Russian Journal of Genetics, Vol. 40, No. 6, 2004, pp. 644655. Translated from Genetika, Vol. 40, No. 6, 2004, pp. 794807.Original Russian Text Copyright 2004 by Trut, Plyusnina, Oskina.
Although the appearance of dogs is a very ancientevent in the human history, the issue of the evolution oftheir ancestors is still under debate. What caused thetransformation of the wolf into the dog? When andwhere was the dog domesticated? What directed evolu-tion of this species? Which factors determined thegenetic loss of wild reactions in dogs behavior and theformation of strong devotion of dogs to humans? Whatmade dogs so diverse although, as has been established,they all have originated from the standard and uniformwolves? Why their evolutionary rate has dramaticallyincreased in the process of domestication? These keyissues have been already addressed by Darwin , buteven today they are still under debate . The hypo-thetic significance of mutation and founder effect indomestication has been often discussed . Evolu-tionists admit the existence of rapid, saltatory genetictransformations . But even in view of this, 10 00015 000 years that have elapsed since the appearance offirst dogs [2, 7, 9, 11] seem to be too short an interval toaccumulate all mutational changes that are needed tocreate the diversity of the extant breeds.
The role of temporal parameters of development inrapid evolutionary change, typical for domestic ani-mals, is widely recognized [12, 13]. Neoteny, which isthe process of retardation of development of somaticcharacters so that the animals retain juvenile traits atmaturity, is regarded as a possible mechanism underlyingevolutionary change during domestication [6, 14, 15]. Thephenomenon of neoteny has been discovered anddescribed for invertebrates whose larvae acquired theability for sexual reproduction . The term
is used to describe temporal shifts related to dissociatedrates of development of reproductive and somatic traits.However, this term is widely used in evolutionary liter-ature as a synonym of
, which also
refers to delayed rates of somatic development accom-panying accelerated sexual maturation.
Many adult dogs in fact are similar to puppies intheir behavior and morphology. Moreover, the breeddifferentiation in the dog is believed to be possibly basedon neotenic processes so that some morphological orphysiological traits delayed at a particular developmentalstage may become characteristics of the breed [6, 14, 15].The issue on the appearance of neotenic processes is ofprincipal importance. Since variability of developmen-tal rates has an important role in evolution, a mecha-nism must exist that would protect this variability fromdirect selection pressure. However, in some cases, thisvariability hidden from direct selection may be sub-jected to indirect selection aimed at other traits.
D.K. Belyaev [3, 4] has paid particular attention tothe role of developmental processes and their regula-tion in the evolution of domestic animals. He believedthat the main processes affecting ontogenetic systemsand principally determining the progression of domes-tication occurred at its very beginning, at the time of thefirst stage of establishing relationships between the ani-mal and human as the novel environmental factor.Regardless of the scenario of this early period, the pro-gression of the process was largely determined byselection on behavior of the animals and their ability foradaptation to and coexistence with humans. During thisearliest stage of domestication, natural selection pre-served its dominating role. The man was only a factorthat shifted the direction of selection to behavior andability of the dog to exist in the new, anthropogenicenvironment. It was only gradually that natural selec-tion surrendered its role to artificial one. At firstprob-ably, since very ancient timeshumans practicedunconscious, unsystematic artificial selection, whichgradually gave way to directed systematic selectivebreeding. The exact time of this is hard to determine,
An Experiment on Fox Domesticationand Debatable Issues of Evolution of the Dog
L. N. Trut, I. Z. Plyusnina, and I. N. Oskina
Institute of Cytology and Genetics, Russian Academy of Sciences, Novosibirsk, 630090 Russia;fax: (3832) 33-12-78; e-mail: email@example.com
Received August 12, 2003
This paper is a review of the results of the authors obtained in a long-term experiment on fox domes-tication. Debatable issues of dog evolution are discussed in light of these results. It is demonstrated that geneticphysiological mechanisms of the behavior transformation during selection and the nature of the arising pheno-typic changes are associated with retarded development of corresponding ontogenetic processes. As a result ofthis retardation, the adult animals retain juvenile traits of behavior and morphology (the phenomenon of neo-teny). The role of hormonal changes caused by domestication in the evolutionary origin of neoteny is discussed.
THEORETICAL PAPERS AND REVIEWS
RUSSIAN JOURNAL OF GENETICS
AN EXPERIMENT ON FOX DOMESTICATION AND DEBATABLE ISSUES 645
but one thing is certain: target artificial selectionopened a new era in the evolution of domestic animals.These issues were thoroughly examined by Kislovskiiin his work The Problem of Controlling Evolution ofDomestic Animals .
Ascribing major evolutionary significance to theearliest stage of domestication and selection (first natu-ral, and them artificial unconscious), which has beenoperating for thousands of years, D.K.Belyaev main-tained that its consequences could be reproduced in avery short evolutionary time interval of extremelystrong directional selection. This selection for the abil-ity to adapt to a new social factor, humans, primarilytransformed the animal behavior towards domestica-tion. However, as it involved key loci of gene networks or functionally coordinated gene groups regulatingdevelopment, it could destabilize ontogeny and its tem-poral parameters.
The aim of the present study is evaluating possibleroles of each of the disputed factors in the evolution ofthe domestic dog on the basis of the evidence accumu-lated during a long-term domestication experiment withsilver foxes
This experiment started, asis generally known, at the initiative of D.K. Belyaev inthe early years of Siberian Division of the RussianAcademy of Sciences, has been ever since in progressat the Institute of Cytology and Genetics.
ONTOGENETIC MECHANISMS OF TRANSFORMATION OF FOX BEHAVIORThe prerequisite for setting up a selection experi-
ment was the genetic character of polymorphism for theexpression of defensive reactions to humans, which hadbeen found in farm populations of silver foxes .In some animals, the number of which constituted lessthan 10% of the total number, the expression of thesereactions was weak bordering on zero. In the early1960s, these animals (100 females and 30 males) wereselected from various farm populations to be the par-ents of the first generation of the experimental popula-tion. The main task at this stage of selection was elimi-nating defensive reactions to humans. In order to revealvariability in the expression of these reactions morecompletely, the animals in the selected population weresubjected to more intensive contacts with humans thanin usual practice. During these contacts, the pups weresubjected to a number of tests: the experimenterattempted to hand feed, stroke or handle them. Thistype of humananimal communication continued forthe first three to four months of life of the animals. As aresult, the emotionally negative defensive reactions tohumans in these foxes weakened, disappeared or, insome of the animals, emotionally positive reactionswere formed. The foxes that retained aggressivefear-ful reactions to humans in spite of the 3-month periodof human contacts with them, were eliminated by selec-tion from the population as soon as in 2 to 3 generations. In generation 4 of selection, the first pups
appeared that did not form aggressivefearful reactionsto humans as a result of positive contacts with them. Onthe contrary, these pups demonstrated emotionally pos-itive response to humans: when the experimenterapproached them, they whined and wagged their tailsanticipating a positive contact. The task of further selec-tion was enhancing the expression of this response toselectively form in the foxes the type of behavior similarto that of the domestic dog. Throughout the experiment,about 10 000 animals was used that produced more than50 000 pups. All pups that survived to the age of 78 months were tested for their response to human con-tact. Behavioral and autonomous components of thisresponse were evaluated in different situations: theexperimenter (1) approaches the cage; (2) stands by thecage; (3) opens the cage; (4) touches the animal; and(5) closes the cage. Selection was based primarily onthe quantitative measurements of this response [19, 20].Figure 1 illustrates the result of selection. The animalsshown in this figure were the domestication elite. Theirbehavior is similar to that of the domestic dog: they donot escape humans but actively seek contact with them.Seeing a human even at a distance, they whine, yelp,and wag their tail anticipating a contact; during the con-tact, they try to lick the experimenters face and hands.Some of these foxes, being out of cage, follow theexperimenter like a dog (Fig. 1). The first pups assignedto domestication elite category appeared in generation6 of selection. Note that this behavior was expressed inthe early development of these pups (at the age of aboutthree weeks), i.e., it was formed without any specificcontacts with the experimenter. Because of this, suchregular contacts were not conducted in further experi-ment. The contacts of animals with humans wererestricted to those with the maintenance staff. Table 1shows the dynamics of changes in the proportion of thedomestication elite at some stages of selection. In gen-eration 6, when they appeared, they constituted only1.8%; in generation F
, 17.8%; in generation F
already 35%; and in generation F
, the elite pups con-stituted nearly half of the progeny (49%). At presentalmost 70% of the progeny are assigned to the domes-tication elite. Note that selection pressure was extreme:only 3% of males and no more than 8 to 10% of femaleswere used as the parents of the next generation .
Thus, using strong systematic selection for behav-ior, a unique population of domesticated foxes was cre-ated that has a complex of behavioral reactions charac-teristic of the domestic dog rather than the fox as a spe-cies. This is a key formative consequence of selectionfor domestication.
What underlies the genetic transformation of foxbehavior produced by selection? Which physiologicalmechanisms transformed the aggressive and fearfulanimals into domesticated ones, i.e., more adapted tothe new social environment and to humans? The sensi-tive period of such adaptation (or primary socialization)is known to begin in the early postnatal development byfunctional maturation of sensory systems and locomo-
RUSSIAN JOURNAL OF GENETICS
tion. Owing to that, the animals can perceive the envi-ronment and respond to it. The development of the fear-ful response is regarded as a factor that substantiallycomplicates (if not altogether blocks) the processes oftheir exploration of the social environment and adapta-tion to it, which is true for all animals regardless of thelevel of their social organization . In dogs, thefearful response manifests mainly by the age of 4 to 6months (in some of them, by 8 to 10 months), whereasin wolf pups this response is formed by 1.5 months ofage [22, 24]. It is noteworthy that such differences in
the time of formation of fearful response in the earlypostnatal development are also observed in domesti-cated and non-domesticated foxes, which means thatselection for tame behavior involves systems control-ling the rate of development of this response. Normally,in fox pups from the non-selected population, the fear-ful response is formed on average by the age of 45 days,similarly to wolf pups. Our experiments have shownthat this age is characterized by a reduction in explor-atory activity in an novel environment (Fig. 2). By con-trast, in domesticated fox pups from generations 2830of selection (19881990), this does not happen even atthe age of 3 months. We did not estimate parameters ofexploratory behavior either in domesticated pups or inpups from the unselected population at the beginning ofselection. However, a comparison of these parametersin domesticated pups from different generations ofselection indicates that their changes result from selec-tion. For instance, the total time of motor activity testedin novel environment in domesticated representativesof generations 1718 (in 19771978) at the age of 6570 days was 82
9 s, and in representatives of genera-tions 2830, 175
5 s (1991). The high parameters ofexploratory activity in pups from generations 2830 didnot decrease by the age of 3 to 4 months being equal to188
6 s in 3-month-old and 180
8 s, in 4-month-oldpups. Due to these changes in the rates of behaviordevelopment, the sensitive period and effectiveness ofsocial adaptation increase.
Genetic effects on fear response are currently exam-ined at the molecular level. Experiments with rats haveshown that variation in the expression of this responsetested in a wide range of conditions is a direct effect of aQTL (quantitative trait locus) located on chromosome 5. Detection and localization in foxes of such QTLswhose expression would correlate with fearful responseseems to be possible in near future (Kukekova
, inpress). Some data indicate that regulation of the expres-sion of these alleles involves adrenocortical hormones.The fact is that the age of 45 days in fox pups fromunselected populations is characterized not only by theappearance of fearful response in ontogeny but also bya sharp increase of glucocorticoids in the peripheralblood (Fig. 2). In pups of the same age from the domes-ticated population, neither fearful response appearedand exploratory activity is diminished nor glucocorti-coid content is increased. Glucocorticoids may beinvolved in the determination of development rates andchanges of these rates during domestication.
HYPOTHALAMUSHYPOPHYSIS-ADRENAL SYSTEM AND BRAIN NEUROTRANSMITTERS
DURING SELECTION OF FOXESFOR DOMESTICATION
Together with the genetic transformation of foxbehavior, selection significantly decreased the func-tional activity of the pituitaryadrenal system (PA).Selection for domestication decreases the total gluco-
The behavior of an animal from the domesticatedpopulation.
The number and proportion of...