Geographical Origin of Advanced article the Domestic Dog Article Contents . Introduction Cornelya FC Klu¨tsch, KTH-Royal Institute of Technology, Gene Technology, Stockholm, . Archaeological Record . Measurements in Archaeological Research Sweden . Genetical Studies

Peter Savolainen, KTH-Royal Institute of Technology, Gene Technology, Stockholm, Sweden . Genome-wide SNP Data

. Ancient DNA Studies

. Conclusion

Online posting date: 15th April 2011

The domestic dog (Canis familiaris) is considered to be the et al., 2003). Until 20 years ago, the origin of the domestic oldest domestic animal in the world. World-wide mito- dog was a total riddle. It was not clear if the wild ancestor chondrial deoxyribonucleic acid (mtDNA) studies clearly was wolf (Canis lupus), jackal (Canis aureus; Clutton- indicate a single origin in time and place in Southeast Asia Brock, 1995) or coyote (Canis latrans) or if all three or a combination of two of these contributed to the gene pool of less than 16 000 years ago including a high number of the dog. Even more contention existed in the field when it female foundation wolves resulting in 10 subhaplogroups came to the questions ‘where’, ‘when’ and ‘how many within three haplogroups. Later hybridisation events in times’ the dog was domesticated. Since then, especially East Asia, the Middle East, Scandinavia and possibly North through genetical studies, the picture has become increas- America formed 3–4 small haplogroups. In contrast, the ingly clearer, for example, most importantly, pointing out archaeological record, favours other, sometimes multiple, that the wolf is the only ancestor of the dog (Figure 1; Olsen regions for dog domestication (mainly Europe and Middle and Olsen, 1977; Wayne, 1993; Vila` et al., 1997; Leonard East), but suffers from a lack of samples from Southeast et al., 2002; Lindblad-Toh et al., 2005). However, the Asia and is problematic, because of the difficulty of dis- tinguishing between wolf and dog in fossil remains. Future studies including, for example, paternal markers such as the Y-chromosome, autosomal markers like SNPs, and ancient DNA samples of both wolves and dogs may give new insights into early domestication history and the dog’s migration routes.

Introduction

The domestication of plants and animals was the crucial step in the Neolithic transition from a hunter-gather to a sedentary lifestyle which led to the advanced agricultural societies and eventually to modern civilisations (Diamond, 2002). Therefore, domestication is considered to be the most significant innovation in history for human society (Diamond, 2002). The study of domestication may tell us not only something about the domestic animal, but also about the human cultural context, thus giving insights about human societies throughout the world (Bruford

ELS subject area: Evolution and Diversity of Life

How to cite: Klu¨tsch, Cornelya FC; and Savolainen, Peter (April 2011) Geographical Origin of the Domestic Dog. In: Encyclopedia of Life Sciences (ELS). John Wiley & Sons, Ltd: Chichester. DOI: 10.1002/9780470015902.a0022867 Figure 1 The dog’s ancestor, the wolf (Canis lupus). Image courtesy of Gary Kramer, US Fish and Wildlife Service.

ENCYCLOPEDIA OF LIFE SCIENCES & 2011, John Wiley & Sons, Ltd. www.els.net 1 Geographical Origin of the Domestic Dog domestic dog is a good illustration of how different scien- that these areas are more likely to be areas of origins for the tific evidence (of different nature) may result in diverse dog just because a larger number of early Canid specimens interpretations. Therefore, there is not full agreement have been found there (see Table 1 for a review of the earliest about the details of dog domestication, but the picture archaeological findings proposed to be dog). becomes increasingly clearer. However, wolves were formerly distributed throughout The dog is generally considered to be the earliest North America and Eurasia (Nowak, 2003; Napierala and domestic animal (e.g. Zeder et al., 2006; Wang and Ted- Uerpmann, 2010). Logically it follows that a comprehen- ford, 2008) and has been unparalleled in its success as a sive picture of dog domestication can only be validated if companion in human society. It is the only animal which in fossil of dog/wolf remains from all parts of the Old World ancient times accompanied humans to all continents and, are studied. Nevertheless, it is generally agreed by many in contrast to other domestic animals, it serves a large scientists that either Chinese or Middle Eastern wolves number of different ‘roles’ as hunting, herding, guarding, were the ancestors of the domestic dog (Olsen and Olsen, guiding, searching and rescuing companion (Figure 2). In 1977), because these southern wolves are smaller in size and some countries it is also a source of food and fur and it has might have been easier to handle than the much larger wolf lately also become a model organism for medical and types from Northern Europe, Russia and Siberia. genetical research, because it has, in addition to a countless morphological variety, diseases similar to those of humans (Sutter and Ostrander, 2004; Ostrander and Wayne, 2005). Scientists as nonscientists alike are therefore fascinated by Measurements in Archaeological the history of the domestic dog. Research However, there is still considerable controversy among scientists as to whether the dog has a single or multiple Measurements of bones and teeth have been among the origins, and to where and when the dog was domesticated. most commonly used techniques in osteological research In Eurasia, two principal ancient domestication centres (e.g. Street, 2002; Germonpre´ et al., 2009; Detry and have been identified in which most of the current livestock Cardoso, 2010; Napierala and Uerpmann, 2010). Since species were domesticated: The Middle East (mainly domestication was accompanied by a decrease in body Fertile Crescent) and China/Southeast-Asia (Diamond, size (Davis, 1981; Dayan, 1994), somewhat smaller fossil 2002; Bruford et al., 2003). However, the dog is not a remains have been interpreted as being those of domestic typical domestic animal, because it was probably not dog. However, two of the earliest findings, from Israel domesticated solely as livestock and it is a carnivore and (Davis and Valla, 1978) and Bonn-Oberkassel in Germany therefore might have been domesticated outside the clas- (Table 1; Street, 2002; Napierala and Uerpmann, 2010) are sical domestication centres. Here, we review the different from subadult specimens. Therefore, smaller body size in lines of evidence (the archaeological record and genetical these cases is possibly related to younger age rather than studies), discuss the advantages and limitations of each evidence for domestication. Further, Napierala and approach, and propose a probable scenario for dog origins. Uerpmann (2010) point out that the finds from, for example Eliseevichi I, and Kniegrotte (Table 1) cannot be clearly assigned to domestic dogs, because tooth meas- Archaeological Record urements (upper carnassials, P4) are within the range of variability found in teeth size of wolf populations. In Studies of the archaeological record may give insights contrast, the Kesslerloch Cave find is clearly smaller. into the history of domestication and potentially deliver Furthermore, the Eliseevichi I find from western Russia, answers about ‘where’, ‘when’ and ‘how many times’ in layers C-14 dated to 13 000–17 000 BP (Sablin and domestication occurred. However, the archaeological Khlopachev, 2002) are questioned by Wang and Tedford record for dogs is fragmentary (Napierala and Uerpmann, (2008) since only some of the morphological characteristics 2010) and only a few fossils considered to be remains of are dog-like (similar to a Siberian husky), while others are potential early domesticated dogs have been described more wolf-like. The same applies to the very early Belgian (Table 1). Further, there is an ongoing intensive discussion canid (Table 1; Germonpre´ et al., 2009). Napierala and about the question: Is this fossil really a dog or is it a small Uerpmann (2010) emphasise that the find by Germonpre´ ancient wolf? Naturally, the further we go back in time, the et al. (2009) must therefore be re-evaluated in comparison more similar the early domestic dogs would have been to to unambiguous wolf and dog samples rather than to their ancestors, the wolves (Figure 1; Wang and Tedford, questionable finds from Eliseevichi I (Russia), and other 2008). The situation is further complicated by the fact that findings from Mezin and Mezhirich (Ukraine). Likewise, remains are normally incomplete, consisting only of a few the feature of a broadened snout, which is often taken as a bones or teeth (Wang and Tedford, 2008), which makes a sign of domestication, is an unreliable morphological comprehensive examination impossible. Additionally, character. Napierala and Uerpmann (2010) suggest other there is considerable bias in the geographical distribution influences, such as individual age or changes in predator– of excavations carried out with a strong focus on Europe prey relations, to potentially impact the modification of and the Middle East, giving the perhaps false impression this character. There are additional osteological traits used

2 ENCYCLOPEDIA OF LIFE SCIENCES & 2011, John Wiley & Sons, Ltd. www.els.net Geographical Origin of the Domestic Dog

(a) (b)

(c) (d)

(e) (f) (g)

Figure 2 The domestic dog (Canis familiaris) comes in different shapes and sizes. The (a) Tibetan Mastiff (guarding dog) originates from Central Asia and is a representative of an ancient dog group, the Mastiffs. (b) The Sloughi, a representative of the ancient group of Hounds, originates from (North-) Africa and is traditionally used for hunting. Representatives of modern breeds are the (c) Fox Terrier (fox hunting), (d) Swedish Vallhund (herding/guarding) and (e) Wirehaired Dachshund (hunting). Other presumably old dog breeds include the (f) Bergamasco Sheepdog (herding) and (g) Naked Peruvian Inca Orchid (unclear role; probably connected to spiritual rituals and food), the latter has no fur except for the head hair. (a) Reproduced from http:// commons.wikimedia.org/wiki/File:Candra_Du_Domaine_De_Toundra,_CAC.jpg under the wikimedia commons license; (b) and (c) Copyright Dr. Dominique de Caprona, reproduced with permission; (d) reproduced from http://commons.wikimedia.org/wiki/File:V%C3%A4stg%C3%B6taspets1.jpg under the wikimedia creative commons license; (e) reproduced from http://commons.wikimedia.org/wiki/File:Wire-haired_Dachshund_R_01.JPG under the wikimedia creative commons license; (f) reproduced from http://en.wikipedia.org/wiki/File:Ortensia_di_Valle_Scrivia.jpg under the wikimedia creative commons license and (g) reproduced from http://commons.wikimedia.org/wiki/File:Perro_sin_pelo_del_Per%C3%BA.JPG under the wikimedia creative commons license.

ENCYCLOPEDIA OF LIFE SCIENCES & 2011, John Wiley & Sons, Ltd. www.els.net 3 Geographical Origin of the Domestic Dog

Table 1 An overview of the earliest canid remains proposed to be dog Place Potential date Remarks Reference Bonn-Oberkassel  14 000 BP Subadult specimen; Nobis (1979); Street (2002) (Germany) wolf-sized specimen (Deguilloux et al., 2009) Hayonim Terrace  11 500 BP Two fairly complete canid Tchernov and Valla (1997) (northern Israel) remains Eliseevichi I (Russia) 13 000217 000 BP Canid cranium; wolf-sized Sablin and Khlopachev specimen (Deguilloux et al., (2002) 2009) Kesslerloch cave 14 100214 600 BP Skull fragment includes a Napierala and Uerpmann (Switzerland) large part of the right (2010) maxilla and a small portion of the zygomaticum and the palatinum; adult Goyet (Belgium) 31 680 + 250 BP A ‘fossil large canid’ Germonpre´ et al. (2009) northern Israeli sites of Ein 11 310212 310 BP Three canid finds, Davis and Valla (1978) Mallaha (Eynan) and consisting of a diminutive Hayonim terrace (Israel) carnassial and mandible, and a canid puppy skeleton buried with a human Kniegrotte Cave, Kniegrotte: 12 230 + 90 Several canid remains Musil (2000) Teufelsbru¨ cke Cave, and BP; 13 585 + 165 BP; Oelknitz (Germany) Teufelsbru¨ cke: 12 300 + 85213 025 + 85; Oelknitz: 10 990 + 852 12 545 + 80 BP

Note:BP5 before present.

to distinguish domestic dogs from wolves, for example 1981). Domestication itself is also characterised by crowding of the teeth caused by shortening of the facial decrease in body size, possibly because of selection of region (Clutton-Brock, 1995). However, these traits are smaller individuals in the ancestral population (giving sometimes found also among wolves (Musil, 2000), and the easier handling, and more specimens in a confined space) extent of the variation in the ancient wolf populations is and/or insufficient food supply (Davis, 1981). Thus, with often not well studied. Musil (2000) and Davis (1981) fur- two possible causes for small body size in canids, the dis- ther mention the impact of sexual dimorphism on size tinction between small wolves and early dogs gets even measurements. Since female wolves are on average smaller more complicated. Olsen and Olsen (1977) point out that than male wolves, this fact may add to the difficulties in Chinese wolves (but not larger wolf subspecies) share a distinguishing between wolves and early dogs. It is in fact morphological feature with dogs ‘the turned-back apex of often impossible to find out from a few bones, if the canid the coronoid process of the ascending ramus’, giving an was a female or a male. Further, Musil (2000) writes that all indication that dogs originated from Chinese wolves. A few potential dog characteristics (reduced body size, shorter studies by Jing (2008) and others (Underhill, 1997; Jin and snout and facial parts of the skull, crowded teeth, reduced Xu, 1992) pushes the earliest record for domestic dog to brain volume) can develop within a few generations in wild 10 000 years before present in North China. However, wolves kept in captivity. Finally, other authors (Davis and relatively little archaeological work has been performed Valla, 1978; Davis, 1981; Dayan, 1994) discuss the prob- and the soil in many regions of South China is acidic. lematic use of body size and associated measurements from Pottery and stone objects can survive these acid conditions, a different perspective. Temperature rose at the end of the but animal remains are highly degraded. In addition, in Pleistocene, resulting in a decrease in body size of wolves previous excavations in South China, archaeologists often and other species around 10 000–12 000 years BP (Davis, considered only wolves and not dogs to be present and as a

4 ENCYCLOPEDIA OF LIFE SCIENCES & 2011, John Wiley & Sons, Ltd. www.els.net Geographical Origin of the Domestic Dog result, many dog remains may have been incorrectly Genetical Studies assigned to Canis lupus instead to Canis familiaris (Jing Y, personal communication). Further excavations in South Another way to tackle the domestication history of the dog China may shed light on the earliest domestication history and other domesticates, is the use of genetical markers, of dogs in this region. such as mitochondrial deoxyribonucleic acid (mtDNA; To conclude, the archaeological record is very difficult to Table 2). This marker has been widely used to study interpret and there is considerable controversy about the domestication (e.g. pig: Larson et al., 2010; horse: Lei et al., authenticity of many of the earliest findings proposed to be 2009; dog: Savolainen et al., 2002; Pang et al., 2009). dog. However, from around 10 000 years ago, the arch- In principal, these analyses involve comparing mtDNA aeological findings of dogs and especially burials (Morey, haplotypes among all dogs and possibly also wolves, 2006) become numerous and an established relationship worldwide. If all dog populations have a similar set of between humans and dogs seems unambiguous. Therefore, haplotypes, this would indicate a single common origin of the archaeological record shows that domestic dogs existed all dog populations, and the geographical region with the 10 000 years ago, and this implies that the domestication most complete set of haplotypes would be considered event might have been a few thousand years earlier. the likely origin. Derived populations would have only a However, where and when exactly the dog was domesti- subset of all haplotypes since some haplotypes would have cated remains unclear based on the archaeological record been lost as dog spread all over the world. If instead dog alone. populations throughout the world would have largely

Table 2 Overview of the different markers and approaches used in the study of dog domestication Marker/type of study Inheritance Properties/advantages Disadvantages mtDNA Maternal – High copy number; therefore usable – Can be poor estimator of total for all kind of tissues including genomic diversity, because limited to ancient/low-quality DNA material one locus – Highly variable (  10 Â higher than – No information about nuclear and nuclear genome) paternal variation (mtDNA is an – Haploid; phylogenetic analysis is extra-nuclear genetic marker) relatively straightforward (Bruford et al., 2003) Y-chromosome Paternal – Haploid; analysis of haplotypes – Low diversity straightforward – Often used as additional marker to complement mtDNA data SNP Autosomal/ – Next-generation sequencing makes it – Have to be recovered from a biparental possible to recover a high number of representative sample of the SNPs in the genome population, otherwise ascertainment – Gives information about nuclear bias can affect analysis severely genome (Albrechtsen et al., 2010) Modern DNA NA – High sample numbers – Past events such as replacement of studies – Thereby, giving a detailed picture of lineages may be hard to detect today’s genetic variation and reconstruction of past events possible to a certain degree aDNA NA – Additional estimates about original – Often few samples genetic diversity – Error-prone due to high – May give temporal change of genetic contamination risk (Malmstro¨ m diversity estimates (e.g. identification et al., 2005; Linderholm et al., 2008) of hybridisation and/or local and low-quality DNA domestication leading to partial or – High conservation variability of total replacement of lineages) DNA among geographical locations and different body parts

Note:NA5 not applicable.

ENCYCLOPEDIA OF LIFE SCIENCES & 2011, John Wiley & Sons, Ltd. www.els.net 5 Geographical Origin of the Domestic Dog nonoverlapping gene pools, then this would indicate sev- dingoes originate from domestic dogs in Southeast Asia, eral origins from different wolf populations. that they had a very small founder population, and have The most comprehensive study based on mtDNA to lived largely isolated from other dog populations for sev- date, including 1543 dogs from all over Eurasia and Africa eral thousand years. Interestingly, the New Guinea Singing (Figure 3 and Figure 4; Pang et al., 2009), is the only genetic dog (NGSD), another wild dog living in the New Guinea study with a global sampling of mostly nonbreed dogs. highlands, share mtDNA haplotype A29 with the dingo This study indicates that all dogs worldwide have a single and dogs from Southeast Asia, and also some alleles of origin from Southeast Asia, South of the Yangtze River another genetical marker, the Dog leucocyte antigen class (Figure 4b), less than 16 000 years ago. The study focused on II (Runstadler et al., 2006) as well as some morphological dogs only, because it is unclear to date, if the contemporary and behavioural characteristics (Koler-Matznick et al., wolf populations are representative for the wolf popu- 2003). Therefore, it seems likely that these two populations lations at the time of domestication. have a common origin from a migration of early domestic It showed that almost 100% of dogs in all regions share dogs from Southeast Asia. haplogroups A, B and C (Figure 3a and Figure 4a), indicating At least three events of subsequent local dog–wolf a single geographical origin. Genetic variation is clearly hybridisation are also indicated by mtDNA data, one each highest South of the Yangtze River in China (Figure 3b and in Japan (forming haplogroup F), Middle East (haplogroup Figure 4b). Outside this region, all other regions (Africa, d2) and Scandinavia (haplogroup d1; Klu¨ tsch et al., 2011). Europe, Middle East and even North China) show only a These minor subhaplogroups consist of one or only a few subset of genetic diversity in terms of unique haplotypes haplotypes and are found only regionally which is why dog– and haplogroups (Figure 3b and Figure 4b). For example, all wolf hybridisation is their most likely source. Additionally, 10 principal subhaplogroups of haplogroups A, B and C there is one possible hybridisation event identified in North among dogs are found only South of Yangtze River, but in America by Koop et al. (2000). In their phylogenetic tree, China North of Yangtze River only 7 haplogroups are sequences of ‘ancient’ (a couple of hundred years old) found, in SW Asia and Africa 5, and in Europe 4. This samples of the North American Tahltan dog and Wool dog indicates that the full set of 10 haplogroups originated clustered with a wolf sample from Yukon in a separate South of Yangtze River and when dogs migrated to other group, indicating a close phylogenetic relationship between parts of the world, only a subset of the subhaplogroups was the two indigenous dog types and the North American wolf. spread. Importantly, dogs worldwide have the same subset Therefore, the authors concluded that a separate local of haplogroups, indicating a scenario that includes an domestication might have occurred in North America. expansion from a single region of origin. The improved However, hybridisation of Old World dogs with a New sampling (in comparison to earlier studies by Vila` et al., World wolf is a more probable scenario, considering that no 1997 and Savolainen et al., 2002) and analysis of whole evidence for a separate domestication has been found in mtDNA genomes refined molecular dating for the dog North American archaeological sites (Koop et al., 2000) origin which is now more in agreement with archaeological and that in genetical analyses of ancient and modern dog data (e.g. Jing, 2008; Davis and Valla, 1978; Street, 2002) samples, Leonard et al. (2002) found that all New World than older estimates by Vila` et al. (1997). Further, the data dogs have an origin from Old World wolves. suggest that a high number of female wolves (  50 or more) Importantly, hybridisation is not part of de novo were domesticated, for the first time giving some insight domestication, but may nevertheless contribute to the into the earliest stages of dog domestication. gene pool of both species. Hybridisation between dogs Savolainen et al. (2004) investigated the origins of the and wolves as an intentional act by humans is part of dingo (Figure 5), a feral dog living in Australia, as an the domestication history of dogs, but still not domesti- example for an early migration from the centre of domes- cation de novo. The distinction between hybridisation and tication in Southeast Asia. In their study, 211 Australian domestication might be challenging in different areas of dingoes sampled in all states of Australia were sequenced research and ancient DNA (aDNA) analysis may help to for the mtDNA control region and compared to sequences establish original genetic diversity in both wolves and dogs from 676 dogs from all continents and 38 Eurasian wolves at a given locality before hybridisation. Thus, aDNA may (Savolainen et al., 2002). All sequences among dingoes give more detailed insights into temporal local changes, were either identical to or differing by a single substitution including hybridisation and/or local domestication events. from one mtDNA haplotype, called A29. This haplotype Notably, these studies of modern and ancient dogs do not was found in more than 50% of all sampled dingoes and is imply that there have been no additional attempts some- otherwise found only among dogs in Southeast Asia, where in the world to breed wolves in captivity, and ultim- Siberia and Arctic America. Of 19, 18 other identified ately domesticate them. However, if this is the case, they have haplotypes were found in dingoes only, indicating that not contributed to the mtDNA gene pool of modern dogs, these haplotypes derived from the central A29 haplotype. possibly because the population went extinct at some point of Based on the mean genetic distance of these derived dingo time. Additionally, mtDNA displays only the maternal side sequences to A29, the arrival of the dingo in Australia was of the story and other markers, such as the Y-chromosome calculated to 5000 YBP which is roughly in congruence (paternally inherited; Natanaelsson et al., 2006) and auto- with the archaeological record. This study suggests that somal SNPs are needed to complement the picture.

6 ENCYCLOPEDIA OF LIFE SCIENCES & 2011, John Wiley & Sons, Ltd. www.els.net Geographical Origin of the Domestic Dog

East Asia West Europe A A A

8 3 3

55 44 18

C C C B B B

A

SW Asia N China C China A A A

C C C B B B

F Sib,Mon,Kor,Jap ASY A A A B 3 22

E

C C C C B B B D

Coy (a) (b) (c)

Figure 3 Genetic relationships between the mtDNA CR sequences (582 bp). (a) Phylogenetic tree for dog (unlabeled), wolf (filled square) and coyote (Coy) haplotypes. The six main phylogenetic haplogroups (A–F) of dog haplotypes, discussed in the text, are indicated. Note the close relation between dog and wolf, compared to coyote (the closest wild relative of wolf), clearly indicating that dogs originate from wolf. (b) Minimum spanning networks showingthe relationships between the haplotypes in dog haplogroup A, B and C, and the representation in different geographical regions. The haplotypes are symbolised by circles (coloured or white), and are separated by one mutational step; black dots are hypothetical intermediates. Fourteen virtually universally occurring haplotypes (‘Universal Types’: UTs) are indicated with bold lining. The representation of haplotypes in the geographical regions is shown by the colour and size of the circles. A coloured circle indicates the presence of the haplotype (blue for haplotypes shared with other regions and orange for haplotypes unique to the region); white denotes nonrepresented haplotypes. The size of the circle is proportional to the frequency of the haplotype in the region. Note the difference in representation between the eastern (East Asia) and western (West: Europe, SW Asia and Africa) parts of the Old World, and that Asia South of Yangtze River (ASY) is the only subregion with virtually complete representation in all parts across the networks. (c) The 10 principal subhaplogroups in haplogroups A, B and C (six for haplogroup A, two each for B and C), identified by analysis of the entire mtDNA genomes for the samples indicated by colour. Haplogroup A: subclade a1 orange, a2 green, a3 red (upper part of network), a4 red (lower part of network), a5 blue and a6 yellow. B: subclade b1 orange and b2 green. C: subclade c1 green and c2 orange. See geographical representation of the 10 subhaplogroups in Figure 4, compare also with Figure 3b. Reproduced from Pang et al. (2009) with permission from Oxford University Press.

ENCYCLOPEDIA OF LIFE SCIENCES & 2011, John Wiley & Sons, Ltd. www.els.net 7 Geographical Origin of the Domestic Dog

A BC Britain UT: 73.1 UTd: 99.1 Siberia SC: - UT: 46.7 UTd: 75.0 SC: a2 Eur N Cont UT: 94.5 N China UTd: 100 SC: - UT: 79.6 UTd: 89.8 Eur S Cont SC: a2 UT: 75.4 SW Asia UTd: 96.5 C China SC: - UT: 81.5 UT: 70.9 UTd: 94.6 UTd: 85.8 Japan SC: b2 SC: a2, a4 UT: 58.5 a5 UTd: 84.7 Africa SC: a2, a4 UT: 57.1 b2 UTd: 91.1 ASY SC: a2 India UT: 40.8 UT: 54.2 UTd: 53.0 Utd: 81.4 SC: a2, a3, SC: a2 a4, a5, (a) a6, b2

Heilongjiang UT: 90.4 UTd: 90.4 SC: a2

LHS UT: 67.4 UTd: 89.1 Liaoning SC: a2 Yellow River Korea UT: 63.3 UTd: 77.8 Hebei SC: a2, a4, Shanxi b2 UT: 86.8 UTd: 93.4 SC: a2, a4 Japan Shaanxi UT: 58.5 UTd: 84.7 Tibet Sichuan Yangtze River SC: a2, a4, UT: 35.1 UT: 39.6 b2 UTd: 51.4 UTd: 70.8 SC: a2, a4 SC: a2, a5 Hunan UT: 44.4 Jiangxi UT: 37.0 Guizhou UTd: 48.1 UTd: 47.8 UT: 47.4 SC: a2, a3 SC: a2, a5, UTd: 64.9 b2 UT: 54.7 SC: a2, a3, UTd: 65.3 a5, b2 GG Guangdong SC: a2, a3 Yunnan Guangxi UT: 18.4 a5, a6, UTd: 32.7 b2 Vietnam SC: a2, a3, a5, b2 SE Asia UT: 35.1 UTd: 50.9 Thailand SC: a2, a3, a4, a5, (b) Cambodia b2

8 ENCYCLOPEDIA OF LIFE SCIENCES & 2011, John Wiley & Sons, Ltd. www.els.net Geographical Origin of the Domestic Dog

Genome-wide SNP Data A recent study by vonHoldt et al. (2010) aimed at deter- mining the primary centre of dog domestication by a gen- ome-wide SNP survey of pure-bred dogs (mainly from Europe/the United States; 66/85 breeds) and wolves from Europe, the Middle East and North/Central China. They found highest haplotype sharing between modern dogs and wolves from the Middle East, and therefore suggested the ‘Middle East as a primary source of genetic variation in the dog’ (vonHoldt et al., 2010). This clearly contradicts the comprehensive studies of mtDNA discussed earlier, which suggest that dogs were domesticated in Southeast Asia. However, the SNP study suffers a similar bias as the archaeological record, namely a lack of representative samples, for both dogs and wolves, from many parts of the world, and especially from southern East Asia. Clearly, the lack of samples from southern East Asia may explain the discrepancy between the results of vonHoldt et al. (2010) and those of Pang et al. (2009). Actually, if dogs originated in southern East Asia, this could not have been detected by this study, since no wolves from southern East Asia were analysed. Furthermore, the source for SNPs used for the analyses was almost exclusively European breed dogs, and not a single dog or wolf from southern East Asia was included. Therefore, the SNPs were recovered from only a small part of the total gene pool of dogs. As a result, the SNPs used are likely to suffer a phenomenon called ascertainment bias (e.g. Morin et al., 2004; Albrechtsen et al., 2010) and hence are not suitable for an objective comparison of genetic diversity in dogs and wolves worldwide. To reconcile the results of the different genetic studies, it is vital to recall that mtDNA data indicated regional hybridisations of dogs and wolves, one of them in the Middle East. Hence, there is a possibility that the signal picked up by the SNP survey of vonHoldt et al. (2010) may result from dog–wolf admixture rather than ancestry of dogs in Southwest Asia. Autosomal (e.g. SNP) markers are undoubtedly an important tool for future investigations, but studies of the origins of dogs must be based on samples with practically complete geographical coverage. Figure 5 (a) and (b) The dingo, an early domesticated, now feral, dog indigenous to Australia. Image courtesy of Sam Fraser-Smith. To conclude, mtDNA gives a very distinct indication Ancient DNA Studies that dogs originated in a single region, Southern East Asia, To study early domestication history and migration routes 10 000–16 000 years ago, and that the population is based from southern East Asia in more detail, it would be crucial on a founder population of at least 50 wolf females and to look at ancient DNA of early dog and early wolf some subsequent regional hybridisation events. remains and compare these with contemporary samples.

Figure 4 Genetic diversity for CR data, among regions across the Old World. (a) Genetic diversity across the Old World. Pie diagrams show the proportion of individuals having haplogroups A (blue), B (red) and C (yellow). Note the similar proportions of A, B and C among regions, which indicates that haplogroups A, B and C (and consequently dogs) originated from a single geographical region. Boxes show (i) UT: the proportion of individuals having one of the 14 ‘universally occurring’ haplotypes of haplogroups A, B and C. (ii) UTd: the proportion having a UT-derived haplotype, i.e. a haplotype which is either a UT or differs by a single mutation from a UT. Note that 95–100% of the dogs in Europe and SW Asia have a UTd, while ASY has only 47% UTd, a consequence of the difference in representation across the networks in Figure 3b. (iii) SC: the representation of the 6 nonuniversal (the other 4 are found in all populations) of the totally 10 subhaplogroups of haplogroups A, B and C (see the 10 groups in Figure 3c). Note that only ASY harbours all 10 subhaplogroups. (Eur N/S cont – Europe North/South continent; N/C China – North/Central China [north of Yellow River/between Yellow and Yangtze River]; ASY – Asia South of Yangtze River). (b) Genetic diversity in Southeastern Asia. Boxes show UT, UTd and SC as defined in Figure 5a. Sampled regions are indicated by colour. LHS – Liaoning, Hebei and Shanxi; GG – Guanxi and Guangdong. SE Asia – Vietnam, Thailand and Cambodia. Tibet includes samples from Qinghai and Nepal. Reproduced from Pang et al. (2009) with permission from Oxford University Press.

ENCYCLOPEDIA OF LIFE SCIENCES & 2011, John Wiley & Sons, Ltd. www.els.net 9 Geographical Origin of the Domestic Dog

For example, this may show whether the earliest dogs haplotypes are under-represented (only one canid each in universally had the same mix of haplogroups A, B and C as Verginelli et al., 2005; Leonard et al., 2002). Whether this found among today’s dogs. This is imperative also because under-representation of clade B haplotypes in ancient today’s wolf population represents only a fraction of the samples is caused by sample bias or mirrors a real pattern wolf population that existed in the past. Consequently, warrants investigation of a larger number of ancient DNA genetic diversity is probably reduced and not represen- materials. Nevertheless, the most important finding is that tative of earlier wolf populations and in many regions ancient and modern dogs carry the same three major the wolf populations are completely exterminated (Gao, haplogroups A, B and C, confirming a genetic continuity 2006). Also, because there is increasing evidence of through time. Hence, DNA data in ancient and modern recent hybridisation between dogs and wolves in the wild dogs do not support the hypothesis of lineage replacement, (Wronski and Macasero, 2008; Anderson et al., 2009; that is, ancient European and American dogs having a Iacolina et al., 2010), analysis of ancient wolf samples may different genetic makeup (different haplogroups) than be necessary for reconstructing the original genetic diver- modern European/American dogs. sity in prehistoric wolves. Likewise, unresolved issues such as the origin of haplogroup D, including subhaplogroups d1 (mainly found in Scandinavia; Klu¨ tsch et al., 2011) and Conclusion d2 (only found in the Middle East and along the Medi- terranean Sea; Pang et al., 2009; Klu¨ tsch et al., 2011) could Recent mtDNA studies of modern dog samples indicate a be tackled by ancient DNA approaches. So far, neither single origin of dogs in southern East Asia, 10 000–16 000 contemporary wolves nor historical/ancient dog and wolf years ago, and ancient mtDNA studies showing a similar samples were found to carry haplotypes from these sub- mix of haplogroups A, B and C as modern dogs do not haplogroups (Flagstad et al., 2003; Malmstro¨ m et al., contradict these findings. The mtDNA data also indicate 2008). that at least 50 and potentially hundreds of wolves were In order to determine the origin of North American domesticated. Two other lines of evidence, the archaeo- dogs, Leonard et al. (2002) analysed mtDNA for 24 pre- logical record and the genome-wide SNP data set by von- European dogs collected from archaeological sites in Holdt et al. (2010), have mostly indicated that either the Alaska, Mexico, Peru and Bolivia. With this approach, the Middle East and/or Europe might have been geographical problem that modern European dogs might have contrib- regions of dog origins. However, both datasets are geo- uted to the gene pool of modern New World dogs was graphically biased, having no or only a few samples from avoided. These sequences were compared to sequence data southern East Asia. Therefore, if the dog origin is in of modern wolves and dogs. This showed that native southern East Asia, this would have gone undetected by American dogs originate from multiple Old World dog these studies. lineages (haplogroups A, B and C), indicating that all Although the details of the earliest dog domestication ancient and modern dogs worldwide have a common origin history are not yet completely clear, recent studies have from Old World wolf ancestors. established a plausible scenario. Further studies of geo- Additional ancient DNA studies of dogs have focused on graphically comprehensive modern sample collections as local comparisons of ancient dog and wolf samples in well as key archaeological specimens, using additional Europe. Deguilloux et al. (2009) retrieved three haplotypes, genetical markers, for example Y-chromosome sequence all belonging to the previously identified haplogroup C and autosomal SNPs, promises a precise picture of how the (Vila` et al., 1997; Savolainen et al., 2002; Pang et al., 2009), domestic dog originated and subsequently spread around from one location (Villeneuve–Tolosane) in Southeast the world. See also: Ancient DNA: Phylogenetic Appli- France from the Middle Neolithic (6000–5800 YBP). In a cations; Ancient DNA: Recovery and Analysis; Chromo- study of 24 ancient dog samples from Sweden (Malmstro¨ m some Y; History of Scientific Agriculture: Animals; et al., 2008), including dogs from Neolithic (  5300–4500 Microsatellite Instability; Microsatellites; Mitochondrial YBP) and Medieval sites (  1000–500 YBP), haplogroups DNA Polymorphisms; Mitochondrial Genome; Single A and C were found. Finally, Verginelli et al. (2005) ana- Nucleotide Polymorphism (SNP); Y Chromosome lysed five Italian prehistoric canid remains. Two dog samples from the Eneolithic and the Bronze age (from 4110+40 YBP to 3040+40 YBP) had haplotypes from References haplogroups A and C, in agreement with the findings of Albrechtsen A, Nielsen FC and Nielsen R (2010) Ascertainment Malmstro¨ m et al. (2008) and Deguilloux et al. (2009). The bias in SNP chips affects measures of population divergence. other three canids could not be definitely identified as dogs Molecular Biology and Evolution 27(11): 2534–2547. or wolves. Two samples from around 10 000 YBP had Anderson TM, vonHoldt BM, Candille SI et al. (2009) Molecular haplogroup A and B haplotypes, and the third from around and evolutionary history of melanism in North American gray 15 000 YBP a haplotype from haplogroup C. Therefore, wolves. Science 323(5919): 1339–1343. ancient DNA studies suggest that in the Middle Neolithic, Bruford MW, Bradley DG and Luikart G (2003) DNA markers haplogroups A and C were well established in European reveal the complexity of livestock domestication. Nature as well as North American dogs. Remarkably, clade B Reviews. Genetics 4: 900–910.

10 ENCYCLOPEDIA OF LIFE SCIENCES & 2011, John Wiley & Sons, Ltd. www.els.net Geographical Origin of the Domestic Dog

Clutton-Brock J (1995) Origins of the dog: domestication and Lei CZ, Su R, Bower MA et al. (2009) Multiple maternal origins of early history. In: Serpell J (ed.) The Domestic Dog, Its Evolution, native modern and ancient horse populations in China. Animal Behavior and Interactions With People, pp. 7–20. Cambridge: Genetics 40(6): 933–944. Cambridge University Press. Leonard JA, Wayne RK, Wheeler J et al. (2002) Ancient DNA Davis SJM (1981) The effects of temperature change and evidence for Old World origin of New World dogs. Science domestication on the body size of Late Pleistocene to Holocene 298(5598): 1613–1616. mammals of Israel. Paleobiology 7(1): 101–114. Lindblad-Toh K, Wade CM, Mikkelsen TS et al. (2005) Genome Davis SJM and Valla FR (1978) Evidence for domestication of sequence, comparative analysis and haplotype structure of the the dog 12,000 years ago in the Natufian of Israel. Nature 276: domestic dog. Nature 438: 803–819. 608–610. Linderholm A, Malmstro¨ m H, Lide´ nKet al. (2008) Cryptic Dayan T (1994) Early domesticated dogs of the Near East. Journal contamination and making phylogenetic nonsense. PLoS ONE of Archaeological Science 21: 633–640. 3(5): e2316. Deguilloux MF, Moquel J, Pemonge MH and Colombeau G Malmstro¨ m H, Stora˚ J, Dale´ nLet al. (2005) Extensive human (2009) Ancient DNA supports lineage replacement in European DNA contamination in extracts from ancient dog bones dog gene pool: insight into Neolithic southeast France. Journal and teeth. Molecular Biology and Evolution 22(10): 2040–2047. of Archaeological Science 36(2): 513–519. Malmstro¨ m H, Vila` C, Gilbert MT et al. (2008) Barking up the Detry C and Cardoso JL (2010) On some remains of dog wrong tree: Modern northern European dogs fail to explain (Canis familiaris) from the Mesolithic shell-middens of their origin. BMC Evolutionary Biology 8: 71. Muge, Portugal. Journal of Archaeological Science 37(11): Morey DF (2006) Burying key evidence: the social bond 2762–2774. between dogs and people. Journal of Archaeological Science Diamond J (2002) Evolution, consequences and future of plant 33(2): 158–175. and animal domestication. Nature 418: 700–707. Morin PA, Luikart G, Wayne RK and the SNP workshop group Flagstad Ø, Walker CW, Vila` C et al. (2003) Two centuries of the (2004) SNPs in ecology, evolution and conservation. Trends in Scandinavian wolf population: patterns of genetic variability Ecology and Evolution 19(4): 208–216. and migration during an era of dramatic decline. Molecular Musil R (2000) Evidence for the domestication of wolves in central Ecology 12(4): 869–880. European Magdalenian sites. In: Dogs through Time: An Gao ZX (2006) Review of the research on wolf in China. Chinese Archaeological Perspective (ed. Susan J. Crockford). BAR Journal of Zoology 41: 134–136, (in Chinese). International series 889: 21–28. Germonpre´ M, Sablin MV, Stevens RE et al. (2009) Fossil dogs Napierala H and Uerpmann H-P (2010) A ‘New’ palaeolithic dog and wolves from Palaeolithic sites in Belgium, the Ukraine and from Central Europe. International Journal of Osteoarchaeol- Russia: osteometry, ancient DNA and stable isotopes. Journal ogy, early view [Epub ahead of print]. of Archaeological Science 36(2): 473–490. Natanaelsson C, Oskarsson MC, Angleby H et al. (2006) Dog Y vonHoldt BM, Pollinger JP, Lohmueller KE et al. (2010) Gen- chromosomal DNA sequence: identification, sequencing and ome-wide SNP and haplotype analyses reveal a rich history SNP discovery. BMC Genetics 7: 45. underlying dog domestication. Nature 464(7290): 898–902. Nobis G (1979) Der a¨ lteste Haushund lebte vor 14000 Jahren. Iacolina L, Scandura M, Gazzola A et al. (2010) Y-chromosome Umschau 79: 610. microsatellite variation in Italian wolves: a contribution to the Nowak RM (2003) Wolf evolution and taxonomy. In: Mech LD study of wolf-dog hybridization patterns. Mammalian Biology and Boitani L (eds) Wolves: Behavior, Ecology, and Conser- 75(4): 341–347. vation, pp. 239–258. Chicago: The University of Chicago Press. Jin J and Xu H (1992) Opinions about the early Neolithic site of Olsen SJ and Olsen JW (1977) The Chinese wolf, ancestor of New Nanzhuangtou at Xushui. Kaogu 11: 1018–1022, (In Chinese). World dogs. Science 197(4303): 533–535. Jing Y (2008) The origins and development of animal domesti- Ostrander EA and Wayne RK (2005) The Canine genome. Gen- cation in China. Chinese Archaeology 8: 1–7. ome Research 15: 1706–1716. Klu¨ tsch CFC, Seppa¨ la¨ EH, Lohi H et al. (2011) Regional occur- Pang JF, Kluetsch C, Zou XJ et al. (2009) mtDNA data indicate a rence, high frequency but low diversity of mitochondrial DNA single origin for dogs South of Yangtze River, less than 16,300 haplogroup d1 suggests a recent dog-wolf hybridization in years ago, from numerous wolves. Molecular Biology and Scandinavia. Animal Genetics 42(1): 100–103. Evolution 26(12): 2849–2864. Koler-Matznick J, Lehr Brisbin I Jr, Feinstein M and Bulmer S Runstadler JA, Angles JM and Pedersen NC (2006) Dog leucocyte (2003) An updated description of the New Guinea singing dog antigen class II diversity and relationships among indigenous (Canis hallstromi, Troughton 1957). Journal of Zoology, London dogs of the island nations of Indonesia (Bali), Australia and 261: 109–118. New Guinea. Tissue Antigens 68(5): 418–426. Koop BF, Burbridge M, Byun A et al. (2000) Ancient DNA Sablin MV and Khlopachev GA (2002) The earliest ice age dogs: evidence of a separate origin for North American indigenous evidence from Eliseevichi I. Current Anthropology 43(5): 795–799. dogs. In: Dogs through Time: An Archaeological Perspective Savolainen P, Leitner T, Wilton AN et al. (2004) A detailed pic- (ed. Susan J. Crockford). BAR International series 889: ture of the origin of the Australian dingo, obtained from the 271–285. study of mitochondrial DNA. Proceedings of the National Larson G, Liu R, Zhao X et al. (2010) Patterns of East Asian pig Academy of Sciences of the USA 101(33): 12387–12390. domestication, migration, and turnover revealed by modern Savolainen P, Zhang YP, Luo J, Lundeberg J and Leitner T (2002) and ancient DNA. Proceedings of the National Academy of Genetic evidence of an East Asian origin of dogs. Science Sciences of the USA 107(17): 7686–7691. 298(5598): 1610–1613.

ENCYCLOPEDIA OF LIFE SCIENCES & 2011, John Wiley & Sons, Ltd. www.els.net 11 Geographical Origin of the Domestic Dog

Street M (2002) Ein Wiedersehen mit dem Hund von Bonn China. Proceedings of the National Academy of Sciences of the Oberkassel. Bonner zoologische Beitra¨ge 50(3): 269–290, (in USA 106(14): 5523–5528. German). Benecke N (1987) Studies on early dog remains from Northern Sutter NB and Ostrander EA (2004) Dog star rising: the canine Europe. Journal of Archaeological Science 14(1): 31–49. genetic system. Nature Review Genetics 5(12): 900–910. Benecke N (1993) Zur Kenntnis der mesolithischen Hunde des Tchernov E and Valla FF (1997) Two new dogs, and other su¨ dlichen Ostseegebietes. Zeitschrift fu¨r Archa¨ologie 27: 39–65, Natufian dogs, from the Southern Levant. Journal of Arch- (in German). aeological Science 24(1): 65–95. Ishiguro N, Inoshima Y and Shigehara N (2009) Mitocho- Underhill AP (1997) Current issues in Chinese Neolithic ndrial DNA analysis of the Japanese wolf (Canis Lupus Hodo- Archaeology. Journal of World Prehistory 11(2): 103–160. philax Temminck, 1839) and comparison with representative Verginelli F, Capelli C, Coia V et al. (2005) Mitochondrial DNA wolf and domestic dog haplotypes. Zoological Science 26(11): from prehistoric Canids highlights relationships between dogs 765–770. and Southeast European wolves. Molecular Biology and Evo- Morey DF (1994) The early evolution of the domestic dog. lution 22(12): 2541–2551. American Scientist 82(4): 336–347. Vila` C, Savolainen P, Maldonado JE et al. (1997) Multiple and Ostrander EA, Giger U and Lindblad-Toh K (eds) (2006) The Dog ancient origins of the domestic dog. Science 276(5319): 1687–1689. and Its Genome. Cold Spring Harbor Monograph Series 44, Wang X and Tedford RH (2008) Dogs, their Fossil Relatives and 584pp. Cold Spring Harbour, NY: Cold Spring Harbour Evolutionary History. New York: Columbia University Press. Laboratory Press. Wayne RK (1993) Molecular evolution of the dog family. Trends Tanabe Y (1991) The Origin of Japanese dogs and their in Genetics 9(6): 218–224. association with Japanese people. Zoological Science 8: Wronski T and Macasero W (2008) Evidence for the persistence of 639–651. Arabian Wolf (Canis lupus pallipes) in the Ibex Reserve, Saudi Trut LN (1999) Early Canid domestication: The Farm-Fox Arabia and its preferred prey species. Zoology in the Middle East experiment. American Scientist 87(2): 160–169. 45: 11–18. Trut LN, Plyusnina IZ and Oskina IN (2004) An experiment on Zeder MA, Emshwiller E, Smith BD and Bradley DG (2006) fox domestication and debatable issues of evolution of the dog. Documenting domestication: the intersection of genetics and Russian Journal of Genetics 40(6): 644–655. archaeology. Trends in Genetics 22(3): 139–155. Vigne JD (2005) L’humerus de chien magdale´ nien de Errala (Gipuzkoa, Espagne) et la domestication tardiglaciaire du loup en Europe. Munibe (Antropologia-Arkeologia) 57(1): 279–287, Further Reading (in French). Barton L, Newsome SD, Chen F-H et al. (2009) Agricultural origins and the isotopic identity of domestication in northern

12 ENCYCLOPEDIA OF LIFE SCIENCES & 2011, John Wiley & Sons, Ltd. www.els.net