AM. ZOOLOGIST, 7:373-381 (1967). The Evolution of Social Behavior in Dogs and Wolves J. P. SCOTT Department of Psychology, Boioling Green State University, Bowling Green, Ohio SYNOPSIS. Patterns of social behavior in the Family Canidae appear to be conservative traits in evolution. Darwin's concept of sexual selection may be broadened to "social selection," including all factors in the social environment which exert selection pressure. In a highly social species, the social environment tends to be stable, generation after generation, accounting for the stability of social behavioral patterns. The evolution of social behavior is also related to development, with different and sometimes opposite selection pressures acting at different periods in life. Some myths and misconceptions regarding dog-wolf behavior are described, and some problems for future research outlined. One of the things which distinguishes different members of the Family Canidae from each other is the differential development of social behavior. Among the close relatives of the domestic dog, the most highly social species is the wolf, and it is this characteristic which enables us to say with a good deal of certainty that dogs are most probably descended from wolves rather than from other forms, the coyote and jackal, which normally form no group larger than a mated pair or, for a brief period, the litter born to this pair. In the process of preparing an ethogram for the dog we listed some 90 behavior patterns observed in the laboratory, nursery, and free situations (Scott and Fuller, 1965). We then examined the available descriptions of behavior of the wolf, particularly those of Murie (1944), Young and Goldman (1944), and Schenkel (1947), and found all but 19 of the same patterns observed in dogs. Most of the missing ones were minor patterns of behavior that almost undoubtedly do occur in wolves and had simply escaped observation under field and zoo conditions. Conversely, the very few behavior patterns reported in wolves but not in dogs were those observed in special hunting situations and hence not likely to occur in dogs living under our conditions. The only major behavior pattern of wolves that has not yet been reported in dogs is that of a dominant wolf pinning a subordinate one to the ground by the neck (Ginsburg, 1965). We felt justified, therefore, in concluding that the dog could only have been domesticated from the wolf. As further evidence comes in from detailed behavioral studies of other Canidae, this evidence cannot be given as much weight, although the data now available for broader comparisons in no way conflict with our original conclusion. Tembrock (1957) did a detailed ethogram of the red fox and was able to identify many of the same patterns of behavior as in the dog and wolf, with minor modifications such as the tone of the bark. Significantly, Tembrock did not find the dog-wolf postures of dominance and subordination. Foxes are even more solitary than coyotes and jackals, and they apparently have never evolved patterns of agonistic behavior which would permit them to live in larger groups than the mated pair. Since the fox belongs to a different genus from the dog and wolf, we must conclude that patterns of social behavior, at least within the Family Canidae, are conservative evolutionary traits. Let us now consider the evolution of social behavior in relation to natural selection. In this context Darwin chiefly thought in terms of sexual selection, accounting for the bright colors of male birds as the result of rivalry over females. He was not greatly concerned with the evolution of social behavior itself, and it has been only within the last half century that naturalistic studies of animal behavior have revealed the fact that birds and mammals are capable of developing a high degree of social organization. It is not just man and the social insects that are capable of social behavior. (373)
374 j. P. SCOTT We can, on the basis of present knowledge, consider sexual selection and other social factors in evolution under the more inclusive title of social selection. Selection is going on not only as a matter of individual survival against physical forces, as competition between species, and as competition between individuals within species, but also for survival in relation to the social environment. For a highly social animal the social environment may be the most stable part of the environment from generation to generation and so exerts a relatively constant selection pressure over long periods of time. This constancy would, in turn, account for the conservative nature of social behavioral patterns noted in the paragraphs above. The selection pressure exerted by the social environment may be at variance with selection exerted by other environmental factors. A well known example is the antlers of deer, which have no use except during the rutting season in the pushing contests between males, and are at other seasons of the year not only a handicap in dealing with predators but are also a considerable drain on the animal's metabolism. In wolves, the defense of the den and young cubs against predators such as bears is highly adaptive from the social viewpoint, since it leads to the survival of the young, but it has little benefit for the individual, which may be injured severely as a result. From the viewpoint of population genetics the important thing is not the survival of the individual but the survival of the gene. An individual may survive to a ripe old-age but unless he succeeds in passing along his genes to the next generation he will have suffered "genetic death." Opposing selection-pressures should, therefore, produce a state of equilibrium (Hamilton, 1964). Genes do not occur as individual entities but in combinations peculiar to individuals, in shared combinations in close social relatives, and in gene pools common to populations. The processes of selection and differential survival act at all three of these levels of organization. Therefore, the survival of genes and organized groups of genes is dependent on the survival of individuals, of social groups, and of breeding populations, or species. In addition to these three different kinds of selection-pressure, the characteristics of the social environment may change as the animal develops. For example, young foxes live together in a litter of both sexes in the den of their parents, but as adults they may be associated with only one other animal, an adult of the opposite sex, during most of the year. It is appropriate, therefore, to examine the evolution of behavior in relation to periods of development. When we do this, we find that behavior which is appropriate and adaptive at one time in development is not necessarily useful in another. SELECTION IN RELATION TO DEVELOPMENT Prenatal development. We have no good figures for either dogs or wolves on prenatal mortality, but in human populations this is at least 20% of conceptions, and figures from various domestic and laboratory animals indicate that the true figure is in the neighborhood of 40% to 60%. While some of this mortality is attributed to environmental factors, it is highly probable that a large proportion of these prenatal deaths is caused by lethal genes. It would be interesting to compare the relative figures in dogs and wolves and thus obtain some indication of the relative levels of lethal genes in a wild and domestic species. This could perhaps be done by inbreeding captive wolves. Postnatal mortality. Again we have no figures for wolves, but records of dog populations indicate that a postnatal mortality rate of 15% is a relatively good one, and that many breed-populations show higher rates. Once this period is passed, the death rates are very low except for accidental causes. The interesting point here is that even with a highly protected domestic animal given care nearly as good as that given to human infants, there is still a relatively high postnatal rate of mortality. This means that natural selection has by no means been eliminated. If we estimate that
EVOLUTION OF SOCIAL BEHAVIOR 375 half the prenatal and postnatal mortality is due to environmental factors and further assume that the minimal figures given above are correct, 16% of the original population at conception has been eliminated because of genetic defects by the time the puppies have passed through early postnatal life. The actual figures are probably much higher. Evolution of behavior in the neonatal period- As indicated above, the neonatal period is one in which a high percentage of mortality takes place and hence a considerable amount of selection. If we look at the behavior of neonatal dog and wolf puppies we realize that their behavior is almost entirely adapted to the conditions of neonatal existence. The puppies are born in a very immature state, with both the eyes and ears closed. Behavior consists chiefly of nursing, a neonatal form of ingestive behavior. Elimination is accomplished in a reflex fashion in response to licking by the mother, an adaptation which keeps the nest or den area clean. The puppies vocalize in response to pain or discomfort, and almost the only other form of behavior is a slow crawl, throwing the head from side to side. An isolated puppy will continue this crawl until he touches some warm object, which may be the mother's body or the bodies of litter mates. When the mother leaves the puppies they remain in a close group or pile, showing only occasional twitching movements which probably serve to promote circulation. Investigatory movements and nursing are only aroused by tactile stimulation from the mother when she returns. It is obvious that this behavior is adapted for the conditions of neonatal existence in which the mother provides a great deal of care and protection. There follows a transition period of approximately one week, during which many of the adult forms of behavior appear at least in rudimentary forms. We have here a phenomenon which is commonplace in the insects, namely evolution toward life in two different habitats, with a period of metamorphosis in between. Many other mammals show a similar phenomenon, including man, and its theoretical consequences need to be explored. One of these is that behavior in the neonatal period does not necessarily lead into or form the foundation for later behavior. Therefore the modification of neonatal behavior by experience may leave adult behavior unchanged. In human development it has usually been assumed that each experience is important for that which follows, but this is not necessarily so, and it may explain the fact that highly different sorts of early infantile feeding and rearing seem to have relatively little effect on later behavior. Evolution of behavior in the socialization period. At the outset of this period the puppy is capable of awkward but recognizably adult forms of locomotion, feeding, and elimination. It also begins to show agonistic social behavior in the form of playful fighting, and a little later, patterns of playful sexual behavior. However, the unique behavior in this period is associated with the formation of emotional attachments to places and individuals. For a young member of a highly social species, it is essential for survival to form such an attachment rapidly and firmly. Evolution has obviously proceeded in this direction in the case of the dog and wolf, and the process is so fundamental that there is little genetic variation even between domestic breeds of dogs. All puppies are capable of forming firm attachments to people and places beginning at about three weeks of age. While we know relatively little about this process in wolf puppies, it is known that they will rapidly form close attachments to human beings if taken from the den soon after the eyes open. The chief difference between dogs and wolves appears to be the early development in the wolf of agonistic behavior which discourages most people from adopting wolf puppies at somewhat later ages. However, as Woolpy and Ginsburg (1967) have shown, it is possible for wolves to become socialized even as adults, given prolonged contact with human beings. As to the behavioral mechanisms involved, the chief one is that of distress-
376 J. P. SCOTT vocalization in the absence of familiar places and/or individuals. This and the associated internal reactions have two adaptive functions- In the first place the signal of distress attracts the attention of parents or care-taking animals, who are then able to find the puppy and relieve its distress. Under natural conditions such distress signals would attract predators as well as parental animals, and it would be interesting to know how wolf puppies respond in this respect. Of course, under ordinary conditions the wolf is a dominant predator and the puppy would be unlikely to wander far away from the familiar den which the adult wolves defend as a territory. The second adaptive function of this behavior is that it reinforces social attachment. Distress-vocalization is almost certainly accompanied by a form of internal emotional distress, and the puppy must soon learn that he feels distressed when away from familiar individuals and objects, and that this distress is relieved when he returns to them. This occurs over and over again in the life of a young puppy and should result in the formation of a firm social bond. Evolution of behavior in the juvenile period. So far I have been talking from a basis of firmly established facts, but from now on I must talk chiefly of unsolved problems. There have been relatively few observations of wolf cubs during the juvenile period. We know from the work of other authors in this symposium that adult wolves show a great deal of flexibility and adaptability in pack formation. The lone wolf is a rarity, but wolves can survive as isolated mated pairs, and packs may range in size from the relatively small ones observed by Murie in the caribou-hunting wolves of Alaska to the large pack of moosehunting wolves on Isle Royale described by Mech (1966). Murie saw a group of young wolves leave the parent pack and go off by themselves. Is this the normal method of pack formation? Also, how do older packs replenish their numbers as the members grow old and die, or do such packs simply break up and disappear? A further problem is the way in which the young wolves learn to hunt, especially if they leave the parent pack at a relatively early age. Adult wolves feed their young for long periods, bringing back pieces of meat which are cached near the den, and also vomiting food for the young animals. However, this gives no opportunity for the young wolf to participate in hunting, and it is doubtful that wolves at four months of age would be able to cope with large prey. The pubertal period. The sexual behavior of the Canidae is unique in that it includes the puzzling phenomenon of the copulatory tie. The mating male and female remain attached to each other in a helpless and awkward position for many minutes and sometimes for half an hour or so. There are reports that early settlers sometimes hunted wolves in this fashion, by tying out a female dog in estrus and then clubbing the helpless male wolf which mated with her. This contrasts with the mating behavior of herbivorous animals, which is usually very rapid and, in the case of cattle, accomplished within a few seconds. The wolf is, of course, a dominant predator and seldom attacked, so that its sexual behavior under most conditions is not non-adaptive. Even so, the possible adaptive function is not obvious and needs to be investigated. This is probably another case of behavior having a social function not correlated with individual survival. Parental period. Reproduction in the dog is obviously adapted to the life of a carnivorous animal, and presumably this is much the same in the wolf. Implantation is delayed, taking place long after conception, and growth of the fetus is slow until approximately half way through pregnancy. The gestation period is relatively short for such a large animal, being approximately 63 days. As a result, the female is heavy with young for only a short period of time. There follows a period of intense maternal care, which begins to slacken off during the transition period from two to three weeks, with the result that during the
EVOLUTION OF SOCIAL BEHAVIOR 377 period of socialization the young are more and more in one another's company while the mother leaves for long periods. The strongest social bonds should therefore be formed with the litter mates rather than the mother, and thus should lay the foundation for future pack formation. In the case of wolves, adults other than the mother join in feeding the young, and Crisler (1958) has described a captive yearling pair, not yet sexually mature, vomiting food for young cubs. Food-caching at the den is also of help to the nursing mother. The annual breeding cycle of the wolf results in the cubs being born in late April and May, at a time when many of the large herd animals are also producing their young and, consequently, food is plentiful. In most breeds of dogs, the estrus cycles of females occur at roughly 6-month intervals, with no relationship to season. An exception is the African basenji, which like the dingo has an annual cycle beginning near the autumnal equinox in northern latitudes. Presumably this cycle was modified from that of the wolf and is adapted to biotic conditions in the tropics, but we have no evidence on this point. MYTHS AND MISCONCEPTIONS CONCERNING WOLF BEHAVIOR Before continuing to list some of the important research problems connected with this species it is appropriate to list certain issues which now appear to be definitely settled. Many of these have arisen out of myths, hunters' tales, folklore, and even modern romantic fiction. No authentic cases of "wolf children." Beginning with the myth of Romulus and Remus and continuing in modern times with Kipling's Jungle Books, there have been repeated stories of young children who have grown up with wolf packs. From what we know of the way in which wolves rear their young it would be impossible for a young human infant to survive even if a female wolf adopted it. Wolves and dogs nurse their young for approximately seven weeks and thereafter feed their young on vomited food. The likelihood of a baby surviving under these conditions is so small as to be almost nonexistent. The reports of "wolf children" at the most indicate that juvenile humans have been seen in the company of wolves (Ogburn and Bose, 1959). There is no evidence whatever that these children were actually reared by wolves nor indeed that they had any close association with them. These reports come from areas where poverty and child neglect are common and, assuming that they have any basis at all, the most probable explanation is that feeble-minded children were abandoned by their parents and survived by scavenging. No evidence of any but wolf ancestry for the dog. Darwin (1859) thought that dogs must have come from two species because of their great amount of variation. However, his statement was based on the early and mistaken idea of blending inheritance. He was, of course, unaware of Mendelian inheritance and mutations as sources of variation within a species. The hypothesis of dual ancestry has been repeated in a popular book by Lorenz (1955), who even attempted to identify "jackal breeds" and "wolf breeds." Aside from the fact that historical evidence shows that cross-breeding has taken place between most, if not all European breeds in historical times, the behavioral evidence is overwhelmingly in the direction of pure wolf ancestry. Wolves are highly social animals, as are dogs, whereas jackals are like coyotes, ordinarily forming groups no larger than a mated pair. Dog vocalizations and wolf vocalizations are very similar and include nothing like the complex vocalization patterns of the jackal (Van der Merwe, 1953). The latter evidence convinced Lorenz, and he has withdrawn his hypothesis of jackal ancestry. The other possible ancestor of the dog is the coyote of North America, and here our negative evidence is less complete since almost all of the native Amerindian dogs have become extinct. However, no present domestic dogs show the peculiar vocalization pattern of the coyote. Incidentally, Darwin also concluded from trappers' and hunters' reports that wolves
378 J. P. SCOTT never bark, and if they did they must have learned it from dogs. All modern observers agree that wolves bark no matter how near or remote from domestic dogs. The basis of this myth is simply that barking is an alarm signal normally given only near the den, whereas howling is a form of communication employed on the hunt, chiefly by animals that have become separated from the pack and are re-establishing connections. Human hunters are therefore much more likely to hear wolves howling than barking. It is also undoubtedly true that many breeds of dog bark more than wolves, having probably been selected for their tendency to bark frequently. Since dogs are capable of fertile crossing with both coyotes and jackals, we cannot rule out the possibility that such crosses may have taken place in the past. However, even if this is true, the proportion of jackal or coyote genes has never reached noticeable proportions. In general, all evidence is in favor of domestication of the dog from a local population of small wolves somewhere in the Near East or Central Europe. Although modern experimenters such as Ginsburg (1965), Fentress (1967), and Murie (1944) have demonstrated how easy it is to domesticate a wolf pup, it is highly likely that domestication took place only once. The universal occurrence of sickle-shaped or curly tails in domestic dogs indicates that all domestic dogs have a common ancestor. This mutation probably occurred early in the history of domestication and was preserved because it was useful as a method of distinguishing wild from domestic animals. A T o attacks on human beings by North American wolves. According to European folklore, wolves are fierce, slavering animals with red eyes that attack human beings on sight, and there are many American stories of persons who have been frightened by wolves. Investigating these stories, Young and Goldman (1944) were able to find only one authentic case where a wolf actually attacked a human being, and this was a case where a girl was milking a cow at dusk, and where the wolf might understandably have confused the girl and the cow. Wolves do hang around hunting camps and eat refuse and offal and they have been known to follow persons on foot, but never to have actually attacked them. This raises the question as to whether the reputation of European wolves is based on fact. Dr. Pulliainen (1966) assures me that authentic cases of wolves attacking and killing people have occurred in Finland and Russia within recent years. If so, the North American and European wolves must be different from each other in their predatory behavior. On both continents men and wolves often occupy the same ecological niche as hunters of the large herd animals, and so compete with one another, competition being intensified when these same herd animals are domesticated and thus become the helpless prey of either species. Men prey on wolves on both continents, but why should only the European wolves prey on men? This raises a more general research problem, namely, how different are North American and Eurasian wolves in other respects? All taxonomists agree that they are the same species, but the two populations must have been separated for many thousands of years with little interchange of hereditary material across the Bering Strait. One would expect that the wolves of Northern Europe and Greenland would be most different from each other. Although geographically close they are genetically connected only tenuously through the distant Bering Strait. No exposure of vital parts by subordinate animals. Lorenz (1955) has suggested that the loser of a fight, or a subordinate animal, exposes the jugular vein in his neck to the dominant individual. I have observed dogs living in groups for over 20 years and recorded many different forms of behavioral adjustment between subordinate and dominant animals. Neither in these laboratory animals nor in street clogs have I ever observed anything that could be interpreted as exposing the jugular vein. Rather than exposing vital parts, the subordinate animal tends to take up defensive and protective postures. Instead of the jugular vein, the dominant dog is most
EVOLUTION OF SOCIAL BEHAVIOR 379 likely to be presented with a mouthful of snapping teeth. Schenkel, in this symposium, shows how this misconception may have arisen in the case of wolves.. SOME PROBLEMS OF SOCIAL EVOLUTION IN WOLVES Variation and polymorphism. Numerous reports indicate that wolves are morphologically variable (Jolicoeur, 1959). Murie (1944) was able to recognize easily different members of the pack he studied by differences in color and appearance. Furthermore, there are large differences between local populations. One thing that we do not know is how much variation exists between individuals and how much variation in social organization occurs in different populations and under different environmental conditions. We have reports of a great deal of variation in pack size but otherwise little detail, except that Allen and Mech (1966) believe that the so-called "lone wolf" has a short life under the conditions in which their pack was studied. Inbreeding. The degree of inbreeding within a social group is obviously related to the evolution of local populations. This problem can only be solved by identifying the individuals within a pack and following them over a period of years. This problem is also related to that of pack formation. If the normal method of forming a new pack is for a litter to move out in a body from an old pack, there will obviously be a high degree of inbreeding. On the other hand, if packs are formed by individuals from several packs, the inbreeding coefficients will be low. The evidence we have so far is that packs tend to reject outsiders. Population control. Rabb, Woolpy, and Ginsburg (1967) have found in a group of captive wolves that there are rarely over one or two litters born per year, chiefly because the dominant female in the pack will not allow others to mate. Mech (1966) has found that the Isle Royale pack remains at a very constant level of numbers. Whether or not the mechanisms are the same in both cases, it would appear that wolf packs can exercise some sort of population control. On the other hand, as Rausch (1967) has shown in an Alaskan game management area, wolves can populate an empty area quite rapidly. This is an appropriate place to mention some of the discrepancies between the reports of wolf behavior by hunters and managers of control programs and the accounts of workers who study wolves closely and with a minimum of disturbance. The latter report wolves as generally friendly, cooperative, and peaceable within their own packs, whereas the former are likely to report cases of destructive aggression and even cannibalism. The probable explanation is that hunters and those working with control programs are inevitably observing the effects of a disruption and breakdown of social organization which they themselves have brought about. This is also an appropriate place to reiterate the necessity of marking animals in the field so that they can be identified. Without knowing the identity of individuals and the social relationships between them, it is impossible to interpret their behavior in any but hypothetical ways. Scent posts and territoriality. Both dogs and wolves make extensive use of scent posts, and both guard small areas around their dens or living places. Many persons have alleged that scent posts are used to establish boundary lines which are respected by other individuals. There is no evidence that domestic dogs respect such marks as boundaries, and the usual reaction of any dog is simply to add his own contribution of urine or feces. Such scent posts do communicate the estrus condition of females, and the most plausible hypothesis is to consider this behavior as a form of communication which merely indicates that a dog has been there recently. Whether or not wolves respect scent posts as territorial boundaries is another question, but in view of the enonnous hunting ranges used by wolves it is obviously impossible for them to effectively patrol the boundaries against intruders. The answer to this question should come eventually from the detailed studies of the wanderings of wolf packs as described by Kolenosky
380 J. P. SCOTT and Johnston in this symposium. Adaptive function of behavior patterns. There are many patterns of behavior observed in domestic dogs which appear to be functionless from a human point of view. For example, dogs frequently turn around several times before lying down, and this has been imaginatively interpreted as tramping down grass or snow. It could just as easily be imagined that the dog is feeling the ground for rough places before lying down. As Tinbergen (1965) has pointed out, adaptive functions can be verified experimentally and thus become more than enlightened guesses. One method of verification would be observation of wolf behavior and the designing of some simple field experimental situations which would test the resulting hypotheses. Communication. Wolves obviously communicate a great deal, partly by visual signals arising from body posture and movements, partly through scent, and partly through vocalization. As howling and barking are studied in closer detail and variations noted, it may be possible to discover just how much is communicated. It is interesting to note in the report by Theberge and Falls (1967) that wolves can discriminate easily between recorded vocalization and the real thing, and that wolves respond more readily to human imitations than to recordings. With this kind of discrimination it should be possible for wolves to recognize each other's individual calls. Using human vocalizations it certainly should be possible to experiment with intrusions into the hunting range and thus determine whether or not it is defended as a territory. CONCLUSIONS 1. Although complete ethograms are not yet available for the coyote and jackal, a comparison of patterns of social behavior in the dog and wolf leads to the conclusion that the wolf is the sole ancestor of the dog. Other evidence indicates that domestication occurred only once. 2. Comparison of fox with dog-wolf ethograms indicates that social behavioral patterns tend to be conservative in evolution; this is accounted for by the stability of the social environment from generation to generation. 3. Darwin's concept of sexual selection may be broadened to that of "social selection," including all factors in the social environment which exert selection pressure. 4. Considered developmentally, evolution of behavior in the dog and wolf has proceeded in two different directions at different periods of life: toward adaptation for a highly protected neonatal existence, and toward adaptation for more independent adult existence. 5. Neonatal patterns of behavior, therefore, do not necessarily lead into adult patterns, and their modification in early experience may have little effect on adult patterns of behavior. 6. Various evolutionary problems are listed in relation to developmental periods. The adaptive significance of the copulatory tie is not understood; this appears to be an example of a behavioral pattern having a social function but being somewhat maladaptive for individual survival. 7. Various myths and misconceptions regarding wolf behavior are briefly discussed. 8. A number of basic facts regarding wolf behavior are still unavailable, such as: amount of variation in social organization among local populations, degree of inbreeding in wild populations, social methods of population control, the degree of territoriality with respect to hunting ranges, the adaptive significance of many behavioral patterns observed in dogs, and the effectiveness of vocal communication between individuals and between packs. REFERENCES Allen, D. L., and L. D. Mech. 1966. Private communication. Crisler, L. 1958. Arctic wild. Harper, New York. Darwin, C. 1859. Origin of species. Rand McNally, Chicago (Reprinted edition). Fentress, J. C. 1967. Observations on the behavioral development of a hand-reared male timber wolf. Am. Zoologist 7: 339-351. Ginsburg, B. E. 1965. Coaction of genetical and nongenetical factors influencing sexual behavior. In F. A. Beach, [ed.], Sex and behavior. John Wiley, New York.
EVOLUTION OF SOCIAL BEHAVIOR 381 Hamilton, W. D. 1964. Genetical evolution of social behavior. J. Theoretical Biol. 7:1-52. Jolicoeur, P. 1959. Multivariate geographical variation in the wolf Canis lupus L. Evolution 13: 283-299. Lorenz, K. 1955. Man meets dog. Houghton-Mifflin, Boston. Mech, L. D. 1966. The wolves of Isle Royale. U. S. Natl. Park Serv., Fauna Ser. 7. U. S. Govt. Printing Office. Washington, D. C. Murie, A. 1944. The wolves of Mt. McKinley. U. S. Natl. Park Serv., Fauna Ser. 5. U. S. Govt. Printing Office, Washington, D. C. Ogburn, W. F., and W. K. Bose. 1959. On the trail of the wolf-children. Genet. Psychol. Monogr. 60: 117-193. Pulliainen, E. 1966. Private communication. Rabb, G. B., J. H. Woolpy, and B. E. Ginsburg. 1967. Social relationships in a group of captive wolves. Am. Zoologist 7:305-311. Rausch, R. A. 1967. Some aspects of the population ecology of wolves, Alaska. Am. Zoologist 7:253-265. Schenkel, R. 1947. Ausdrucks-Studien an Wolfen. Behaviour 1:81-129. Scott, J. P., and J. L. Fuller. 1965. Genetics and the social behavior of the dog. University of Chicago Press, Chicago. Tembrock, G. 1957. Zur Ethologie des Rotfuchses (Vulpes vulpes L.) unter besondere Beriicksichtigung des Fortpflanzung. Zool. Garten Leipzig 23: 289-532. Theberge, J., and J. B. Falls. 1967. Howling as a means of communication in timber wolves. Am. Zoologist 7:331-338. Tinbergen, N. 1965. Behavior and natural selection. In A. J. Moore [ed.], Ideas in modern biology. Natural History Press, Garden City, N. Y. Van der Merwe, N. J. 1953. The jackal. Fauna and Flora, Transvaal Prov. Admin. Publ. 4. Woolpy, J. H., and B. E. Ginsburg. 1967. Wolf socialization; a study of temperament in a wild social species. Am. Zoologist 7:357-363. Young, S. P., and E. A. Goldman. 1944. The wolves of North America. Am. Wild). Inst., Washington, D. C.