Comparative analysis of morphological and behavioral characters in the domestic dog and their importance in the reconstruction of phylogenetic relationships in canids

Comparative analysis of morphological and behavioral characters in the domestic dog and their importance in the reconstruction of phylogenetic relationships in canids

J. Jordana¹, X. Manteca² and O. Ribo³

¹Unitat de Genètica i Millora Animal, Dep. de Patologia i de Producció Animals, Facultat de Veterinària, Universitat Autònoma de Barcelona, 08193 Bellaterra, Barcelona, Spain. Send correspondence to J.J. Fax: [phone]; E-mail:[e-mail]

²Unitat de Fisiologia i Etologia, Dep. de Fisiologia, Facultat de Veterinària, Universitat Autònoma de Barcelona, Barcelona, Spain.

³Unitat de Producció Animal, Dep. de Patologia i de Producció Animals, Facultat de Veterinària, Universitat Autònoma de Barcelona, Barcelona, Spain.

ABSTRACT

Relationships among 25 dog breeds, classified a priori by their respective ancestral trunks, were studied using data from 29 morphological and 13 behavioral characteristics. Although a certain correlation was found between both types of traits (r = 0.13; P < 0.05), this relationship was not manifested, regarding the level of racial classification, in the obtained dendrograms. The relationships between breeds obtained from morphological data were more congruent than those obtained from behavioral data when compared with phylogenies from other sources of information (mainly electrophoretic analysis). This indicates that the morphological characters could give more and better complementary information than the behavioral ones in the reconstruction of the phylogenetic relationships of canids. The mean character difference (MCD), used as a measure of taxonomic resemblance between breeds, had a value of 0.53 (± 0.12 STD), and was of a magnitude very similar to that obtained in other domestic animal species (cattle, horse, sheep and goats), indicating that a similar degree of morphological differences between breeds of these species exists.

INTRODUCTION

The great diversity of canine breeds that exist in the world, and attempts at systematic classification of these breeds into closely related groups are widely known. Studies of filiations or possible genetic relationships among the different dog breeds differ substantially, because not enough archaeological or historical documents are provided to allow one to reconstruct the diversification of dog from its origin. As Peters (1969) said, "the records compiled prior to the middle of the 19th century are so few, incomplete, and inaccurate that a person can prove almost anything he cares to, regarding specific breed ancestry."

At present, it is accepted that the dog (Canis familiaris) descended from a sole ancestor, the gray wolf (Canis lupus). Some authors designate it with the name of Canis lupus, form familiaris. This hypothesis is based on comparative studies of dental morphology (Olsen and Olsen, 1977), behavior (Scott, 1968), and cranial morphology between domestic and wild canids (Wayne, 1986). Allozyme divergence studies suggest that Canis familiaris separated from the common trunk of Canis lupus between 325,000 to 1 million years ago (Wayne and O'Brien, 1987). The domestic dog is an extremely close relative of the gray wolf, differing from it by, at most, 0.2% mitochondrial DNA sequence (Wayne and Jenks, 1991). In comparison, the gray wolf differs from its closest wild relative, the coyote (Canis latrans), by about 4% mtDNA sequence (Wayne, 1993). Mitochondrial DNA control region sequences were analyzed by Vilà et al. (1997) from 162 wolves at 27 localities worldwide and from 140 domestic dogs representing 67 breeds. Sequences from both dogs and wolves showed considerable diversity and supported the hypothesis that wolves were the ancestors of dogs. The sequence divergence suggested that dogs originated more than 100,000 years before the present. Dogs are gray wolves, despite their diversity in size and proportion; the wide variation in their adult morphology probably results from simple changes in developmental rate and timing (Wayne, 1986, 1993). Nevertheless, the great diversity of findings from fossil remains (Scott, 1968; Olsen and Olsen, 1977; Davis and Valla, 1978) suggests multiple domestication events at different times and places. Dogs may be derived from several different ancestral gray wolf populations (Wayne, 1993). Nevertheless, according to Vilà et al. (1997), after the origin of dogs from a wolf ancestor, dogs and wolves may have continued to exchange genes. Backcrossing events could have provided part of the raw material for artifical selection and for the extraordinary degree of phenotypic diversity in the domestic dog.

In the species Canis familiaris, phylogenetic studies carried out by numerical taxonomy (Sneath and Sokal, 1973) are even more scarce. The main tool for a reliable reconstruction of phylogenies during the last decades has been analysis of biochemical polymorphism of structural genic loci (Tanabe et al., 1974, 1977, 1978, Sugiura et al., 1977; Hashimoto et al., 1984; Kobayashi et al., 1987; Jordana et al., 1992a). Nevertheless, possible interpretations derived at breed level should be taken with caution, because of the assumptions, limitations and biases that these methodologies could have (Jordana et al., 1992a). Secondly, an attempt to analyze the evolutionary divergence between breeds is extremely difficult, due to the short amount of time that has lapsed in their evolution, which means that genetic, racial differences are due essentially to changes in genic frequencies rather than to fixation or loss of several alleles in the genes and analyzed populations. At present, the possibility of analyzing the great quantity of polymorphisms that exist at the DNA level, in some determined hypervariable regions of the genome named "minisatellites" and "microsatellites", will doubtlessly help clarify the phylogenetic relationships among breeds (Bowcock et al., 1994), as preliminary results demostrate (Altet et al., 1996).

Olsen and Olsen (1977) and Clutton-Brock (1984), according to comparative studies of dental morphology from archaeological and present findings, suggest that the small western Asiatic wolf, Canis lupus arabs, was probably the progenitor of most European and southern Asiatic dogs. The small Chinese wolf, Canis lupus chanco, could be the ancestor of the early Chinese and Mongolian dogs, while the large northern wolves, Canis lupus lycaon, could be the ancestors of the so-named Dog of the Turberas or Canis familiaris palustris, main progenitor of the Eskimo breeds of Spitz type. From Canis lupus arabs, four new primitive types of dogs appeared: Canis familiaris metris-optimae, ancestor of the shepherd dogs; Canis familiaris inostranzewi, which would be converted into the predecessor of the moloses from northern India; Canis familiaris intermedius, from which would proceed the braques, and lastly the medium-oriental trunk, the Canis familiaris leineri, which would give rise to the harriers.

The main objective of this study was to analyze several morphological and behavioral characters of a series of breeds, lumped together historically in the ancestral trunks previously commented (Antonius, 1922; Rousselet-Blanc, 1983), in order to improve the knowledge of relationships among the current canine breeds. Do the two types of characters group the breeds similarly? What is the degree of existent correlation between the morphological and behavioral characters? What characters would be more useful in order to supplement the reconstruction of phylogenies?

Lastly, it must be pointed out that the results obtained in this study attempted to show only the degree of relationship and morphological and behavioral similarity among current dog breeds, which may or may not be indicative of the true evolutionary history of the populations. It ought to be considered that morphological and behavioral characters have been subjected to artificial selection over a long period of time. Also, there has been genetic migration among some of these populations.

This work presents studies from qualitative and quantitative data analyses, using statistical methods (SAS, 1988) and available computing packages specially designed for such analyses (Swofford, 1993).

MATERIAL AND METHODS

Breeds studied

A total of 25 present-day canine breeds have been studied and classified a priori in different ancestral trunks. Canis lupus chanco: Lhasa apso (LA), Pekingese (PK) and Shih-Tzu (ST). Canis lupus lycaon: which would give rise to the Canis familiaris palustris from which would derive breeds such as: Akita-Inu (AI), Alaskan malamute (AM), Chow-chow (CC), Pomeranian (PM), Samoyed (SA), and Siberian Husky (SH). Canis lupus arabs: which would give rise to four primitive types of dogs:

Canis familiaris metris-optimae: German shepherd dog (GS), Old English sheepdog (OE), Rough collie (RC), and Shetland sheepdog (SS).

Canis familiaris leineri: Afghanhound (AF).

Canis familiaris intermedius: Basset hound (BH), Beagle (BE), Dalmatian (DA), Golden retriever (GR), Irish red setter (IS), Labrador retriever (LR), and Pointer (PO).

Canis familiaris inostranzewi: Great Dane (GD), Newfoundland (NF), Rottweiler (RO), and Saint Bernard (SB).

Characters and analyses

A total of 29 morphological characters were studied using an ideal specimen for each of the 25 dog breeds. The state of each of the characters for each breed was established according to descriptions offered by official racial standards (Gómez-Toldrà, 1985). Numbers were assigned to each state of the different characteristics in an arbitrary manner. These numbers did not represent any specific weight. The number of states for each character was established depending on the number of distinguishable phenotypic classes. The characters used and their state numbers are shown in Table I. Continuous quantitative characters (Z, A, B and C characters in Table I) may be split into a small number of classes, each representing one of the states of the character in the data matrix. The original matrix of morphological resemblances is shown in Table II.

Table I
- Characters and their states, used for the construction of the morphological resemblance matrix.

(A) Size	(B) Length/width proportions	(C) Cranial profile	(D) Occipital crest	(E) Stop
0. Medium	0. Longilinear	0.Subconvex	0. Pronounced	0. Pronounced
1. Large	1. Mesolinear	1. Convex	1. Slightly pronounced	1. Moderate
2. Small	2. Brevilinear	2. Rectilinear	2. Null	2. Slightly pronounced or null
(F) Face profile	(G) Tooth position	(H) Ear insertion	(I) Ear position	(J) Ear size in relation to head
0. Straight	0. Scissors	0. Up (above eye line)	0. Fallen	0. Small
1.Subconvex	1. Pincers	1. Middle (same height)	1. Erect	1. Middle (proportionate)
2. Convex	2. Both	2. Down (below eye line)	2. Folded	2. Large
(K) Eye form	(L) Eye color	(M) Truffle color	(N) Lips	(O) Neck dewlap
0. Rounded	0. Dark (black/brown)	0. Black	0. Straight	0. Present
1. Almond	1. Bright (amber/hazelnut)	1. Meat	1. Superior just covers inferior	1. Absent
2. Oblique	2. Various	2. Brown	2. Superior covers inferior
(P) Dorsal line	(Q) Back	(R) Withers	(S) Skin	(T) Hair length
0. Straight	0. Oblique	0. Slightly pronounced	0. Tensed	0. Long
1. Slightly saddled	1. Straight	1. Well defined	1. With pleat	1. Middle length
2. Slightly carped				2. Short
				3. Various
(U) Belly	(V) Tail insertion	(W) Tail form	(X) Tail length	(Y) Foot form
0. Slightly contracted	0. Up	0. Fallen (sword)	0. Short	0. Cat (rounded)
1. Contracted	1. Middle	1. Sabre	1. Middle length	1. Hare (oval)
2. Very contracted	2. Down	2. Sickle	2. Long
		3. Back level
		4. Upright
		5. Without form
(Z) Live weight in males	(A) Live weight in females	(B) Wither height in males	(C) Wither height in females
0. £ 10 kg	0. < 10 kg	0. < 30 cm	0. < 30 cm
1. 11-20 kg	1. 11-20 kg	1. 31-46 cm	1. 31-44 cm
2. 21-28 kg	2. 21-30 kg	2. 47-57 cm	2. 45-52 cm
3. 29-37 kg	3. 31-40 kg	3. 58-65 cm	3. 53-62 cm
4. 38-45 kg	4. > 40 kg	4. 66-72 cm	4. 63-68 cm
5. > 45 kg		5. > 72 cm	5. > 68 cm

Also, a total of 13 behavioral characters described by Hart and Miller (1985) and by Hart and Hart (1985) were analyzed. Basing ourselves on the scale assigned by these authors to each breed for each character (ranging 0 to 10), three states for each of the characters were assigned: 0 = low, 1 = medium, and 2 = high. The analyzed characters were: (A) excitability, (B) general activity, (C) snapping at children, (D) excessive barking, (E) affection demand, (F) territorial defense, (G) watchdog barking, (H) aggression to dogs, (I) dominance over owner, (J) obedience training, (K) housebreaking ease, (L) destructiveness, and (M) playfulness. The original matrix of behavioral resemblances is shown in Table III.

Table III
- Behavioral resemblance matrix.

Breeds\Characters	A	B	C	D	E	F	G	H	I	J	K	L	M
Afghanhound (AF)	1	1	2	1	0	2	0	2	2	0	1	2	0
Akita-Inu (AI)	0	0	0	0	0	2	1	2	1	2	2	0	0
Alaskan malamute (AM)	0	0	1	0	1	2	0	2	2	1	1	2	0
Basset hound (BH)	0	0	0	1	0	0	0	1	0	0	0	0	0
Beagle (BE)	2	2	1	2	0	0	1	1	2	0	0	2	1
Chow chow (CC)	1	0	2	1	0	2	1	2	2	0	0	0	0
Dalmatian (DA)	1	1	2	1	0	1	1	2	2	1	0	2	1
German shepherd dog (GS)	1	1	1	1	0	2	2	2	2	2	2	2	2
Golden retriever (GR)	0	1	0	0	2	0	0	0	0	2	1	0	2
Great Dane (GD)	0	0	0	0	0	2	1	1	1	1	1	1	0
Irish red setter (IS)	2	2	1	2	2	0	1	0	1	0	1	2	2
Labrador retriever (LR)	0	0	0	0	1	1	1	0	0	2	2	0	2
Lhasa apso (LA)	2	1	2	2	2	1	2	1	2	1	0	1	0
Newfoundland (NF)	0	0	0	0	1	0	0	0	0	1	2	0	0
Old English sheepdog (OE)	0	0	1	0	1	0	0	1	1	0	1	2	0
Pekingese (PK)	2	1	2	2	2	1	1	2	2	0	0	0	0
Pointer (PO)	1	2	0	1	1	1	0	0	1	1	1	2	2
Pomeranian (PM)	2	2	2	2	2	1	2	1	1	0	0	2	0
Rottweiler (RO)	0	0	1	0	0	2	2	2	2	2	2	0	0
Rough collie (RC)	0	0	0	0	0	1	1	0	0	2	2	0	1
Saint Bernard (SB)	0	0	1	0	0	2	0	1	2	0	1	0	0
Samoyed (SA)	1	0	1	1	0	2	0	1	2	0	1	0	0
Shetland sheepdog (SS)	0	2	1	2	2	2	2	0	0	2	2	0	2
Shih-Tzu (ST)	2	2	1	1	2	1	2	0	1	1	2	1	2
Siberian Husky (SH)	1	2	2	1	0	1	0	2	2	1	0	2	1

For qualitative analysis of morphological and behavioral characters, the PAUP package (Phylogenetic Analysis Using Parsimony) (Swofford, 1993) was used for the discrete characters shown in Tables II and III. This analysis was based on the parsimony principle, and the criterion was to find the tree (dendrogram) that required the least number of changes. The method used was Fitch parsimony (Fitch, 1971). To give an evolutionary direction, resulting trees were rooted using the midpoint rooting method (Farris, 1972). The PAUP package also allows one to compute the confidence limits of the topology by means of a bootstrap analysis (Efron, 1979), adapted to the inference of phylogenies (Felsenstein, 1985). One hundred bootstrap replicates were made, and a consensus tree was obtained based on the majority-rule method (Margush and McMorris, 1981) and produced by global branch swapping algorithm (Hendy and Penny, 1982).

For quantitative analysis, qualitative data were transformed and processed into a matrix of distances. The mean character difference (MCD) proposed by Cain and Harrison (1958) was calculated as a measure of taxonomic resemblance.

MCD = 1/n S_i oX_ij - X_iko

MCD varied between 0 and 1, where n was the number of traits and (X_ij - X_ik) = alternative values (0 or 1) for the differences between j and k breeds within the character i.

X_ij - X_ik = 0 if X_ij = X_ik

X_ij - X_ik = 1 if X_ij¹ X_ik

Also, from MCD values obtained between populations, the coefficient of correlation (Pearson's moment product) between both types of characters (behavior vs. morphology) was calculated, using a statistical software program (SAS, 1988).

RESULTS AND DISCUSSION

Qualitative analysis

The dendrogram resulting from the application of the Fitch parsimony method to the morphological traits (Table II) is shown in Figure 1. Fitch parsimony needed 161 steps (total length of the tree) to rearrange the characters and to obtain the maximum parsimonious tree. The consistency index (a measure of homoplasy) was 0.391. Branch and internodal distances were proportional to the number of character-stage changes required.

Figure 1
- Qualitative analysis of morphological data (Table II). Dendrogram produced by PAUP analysis, resulting from the application of Fitch parsimony method (Fitch, 1971). Branch and internodal distances are proportional to the number of character-stage changes required. The tree was rooted at the midpoint.

Two large, perfectly definite clusters can be observed in this tree. Cluster A is formed by all the breeds that had previously been integrated into the groups C. f. leineri, C. f. intermedius and C. f. inostranzewi, all of which descended from the ancestral trunk C. l. arabs. Cluster B includes descendants of C. f. metris-optimae, and descendants of both C. l. chanco and C. l. lycaon ancestral trunks.

The fact that C. f. metris-optimae was assigned to cluster B and not A, which would seem more logical if we consider that this group is also a direct descendent of the small western Asiatic wolf (C. l. arabs), attracts attention.

However, it must be remembered that morphological characters have been subjected to a great pressure of selection, both natural and artificial. Selection has had evolutionary strength, and would have been a great influence in the process of breed differentiation. For more reliable studies of evolutionary divergence, neutral structural genes, with a high rate of polymorphism and no relationship with the fitness of the individuals, would be more appropriate, especially for alleles that appeared by mutation and fixation or lost by drift, specifically the markers of DNA, minisatellites and microsatellites (Bruford and Wayne, 1993; Bowcock et al., 1994). Nevertheless, analysis of morphological characters indicated the relationships or similarity, at this level, between current dog breeds. The dendrogram shows that the shepherd canine breeds (C. f. metris-optimae) have maintained greater gene flow with descendants of both C. l. chanco and C. l. lycaon than with other representatives of C. l. arabs. If migration were not considered very probable, then the possibility of convergence evolution (Sneath and Sokal, 1973) for some morphological traits among several trunks could be argued.

At the individual breed level, morphological analysis perfectly assigned each population into their ancestral trunk. The greatest degree of disagreement was observed within cluster A, where representatives of C. f. intermedius were related more closely with representatives of C. f. inostranzewi, specifically the Basset hound and Beagle breeds, which form a very definite cluster with Great Dane, Saint Bernard, Newfoundland and Rottweiler breeds. Similar results of this close relationship between C. f. intermedius and C. f. inostranzewi were also obtained by Jordana et al. (1992b) using exclusively Spanish dog breeds.

According to Olsen and Olsen (1977) as well as our morphological results, a close evolutionary relationship could exist between the C. l. chanco and C. l. lycaon ancestral trunks. The small Chinese, or Asian, wolf (C. l. chanco) is the most likely ancestor of the early domestic dog in China and Mongolia, as well as of those dogs that probably accompanied humans to North America over the Bering Strait. Large northern wolves (C. l. lycaon) were the ancestors of the more recent large dogs of the circumpolar areas. According to Olsen and Olsen (1977), some link between these animals and the first tamed small wolves must still be sought.

Figure 2 represents the majority-rule consensus tree formed after one hundred bootstrap replicates; the values in the tree indicate the number of replicates from the bootstrap analysis (loosely, the width of the confidence interval). Both the parsimony tree and the majority-rule consensus tree are similar enough, but they do differ in some respects, which is frequently the case (Sanderson, 1989). In the bootstrap tree, an unresolved trichotomy was present between Afghanhound, Irish red setter, and other breeds. Moreover, the cluster formed by the Basset hound and Beagle breeds (C. f. intermedius) was no longer related very closely with descendants of the C. f. inostranzewi trunk. However, low significance levels indicate that there is little confidence in this arrangement.

Figure 2
- Majority-rule consensus tree and bootstrap replicates by PAUP analysis from morphological data.

The dendrogram resulting from the application of the Fitch parsimony method to the behavioral traits (Table III) is shown in Figure 3. Fitch parsimony needed 86 steps to rearrange the characters and to obtain the maximum parsimonious tree. The consistency index was 0.302.

Figure 3
- Dendrogram obtained by PAUP analysis, using qualitative behavioral data.

In this dendrogram, two large clusters were also formed, but the pattern of behavioral relationships among breeds was confused, and totally incompatible with those obtained from morphological and electrophoretical analysis (Tanabe et al., 1974; Kobayashi et al., 1987). These traits were apparently those most severely influenced by natural and/or artificial selection, stochastic factors, and local environmental pressures. Not even at the ancestral trunk level could a certain tendency be suggested for the evolution of the breeds; perhaps if the number of analyzed characters had been larger any tendency, more or less logical, could have been argued for the ancestral relationships.

With regard to degree of relationship among individualized breeds, no clear and consistent relationships were apparent. Figure 4 shows the majority-rule consensus tree formed after one hundred bootstrap replicates. The very low levels of significance indicate that there is very little confidence in this arrangement. The only minimally stable relationships would be between the breeds of Labrador retriever (C. f. intermedius) and Rough collie (C. f. metris-optimae) with a confidence level of 58%, and between Dalmatian (C. f. intermedius) and Siberian Husky (C. f. palustris) breeds with a confidence level of 52%. Only Hart and Hart (1985), analyzing the same 13 behavioral traits in a total of 56 dog breeds by simple cluster analysis, argued that the clusters obtained could reflect the conventional groupings of dogs to some degree into working, sporting, hound and terrier breeds. Nevertheless, in conclusion, it appears that behavioral characters are of little value in determining the phylogeny of the Canis familiaris. Results and similar conclusions, with regard to behavioral traits, as a complementary tool in order to infer phylogenies, were obtained by Spicer (1992) in the evaluation of the phylogeny of the Drosophila virilis species group. He compared the obtained relationships using several sources: morphological, behavioral, chromosomal, developmental, and molecular data.

Figure 4
- Majority-rule consensus tree and bootstrap replicates by PAUP analysis from behavioral data.

Quantitative analysis

Values for the mean character differences (MCD) between dog breeds for morphological and behavioral characters are shown in Table IV.

For the morphological traits, the average MCD between breeds had a value of 0.53 (± 0.12 STD), with extreme values of 0.03 between the Lhasa apso vs. Shih-Tzu pair, and 0.83 for the Akita-Inu vs. Basset hound pair. This average MCD value between dog breeds has a very similar magnitude to that obtained between other domestic animal species; for instance, bovine, equine, ovine, caprine and Spanish dogs had MCD values of 0.57 ± 0.12 (Jordana et al., 1991), 0.57 ± 0.17 (Jordana et al., 1995), 0.55 ± 0.12 (Jordana and Ribó, 1991), 0.66 ± 0.10 (Jordana et al., 1993) and 0.57 ± 0.15 (Jordana et al., 1992b), respectively, which indicates that a similar degree of morphological differences exists between the breeds of these species.

If the existent morphological variability within each ancestral trunk is analyzed by calculating the corresponding MCDs, intratrunk MCD average values could be obtained which oscillate between 0.34 for C. f. inostranzewi and 0.44 for C. f. intermedius. The only value that moves away from this range corresponds to the representatives of C. l. chanco (MCD = 0.18). The F test for heterogeneity of means by the Student-Newman-Keuls test (SAS, 1988) indicates that significant differences exist only between C. l. chanco and the remaining ancestral trunks (F = 2.99; P < 0.05). This result would indicate that descendant breeds of C. l. chanco (Lhasa apso, Pekingese and Shih-Tzu breeds) have a greater relationship of morphological similarity than the representatives of the other ancestral trunks, corroborating the close, observed relationships in the dendrogram of Figure 1. Nevertheless, this assertion should be taken with caution, due to the reduced number of analyzed breeds and to the great similarity that exists between two of them, in particular Lhasa Apso vs Shih-Tzu (MCD = 0.03).

For the behavioral traits, the average MCD between breeds had a value of 0.66 (± 0.18 STD), with extreme values of 0.15 between the Labrador retriever vs. Rough collie pair, and 1.00 for the Afghanhound vs. Labrador retriever, Dalmatian vs. Shetland sheepdog, Golden retriever vs. Samoyed, and Pointer vs. Rottweiler pairs.

If the breeds were assembled in their hypothetical ancestral trunks in a similar way to that carried out with morphological characters, the existent behavioral variability within each ancestral trunk could be analyzed by calculating the corresponding MCDs. Intratrunk MCD average values oscillated between 0.48 for C. f. intermedius and 0.72 for C. f. metris-optimae. The F test for heterogeneity of means by the Student-Newman-Keuls test did not show significant differences between any group, indicating that the degree of behavioral intratrunk variability was similar for all groups.

Although a certain degree of significant correlation between both types of traits exists (r = 0.13; P < 0.05), this relationship was not manifested, at the level of racial classification, in the obtained dendrograms. Therefore, although the two types of traits could have been influenced by natural and/or artificial selection, stochastic factors, and local environmental pressures, the pattern of morphological relationships among breeds was not incompatible with those obtained from electrophoretic analysis (Tanabe et al., 1974; Kobayashi et al., 1987; Jordana et al., 1992a). This could be due to the higher heritability values (h²) that these characters show, ranging between 0.3 to 0.8 on average (Verryn and Geerthsen, 1987), compared to those of behavior, whose h² ranges between 0.0 to 0.2 on average (Mackenzie et al., 1986; Karjalainen et al., 1994). This makes the phenotype (P) a better indicator of genotype (G) due to environmental component (E) not having the same influence as is the case of the behavior characters, according to a formula proposed by Johannsen (1903) where phenotype = genotype + environment (P = G + E). In conclusion, it could be said that morphological characters could give more and better complementary information than that of behavioral ones in the reconstruction of the phylogenetic relationships of Canids.

ACKNOWLEDGMENTS

The suggestions of the reviewers were greatly appreciated. Also, we would like to thank Chuck Simmons for assistance with the preparation of this manuscript.

RESUMO

As relações entre 25 raças de cães, classificadas "a priori" pelos seus respectivos troncos ancestrais, foram estudadas usando-se dados de 29 características morfológicas e 13 comportamentais. Embora tenha sido encontrada uma certa correlação entre ambos os tipos de caracteres (r = 0.13; P < 0.05), esta relação não se manifestou, com relação ao nível de classificação racial, nos dendrogramas obtidos. As relações entre as raças obtidas dos dados morfológicos foram mais congruentes do que as obtidas dos dados comportamentais, quando comparadas com filogenias de outras fontes de informação (análise eletroforética, principalmente). Isso indica que os caracteres morfológicos poderiam fornecer informações complementares melhores e em maior número do que os caracteres comportamentais, na reconstrução das relações filogenéticas dos canídeos. A diferença média de caráter (MCD), usada como uma medida de semelhança taxonômica entre raças, teve um valor de 0,53 (± 0,12 desvios padrão) e foi de uma magnitude muito semelhante à obtida em outras espécies de animais domésticos (bovinos, eqüinos, ovinos e caprinos), indicando que existe um grau similar de diferenças morfológicas entre as raças destas espécies.

(Received August 8, 1997)

• Altet, L., Francino, O. and Sánchez, A. (1996). Microsatellite typing in Spanish populations of pure breed dogs for paternity testing and animal identification. Anim. Genet. 27: 87.
• Antonius, O. (1922). Grundzüge einer Stammesgesichte der Haustiere. Gustav Fischer, Jena, pp. 75-138.
• Bowcock, A.M., Ruiz-Linares, A., Tomfohrde, J., Minch, E., Kidd, R.J. and Cavalli-Sforza, L.L. (1994). High resolution of human evolutionary trees with polymorphic microsatellites. Nature 368: 455-457.
• Bruford, M.W. and Wayne, R.K. (1993). Microsatellites and their application to population genetic studies. Curr. Op. Genet. Devel. 3: 939-943.
• Cain, A.J. and Harrison, G.A. (1958). An analysis of the taxonomist's judgement of affinity. Proc. Zool. Soc. Lond. 131: 85-98.
• Clutton-Brock, J. (1984). Dog. In: Evolution of Domesticated Animals (Mason, I.L., ed). Longman, London and New York, pp. 198-211.
• Davis, S.J.M. and Valla, F.R. (1978). Evidence for domestication of the dog 12,000 years ago in the Natufian of Israel. Nature 276: 608-610.
• Efron, B. (1979). Bootstrap methods: another look at the jackknife. Ann. Stat. 7: 1-26.
• Farris, J.S. (1972). Estimating phylogenetic trees from distance matrices. Am. Nat. 106: 645-668.
• Felsenstein, J. (1985). Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783-791.
• Fitch, W.M. (1971). Toward defining the course of evolution: minimal change for a specific tree topology. System. Zool. 20: 406-416.
• Gómez-Toldrŕ, S. (1985). Los Perros. Gran Enciclopedia Canina. Planeta-De Agostini, Barcelona.
• Hart, B.L. and Hart, L.A. (1985). Selecting pet dogs on the basis of cluster analysis of breed behavior profiles and gender. J. Am. Vet. Med. Assoc. 186: 1181-1185.
• Hart, B.L. and Miller, M.F. (1985). Behavioral profiles of dog breeds. J. Am. Vet. Med. Assoc. 186: 1175-1180.
• Hashimoto, Y., Yamakawa, T. and Tanabe, Y. (1984). Further studies on the red cell glycolipids of various breeds of dogs. A possible assumption about the origin of Japanese dogs. J. Biochem. 96: 1777-1782.
• Hendy, M.D. and Penny, D. (1982). Branch and bound algorithms to determine minimal evolutionary trees. Math. Biosci. 59: 277-290.
• Johannsen, W. (1903). Über Erblichkeit in Populationen und in reinen Linien. Gustav Fisher, Jena.
• Jordana, J. and Ribó, O. (1991). Relaciones filogenéticas entre razas ovinas espańolas obtenidas a partir del estudio de caracteres morfológicos. Invest. Agrar.: Prod. Sanid Anim. 6: 51-63.
• Jordana, J., Pelegrín, M. and Piedrafita, J. (1991). Relaciones genéticas en bovinos espańoles, obtenidas a partir del estudio de caracteres morfológicos. Inf. Téc. Econ. Agrar. 80: 50-64.
• Jordana, J., Piedrafita, J. and Sánchez, A. (1992a). Genetic relationships in Spanish dog breeds. II. The analysis of biochemical polymorphism. Genet. Sel. Evol. 24: 245-263.
• Jordana, J., Piedrafita, J. and Sánchez, A. (1992b). Genetic relationships in Spanish dog breeds. I. The analysis of morphological characters. Genet. Sel. Evol. 24: 225-244.
• Jordana, J., Ribó, O. and Pelegrín, M. (1993). Analysis of genetic relationships from morphological characters in Spanish goat breeds. Small Ruminant Res. 12: 301-314.
• Jordana, J., Parés, P.M. and Sánchez, A. (1995). Analysis of genetic relationships in horse breeds. J. Equine Vet. Sci. 15: 320-328.
• Karjalainen, L., Ojala, M. and Vilva, V. (1994). Estimates of heritability for measures of hunting performance in the Finnish Spitz. In: Proceedings of the 5th World Congress on Genetics Applied to Livestock Production (Smith, C., Gavora, J.S., Benkel, B., Chesnais, J., Fairfull, W., Gibson, J.P., Kennedy, B.W. and Burnside, E.B., eds). Vol. 20. University of Guelph, Guelph, pp. 165-168.
• Kobayashi, R., Miyakawa, H., Tanabe, Y., Hashimoto, Y., Ota, K. and Faruque, M.O. (1987). Blood proteins polymorphism in the Bangladesh native dogs. Genet. Stud. Breed Diff. Native Domest. Anim. Bangladesh 2: 93-103.
• Mackenzie, S.A., Oltenacu, E.A.B. and Houpt, K.A. (1986). Canine behavioral genetics: a review. App. Anim. Behav. Sci. 15: 365-393.
• Margush, T. and McMorris, F.R. (1981). Consensus n-trees. Bull. Math. Biol. 43: 239-244.
• Olsen, S.J. and Olsen, J.W. (1977). The Chinese wolf, ancestor of New World dogs. Science 197: 533-535.
• Peters, J.A. (1969). Canine breed ancestry. J. Am. Vet. Med. Ass. 155: 621-624.
• Rousselet-Blanc, P. (Ed.) (1983). Nuestro Amigo el Perro. Selecciones del Reader's Digest (Iberia), Madrid.
• Sanderson, M.J. (1989). Confidence limits on phylogenies: the bootstrap revisited. Cladistics 5: 113-129.
• SAS (1988). SAS User's Guide: Statistics. Version 6.03. SAS Institute Inc., Cary.
• Scott, J.P. (1968). Evolution and domestication of the dog. Evol. Biol. 2: 243-275.
• Sneath, P.H.A. and Sokal, R.R. (1973). Numerical Taxonomy. Freeman, San Francisco.
• Spicer, G.S. (1992). Reevaluation of the phylogeny of the Drosophila virilis species group (Diptera: Drosophilidae). Ann. Entomol. Soc. Am. 85: 11-25.
• Sugiura, S., Tanabe, Y. and Ota, K. (1977). Genetic polymorphism of eserine resistant esterases in canine plasma. Anim. Blood Groups Biochem. Genet. 8: 121-126.
• Swofford, D.L. (1993). PAUP: Phylogenetic Analysis Using Parsimony. Version 3.1.1. Computer program distributed by the Illinois Natural History Survey, Champaign, Illinois.
• Tanabe, Y., Sugiura, S., Asanoma, M. and Ota, K. (1974). Genetic polymorphism of leucine aminopeptidase in canine plasma. Anim. Blood Groups Biochem. Genet. 5: 225-230.
• Tanabe, Y., Omi, T. and Ota, K. (1977). Genetic variants of glucose phosphate isomerase (E.C. 5.3.1.9) in canine erythrocytes. Anim. Blood Groups Biochem. Genet. 8: 191-195.
• Tanabe Y., Omi T. and Ota K. (1978). Genetic variants of hemoglobin in canine erythrocytes. Anim. Blood Groups Biochem. Genet. 9: 79-83.
• Verryn, S.D. and Geerthsen, J.M.P. (1987). Heritabilities of a population of German Shepherd Dogs with a complex interrelationships structure. Theor. Appl. Genet. 75: 144-146.
• Vilŕ, C., Savolainen, P., Maldonado, J.E., Amorim, I.R., Rice, J.E., Honeycutt, R.L., Crandall, K.A., Lundeberg, J. and Wayne, R.K. (1997). Multiple and ancient origins of the domestic dog. Science 276: 1687-1689.
• Wayne, R.K. (1986). Cranial morphology of domestic and wild canids: the influence of development on morphological change. Evolution 40: 243-261.
• Wayne, R.K. (1993). Molecular evolution of the dog family. Trends Genet. 9: 218-224.
• Wayne, R.K. and Jenks, S.M. (1991). Mitochondrial DNA analysis implying extensive hybridization of the endangered red wolf Canis rufus Nature 351: 565-568.
• Wayne, R.K. and O'Brien, S.J. (1987). Allozyme divergence within the canidae. Syst. Zool. 36: 339-355.

Publication Dates

• Publication in this collection
  02 June 1999
• Date of issue
  Mar 1999

History

• Received
  08 Aug 1997