The domestic dog is a member of the family Canidae, a diverse group of carnivores which contains 36 extant species. These include wolves, jackals, and foxes, as well as the coyote, the maned wolf, the bush dog, the African wild dog, the Dhole and the racoon dog. Within this family, the dog is most closely related to wolves, jackals and the coyote, as these canines all have the same number of chromosomes and are all capable of interbreeding to produce fertile offspring. This makes all of these species potential dog ancestors, and all have been suggested at one point to have played a part in the dog's ancestry. In fact, because of the extreme variation that exists between different dog breeds, it has been assumed in the past that multiple species of canids have shared in the ancestry of the domestic dog. For example, the Nobel prize winning ethologist Konrad Lorenz once theorized that both wolves and jackals were the ancestors of the modern dog, with some breeds being the descendants of wolves, and others being the descendants of jackals (Lorenz, 1954). However, many morphological and behavioural studies have suggested that dogs are likely the descendants of wolves (see Zimen, 1981).
Recently, the study of DNA sequences has been used to address the question of the dog's ancestry, and studies using this approach have all pointed to the wolf as being the dog's closest wild relative. For example, a study done by Vila et al. (1997) examined mitochondrial DNA (mtDNA) sequences from several canines to investigate the origins of the domestic dog. mtDNA sequences were obtained from 162 wolves representing 27 populations from Europe, the Middle East, North America and Asia, 140 dogs representing 67 breeds, and 5 coyotes, 2 golden jackals, 2 black-backed jackals and 8 Abyssinian jackals. mtDNA is inherited only maternally, unlike nuclear DNA (which is found in the nucleus of the cell), which is inherited from both the mother and the father. The region of mtDNA sequenced (the control region) region was chosen because it evolves very rapidly, and since dogs are such a young species, only a rapidly evolving region of DNA would show any differences between different dog types.
The study found that dog sequences varied from jackal or coyote sequences in at least 20 places and that wolf and dog sequences were very similar - identical in one case. This suggests that the wolf is the ancestor of the dog. A analysis of the dog and wolf sequences showed that dogs may have come from at least two unique common ancestors, which suggests that the domestication of the dog may have occurred at least twice. It is also interesting to note that individuals of the same breed sometimes had very different sequences, so it seems that there was a great deal of hybridization between various dog types before modern pure breeds were established. The research team could not figure out which groups of modern wolves are the ancestors of today's wolves and noted that this ancestral wolf population may now be extinct. Other studies on domestic dog evolution which also used sequences of mitochondrial DNA to address the problem also concluded that the wolf is the ancestor of the dog and that the dog was domesticated more than once (see Tsuda et al., 1997).
Today there is little doubt that the dog did evolve from the wolf, although it may be possible that occasional interbreeding between dogs and jackals or coyotes has occurred throughout time. Dogs and wolves are considered by some biologists to be the same species, with the dog (Canis lupus familiaris) being a domestic variant or subspecies of the wolf (Canis lupus). However, Canis familiaris is still used by most biologists as the scientific name for the dog.
The study by Vila et al. (1997) described above used the sequences generated to estimate the age of the domestic dog by using what is called a "molecular clock." Since the DNA sequences of two evolutionary lineages accumulate change with time, the number of nucleotide differences in the DNA sequences of two different species can be used to estimate the amount of time that has lapsed since the two species diverged. The clock used by Vila et al. (1997) was calibrated using the sequence differences between wolves and coyotes, two species which diverged about one million years ago. The number of differences between the wolf and coyote DNA was used with the approximate age of divergence between the two species to determine the average number of nucleotide changes that would occur in a given amount of time for two different canine species. A comparison of wolf and dog sequences was considered in light of the average number of changes expected to occur and it was estimated that the earliest dogs may had been domesticated about 135 000 years ago. This age is far older than any other previous estimate of the dog's age.
The 135 000 year date has been questioned by many wolf biologists and molecular geneticists (see Federoff and Nowak, 1997). Since mtDNA sequences have high and variable rates of change throughout time, the estimate that dogs may have been domesticated over 100 000 years ago may be somewhat inaccurate. The rate of mtDNA change between dogs and wolves may differ greatly from the rate of mtDNA change between wolves and coyotes because dogs breed twice a year and wolves and coyotes can only breed once. The 135 000 year date also disagrees with a date recovered from a larger data set of dog and wolf mtDNA examined by Savolainen et al. (2002). This study suggested that dogs likely originated about 15 000 years before the present (B.P.).
The date also contradicts archaeological evidence that suggests dogs were domesticated about 14 000 years ago. The oldest discovered fragment of dog bone is from Germany and dates back to about 14 000 years B.P. Slightly older dog remains which date to 12 000 years B.P. have also been found in Israel. The oldest dog remains found in the Americas are from Danger Cave, Utah and date back to 9 000 - 10 000 years B.P. However, early dogs most likely resembled wolves morphologically, so it is possible that very old dog fossils are often classified as wolf fossils. Overall, the early archaeological record of dogs is rather poor, since most Palaeolithic sites contain very few canid remains, and the ones that are found tend to consist of only small fragments of bone which are difficult to accurately identify.
There is no doubt that dogs were associated with humans by the early Neolithic, as a great deal of rock art depicting dogs with humans and clay sculptures of dogs have been found in southwest Asia, Iraq, Turkey and to a lesser extent, Africa, England and Denmark. One example is a Neolithic rock painting from Iraq showing people hunting deer while accompanied by dogs with curly tails. Another Neolithic painting found in Algeria shows an ox-like creature surrounded by dogs with curly tails and a human hunter holding a spear. The Neolithic archaeological record of dogs is much more extensive that the Palaeolithic one. For instance, 53 cranial and mandibular remains of dogs from 6 600 B.C.E. have been found in the Jarno mountains between Iran and Iraq. Clay statues of dogs have been found there as well. By 5400-4600 B.C.E., there were likely some morphological differences between dogs found throughout the world, because dog remains found by the Danube river in Romania, which date back to this time, are far smaller than dog remains found at many other sites of a similar age (Thurston, 1996).
The dog is the only domesticated species that was found throughout the world before the fifteenth century, and the dog likely evolved from the wolf, a species with a holarctic distribution. These facts confound the problem of determining where the dog originated, and whether or not New World dogs are the descendants of Old World dogs or are the result of a separate domestication event. Two recent studies have addressed this problem using molecular data.
Savolainen et al. (2002) sequenced 582 base pairs of mtDNA obtained from 38 Eurasian wolves, and of 654 dogs from Europe, Asia, Africa, and Arctic America. The pattern of genetic variation among the dogs was examined and used to infer the location of the origin of the dog. This approach was used because in a comparison of the genetic diversity of an ancestral population and a population which was derived from a subset of the ancestral population, one would expect the genetic diversity to be greater in the ancestral population than in the derived population. This is because the derived population will contain only a portion of the genetic diversity found in the ancestral one. It was noted that dogs from east Asia possessed significantly more genetic variation than dogs from other parts of the world, which suggests that the domestic dog may have originated there. Morphological evidence also points to an east Asian origin of the domestic dog, because an osteological jaw feature found in domestic dogs is also found in east Asian wolves, but is absent in all other wolves. (Olsen and Olsen, 1977). The study by Savolainen et al. (2002) also noted that all dogs likely originated from at least five female wolf lines, as a statistical analysis of the data placed the dog sequences into five different, distinct clades.
The question as to whether or not there were two domestication events in the Old World and the New World has also been addressed recently through the use of molecular data. Leonard et al. (2002) isolated DNA from the bones of 37 dogs found at archaeological sites in Mexico, Peru and Bolivia. All bones were deposited before the arrival of Columbus in the New World. A 425 base pair fragment of mtDNA was sequenced from 13 of the specimens. Ancient dog specimens were used to represent American dogs because modern dogs in the Americas may contain mtDNA from dogs imported from the Old World because they likely interbred with them in the past. The ancient Latin American dog sequences were compared to fragments from a number of modern dogs and grey wolves. The ancient Latin American dogs did not appear to be as closely related to North American grey wolves as they were to Eurasian grey wolves and dogs, suggesting that North American dogs are the descendants of Eurasian grey wolves.
However, since the remains examined came from regions of Latin America where wolves were either rare or absent, remains from dog bones preserved in Alaska were used to obtain DNA for the analysis of the same fragment of mtDNA examined in the Latin American dogs. These dog remains are from 1450-1675 C.E., after Europeans had began to colonize North America. However, they date from before the first sighting of Alaska by Europeans, so these remains are likely from true North American dogs. Eight different sequences were found in the eleven samples examined, and five were unique to these dogs. As was the case with the Latin American dogs, an analysis of the sequences suggested that the ancient Alaskan dogs were likely the descendants of Eurasian wolves and not North American ones.
These results suggest that both ancient and modern dogs throughout the world are descendants of Old World wolves. This means that the people who colonized the Americas 12 -14 000 years B.P. must have been accompanied by dogs, as no separate domestication event occurred in the New World. If dogs originated about 12 - 14 000 years ago, they must have spread across several continents in a relatively short period of time. In addition, the large amount of variation found in the ancient dog mtDNA sequences suggests that the Eurasian dogs brought to the Americas were already quite genetically diverse. It is also interesting to note that some of the ancient North American dogs possessed mtDNA sequences that were not seen in any of over 350 modern dogs. The authors of the study suggested that this is because North American dogs have essentially been replaced by Old World dogs as North American dogs were crossed with Old World dogs. A study involving modern Mexican hairless dogs (also called the Xoloitzcuintli) also supports this (Vila et al., 1997). This breed has existed in Mexico for at least 2 000 years, but modern representatives of the breed possess mtDNA haplotypes that are indistinguishable from the haplotypes of European dogs, suggesting that this breed was crossed with Old World dogs extensively in the past.
