Canid : Early The Farm- Experiment

Foxes hredfortamability in a AO-yearexperiment exhibit remarkable transformations thatsuggest an interplaybetween behavioral genetics and development

Lyudmila N. Trut

scientists ponder how animals came mon? Do they stem from a single cause, and if When to be domesticated, they almost in so, what is it? In the case of the , Morey evitably wind up thinking about . The dog identifies one common factor as pedomorphosis, was probably the firstdomestic animal, and it is the retention of juvenile traitsby adults. Those the one inwhich domestication has progressed traits include both morphological ones, such the furthest?far enough to turnCants lupus into as skulls that are unusually broad for their Canis famili?ris.Evolutionary theoristshave long length, and behavioral ones, such as whining, speculated about exactly how dogs' association barking and submissiveness?all characteris with human beings may have been linked to tics thatwolves outgrow but that dogs do not. theirdivergence from theirwild forebears, Morey considers pedomorphosis in dogs a by a topic that anthropologist Darcy Morey has dis product of natural selection for earlier sexual cussed in some detail in the pages of thismaga maturity and smaller body size, features that, zine (July-August 1994). according to evolutionary theory, ought to in As Morey pointed out, debates about the crease the fitness of animals engaged in colo origins of animal domestication tend to focus nizing a new ecological niche. on "the issue of intentionality"?the extent to The common patterns are not confined to a which domestication was the result of deliber single species. In a wide range ofmammals? ate human choice. Was domestication actually herbivores and predators, large and small? "self-domestication," the colonization of new domestication seems to have brought with it ecological niches by animals such as ? strikinglysimilar changes in appearance and be Or did it result from intentional decisions by havior: changes in size, changes in coat color, answer even in the animals' human beings? How you those ques changes reproductive cy tionswill determine how you understand the cles. Our research group at the Institute of Cy and that and Genetics in has Lyudmila N. Trut is head of morphological physiological changes tology Novosibirsk, , as the research group at the domestication has brought about?whether spent decades investigating such patterns and Institute of Cytology and the results of the pressure of natural selection other questions of the early evolution of domes Genetics the Siberian of in a new niche, or as deliberately cultivated ad tic animals. Our work grew out of the interests theRussian our Department of vantageous traits. and ideas of the late director of institute, the Sciences, in Academy of Inmany ways, the question of inten K. Novosibirsk. She received her though, geneticist Dmitry Belyaev. is beside the Domestication was Like believed that the doctoral in 1980. Her tionality point. Morey, Belyaev pat degree not a event but rather a terns of observed in domesticated ani current research interests are single long, complex changes process. Natural selection and artificial selection mals resulted from that oc the patterns of evolutionary genetic changes both have at different times or curred in the course of selection. transformations at the early may operated Belyaev, even at the same time. For even if believed that the factor selected steps of animal domestication. example, pre however, key Her is historic set out to was not size or research group develop people deliberately domesti for reproduction, but behavior? ing the problem of domestica cate wolves, natural selection would still have specifically amenability to domestication, or tion as an event evolutionary been at work. The selective regime may have tamability.More than any other quality, Belyaev with the use of experimental changed drastically when wolves started living believed, tamabilitymust have determined how models, the silver including with but selective pressure continued well an animal would to life among hu theAmerican the people, adapt fox, mink, of Homo chose to do. man Because behavior is rooted in biol river otter and thewild regardless anything sapiens beings. gray Another with the debate over in for tameness and rat. Address: Institute problem ogy, selecting against aggres of is that it can overshadow other im sion means for Cytologyand Genetics of the tentionality selecting physiological changes For in in the that the hormones Russian Academy of Sciences, portant questions. example, becoming systems govern body's a 630090Novosibirsk 90, domesticated, animals have undergone host and neurochemicals. Those changes, in turn, Russia. Internet: of changes inmorphology, physiology and be could have had far-reaching effects on the de [email protected]. havior. What do those changes have in com velopment of the animals themselves, effects

160 American Scientist, Volume 87

This content downloaded from 128.208.80.176 on Mon, 10 Feb 2014 21:02:47 PM All use subject to JSTOR Terms and Conditions 1. In a a Figure the late 1950s, the Russian geneticist Dmitry K. Belyaev began decades-long effort to breed population of tame . Belyaev, then director of the Institute of Cytology and Genetics of the U.S.S.R. (now Russian) Academy of Sciences in Novosibirsk, Siberia, hoped to as show that physical and morphological changes in domestic animals such dogs could have resulted from selection for a single behavioral trait, friendliness toward people. Fourteen years after his death, the experiment continues, and the results appear to support Belyaev's hypothesis. (All photographs courtesy of the author.)

thatmight well explain why different animals years after his death, it is still in progress. would respond in similar ways when subjected Through genetic selection alone, our research same to the kinds of selective pressures. group has created a population of tame foxes To test his hypothesis, Belyaev decided to fundamentally different in temperament and be turn back the clock to the point at which ani havior from theirwild forebears. In the process we mals received the first challenge of domestica have observed some striking changes in tion. By replaying theprocess, he would be able physiology, morphology and behavior, which see to how changes inbehavior, physiology and rnirrorthe changes known in other domestic an came morphology first about. Of course, repro imals and bear out many of Belyaev's ideas. ducing the ways and means of those ancient transformations, even in the roughest outlines, Belyaev's Hypothesis a as would be formidable task. To keep things Belyaev began his experiment in 1959, a time clear and simple as possible, Belyaev designed a when Soviet genetics was starting to recover a selective-breeding program to reproduce sin from the anti-Darwinian ideology of Trofim gle major factor, strong selection pressure for Lysenko. Belyaev's own career had suffered. In tamability.He chose as his experimental model 1948 his commitment to orthodox genetics had a species taxonomically close to the dog but cost him his job as head of the Department of never before domesticated: vulpes, the sil Fur Animal Breeding at the Central Research ver fox. Belyaev's fox-breeding experiment oc Laboratory of Fur Breeding inMoscow. During cupied the last 26 years of his life. Today, 14 the 1950s he continued to conduct genetic re

1999 March-April 161

This content downloaded from 128.208.80.176 on Mon, 10 Feb 2014 21:02:47 PM All use subject to JSTOR Terms and Conditions wavy or curly hair

rolled tails

shortened tafIs, fewervertebrae

floppyears

changes in reproduc? cycle

similar mor Figure 2. Early in the process of domestication, Darwin noted long ago, animals often undergo selection phological and physiological changes. Because behavior is rooted in biology, Belyaev believed that broader effects on the ani for behavior implied selection for physiological characteristics that would have to mals' development. These effects might explain patterns in the responses of various animals domestication.

search under the guise of studying animal except the elephant. Another major evolution physiology. He moved toNovosibirsk, where ary consequence of domestication is loss of the an he helped found the Siberian Department of seasonal rhythmof reproduction. Most wild are the Soviet (now Russian) Academy of Sciences imals inmiddle latitudes genetically pro once a and became the director of the Department's grammed tomate year, during mating Institute of Cytology and Genetics, a post he seasons cued by changes in daylight. Domestic held from 1959 until his death in 1985. Under animals at the same latitudes, however, now can more once a his leadership the institute became a center of mate and bear young than year season. basic and applied research in both classical ge and in any netics and modern molecular genetics. His Belyaev believed that similarity in the pat was own work included ground-breaking investi ternsof these traits the result of selection for re gations of evolutionary change in animals un amenability to domestication. Behavioral are a der extreme conditions (including domestica sponses, he reasoned, regulated by fine tion) and of the evolutionary roles of factors balance between neurotransmitters and hor such as stress, selection for behavioral traits mones at the level of thewhole organism. The a and the environmental photoperiod, or dura genes that control that balance occupy high tion of natural daylight. Animal domestication level in the hierarchical system of the genome. were was his lifelong project, and fur bearers Even slight alterations in those regulatory genes can a in his favorite subjects. give rise to wide network of changes Early in the process of domestication, thedevelopmental processes theygovern. Thus, Belyaev noted, most domestic animals had un selecting animals forbehavior may lead to oth devel dergone the same basic morphological and er, far-reaching changes in the animals' physiological changes. Their bodies changed in opment. Because mammals fromwidely differ size and proportions, leading to the appearance ent taxonomic groups share similar regulatory of dwarf and giant breeds. The normal pattern mechanisms forhormones and neurochemistry, of coat color thathad evolved as camouflage in it is reasonable tobelieve that selecting them for the wild altered as well. Many domesticated similar behavior?tameness?should alter those animals are piebald, completely lacking pig mechanisms, and the developmental pathways mentation in specific body areas. Hair turned they govern, in similar ways. wavy or curly, as it has done in Astrakhan For Belyaev's hypothesis tomake evolution more must true. sheep, poodles, domestic donkeys, horses, pigs, ary sense, two things be Varia goats and even laboratory mice and guinea tions in tamabilitymust be determined at least an pigs. Some animals' hair also became longer partly by animal's genes, and domestication (Angora type) or shorter (rex type). must place that animal under strong selective Tails changed, too.Many breeds of dogs and pressure. We have looked into both questions. our pigs carry their tails curled up in a circle or In the early 1960s team studied the patterns semicircle. Some dogs, cats and sheep have and nature of tamability in populations of farm short tails resulting from a decrease in the num foxes.We cross-bred foxes of differentbehavior, even ber of tail vertebrae. Ears became floppy. As cross-fostered newborns and transplanted Darwin noted in chapter 1 of On theOrigin of embryos between donor and host mothers to Species, "not a single domestic animal can be known to react differently human beings. Our named which has not in some country drooping studies showed that about 35 percent of thevari ears"?a feature not found in any wild animal ations in the foxes' defense response to the ex

162 American Scientist, Volume 87

This content downloaded from 128.208.80.176 on Mon, 10 Feb 2014 21:02:47 PM All use subject to JSTOR Terms and Conditions perimenter are genetically detenriined. To get some idea of how powerful the selective pres sures on those genes might have been, our group has domesticated other animals, includ ingriver otters (Ultra lutra)and gray rats (Rattus norvegicus) caught in thewild. Out of 50 otters caught during recent years, only eight of them (16 percent) showing weak defensive behavior made a genetic contribution to the next genera tion.Among the gray rats, only 14 percent of the wild-caught yielded offspring living to adult hood. If our numbers are typical, it is clear that domestication must place wild animals under severe extreme stress and selective pressure.

The Experiment In setting up our breeding experiment, Belyaev bypassed that initial trauma. He began with 30 male foxes and 100 vixens, most of them from a commercial fur farm in . The founding foxes were already tamer than theirwild rela tives.Foxes had been farmed since thebeginning of this century, so the earliest steps of domestica tion?capture, caging and isolation from other wild foxes?had already lefttheir marks on our foxes' genes and behavior. From the outset, Belyaev selected foxes for tameness and tameness alone, a criterion we have scrupulously followed. Selection is strict; in recent years, typically not more than 4 or 5 percent ofmale offspring and about 20 percent of female offspring have been allowed to breed. To ensure that their tameness results 3. Piebald coat color is one of the most from genetic selection, we do not train the fox Figure striking mutations among domestic animals. The pattern is es.Most of them their lives in cages and spend seen in (border collie, are allowed brief "time dosed" contacts frequently dogs top right), pigs, only cows. horses and Belyaev's hypothesis predicted that with human are with their a beings. Pups caged similar mutation he called Star, seen occasionally in mothers until are VA to 2 months old. occur they farmed foxes, would with increasing frequency Then are with their littermates but they caged in foxes selected for tamability. The photograph without theirmothers. At threemonths, each above shows a fox in the selected population with the pup ismoved to its own cage. Star mutation. To evaluate the foxes for tameness, we give them a series of tests.When a pup is one month ging their tails and whining. In the sixth gener old, an experimenter offers it food from his ation bred for tameness we had to add an even hand while trying to stroke and handle the pup. higher-scoring category. Members of Class IE, The pups are tested twice, once in a cage and the "domesticated elite," are eager to establish once while moving freelywith other pups in an human contact,whimpering to attract attention enclosure, where they can choose tomake con and sniffing and licking experimenters like tact eitherwith the human experimenter orwith dogs. They start displaying this kind of behav another pup. The test is repeated monthly until ior before they are one month old. By the tenth the pups are six or seven months old. generation, 18 percent of foxpups were elite; by At seven or eight months, when the foxes the 20th, the figure had reached 35 percent. To reach sexual maturity, they are scored for tame day elite foxesmake up 70 to 80 percent of our ness and assigned to one of three classes. The experimentally selected population. least domesticated foxes, those thatflee from ex Now, 40 years and 45,000 foxes after Belyaev perimenters or bite when stroked or handled, began, our experiment has achieved an array are assigned to Class III. (Even Class III foxes of concrete results. The most obvious of them is are tamer than the calmest farm-bred foxes. a unique population of 100 foxes (at latest Among other things, they allow themselves to count), each of them the product of between 30 be hand fed.) Foxes inClass II let themselves be and 35 generations of selection. They are un petted and handled but show no emotionally usual animals, docile, eager to please and un friendly response to experimenters. Foxes in mistakably domesticated. When tested in Class I are friendly toward experimenters, wag groups in an enclosure, pups compete foratten

This content downloaded from 128.208.80.176 on Mon, 10 Feb 2014 21:02:47 PM All use subject to JSTOR Terms and Conditions as ears are Figure 4. In typical silver foxes, such those in the founding population of Belyaev's breeding experiment, erect, the tail is low slung and the fur is silver-black, save for the tip of the tail. (All drawings of foxes were made from the author's photographs.)

tion, snarling fiercely at one another as they behavior was, the later it showed the fear re seek the favor of theirhuman handler. Over the sponse and the later came the surge in itsplas years several of our domesticated foxes have ma corticosteroids. Thus, selection fordomesti escaped from the fur farm fordays. All of them cation gives rises to changes in the timing of the eventually returned. Probably theywould have postnatal development of certain physiological been unable to survive in thewild. and hormonal mechanisms underlying the for mation of social behavior. Physical Changes Other physical changes mirror those in dogs Physically, the foxes differmarkedly from their and other domesticated animals. In our foxes, wild relatives. Some of the differences have ob novel traits began to appear in the eighth to vious links to the changes in their social be tenth selected generations. The firstones we not havior. In dogs, for example, it iswell known ed were changes in the foxes' coat color, chieflya that the firstweeks of life are crucial for form loss of pigment in certain areas of thebody, lead ing primary social bonds with human beings. ing in some cases to a star-shaped pattern on the The "window" of bonding opens when a pup face similar to that seen in some breeds of dog. py becomes able to sense and explore its sur Next came traits such as floppy ears and rolled roundings, and it closes when the pup starts tails similar to those in some breeds of dog. After to fear unknown stimuli. According to our 15 to 20 generations we noted the appearance of studies, nondomesticated fox pups start re foxes with shorter tails and legs and with un sponding to auditory stimuli on day 16 after derbites or overbites. The novel traits are still birth, and their eyes are completely open by fairly rare.Most of them show up in no more day 18 or 19.On average, our domesticated fox than a few animals per 100 to a few per 10,000. pups respond to sounds two days earlier and Some have been seen in commercial popula open their eyes a day earlier than their nondo tions, though at levels at least a magnitude low mesticated cousins. Nondomesticated foxes er thanwe recorded in our domesticated foxes. first show the fear response at 6 weeks of age; domesticated ones show it after 9 weeks or Alternative Explanations even later. (Dogs show it at 8 to 12weeks, de What might have caused these changes in the pending on the breed.) As a result, domesticat fox population? Before discussing Belyaev's ed pups have more time to become incorpo explanation, we should consider other possi rated into a human social environment. bilities.Might rates and patterns of changes ob Moreover, we have found that the delayed served in foxes be due, for example, to in development of the fear response is linked to breeding? That could be true ifenough foxes in changes in plasma levels of corticosteroids, hor Belyaev's founding population carried a reces mones concerned with an animal's adaptation sive mutant gene from the trait along with a to stress. In foxes, the level of corticosteroids ris dominant normal gene thatmasked its effects. es sharply between the ages of 2 to 4 months Such mixed-gene, or heterozygous, foxeswould and reach adult levels by the age of 8 months. have been hidden carriers, unaffected by the One of our studies found that the more ad mutation themselves but capable of passing it an on to vanced animal's selection for domesticated later generations.

164 American Scientist, Volume 87

This content downloaded from 128.208.80.176 on Mon, 10 Feb 2014 21:02:47 PM All use subject to JSTOR Terms and Conditions As Morey pointed out, inbreeding might well have been rampant during the early steps of dog domestication. But it certainly cannot explain the novel traitswe have observed in our foxes, for two reasons. First, we designed the for our experimental fox population to prevent it.Through outbreeding with foxes from commercial fox farms and oth er standard methods, we have kept the in breeding coefficients forour foxpopulation be tween 0.02 and 0.07. That means thatwhenever a fox pup with a novel traithas been born into the herd, the probability that it acquired the trait through inbreeding (that is, by inheriting both of itsmutant genes from the same ances tor) has varied between only 2 and 7 percent. Second, some of the new traits are not reces or sive: They are controlled by dominant in one completely dominant genes. Any foxwith of those genes would have shown its effects; 5. in were selected to breed there could have been no "hidden carriers" in Figure Foxes Belyaev's experimental group depending on how reacted to their human Vicious foxes were excluded the they keepers. (top left) original population. no from the experimental population. Foxes showing slight fear and viciousness Another, subtler is that the novel possibility toward humans were used in cross-breeding for the next generation (top right). ties in our domesticated are classic were no population Their offspring (photograph, bottom) calm and showed negative emotional of selection for a by-products strong quantita responses to people. tive trait. In genetics, quantitative traits are characteristics that can vary over a range of gene might upset the genetic balance in some to new possibilities; unlike Gregor Mendel's peas, animals, causing them show unusual which were either smooth or wrinkly with no traits,most of them harmful to the fox.Note as an not matter middle ground, quantitative traits such that in this argument, it does animal's size, the amount of milk it produces whether the traitbeing selected for is tameness or its overall friendliness toward human be or some other quantitative trait.Any breeding a ings can be high, low or anywhere in between. program that affects polygene might have What makes selecting forquantitative traits so similar effects. is it perilous is that they (or at least the part of them The problem with that explanation that we see that is genetic) tend to be controlled not by sin does not explain why the particular we see. gle genes but by complex systems of genes, mutations do If disrupted polygenes a known as polygenes. Because polygenes are so are responsible, then the effects of selection on intricate, anything that tampers with them experiment ought to depend strongly runs the risk of upsetting other parts of an or which mutations already existed in the popu case our ganism's genetic machinery. In the of lation. If Belyaev had started with 130 foxes foxes, a breeding program that alters a poly from, say, , then their descen

1999 March-April 165

This content downloaded from 128.208.80.176 on Mon, 10 Feb 2014 21:02:47 PM All use subject to JSTOR Terms and Conditions dants today would have ended up with a com After 12 generations of , the pletely different set of novelties. Domesticat basal levels of corticosteroids in theblood plasma ing a population of wolves, or pigs, or cattle of our domesticated foxes had dropped to slight a ought to produce novel traitsmore different lymore than half the level in control group. Af still. Yet as Belyaev pointed out, when we look ter28 to 30 generations of selection, the level had our at the changes in other domesticated animals, halved again. The adrenal cortex in foxes also the most striking things about them are not responds less sharplywhen the foxes are subject even how diverse they are, but how similar. Differ ed to emotional stress.Selection has affected ent animals, domesticated by different people theneurochemistry of our foxes' brains. Changes at different times in different parts of the have taken place in the serotonin system, thought world, appear to have passed through the tobe the leadingmediator inhibitinganimals' ag a same morphological and physiological evolu gressive behavior. Compared with control con tionary pathways. How can that be? group, thebrains of our domesticated foxes According to Belyaev, the answer is not that tain higher levels of serotonin; of itsmajor domestication selects for a quantitative traitbut metabolite, 5-oxyindolacetic acid; and of trypto that it selects for a behavioralone. He considered phan hydroxylase, the key enzyme of serotonin genetic transformationsof behavior tobe thekey synthesis. Serotonin, likeother neurotransmitters, factor entraining other genetic events. Many of is critically involved in shaping an animal's de thepolygenes determining behavior may be reg velopment from itsearliest stages. ulatory, engaged in stabilizing an organism's ear lydevelopment, or ontogenesis.Ontogenesis is an Selection and Development extremely delicate process. In principle, even Evidently, then, selecting foxes for domestica slight shifts in the sequence of events could throw tionmay have triggered profound changes in it into chaos. Thus thegenes thatorchestrate those themechanisms that regulate their develop events and keep them on trackhave a powerful ment. In particular, most of the novel traits and seem role to play Which genes are they?Although nu other changes in the foxes to result from merous genes interact to stabilize an organism's shifts in the rates of certain ontogenetic development, the lead role belongs to the genes processes?in other words, from changes in that control the functioningof the neural and en timing. This fact is clear enough for some of docrine systems. Yet those same genes also gov the novelties mentioned above, such as the ear ern the systems that control an animal's behavior, lier eye opening and response to noises and un including its friendliness or hostility toward hu the delayed onset of the fear response to can some man beings. So, inprinciple, selecting animals for known stimuli. But it also explain of exam behavioral traitscan fundamentally alter the de the less obvious ones. Hoppy ears, for are velopment of an organism. ple, characteristic of newborn fox pups but As our breeding program has progressed, we may get carried over to adulthood. have indeed observed changes in some of the an Even novel coat colors may be attributable to imals' neurochemical and neurohormonal mech changes in the timing of embryonic develop we anisms. For example, we have measured a steady ment. One of the earliest novel traits ob was a drop in the hormone-producing activity of the served in our domesticated foxes loss of foxes' adrenal glands. Among several other roles pigment in parts of the head and body Belyaev in the body, the adrenal cortex comes into play determined that thispiebald pattern is governed when an animal has to adapt to stress. It releases by a gene that he named Star. Later my col hormones such as corticosteroids,which stimu league Lyudmila Prasolova and I discovered late thebody to extract energy from its reserves of that the Star gene affects themigration rate of fatsand proteins. melanoblasts, the embryonic precursors of thepig

166 American Scientist, Volume 87

This content downloaded from 128.208.80.176 on Mon, 10 Feb 2014 21:02:47 PM All use subject to JSTOR Terms and Conditions ment cells (melanocytes)that give color to an ani mal's fur.Melanocytes form in the embryonic fox's neural crest and latermove tovarious parts of the embryo's epidermis. Normally thismigra tion starts around days 28 to 31 of the embryo's development. In foxes that carry even a single copy of the Star gene, however, melanoblasts pass into thepotentially depigmented areas of the epi dermis two days later, on average. That delay may lead to the death of the tardymelanoblasts, thus altering the pigmentation inways thatgive rise to thedistinctive Star pattern. One developmental trend towhich we have devoted particular attention has to do with the growth of the skull. In 1990 and 1991, after noticing abnormal developments in the skulls and jaws of some of our foxes,we decided to study variations in the animals' cranial traits.Of course, changes in the shape of the skull are among themost obvious ways inwhich dogs differ from wolves. As I mentioned earlier, Morey believes that they are a result of selec tion (either natural or artificial) for reproductive timing and smaller body size. In our breeding experiment, we have selected foxes only forbehavior, not size; ifanything, our foxesmay be slightly longer,on average, than the ones Belyaev startedwith 40 years ago. Never theless, we found that their skulls have been changing. In our domesticated foxes of both sex es, cranial height and width tended tobe smaller, and snouts tended tobe shorter and wider, than those of a control group of farmed foxes. Another interesting change is that the cra nial morphology of domesticated adult males became somewhat ''feminized/' In farmed fox es, the crania of males tended to be larger in volume than those of females, and various oth er proportions differed sharply between the sexes. In the domesticated foxes the sexual di were morphism decreased. The differences in vol Figure 7. Changes in the foxes' coat color the ume remained, but in other respects the skulls first novel traits noted, appearing in the eighth to of males became more like those of females. tenth selected generations. The expression of the traits varied, classical rules of In of cranial allometry showed that the following genetics. Analysis a areas of in skull result either from fox homozygous for the Star gene, large changes proportions similar to those in some breeds in the of the first of depigmentation dog changes timing appearance are seen In addition some foxes the structures or from in their (top). displayed particular changes brown mottling seen in some dogs, which appeared we the skulls growth rates. Because studied as a semirecessive trait. only of adult foxes, however, we cannot judge whether any of these changes are pedomor tion. Fur farmershave tried fordecades tobreed more an phic, as Morey believes they are in dogs. foxes thatwould reproduce often than The most significant changes in developmen nually, but all theirattempts have failed. tal timing in our foxesmay be the smallest ones: In our experimental fox population, howev a those that have to do with reproduction. In the er, some reproductive traits have changed in wild, foxes reach sexual maturity when they are correlated manner. The domesticated foxes a about 8 months old. They are strict seasonal reach sexual maturity about month earlier breeders, mating once a year in response to than nondomesticateci foxes do, and they give on one changes in the length of the day (in Siberia the birth to litters that are, average, pup season mating season runs from late January to late larger. The mating has lengthened. March) and giving birth to litters ranging from Some females breed out of season, inNovem one to thirteenpups, with an average of four or ber-December or April-May, and a few of five. Natural selection has hard-wired these them have mated twice a year. Only a very traits into foxes with little or no genetic varia small number of our vixens have shown such

1999 March-April

This content downloaded from 128.208.80.176 on Mon, 10 Feb 2014 21:02:47 PM All use subject to JSTOR Terms and Conditions an incidence of certain other Figure 8. Foxes in the domesticated population show unusually high changes, curled like and including (clockwise from top left) floppy ears, shortened legs and tails, tails upward dogs', underbites and overbites. The rates of some common aberrations are compared in the table. In addition to the was most marked. Star depigmentation pattern, the increased incidence of doglike tail characteristics

a unusual behavior, and in 40 years, no offspring new morphological traitsemerge, process pre of an extraseasonal mating has survived to viously known only from archaeological evi can adulthood. Nevertheless, the striking fact is dence. Now we know that these changes a that, to our knowledge, out-of-season mating burst into population early in domestication, stresses that has never been previously observed in foxes triggered by the of captivity, and in the experiencing a natural photoperiod. many of them result from changes timing some cases of developmental processes. In the Lessons Learned changes in timing, such as earlier sexual matu or somatic Forty years into our unique lifelong experiment, rity retarded growth of characters, we believe that Dmitry Belyaev would be resemble pedomorphosis. remain.We be pleased with its progress. By intense selective Some long-standing puzzles to breeding, we have compressed into a few lieved at the start that foxes could be made a decades an ancient process that originally un reproduce twice year and all year round, like folded over thousands of years. Before our eyes, dogs. We would like to understand why this so. are "the Beast" has turned into "Beauty," as the ag has turned out not to be quite We also gressive behavior of our herd's wild progeni curious about how the vocal repertoire of foxes the calls tors entirely disappeared. We have watched changes under domestication. Some of

168 American Scientist, Volume 87

This content downloaded from 128.208.80.176 on Mon, 10 Feb 2014 21:02:47 PM All use subject to JSTOR Terms and Conditions of our adult foxes resemble those of dogs and, like those of dogs, appear to be holdovers from puppyhood, but only further study will reveal the details. The biggest unanswered question is just how much furtherour selective-breeding experiment can go. The domestic fox is not a domestic dog, but we believe that ithas the genetic potential to become more and more doglike. We can contin ue to increase that potential through further breeding, but the foxeswill realize it fully only through close contactwith human beings. Over the years, other investigators and I have raised several fox pups in domestic conditions, either in the laboratory or at home as pets. They have shown themselves to be good-tempered crea tures, as devoted as dogs but as independent as cats, capable of formingdeep-rooted pair bonds with human beings?mutual bonds, as those of us who work with them know. If our experi ment should continue, and if foxpups could be raised and trained the way dog puppies are now, there is no tellingwhat sort of animal they one become. might day 9. into the between 70 and 80 of the foxes Whether thatwill remains tobe seen. Figure Forty years experiment, percent happen bred for tameness are members of the "domesticated elite." When For the first time in 40 the future of our human-friendly years, as are raised pets, they devoted, affectionate and capable of forming strong social domestication is in doubt, experiment jeopar bonds with people (here, with technical assistant Marina Nurgalieva). The foxes dized the crisis of the Russian by continuing seek out human contact and lick experimenters' hands and faces. The friendly In 1996 the of our breed economy. population behavior is evident before the fox pups are a month old. ing herd stood at 700. Last year, with no funds to feed the foxes or to pay the salaries of our help during preparation of themanuscript and to staff,we had to cut the number to 100. Earlier YekaterinaOmelchenko for technicalassistance. we were able to cover most of our expenses by the of the foxes culled from the Bibliography selling pelts D. . 1969. Domestication of animals. Science herd. Now that source of revenue has Belyaev, breeding Journal (U.K.) 5:47-52. all but dried us as a up, leaving increasingly depen Belyaev, D. K. 1979. Destabilizing selection factor in dent on outside at a timewhen shrink funding domestication. The Journal ofHeredity 70:301-308. and in the D. A. L. In ing budgets changes grant-awarding Belyaev, K., O. Ruvinsky and N. Trut. 1981. herited activation-inactivation of the star in fox system in Russia are making long-term experi gene es. ments such as ours harder and harder to sus The Journal ofHeredity 72:264-274. D. K., and L. N. Trut. 1982. evolu tain. Like many other in our Belyaev, Accelerating enterprises country, tion. Science in theUSSR 5:24-29,60-64. we are more Recent becoming entrepreneurial. Falconer, D. S. 1981. Introduction toQuantitative Genetics. we have sold some of our foxes to Scandina ly New York: Longman. vian fur who have been breeders, pressured by Logvinenko, N. S., P. M. Krass, L. N. Trut, L. N. Ivanova animal-rights groups to develop animals that and D. K. Belyaev. 1979. Genetics and phenogenetics do not suffer stress in We also to of hormonal characteristics of animals. V. Influence of captivity. plan domestication on of and market as house a commercial ven ontogenesis estrogen- prog pups pets, functions of ovaries and adrenals ture that should lead to some if in esterone-secreting interesting, in silver fox females. Genetica 15:320-326 (in Russian). in its own av formal, experiments right.Many Morey, D. F. 1994. The early evolution of the domestic enues American Scientist 82:336-347. of both applied and basic research remain dog. forus to pursue, provided we save our unique Raff, R. A., and T. C. Kaufman. 1983. Embryos, Genes and Evolution. New York: Macmillan fox population. Publishing Company. Scott, J.P., and J.L. Fuller. 1965. Genetics and theSocial Behav Links to Internet ior of theDog. Chicago: University of Chicago Press. resources for further Acknowledgments Trut, L. N. 1988. The variable rates of evolution transforma of This article is dedicated to the K. exploration "Early memory ofDmitry tions and theirparallelism in terms of destabilizing selec Canid Domestication: The research was Belyaev. supported by grants tion. Journal ofAnimal Breeding and Genetics 105:81-90. The Farm-Fox RBD000 and RBD300 the International Sci Trut, L. N. 1996. Sex ratio in silver foxes: effects of do are from Experiment'' avail mestication and the star Theoretical and entificFunds, grants 93-04-06936 and 96-04-49972 gene. Applied able on the American Genetics 92:109-115. from theRussian Fund of Fundamental Research, Scientist Web site: Wayne, R. K. 1993. Molecular evolution of the dog fam and grant 1757 of theRussian University Fund. ily. Trends inGenetics 9:218-224. http: / /www.amsci.org/ The author expresses her toAnna Fadeeva gratitude Zeuner, F. T. 1954. Domestication of animals. InA History of amsei/articles/ translation the Russian into for of manuscript from Technology, Volume I, ed. C. Singer, E. J.Holmyard, A. R. articles99/trut.html . English. She is also grateful to Irina Plysnina for Hall and I.Williams. Oxford: Oxford University Press.

1999 March-April 169

This content downloaded from 128.208.80.176 on Mon, 10 Feb 2014 21:02:47 PM All use subject to JSTOR Terms and Conditions