Mitochondrial DNA Indicates a Single Origin to the Dog in Southern China 5.400 to 16.300 Years Ago
Summary by Dr. Dominique de Caprona
© de Caprona 2009

Arctic Wolf (Canada) (Canis lupus arctos) © R. Schmode  ~  Eurasian Wolf (Canis lupus lupus) © G.Volker

Russian wolves (Canis lupus communis) © V.Bologov  ~  Arabian Wolf (Canis lupus arabs) critically endangered © S. Rerucha

All photographs copyrighted to their photographer. Please do not use for any purpose without asking.

In their article entitled "mtDNA Data indicates a Single Origin for Dogs South of the Yangtze River, less than 16,300 Years Ago, from Numerous Wolves", Pang et al. (2009) analyse the mitochondrial DNA (mtDNA) of a large sample of dogs to determine when, where and how the wolf was domesticated to become the domestic dog. 
The study of samples from ancient dogs have shown that ancient and modern dogs share the same mtDNA, and that American dogs descend from the Old World dogs, hence the focus of this study on the Old World dogs.

An indepth survey of the patterns of geographical diversity for dog mtDNA was carried out by anlysing 582bp of the Control Region for 1.543 domestic dogs across the Old World. Because it had been earlier indicated that the origin of dogs was somewhere in East Asia (Savolainen et al 2002), this area was densely sampled. 
This study includes samples from Europe (North and South): Britain, Scandinavia, Miscellaneous; Southwest Asia:Israel, Iran, Turkey, Miscellanous; Africa: North Africa, Central Africa, South Africa, Miscellaneous;South East Asia: Thailand, Vietnam, Cambodia; North China (Heilongjian, Liaoning, Hebei, Shanxi); Central China (Shaanxi, Shichuan); South China (Guangdong, Guanhxi, Hunan, Guizhou, Jiangxi, Yunnan; Tibet (Quinghai and Nepal) and Hainansanya. Samples were taken on dogs in rural areas as they were assumed to have little influx of foreign dogs or on dogs which belonged to breeds with known historical origins. Most of the "non-breed" dogs had specialized morphology and were kept/bred for specific purposes (guard, pet, herding, meat), they were not typical strays. 
The wolf samples analysed for the 582bp region were from Afghanistan, Canada, China, Mongolia , Romania, Russia, Saudi Arabia, and Yugoslavia.

Based on this survey, a subsample of 169 dogs representative across mtDNA diversity were selected for an analysis of almost the entire mtDNA genome.
The samples represented 5 of the 6 main phylogenic groups or clades: A,B,C,D,F and E. The study focuses on the first 3, as the clades D, E,and F are rare, geographically restricted and could be derived from post-domestication wolf-dog hybridisation.

The 2 major dog clades, A and B, contained wolf haplotypes. 
1) Clade A had 3 haplotypes from wolves of North China and Mongolia, which differed by 1 or 2 substitutions from the closest dog haplotypes; 
2) In Clade B, 4 wolf haplotypes were found- 2 from North Chinese and Romanian wolves were identical to dog haplotypes, and 2 from Afghani and Yugoslavian wolves differed by 2 substitutions from the closest dog haplotypes.
The study of wolves haplotypes remains incomplete because of the extinction of wolves in large parts of Europe and southern Asia and for this reason cannot give any tangible insight as to the diversity of wolf haplotypes at the time of domestication.
Therefore the authors focused on comparing the distribution of genetic diversity among the dogs themselves across the world to determine the place and time of the domestication of the wolf.

Tamed Mongolian wolf pup (Canis lupus chanco) 
Bayan-Olgii, Kazakh Golden Eagle Festival, West Mongolia
© Alexander Choi Jackson 2006

The authors find that the Old World dogs share a homogenous gene pool:
1) Clades A, B, and C were found in every population, from 97.4%-100% of all dogs.
2) The proportions of individuals with these haplotypes was similar across the Old World.
3) 14 of the haplotypes (9 in clade A, 2 in clade B, 3 in clade C) were termed Universal Types or "UTs" 
4) The UTs were very frequent in the western part of the Old World (West of Ural and Himalayas): 81.5% of dogs in Europe and Southwest Asia had a UT which is the same as a dog in Asia.
5) The UTs were much fewer in East Asia - 54.2% of the dogs there have them, and in the extreme Southeastern part of China (South of the Yangtze River)- 40.8% of the dogs have UTs with a range from 18.4% to 54.7% among the provinces.
6) Interestingly, in the West, all other haplotypes apart from the UTs differ by a single mutation from a UT. 98.7% of European dogs and 94.6% of Southwest Asian dogs either had a UT or a UT derived haplotype (UTd). This means that almost all the dogs in the West had a haplotype traced back to a haplotype also found in East Asia
7) In contrast, only 69.2% of East Asian dogs and 53.0% of dogs South of the Yangzte river had UTds.
8) The genetic diversity follows a gradient, from maximum levels at one end of the continent South of the Yangtze River, decreasing through East Asia and Eurasia to reach the lowest levels in Europe at the other end of the continent. This affects Clade A in particular, Western populations lacking several mostly identical parts of Clade A, the populations having an almost complete representation across clade A being those South of the Yantze River.
9) The Western dogs have a haplotype pool which consists almost exclusively of 14 UTs and surrounding haplotypes. East Asian population particularly in Southern China have a large proportion of dogs with unique haplotypes a large distance away from the UTs, therefore also from the haplotypes found in the West.
10) The complete mtDNA genome analysis shows that clades A, B and C consist of several subgroups. Clade A has 6 major subclades, Clade B and C have 2 subclades each, for a total of 10 subclades or haplogroups.
The complete set of 10 subclades is found only in Southern Eastern China, 7 are found in Central China, 5 in Northern China/India/SWAsia, and only 4 in Europe. 5 of the 6 Clade A subclades were found only in East Asia. In Southern China, no province had all 10 subclades, but in Yunnan, Southeast Asia and Guizhou, 9,9, and 8 of these subclades were found. The smallest region containing all 10 subclades is in the Southwest area South of the Yangtze River.
11) The substitution rate for the mtDNA data was calculated to estimate the age of the phylognetic groups. The fossil record is incomplete and the separation time for dog/wolf and coyote cannot be calibrated exactly. The data indicate that clades A,B and C spread simultaneously across the Old world 5.400 to 16.300 years ago.
12) The minimum number of lineages was estimated for the whole mtDNA genome data which existed at the time of the spread of the dogs across the Old World. Haplotypes separated by more than 12 substitutions from each other should have originated from different founders at the time of the spread. In counting such haplotypes, the authors find a minimum of 51 different female wolf lineages behind today's dogs' haplotypes. This estimate is conservative and the authors assume that several hundred female wolves were domesticated. They base their assumption on the fact that
A. some haplotypes differing by less than 12 substitutions could be from different female wolf founders
B. some lineages may be present that are not detected in this data set
C. some lineages may have become extinct since the domestication of the wolf
D. some founder female wolves may have had the same haplotype
13) The exact area South of the Yangtze where the domestication of the dog may have taken place is not clear. The full representation of all 10 subclades is found in a region which includes Yunnan, and South East Asia. Guizhou has more haplotypes than Yunnan and only one subclade less. 


1) The simplest explanation for the geographical distribution of the 10 subclades of clades A, B and C is that they had a single origin within the area close to the South of the Yangtze river and South Eastern Asia. From there a subset of the original gene pool spread to the rest of the world.
The similar proportion of these 3 clades across populations of the old World indicate that these populations originate from a founder population with a similar proportion of the 3 clades. Had the domestcation of the dogs happened independently in various areas, one would expect a large proportion of regionally unique haplotypes and varying proportions of the clades would be expected.
Populations genetic simulations showed that multiple origins in time would have required extreme migration rates and thorough mixing of the populations to end up with the observed proportions of the clades, an unlikely event. The only scenario which includes moderate migration is that of a single origin in time and space for the 3 clades.
2) It is not possible to date the origins of dogs from these data, but the date of the spread from the center of origin can be estimated. If the dog populations spread shortly after the domestication of wolves, the genetic data indicate that dogs originated approximately 5.400 to 16.300 years ago.
Because one substitution in the mtDNA genome corresponds to 3.200-9.600 years, the sharing of subclades and UTs across Eurasia is not caused by recent mixing of populations but results from ancient events.
3) These results show for the first time reasonable estimates for the time of origin of dogs and the number of female founder wolves. They dismiss the assertion by Vila et al 1997 that mtDNA indicate an origin of dogs much earlier than 10.000 to 15.000 years ago.
4) The European dogs lack 6 of the 10 subclades. They have a special history compared to other dog populations in that they have specialized in a very large number of dog breeds. For this reason one could think that the low genetic diversitry of European dog breeds results from severe bottlenecks during the development of the many different breeds. However the gene pool is almost identical to that of the dogs from South West Asia. Therefore the 6 subclades must have been missing already before the development of the many different breeds in Europe. In spite of the largest morphological variation in all of their many breeds, the European dogs have the lowest genetic variation, and this fact places them at the periphery of the Eurasian dog populations, and does not reflect modern breeding history.
5) The analysis of the whole mtDNA genome, as opposed to the mtDNA bp analysis, brings the necessary finer resolution needed to time the dog domestication more appropriately in good agreement with archaelogical finds.
6) It is possible that the domestication of wolves was a widespread practice which took place in various regions South of the Yangtze River. The North Chinese dogs' genetic diversity is very low and comparable to the European dogs with only 5 subclades and 89.8% UTds. The wolf was present in South Eastern China until the 1950s although it is now extinct there. Interestingly Olsen and Olsen (1977) found that a morphological feature typical of the dog's jaw is found also in Chinese wolves ( the turned-back apex of the coronoid process of the ascending ramus). Added to the relatively small size of the Chinese wolf, these 2 authors had concluded that the Chinese wolf was a likely ancestor of the dog.
7) This study focuses only on the mtDNA of today's dogs and therefore only on the domestication of the female wolves which were ancestral to today's dogs. 
If there was any taming or domestication of wolves or other canids which are not ancestors of today's dogs, this study does not reveal it.

These results give for the first time some indication as to which human culture may have tamed the wolves. The time and place of the domestication of the wolves coincides with the transition from hunting to farming and the emergence of rice culture. This study shows that a large number of wolves were domesticated suggesting that the domestication was not a chance/isolated event but probably a widepread custom - it also shows that the domestication of wolves happened only once.
The authors hypothesize that the wolf was first tamed as a source of food, as dogs have been eaten on a large scale in Southern East Asia until today. They suggest that the use of dogs for herding, hunting or companionship was perhaps the result of selection by European non-dog eating peoples.
The authors state that more thorough studies in South East Asia of archaelogical canid remains, and of genetics of modern and ancient canid samples should help to ascertain the validity of this hypothesis. The authors also stress that a similar study on the male lineages (Y chromosome) would complement this study of the mitochondrial DNA of female wolf and dog lineages (mtDNA being only inherited through female lines)


I thank Peter Savolainen and Arman Ardalan for their positive comments, Vladimir Bologov for the Russian wolves picture, Simon Rerucha for the Arabian wolf picture, Alexander Choi Jackson for his picture of the Mongolian Wolf and boy, and Philippe Touret for providing the invaluable pictures of Chinese dogs and the information pertaining to them. 


Olsen SJ, Olsen JW (1977): "The Chinese Wolf, Ancestor of New World Dogs", Nature 197: 533-535

Pang JF, Kluetsch C, Zou XJ, Zhang AB, Luo LY, Angleby H, Ardalan A, Ekström C, Sköllermo A, Lundeberg J, Matsumura S, Leitner T, Zhang YP, Savolainen P.(2009):"MtDNA Data Indicates a Single Origin for Dogs South of Yangtze River, less than 16,300 Years Ago, from Numerous Wolves." Mol Biol Evol. 2009 Epub. Sep 1. 

Savolainen P, Zhang Y, Luo J, Lundeberg J, Leitner T (2002): "Genetic evidence for an East Asian Origin of Domestic Dogs." Science 298:1610–1613.

Examples of modern Chinese dogs, photographed in the Guangxi and Guizhou areas, South of the Yangtze River in August of 2009
(none of these dogs were strays, they all belonged to a specific house and someone who took care of them; the information in the photos' captions is given by Philippe Touret)

© Philippe Touret 2009
Left: Bitch "Xiasi Quan" in the small horned Miao* village of Boji. The breed, which is named after the village of Xiasi, 
prefecture of Kaiili, is now accepted in most Chinese dog shows.
Right: dog from the Matang village of the Gejia people (Guizhou). All the dogs of this area were taller, more powerful and long-legged and of that off-white color. 
In a collection of thousands of pictures of this region's dogs, the first time these dogs are documented.

© Philippe Touret 2009
Left: photograph taken in the small town of Yangxi (Guangxi)
Front :a "Pig-Dog" of the minority Dong. Its name relates to the fact that once it has been fattened, slaughtered and skinned, 
this dog looks like a piglet, its body being covered with a skin of fat.
Back: a typical hunting dog of the Guangxi region. The curled tail is not the tail carriage liked by the local hunters 
who prefer a straight tail carried high, which enables them to see the dog in high grass.
Right: dogs of Biasha (Guizhou), village inhabited by black Miaos*.
The local dogs represent an intermediate type between the dog types of Guangxi and Guizhou

© Philippe Touret 2009
Left: dogs of Gaoyao (Guizhou), only village during that trip where brindled dogs were seen. 
The puppy resembles a traditional Chow-Chow, his mother has the characteristics of an official Xiasi Quan, except her color. 
The white color for this breed is imposed by the standard but other coat colors can be found ( black, fawn, slate blue, brindle)
Right: bitch photographed in small horned Miao* village of Qingman (Guizhou). Type close to the breeds hunting boar in neighboring Guangxi.

© Philippe Touret 2009 ~   © Jean-Maurice Touret 2009 
Left: Photograph taken in Dingwei (Guizhou). Dog of original Chow-Chow type, roots of this breed to be found towards Guangdong
Right: photograph taken in the Miao village of Gaoyao (Guizhou). 
Dog close to the Xiasi Quan type, slightly wire haired, 
particolored coat unusual for the region where solid colors are the absolute majority.

Puppies photographed in black Miao* village of Biasha (Guizhou) © Philippe Touret 2009


Author's notes:
*The Miaos are a group of linguistically related people in Southwest China. They are referred to in Chinese as "small horned" or "long-horned" Miaos depending on the length of the horns of the headpiece they wear, or black, white, red, striped, small flowery, big flowery, green, blue Miaos depending on the most characteristic color of the women's clothes.


| Copyright: Copyright 2008-2012, all rights of all pages under this chapter reserved to Dominique de Caprona.|