Origin of the domestic dog

During the Late Pleistocene glaciation, a vast mammoth steppe stretched from Spain eastwards across Eurasia and over the Bering land bridge into Alaska and the Yukon. The Late Pleistocene was characterized by a series of severe and rapid climate oscillations with regional temperature changes of up to 16 °C (29 °F), which has been correlated with megafaunal extinctions. There is no evidence of megafaunal extinctions at the height of the Last Glacial Maximum, indicating that increasing cold and glaciation were not factors. Multiple events appear to have caused the rapid replacement of one species by another one within the same genus, or one population by another within the same species, across a broad area. As some species became extinct, so too did the predators that depended on them (coextinction).[10]

Diagram of a wolf skull with key features labelled

The origin of dogs is couched in the paleobiogeography of wolf populations during the Late Pleistocene. The earliest fossils of Canis lupus were found in what was once eastern Beringia at Old Crow, Yukon, Canada and at Cripple Creek Sump, Fairbanks, Alaska. The age is not agreed but could date 1 million YBP. Considerable morphological diversity existed among grey wolves by the Late Pleistocene. These are regarded as having been more cranio-dentally robust than modern grey wolves, often with a shortened rostrum, the pronounced development of the temporalis muscle, and robust premolars. It is proposed that these features were specialized adaptations for the processing of carcass and bone associated with the hunting and scavenging of Pleistocene megafauna. Compared with modern wolves, some Pleistocene wolves showed an increase in tooth breakage that is similar to that seen in the extinct dire wolf. This suggests that these either often processed carcasses, or that they competed with other carnivores and needed to quickly consume their prey. The frequency and location of tooth fractures found in these wolves compared with the modern spotted hyena indicates that these wolves were habitual bone crackers.[1]

Grey wolves suffered a species-wide population bottleneck (reduction) approximately 25,000 years before present (YBP) during the Last Glacial Maximum. This was followed by a single population of modern wolves expanding out of a Beringia refuge to repopulate the wolf's former range, replacing the remaining Late Pleistocene wolf populations across Eurasia and North America as they did so.[11][12][13] This source population probably did not give rise to dogs, but admixed with dogs which allowed them to gain coat colour genes that are also related to immunity, and provided dogs with genes which allowed them to adapt to high-altitude environments (e.g. Tibet). This suggests that the genetic divergence of European and East Asian dogs could be based on admixture with different sub-populations of wolves.[13]

There is little genetic information available on the ancient wolves that existed prior to the bottleneck. However, studies show that one or more of these ancient populations is more directly ancestral to dogs than are modern wolves, and conceivably these were more prone to domestication by the first humans to invade Eurasia.[13]

An apex predator sits on the top trophic level of the food chain, while a mesopredator sits further down the food chain and is dependent on smaller animals. Towards the end of the Pleistocene era, most of today's apex predators were mesopredators and this included the wolf. During the ecological upheaval associated with the close of the Late Pleistocene, one type of wolf population rose to become today's apex predator and another joined with humans to become an apex consumer.[14]

It was such a long standing view that the gray wolf that we know today was around for hundreds of thousands of years and that dogs derived from them. We're very surprised that they're not.

— Robert K. Wayne[15]

Genetic studies indicate that the grey wolf is the closest living relative of the dog, with no evidence of any other canine species having contributed. Attempting to reconstruct the dog's lineage through the phylogenetic analysis of DNA sequences from modern dogs and wolves has given conflicting results for several reasons. Firstly, studies indicate that an extinct Late Pleistocene wolf is the nearest common ancestor to the dog, with modern wolves not being the dog's direct ancestor.[3] Secondly, the genetic divergence between the dog and modern wolves occurred over a short period of time, so that the time of the divergence is difficult to date (referred to as incomplete lineage sorting). This is complicated further by the cross-breeding that has occurred between dogs and wolves since domestication (referred to as post-domestication gene flow). Finally, there have been only tens of thousands of generations of dogs since domestication, so that the number of mutations between the dog and the wolf are few and this makes the timing of domestication difficult to date.[3]

In 2013, the whole genome sequencing of modern dogs and wolves indicated a divergence time of 32,000 YBP. In 2014, another study indicated 11,000-16,000 YBP based on the modern wolf's mutation rate. The first draft genome sequence of a Pleistocene wolf was published in 2015. This wolf from the Taymyr Peninsula belonged to a population that had diverged from the ancestors of both modern wolves and dogs. Radiocarbon dating indicates its age to be 35,000 YBP, and this age could then be used to calibrate the wolf's mutation rate, indicating that the genetic divergence between dogs and wolves occurred before the Last Glacial Maximum, between 27,000–40,000 YBP. When the Pleistocene wolf's mutation rate was applied to the timing of the earlier 2014 study which had originally used the modern wolf's mutation rate, that study gave the same result of 27,000–40,000 YBP.[1] In 2017, a study compared the nuclear genome sequences of three ancient dog specimens and found evidence of a single dog-wolf divergence occurring between 36,900-41,500 YBP.[16]

Prior to genetic divergence, the population of wolves ancestral to the dog outnumbered all other wolf populations, and after divergence the dog population underwent a population reduction to be much lower.[17][18]

Numerous genetic studies have found that the dogs from Southeast Asia and South China show greater genetic diversity than those dogs from other regions, suggesting that this was the place of their origin.[19][20][21][22][23][24] A similar study found greater genetic diversity in African village dogs than in breed dogs.[25] In 2015, a whole genome analysis of indigenous dogs from China and its border with Vietnam were compared with indigenous dogs from Africa and dog breeds from other regions. Based on the higher genetic diversity of the East Asian dogs, the study concluded that dogs originated in southern East Asia, which was followed by a migration of a subset of ancestral dogs 15,000 YBP towards the Middle East, Africa and Europe, then reaching Europe 10,000 YBP. Then, one of these lineages migrated back to northern China and admixed with endemic Asian lineages before migrating to the Americas.[24]

An East Asian origin has been questioned because dog fossils have been found in Europe dating around 15,000 YBP but only 12,000 YBP in far eastern Russia.[26] The reply is that archaeological studies in East Asia lag behind those in Europe, and that the environmental conditions in southern East Asia do not favor the preservation of fossils. Although primitive forms of the dog may have existed in Europe in the past, the genetic evidence indicates that these were later replaced by dogs that have migrated from southern East Asia.[24] In 2017, a literature review found that this East Asian study sampled only east Asian indigenous dogs and compared their patterns of genetic diversity to those of breed dogs from other geographic regions. As it is known that the genetic bottlenecks associated with formation of breeds strongly reduce genetic diversity, this was not an appropriate comparison.[3]

In 2010, a study using single nucleotide polymorphisms indicated that dogs originated in the Middle East due to the greater sharing of haplotypes between dogs and Middle Eastern grey wolves, indicating that Middle Eastern wolves were the source of domestic dogs and not East Asian wolves. Else there may have been significant admixture between some regional breeds and regional wolves.[27] In 2011, a study found that there had been dog-wolf hybridization and not an independent domestication,[28][21] with southern East Asia being the most likely origin of dogs based on their higher level of genetic diversity.[21][22] In 2012, a study found that the nuclear genome of dogs derived from wolves originating in the Middle East and Europe.[29]

In 2015, a DNA study looked at autosomal, maternal mitochondrial DNA (mDNA) and paternal Y chromosome (yDNA) diversity in purebred and village dogs from 38 countries. Some dog populations in the Neotropics and the South Pacific are almost completely derived from European dogs, and other regions show clear admixture between indigenous and European dogs. The indigenous dog populations of Vietnam, India, and Egypt show minimal evidence of European admixture, and exhibit high diversity and low linkage disequilibrium consistent with a Central Asian domestication origin, followed by a population expansion in East Asia. The study could not rule out the possibility that dogs were domesticated elsewhere and subsequently arrived in and diversified from Central Asia. Studies of extant dogs cannot exclude the possibility of earlier domestication events that subsequently died out or were overwhelmed by more modern populations.[30] In 2016, this finding was questioned by a whole genome study that included linkage disequilibrium data from east Asian indigenous dogs and found these exhibited a lower level than those of the central Asian dogs, indicating an East Asia origin.[31] This assessment was then called to question because of a negative bias caused by the low coverage used in the genome sequences.[32] In 2017, a literature review found that because it is known that the genetic bottlenecks associated with formation of breeds raise linkage disequilibrium, the comparison of purebred with village dogs was not appropriate.[3]

Most genetic studies conducted over the last two decades were based on modern dog breeds and extant wolf populations, with their findings dependent on a number of assumptions. These studies assumed that the extant wolf was the ancestor of the dog, did not consider genetic admixture between wolves and dogs, nor the impact of incomplete lineage sorting. These pre-genomic studies have suggested an origin of dogs in Southeast Asia, East Asia, Central Asia, the Middle East, or Europe. More recently, the field of Paleogenomics applies the latest molecular technologies to fossil remains that still contain useful DNA.[1]

In 2013, a study looked at the well-preserved 33,000-year-old skull and left mandible of a dog-like canid that was excavated from Razboinichya Cave in the Altai Mountains of southern Siberia (Central Asia). The mDNA analysis found it to be more closely related to dogs than wolves.[33] Later in 2013, another study found that the canid could not be classified as it fell between both dogs and wolves.[34] In 2017, evolutionary biologists reviewed all of the evidence available on dog divergence and supported the specimens from the Altai mountains as being those of dogs from a lineage that is now extinct, and that was derived from a population of small wolves that is also now extinct.[3]

See further: Altai dog

The 14,500-year-old upper-right jaw of a Pleistocene wolf found in the Kessleroch cave near Thayngen in the canton of Schaffhausen, Switzerland, is the sister to 2/3 of modern dogs[34]

In 2013, a study sequenced the complete and partial mitochondrial genomes of 18 fossil canids dated from 1,000 to 36,000 YBP from the Old and New Worlds, and compared these with the complete mitochondrial genome sequences from modern wolves and dogs. Phylogenetic analysis showed that modern dog mDNA haplotypes resolve into four monophyletic clades with strong statistical support, and these have been designated by researchers as clades A-D.[34][35][36] Based on the specimens used in this study, clade A included 64% of the dogs sampled and these were sister to a 14,500 YBP wolf sequence from the Kessleroch cave near Thayngen in the canton of Schaffhausen, Switzerland, with a most recent common ancestor estimated to 32,100 YBP. This group of dogs matched three fossil pre-Columbian New World dogs dated between 1,000 and 8,500 YBP, which supported the hypothesis that pre-Columbian dogs in the New World share ancestry with modern dogs and that they likely arrived with the first humans to the New World. Clade B included 22% of the dog sequences and was related to modern wolves from Sweden and the Ukraine, with a common recent ancestor estimated to 9,200 YBP. However, this relationship might represent mitochondrial genome introgression from wolves because dogs were domesticated by this time. Clade C included 12% of the dogs sampled and these were sister to two ancient dogs from the Bonn-Oberkassel cave (14,700 YBP) and the Kartstein cave (12,500 YBP) near Mechernich in Germany, with a common recent ancestor estimated to 16,000–24,000 YBP. Clade D contained sequences from 2 Scandinavian breeds (Jamthund, Norwegian Elkhound) and were sister to another 14,500 YBP wolf sequence also from the Kesserloch cave, with a common recent ancestor estimated to 18,300 YBP. Its branch is phylogenetically rooted in the same sequence as the "Altai dog" (not a direct ancestor). The data from this study indicated a European origin for dogs that was estimated at 18,800–32,100 YBP based on the genetic relationship of 78% of the sampled dogs with ancient canid specimens found in Europe.[37][34] The data supports the hypothesis that dog domestication preceded the emergence of agriculture[35] and was initiated close to the Last Glacial Maximum when hunter-gatherers preyed on megafauna.[34][38]

The study found that three ancient Belgium canids (the 36,000 YBP "Goyet dog" cataloged as Canis species, along with Belgium 30,000 YBP and 26,000 YBP cataloged as Canis lupus) formed an ancient clade that was the most divergent group. The study found that the skulls of the "Goyet dog" and the "Altai dog" had some dog-like characteristics and proposed that the may have represented an aborted domestication episode. If so, there may have been originally more than one ancient domestication event for dogs[34] as there was for domestic pigs.[39]

One theory is that domestication occurred during one of the five cold Heinrich events that occurred after the arrival of humans in West Europe 37 000, 29 000, 23 000, 16 500 and 12 000 YBP. The theory is that the extreme cold during one of these events caused humans to either shift their location, adapt through a breakdown in their culture and change of their beliefs, or adopt innovative approaches. The adoption of the large wolf/dog was an adaptation to this hostile environment.[40]

A criticism of the European proposal is that dogs in East Asia show more genetic diversity. However, dramatic differences in genetic diversity can be influenced both by an ancient and recent history of inbreeding.[24] A counter-comment is that the modern European breeds only emerged in the 19th century, and that throughout history global dog populations experienced numerous episodes of diversification and homogenization, with each round further reducing the power of genetic data derived from modern breeds to help infer their early history.[26]

In 2019, an mDNA study of 19 Late Pleistocene-Holocene wolf samples from northern Italy found that these fell within mDNA haplogroup 2 except for one sample. One canid specimen from the Cava Filo archaeological site near San Lazzaro di Savena, Bologna fell within the domestic dog clade A haplotype — it was radio-carbon dated 24,700 YBP.[41]

Mandible of Canis c.f. variabilis from northeastern Siberia dated 360,000-400,000 years old.[42]

In 2015, a study recovered mDNA from ancient canid specimens that were discovered in arctic northeastern Siberia (which was once western Beringia). These specimens included the mandible of a 360,000-400,000 YBP Canis c.f. variabilis (where c.f. is a Latin term meaning uncertain). Phylogenetic analyses of these canids revealed nine mDNA haplotypes not detected before. The Canis c.f. variabilis specimen clustered with other wolf samples from across Russia and Asia. The mDNA haplotypes of one 8,750 YBP specimen and some 28,000 YBP specimens matched with those of geographically widely-spread modern dogs. One 47,000 YBP canid was distinct from wolves but was only a few mutations away from those haplotypes found in modern dogs. The authors concluded that the structure of the modern dog gene pool was contributed to from ancient Siberian wolves and possibly from Canis c.f. variabilis.[42][43]

Dogs show both ancient and modern lineages. The ancient lineages appear most in Asia but least in Europe because the Victorian era development of modern dog breeds used little of the ancient lineages.[27][30][17] All dog populations (breed, village, and feral) show some evidence of genetic admixture between modern and ancient dogs. Some ancient dog populations that once occupied Europe and the New World no longer exist.[34][2][17][44] This implies that some ancient dog populations were entirely replaced and others admixed over a long period of time.[45] European dog populations have undergone extensive turnover during the last 15,000 years which has erased the genomic signature of early European dogs,[30][46] the genetic heritage of the modern breeds has become blurred due to admixture,[26] and there was the possibility of past domestication events that had gone extinct or had been largely replaced by more modern dog populations.[30]

In 2016, a study compared the mDNA and whole-genome sequences of a worldwide panel of modern dogs, the mDNA sequences of 59 ancient European dog specimens dated 14,000-3,000 YBP, and the nuclear genome sequence of a dog specimen that was found in the Late Neolithic passage grave at Newgrange, Ireland and radiocarbon dated at 4,800 YBP. A genetic analysis of the Newgrange dog showed that it was male, did not possess genetic variants associated with modern coat length nor color, was not as able to process starch as efficiently as modern dogs but more efficiently than wolves, and showed ancestry from a population of wolves that could not be found in other dogs nor wolves today. As the taxonomic classification of the "proto-dog" Paleolithic dogs as being either dogs or wolves remains controversial, they were excluded from the study. The phylogenetic tree generated from mDNA sequences found a deep division between the Sarloos wolfdog and all other dogs, indicating that breed's recent deriving from the German Shepherd and captive grey wolves. The next largest division was between eastern Asian dogs and western Eurasian (Europe and the Middle East) dogs that had occurred between 14,000-6,400 YBP, with the Newgrange dog clustering with the western Eurasian dogs.[39]

The Newgrange and ancient European dog mDNA sequences could be largely assigned to mDNA haplogroups C and D but modern European dog sequences could be largely assigned to mDNA haplogroups A and B, indicating a turnover of dogs in the past from a place other than Europe. As this split dates older than the Newgrange dog this suggests that the replacement was only partial. The analysis showed that most modern European dogs had undergone a population bottleneck (reduction) which can be an indicator of travel. The archaeological record shows dog remains dating over 15,000 YBP in western Eurasia, over 12,500 YBP in eastern Eurasia, but none older than 8,000 YBP in Central Asia. The study proposes that dogs may have been domesticated separately in both eastern and western Eurasia from two genetically distinct and now extinct wolf populations. East Eurasian dogs then made their way with migrating people to western Europe between 14,000-6,400 YBP where they partially replaced the dogs of Europe.[17][47] Two domestication events in western Eurasia and eastern Eurasia has recently been found for the domestic pig.[17]

The hypothesis is that two genetically different, and possibly now extinct, wolf populations were domesticated independently in eastern and western Eurasia to produce paleolithic dogs.[17] The eastern Eurasian dogs then dispersed westward alongside humans, reaching western Europe 6,400–14,000 YBP where they partially replaced the western paleolithic dogs.[37][17][1] A single domestication is thought to be due to chance, however dual domestication on different sides of the world is unlikely to have happened randomly and it suggests that external factors - an environmental driver - may have forced wolves to work together with humans for survival. It is possible that wolves took advantage of resources that humans had, or humans may have been introduced to wolves in an area in which they didn't previously live.[48]

In 2017, a study compared the nuclear genome sequences of three ancient dog specimens from Germany and Ireland with sequences from over 5,000 dogs and wolves. These Neolithic dog specimens included a dog sample from the Early Neolithic site in Herxheim, Germany dated 7,000 YBP, one from the Late Neolithic site of Kirschbaum (Cherry Tree) Cave near Forchheim, Germany dated 4,700 YBP, and a dog from Newgrange, Ireland dated 4,800 YBP. The study found that modern European dogs descended from their Neolithic ancestors with no evidence of a population turnover. There was evidence of a single dog-wolf divergence occurring between 36,900-41,500 YBP, followed by a divergence between Southeast Asian and Western Eurasian dogs 17,500-23,900 YBP and this indicates a single dog domestication event occurring between 20,000-40,000 YBP. The 3 dogs indicated ancestry that could be found in South East Asian dogs. Additionally, the Cherry Tree Cave dog showed ancestry that could be found in the Middle East, India and Central Asia.[16] The study did not support a dual domestication event, and detected admixture between the ancestors of modern European and Southeast Asian dogs.[16][1]

A 2018 study of mDNA sequences shows that the pre-Neolithic dogs of Europe all fell under haplogroup C. The Neolithic and Post-Neolithic dogs from Southeastern Europe that are associated with farmers fell under haplogroup D. In Western and Northern Europe, haplogroup D became diluted into the native dog population. This implies that haplogroup D arrived in Europe 9,000 YBP from the Near East along with pigs, cows, sheep, and goats.[49] Later in 2018, another study looked at the y-chromosome male lineage of the ancient fossils of the Herxheim, Kirschbaum, and Newgrange dogs along with other canines. The study identified six major dog yDNA haplogroups, of which two of these include the majority of modern dogs. The Newgrange dog fell into the most commonly occurring of these haplogroups. The two ancient German dogs fell into a haplogroup commonly found among dogs from the Middle East and Asia, with the Kirschbaum dog sharing a common male lineage with the extant Indian wolf. The study concluded that at least 2 different male haplogroups existed in ancient Europe, and that the dog male lineage diverged from its nearest common ancestor shared with the grey wolf sometime between 68,000-151,000 YBP.[50]

The questions of when and where dogs were first domesticated have taxed geneticists and archaeologists for decades.[2] Identifying the earliest dogs is difficult because the key morphological characters that are used by zooarchaeologists to differentiate domestic dogs from their wild wolf ancestors (size and position of teeth, dental pathologies, and size and proportion of cranial and postcranial elements) were not yet fixed during the initial phases of the domestication process. The range of natural variation among these characters that may have existed in ancient wolf populations, and the time it took for these traits to appear in dogs, are unknown.[26]

The domestication of animals commenced over 15,000 YBP, beginning with the grey wolf by nomadic hunter-gatherers. It was not until 11,000 YBP that people living in the Near East entered into relationships with wild populations of aurochs, boar, sheep, and goats. A domestication process then began to develop. The grey wolf most likely followed the commensal pathway to domestication. When, where, and how many times wolves may have been domesticated remains debated because only a small number of ancient specimens have been found, and both archaeology and genetics continue to provide conflicting evidence. The most widely accepted, earliest dog remains date back 15,000 YBP to the Bonn–Oberkassel dog. Earlier remains dating back to 30,000 YBP have been described as Paleolithic dogs, however their status as dogs or wolves remains debated.[4]

There are a number of recently discovered specimens which are proposed as being Paleolithic dogs, however their taxonomy is debated. These have been found in Europe and Siberia dated 40,000-17,000 YBP. They include Hohle Fels in Germany, Goyet Caves in Belgium, Predmosti in the Czech Republic, and four sites in Russia: Razboinichya Cave in the Altai Republic, Kostyonki-Borshchyovo archaeological complex on the Don river in western Russia, Ulakhan Sular in the Sakha Republic, and Eliseevichi 1 in the Dnieper river basin. Paw-prints from Chauvet Cave in France dated 26,000 YBP are suggested as being those of a dog, however these have been challenged as being left by a wolf.[1]

Razboinichya Cave "Altai dog" that is dated 33,500 years old

There are also a number of later proposed Paleolithic dogs whose taxonomy has not been confirmed. These include a number of specimens from Germany (Kniegrotte, Oelknitz, Teufelsbrucke), Switzerland (Monruz, Kesslerloch, Champre-veyres-Hauterive), and Ukraine (Mezin, Mezhirich). A set of specimens dating 15,000-13,500 YBP have been confidently identified as domesticated dogs, based on their morphology and the archaeological sites in which they have been found. These include Spain (Erralla), France (Montespan, Le Morin, Le Closeau, Pont d’Ambon), and Germany (Bonn-Oberkassel). After this period, the remains of domesticated dogs have been identified from archaeological sites across Eurasia.[1]

Possible dog domestication between 15,000-40,000 YBP is not clear due to the debate over what the Paleolithic dog specimens represent. This is due to the flexibility of genus Canis morphology, and the close morphological similarities between Canis lupus and Canis familiaris. It is also due to the scarcity of Pleistocene wolf specimens available for analyses and so their morphological variation is unknown. Habitat type, climate, and prey specialization greatly modify the morphological plasticity of grey wolf populations, resulting in a range of morphologically, genetically, and ecologically distinct wolf morphotypes. With no baseline to work from, zooarchaeologists find it difficult to be able to differentiate between the initial indicators of dog domestication and various types of Late Pleistocene wolf ecomorphs, which can lead to the mis-identification of both early dogs and wolves. Additionally, the ongoing prehistoric admixture with local wolf populations during the domestication process may have led to canids that were domesticated in their behavior but wolflike in their morphology. Attempting to identify early tamed wolves, wolfdogs, or proto-dogs through morphological analysis alone may be impossible without the inclusion of genetic analyses.[1]

A domestication study looked at the reasons why the archeological record that is based on the dating of fossil remains often differed from the genetic record contained within the cells of living species. The study concluded that our inability to date domestication is because domestication is a continuum and there is no single point where we can say that a species was clearly domesticated using these two techniques. The study proposes that changes in morphology across time and how humans were interacting with the species in the past needs to be considered in addition to these two techniques.[51]

..."wild" and "domesticated" exist as concepts along a continuum, and the boundary between them is often blurred — and, at least in the case of wolves, it was never clear to begin with.

— Raymond Pierotti[8]

Evolution of temperatures in the postglacial period, after the Last Glacial Maximum (LGM), showing very low temperatures for the most part of the Younger Dryas, rapidly rising afterwards to reach the level of the warm Holocene, based on Greenland ice cores.[55]

The domestication of animals and plants was triggered by the climatic and environmental changes that occurred after the peak of the Last Glacial Maximum around 21,000 YBP and which continue to this present day. These changes made obtaining food difficult. The first domesticate was the grey wolf (Canis lupus) at least 15,000 YBP. The Younger Dryas that occurred 12,900 YBP was a period of intense cold and aridity that put pressure on humans to intensify their foraging strategies. With the closing of the Younger Dryas at the beginning of the Holocene around 11,700 YBP, favorable climatic conditions and increasing human populations led to small-scale animal and plant domestication, which allowed humans to augment the food that they were obtaining through hunter-gathering. The Neolithic transition led to agricultural societies emerging in locations across Eurasia, North Africa, and South and Central America.[56]

Watercolor tracing made by archaeologist Henri Breuil from a cave painting of a wolf-like canid, Font-de-Gaume, France dated 19,000 years ago.

In August 2015, a study undertook an analysis of the complete mitogenome sequences of 555 modern and ancient dogs. The sequences showed an increase in the population size approximately 23,500 YBP, which broadly coincides with the proposed genetic divergence of the ancestors of dogs and present-day wolves before the Last Glacial Maximum. A ten-fold increase in the population size occurred after 15,000 YBP, which may be attributable to domestication events and is consistent with the demographic dependence of dogs on the human population.[57]

Humans and wolves both exist in complex social groups. How humans and wolves got together remains unknown. One view holds that domestication as a process that is difficult to define. The term was developed by anthropologists with a human-centric view in which humans took wild animals (ungulates) and bred them to be "domestic", usually in order to provide improved food or materials for human consumption. That term may not be appropriate for a large carnivore such as the dog. This alternate view regards dogs as being either socialized and able to live among humans, or unsocialized. There exists today dogs that live with their human families but are unsocialized and will threaten strangers defensively and aggressively no different from a wild wolf. There also exists a number of cases where wild wolves have approached people in remote places, attempting to initiate play and to form companionship.[58] One such notable wolf was Romeo, a gentle black wolf that formed relationships with the people and dogs of Juneau, Alaska.[59] This view holds that before there could have been domestication of the wolf, there had to have been its socialization.[58][60]

Mammoth bone dwelling, Mezhirich site, Ukraine

Animal domestication is a coevolutionary process in which a population responds to selective pressure while adapting to a novel niche that included another species with evolving behaviors.[2]

See further: Convergent evolution between dogs and humans

The dog is a classic example of a domestic animal that likely traveled a commensal pathway into domestication. The dog was the first domesticant, and was domesticated and widely established across Eurasia before the end of the Pleistocene, well before cultivation or the domestication of other animals.[26] It may have been inevitable that the first domesticated animal came from the order of carnivores as these are less afraid when approaching other species. Within the carnivores, the first domesticated animal would need to exist without an all-meat diet, possess a running and hunting ability to provide its own food, and be of a controllable size to coexist with humans, indicating the family Canidae, and the right temperament[61] with wolves being among the most gregarious and cooperative animals on the planet.[62][63]

See further: Commensal pathway

Ancient DNA supports the hypothesis that dog domestication preceded the emergence of agriculture[34][35] and was initiated close to the Last Glacial Maximum 27,000 YBP when hunter-gatherers preyed on megafauna, and when proto-dogs might have taken advantage of carcasses left on site by early hunters, assisted in the capture of prey, or provided defense from large competing predators at kill-sites.[34] Wolves were probably attracted to human campfires by the smell of meat being cooked and discarded refuse in the vicinity, first loosely attaching themselves and then considering these as part of their home territory where their warning growls would alert humans to the approach of outsiders.[64] The wolves most likely drawn to human camps were the less-aggressive, subdominant pack members with lowered flight response, higher stress thresholds, less wary around humans, and therefore better candidates for domestication.[65] The earliest sign of domestication in dogs was the neotenization of skull morphology[65][66][67] and the shortening of snout length that results in tooth crowding, reduction in tooth size, and a reduction in the number of teeth,[68][65] which has been attributed to the strong selection for reduced aggression.[65][66] This process may have begun during the initial commensal stage of dog domestication, even before humans began to be active partners in the process.[2][65]

Montage showing the morphological variation of the dog.

A maternal mDNA, paternal yDNA, and microsatellite assessment of two wolf populations in North America and combined with satellite telemetry data revealed significant genetic and morphological differences between one population that migrated with and preyed upon caribou, and another territorial ecotype population that remained in a boreal coniferous forest. Though these two populations spend a period of the year in the same place, and though there was evidence of gene flow between them, the difference in prey–habitat specialization has been sufficient to maintain genetic and even coloration divergence.[2][69] A study has identified the remains of a population of extinct Pleistocene Beringian wolves with unique mDNA signatures. The skull shape, tooth wear, and isotopic signatures suggested these were specialist megafauna hunters and scavengers that became extinct while less specialized wolf ecotypes survived.[2][70] Analogous to the modern wolf ecotype that has evolved to track and prey upon caribou, a Pleistocene wolf population could have begun following mobile hunter-gatherers, thus slowly acquiring genetic and phenotypic differences that would have allowed them to more successfully adapt to the human habitat.[2][71]

Even today, the wolves on Ellesmere Island do not fear humans, which is thought to be due to them seeing humans so little, and they will approach humans cautiously, curiously and closely.[72][73][74][75]

See further: Megafaunal wolf

Since domestication, dogs have traveled alongside humans across most of the planet, often hybridizing with local wolves. This has resulted in complex patterns of ancient and recent admixture among both the wild and the domestic canids.[54] The β-defensin gene responsible for the black coat of North American wolves was the result of a single introgression from early Native American dogs in the Yukon between 1,600-7,200 YBP.[76]

Studies of whole-genome sequences indicate admixture between the dog-wolf ancestor and golden jackals. There is evidence of admixture that occurred after domestication that is common within local populations of wolves and dogs. This implies that the genomic diversity found in dogs may represent interbreeding with local wolf populations and not their descent from them, which confounds the ability to infer the dog's origin.[3] The short divergence time between dogs and wolves followed by their continuous admixture has led to 20% of the genome of East Asian wolves and 7–25% of the genome of European and Middle Eastern wolves showing contributions from dogs.[1] Whole genome sequencing indicates that while there has been wide-spread geneflow from dogs into different wolf populations, the world's dog population forms a homogenous group with little evidence of outbreeding with wolves, apart from deliberate crossings such as the Sarloos wolfdog. Ancient DNA indicates that dogs in Europe dating back to 5,000 YBP also did not outbreed with wolves.[77]

In 2018, a study found there was introgression of the male dog into the wolf, but only one hybrid detected which was the result of a male wolf crossed with a female dog. Wolves have maintained their phenotype differences from the dog, which indicates low-frequency hybridization. There was almost no admixture detected in the North American specimens.[78]

The Greenland dog carries 3.5% genetic material inherited from a 35,000 years-old wolf from the Taymyr Peninsula, arctic Siberia.

In 2015, a study mapped the first genome of a 35,000 YBP Pleistocene wolf fossil found in the Taymyr Peninsula, arctic Siberia and compared it with those of modern dogs and grey wolves. The Taimyr wolf diverged from the dog–grey wolf ancestor immediately before the dog and grey wolf diverged from each other, which implies that the majority of grey wolf populations today stems from an ancestral population that lived less than 35,000 years ago but before the inundation of the Bering Land Bridge with the subsequent isolation of Eurasian and North American wolves.[79]

The Taimyr wolf shared more alleles (i.e. gene expressions) with those breeds that are associated with high latitudes and arctic human populations - the Siberian husky and Greenland dog, and to a lesser extent the Shar Pei and Finnish spitz. The Greenland dog shows 3.5% Taimyr wolf ancestry, which indicates admixture between the Taimyr wolf population and the ancestral dog population of these four high-latitude breeds. These results can be explained either by a very early presence of dogs in northern Eurasia or by the genetic legacy of the Taimyr wolf being preserved in northern wolf populations until the arrival of dogs into high latitudes. This introgression could have provided early dogs living in high latitudes with adaptations to the new and challenging environment. It also indicates that the ancestry of present-day dog breeds descends from more than one region.[79]^:3–4 An attempt to explore admixture between the Taimyr wolf and grey wolves produced unreliable results.[79]^:23

As the Taimyr wolf had contributed to the genetic makeup of the Arctic breeds, this indicates that the descendants of the Taimyr wolf survived until dogs were domesticated in Europe and arrived at high latitudes where they mixed with local wolves, and these both contributed to the modern Arctic breeds. Based on the most widely accepted oldest zooarchaeological dog remains, domestic dogs most likely arrived at high latitudes within the last 15,000 years. The mutation rates calibrated from both the Taimyr wolf and the Newgrange dog genomes suggest that the modern wolf and dog populations diverged from a common ancestor between 20,000 and 60,000 YBP. This indicates that either dogs were domesticated much earlier than their first appearance in the archaeological record, or they arrived in the Arctic early, or both.[5] Another view is that because northern breeds can trace at least some of their ancestry back to the Taimyr wolf, this indicates the possibility of more than one domestication event.[1]

Reduction in size under selective breeding - grey wolf and chihuahua skulls.

The wolf's family portrait reveals a diversity of form among breeds of domestic dogs.

Charles Darwin recognized the small number of traits that made domestic species different from their wild ancestors. He was also the first to recognize the difference between conscious selective breeding in which humans directly select for desirable traits, and unconscious selection where traits evolve as a by-product of natural selection or from selection on other traits.[80][81] Domestic animals have variations in coat color as well as texture, dwarf and giant varieties, and changes in their reproductive cycle, and many others have tooth crowding and floppy ears.

Although it is easy to assume that each of these traits was uniquely selected for by hunter-gatherers and early farmers, beginning in 1959 Dmitry Belyayev tested the reactions of silver foxes to a hand placed in their cage and selected the tamest, least aggressive individuals to breed. His hypothesis was that, by selecting a behavioral trait, he could also influence the phenotype of subsequent generations, making them more domestic in appearance. Over the next 40 years, he succeeded in producing foxes with traits that were never directly selected for, including piebald coats floppy ears, upturned tails, shortened snouts, and shifts in developmental timing.[66][82][83] In the 1980s, a researcher used a set of behavioral, cognitive, and visible phenotypic markers, such as coat colour, to produce domesticated fallow deer within a few generations.[82][84] Similar results for tameness and fear have been found for mink[85] and Japanese quail.[86] In addition to demonstrating that domestic phenotypic traits could arise through selection for a behavioral trait, and domestic behavioral traits could arise through the selection for a phenotypic trait, these experiments provided a mechanism to explain how the animal domestication process could have begun without deliberate human forethought and action.[82]

The genetic difference between domestic and wild populations can be framed within two considerations. The first distinguishes between domestication traits that are presumed to have been essential at the early stages of domestication, and improvement traits that have appeared since the split between wild and domestic populations.[2][87][88] Domestication traits are generally fixed within all domesticates and were selected during the initial episode of domestication, whereas improvement traits are present only in a proportion of domesticates, though they may be fixed in individual breeds or regional populations.[2][88][89] A second issue is whether traits associated with the domestication syndrome resulted from a relaxation of selection as animals exited the wild environment or from positive selection resulting from intentional and unintentional human preference. Some recent genomic studies on the genetic basis of traits associated with the domestication syndrome have shed light on both of these issues.[2] A study published in 2016 suggested that there have been negative genetic consequences of the domestication process as well, that enrichment of disease-related gene variants accompanied positively selected traits.[90]

In 2010, a study identified 51 regions of the dog genome that were associated with phenotypic variation among breeds in 57 traits studied, which included body, cranial, dental, and long bone shape and size. There were 3 quantitative trait loci that explained most of the phenotypic variation. Indicators of recent selection were shown by many of the 51 genomic regions that were associated with traits that define a breed, which include body size, coat characteristics, and ear floppiness.[91] Geneticists have identified more than 300 genetic loci and 150 genes associated with coat color variability.[82][92] Knowing the mutations associated with different colors has allowed the correlation between the timing of the appearance of variable coat colors in horses with the timing of their domestication.[82][93] Other studies have shown how human-induced selection is responsible for the allelic variation in pigs.[82][94] Together, these insights suggest that, although natural selection has kept variation to a minimum before domestication, humans have actively selected for novel coat colors as soon as they appeared in managed populations.[82][95]

In 2015, a study looked at over 100 pig genome sequences to ascertain their process of domestication. A model that fitted the data included admixture with a now extinct ghost population of wild pigs during the Pleistocene. The study also found that despite back-crossing with wild pigs, the genomes of domestic pigs have strong signatures of selection at genetic loci that affect behavior and morphology. The study concluded that human selection for domestic traits likely counteracted the homogenizing effect of gene flow from wild boars and created domestication islands in the genome. The same process may also apply to other domesticated animals.[39][96]

In 2014, a whole genome study of the DNA differences between wolves and dogs found that dogs did not show a reduced fear response but did show greater synaptic plasticity. Synaptic plasticity is widely believed to be the cellular correlate of learning and memory, and this change may have altered the learning and memory abilities of dogs in comparison to wolves.[97]

Unlike other domestic species which were primarily selected for production-related traits, dogs were initially selected for their behaviors.[98][99] In 2016, a study found that there were only 11 fixed genes that showed variation between wolves and dogs. These gene variations were unlikely to have been the result of natural evolution, and indicate selection on both morphology and behavior during dog domestication. There was evidence of selection during dog domestication of genes that affect the adrenaline and noradrenaline biosynthesis pathway. These genes are involved in the synthesis, transport and degradation of a variety of neurotransmitters, particularly the catecholamines, which include dopamine and noradrenaline. Recurrent selection on this pathway and its role in emotional processing and the fight-or-flight response[99][100] suggests that the behavioral changes we see in dogs compared to wolves may be due to changes in this pathway, leading to tameness and an emotional processing ability.[99] Dogs generally show reduced fear and aggression compared to wolves.[99][101] Some of these genes have been associated with aggression in some dog breeds, indicating their importance in both the initial domestication and then later in breed formation.[99]

In 2018, a study identified 429 genes that differed between modern dogs and modern wolves. As the differences in these genes could also be found in ancient dog fossils, these were regarded as being the result of the initial domestication and not from recent breed formation. These genes are linked to neural crest and central nervous system development. These genes affect embryogenesis and can confer tameness, smaller jaws, floppy ears, and diminished craniofacial development, which distinguish domesticated dogs from wolves and are considered to reflect domestication syndrome. The study proposes that domestication syndrome is caused by alterations in the migration or activity of neural crest cells during their development. The study concluded that during early dog domestication, the initial selection was for behavior. This trait is influenced by those genes which act in the neural crest, which led to the phenotypes observed in modern dogs.[102]

Similar to humans, wolves show strong social and emotional bonds within their groupings, and this relationship might have been the foundation for the evolution of dog-human bonding.[103][104] In 2019, a literature review led to a new theory named Active Social Domestication, in which the social environment of the dog ancestor induced neuro-physiological changes that caused an epigenetic cascade, which led to the rapid development of domestication syndrome. [103][105]

The difference in overall body size between a Cane Corso (Italian mastiff) and a Yorkshire terrier is over 30-fold, yet both are members of the same species.

AMY2B (Alpha-Amylase 2B) is a gene that codes a protein that assists with the first step in the digestion of dietary starch and glycogen. An expansion of this gene in dogs would enable early dogs to exploit a starch-rich diet as they fed on refuse from agriculture.[28][99] In a study in 2014, the data indicated that the wolves and dingo had just two copies of the gene and the Siberian Husky that is associated with hunter-gatherers had just three or four copies, whereas the Saluki that is associated with the Fertile Crescent where agriculture originated had 29 copies. The results show that on average, modern dogs have a high copy number of the gene, whereas wolves and dingoes do not. The high copy number of AMY2B variants likely already existed as a standing variation in early domestic dogs, but expanded more recently with the development of large agriculturally based civilizations. This suggests that at the beginning of the domestication process, dogs may have been characterized by a more carnivorous diet than their modern-day counterparts, a diet held in common with early hunter-gatherers.[28] A later study indicated that because most dogs had a high copy number of the AMY2B gene but the arctic breeds and the dingo did not, that the dog's dietary change may not have been caused by initial domestication but by the subsequent spread of agriculture to most - but not all - regions of the planet.[106] In 2017, an analysis of 3 ancient dog remains from Germany dated 7,000-4,700 YBP showed no increase in the AMY2B gene copy number although agriculture had reached this region by that time, questioning the role of agriculture in dog domestication.[16]

In 2016, a study of the dog genome compared to the wolf genome looked for genes that showed signs of having undergone positive selection. The study identified genes relating to brain function and behavior, and to lipid metabolism. This ability to process lipids indicates a dietary target of selection that was important when proto-dogs hunted and fed alongside hunter-gatherers. The evolution of the dietary metabolism genes may have helped process the increased lipid content of early dog diets as they scavenged on the remains of carcasses left by hunter-gatherers.[107] Prey capture rates may have increased in comparison to wolves and with it the amount of lipid consumed by the assisting proto-dogs.[107][38][108] A unique dietary selection pressure may have evolved both from the amount consumed, and the shifting composition of, tissues that were available to proto-dogs once humans had removed the most desirable parts of the carcass for themselves.[107] A study of the mammal biomass during modern human expansion into the northern Mammoth steppe found that it had occurred under conditions of unlimited resources, and that many of the animals were killed with only a small part consumed or left unused.[109]

See further: Phenotypic plasticity

As a result of the domestication process there is also evidence of convergent evolution having occurred between dogs and humans.[110] The history of the two is forever intertwined.[111] Dogs suffer from the same common diseases - such as cancer, diabetes, heart disease, and neurological disorders - as do humans. The underlying disease pathology is similar to humans, as is their responses and outcomes to treatment.[54]

Montage showing the coat variation of the dog.

There are patterns of genes which are related by their function and these patterns can be found in both dogs and humans. This fact can be used to study the coevolution of gene function. Dogs accompanied humans when they first migrated into new environments. Both dogs and humans have adapted to different environmental conditions, with their genomes showing parallel evolution. These include adaptation to high altitude, low oxygen hypoxia conditions, and genes that play a role in digestion, metabolism, neurological processes, and some related to cancer. It can be inferred from those genes which act on the serotonin system in the brain that these have given rise to less aggressive behavior when living in a crowded environment.[1]

In 2007, a study found that dog domestication was accompanied by selection at three genes with key roles in starch digestion: AMY2B, MGAMand SGLT1, and was a striking case of parallel evolution when coping with an increasingly starch-rich diet caused similar adaptive responses in dogs and humans.[112][113]

Convergent evolution is when distantly related species independently evolve similar solutions to the same problem. For example, fish, penguins and dolphins have each separately evolved flippers as a solution to the problem of moving through the water. What has been found between dogs and humans is something less frequently demonstrated: psychological convergence. Dogs have independently evolved to be cognitively more similar to humans than we are to our closest genetic relatives.[110]^:60 Dogs have evolved specialized skills for reading human social and communicative behavior. These skills seem more flexible – and possibly more human-like – than those of other animals more closely related to humans phylogenetically, such as chimpanzees, bonobos and other great apes. This raises the possibility that convergent evolution has occurred: both Canis familiaris and Homo sapiens might have evolved some similar (although obviously not identical) social-communicative skills – in both cases adapted for certain kinds of social and communicative interactions with human beings.[114]

The pointing gesture is a human-specific signal, is referential in its nature, and is a foundation building-block of human communication. Human infants acquire it weeks before the first spoken word.[115] In 2009, a study compared the responses to a range of pointing gestures by dogs and human infants. The study showed little difference in the performance of 2-year-old children and dogs, while 3-year-old children's performance was higher. The results also showed that all subjects were able to generalize from their previous experience to respond to relatively novel pointing gestures. These findings suggest that dogs demonstrating a similar level of performance as 2-year-old children can be explained as a joint outcome of their evolutionary history as well as their socialization in a human environment.[116]

Later studies support coevolution in that dogs can discriminate the emotional expressions of human faces,[117] and that most people can tell from a bark whether a dog is alone, being approached by a stranger, playing, or being aggressive,[118] and can tell from a growl how big the dog is.[119]

In 2015, a study found that when dogs and their owners interact, extended eye contact (mutual gaze) increases oxytocin levels in both the dog and its owner. As oxytocin is known for its role in maternal bonding, it is considered likely that this effect has supported the coevolution of human-dog bonding.[120] One observer has stated, "The dog could have arisen only from animals predisposed to human society by lack of fear, attentiveness, curiosity, necessity, and recognition of advantage gained through collaboration....the humans and wolves involved in the conversion were sentient, observant beings constantly making decisions about how they lived and what they did, based on the perceived ability to obtain at a given time and place what they needed to survive and thrive. They were social animals willing, even eager, to join forces with another animal to merge their sense of group with the others' sense and create an expanded super-group that was beneficial to both in multiple ways. They were individual animals and people involved, from our perspective, in a biological and cultural process that involved linking not only their lives but the evolutionary fate of their heirs in ways, we must assume, they could never have imagined. Powerful emotions were in play that many observers today refer to as love – boundless, unquestioning love."[9]

In 2002, a study proposed that immediate human ancestors and wolves may have domesticated each other through a strategic alliance that would change both respectively into humans and dogs. The effects of human psychology, hunting practices, territoriality and social behavior would have been profound.[121]

Early humans moved from scavenging and small-game hunting to big-game hunting by living in larger, socially more-complex groups, learning to hunt in packs, and developing powers of cooperation and negotiation in complex situations. As these are characteristics of wolves, dogs and humans, it can be argued that these behaviors were enhanced once wolves and humans began to cohabit. Communal hunting led to communal defense. Wolves actively patrol and defend their scent-marked territory, and perhaps humans had their sense of territoriality enhanced by living with wolves.[121] One of the keys to recent human survival has been the forming of partnerships. Strong bonds exist between same-sex wolves, dogs and humans and these bonds are stronger than exist between other same-sex animal pairs. Today, the most widespread form of inter-species bonding occurs between humans and dogs. The concept of friendship has ancient origins but it may have been enhanced through the inter-species relationship to give a survival advantage.[121][122]

In 2003, a study compared the behavior and ethics of chimpanzees, wolves and humans. Cooperation among humans' closest genetic relative is limited to occasional hunting episodes or the persecution of a competitor for personal advantage, which had to be tempered if humans were to become domesticated.[62][123] The closest approximation to human morality that can be found in nature is that of the grey wolf, Canis lupus. Wolves are among the most gregarious and cooperative of animals on the planet,[62][63] and their ability to cooperate in well-coordinated drives to hunt prey, carry items too heavy for an individual, provisioning not only their own young but also the other pack members, babysitting etc. are rivaled only by that of human societies. Similar forms of cooperation are observed in two closely related canids, the African wild dog and the Asian dhole, therefore it is reasonable to assume that canid sociality and cooperation are old traits that in terms of evolution predate human sociality and cooperation. Today's wolves may even be less social than their ancestors, as they have lost access to big herds of ungulates and now tend more toward a lifestyle similar to coyotes, jackals, and even foxes.[62] Social sharing within families may be a trait that early humans learned from wolves,[62][124] and with wolves digging dens long before humans constructed huts it is not clear who domesticated whom.[125][62][123]

Bison surrounded by grey wolf pack

On the mammoth steppe the wolf's ability to hunt in packs, to share risk fairly among pack members, and to cooperate moved them to the top of the food chain above lions, hyenas and bears. Some wolves followed the great reindeer herds, eliminating the unfit, the weaklings, the sick and the aged, and therefore improved the herd. These wolves had become the first pastoralists hundreds of thousands of years before humans also took to this role.[125] The wolves' advantage over their competitors was that they were able to keep pace with the herds, move fast and enduringly, and make the most efficient use of their kill by their ability to "wolf down" a large part of their quarry before other predators had detected the kill.

The study proposed that during the Last Glacial Maximum, some of our ancestors teamed up with those pastoralist wolves and learned their techniques.[62][126] Many of our ancestors remained gatherers and scavengers, or specialized as fish-hunters, hunter-gatherers, and hunter-gardeners. However, some ancestors adopted the pastoralist wolves' lifestyle as herd followers and herders of reindeer, horses, and other hoofed animals. They harvested the best stock for themselves while the wolves kept the herd strong, and this group of humans was to become the first herders and this group of wolves was to become the first dogs.[125][62]

In 1914, on the eve of the First World War, two human skeletons were discovered during basalt quarrying at Oberkassel, Bonn in Germany. With them were found a right mandible of a "wolf" and other animal bones.[127] After the end of the First World War, in 1919 a full study was made of these remains. The mandible was recorded as "Canis lupus, the wolf" and some of the other animal bones were assigned to it.[128] The remains were then stored and forgotten for fifty years. In the late 1970s there was renewed interest in the Oberkassel remains and the mandible was re-examined and reclassified as belonging to a domesticated dog.[129][130][131] The mitochondrial DNA sequence of the mandible was matched to Canis lupus familiaris - dog,[34] and confirms that the Oberkassel dog is a direct ancestor of today's dogs.[132] The bodies were dated to 14,223 YBP.[133] This implies that in Western Europe there were morphologically and genetically "modern" dogs in existence around 14,500 YBP.[134]

Later studies assigned more of the other animal bones to the dog until most of a skeleton could be assembled.[134] The humans were a man aged 40 years and a woman aged 25 years. All three skeletal remains were found covered with large 20 cm thick basalt blocks and were sprayed with red hematite powder.[133] The consensus is that a dog was buried along with two humans.[134] A tooth belonging to a smaller and older dog was also identified but it had not been sprayed with red powder.[133] The cause of the death of the two humans is not known.[134] A pathology study of the dog remains suggests that it had died young after suffering from canine distemper between ages 19 and 23 weeks.[133] The dog could not have survived during this period without intensive human care.[134][133] During this period the dog was of no utilitarian use to humans,[133] and suggests the existence of emotional or symbolic ties between these humans and this dog.[134] In conclusion, near the end of the Late Pleistocene at least some humans regarded dogs not just materialistically, but had developed emotional and caring bonds for their dogs.[133]

Shortly after - around 14,000 YBP - two dogs were buried together in the Grotte du Moulin cave in Troubat, France by the Azilian culture.[135]

Saharan rock art depicting two dogs attacking a mouflon, Algeria during the Horse Period 3,200-1,000 YBP

During the Upper Paleolithic (50,000-10,000 YBP), the increase in human population density, advances in blade and hunting technology, and climate change may have altered prey densities and made scavenging crucial to the survival of some wolf populations. Adaptations to scavenging such as tameness, small body size, and a decreased age of reproduction would reduce their hunting efficiency further, eventually leading to obligated scavenging.[30][136] Whether these earliest dogs were simply human-commensal scavengers or they played some role as companions or hunters that hastened their spread is unknown.[30]

Researchers have proposed that in the past a hunting partnership existed between humans and dogs that was the basis for dog domestication.[137][138][139] Petroglyph rock art dating to 8,000 YBP at the sites of Shuwaymis and Jubbah, in northwestern Saudi Arabia, depict large numbers of dogs participating in hunting scenes with some being controlled on leashes.[140] The transition from the Late Pleistocene into the early Holocene was marked by climatic change from cold and dry to warmer, wetter conditions and rapid shifts in flora and fauna, with much of the open habitat of large herbivores being replaced by forests.[139] In the early Holocene, it is proposed that along with changes in arrow-head technology that hunting dogs were used by hunters to track and retrieve wounded game in thick forests.[138][139] The dog's ability to chase, track, sniff out and hold prey can significantly increase the success of hunters in forests, where human senses and location skills are not as sharp as in more open habitats. Dogs are still used for hunting in forests today.[139]

Sled dog types, sketched in 1833

In 2017, a study showed that 9,000 YBP the domestic dog was present at what is now Zhokhov Island. The dogs were selectively bred as either sled dogs or as hunting dogs, which implies that a sled dog standard and a hunting dog standard existed at that time. The optimal maximum size for a sled dog is 20–25 kg based on thermo-regulation, and the ancient sled dogs were between 16–25 kg. The same standard has been found in the remains of sled dogs from this region 2,000 YBP and in the modern Siberian husky breed standard. Other dogs were more massive at 30 kg and appear to be dogs that had been crossed with wolves and used for polar bear hunting. At death, the heads of the dogs had been carefully separated from their bodies by humans, probably for ceremonial reasons.[141]

The study proposes that after having diverged from the common ancestor shared with the grey wolf, the evolution of the dog proceeded in three stages. The first was natural selection based on feeding behavior within the ecological niche that had been formed through human activity. The second was artificial selection based on tamability. The third was directed selection based on forming breeds that possessed qualities to help with specific tasks within the human economy. The process commenced 30,000-40,000 YBP with its speed increasing in each stage until domestication became complete.[141]

In North America, the earliest dog remains were found in Illinois and radiocarbon dating indicates 9,900 YBP. These include three isolated burials at the Koster Site near the lower Illinois River in Greene County, and one burial 35 km away at the Stilwell II site in Pike County. These dogs were medium-sized adults around 50 cm (20 in) in height and around 17 kilograms (37 lb) in weight, with very active lifestyles and varied morphologies. Isotope analysis can be used to identify some chemical elements, allowing researchers to make inferences about the diet of a species. An isotope analysis of bone collagen indicates a diet consisting largely of freshwater fish. Similar dog burials across Eurasia are thought to be due to the dog's importance in hunting to people who were trying to adapt to the changing environments and prey species during the Pleistocene-Holocene transition. In these places, the dog had gained an elevated social status. [142]

In 2018, a study compared sequences of North American dog fossils with Siberian dog fossils and modern dogs. The nearest relative to the North American fossils was a 9,000 YBP fossil discovered on Zhokhov Island, arctic north-eastern Siberia, which was connected to the mainland at that time. The study inferred from mDNA that all of the North American dogs shared a common ancestor dated 14,600 YBP, and this ancestor had diverged along with the ancestor of the Zhokhov dog from their common ancestor 15,600 YBP. The timing of the Koster dogs shows that dogs entered North America from Siberia 4,500 years after humans did, were isolated for the next 9,000 years, and after contact with Europeans these no longer exist because they were replaced by Eurasian dogs. The pre-contact dogs exhibit a unique genetic signature that is now gone, with nDNA indicating that their nearest genetic relatives today are the arctic breed dogs - Alaskan malamutes, Greenland dogs, and Alaskan huskies and Siberian huskies.[143]

In 2019, a study found that those dogs brought initially into the North American Arctic from northeastern Siberia were later replaced by dogs accompanying the Inuit during their expansion beginning 2,000 years ago. These Inuit dogs were more genetically diverse and more morphologically divergent when compared with the earlier dogs. Today, Arctic sledge dogs are the last descendants in the Americas of this pre-European dog lineage.[144]

In Japan, temperate deciduous forests rapidly spread onto the main island of Honshu and caused an adaption away from hunting megafauna (Naumann's elephant and Yabe's giant deer) to hunting the quicker sika deer and wild boar in dense forest. With this came a change in hunting technology, including a shift to smaller, triangular points for arrows. A study of the Jōmon people that lived on the Pacific coast of Honshu during the early Holocene shows that they were conducting individual dog burials and were probably using dogs as tools for hunting sika deer and wild boar, as hunters in Japan still do today.[139]

Hunting dogs make major contributions to forager societies and the ethnographic record shows them being given proper names, treated as family members, and considered separate to other types of dogs.[139][145] This special treatment includes separate burials with markers and grave-goods,[139][146][147] with those that were exceptional hunters or that were killed on the hunt often venerated.[139][148] A dog's value as a hunting partner gives them status as a living weapon and the most skilled elevated to taking on a "personhood", with their social position in life and in death similar to that of the skilled hunters.[139][149]

Intentional dog burials together with ungulate hunting is also found in other early Holocene deciduous forest forager societies in Europe[150] and North America,[151][152] indicating that across the Holarctic temperate zone hunting dogs were a widespread adaptation to forest ungulate hunting.[139]

In 2020, an mDNA study of ancient dog fossils from the Yellow River and Yangtze River basins of southern China showed that most of the ancient dogs fell within haplogroup A1b, as do the Australian dingoes and the pre-colonial dogs of the Pacific, but in low frequency in China today. The specimen from the Tianluoshan archaeological site, Zhejiang province dates to 7,000 YBP and is basal to the entire lineage. The dogs belonging to this haplogroup were once widely distributed in southern China, then dispersed through Southeast Asia into New Guinea and Oceania, but were replaced in China 2,000 YBP by dogs of other lineages.[153]

Dogs are the most variable mammal on earth, with artificial selection producing around 450 globally recognized dog breeds. These breeds possess distinct traits related to morphology, which include body size, skull shape, tail phenotype, fur type and colour. Their behavioural traits include guarding, herding, and hunting, and personality traits such as hypersocial behavior, boldness, and aggression. Most breeds were derived from small numbers of founders within the last 200 years. As a result, today dogs are the most abundant carnivore species and are dispersed around the world.[54]