How did Man’s Best Friend come about?  With 400 million dogs in the world, this question has been asked often.  The more time I’ve spent exploring the question, the more odd answers I’ve found.

Humans and their dogs have been best friends for a long time, it seems.  But just how long is a matter of dispute.  Conventional wisdom says that dogs emerged as humans were settling down in villages and beginning agriculture, about 10,000-12,000 years ago.

But analysis of mitochondrial DNA (mtDNA) tells a different story. MtDNA in the cells of living things reproduces asexually, and comes only from the mother.  The only way it ever changes is through mutation, and that rate of mutation is slow: about 1-2% every 100,000 years. Studies which compare the mtDNA of modern wolves, dogs, jackals, and other canids around the world shows that dogs are essentially gray wolves, and the relationship of the two is far closer than between either and any other canid.  Dogs and wolves differ from wolves by at most 0.2% of the mtDNA sequence, as compared to the 4% difference between the gray wolf and the coyote1.

The Gray Wolf was once the world’s most widely distributed mammal2, native to everywhere but Australia, Central and South America, and Africa.  Because wolves are not found in Africa, the human-wolf relationship could not have begun until humans began to migrate out of Africa, about 60,000 years ago3.

As humans left Africa, they first occupied the Middle East. Dogs seem to share more genetic similarity with Middle Eastern gray wolves than with any other wolf population worldwide4.  The Middle East is the where domestic cats and most of our livestock originated, as well. A reasonable conclusion is that dogs originated in the Middle East, derived from Middle Eastern gray wolves, at about the time that Homo sapiens began arriving 60,000-80,000 years ago.


As of 2013, the oldest fossil specimens genetically linked to the modern dog’s lineage date to approximately 33,000-36,000 years ago5.  Why have earlier fossils not been found? It may be that earlier canids continued to resemble wolves, and archeologists have misclassified them.  Or it may be that earlier canids were significantly less tame, and that it was only around 33,000-36,000 years ago that these early canids began to be man’s best friend.


In Africa, where human life originated, large animals coevolved with humans, and so as one species became more skillful as a predator, others became more skilled at avoiding becoming prey.  Thus large mammals in Africa were reasonably safe from humans.  But as humans spread across the globe, they encountered species that had not had the luxury of this coevolution.

Animals all seem to know that larger animals pose greater threats to their survival, and smaller ones are generally safer.  As humans spread out of Africa and into new areas, they encountered large herbivores that probably regarded them with interest, but not fear.  These herbivores were probably easier prey for early humans than smaller animals such as squirrels, which would have feared them because of their size difference.

It seems likely that humans killed the largest herbivores first: Columbian Mammoth, Gomphotheres, Mastodon, Woolly Mammoth, Giant Armadillo, Glyptodont, Tapir, Harlan’s Ground Sloth, Giant Beaver, Giant Capybara, and Woolly Rhino.  When they were gone, humans might have shifted to  large ungulates:  moose, caribou, deer, elk, and wild boar.

Humans  likely avoided the predators (such as the American Lion, Saber-Toothed Cats, Cave Lions, Scimitar Cat).  With the large prey gone, the large predators would have struggled to put food on the table.  The extinction of the large predators could have followed the collapse of lower levels of their food chain.

Removing the herbivores would also have impacted the vegetation, and changes in vegetation could produce changes in the frequency of fire…  Removing a keystone herbivore, such as the mammoth or mastodon, could have triggered ecological change that affected the survival of other herbivores6.  So while the collapse of megafauna worldwide follows the arrival of man, overhunting is not necessarily the direct cause of each extinction.

Large mammals reproduce slowly, and normally live a long time.  Such low reproductive rates make them more vulnerable to a change in circumstances, such as the addition of a new predator.  As megafauna killers, it has been hypothesized that humans generally lived in an area just a year or two, and advanced about 10 miles a year to new hunting grounds7.

We can’t prove that these extinctions were caused by man.  Homo did not leave many traces, and might have begun hunting an area thousands of years before they left evidence of their presence there. So while there is evidence that humans dined on megafauna, we don’t have many such smoking guns.  But modern dating techniques remove climate change as a possible explanation of most of these extinctions.

“New evidence based on accurate optically stimulated luminescence and uranium-thorium dating of megafaunal remains suggests that humans were the ultimate cause of the extinction of megafauna in Australia.8 The dates derived show that all forms of megafauna on the Australian mainland became extinct in the same rapid timeframe — approximately 46,000 years ago9 — the period of time in which humans first arrived in Australia. Analysis of oxygen and carbon isotopes from teeth of megafauna indicate the regional climates at the time of extinction were similar to arid regional climates of today, and that the megafauna were well adapted to arid climates.10.”11

Wikipedia summarizes the overkill explanation:

Megafaunal extinctions followed a distinctive landmass-by-landmass pattern that closely parallels the spread of humans into previously uninhabited regions of the world, and which shows no correlation with climatic history… Australia was struck first around 45,000 years ago, followed by Tasmania about 41,000 years ago (after formation of a land bridge to Australia about 43,000 years ago), Japan apparently about 30,000 years ago, North America 13,000 years ago, South America about 500 years later, Cyprus 10,000 years ago, the Antilles 6000 years ago,New Caledonia and nearby islands 3000 years ago,Madagascar 2000 years ago, New Zealand 700 years ago, the Mascarenes 400 years ago, and the Commander Islands 250 years ago.12.

Camp Followers

During this period of megafauna hunting, and before the extinctions, wolves could have learned that man was a hungry wolf’s best friend.  Today, the majority of a wolf’s diet consists of large ungulates as well as smaller birds and mammals, carrion, and garbage13. Wolves could harrass a mammoth, but could not bring it down. Humans could kill it with spears or by driving it off a cliff.

And Homo likely learned that wolves could find prey, and run them to exhaustion.

Five men could kill a Woolly Mammoth or Giant Sloth, but five could not eat it all.  Whether they helped themselves to what they could carry, or simply ate it on the spot, there would have been considerable remains.  For their next meal, these men could simply kill another large mammal.

Wolves will eat carrion, and hunters could have provided all or some of their diets.  Tracking hunters or finding their kills would have been straightforward for a wolf, whose sense of smell is far more sensitive that that of humans.

Eventually, some wolves might have taken to walking with the hunters, fanned out in the area searching for prey.  The kill might have been done by a combination of humans and canids, and the quarry then shared.

Discovered Mutualism: The Utility of a Partnership

Early canids would have found humans useful.  They could kill megafauna. They could provide a family structure that the wolf was familiar with: mother, father, siblings.

But humans would find canids useful, and value them for their utility.

  • Canids could carry a load, like a small pack horse.  In a 7,000 year old grave in Siberia, a dog was found buried alongside humans.  Its skeleton suggested that it was repeatedly used to transport loads14.
  • Canids could help with the hunt. The Siberian dog suffered numerous broken bones during its lifetime, possibly from the feet of red deer during hunting outings15.
  • Canids might bark at approaching danger. The bark must have been valued by bands of Homo when they were attacked in the night by other bands.  If a dog barked a warning, both it and its family of humans might have had a better chance to live and reproduce.
  • Canids could eat what humans ate. A DNA and stable isotope analysis of the Siberian dog determined that he ate exactly what humans at the site consumed: fish, freshwater seal meat, deer, small mammals, and some plant foods.16.

In the Czech Republic, a canine skull has been found with a mammoth bone inserted in its mouth, and the dog’s brains removed, as in a ritual burial. That skull, and two others, was shorter than a wolf’s, with a short snout, wide palate, and wide braincase17.  From the skulls, the canids were dated to the Gravettian toolmaking culture, which ranged from 32,000 years ago and 22,000 years ago18.  Modern dogs have shorter snouts, wider palates, and wider braincases than wolves, and they are respected and loved.  This appears to have been true 22,000 BP to 32,000 BP as well.

Early Domestication: Overcoming Fearfulness

Dogs of the period 60,000 BP (Before Present) to 30,000 BP might have resembled wolves, and differed primarily in their willingness to approach humans more closely than their wild cousins.  Those most willing to approach humans would have been the most likely to be able to dine after the hunt, and the most likely to pass on their genes. Between the period when Homo reached a landmass and when megafauna became extinct, it is likely that some wolves became camp followers.

Domestication was likely a complex affair.  Both wolves and humans could have found each other generally beneficial, but there were risks. If humans are dangerous to large mammals, they could be dangerous to wolves, too. So some wolves must have learned to read tone of voice, body language, and other properties to discern whether they should approach or avoid.  A number of studies conclude that dogs can do what dog owners have long known:  they can read our emotions.

One reason dogs are so good at reading facial expression may be that in humans, the right side of the face has been found to more faithfully present our emotions.  Somewhere along the line, Canids learned this, and look toward their left — the right side of a human face — when looking at a human face, though they show no left-right preference when looking at the faces of dogs or other animals. Dogs can recognize smiles19 and the other key emotions of anger and fear.  And from a bark alone, humans can usually distinguish six emotions, including anger, fear, happiness, and despair.

The friendliest wolves would have had the greatest success with humans, getting to the kill sooner.  They would have been quicker to learn human ways, and human emotions.  Their emerging skill at reading human emotions would have been useful to their survival.  Some early part of the emergence of dogs from wolves must have been self-domestication.

Survival of the Friendliest

For thousands of years, humans preferentially favored the friendliest wolves.  Humans may have shown their preferences both by dining on the fearsome, and caring for the friendly. Through this artificial selection, surviving wolves became friendlier, though there were genetic setbacks, as when a partially domesticated canid would fall in love with a wolf, or when the humans would become too hungry for kindness.  But by selectively favoring the friendliest, humans increased the chance that two friendly canids would mate, and pass on whatever genetic predisposition to friendliness they might have.

Favoring the friendliest was easy: just choose those who appear juvenile and cute to humans.  Domesticated animals of all sorts — dogs, cows, pigs, and rats — are likely to have floppier ears and curlier tails than their wild counterparts. Domesticated animals sometimes have spotted coats.  These various qualities are sometimes called the domestication phenotype.

Favoring the friendliest may have led to neoteny; the retention of juvenile characteristics into adulthood.  Juvenile wolves are friendly and curious. They bark and play.  In wolves, these characteristics melt away after childhood.  In dogs, they do not. In humans, there are neotenous traits as well: flattened face, broadened face, large brain, hairless body, hairless face, small nose, reduction of brow ridge, small teeth, small upper jaw (maxilla), small lower jaw (mandible), thinness of skull bones, limbs proportionately short compared to torso length, longer leg than arm length, larger eyes, and upright stance. Physical neotenization in humans may have caused psychologically neotenous traits in humans: curiosity, playfulness, affection, sociality and an innate desire to cooperate20.

Dmitry Belyaev and Lyudmila Trut have conducted a 40 year experiment in the domestication of silver foxes.  They began with 130 foxes from Siberian fur farms, bred them, tested the offspring for approachability and friendliness, and bred those with the highest scores.  The results of this experiment are foxes that closely resemble man’s best friend.  The foxes seek human company, wag their tails, roll on their backs, and pant eagerly in anticipation of attention.  And their appearance changed, too:  by selecting for approachability and friendliness, the researchers created a fox that looked like the domesticated dog, with floppy ears and many color variations (see photo.)  The dramatic changes in the foxes were rapid: by the fourth generation, the tamest would jump into the researchers arms and lick their faces21.  And the friendly foxes prove to be smart, and appear to be as good dogs at reading human faces.

Foxes bred through generations to be as human-friendly as dogs get a boost from Lyudmila Trut (center) and other staff at the Institute of Cytology and Genetics, in Novosibirsk, Siberia. Photograph by Vincent J. Musi. Reprinted without permission from National Geographic.

Early man did not have the luxury of a farm and a researcher to choose and breed the friendliest.  But they did have time on their side.  Ten thousand years of fumbling can surely achieve what 40 years of science can do.

Shifting to Starch

Around 10,000 years ago, humans shifted from hunting and gathering to farming.  The shift was gradual, and came about as local human populations increased, and local wildlife populations decreased.  Humans needed food, and hunting wasn’t meeting their needs. Primitive agriculture began, with humans assisting their favorite wild plants in growth and reproduction.  Human diet shifted to starches, and continued to discard table scraps, now with more starch and less meat. This shift in diet was hard on both humans and their canid friends who now counted on table scraps. Natural selection favored those humans and canids who were able to digest this new diet. This is an example of parallel evolution, in which one selection pressure had similar effects on two species.

Under the hood, there were genetic changes to allow digestion of starch by dogs22. Dogs have:

  • extra copies of the gene for amylase — an enzyme that converts starch to maltose — and can produce 28 times more of that protein than wolves.
  • mutations in the gene that allows for the conversion of this maltose to sugar, resulting in 12 times more of that enzyme than wolves.
  • mutation in a third gene that improves the function of a protein that absorbs sugar through the gut.

These same genes that improve starch digestion in dogs also changed in humans, so that we could digest starch.

The increased consumption of starch might have triggered an increase in estrogen and testosterone levels. As these levels go up, the age of puberty drops.  Earlier maturation — neoteny — might be the result of changes in diet23  As civilization emerged, humans may have selected each other for reduced fear and aggression.  In the wild, there is a place for fear and aggression, but in crowded living, domesticating the wild in us makes good sense.


For the thousands of years that humans and canids have lived together, each playing a role in the artificial selection of the other.  One force of humans was to bring out the bark in canids.  Adult wolves don’t bark, but contain the same vocal apparatus as dogs and wolf puppies, so could bark as their pups do.  Studies have shown that the acoustic structure of alarm barks is the same across all dogs, whereas the sounds they make at play varies widely24.  Homo would not have cared what they sounded like at play, but would have valued an easily recognized alarm bark. Research now indicates that humans can classify dog barks recorded in different situations25, and that different barks carry different emotional information for humans26.

Neoteny could help explain barking: adult canids who were neotenic not only retained their playfulness and friendliness, but retained the bark of wolf pups.

Barking at approaching strangers would have benefited both Homo and canid. Dogs that barked to warn of approaching trouble were richly rewarded by the survivors; those that did not were sometimes eaten by the attackers.  Barking at strangers further distinguished dogs from wolves.

Multiple Origins?

With so many wolves, and so much time to work with, we might expect that dogs originated at different times for these different places — continental Asia, Australia, Tasmania, Japan, North America, and South America in succession.  But genetic evidence is now pointing toward a single line, beginning in the Middle East, and tracking along with the spread of humans around the globe.  Those early dogs who reached North and South America seem to have been displaced by the dogs of the Europeans who colonized the Americas, but those European dogs, too, began their careers in the Middle East.  It might be that the jump start those early dogs had in starch digestion, friendliness, barking and brain size always gave them a competitive advantage over any dogs to have more recently emerged from the wild wolf.


Show 26 footnotes

  1. Robert K. Wayne. Molecular Evolution of the Dog Family. published online.
  2.  Canis
  3. Stringer, C. (2012). “AOP”. Nature 485 (7396): 33–35. doi:10.1038/485033a. PMID 22552077.
  4. Stuart Wolpert Dogs likely originated in the Middle East, new genetic data indicate. Findings based on analysis of largest set of genetic markers ever studied. UCLA Newsroom. March 17, 2010.
  5. Druzhkova AS, Thalmann O, Trifonov VA, Leonard JA, Vorobieva NV, et al. (2013) Ancient DNA Analysis Affirms the Canid from Altai as a Primitive Dog. PLoS ONE 8(3): e57754. doi:10.1371/journal.pone.0057754; Mietje Germonpréa; Mikhail V. Sablinb; Rhiannon E. Stevensc; Robert E.M. Hedgesd; Michael Hofreitere; Mathias Stillere; Viviane R. Desprése (February 2009). “Fossil dogs and wolves from Palaeolithic sites in Belgium, the Ukraine and Russia: osteometry, ancient DNA and stable isotopes”. Journal of Archaeological Science 36 (2): 473–490. doi:10.1016/j.jas.2008.09.033.
  6. The Late Pleistocene Extinctions. Illinois State Museum.
  7. Stevens, William K. 1997. New Suspect in Ancient Extinctions of the Pleistocene Megafauna: Disease. New York Times,   4-29-97.
  8.  Prideaux, G. J.; Long, J. A.; Ayliffe, L. K.; Hellstrom, J. C.; Pillans, B.; Boles, W. E.; Hutchinson, M. N.; Roberts, R. G.; Cupper, M. L.; Arnold, L. J.; Devine, P. D.; Warburton, N. M. (2007-01-25). “An arid-adapted middle Pleistocene vertebrate fauna from south-central Australia”. Nature 445 (7126): 422–425. doi:10.1038/nature 05471. PMID 17251978.
  9. Roberts, R. G.; Flannery, T. F.; Ayliffe, L. K.; Yoshida, H.; Olley, J. M.; Prideaux, G. J.; Laslett, G. M.; Baynes, A.; Smith, M. A.; Jones, R.; Smith, B. L. (2001-06-08). “New Ages for the Last Australian Megafauna: Continent-Wide Extinction About 46,000 Years Ago”. Science 292 (5523): 1888–1892. doi:10.1126/science.1060264. PMID 11397939. 
  10.  Prideaux, G. J.; Long, J. A.; Ayliffe, L. K.; Hellstrom, J. C.; Pillans, B.; Boles, W. E.; Hutchinson, M. N.; Roberts, R. G.; Cupper, M. L.; Arnold, L. J.; Devine, P. D.; Warburton, N. M. (2007-01-25). “An arid-adapted middle Pleistocene vertebrate fauna from south-central Australia”. Nature 445 (7126): 422–425. doi:10.1038/nature 05471. PMID 17251978 .
  11. Australian Megafauna. Wikipedia.
  12. Megafauna. Wikipedia.
  13. Canis
  14. Jennifer Viegas, 2011. Prehistoric Dog Lived, Died Among Humans. Discovery News.
  15. Jennifer Viegas, op.cit
  16. Jennifer Viegas, op.cit
  17. Mietje Germonpréa, Martina Lázničková-Galetováb, Mikhail V. Sablinc. Palaeolithic dog skulls at the Gravettian Předmostí site, the Czech Republic. Journal of Archaeological Science. Volume 39, Issue 1, January 2012, Pages 184–202.
  18.  GravettianWikipedia.
  19. Dogs can discriminate human smiling faces from blank expressions. Nagasawa, Miho; Murai, Kensuke; Mogi, Kazutaka; Kikusui, Takefumi. Animal Cognition vol. 14 issue 4 July 2011. p. 525 – 533
  20. Neoteny. Wikipedia.
  21. Evan Ratliff. Animal Domestication. National Geographic.
  22. Axelsson E, Ratnakumar A, Arendt ML, Maqbool K, Webster MT, Perloski M, Liberg O, Arnemo JM, Hedhammar A, Lindblad-Toh K. “The genomic signature of dog domestication reveals adaptation to a starch-rich diet,” Nature, 2013: doi:10.1038/nature11837, 2013.
  23. Andrew Lehman. What is Neoteny?
  24. Brandon Keim. Humans Guided Evolution of Dog Barks. Wired.
  25. Pongrácz, P., Miklósi, Á., Molnár, Cs., Csányi, V. 2005. Human listeners are able to classify dog barks recorded in different situations. Journal of Comparative Psychology, 119: 136-144.
  26. Pongrácz, P., Molnár, Cs., Miklósi, Á. 2006. Acoustic parameters of dog barks carry emotional information for humans. Applied Animal Behaviour Science, 100: 228-240.