It is widely agreed that the evolutionary lineage of thegrey wolf can be traced back 2 million years to theEarly Pleistocene speciesCanis etruscus, and its successor theMiddle PleistoceneCanis mosbachensis.^[2][3] The grey wolfCanis lupus is a highly adaptable species that is able to exist in a range of environments and which possesses a wide distribution across theHolarctic. Studies of modern grey wolves have identified distinct sub-populations that live in close proximity to each other.^[4][5] This variation in sub-populations is closely linked to differences in habitat – precipitation, temperature, vegetation, and prey specialization – which affect cranio-dental plasticity.^[6][7][8][9]

The earliest specimens of the modern grey wolf date to around 400,000 years ago,^[10] or possibly earlier to 1 million years ago.^[11] Most modern wolves share most of their common ancestry within the last 25-23,000 years from earlier Siberian wolf populations.^{[12][13][14][15]} While some sources have suggested that this is the result of apopulation bottleneck, others suggest that this is a normal consequence ofgene flow homogenising wolf genomes across their range.^[16]

The fossil record for ancient vertebrates is composed of rarely occurring fragments from which it is often impossible to obtain genetic material. Researchers are limited tomorphologic analysis but it is difficult to estimate the intra-species and inter-species variations and relationships that existed between specimens across time and place. Some observations are debated by researchers who do not always agree, and hypotheses that are supported by some authors are challenged by others.^[17]

TheCretaceous–Paleogene extinction event occurred 66 million years ago and brought an end to the non-avian dinosaurs and the appearance of the first carnivorans.^[18] The name carnivoran is given to a member of the orderCarnivora. Carnivorans possess a common arrangement of teeth calledcarnassials, in which the first lowermolar and the last upperpremolar possess blade-like enamelcrowns that act similar to a pair of shears for cutting meat. This dental arrangement has been modified by adaptation over the past 60 million years for diets composed of meat, for crushing vegetation, or for the loss of the carnassial function altogether as in seals, sea lions, and walruses. Today, not all carnivorans arecarnivores, such as the insect-eatingaardwolf.^[19]

The carnivoran ancestors of the dog-likecaniforms and the cat-likefeliforms began their separate evolutionary paths just after the end of the dinosaurs. The first members of the dog familyCanidae appeared 40 million years ago,^[20] of which only its subfamily theCaninae survives today in the form of the wolf-like and fox-like canines. The caniforms included the fox-like genusLeptocyon whose various species existed from 34 million YBP before branching 11.9 million YBP intoVulpes (foxes) andCanini (canines). The jackal-sizedEucyon existed in North America from 10 million YBP and by theEarly Pliocene about 6–5 million YBP the coyote-likeEucyon davisi^[21] invaded Eurasia. In North America it gave rise to earlyCanis which first appeared in theMiocene (6 million YBP) in south-western US and Mexico. By 5 million YBP the largerCanis lepophagus appeared in the same region.^[22]

The canids that had immigrated from North America to Eurasia –Eucyon,Vulpes, andNyctereutes – were small to medium-sized predators during the Late Miocene and Early Pliocene but they were not the top predators. The position of the canids would change with the arrival ofCanis to become a dominant predator across theHolarctic. The wolf-sizedC. chihliensis appeared in northern China in the Mid-Pliocene around 4–3 million YBP.^[23] The large wolf-sizedCanis appeared in the MiddlePliocene about 3 million years ago in the Yushe Basin,Shanxi Province, China. By 2.5 million years ago its range included the Nihewan Basin inYangyuan County,Hebei, China and Kuruksay, Tadzhikistan.^[24] This was followed by an explosion ofCanis evolution across Eurasia in the Early Pleistocene around 1.8 million YBP in what is commonly referred to as thewolf event. It is associated with the formation of themammoth steppe and continental glaciation.Canis spread to Europe in the forms ofC. arnensis,C. etruscus, andC. falconeri.^[23] Other studies state that the oldestCanis remains that have been found in Europe were from France and dated to 3.1 million YBP,^[25] followed byCanis cf. etruscus (wherecf. in Latin means confer, uncertain) from Italy dated to 2.2 million YBP.^[26]

The fossil record is incomplete but it is likely that wolves arose from a population of small, early canids.^[27]: p241 Morphological evidence^{[27]: p239 [28]} and genetic evidence^[29] both suggest that wolves evolved during thePliocene and EarlyPleistocene eras from the same lineage that also led to the coyote,^[27]: p239 with fossil specimens indicating that the coyote and the wolf diverged from a common ancestor 1.5 million years ago.^{[27]: p240 [28]} The ancestor of the jackal and the other extant members of the genusCanis had split from the lineage before this time.^[27]: p240

After this separation from a common ancestor the species that were believed to be involved in the further evolution of the wolf and coyote – and the beliefs of somepaleontologists – diverged.^[27]: p240 A number of researchers believed that the lines ofC. priscolatrans,C. etruscus,C. rufus,C. lycaon, andC. lupus were components involved in some way that lead to the modern wolf and coyote.^{[27]: p240 [30][31][32][33][34][35]}

Canis lepophagus lived in the early Pliocene in North America.^[36]Kurten proposed that theBlancanC. lepophagus^[37] derived from smallerMioceneCanis species in North America. It then became widespread across Eurasia where it was either identical to, or closely related with,C. arnensis of Europe.^{[27]: p241 [38][39]}

Johnston describesC. lepophagus as having a more slender skull and skeleton than in the modern coyote.^[40]: 385 Robert M. Nowak found that the early populations had small, delicate and narrowly proportioned skulls that resemble small coyotes and appear to be ancestral toC. latrans.^[27]: p241 Johnson noted that some specimens found in Cita Canyon, Texas, had larger, broader skulls,^[40] and along with other fragments Nowak suggested that these were evolving into wolves.^{[27]: p241 [28]}

Tedford disagreed with previous authors and found that its cranio-dental morphology lacked some characteristics that are shared byC. lupus andC. latrans, and therefore there was not a close relationship but it did suggestC. lepophagus was the ancestor of both wolves and coyotes.^[11]: p119

The North American wolves became larger, with tooth specimens indicating thatC. priscolatrans diverged into the large wolfC. armbrusteri.^{[27]: p242 [41]} during the Middle Pleistocene in North America.^[28] Robert A. Martin disagreed, and believed thatC. armbrusteri^[42] wasC. lupus.^[33] Nowak disagreed with Martin and proposed thatC. armbrusteri was not related toC. lupus butC. priscolatrans, which then gave rise toA. dirus. Tedford proposed that the South AmericanC. gezi andC. nehringi share dental and cranial similarities developed for hypercarnivory, suggestingC. armbrusteri was the common ancestor ofC. gezi,C. nehringi andA. dirus.^[11]: 148

In 1908 the paleontologistJohn Campbell Merriam began retrieving numerous fossilized bone fragments of a large wolf from the Rancho La Brea tar pits. By 1912 he had found a skeleton sufficiently complete to be able to formally recognize these and the previously found specimens under the nameC. dirus (Leidy 1858)^{[citation needed]}.

Canis dirus,^[43] or as it is now widely recognized,Aenocyon dirus, lived in the late Pleistocene to early Holocene in North and South America.^[44] In 1987, a new hypothesis proposed that a mammal population could give rise to a larger form called a hypermorph during times when food was abundant, but when food later became scarce the hypermorph would either adapt to a smaller form or go extinct. This hypothesis might explain the large body sizes found in many Late Pleistocene mammals compared to their modern counterparts. Both extinction andspeciation – a new species splitting from an older one – could occur together during periods of climatic extremes.^[45][46] Gloria D. Goulet agreed with Martin and further proposed that this hypothesis might explain the sudden appearance ofA. dirus in North America, and that because of the similarities in their skull shapes thatC. lupus gave rise to theA. dirus hypermorph due to abundant game, a stable environment, and large competitors.^[47] Nowak, Kurten and Berta disagreed with Goulet and proposed thatA. dirus was not derived fromC. lupus.^[28][39][48] The three noted paleontologists Xiaoming Wang, R. H. Tedford and R. M. Nowak have all proposed thatA. dirus had evolved fromC. armbrusteri,^{[49][11]: 181} with Nowak stating that there were specimens from Cumberland Cave, Maryland, that indicatedC. armbrusteri diverging intoA. dirus.^{[27]: p243 [50]} The two taxa share a number of characteristics (synapomorphy), which suggests an origin ofA. dirus in the lateIrvingtonian in the open terrain in the midcontinent, and then later expanding eastward and displacing its ancestorC. armbrusteri.^[11]: 181

However, in 2021, a study indicated the dire wolf to be a highly divergent lineage which last shared amost recent common ancestor with the wolf-like canines 5.7 million years ago. The morphological similarity between dire wolves and gray wolves was concluded to be due toconvergent evolution. This finding indicates that the wolf and coyote lineages evolved in isolation from the dire wolf lineage. The study proposes an early origin of the dire wolf lineage in the Americas, and that this geographic isolation allowed them to develop a degree ofreproductive isolation since their divergence 5.7 million years ago. Coyotes, dholes, gray wolves, and the extinctXenocyon evolved in Eurasia and expanded into North America relatively recently during the Late Pleistocene, therefore there was no admixture with the dire wolf. As a result, the study found that the correct binomial name of the dire wolf isAenocyon dirus, as proposed by Merriam in 1918. The long-term isolation of the dire wolf lineage implies that other American fossil taxa, includingC. armbrusteri andC. edwardii, may also belong to the dire wolf's lineage.^[51]

Wang andTedford proposed that the genusCanis was the descendant of the coyote-likeEucyon davisi, and its remains first appeared in theMiocene (6 million YBP) in south-western USA and Mexico. By thePliocene (5 million YBP), the largerCanis lepophagus appeared in the same region and by theEarly Pleistocene (1 million YBP)Canis latrans (thecoyote) was in existence. They proposed that the progression fromEucyon davisi toC lepophagus to the coyote was linear evolution.^[22] Additionally,C. edwardii,C. latrans andC. aureus form together a small clade and becauseC. edwardii appeared earliest spanning the mid-Blancan (late Pliocene) to the close of the Irvingtonian (late Pleistocene) it is proposed as the ancestor.^{[11]: p175, 180}

Nowak and Tedford also believed that it was possible forC. lupus to have been derived from a Miocene or Pliocene canid line that preceded and was separate fromC. lepophagus.^[28][52] Based on morphology from China, thePliocene wolfC. chihliensis may have been the ancestor for bothC. armbrusteri andC. lupus before their migration into North America.^{[23][11]: p181}C. chihliensis appears to be more primitive and smaller thanC. lupus, and measurements of its skull and teeth are similar toC. lupus but those of its postcranial elements are smaller.^[53]C. armbrusteri appeared in North America in theMiddle Pleistocene and is a wolf-like form larger than anyCanis at that time.^[28] At the end of the most recent glacial retreat during the past 30,000 years, warming melted the glacial barriers across northern Canada allowing arctic mammals to extend their range into mid-latitude North America, including elk, caribou, bison, and the grey wolf.^[54]

In Eurasia during the Middle Pleistocene,C. falconeri gave rise to the hypercarnivore genusXenocyon, which then gave rise to genusCuon (the dhole) and genusLycaon (the African hunting dog).^[55][56] Just before the appearance ofC. dirus, North America was invaded by genusXenocyon that was as large asC. dirus and more hypercarnivorous. The fossil record shows them as rare and it is assumed that they could not compete with the newly derivedC. dirus.^[57]

The large wolfC. antonii from late Pliocene to early Pleistocene China was assessed as being a variation withinC. chihliensis,^[11]: p197 and the large wolfC. falconeri occurred abruptly in Europe in the Early Pleistocene, perhaps representing a westward extension ofC. antonii.^[11]: p181

In 2020, researchers named a new speciesC. borjgali that was found inDmanisi, Georgia in a site dated 1.8—1.75 million years ago.^[58] This specimen did not show the peculiarities ofC. etruscus but appears to be closer to a primitive form ofC. mosbachensis and is proposed as the ancestor ofC. mosbachensis.^[59][58] The wolfC. borjgali is probably the ancestor of the wolf-likecrown-clade speciesC. lupus,C. latrans, andC. lupaster.^[59]

Canis mosbachensis, sometimes known as the Mosbach wolf, is an extinct small wolf that once inhabited Eurasia from theMiddle toLate Pleistocene.^[60] The phylogenetic descent of the extant wolfC. lupus fromC. etruscus throughC. mosbachensis is widely accepted.^{[60][27]: 239–245} In 2010, a study found that the diversity of theCanis group decreased by the end of theEarly Pleistocene toMiddle Pleistocene and was limited in Eurasia to the small wolves of theC. mosbachensis–C. variabilis group that were a comparable size to the extantIndian wolf (Canis lupus pallipes), and the large hypercarnivorousCanis (Xenocyon)lycanoides that was comparable in size to extant northern grey wolves.^[58]

The earliestCanis lupus specimen was a fossil tooth discovered atOld Crow, Yukon, Canada. The specimen was found in sediment dated 1 million YBP,^[11] however the geological attribution of this sediment is questioned.^[11][61] Slightly younger specimens were discovered at Cripple Creek Sump,Fairbanks, Alaska, in strata dated 810,000 YBP. Both discoveries point to an origin of these wolves in eastBeringia during theMiddle Pleistocene.^[11]

In France, the subspeciesC. l. lunellensis Bonifay, 1971^[62] discovered atLunel-Viel,Hérault dated 400–350,000 YBP,C. l. santenaisiensis Argant, 1991^[63] fromSantenay, Côte-d'Or dated to 200,000 YBP, andC. lupus maximus Boudadi-Maligne, 2012^[64] from Jaurens cave,Nespouls,Corrèze dated 31,000 YBP, show a progressive increase in size and are proposed to bechrono-subspecies.^[65][17] In Italy, the earliestCanis lupus specimens were found at La Polledrara di Cecanibbio, 20 km north-west of Rome in strata dated 340,000–320,000 YBP.^[66][67] In 2017, a study found that the dimensions of the upper and lower carnassial teeth of the early Holocene Italian wolf are close to those ofC. l. maximus. Fluctuations in the size ofC. lupus carnassial teeth correlate with the spread of megafauna. The Italian wolf underwent a reduction in body size with the loss of the red deer in Italy during the Renaissance.^[17] The proposed lineage is:

C. etruscus → C. mosbachensis → C. l. lunellensis → C. l. santenaisiensis → C. l. maximus → C. l. lupus^[17]

In 2022, a new subspeciesCanis lupus bohemica wastaxonomically described after having been discovered in the Bat Cave system located nearSrbsko, Central Bohemia, Czech Republic. The Bohemian wolf is an extinct short-legged wolf that first appeared 800,000 years ago (MIS 20, the Rhumian Interglacial of the earlyCromerian stage, Middle Pleistocene) and once inhabited what was part of the mammoth steppe. It is proposed as the ancestor ofCanis lupus mosbachensis. In comparison,C. etruscus appears to be the ancestor of theAfro-Eurasian jackal.^[68]

In Hungary in 1969, a tooth (thepremolar of theMaxilla) was found which dated to the Middle Pleistocene, and was assessed as being midway between that ofCanis mosbachensis and thecave wolfCanis lupus spelaeus, but leaning towardsC.l. spelaeus.^[69] During the late Middle Pleistocene around 600,000 years ago, the Bohemian wolf diversified into two wolf lineages that specialized for different environmental and climatic conditions. One was a southern interglacial (warm climate) grey wolf of Europe which was to become the Mosbach wolf, and the other a northern glacial white wolf of Eurasia which was to becomeC. l. spelaeus.^[68]

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 either Europe or Siberia and date 40,000-17,000 YBP. They includeHohle Fels in Germany,Goyet Caves in Belgium,Predmosti in the Czech Republic, and four sites in Russia: Razboinichya Cave,Kostyonki-8, Ulakhan Sular, and Eliseevichi 1. Paw-prints fromChauvet 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.^[70] Paleolithic dogs were directly associated with human hunting camps in Europe over 30,000 (YBP) and it is proposed that they were domesticated. They are also proposed to be either a proto-dog and the ancestor of the domestic dog or a type of wolf unknown to science.^[71]

In 2002, a study was undertaken into the fossil skulls of two large canids that had been found buried within meters of the doorway of what was once a mammoth-bone hut at the Eliseevichi-IUpper Paleolithic site in theBryansk Region on the Russian Plain, and using an accepted morphologically based definition of domestication declared them to be "Ice Age dogs". The carbon dating gave a calendar-year age estimate that ranged between 16,945 and 13,905 YBP.^[72] In 2013, a study looked at one of these skulls and its mitochondrial DNA sequence was identified asCanis familiaris.^[73]

In 2015, a zooarchaeologist stated that "In terms of phenotypes, dogs and wolves are fundamentally different animals."^[74]

In 1986, a study of skull morphology found that the domestic dog is morphologically distinct from all other canids except the wolf-like canids. "The difference in size and proportion between some breeds are as great as those between any wild genera, but all dogs are clearly members of the same species."^[75] In 2010, a study of dog skull shape compared to extantcarnivorans proposed that "The greatest shape distances between dog breeds clearly surpass the maximum divergence between species in the Carnivora. Moreover, domestic dogs occupy a range of novel shapes outside the domain of wild carnivorans."^[76]

The domestic dog compared to the modern wolf shows the greatest variation in the size and shape of the skull (Evans 1979) that range from 7 to 28 cm in length (McGreevy 2004). Wolves aredolichocephalic (long skulled) but not as extreme as some breeds of such as greyhounds and Russian wolfhounds (McGreevy 2004). Caninebrachycephaly (short-skulledness) is found only in domestic dogs and is related topaedomorphosis (Goodwin 1997). Puppies are born with short snouts, with the longer skull of dolichocephalic dogs emerging in later development (Coppinger 1995). Other differences in head shape between brachycephalic and dolichocephalic dogs include changes in the craniofacial angle (angle between thebasilar axis andhard palate) (Regodón 1993), morphology of thetemporomandibular joint (Dickie 2001), and radiographic anatomy of thecribriform plate (Schwarz 2000).^[77]

Nowak indicated that orbital angle of the eye socket is an important characteristic defining the difference between the dog and the wolf, with the wolf having the lower angle. Nowak compared the orbital angles of four North Americancanines (including the Indian dog) and produced the following values in degrees: coyote-42.8, wolf-42.8, dog-52.9 dire wolf-53.1. The orbital angle of the eye socket was clearly larger in the dog than in the coyote and the wolf; why it was almost the same as that of the dire wolf was not commented on.^[28]

Many authors have concluded that compared to the adult extant wolf, the adult domestic dog has a relatively reduced rostrum (front part of the skull), an elevatedfrontal bone, a widerpalate, a broadercranium, and smaller teeth (Hildebrand1954; Clutton-Brock, Corbet & Hills 1976; Olsen 1985; Wayne 1986; Hemmer 1990; Morey 1990). Other authors have disagreed and have stated that these traits can overlap and vary within the two (Crockford 1999; Harrison 1973). Wolf cubs have similar relative skull proportions as adult dogs and this was proposed as evidence that the domestic dog is aneotenic wolf. This was proposed to be due to either human selection for juvenile appearance or due to apleiotropic effect as a result of selection for juvenile behavior (Clutton-Brock 1977; Belyaev 1979; Wayne 1986; Coppinger and Schneider 1995). Wayne (1986) concluded that his dog samples did not have significant relative shortening of the rostrum compared to wolves, calling this identifying feature into question.^[78] A 2004 study that used 310 wolf skulls and over 700 dog skulls representing 100 breeds concluded that the evolution of dog skulls can generally not be described by heterochronic processes such as neoteny although some pedomorphic dog breeds have skulls that resemble the skulls of juvenile wolves.^[79] "Dogs are not paedomorphic wolves."^[80]

Compared to the wolf, dog dentition is relatively less robust (Olsen 1985; Hemmer 1990), which is proposed to be due to the relaxation of natural selection when wolves became commensal scavengers, or to artificial selection (Olsen 1985; Clutton-Brock 1995). However, Kieser and Groeneveld (1992) compared the mandibulo-dental measurements of jackals (C. adustus, C. mesomelas) and Cape foxes (Vulpes chama) to equivalent-sized dogs and found that the canines of these other canids tended to be slightly smaller and their second molars larger compared to dogs, otherwise the proportions were essentially the same in all species. They concluded that "...the teeth of canids appear to have evolved in concert with one another and relatively independently of differences in dimorphism, size or functional demands". This calls into question the assumption that dog teeth are relatively small due to recent selection, suggesting that dog dentition is plesiomorphic from an ancestor that was smaller than the wolf.^[78]

The reduced body size of the early dog compared to a wolf is thought due to niche selection (Olsen 1985; Morey 1992; Coppinger &Coppinger 2001). Morey (1992:199) states that "Results...are consistent with a hypothesis that early domestic dogs are evolutionary paedomorphs, products of strong selection for ontogenetically channeled size reduction and alterations of reproductive timing associated with the new domestic way of life."^[78] However, in a domestication experiment the domesticated foxes remained the same size as unselected foxes (Trutt 1999:167).^[74]

Wayne (1986) concluded that the dog is closer in skull morphology toC. latrans, C. aureus, C. adustus, C. mesomelas, Cuon alpinus andLycaon pictus than to the wolf. Dahr (1942) concluded that the shape of the dog brain case is closer to that of the coyote than to that of the wolf. Manwell and Baker (1983) reviewed Dahr's work with the addition of dental data for canids and concluded that the dog ancestor was probably within the range of 13.6–20.5 kg, which is smaller than the range 27–54 kg for extant wolves (Mech 1970) and is comparable with theDingo.^[78]

The auditorybulla of the dog is relatively smaller and flatter than that of the wolf (Harrison 1973; Clutton-Brock, Corbet & Hill 1976; Nowak 1979; Olsen 1985; Wayne 1986), which is proposed to be due to relaxed selection under domestication as the dog no longer required the acute hearing of the wolf. However, bulla shape has been shown to facilitate increased sensitivity to specific frequencies but shape and size may not be correlated with acuity (Ewer 1973). Therefore, the observed difference could be that the dog bulla has retained its ancestral shape.^[78]

The ventral edge of the dog's horizontalramus of the mandible has a convex curve that does not exist in the wolf (Olsen 1985; Clutton-Brock 1995), and no discussion of this difference could be found in the literature. However, Biknevicius and Van Valkenburgh (1997) noticed that the horizontal ramus of bone-processing predators is thicker dorso-ventrally at the pointcaudal to the site of bone processing. This thickening may have been a function for niche adaptation by the dog's ancestor.^[78]

A description of the superficial brain morphology of jackals (C. mesomelas, C. aureus), coyotes (C. latrans), wolves (C. lupus, C. rufus), and dogs indicated that thecerebellum of the dog closely approximates that of the coyote, which is closely aligned with the jackals, and that the wolves show numerous brain traits distinct from the other species (Atkins and Dillon 1971). Wolves also have serological and biochemical traits distinct from dogs (Leone and Wiens 1956; Lauer, Kuyt & Baker 1969).^[78]

The tails of domestic dogs tend to curl upwards, which is not found in other canid.^[81] Dogs possess a domed forehead, which is easy to see in a profile view with a bulge above theorbit (the bony socket) of the eye. Thetemporalis muscle that closes the jaws is more robust in wolves.^[82] Compared with equally sized wolves, dogs possess 20% smaller skulls, and compared with equally weighted wolves dogs possess 20% smaller brains.^[83]: 47 Thetympanic bullae are large, convex, and almost spherical in wolves, while the bullae of dogs are smaller, compressed, and slightly crumpled.^[84] Dogs have a distinctive "stop" between the forehead and nose.^[85] Domestic dogs are distinguishable from wolves by starch gelelectrophoresis of red blood cell acidphosphatase.^[86]

During the Last Glacial Maximum, there was greater wolfgenetic diversity than there is today,^[12][73] and within the Pleistocene grey wolf population the variations between local environments would have encouraged a range of wolf ecotypes that were genetically, morphologically and ecologically distinct from one another.^[87] One author has proposed that the most likely explanation for the different morphological characteristics of the dog compared to the wolf is that the dog's ancestor was adapted to a different niche than the wolf.^[78]

Themitochondria within eachcell contain many copies of a small circularDNA genome and in mammals it is 16,000–18,000base pairs in length. A cell contains hundreds or thousands of mitochondria and therefore thegenes contained within those mitochondria are more abundant than the genes that occur in the nucleus of the cell.^[88][89] The abundance ofmitochondrial DNA (mDNA) is useful for the genetic analysis of ancient remains where the DNA has degraded.^[89][90]

MitochondrialDNA sequences have a highermutation rate than the mutation rate of nuclear genes and for mammals this rate is 5–10 times faster.^[89][91][92] The mitochondrial protein-coding genes evolve much faster and are powerful markers for inferring evolution history at category levels such as families, genera, and species. However, they have evolved at a faster rate than other DNA markers and there is a timing difference in its molecular clock that needs to be validated against other sources. The taxonomic status of uncertain species is better resolved through usingnuclear DNA from the nucleus of the cell, which is more suitable for analyzing the recent history.^[93] In most cases, mDNA is inherited from the maternal ancestor.^[89][94] Therefore,phylogenetic analysis of mDNA sequences within species provides a history of maternal lineages that can be represented as aphylogenetic tree.^[89][95][96]

The mDNA sequences of the dog and wolf differ by only 0–12 substitutions within 261 base-pairs, whereas dogs always differed from coyotes and jackals by at least 20 substitutions.^[89][99] This finding implies that the dog derived from the wolf and that there has been repeated back-crossing,^[99] or that the dog may have descended from a now extinct species of canid whose closest living relative is the modern wolf.^[100]

Different DNA studies may give conflicting results because of the specimens selected, the technology used, and the assumptions made by the researchers.^[101] Any one from a panel ofgenetic markers can be chosen for use in a study. The techniques used toextract,locate andcompare genetic sequences can be applied using advances in technology, which allows researchers to observe longer lengths ofbase pairs that provide more data to give betterphylogenetic resolution.^[102]Phylogenetic trees compiled using differentgenetic markers have given conflicting results on the relationship between the wolf, dog and coyote. One study based onSNPs^[103] (a singlemutation), and another based onnuclear gene sequences^[104] (taken from thecell nucleus), showed dogs clustering with coyotes and separate from wolves. Another study based on SNPS showed wolves clustering with coyotes and separate from dogs.^[105] Other studies based on a number of markers show the more widely accepted result of wolves clustering with dogs separate from coyotes.^[99][106] These results demonstrate that caution is needed when interpreting the results provided by genetic markers.^[103]

There are two key assumptions that are made for dating the divergence time for species: thegeneration time and the geneticmutation rate per generation. The time between generations for wolves is assumed to be three years based on the extant grey wolf, and two years for the dog based on the extant dog.^[97] One recent major study assumed a generation time of 2 years for the dog for as far back as 10,000 years ago, and then assumed a generation time of 3 years (the same as the wolf) before that to calculate a proposed divergence time between the two.^[12] In 2017, the wolf research scientistL. David Mech queried why evolutionary biologists were calculating the approximate time of the dog diverging from the wolf through using a wolf generation time of three years when published works using large data sets demonstrate a figure of 4.2–4.7 years. They were encouraged to recalculate their divergence dates accordingly.^[107]

DNA studies are conducted but with "the mutation rate as the dominant source of uncertainty."^[12] In 2005, Lindblad-Toh sequenced the first draft genome of the extant dog, and calculated a proposed mutation rate of 1x10⁻⁸ mutations per generation.^[97] In 2015, Skoglund was able to sequence the first draft genome of the 35,000 YBPTaimyr wolf and used its radio-carbon date to validate a proposed genetic mutation rate of 0.4x10⁻⁸ mutations per generation.^[14] The difference is a timing factor of 2.5, however another study stated that because only one Pleistocene wolf specimen has so far been sequenced, then the result should be treated with caution, with that study then providing both estimates to calculate the proposed divergence times between the wolf and dog.^[13] However, in 2016 the mutation rate of the 4,800 YBPNewgrange dog matched that of the Taimyr wolf.^[108]

The wolf-like canids (the canid subfamilyCaninae) are a group of large carnivores that are genetically closely related because theirchromosomes number 78. The group includes genusCanis,Cuon andLycaon. The members are thedog(C. familiaris),grey wolf (C. lupus),coyote (C. latrans),golden jackal (C. aureus),Ethiopian wolf (C. simensis),black-backed jackal (Lupulella mesomelas),side-striped jackal (Lupulella adusta),dhole (Cuon alpinus), andAfrican wild dog (Lycaon pictus).^{[109][110][111]} Newly proposed members include thered wolf (Canis rufus),eastern wolf (Canis lycaon), andAfrican golden wolf (C. anthus). As they possess 78 chromosomes, all members of the genusCanis (coyotes, wolves, jackals) arekaryologically indistinguishable from each other, and from the dhole and the African hunting dog.^{[89]: p279 [112]} The members ofCanis can potentiallyinterbreed^[100] and there is evidence that the Ethiopian wolf has hybridized with dogs.^[113] According to zoologistReginald Pocock, a dhole interbred with a golden jackal.^[114] The African hunting dog is large, highly mobile, known to disperse over large distances and are rare throughout much of their geographical range,^[115] making opportunities for hybridization difficult. A study of the maternalmitochondrial DNA of the black-backed jackal could find no evidence ofgenotypes from the most likely mates – the side-striped jackal nor the golden jackal – indicating that male black-backed jackals had not bred with these.^[116] A search of the scientific literature could not find evidence of hybridization for the rare side-striped jackal.

A DNA sequence alignment for the wolf-like canids gave a phylogenetic tree with the grey wolf and dog being the most closely related, followed by a close affiliation with the coyote, golden jackal and Ethiopian wolf, and the dog can hybridize in the wild with these three species. Next closest to this group are the dhole and African wild dog that both have unique meat-slicing teeth, suggesting that this adaptation was later lost by the other members.^[97] The two African jackals are shown as the mostbasal members of this clade, which means that this tree is indicating an African origin for the clade.^[97][117] The tree illustrates thegenotype–phenotype distinction, where agenotype is an organism's fullhereditary information and aphenotype is an organism's actual observed properties, such asmorphology,development, orbehavior. By phenotype, the dhole (genusCuon) and the African hunting dog (genusLycaon) are not classified as members of the genusCanis, but by genotype they are closer to dogs, wolves and coyotes than are the two genusCanis jackals – the Side-striped jackal (C. adustus) and the Black-backed jackal (C. mesomelas).

In 2015, a study of mitochondrial genome sequences and nuclear genome sequences of African and Eurasian canids indicated that extant wolf-like canids had colonized Africa from Eurasia at least 5 times throughout the Pliocene and Pleistocene, which is consistent with fossil evidence suggesting that much of the African canid diversity resulted from the immigration of Eurasian ancestors, likely coincident with Plio-Pleistocene climatic oscillations between arid and humid conditions.^[98]

The phylogenetic tree for the wolf-like canids may give conflicting positions for the black-backed jackal and the side-striped jackal relative to the genusCanis members depending on whether the genetic markers were based on mitochondrial DNA or nuclear DNA. The explanation proposed for this mito-nuclear discord is that mitochondrial DNAintrogression occurred from an ancient ancestor of genusCanis into the lineage that led to the black-backed jackal around 6.2–5.2 million years ago.^[118]

Phylogenetic tree of the extant wolf-like canids,^[a] with the pink shading representing the speciesCanis lupus.

	Side-striped jackal Black-backed jackal 2.62 mya African wild dog Dhole Ethiopian wolf Golden jackal African golden wolf Coyote Himalayan wolf Indian plains wolf †Late Pleistocene   wolf Modern grey wolf Dog 25 kya 80 kya 120 kya 240 kya 380 kya 600 kya 960 kya 2.0 mya 2.5 mya 3.0 mya
3.5 mya

In 2018,whole genome sequencing was used to compare members of genusCanis, along with the dhole (Cuon alpinus) and the African hunting dog (Lycaon pictus). There is evidence of gene flow betweenAfrican golden wolves,golden jackals, and grey wolves. The study suggests that the African golden wolf is a descendant of a geneticallyadmixed canid of 72% greywolf and 28%Ethiopian wolf ancestry, and that the Ethiopian wolf once had a wider range in Africa. One African golden wolf from the EgyptianSinai Peninsula showed high admixture with the Middle Eastern grey wolves and dogs, highlighting the role of the land bridge between the African and Eurasian continents in canid evolution. There is evidence of gene flow between golden jackals and Middle Eastern wolves, less so with European and Asian wolves, and least with North American wolves. The study proposes that the golden jackal ancestry found in North American wolves may have occurred before the divergence of the Eurasian and North American grey wolves. The study indicates that the common ancestor of thecoyote and grey wolf has geneticallyadmixed with aghost population of an extinct unidentified canid. The canid is genetically close to thedhole and has evolved after the divergence of the African hunting dog from the other canid species. Thebasal position of the coyote compared to the wolf is proposed to be due to the coyote retaining more of the mitochondrial genome of this unknown canid.^[119]

A genomic study on the wolves of China included museum specimens of wolves from southern China that were collected between 1963 and 1988. The wolves in the study formed 3 clades: north Asian wolves that included those from northern China and eastern Russia,Himalayan wolves from the Tibetan Plateau, and a unique population from southern China. One specimen located as far southeast asJiangxi province shows evidence of being admixed between Tibetan-related wolves and other wolves in China. One specimen fromZhejiang province in eastern China shared gene flow with the wolves from southern China, however its genome was 12-14 percent admixed with a canid that may be the dhole or an unknown canid that predates the genetic divergence of the dhole. The wolf population from southern China is believed to be still existing in that region.^[132]

Grey wolves suffered a species-widepopulation bottleneck (reduction) approximately 25,000 YBP during the Last Glacial Maximum. This was followed by a single population of modern wolves expanding out of aBeringia refuge to repopulate the wolf's former range, replacing the remaining Late Pleistocene wolf populations across Eurasia and North America as they did so.^{[15][130][131]} 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.^[131]

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.^[131]

As of 2020, the oldest known intact wolf remains belongs to a mummified pup dated 56,000 YBP that was recovered from the permafrost along a small tributary of Last Chance Creek nearDawson City, Yukon, Canada. A DNA analysis showed that it belonged to the Beringian wolf clade, that themost recent common ancestor of this clade dates to 86,700–67,500 YBP, and that this clade was basal to all other wolves except for the Himalayan wolf.^[133]

In 2016, a study built on the work of another major study^[73] and analyzed the sequences of 12 genes that are located on theheavy strand of the mitochondrial genome of extinct and modernC. lupus. The study excluded the sequences of the divergentHimalayan wolf and theIndian grey wolf. The ancient specimens wereradiocarbon dated andstratagraphically dated, and together with the sequences generated a time-based phylogenetic tree. From the tree, the study was able to infer the most recent common ancestor for all otherC. lupus specimens – modern and extinct – was 80,000 YBP and this date concurred with the earlier study.^[73][129] The study could find no evidence of a population bottleneck for wolves until a few thousand years ago.^[129]

The phylogenetic tree showed thepolyphyly of American wolves, the Mexican wolf was divergent from other North American wolves, and these other North American wolves formed two closely related clades. A scenario consistent with the phylogenetic, ice sheet and sea-level data was that during the Ice Age when sea levels were at their lowest, there was a single wave of wolf colonization into North America starting with the opening of theBering land bridge 70,000 YBP and closing during theLate Glacial Maximum of the Yukon corridor that ran through the division between theLaurentide Ice Sheet and theCordilleran Ice Sheet 23,000 YBP. Mexican wolves were part of the single wave and either diverged from the other wolves before entering North America or once in North America due to the change in its environment.

As wolves had been in the fossil record of North America but modern wolves could trace their ancestry back only 80,000 years, the wolf haplotypes that were already in North America were replaced by these invaders, either through competitive displacement or through admixture.^[129] The replacement in North America of a basal population of wolves by a more recent one supported the findings of earlier studies.^{[126][134][125][129]} There possibly existed a panmictic wolf population with gene flow spanning Eurasia and North America until the closing of the ice sheets.^{[126][135][129]} Once the sheets closed, the southern wolves were isolated and north of the sheets only theBeringian wolf existed. The land bridge became inundated by the sea 10,000 YBP, the sheets receded 12,000–6,000 YBP, the Beringian wolf went extinct and the southern wolves expanded to recolonize the rest of North America. All North American wolves are descended from those that were once isolated south of the ice sheets. However, much of their diversity was later lost during the twentieth century.^[129]

Studies usingmitochondrial DNA have indicated that the wolves of coastal south-east Alaska are genetically distinct from inland grey wolves, reflecting a pattern also observed in other taxa. They show a phylogenetic relationship with extirpated wolves from the south (Oklahoma), indicating that these wolves are the last remains of a once widespread group that has been largely extirpated during the last century, and that the wolves of northern North America had originally expanded from southern refuges below theWisconsin glaciation after the ice had melted at the end of theLast Glacial Maximum.^{[136][137][138]} Awhole-genome DNA study indicated that all North American wolves weremonophyletic and therefore are the descendants of a common ancestor.^[139]

During the same period, theSoya Strait betweenHokkaido andSakhalin Island was dry for 75,000 years and it was proposed that the extinctEzo wolf (C. l. hattai) arrived on Hokkaido from Sakhalin.^{[140][129][141]} However, the sequences indicated that it arrived in Hokkaido less than 10,000 YBP. The Ezo wolf was closely related to one of the North American clades,^{[140][129][142]} but different from the more southerlyJapanese wolf (C. l. hodophilax) that was basal to modern wolves.^[140][129] The Japanese wolf inhabited Kyushu, Shikoku, and Honshu islands^[143][144] but not Hokkaido Island.^[144] This indicates that its ancestor may have migrated from the Asian continent through the Korean Peninsula into Japan.^[140][144] The past sea levels of theKorean Strait together with the timing of the Japanese wolf sequences indicated that it arrived to the southern islands less than 20,000 YBP.^[129] In 2020, a genomic study found that the Japanese wolf was the last of the Siberian Pleistocene wolves, which were thought to have gone extinct at the end of the Late Pleistocene (11,700 years ago). Some of these had survived into the 20th Century and had admixed with Japanese dogs.^[145]

The dog was a very successful invader of North America and had established a widespreadecological niche by the Early–Middle Holocene. There was no overlap in niche between the dog and the wolf in comparison to the dog and other North American canids. By the Late Holocene, the dog's niche area was less in size than researchers had expected to find, indicating that it was limited bybiotic factors. These regions include the northeast and northwest of the United States that correlate with the greatest densities of early human occupation, indicating that the dog had "defected" from the wolf niche to the human niche and explains why the dog's niche area was not as large as expected. The separation between dog and wolf may reflect the rapid rate in which domestication occurred,^[146] including the possibility of a second domestication event occurring in North America.^[147][146] Packs of wolves and hunter-gatherers hunt similar prey in a similar way within a similar group social structure that may have facilitated wolf domestication.^[148][149]

The wolf was exterminated in the southern part of their historic geographical range in North America by the middle of the 20th century. An mDNA study of 34 wolf remains from North America dated between 1856 and 1915 found their genetic diversity to be twice that of modern wolves in these regions, and two thirds of the haplotypes identified were unique. These results indicate that a historic population of several hundred thousand wolves once existed in Mexico and the western US.^[150][128]

In 1993, a study proposed that the wolves of North America display skull traits more similar towards the coyote than those wolves from Eurasia.^[47] In 2016, awhole-genome DNA study proposed, based on the assumptions made, that all of the North American wolves and coyotes diverged from a common ancestor less than 6,000–117,000 years ago. The study also indicated that all North American wolves have a significant amount of coyote ancestry and all coyotes some degree of wolf ancestry, and that thered wolf andeastern wolf are highlyadmixed with different proportions of grey wolf and coyote ancestry. However, the test that was used cannot differentiate between ancient and recent hybridization. One test indicated a wolf/coyote divergence time of 51,000 years before present that matched other studies indicating that the extant wolf came into being around this time. Another test indicated that the red wolf diverged from the coyote between 55,000 and 117,000 years before present and the Great Lakes region wolf 32,000 years before present. But a 2000 study on the haplotypes of North American canids showed that the divergences between coyotes, grey wolves, red wolves, and Algonquin (eastern) wolves happened much earlier. This study showed that there was a 3.2 % divergence between the eastern Canadian wolf and coyote gene sequences, and a 2.3 % sequence divergence between the red wolf and coyote haplotypes. However, the divergence was a much greater 8.0% between gray wolf (C. lupus) mtDNA and eastern/red wolf haplotypes, and 10.0% between gray wolf and coyote haplotypes. The sequence difference observed between eastern Canadian wolf sequences and coyote sequences is consistent with a separation of 150 000 – 300 000 years, using a divergence rate of 1–2% per 100 000 years.^[151] Additionally, conflicting with the recent divergent hypothesis, the fossil record that indicates a coyote-like specimen dated to 1 million years before present.^[36]

The modern grey wolf expanded out of Beringia 25,000 years ago.^[131] The modern coyote appeared around 10,000 years ago, and most likely descended from thePleistocene coyote. The most genetically basal coyote mDNA clade pre-dates theLast Glacial Maximum and is a haplotype that can only be found in the Eastern wolf. This implies that the large, wolf-likePleistocene coyote was the ancestor of theEastern wolf andRed wolf, which diverged from each other later in time.

The domestic dog (Canis lupus familiaris) is the most widely abundant large carnivore.^{[73][13][152]} Over the past million years, numerous wolf-like forms existed but their turnover has been high, and modern wolves are not the lineal ancestors of dogs.^{[73][12][13][150]} Although research had suggested that dogs and wolves were genetically very close relatives,^{[99][100][109]} laterphylogenetic analysis strongly supported the hypothesis that dogs and wolves are reciprocallymonophylictaxa that form two sisterclades.^{[99][12][153]} This suggests that none of the modern wolf populations are related to the wolves that were first domesticated and the wolf ancestor of dogs is therefore presumed extinct.^[12][153] Recentmitochondrial DNA analyses of ancient and modern grey wolf specimens supports a pattern of population reduction and turnover.^{[73][126][125]} An alternate proposal is that during the ecological upheavals of the Late Pleistocene all of the remaining members of a dwindling lineage joined humans.^[154][74]

In 2016, a study investigated for the first time the population subdivisions, demography, and the relationships of grey wolves based on theirwhole-genome sequences. The study indicated that the dog was a divergent subspecies of the grey wolf and was derived from a now-extinctghost population of Late Pleistocene wolves,^[73][12][13] and the dog and the dingo are not separate species.^[13] The genome-widephylogenetic tree indicated agenetic divergence between New World and Old World wolves, which was then followed by a divergence between the dog and Old World wolves 27,000YBP^[14][13] – 29,000 YBP.^[13] The dog forms a sister taxon with Eurasian grey wolves but not North American wolves. The dog had considerable pre-ancestry after its divergence from the Old World wolves before it separated into distinct lineages that are nearly as distinct from one another as they are from wolves.^[13] The study suggested that previous datings based on the divergence between wolves and coyotes of one million years ago using fossils of what appeared to be coyote-like specimens may not reflect the ancestry of the modern forms.^{[98][12][14][13]}

The study indicated that theMexican wolf was also a divergent form of grey wolf, suggesting that may have been part of an early invasion into North America.^[13][150] The Tibetan wolf was found to be the most highly divergent of the Old World wolves, had suffered a historical population bottleneck and had only recently recolonized the Tibetan Plateau. Glaciation may have caused its habitat loss, genetic isolation then local adaption.^[13]

The study indicated that there has been extensivegenetic admixture between domestic dogs and wolves, with up to 25% of the genome of Old World wolves showing signs of dog ancestry, possibly as the result ofgene flow from dogs into wolves that were ancestral to all modern wolves. There was evidence of significant gene flow between the European wolves plus the Israeli wolf with thebasenji andboxer, which suggests admixture between the lineages ancestral to these breeds and wolf populations.^[12][13] For the lowland Asian wolves: the Central Russian and East Russian wolves and all of the lowland Chinese wolves had significant gene flow with the Chinese indigenous dogs, theTibetan Mastiff and thedingo. For the highland Asian wolves: The Tibetan wolves did not show significant admixture with dogs; however, theQinghai wolves had gene flow with the dingo and one of them had gene flow with the Chinese dogs. The New World wolves did not show any gene flow with the boxer, dingo or Chinese indigenous dogs but there was indication of gene flow between the Mexican wolf and the African basenji.^[13] All species within the genusCanis, thewolf-like canids, are phylogenetically closely related with 78chromosomes and can potentiallyinterbreed.^[100] There was indication of gene flow into thegolden jackal from the population ancestral to all wolves and dogs (11.3%–13.6%) and much lower rates (up to 2.8%) from extant wolf populations.^[12][13]

The data indicated that all wolves shared similar population trajectories, followed by population decline that coincided with the expansion of modern humans worldwide and their technology for capturing large game.^[13][155] Late Pleistocene carnivores would have been social living in large prides, clans and packs in order to hunt the larger game available at that time, and these larger groups would have been more conspicuous targets for human persecutors.^[155] Large dogs accompanying the humans may have accelerated the rate of decline of carnivores that competed for game,^[13][156] therefore humans expanded across Eurasia, encountered wolves, domesticated some and possibly caused the decline of others.^[13]

The study concluded that admixture had confounded the ability to make inferences about the place of dog domestication. Past studies based on SNPs, genome-wide similarities with Chinese wolves, and lowerlinkage disequilibrium might reflect regional admixture between dogs with wolves and gene flow between dog populations, with divergent dog breeds possibly maintaining more wolf ancestry in their genome. The study proposed that analysis of ancient DNA might be a better approach.^[13]

In the same year, a study found that there were only 11 fixed genes that showed variation between wolves and dogs. These genes are thought to affect tameness and emotional processing ability.^[157] Another study provided a listing of all of the grey wolf and dog mDNA haplotypes combined in the one phylogenetic tree.^[158]

In 2018, a study compared the sequences of 61,000Single-nucleotide polymorphisms (mutations) taken from across the genome of grey wolves. The study indicated that there exists individual wolves of dog/wolf ancestry in most of the wolf populations of Eurasia but less so in North America. The hybridization has been occurring across different time scales and was not a recent event. Low-level hybridization did not reduce the wolf distinctiveness.^[159]

The dingo (Canis familiaris dingo) refers to the dog found inAustralia. The dingo is a divergent subspecies of the grey wolf and is not a separate species,^[13] and is considered genetically to be abasal member of the domestic dog clade.^[13][98][12] The genetic evidence indicates that the dingo originated from East Asian domestic dogs and was introduced through the South-East Asian archipelago into Australia,^[160][161] with a common ancestry between the Australian dingo and theNew Guinea Singing Dog.^[161][162]

In 2015, a study looked at themitochondrial control region sequences of 13 ancient canid remains and one modern wolf from five sites across Arctic north-east Siberia. The fourteen canids revealed nine mitochondrialhaplotypes, three of which were on record and the others not reported before. The phylogentic tree generated from the sequences showed that four of the Siberian canids dated 28,000 YBP and oneCanis c.f. variabilis dated 360,000 YBP were highly divergent. The haplotype designated as S805 (28,000 YBP) from theYana River was one mutation away from another haplotype S902 (8,000 YBP) that represents Clade A of the modern wolf and domestic dog lineages. Closely related to this haplotype was one that was found in the recently extinctJapanese wolf. Several ancient haplotypes were oriented around S805, includingCanis c.f. variabilis (360,000 YBP), Belgium (36,000 YBP – the "Goyet dog"), Belgium (30,000 YBP), and Konsteki, Russia (22,000 YBP). Given the position of the S805 haplotype on the phylogenetic tree, it may potentially represent a direct link from theprogenitor (includingCanis c.f. variabilis) to the domestic dog and modern wolf lineages. The grey wolf is thought to be ancestral to the domestic dog, however its relationship toC. variabilis, and the genetic contribution ofC. variabilis to the dog, is the subject of debate.^[163]

TheZhokhov Island (8,700 YBP) and Aachim (1,700 YBP) canid haplotypes fell within the domestic dog clade, cluster with S805, and also share their haplotypes with – or are one mutation away from – theTibetan wolf (C. l. filchneri) and the recently extinct Japanese wolf (C. l. hodophilax). This may indicate that these canids retained the genetic signature of admixture with regional wolf populations. Another haplotype designated as S504 (47,000 YBP) fromDuvanny Yar appeared on the phylogenetic tree as not being connected to wolves (both ancient and modern) yet ancestral to dogs, and may represent a genetic source for regional dogs.^[163]

The authors concluded that the structure of the modern doggene pool was contributed to from ancient Siberian wolves and possibly fromCanis c.f. variabilis.^[163][164]

In 2015, a study looked at the paleoecology of large carnivores across theMammoth steppe during theLate Pleistocene by using stable isotope analysis of their fossilcollagen to reconstruct their diets. Based on testing in Belgium, around 40,000 YBP theCave hyenas preyed on mammoth, woolly rhinoceros, horses and reindeer, withcave lions taking reindeer and youngcave bears. Wolves appear to have been out-competed bycave hyenas and had their diet restricted to chamois, giant deer and red deer. However, after the Last Glacial Maximum around 14,000 YBP, wolves had access to all prey species, the cave lion was restricted to reindeer, and the cave hyena had gone extinct.^{[165][166][167]} The data suggests that the extinction of the cave hyena allowed the wolf to become the dominant predator rather than the cave lion, just before the cave lion's extinction.^[167] Another study indicated that the wolf thrived compared to the cave hyena when there was greater snow cover.^[168]

The grey wolfCanis lupus is a highly adaptable species that is able to exist in a range of environments and which possesses a wide distribution across theHolarctic. Studies of modern grey wolves have identified distinct sub-populations that live in close proximity to each other.^[4][5] This variation in sub-populations is closely linked to differences in habitat – precipitation, temperature, vegetation, and prey specialization – which affect cranio-dental plasticity.^[6][7][8][9] The archaeological and paleontological records show their continuous presence for at least the last 300,000 years.^[58] This continuous wolf presence contrasts withgenomic studies, which suggest that all modern wolves and dogs descend from a common ancestral wolf population^[12][14][13] that existed as recently as 20,000 years ago.^[12] These studies indicate that apopulation bottleneck was followed by a rapid radiation from an ancestral population at a time during, or just after, theLast Glacial Maximum. This implies that the original wolf populations were out-competed by a new type of wolf which replaced them.^[129] However, the geographic origin of this radiation is not known.

Apart from domestication, humans have harmed the wolf by restricting its habitat through persecution. This has caused a dramatic decrease in its population size over the last two centuries.^[169][170] The shrinking of its habitats that overlap with those of close-relatives such as dogs and coyotes have led to numerous occurrences of hybridization.^[171][172] These events, in addition to recent turnovers (extinctions and repopulations by other geneotypes), has made the unravelling of thephylogeographic history of the wolf difficult.^[120]

Anecotype is a variant in which thephenotypic differences are too few or too subtle to warrant being classified as a subspecies. These can occur in the same geographic region where distinct habitats such as meadow, forest, swamp, and sand dunes provide ecological niches. Where similar ecological conditions occur in widely separated places it is possible for a similar ecotype to occur. This is different from a subspecies, which may exist across a number of different habitats. In animals, ecotypes can be regarded as micro-subspecies that owe their differing characteristics to the effects of a very local environment.^[173] Ecotypes have notaxonomic rank.

Grey wolves have a wide, natural distribution across theHolarctic that includes many different habitats, which can vary from the high arctic to dense forests, open steppe and deserts. The genetic differences between different populations of grey wolves is tightly linked to the type of habitat in which they live.^[9] Differences in genetic markers among the Scandinavian wolf population has arisen in only just over a decade due to their small population size,^[9][174] which indicates that these differences are not dependent on a long time spent in isolation and that larger population patterns can evolve in just a few thousand years.^[9] These differences can also include fur color and density, and body size.^[9][175][4] The differences can also include behavior, as coastal wolves eat fish^[9][175] and tundra wolves migrate.^[9][4] These differences have been observed between two wolf populations that are living in close proximity. It has been shown that mountain wolves do not interbreed with nearby coastal wolves, and the Alps of France and Switzerland have been repopulated with wolves from the mountains of nearby Italy^[9][176] and from the far away mountains of Croatia^[9][177] rather than from the nearer lowlands, which indicates that distance is not the driving force in differences between the twoecomorphs.^[9]

In 2013, a genetic study found that the wolf population in Europe was divided along a north–south axis and formed five major clusters. Three clusters were identified occupying southern and central Europe in Italy, the Carpathians, and the Dinaric-Balkans. Another two clusters were identified occupying north-central Europe and the Ukrainian steppe. The Italian wolf consisted of an isolated population with low genetic diversity. Wolves from Croatia, Bulgaria, and Greece formed the Dinaric-Balkans cluster. Wolves from Finland, Latvia, Belarus, Poland and Russia formed the north-central Europe cluster with wolves from the Carpathians cluster a mixture of wolves from the north-central cluster and the Dinaric-Balkans cluster. The wolves from the Carpathians were more similar to the wolves from the Ukrainian Steppe than they were to wolves from north-central Europe. These clusters may have been the result of expansion from glacial refugia, an adaptation to local environments, and landscape fragmentation and the killing of wolves in some areas by humans.^[178]

In 2016, two studies compared the sequences of 42,000single nucleotide polymorphisms in North American grey wolves and found that they formed six ecotypes. These six wolf ecotypes were named West Forest, Boreal Forest, Arctic, High Arctic, Baffin, and British Columbia. The studies found that precipitation and mean diurnal temperature range were the most influential variables on sequence variation.^[5][179] These findings were in accord with previous findings that precipitation influenced morphology,^[180] and that vegetation^[6] and habitat type^[175][181] influenced wolf differences. One of these studies found that the variation in 11 key genes affected wolf vision, sense of smell, hearing, coat color, metabolism, and immunity. The study identified 1,040 genes that are potentially under selection due to habitat variation, and therefore that there was evidence of local adaption of the wolf ecotypes at a molecular level. Most notable was the positive selection of genes that influence vision, coat color, metabolism and immunity in the Arctic and High Arctic ecotypes, and that the British Columbia ecotype also has a unique set of adaptions.^[179] The local adaptation of a wolf ecotype most likely reflects a wolf's preference to remain in the type of habitat that it was born into.^[5]

Ecological factors including habitat type, climate, prey specialization and predatory competition will greatly influence grey wolfgenetic population structure and cranio-dentalplasticity.^{[182][87][9][4][183][184][6][7][8]} During the Last Glacial Maximum, there was greater wolfgenetic diversity than there is today,^[12][73] and within the Pleistocene grey wolf population the variations between local environments would have encouraged a range of wolf ecotypes that were genetically, morphologically and ecologically distinct from one another.^[87]

During theLast Glacial Maximum 20,000 YBP, thePleistocene steppe stretched across northern and central Eurasia and throughBeringia into North America. The Pleistocenewolves of Beringia, and perhaps those across the steppe, were adapted to this habitat. Their tooth and skull morphology indicates that they specialized in preying on now-extinctPleistocene megafauna, and their tooth wear indicates that their behavior was different from modern wolves.^{[125][9][185][186]} This highlights the success ofC. lupus as a species in adapting to different environmental conditions.^[8] This grey wolf ecomorph became extinct at the end of the glaciation, along with the horse and other species on which it depended, and was replaced by wolves from southern North America. This indicates that specialized wolf ecomorphs can become extinct when their environment changes even though the habitat may still support other wolves.^[9] Wolves went through apopulation bottleneck 20,000 YBP that coincides with the Last Glacial Maximum,^{[12][187][87][9]} which indicates that many wolf populations may have gone extinct at the same time as the Beringian wolves.^[9]

There are a small number ofCanis remains that have been found at Goyet Cave, Belgium (36,500 YBP)^[71] Razboinichya Cave, Russia (33,500 YBP)^[188] Kostenki 8, Russia (33,500–26,500 YBP)^[189] Predmosti, Czech Republic (31,000 YBP)^[190] and Eliseevichi 1, Russia (17,000 YBP).^[72] Based on cranial morphometric study of the characteristics thought to be associated with the domestication process, these have been proposed as early Paleolithic dogs.^[189] These characteristics of shortened rostrum, tooth crowding, and absence or rotation of premolars have been documented in both ancient and modern wolves.^{[125][87][8][191][192][193]} Rather than representing early dogs, these specimens may represent "a morphologically distinct local, now extinct, population of wolves".^[87][194]