Top banner


Do the Eyes Have It?

Dog domestication may have helped humans thrive while Neandertals declined

Pat Shipman

We all know the adage that dogs are man’s best friend. And we’ve all heard heartwarming stories about dogs who save their owners—waking them during a fire or summoning help after an accident. Anyone who has ever loved a dog knows the amazing, almost inexpressible warmth of a dog’s companionship and devotion. But it just might be that dogs have done much, much more than that for humankind. They may have saved not only individuals but also our whole species, by “domesticating” us while we domesticated them.

One of the classic conundrums in paleoanthropology is why Neandertals went extinct while modern humans survived in the same habitat at the same time. (The phrase “modern humans,” in this context, refers to humans who were anatomically—if not behaviorally—indistinguishable from ourselves.) The two species overlapped in Europe and the Middle East between 45,000 and 35,000 years ago; at the end of that period, Neandertals were in steep decline and modern humans were thriving. What happened?

A stunning study that illuminates this decisive period was recently published in Science by Paul Mellars and Jennifer French of Cambridge University. They argue, based on a meta-analysis of 164 archaeological sites that date to the period when modern humans and Neandertals overlapped in the Dordogne region of southwest France, that the modern-human population grew so rapidly that it overwhelmed Neandertals with its sheer numbers.

Because not all the archaeological sites in the study contained clearly identifiable remains of modern humans or Neandertals, Mellars and French made a common assumption: that sites containing stone tools of the Mousterian tradition had been created by Neandertals, and those containing more sophisticated and generally later stone tools of the Upper Paleolithic were made by modern humans. This link between tool and toolmaker is well supported by sites that do contain hominin remains, but there is nothing inherent in a stone tool that tells you who made it—not even if you find a skeleton right next to it. Still, stone tools are one of the best available indicators of which species—modern human or Neandertal—inhabited a particular location.

Mellars and French compared the number and sizes of Neandertal and modern-human archaeological sites, as well as the density of tools and the weight per square meter of prey animals, represented by fossils, in those sites. They standardized their results for 1,000-year periods to compensate for the varying amounts of time that the different locations had been occupied. In every respect, modern humans surpassed Neandertals. In fact, the greater success of modern humans was so clear that, according to Mellars and French’s calculations, the human population increased tenfold over the 10,000-year overlap period. Modern humans thrived and Neandertals did not—even though Neandertals had lived in the European habitat for about 250,000 years before modern humans “invaded.” Why weren’t Neandertals better adapted to their environment than the newcomers?

There is no shortage of hypotheses. Some favor climate change, others a modern-human advantage derived from the use of more advanced hunting weapons or greater social cohesion. Now, several important and disparate studies are coming together to suggest another answer, or at least another good hypothesis: The dominance of modern humans could have been in part a consequence of domesticating dogs—possibly combined with a small, but key, change in human anatomy that made people better able to communicate with dogs.

Skulls and Souls

Until 2009, dogs were believed to have been domesticated about 17,000 years ago, long after Neandertals were already extinct. But then Mietje Germonpré of the Royal Belgian Institute of Natural Sciences and her colleagues developed a rigorous statistical method for identifying which fossil canid skulls belonged to wolves and which to dogs—the latter distinguished by their shorter, broader snouts and braincases. As I described in an earlier Marginalia column (“The Woof at the Door,” July–August 2009) the team used this technique to identify three Paleolithic skulls as domestic dogs. The earliest was from the Belgian cave of Goyet and dated to about 32,000 years ago.

Some critics of that study pointed to the long, 15,000-year interlude between those early dog skulls and the next oldest confirmed specimens. But recent discoveries have begun to fill the gap and lend credence to the ancient association between humans and large canids. In early 2012, Germonpré’s team published a study of nine additional canid skulls, three of which represented ancient dogs from Předmostí in the Czech Republic, a site dated to about 27,000 years ago. Another canid skull with many doglike features was recently studied by Nikolai Ovodov, of the Siberian Branch of the Russian Academy of Sciences, and his team. It comes from Razboinichya Cave, Siberia, and is dated to about 29,000 years ago. Although this team did not use the same powerful statistical techniques as Germonpré’s group, they identified the Razboinichya skull as an “incipient dog”—an animal in an early stage of the domestication process.

None of these ancient dog skulls date exactly to the period of modern human–Neandertal overlap, but the domestication process must have been underway even before the first identifiable dog entered the fossil record. The rapidly developing consensus is that dogs were domesticated during the period when both modern humans and Neandertals lived in Europe. So far, all of these early dogs are from modern-human sites. Several lines of evidence suggest that dogs and wolves were especially revered by those humans.

2012-05MargShipmanFA.jpgClick to Enlarge ImageOne of the Předmostí dogs was found with its jaw and cranium still attached to each other in a lifelike position and with a large piece of bone wedged in its mouth. The bone must have been inserted shortly after the dog’s death, while muscles and ligaments still held the jaw to the cranium. The team suggests that in the past, as now, valued hunting dogs were honored and perhaps buried with ritual.

Another indicator of the importance of dogs was that two canine teeth from dogs or wolves at Předmostí were modified to be worn as personal adornments. Rarely did Paleolithic people make jewelry out of parts of food animals, so the high frequency of canid teeth drilled for use as pendants at Předmostí and other Paleolithic sites indicates that they were not considered food. Like humans, canids are very rarely depicted in Paleolithic cave art, also suggesting that the cave artists might have regarded canids as unusually close to humans.

There is something else odd about the early canid skulls: Forty percent of the 20 dog and wolf crania found at Předmostí have been pierced. Citing evidence from northern hunting peoples around the world who ceremonially open the braincases of slain carnivores, Germonpré and her colleagues surmise that the perforation of the Paleolithic dog skulls may have had a ritual significance. “At Předmostí,” the team wrote,

the large number of perforated braincases of large canids and the dog skull holding a bone between its front teeth hint at a specific relationship between humans and large canids, including the possibility of the existence of a wolf/dog ritual that could be connected with the sending of souls.

Earning Their Keep

Would a good dog really have been so important that it would inspire ritual significance—and give modern humans a crucial edge over Neandertals? We can’t observe how ancient, but anatomically modern, humans used dogs in their daily life, but there are some interesting possibilities. We know from their bones that the Paleolithic dogs were very large, with a body mass of at least 32 kilograms and a shoulder height of at least 61 centimeters, about the size of a modern German shepherd. Germonpré and her colleagues suggest that these early dogs might have been beasts of burden. They cite ethnographic examples of peoples like the Blackfeet and Hidatsa of the American West, who bred very large, strong dogs specifically for hauling travois or strapped-on packs.

All but one of the six Paleolithic dog sites that have so far been identified preserve large quantities of mammoth bone which, with meat attached, must have been lugged from the kill site to where the group was living. If the dogs carried the meat, humans would have saved a lot of energy, so each kill would have provided a greater net gain in food—even after feeding the dogs. Additional food generally has marked effects on the health of a group. Better-fed females can have more babies, can provide them with more milk and can have babies at shorter intervals. Before long, using pack dogs could have caused the human population to increase.

Dogs may also have contributed more directly to human hunting success. To discover how big a difference dogs could make, Vesa Ruusila and Mauri Pesonen of the Finnish Game and Fisheries Institute investigated what may be the closest easily studied analog to a mammoth hunt: the Finnish moose hunt. Finns use large dogs such as Norwegian elkhounds or Finnish spitzes to find moose and keep them in place by barking until humans can approach and shoot them. In hunting groups of fewer than 10 people, the average carcass weight per hunter without dogs was 8.4 kilograms per day. With dogs, the yield went up to 13.1 kilograms per hunter per day—an increase of 56 percent.

2012-05MargShipmanFB.jpgClick to Enlarge ImageStudies of modern hunting peoples who use smaller dogs and pursue smaller game also indicate the advantages of working with dogs. Jeremy Koster and Ken Tankersley of the University of Cincinnati studied hunting among the Mayangna and Miskito peoples in Nicaragua. About 85 percent of the mammals taken by the villagers involved the use of dogs. Male dogs—which were larger—brought in more game than females. Male dogs contributed, on average, to the capture of more than 20 kilograms of meat each month, and the most successful of those dogs helped land more than 100 kilograms per month. The most proficient female brought in more than 35 kilograms per month. These dogs weighed, on average, 11 to 12 kilograms, and yet some contributed much more than their body weight in meat each month.

The dogs improved hunting success by increasing the rate at which the hunters encountered game. Finding game “is often the hardest skill to learn for human hunters,” Koster and Tankersley write. The encounter rate for agoutis—rodents weighing 2 to 8 kilograms—was nine times better if dogs were used; for armadillos, the encounter rate was six times better with dogs. The most successful hunts of all involved killing tapirs using both dogs and firearms.

Karen Lupo of Washington State University conducted a similar study among the Bofi and Aka forest hunters of the Central African Republic. These foragers hunt with spears, bows, crossbows and nets; most of their prey animals weigh less than 10 kilograms each. Although dogs had little effect on the proportion of successful hunts, they markedly reduced the time required to make a kill. In hunts that captured giant pouched rats, for instance, dogs reduced the amount of time expended before a kill by 41 percent (29 minutes with dogs versus 49.5 minutes without dogs). Similarly, when porcupine was caught, dogs reduced the time before a kill by 57 percent (44 minutes versus 101.33 minutes). Faster hunts lowered the energy expended by humans during hunting and increased the yield. The dogs were not treated as pets or companions, and the very idea of doing so was considered laughable by Lupo’s informants. Lupo titled her paper “A dog is for hunting.”

Domesticating dogs clearly improves humans’ hunting success and efficiency—whether the game (or the dog) is large or small. The same must have been true in the Paleolithic. If Neandertals did not have domestic dogs and anatomically modern humans did, these hunting companions could have made all the difference in the modern human–Neandertal competition.

I can’t help wondering whether the process of domesticating dogs was connected to changes in human anatomy and communication abilities. Domestication is a two-way street, as we know from examples such as the genetic changes that make adult humans able to digest milk. Those mutations arose several times in different human populations after the domestication of cattle. I have no evidence that the change I am about to discuss did or did not occur between 45,000 and 35,000 years ago. But it might have.

Wide-Eyed Cooperation

2012-05MargShipmanFC.jpgClick to Enlarge ImageA study by Hiromi Kobayashi and Shiro Kohshima of the Tokyo Institute of Technology showed that modern humans are unique among extant primates in having highly visible white sclerae surrounding the colored irises of their eyes, as well as eyelids that expose much of the sclerae. In other primates, the dark sclerae, similarly colored skin and concealing eyelids tend to mask the direction in which the animal is looking, according to the Japanese team. In humans, the white sclerae and open eyelids make the direction of a person’s gaze visible from a distance, particularly if that glance is directed in a more or less horizontal direction. The changes in the human eye may be adaptations to enhance the effectiveness of the gaze signal.

Michael Tomasello and colleagues at the Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany, developed this idea as the “cooperative eye hypothesis.” They suggested that cooperation among humans was facilitated by the ability to recognize where others were looking. Apes will follow gaze less often than human infants, they found. If the direction of gaze and the direction of the head conflicted, apes tended to follow head direction. In a humorous aside, the researchers noted that they tried their experiment with 14 chimpanzees, 4 gorillas, 4 bonobos and 5 orangutans—but dropped the results of tests on three chimps and all five orangutans because they “did not pay attention to the gaze cues sufficiently for their skills to be reliably assessed.” Following gaze was apparently not a high priority to the apes.

2012-05MargShipmanFD.jpgClick to Enlarge ImageThe mutation causing white sclerae is universal in humans, but it turns up occasionally in apes, too. In decades of observations at Gombe National Park in Tanzania, Jane Goodall observed two chimps, probably brothers, who had white sclerae. A third, female chimp developed white sclerae as an adult. But the trait has not spread or reappeared in that population. The advantage of the white sclerae must be related to something that ancient humans did commonly and chimps don’t do or do rarely. Although chimps hunt small prey, often cooperatively, meat makes up less than 2 percent of their diet, whereas Paleolithic humans hunted much larger game that apparently provided a significant part of their diet. Obviously, silent communication among humans would be advantageous for hunting in groups. But there is another skilled gaze-reader: the domestic dog.

A dog will follow the gaze of a videotaped human if the human first attracts the dog’s attention by speaking to it and looking at it, according to results published by Ernõ Téglás, of the Central European University in Budapest, Hungary, and his colleagues. Indeed, dogs perform as well as human infants at following the gaze of a speaker in tests in which the speaker’s head is held still.

Ádám Miklósi of Eötvös Loránd University in Budapest, Hungary, and his team tested dogs and wolves, and found that dogs were far more attentive to human faces than were wolves, even socialized wolves. Although wolves excel at some gaze-following tasks, perhaps suggesting a preadaptation for communicating with humans, dogs tend to look at human faces for cues and wolves do not. Miklósi’s team believes this major behavioral difference is the result of selective breeding during domestication.

Another way of looking at this phenomenon is that the white sclerae became universal among humans because it enabled them to communicate better not only with each other but also with dogs. Once dogs could read a human gaze signal, they would have been even more useful as hunting partners. No genetic study has yet confirmed the prevalence or absence of white sclerae in Paleolithic modern humans or in Neandertals. But if the white sclera mutation occurred more often among the former—perhaps by chance—this feature could have enhanced human-dog communication and promoted domestication. Although some genetic analyses have suggested that modern humans and Neandertals interbred, even the highest estimates of cross-breeding involve very low levels of genetic exchange that might have been inadequate to spread the white sclera trait among Neandertals.

Humans love to look into their dogs’ eyes to “read” their emotions. Dogs apparently feel the same. Maybe—just maybe—this reciprocal communication was instrumental in the survival of our species.


  • Germonpré, M., M. Lázničková-Galetová and M. Sablin. 2012. Palaeolithic dog skulls at the Gravettian Předmostí site, the Czech Republic. Journal of Archaeological Science 39:84–202.
  • Goodall, J. 1986. The Chimpanzees of Gombe: Patterns of Behavior. Cambridge, MA: Belknap Press.
  • Kobayashi, H., and S. Kohshima. 2001. Unique morphology of the human eye and its adaptive meaning: Comparative studies on external morphology of the primate eye. Journal of Human Evolution 40:419–435.
  • Koster, J., and K. Tankersley. 2012. Heterogeneity of hunting ability and nutritional status among domestic dogs in lowland Nicaragua. Proceedings of the National Academy of Sciences of the U.S.A. 109:E463–E470.
  • Lupo, K. 2011. A dog is for hunting. In Ethnozooarchaeology, eds. U. Albarella and A. Trentacoste, pp. 4–12. Oxford: Oxbow Press.
  • Ovodov, N. D., et al. 2011. A 33,000-year-old incipient dog from the Altai Mountains of Siberia: Evidence of the earliest domestication disrupted by the Last Glacial Maximum. PLoS ONE 6(7):e22821.
  • Ruusila, V., and M. Pesonen. 2004. Interspecific cooperation in human (Homo sapiens) hunting: The benefit of a barking dog (Canis familiaris). Annales Zoologici Fennici 41:545–549.
  • Téglás, E., et al. 2012. Dogs’ gaze following is tuned to human communicative signals. Current Biology 22:1–4.
  • Tomasello, M., B. Hare and J. C. Lehmann. 2007. Reliance on head versus eyes in the gaze following of great apes and human infants: The cooperative eye hypothesis. Journal of Human Evolution 52:314–320.



Bottom Banner