.68118 M2W -.24338 -.12902 CD .28448 -.09396 MCW -.243Q7 1.28045 p•L -.28393 -.35493 CD .23697 -.97737 M2W -.25114 -.99912 WP1 .22400 .42442 WP' .23703 .53096 MH .21455 .10354 MH .21515 .13164 PD .20827 .07965 MWP4 .10683 -.09131 LCW .07599 -.52378 M1L -.08694 .40923 M1L -.05502 .44274 LCW .08175 -.56749
• Test 1 C. latrans, C. familiaris, C. /atrans x C. familiaris as knowns. • Test 2 C. /atrans, C. familiaris, C. /atrans x C. familiaris and C. /atrans x C. rufus as knowns. 4.0
• • • N 2.0 /a/rans x famt11ans C ·.c0 u § µ;., -a • • • I :~ 0 u • ••• ••• i:S"' • ... •••••• • .! ••• ·cu • • • a0 Canis /a/rans .. u ·2.0 •
-4.0 -2.0 0 2.0 4.0 Canonical Discriminant Function 1
Figure 6. A discriminant function plot comparing C. /atrans, C. familiaris, and C. /atrans x C. familiaris as knowns, and Kentucky canids as unknowns. Box contours represent the extreme ranges within each known population. Filled circles (•) represent individual unknowns. 29
Table 6. Discriminant classification using C. latrans, C. familiaris, and C. latrans x C. familiaris as knowns and Kentucky coyote-like canids as unknowns.
Test 1 n Predicted Group Membership Group coyotes dogs coydogs
coyotes 122 122 (100%) 0 0 dogs 21 0 20 (95.2%) 1 (4.8%) coydogs 22 9 (40.9%) 0 13 (59.1%) total unknowns 65 59 (86.8%) 1 (1.5%) 5 (7.7%) (institutional + carcasses) unknowns (institutional) 37 32 (86.5%) 1 (2.7%) 4 (89.2%) unknowns (carcasses) 28 27 (96.5%) 0 1 (3.5%) unknowns 27 27 (100%) 0 0 (carcasses w/ skull +DNA)
known groups correctly classified 93.4% 30
classified as a dog (1.5%). Of the 28 unknown canids collected in the wild from which skulls samples were available, 27 were classified as coyotes and one (3.5%) was classified ,is a coydog.
Using combined sexes for coyotes, dogs, coyote x dog hybrids, and coyote x red wolf hybrids as knowns and Kentucky canids as unknowns in a discriminant function analysis, 14 cranial characters were found to be useful discriminators (Test
2 of Table 5). Characters P3L, RW, and PW had the highest discriminant functions for canonical discriminant function 1. The standardized discriminant canonical functions accounted for 100% of the variability. Functions 1 and 2 had high canonical correlations (r = 0.89 and 0.64, respectively) and were highly significant • (P < 0.001). Function 3 had a low canonical correlation (r = 0.39) and was not significant (P > 0.05).
A plot of the discriminant scores for each unknown in relation to known canid groups is given in Figure 7. These results indicate that between known canid groups, there is a distinct separation with little overlap when considering these 14
cranial characters in combination. Of all individuals in the three known groups,
S0.3% were correctly classified (Table 7). One-hundred twenty-one coyotes
(99.2%), 21 domestic dogs (100%), 13 coydogs (59.1%), and three coyote x red
wolf hybrids (33.3%) were correctly classified (one known coyote specimen was
classified as a coydog, seven known coydogs were classified as coyotes and two 31
4.0 •
• • 2.0 •• "' C • • .v2 C ti! ~ ~ .§C ••• ·c: 0 • • • ••• Canis/amllians u • • ••••••••• l5"' .; .. . u • ·2 • • • I a I • ~ u I -2.0 • I •
-4.0
-4.0 -2.0 0 2.0 4.0 Canonical Discriminant Function l
Figure 7. A discriminant function plot comparing C. /atrans, C. familiaris, C. la trans x C. familiaris, and C. /atrans x C. rufus :,is knowns, and Kentucky canids as unknowns. Box contours represent the extreme ranges within each population. Filled circles ( •) represent individual unknowns. 32
Table 7. Discriminant classification using C. /atrans, C. familiaris, C. /atrans x C. familiaris and C. latrans x C. rufus as knowns and Kentucky coyote-like canids as unknowns.
Test2
Predicted Group Membership Group n coyotes dogs coydogs latrans x rufus coyotes 122 121 (99.2%) 0 1 (0.8%) 0 dogs 21 0 21 (100%) 0 0 coydogs 22 7 (31.8%) 2 (9.1%) 13 (59.1%) 0 rufus x latrans 9 6 (66.7%) 0 0 3 (33.3%) total unknowns 65 57 (87.7%) 1 (1.5%) 7 (10.8%) 0 (institutional + carcasses) unknowns (institutional) 37 30(81.1%) 1 (2.7%) 5 (13.5%) 0 unknowns (carcasses) 28 27 0 1 (3.5%) 0 unknowns 27 27 (100%) 0 0 (carcasses w/ skull +DNA) known. groups correctly classified 90.3% 33
were classified as dogs, and six known coyote x red wolf hybrids were classified as coyotes). Of the 65 unknown canids, 57 (87. 7%) were classified as coyotes, 7 were classified as coydogs (10.8%) and 1 was classified as a dog (1.5%). Of the 28 unknown canids collected in the wild from which skull samples were available, 27 were classified as coyotes and one (3.5%), the same individual as in test 1, was classified as a coydog.
Microsatellite DNA Analysis
Figures 8-9 demonstrates the high degree of polymorphism associated with each of the two microsatellite loci surveyed. At locus 263, a total of thirteen canid alleles were identified, which included nine dog alleles and thirteen coyote alleles
(Table 8). Four of the coyote alleles at this locus were species-specific, while all dog alleles were shared by coyotes. In dogs, allele B had the highest frequency
(0.389; Table 9), and in coyotes, allele J was most abundant (0.360). Because nine of the total alleles at this locus were interspecifically shared (A-1), locus 263 was eliminated as a potential genetic marker for hybridization in this study.
At locus 377, a total of fourteen canid alleles were identified, which included seven dog alleles and eight coyote alleles (Table 10). In dogs, alleles A and G were most abundant (0.913 and 0.720, respectively; Table 9), and in coyotes, alleles B and D had the highest frequency (0. 7 42 and 0.355, respectively). Only one allele (L) was interspecifically shared, indicating this locus' potential utility as a marker for hybridization detection. However, discriminant analysis classified these unknown canids as coyotes. Figure 8. Denaturing, polyacrylamide gel electrophoresis of canid microsatellite DNA at locus 263, as revealed by silver staining. Alphanumeric designations refer to canid species, (c) Canis latrans (d) Canis familiaris, and its associated sample number. An Msp 1 DNA digest was used to compare allele migration distances and approximate allele size. 35
.t{t:f,:.f. ?~'::~~\t~ \B» ,~,... ~\~&•:<~-:.❖, ~~ '.\.~ , ~ ',· '$.'$';:.~~-:!:":..JR::t::~,ill~ ',<¥~ t,,\.,<,.. ::...'\:c:Y~, t,QJ <'
· -
~ }{ ' ... ,.. ~. ~~fa:. . . a
s:?::;w,. lj.,
b
Figure 9. Denaturing, polyacrylamide gel electrophoresis of canid microsatellite DNA at locus 377, as revealed by silver staining (a, b, c). Alphanumeric designations refer to canid species, (c) Canis latrans (d) Canis familiaris, and its associated sample number. An Msp 1 DNA digest was used to compare allele migration distances and approximate allele size. 36
Table 8. Individual alleles of canid microsatellite DNA at locus 263.
Coyotes Dogs Sample Alleles Sample Alleles
C1 B,C D1 B,E C2 A,B D2 H,I C3 C,H D7 G,I C4 C,E D8 F,H C5 1,J D9 E,F C6 E,G D12 F,H C9 B,E D13 H,I C10 B,J D14 F,H C11 E,M D15 A,C C12 E,J D16 A,E C13 J,M D17 B,D C14 B,E D18 A,B C15 J,K D19 A,B C16 B,D D20 B,C C17 E,F D21 A,D C18 C,D D22 B,D C19 1,J D23 B,E C20 H,I D24 G,H C21 C,D C22 C,E C23 1,J C24 J,K C25 K,M C26 L,M C27 1,J 37
Table 9. Allele frequencies of canid microsatellite DNA at locus 263 and 377.
Locus 263 Locus 377 Dogs Coyotes Dogs Coyotes (N=18) (N=25) (N=23) (N=31) Allele Allele
A** .278 .040 .A .913 .000 s·· .389 .240 B .000 .742 c·· .111 .120 C .000 .290 D- .167 .320 D .000 .355 E** .222 .040 E .000 .097 F** .222 .040 F .000 .129 G** .167 .080 G .720 .000 H** .278 .080 H .087 .000 1- .167 .200 I .043 .000 J .000 .360 J .000 .290 K .000 .120 K .000 .032 L .000 .040 L** .043 .129 M .000 .160 M .087 .000 N .043 .000
•• denotes interspecific overlap at that allele 38
Table 10; Individual alleles of canid microsatellite DNA at locus 377.
Coyotes Domestic Dogs Sample Alleles Sample Alleles
C1 B,L D1 A,G C2 B,B D2 A,G C3 B,L D3 A,G C4 B,E D4 G,M cs D,D DS A,G C6 D,D D6 1,1 C7 D,F D7 A,G ca D,K DB A,G C9 C,J D9 A,G C10 B,J D10 A,H C11 B,J D11 A,G C12 B,J D12 A,L C13 B,J D13 A,H C14 B,J D15 A,G C15 B,J D16 A,G C16** B,J D17 A,M C17 F,L D18 A,G C18 F,L D19 A,G C19 E,J D20 A,G C20 B,C D21 A,G C21 B,C D22 A,G C22 B,C D23 A,G C23 B,F D24 A,G C24 c,c C25 D,D C26 B,E C27 D,D C28 B,B C29 B,D C30 B,C C31 B,C C32 B,C
"C16 sample locale wa ■ Brown County, OH. 39
CHAPTER4
DISCUSSION
Cranial Morphometric Analysis
Discriminant function analysis revealed that Kentucky wild canids are statistically distinct from known dogs, coydogs, and coyote x red wolf hybrids, but are not statistically distinct from known coyotes. Of the nineteen cranial characters employed in the discriminant analysis, 13 were found to be useful as canid discriminators in both analyses. Both discriminant analyses had a unique additional useful character as well, resulting in a total of 14 discriminating variables for each analysis. Three characters (P'L, RW, and PW) were found to be the best discriminators in both analyses. The discriminating power of palatal width (PW) and
rostral width (RW) reflect tooth size differences between coyotes and dogs. Dogs
and wolves are generally wider across the palate to accommodate larger teeth, while
coyotes often have slightly longer, more narrow teeth (Lawrence and Bossert, 1967),
which may account for the high discriminating power of the third upper premolar
(P'L).
Classification results indicate coyote x dog hybridization to be between 7-
11 %. This finding is similar to coyote x dog hybridization occurrence described in
other states (Gipson et al., 1974, AR; 13%, Freeman, 1976, OK; 13%, Weeks et al.,
1990, OH; 2%). However, of all 28 canids taken in the wild, only one was classified 40
as a coydog, lowering hybridization to 3.5% when institutional samples are excluded. A few of the museum samples examined were labeled as suspected hybrids, which may account for the discrepancy in hybridization percentages between total unknowns (wild and institutional collection unknowns), and wild unKnowns. Also, the preservation of canid specimens by institutions may be biased towards selecting unusual phenotypes or hybrids, and thus may not be representative of the wild population.
Discriminant analysis revealed no evidence of red wolf influence within
Kentucky canids. Other studies in neighboring states, however, have demonstrated
red wolf influence within the wild canid population (Gipson et al., 1974; Elder and
Hayden, 1976; Lydeard et al., 1986).· There have been reports of massive
hybridization between coyotes and red wolves in eastern ~exas, Arkansas and
Louisiana (Paradiso, 1968; Gipson, et al., 1974). However, the fact that the red
wolf may not be a distinct species, but may be instead a coyote x gray wolf hybrid,
questions the validity of these findings (Wayne and Jenks, 1991; Wayne and
Gittleman, 1995). Interestingly, classification results of this study show that only
33% of the known coyote x red wolf hybrids were correctly classified, while the
remaining 66% were "misclassified" as coyotes. All of the coyote x red wolf
samples examined were taken during the red wolf recovery program of the 1970s.
Gipson (1978) reported that no wild canids other than occasional feral dogs
and isolated pockets of red wolves occurred east of the Mississippi River between
1900 and 1965, thus it is unlikely that there is significant red wolf influence within
the wild canid population of Kentucky. Because of the low sample number 41
examined (n=9), and the current confusion over the red wolf's taxonomy, the question of red wolf influence within wild Kentucky canid populations cannot be appropriately addressed at this time.
Microsatellite DNA Analysis
Microsatellite DNA analysis revealed a total of thirteen alleles at locus 263, nine of which were interspecifically shared between coyotes and dogs. Coyotes had 4 alleles at this locus that were species-specific. The distribution of alleles in both species were generally scattered, with allele frequencies never exceeding 40% in either species. Based on these observations, locus 263 was considered too polymorphic to be useful for interspecific hybridization detection in this study. This locus could instead be potentially useful for other genetic studies, such as canine breeding and paternity determination, where greater individual polymorphism at particular loci are desired.
A total of fourteen alleles were detected at locus 377, only one of which was
interspecifically shared between coyotes and dogs (allele L). Coyotes had seven
alleles and dogs had six alleles that were species-specific. The distribution of
alleles in both species occurred at high frequencies at particular alleles, especially
dogs. Almost 96% of dogs possessed allele A, while 72% of dogs had allele G.
Although 72% of coyotes possessed allele B, coyote allele frequencies were more
scattered than dogs at this particular locus. Because of the high frequency and
species specificity of particular markers, locus 377 was utilized for hybridization
detection. All four coyote-like canids with allele L were investigated as potential -12
hybrids. Discriminant function analysis, however, did not confirm these canids as hybrids. In all four cases, these canids were classified as having a greater than
90% probability of being coyotes. While this does not completely eliminate the possibility that these individuals are hybrids, it is more likely that this allele is either common to both species, or that some slight introgression of dog genes has occurred that is not expressed as morphological changes associated with the cranium (Wayne, 1986). Despite these findings, the high frequency of species specific markers make locus 377 an excellent candidate for future canid hybridization studies. Locus 377 has been successfully employed recently in other hybridization studies (Gotelli et al., 1994; Roy et al., 1994).
Both microsatellite loci surveyed revealed a high degree of polymorphism.
Coyotes had greater genetic variability at both loci than dogs. Because dog breeds have been genetically isolated for a long time (Scott, 1968; Olsen and Olsen, 1977;
Davis and Valla, 1978), they may demonstrate a lesser degree of genetic variation than canids such as the coyote and gray wolf, which have higher dispersal rates, and thus, less impedance to gene flow (Wayne, 1993). This high degree of gene flow, however, contributes to low intraspecific variation among widely distributed wild canids (Fischer et al., 1976; Lehman and Wayne, 1991; Roy et al., 1994).
In this study, only 54 canid DNA samples and two microsatellite loci were investigated, thus it is improbable that this genetic survey encompasses all potential alleles and may not be representative of the total genetic diversity of either canid species. 43
CHAPTERS
CONCLUSIONS
The objectives of this study were to employ the techniques of cranial
morphometric and microsatellite DNA analysis to detect hybridization events
between coyotes and dogs in Kentucky, and to assess the utility of each of these
techniques for future hybridization assessment involving canids.
In this study, cranial morphometric analysis revealed a strong statistically
significant separation of known canid groups, with greater than 90% of all known
canids correctly classified. Hybridization between coyotes and domestic dogs in
Kentucky was found to be between 7-11 % when considering both wild and
institutional skulls together, and 3.5% when considering only skulls taken from
carcasses in the wild. This discrepancy may be explained by potential bias in
collection and/or sample retention by institutions. There also appears to be no red
wolf influence within the Kentucky wild canid population, although a greater sample
number may be required to support this conclusion. Based on classification results
. from discriminant function analysis, the wild canids from Kentucky examined in this
analysis are best described as Canis latrans, the coyote.
Genetic analysis of two microsatellite DNA loci revealed a high degree of
polymorphism and genetic variability associated with each locus. Alleles at locus
263 were interspecifically shared at a number of sites and were highly distributed, 44
which eliminated this locus from effective hybridization detection in this study.
Alleles at locus 377 were highly species-specific and occurred at very high frequencies, with only one allele being shared between canid species. Because of these attributes, locus 377 was used to investigate four potential hybrid candidates.
However, discriminant analysis of these four individuals revealed a strong
probability that they are instead coyotes. Despite these results, locus 377 appears
to be a useful marker for hybridization assessment. 45
LITERATURE CITED
Andrews, R.D. and E.K. Boggess. 1978. Ecology of coyotes in Iowa. In M. Bekoff, ed., Coyotes, Biology, Behavior and Management. Academic Press, New York. pp. 191-208.
Barbour R.W. and W.H. Davis. 1974. Mammals of Kentucky. University Press of Kentucky. 322pp.
Beckoff, M. and M.C. Wells. 1980. The social ecology of coyotes. Sci. Amer. 242: 130-142.
Bee, J.W., and E.R. Hall. 1951. An instance of coyote-dog hybridization. Trans. Kansas Acad. Sci. 54: 73-77.
Benirschke, K. 1967. Sterility and fertility of interspecific mammalian hybrids. In K. Benirschke, ed., Comparative Aspects of Reproductive Failure. Springer Verlag, New York. pp. 218-234.
Bowen, W.D. 1980. Variation in coyote social organization: the influence of prey size. Can. J. Zoo/. 59: 639-652.
Carbyn, L.N. 1982. Coyote population fluctuations and spatial distribution in relation to gray wolf territories in Riding Mountain National Park, Manitoba. Can. Field Nat. 96: 176-183.
Carson, H.S. 1962. Coyote, coy-dog, or dog. Maine Fish and Game, Spring 1962: 4-7.
Cook, R. 1952. The coy-dog: hybrid with a future? J. Heredity. 43: 71-73.
Cramer, M. 1995. Furbearer highway mortality survey. Kentucky Department of Fish and Wildlife Resources, Frankfort.
Davis, S., and F.R. Valla. 1978. Evidence for the domestication of the dog 12,000 years ago in the Natufian of Israel. Nature. 276: 608-610.
Dice, LR. 1942. A family of dog-coyote hybrids. J. Mamm. 23: 186-192. 46
Dibello, F.J., S.M. Arthur, and W.B. Krohn. 1990. Food habits of sympatric coyotes, Canis latrans, red foxes, Vulpes vu/pes, and bobcats, Lynx rufus, in Maine. Can. Field. Nat. 104: 403-408.
Elder, W.H., and C. M. Hayden. 1976. Use of discriminant function in taxonomic determination of canids from Missouri. J. Mamm. 58: 17-24.
Fischer, R.M. 1977. A survey on the status of .the coyote (Canis latrans) in Georgia. M.S. Thesis, University of Georgia, Athens. 61 pp.
Fox, M.W. 1975. Evolution of social behavior in canids. In M. Fox, ed., The Wild Canids: Their Systematic Behavioral, Ecology, and Evolution. 'Ian Nostrand Reinhold, New York. pp. 429-460.
Freeman, R.C. 1976. Coyote x Dog Hybridization and Red Wolf Influence in the Wild Canis of Oklahoma. M.S. Thesis, Oklahoma State Univ., Stillwater. 62 pp.
French, T.W., and J.L. Dusi. 1979. Status of the coyote and red wolf in Alabama. J. Alabama Acad. of Sci. 54: 220-224.
Gier, H.T. 1957. Coyotes in Kansas. Bull. Agric. Exp. Sta., Kansas St. College. 393: 1-97.
Gies!, V. 1986. Bergmann's Rule is invalid. Can. J. Zoo/. 65: 1035-1038.
Gipson, P.S., J.A. Sealander, and J.E. Dunn. 1974. The taxonomic status of wild Canis in Arkansas. Syst. Zoo/. 23: 1-11.
Gipson, P.S., and J.A. Sealander. 1974. Changing food habits of wild Canis in Arkansas with emphasis on coyote hybrids and feral dogs. Amer. Midi. Nat. 95: 249-253.
Gipson, P.S., I.K. Gipson, and J.A. Sealander. 1975. Reproductive biology of wild Canis in Arkansas. J. Mamm. 56: 605-612.
Gipson, P.S. 1978. Coyotes and related Canis in the southeastern United States with a comment on Mexican and Central American Canis. In M. Bekoff, ed., Coyotes, Biology, Behavior and Management. Academic Press, New York. pp. 191-208.
Gittleman, J.L. 1985. Carnivore body size: ecological and taxonomical correlates. Oeco/ogia. 67: 540-554. 47
Goertz, J.W., L.V. Fitzgerald, and R.M. Nowak. 1975. The status of wild Canis in Louisiana. Amer. Midi. Nat. 93: 215-218.
Galley, F.B. 1962. Mammals of Georgia. University of Georgia Press, Athens. 281pp.
Gortelli, D., C. Sillero-Zubiri, G.D. Applebaum, M.S. Roy, D.J. Girman, J. Garcia Moreno, E.A. Ostrander, and R.K. Wayne. 1994. Molecular genetics of the most endangered canid: the Ethiopian wolf Canis simensis. Mo/. Ecol. 3: 301-312.
Hall, E.R. 1943. Cranial characters of a dog-coyote hybrid. Amer. Midi. Nat. 29: 371-374.
Hall, E.R. and K.R. Kelson. 1959. The Mammals of North America. 2 vols., Ronald Press, New York. 1083 pp.
Hamilton, M.J., and M.L. Kennedy. 1986. Genie variation in the coyote, Canis latrans, in Tennessee, U.S.A. Genetica. 71: 167-173.
Hilton, H. 1978. Systematics and ecology of the eastern coyote. In M. Bekoff, ed., Coyotes, Biology, Behavior and Management. Academic Press, New York. pp. 209-228.
Howard, W.E. 1949. A means to distinguish skulls of coyotes and domestic dogs. J. Mamm. 30: 169-171. lljen, N.A. 1941. Wolf-dog genetics. J. Genetics. 42: 359-414.
Jolicoeur, P. 1959. Multivariate geographical variation in the wolf Canis lupus. Evolution. 13: 283-299.
Kennedy, M.L., P.L. Leberg, and G.D. Baumgardner. 1986. Morphological variation in the coyote, Canis /atrans, in the southeastern United States. Southwest. Nat. 31: 139-148.
Kennedy, P.K., M. L. Kennedy, P.L. Clarkson., and I. S. Liepins. 1990. Genetic variability in natural populations of the gray wolf, Canis lupus. Can. J. Zoo/. 69: 1183-1188.
Kennedy, J.J., and J.D. Roberts. 1969. Fertility of coyote-dog hybrids. J. Mamm. 50: 830-831.
Kolenosky, G.B. 1971. Hybridization between wolf and coyote. J. Mamm. 52: 446-449. 48
Lariviere, S. and M. Crete. 1993. The size of eastern coyotes (Canis /atrans): a comment. J. Mamm. 74: 1072-1074.
Lawrence, B.L., and W.H. Bossert. 1967. Multiple character analysis of Canis lupus, latrans, and familiaris, with a discussion of the relationships of Canis niger. Amer. Zoo/. 7: 223-232.
Lawrence, B.L., and W.H. Bossert. 1969. The cranial evidence for hybridization in New England Canis. Breviora. 330: 1-12.
Lawrence, B.l., and W.H. Bossert. 1975. Relationships of North American Canis shown by a multiple character analysis of selected populations. In M. Fox, ed., The Wild Canids: Their Systematic Behavioral, Ecology, and Evolution. Van Nostrand Reinhold, New York. pp. 73-86.
Lehman, N., A. Eisenhawer, K. Hansen, L.D. Mech, R.O. Peterson, P. Gogan, and R.K. Wayne. 1991. lntrogression of coyote mitochondrial DNA into sympatric North American gray wolf populations. Evolution. 45: 104-119.
Lehman, N., and R.K. Wayne. 1991. Analysis of coyote mitochondrial DNA genotype frequencies: estimation of the effective number of alleles. Genetics. 128: 405-416.
Lehman, N., P. Clarkson, L.D. Mech, T.J. Meier, and R.K. Wayne. 1992. A study of the genetic relationships within and among wolf packs using DNA fingerprinting and mitochondrial DNA. Behav. Ecol. Sociobiol. 30: 83-94.
Lydeard, C., P.L. Leberg, and G.D. Baumgardner. 1986. Additional notes on the taxonomic status of the coyote (Canis latrans) in Tennessee. J. Tennessee Acad. Sci. 61: 41-45.
Lydeard, C., and M.L. Kennedy. 1988. Morphological assessment of recently founded populations of the coyote, Canis latrans, in Tennessee. J. Mamm. 69: 773-781.
Mahan, B.R., P.S. Gipson, and R.M. Case. 1978. Characteristics and distribution of coyote x dog hybrids collected in Nebraska. Amer. Midi. Nat. 100: 408- 415.
McCarfey, H. 1959. The mammals of eastern Texas. Texas J. Sci. 11: 385-426.
McCarfey, H. 1962. The taxonomic status of wild Canis (Canidae) in the south central United States. Southwest. Nat. 7: 227-235. 49
McGinnis, H.J. 1979. Pennsylvania coyotes and their relationship to other wild Canis populations in the Great Lakes region and the northeastern United States. M.S. Thesis, Pennsylvania State University, University Park.
McKenzie, S. 1988. Genetic variation in nine northern subspecies of the coyote, Canis /atrans. M.S. Thesis, University of Tennessee. 50pp.
Mech, L.D. 1966. The wolves of Isle Royale. United States National Park Fauna Series, 7: 1-210.
Mech, L.D., and M.E. Nelson. 1990. Evidence of prey-caused mortality in three wolves. Amer. Midi. Nat. 123: 207-208.
Mengel, R.M. 1971. A study of dog-coyote hybrids and implications concerning hybridization in Canis. J. Mamm. 52: 316-336.
Messier, F., C. Barrette, and J. Huot. 1986. Coyote predation on a white-tailed deer population in southern Quebec. Can. Field Nat. 64: 1134-1136.
Nei, N.H., C.H. Hull, J.G. Jenkins, K. Steinbrenner, and D.H. Bent. 1975. SPPS: Statistical Package for the Social Sciences. McGraw-Hill, New York.
Nellis, C.H., and L.B. Keith. 1978. Age-related characteristics of coyote canines. J. Wild/. Manage. 42: 680-683.
Nelson, M.E., and L.D. Mech. 1985. Observation of a wolf-killed by a deer. J. Mamm. 66: 187-188.
Nowak, R.M. 1978. Evolution and taxonomy of coyotes and related Canis. In M. Bekoff, ed., Coyotes, Biology, Behavior and Management. Academic Press, New York. pp. 3-16.
Olsen, S.J., and J.W. Olsen. 1977. The Chinese wolf, ancestor of New World dogs. Science. 197: 533-535.
Ostrander, E.A., G. F. Sprague, Jr., and J. Rine. 1993. Identification and characterization of dinucleotide repeat (CA)n markers for genetic mapping in dog. Genomics. 16: 207-213.
Parker, G.R. 1986. The seasonal diet of coyotes, Canis /atrans, in northern New Brunswick. Can. Field Nat. 100: 74-77.
Parker, G.R., and J.W. Maxwell. 1989. Seasonal movements and winter ecology of the coyote, Canis /atrans, in northern New Brunswick. Can. Field Nat. 103: 1-11. OU
Parker, G. R. 1995. Eastern Coyote. Nimbus Publishing, Halifax.
Paradiso, J. 1968. Canids recently collected in east Texas, with comments on the taxonomy of the red wolf. Amer. Midi. Nat. 80: 529-534.
Peterson, R.O. 1977. Wolf ecology and prey relationships on Isle Royale. United States National Park Service Monograph Series. 11: 1-210.
Phillips, M.K. and V.G. Henry. 1992. Comments on red wolf taxonomy. Conserv. Biol. 6: 597.
Rausch, R.A. 1967. Some aspects of the population ecology of wolves, Alaska. Amer. Zoo/. 7: 253-265.
Richens, V., and R. Hugie. 1974. Distribution, taxonomic status, and characteristics of coyotes in Maine. J. Wild/. Manage. 38: 447-454.
Rosenzweig, M.L. 1968. The strategy of body size in mammalian carnivores. Amer. Midi. Nat. 80: 299-315.
Roy, M.S., E. Geffen, D. Smith, E.A. Ostrander, and R.K. Wayne. 1994. Patterns of differentiation arid hybridization in North American wolflike canids, revealed by analysis of microsatellite loci. Mo/. Biol. Evol. 11: 553-570.
Saiki, R.K., D.H. Gelfand, S. Stoffel, S.J. Scharf, R. Higuchi, G.T. Hom, K.B. Mullis, and H.A. Erlich. 1988. Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science. 239: 487-491.
Schmitz, O.J., and G.8. Kolenosky. 1985. Wolves and coyotes in Ontario: morphological relationships and origins. Can. J. Zoo/. 63: 1130-1137.
Schmitz, O.J., and D.M. Lavigne. 1987. Factors affecting body size in sympatric Ontario Canis. J. Mamm. 68: 92-99.
Schultz, V. 1955. Status of the coyote and related forms in Tennessee. J. Tennessee Acad. Sci. 30: 44-46.
Scott, J.P. 1968. Evolution and domestication of the dog. Evol. Biol. 2: 243-275.
Seton, E.T. 1929. Lives of Game Animals. Doubleday, Doran and Co., Inc. New York. 51
Silver, H., and W.T. Silver. 1969. Growth and behavior of the coyote-like canid of northern New England with observations on canid hybrids. Wild. Monogr. 17.
Thurber, J.M. and R.O. Peterson. 1991. Changes in body size associated with range expansion in the coyote (Canis Jatrans). J. Mamm. 72: 750-755.
Wayne, R.K. 1986. Cranial morphology of domestic and wild canids: the influence of development on morphological change. Eva/. 40: 243-261.
Wayne, R.K., and S.J. O'Brien. 1987. Allozyme divergence within the Canidae. Amer. Zoo/. 36: 339-355.
Wayne, R.K., B.Van Valkenburg, P.W. Kat, T.K. Fuller, W.E. Johnson, and S.J. O'Brien. 1989. Genetic and morphological divergence among sympatric canids. J. Heredity. 80: 447-454.
Wayne, R.K. and S.M. Jenks. 1991. Mitochondrial DNA analysis implying extensive hybridization of the endangered red wolf, Canis rufus. Nature. 351: 565-568.
Wayne, R.K., N. Lehman, D. Girman, P.J. Gogan, D.A. Gilbert, K. Hansen, R.O. Peterson, U.S. Seal, A. Eisenhawer, L.D. Mech, and R. J. Krumenaker. 1991. Conservation genetics of the endangered Isle Royale gray wolf. Conserv. Biol. 5: 41-51.
Wayne, R.K., M.W. Allard, and R.L. Honeycutt. 1992. Mitochondrial DNA variability of the gray wolf: genetic consequences of population decline and habitat fragmentation. Conserv. Biol. 6: 559-569.
Wayne, R.K. 1993. Molecular evolution of the <.Jog family. Trends in Genetics. 9: 218-224.
Wayne, R.K., and J.L. Gittleman. 1995. The problematic red wolf. Sci. Amer. July, 1995.
Weeks, J.L., G.M. Tori, and M.C. Shieldcastle. 1990. Coyotes (Canis /atrans) in Ohio. Ohio J. Sci. 90: 142-145.
Young, S.P. and H.H.T. Jackson. 1951. The Clever Coyote. University of Nebraska Press, Lincoln.
Young, S.P. and E.A. Goldman. 1964. The Wolves of North America. Dover Publications, New York.