Systematics of the Genus Capra Inferred from Mitochondrial DNA Sequence Data

Molecular Phylogenetics and Evolution Vol. 13, No. 3, December, pp. 504–510, 1999 Article ID mpev.1999.0688, available online at http://www.idealibrary.com on

Systematics of the Genus Capra Inferred from Mitochondrial DNA Sequence Data Vale´rie Manceau, Laurence Despre´s, Jean Bouvet, and Pierre Taberlet Laboratoire de Biologie des Populations d’Altitude, CNRS UMR 5553, Universite´ Joseph Fourier, B.P. 53, F-38041 Grenoble Cedex 9, France Received March 31, 1998; revised August 21, 1998

Traditional classification in the genus Capra is based mainly on horn morphology. However, previous investigations based on allozyme data are not consistent with this classification. We thus reexamined the evolutionary history of the genus by analyzing mitochondrial DNA (mtDNA) sequence variation. We collected bone samples from museums or dead animals found in the field. Thirty-four individuals were successfully sequenced for a portion of the mtDNA cytochrome b gene and control region (500 bp in total). We obtained a star-like phylogeny supporting a rapid radiation of the genus. In accordance with traditional classification, mtDNA data support the presence of two clades in the Caucasus and the hypothesis of a domestication event in the Fertile Crescent. However, in conflict with morphology, we found that C. aegagrus and C. ibex are polyphyletic species, and we propose a new scenario for Capra immigration into Europe. r 1999 Academic Press Key Words: Capra; molecular systematics; mitochondrial DNA; cytochrome b; control region.

INTRODUCTION The genus Capra includes several forms of wild goats that are present in mountain habitats from Northern Mongolia and Russia to Western Europe and Ethiopia (Fig. 1) and a cosmopolitan domestic form. The basic body pattern of all wild and domestic goats is similar. Moreover, they can freely interbreed in captivity. For these reasons, Couturier (1962) recognized only one species, C. aegagrus. Nevertheless, several species are often recognized in this genus (Schaller, 1977; Corbet, 1978; Nowak, 1991). For example, five wild species have been described by the International Union for Conservation of Nature (IUCN; Shackleton, 1997—see Fig. 1). The taxonomic relationships are based on the analysis of morphological characters, essentially the horn shape of males. The males of C. ibex have scimitarshaped horns with a flat anterior surface broken by 1055-7903/99 $30.00 Copyright r 1999 by Academic Press All rights of reproduction in any form reserved.

prominent transverse ridges. This criterion usually groups together the Siberian, the Nubian or Arabian, the Abyssinian, and the Alpine ibexes (respectively, C. ibex sibirica, C. i. nubiana, C. i. walia, and C. i. ibex—see Fig. 1). Capra spp. that live in the Caucasus are usually separated into two main groups, one in the Western part (Kuban tur) and the other in the Eastern part (Dagestan tur). Schaller (1977) and Shackleton (1997) include the Kuban tur in the ibex group (C. i. caucasica), whereas Corbet (1978) and Nowak (1991) classify it as a separate species: C. caucasica. All these authors agree to accord to the Dagestan tur the status of a separate species: C. cylindricornis. However, Couturier (1962) argued that there is only one taxon in the Caucasus (C. aegagrus caucasica) and that there is an East–West morphological cline. The markhor (C. falconeri) from the southwest of Asia is characterized by its twisted horns, and the Spanish ibex (C. pyrenaica) from Spain is characterized by its horns curved like a lyre. C. aegagrus occupies a large geographic area from Lake Baikal to Afghanistan. Its horns look like those of the ibexes but the anterior surface is sharper (Schaller, 1977). The domestic goat (C. hircus; Schaller, 1977; Nowak, 1991) is generally thought to have originated from C. aegagrus (Harris, 1961) but several authors (Harris, 1961; Corbet, 1978) hypothesize that other Capra spp. could also have been domesticated or inbred with the already domesticated C. aegagrus stock. Several subspecies are defined in each Capra sp. according to horn variations but the status of these subspecies is controversial. Indeed, horn morphology may not be an appropriate character for resolving cladistic issues since it is greatly variable even within a population and can undergo convergent evolution (Schaller, 1977). Few genetic data are available in this genus. Preliminary allozyme data do not support the traditional morphological taxonomy (Hartl et al., 1990, 1992). However, some of the Capra spp. have undergone drastic bottlenecks in the recent past due to hunting or habitat fragmentation so that populations

504

mtDNA SEQUENCE POLYMORPHISM IN THE GENUS Capra

505

FIG. 1. Geographical distribution of the Capra species according to Couturier (1962) and Corbet (1978). The taxonomy is according to Shackleton (1997).

may recently have randomly fixed enzymatic alleles by drift. Thus, the distances calculated from these data may not reflect the long-term history undergone by the species (Stu¨we et al., 1992). In order to clarify the systematics and the evolutionary history of the genus, we analyzed mitochondrial DNA (mtDNA) sequences. Indeed, the rapid evolution of the mitochondrial genome, as well as its lack of recombination and strict maternal inheritance, makes it an attractive marker for inferring phylogeny of closely related species. Moreover, great improvements in DNA extraction and amplification techniques make possible the analysis of very tiny amounts of DNA extracted from bones (Boom et al., 1990; Ho¨ss and Pa¨a¨bo, 1993). Organ or blood sampling was difficult for most Capra species in the wild. We also avoided samples from zoos since animals could have been hybridized (Couturier, 1962) and their geographic origins are often uncertain. We used mostly bone samples from the field or from museum specimens whose geographic origin was well documented. We studied 500-bp sequence portions of the mtDNA cytochrome b gene and the mtDNA control region.

MATERIALS AND METHODS Samples Collection Organs or blood samples were available for European species (C. ibex ibex, n ⫽ 1; C. pyrenaica, n ⫽ 2), domestic goats (C. hircus, n ⫽ 4), and Nubian ibex (C. i. nubiana, n ⫽ 1). We collected a total of 60 bone samples representing all the described Capra spp.: C. aegagrus (n ⫽ 15), C. i. caucasica (n ⫽ 9), C. cylindricornis (n ⫽ 3), C. falconeri (n ⫽ 6), C. i. sibirica (n ⫽ 12), C. i. nubiana (n ⫽ 11), C. pyrenaica (n ⫽ 2), and C. hircus (n ⫽ 2). DNA Extraction DNA was extracted from organ and blood samples by proteinase K digestion followed by phenol/chloroforme extraction and ethanol precipitation. In order to extract DNA from the bones, we first carefully cleaned the bones using sterilized tools and compressed air. A little piece (less than 1 cm3 ) was then crushed in a small mortar with liquid nitrogen. DNA was extracted using a silica-based method modified from Boom et al. (1990) as described by Taberlet and Fumagalli (1996).

506

MANCEAU ET AL.

Amplification and Direct Sequencing Double-stranded amplifications were performed in buffer containing 200 mM (NH4 ) 2SO4, 750 mM Tris– HCl, pH 9, 1.5 mM MgCl2, 1 µM each primer, 50 µM each dNTP, 5 ng of BSA, 0.1 units of Red Goldstar DNA polymerase (Eurogentec). Twenty-five cycles (denaturation: 93°C, 45 s; annealing: 50°C, 45 s; polymerization: 72°C, 90 s) were performed with DNA extracted from blood or tissue. With DNA extracted from bones, we performed a two-step PCR (Ruano et al., 1989). Twentyfive cycles were first performed with diluted primers (0.01 µM) followed by 45 cycles with concentrated primers (1 µM). The primer sets used to amplify the cytochrome b gene were from Irwin et al. (1991). We were able to amplify around 1000 bp of cytochrome b gene using L14841 and H15915 with DNA extracted from organs or blood and only a smaller portion, between L14841 and H15159, with more degraded DNA extracted from the bone samples. We also amplified, for the two types of samples, about 500 bp in the control region between L16284 (58-CACTGGTTCTTACTTCAGG-38) and H385 (58-TAGGCATTTTCAGTGCCTTG-38). L16284 was defined by Southern et al. (1988) in the central domain of the control region and H385 was defined in the tRNA phenylalanine from the analysis of Bovidae sequences found in databases. Double-stranded DNA was purified on agarose gels and used as template for asymetric amplifications. Singlestranded DNA was purified by filtration and sequenced by the method of Sanger et al. (1977) using the T7 Pharmacia manual sequencing kit. The cytochrome b gene was sequenced with L14841 as sequencing primer and the control region with L16284.

MacClade program (version 3; Maddison and Maddison, 1992), we defined tree scores of the most parsimonious trees fitting some a priori phylogenetic hypotheses. Estimation of divergence time. To estimate time of divergence among the Capra spp., we analyzed cytochrome b sequences. This region evolves more slowly than the control region so that homoplasic events are less likely to have occurred. Moreover, we used long sequences to reduce the variance in the confidence intervals of the distance estimates. We were able to analyze sequences only for taxa from which organs or blood samples were available. The evolutionary rate of cytochrome b sequences in the genus Capra has already been estimated (Manceau et al., 1999) using a divergence time of 5–7 million years (Myr) (Hartl et al., 1990). The evolutionary rate is 1.2 to 2.9% per Myr. We tested the molecular clock hypothesis with the likelihood ratio test (see Phylip Package Guidelines, version 3.57c; Felsenstein, 1995). Equal rates of mutations along branches of the maximum likelihood (ML) tree were tested by comparing the likelihoods of the two ML trees with (MLK algorithm) or without (ML algorithm) assuming a molecular clock. In these analyses, we used the mean observed Ts/Tv (⫽3) and A, C, G, and T percentages. The molecular clock hypothesis is accepted if the MLK and ML trees have the same topology and if the likelihood is not significantly decreased in the MLK tree. We defined the 95% confidence interval of genetic distances according to Hillis et al. (1996, p. 532). In order to estimate the interval of the divergence times, we divided the lower value of the 95% genetic distance interval by the highest mutation rate (2.9% per Myr) and the upper value by the smallest mutation rate (1.2% per Myr).

Data Analysis Phylogenetic analysis. Phylogenetic inferences were based on 500 bp combining 234 bp from the cytochrome b gene and 266 bp from the control region. Analyses were conducted either by neighbor-joining (NJ) or by parsimony (MP) approaches using the PAUP* program (test version 4.0d64, written by David L. Swofford). The trees were rooted using homologous sequences of Capreolus pygargus (Accession Nos. AJ000025 and Z70317). For the NJ analysis, distances were calculated with the Tamura–Nei model (Tamura and Nei, 1993) and the robustness was assessed by 2000 bootstrap replicates (Felsenstein, 1985). The parsimony analysis was assessed using the heuristic search method (character optimization ACCTRAN, MULPARS and TBR branch swapping options) since the branch and bound algorithm is too time consuming with all haplotypes. We weighted transversions 1, 3, 10, or 14 times as much as transitions. We estimated the robustness of the MP tree using 1000 bootstrap replicates (Felsenstein, 1985) with one random addition per replicate and saving a maximum of 100 trees per replicate. Then, with the

RESULTS Of the 60 bone samples, 27 successfully yielded PCR amplification product and sequence. The bone (n ⫽ 27) and organ (n ⫽ 7) samples analyzed for the phylogenetic inference are noted in Table 1. Depending on the taxonomic classification, six to nine Capra species are represented and we studied five domestic goat samples. For all the successfully sequenced samples, we were able to analyze 500 bp, combining 266 bp from the control region and 234 bp from the cytochrome b gene. Within the Capra genus, 98 sites were variable (43 within the cytochrome b sequences) and 81 were informative (37 within the cytochrome b sequences). Of the variable sites, 9 were Tv changes and 6 of them were informative. We found 30 haplotypes of the 34 samples sequenced (EMBL Accession Nos.: AJ231402–AJ231421 for cytochrome b, and AJ231422–AJ231445 for control region). The mean pairwise sequence divergence was 4.1% (uncorrected) and 4.3% (corrected). The percentage of A, C, G, and T were, respectively, 28.1, 24.3, 17.8,

507

mtDNA SEQUENCE POLYMORPHISM IN THE GENUS Capra

TABLE 1 Geographic Provenance, Name, Type, and Number of Samples Analyzed Geographic area (*domestic race)

Common name

Species

Type of sample

Origin of sample Jouy-en-Josas (INRA) Jouy-en-Josas (INRA) Jouy-en-Josas (INRA) Jouy-en-Josas (INRA) Field Vienna Museum St. Petersburg Museum St. Petersburg Museum Field St. Petersburg Museum Field Field St. Petersburg Museum Field Field Vienna Museum Field Vienna Museum Vienna Museum Vienna Museum St. Petersburg and Vienna Museum

*Angora

Domestic goat

C. hircus

Blood

Cameroon

Domestic goat

C. hircus

Blood

Morocco

Domestic goat

C. hircus

Blood

French Alps

Domestic goat

C. hircus

Blood

Pakistan Crete Turkmenistan

Domestic goat Cretean wild goat Wild goat or bezoar

C. hircus C. aegagrus C. aegagrus

Bone Bone Bone

Iran

Wild goat or bezoar

C. aegagrus

Bone

Pakistan East Caucasus

Chiltan wild goat Daghestan tur

C. aegagrus C. cylindricornis

Bone Bone

Iberian peninsula Pyrenees Tadzikistan

Spanish ibex Spanish ibex Markhor

C. pyrenaica C. pyrenaica C. falconeri

Organ Bone Bone

Pakistan French Alps Russia China Egypt Soudan Ethiopia West and Central Caucasus

Markhor Alpine ibex Siberian ibex Siberian ibex Nubian ibex Nubian ibex Nubian ibex Kuban tur

C. falconeri C. (ibex) ibex C. (ibex) sibirica C. (ibex) sibirica C. (ibex) nubiana C. (ibex) nubiana C. (ibex) nubiana C. (ibex) caucasica

Bone Blood Bone Bone Bone Bone Bone Bone

and 29.8, and the mean transition/transversion ratio between pairs of sequences was 14.5. We used homologous sequences of Capreolus pygargus to root the tree. The parsimony search with Ts/Tv ⫽ 1 resulted in 2808 most parsimonious trees with a length of 214 steps. The consistency indices were 0.60, 0.60, 0.60, and 0.80 with Ts/Tv of 1, 3, 10, and 14, respectively. Nevertheless, topologies of the four strict consensus trees were the same. Figures 2a and 2b show the bootstrap 50% majority rule consensus trees obtained from MP and NJ analyses. The C. ibex, C. falconeri, and C. aegagrus groups, defined according to horn morphology criteria, were found to be polyphyletic in the two analyses. The most-parsimonious trees that constrain the monophyly of C. ibex, C. falconeri (including C. aegagrus 5, see below), and C. aegagrus/hircus (excluding C. aegagrus 5, see below) groups were 8, 1, and 6 steps longer than the unconstraint trees, respectively. We found six groups that are well supported in bootstrap replicates: (i) C. aegagrus from Iran and Crete, together with the five breeds of domestic goat, (ii) Alpine and Spanish ibexes, (iii) C. caucasica, (iv) C. cylindricornis, (v) C. i. nubiana, and (vi) C. i. sibirica. We obtained long sequences (972 bp) of the cyto-

Sample label C. hircus 1 C. hircus 2 C. hircus 3 C. hircus 4 C. hircus 5 C. aegagrus 1 C. aegagrus 2 C. aegagrus 3 C. aegagrus 4 C. aegagrus 5 C. cylindricornis 1 to 3 C. pyrenaica 1 to 3 C. pyrenaica 4 C. falconeri 1 C. falconeri 2 and 3 C. i. ibex C. i. sibirica 1 C. i. sibirica 2 and 3 C. i. nubiana 1 and 4 C. i. nubiana 2 C. i. nubiana 3 C. i. caucasica 1 to 6

chrome b gene for four haplotypes representing four well-differentiated taxa (C. i. nubiana, C. hircus, C. i. ibex, and C. pyrenaica). The two trees constructed with or without the molecular clock hypothesis had the same topology and their likelihood values were not statistically different (P ⬎ 0.05). Therefore, we can accept the molecular clock hypothesis. The mean distance between these four haplotypes is 6.6% (64 substitutions, 95% confidence interval between 48 and 70 substitutions). We estimated a divergence time between the four haplotypes around 3 Myr, with a confidence interval between 1 and 6 Myr. DISCUSSION The extraction and amplification yield was low (less than 50%) due to the technical difficulty of extracting DNA from bones. However, this type of material allowed us to analyze individuals from several geographical areas where the sampling of tissue from wild animals would have been very difficult. Each Capra species is represented by more than one sample and we found interesting results that validate or question the morphological systematics.

FIG. 2. Phylogenetic relationships within the Capra genus. Bootstrap 50% majority rule consensus tree obtained (a) after 1000 bootstrap replicates based on MP analysis (heuristic search, Ts/Tv ⫽ 1) and (b) after 2000 bootstrap replicates based on NJ analysis (distances calculated with the Tamura–Nei model). The list of individuals analyzed is in Table 1. The phylogenetic inferences are based on 500 bp from portions of the cytochrome b gene and the control region. The bootstrap values are indicated on the internal branches.

508 MANCEAU ET AL.

mtDNA SEQUENCE POLYMORPHISM IN THE GENUS Capra

The five domestic haplotypes (C. hircus 1 to 5) and the C. aegagrus haplotype from Iran (C. aegagrus 4) were monophyletic (bootstrap values of 94% in the NJ and 77% in the MP analyses). This is consistent with the hypothesis of a domestication event in the Fertile Crescent (southwest Asia) in C. aegagrus wild goats (Mason, 1984). The haplotype of C. aegagrus from Crete (C. aegagrus 1) is almost identical to the domestic one, suggesting a feral goat as proposed by Corbet (1978). We found the haplotype of the Chiltan wild goat (C. aegagrus 5) to be genetically close to one markhor (C. falconeri 3). The spiralled horns of the Chiltan wild goats look like those of markhors but according to Schaller (1977), this criteria is not taxonomically informative and their horns have undergone convergent evolution within a small population. Our results do not support this hypothesis and we suggest that the Chiltan wild goat is either a markhor or a hybrid between a wild goat and a markhor since the hybridization could occur between wild species (Couturier, 1962; Corbet, 1978). Two C. aegagrus haplotypes (2 and 3) also appear not to be clustered with other C. aegagrus haplotypes but to be grouped with C. cylindricornis in the NJ analysis (bootstrap value of 58%). This could be due to hybridization between these two taxa since it has been reported by Corbet (1978). More generally, hybridization between Capra spp. may occur fairly frequently in the wild. Notably, hybridization between C. hircus and many other Capra spp., between C. aegagrus and C. falconeri and between C. aegagrus and C. i. nubiana, have been reported (Couturier, 1962; Corbet, 1978). Such hybridizations could be the cause for rejecting the hypotheses of monophyly. We found two mtDNA clades in the Caucasus that are congruent with the two species defined according to the morphological criteria (Schaller, 1977; Corbet, 1978; Nowak, 1991). The monophyly of the six C. i. caucasica haplotypes is supported by bootstrap values of 90 and 75% in the NJ and MP analyses, respectively, and of the three C. cylindricornis haplotypes by bootstrap values of 99 and 96%. Our data support the strong East–West differentiation in the Caucasus and reject the hypothesis of Couturier (1962) who defined only one Caucasian taxon with a morphological cline between the East and the West. However, as female wild goats are more philopatric than males, a strong mtDNA structure can be found without structure for nuclear markers. Further studies on nuclear markers should be performed to confirm the absence of gene flow between these two clades. The monophyly of the two European species (C. pyrenaica and C. i. ibex) is supported by bootstrap values of 93 and 91% in the NJ and MP analyses, respectively. Preliminary results from allozyme data (Hartl et al., 1992) also indicated a much lower genetic

509

distance between the two European species than between the other Capra species studied. The genetic distance between the European species is comparable to that commonly found between subspecies of other ungulates (Hartl et al., 1992). However, according to paleontological studies of Cre´gut-Bonnoure (1992a,b), two independent migration waves of wild goats took place in Europe, the Alpine ibex (C. i. ibex) immigrating some 300,000 years ago followed by the Spanish ibex from Caucasus (C. pyrenaica) some 80,000 years ago. Under this scenario, no close relationships would exist between European species. Our results and those of Hartl et al. (1992) reject this hypothesis and we propose a new scenario: only one wave of immigration of Capra in Europe followed by a vicariance speciation of the two European species. The paleontological description of a Capra sp. dated around 120,000 years ago in Germany (named Capra camburgensis; Toepfer, 1934) which combines the characters of the two European taxa is in accordance with our scenario. The species C. ibex includes the Alpine ibex (C. i. ibex), the Siberian ibex (C. i. sibirica), the Nubian ibex (C. i. nubiana), and sometimes the Kuban tur (C. i. caucasica). The haplotypes of these ‘‘subspecies’’ are polyphyletic in our phylogenetic inferences (eight more steps are required to achieve monophyly in the MP analysis). The scimitar shape of ibex horns appears thus to be either an ancestral or a convergent morphological character. Our results are in accordance with those of Hartl et al. (1990), who calculated genetic distances by biochemical comparison of one individual of each of the taxa C. i. ibex, C. i. nubiana, C. falconeri, and C. aegagrus and found the greatest genetic distance between C. i. ibex and C. i. nubiana. He thus proposed to consider C. i. ibex and C. i. nubiana as separate species. Our results support his proposal. The mean divergence time between the four Capra haplotypes analyzed is estimated at between 1 and 6 million years, supporting a relatively recent (PlioPleistocene) origin of the genus (Hartl et al., 1990; Pilgrim, 1947). Hartl et al. (1990) hypothesized that most Capra species have rapidly radiated during a short time span. The star-like phylogeny that we obtained also supports a rapid radiation. In this context, it is difficult to detect a sequential pattern of cladogenesis. The Capra phylogeny may be better resolved by studying longer sequences and by analyzing many samples per species from throughout the geographic range. Such a study is not tractable in the short term, considering the extreme habitats and political situations of some countries in which caprids occur. It is clear that this type of sampling represents the main future challenge for a better understanding of the evolutionary history of the genus Capra.

510

MANCEAU ET AL.

ACKNOWLEDGMENTS We thank B. Herzig from the Vienna Museum and A. N. Tikhonov from the St. Petersburg Museum, who authorized the collection of samples; D. Vaiman and L. Pe´pin from the INRA of Jouy-en-Josas, who provided samples of domestic goat; and J. P. Crampe, E. Bedin, D. Gauthier, L. Gielly, and Z. Vakaria, who provided samples from the field. We also thank E. Cre´gut-Bonnoure for her great help in collecting samples in Vienna, C. Dubois-Paganon and L. Gielly for technical assistance, and G. Luikart for comments on the manuscript.

REFERENCES Boom, R., Sol, C. J. A., Salimans, M. M. M., Jansen, C. L., Wertheimvan Dillen, P. M. E., and Van Der Noordaa, J. (1990). Rapid and simple method for purification of nucleic acids. J. Clin. Microbiol. 28: 495–503. Corbet, G. B. (1978). ‘‘The Mammals of the Palaearctic Region: A Taxonomic Review,’’ British Museum (Natural History), London. Couturier, M. A. J. (1962). ‘‘Le Bouquetin des Alpes,’’ Allier, Grenoble. Cre´gut-Bonnoure, E. (1992a). Les Caprinae (Mammalia, Bovidae) du Ple´istoce`ne d’Europe: Inte´reˆt biostratigraphique, pale´oe´cologique et arche´ozoologique. Me´m. Soc. Ge´ol. France 160: 85–93. Cre´gut-Bonnoure, E. (1992b). Inte´reˆt biostratigraphique de la morphologie dentaire de Capra (Mammalia, Bovidae). Ann. Zool. Fennici 28: 273–290. Felsenstein, J. (1985). Confidence limits on phylogenies: An approach using the bootstrap. Evolution 39: 783–791. Felsenstein, J. (1995). PHYLIP: Phylogeny Inference Package. 3.57c. Department of Genetics, Univ. of Washington, Seattle. Harris, D. R. (1961). The distribution and ancestry of the domestic goat. J. Linn. Soc. Lond. 173: 79–91. Hartl, G. B., Burger, H., Willing, R., and Suchentrunk, F. (1990). On the biochemical systematics of the Caprini and the Rupricaprini. Biochem. Syst. Ecol. 18: 175–182. Hartl, G. B., Meneguz, P. G., Apollonio, M., Marco-Sanchez, I., Nadlinger, K., and Suchentrunk, F. (1992). Molecular systematics of ibex in Western Europe. In ‘‘Congreso Internacional del Genero Capra en Europa,’’ pp. 21–26. Ronda. Hillis, D. M., Mable, B. K., and Moritz, C. (1996). Applications of molecular systematics: The state of the field and a look to the future. In ‘‘Molecular Systematics’’ (D. M. Hillis, C. Moritz, and B. K. Mable, Eds.), pp. 515–543. Sinauer, Sunderland, MA.

Ho¨ss, M., and Pa¨a¨bo, S. (1993). DNA extraction from Pleistocene bones by a silica-based purification method. Nucleic Acids Res. 21: 3913–3914. Irwin, D. M., Kocher, T. D., and Wilson, A. C. (1991). Evolution of the cytochrome b gene of mammals. J. Mol. Evol. 32: 128–144. Maddison, W. P., and Maddison, D. R. (1992). MacClade: Analysis of Phylogeny and Character Evolution. Version 3. Sinauer, Sunderland, MA. Manceau, V., Crampe, J.-P., Boursot, P., and Taberlet, P. (1999). Identification of evolutionary significant units in the Spanish wild goat Capra pyrenaica (Mammalia, Artiodactyla). Animal Cons. 2: 33–39. Mason, I. L. (1984). Goat. In ‘‘Evolution of Domesticated Animals,’’ pp. 85–89. Longman, New York. Nowak, R. N. (1991). ‘‘Mammals of the World,’’ 5th ed., John Hopkins Press, Baltimore. Pilgrim, G. E. (1947). The evolution of the buffaloes, oxen, sheep and goats. J. Linn. Soc. Lond. 41: 272–286. Ruano, G., Fenton, W., and Kidd, K. K. (1989). Biphasic amplification of very dilute DNA samples via ‘‘booster’’ PCR. Nucleic Acids Res. 17: 5407. Sanger, F., Nicklen, S., and Coulson, A. R. (1977). DNA sequencing with chain-terminating inhibition. Proc. Natl. Acad. Sci. USA 74: 5463–5467. Schaller, G. B. (1977). ‘‘Mountain Monarchs: Wild Sheep and Goats of the Himalaya,’’ Univ. of Chicago Press, Chicago. Shackleton, D. M., and the IUCN/SSC Caprinae Specialist Group. (1997). ‘‘Wild Sheep and Goats and Their Relatives: Status Survey and Conservation Action Plan for Caprinae,’’ IUCN, Gland, Switzerland. Southern, S. O., Southern, P. J., and Dizon, A. E. (1988). Molecular characterization of a cloned dolphin mitochondrial genome. J. Mol. Evol. 28: 32–42. Stu¨we, M., Scribner, K. T., and Alkon, P. U. (1992). A comparison of genetic diversity in Nubian ibex (Capra ibex nubiana) and Alpine ibex (Capra i. ibex). Z. Sa¨ugetierkd. 57: 120–123. Tamura, K., and Nei, M. (1993). Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Mol. Biol. Evol. 10: 512–526. Taberlet, P., and Fumagalli, L. (1996). Owl pellets as a source of DNA for genetic studies of small mammals. Mol. Ecol. 5: 301–305. Toepfer, V. (1934). Ein diluviales Steinbockgeho¨rn aus Thu¨ringen. Pala¨ontol. Zeitschr. 16: 276–281.