Introduction

PHYLOGENY AND ADAPTIVE RADIATION

The infraclass Metatheria represents a unique mammalian radiation that can serve as a basis for comparison with the infraclass Eutheria, or placental mammals. Classically the Metatheria are considered as a single order, the Marsupialia (Simpson, 1945); however, Ride (1964) has argued that the metatherian families may be realistically grouped in four orders. Such a grouping reflects the antiquity of the early adaptive radiation of this taxon. Active investigation on the ethology of marsupials has begun only recently (Marlow, 1961; Grant, 1974; Sharman and Calaby, 1964; McManus, 1967, 1970; Ewer, 1968, 1969; Sorenson, 1970; Kaufman, 1974; Heinsohn, 1966; Russell, 1970, 1973; Stodart, 1966a, 1966b), although anatomical, ecological, and physiological studies have a longer tradition (see Tyndale-Biscoe, 1973, for an excellent review). Yet many of the more profound questions concerning the evolution of mammalian behavior will be answered only by continued comparisons between those eutherians and marsupials that have evolved convergent adaptations for similar ecological niches. The marsupials represent the only "control" group to test our hypotheses concerning the evolution of behavior within the eutherian mammals. The Metatheria deserve a rigorous treatment and we hope further efforts will be made to study the ethology of this group.

It is generally conceded by paleontologists that the Marsupialia and the Eutheria split off as two independent lines of mammalian evolution from the now extinct order Pantotheria. The marsupials as a recognizable group may have separated from this parental stock over 120 million years ago. Although fossil marsupials from the late Cretaceous are distributed in Europe and North America, it would appear that the major adaptive radiations of marsupials occurred in the geographically isolated land masses of Australia and South America. Marsupial evolution in South America involved the development of some rather large carnivorous forms: the Borhyaenidae, which occupied ecological niches similar to eutherian carnivores on the larger continental land masses. Other South American forms remained rather small (the Didelphidae and Caenolestidae), preying upon arthropods and smaller vertebrates.

In Australia the radiation of marsupials involved the production of herbivorous as well as carnivorous forms. The family Dasyuridae maintains a basic carnivore-insectivore adaptation. Some genera, such as Planigale, are rather small and resemble the Holarctic shrews. Other genera either are adapted for feeding on small invertebrates (e.g., Sminthopsis) or are adapted as more general predators (e.g., Dasyurus, Sarcophilus, and Thylacinus). The Australian bandicoots, family Peramelidae, are generalized omnivore-insectivores resembling the continental hedgehogs and armadillos in their feeding and foraging strategies.

Two major herbivorous taxa show extensive adaptive radiations in Australia, including the grazing, browsing kangaroos (family Macropodidae) and the arboreal folivores, comprising the gliders, koalas, and phalangers (superfamily Phalangeroidea). Thus, in the evolutionary history of the marsupials, adaptive radiation produced mammalian forms that replicate in their niche occupancy the major feeding strategies of eutherian mammals, which evolved on the contiguous land masses.

A MORPHOLOGICAL AND DEVELOPMENTAL COMPARISON BETWEEN METATHERIA AND EUTHERIA

These two major taxa of mammals differ in several fundamental respects. The brain structure of the Marsupialia shows some important differences, especially in the lack of a corpus callosum. In addition the relative brain size in many marsupials is considerably less than the brain size of eutherians having comparable body dimensions. The reasons for this discrepancy are difficult to pin down, but Andrew (1962) has hypothesized that brain size tends to be relatively smaller in taxa that have long been isolated from the major continental land masses (see also Jerison, 1973).

The single most profound difference between the Marsupialia and the Eutheria, however, involves structural differences in the reproductive tract (see Tyndale-Biscoe, 1973: 6-8). In addition to the differences in reproductive tract morphology, the marsupials have not evolved a complicated placenta. Most marsupial females develop a yolk sac placenta with a short functional life, although the bandicoots have evolved a placental structure similar to the eutherians One could say, as a generalization, that in eutherian mammals nutrition of the young within the reproductive tract of the female by means of a placenta is more prolonged; hence, the young of eutherians tend to be born in a somewhat more advanced condition than the young of the metatherians are.

Upon being born, the marsupial young transports itself into a teat area, which in most marsupial females is enclosed by a fold of skin, the so-called marsupium. This is not universal, however, since many species of marsupials in the families Dasyuridae and Didelphidae do not show a true pouch development. In all cases, however, the rather altricial marsupial young attaches to a teat and undergoes a great part of its early development extra-utero, reaching postpartum developmental stages over a longer period of time than do the young of comparably sized eutherians. These extended patterns of development in marsupials have some bearing upon the evolution of the signal systems between mother and young.

A Review of Marsupial Interaction Patterns

In order to acquaint the reader with the forms of interaction shown by marsupials, it is essential to offer a few selected examples of marsupial behavior. The maternal neonatal development cycle will be illustrated with Didelphis marsupialis, and the male-female courtship bouts will be illustrated with examples from Didelphis, Phalanger, and Macropus.

A GENERALIZED DEVELOPMENTAL CYCLE FOR DIDELPHIS

In the Virginia opossum (Didelphis marsupialis) the life history sequence appears to exhibit the following pattern (Reynolds, 1952). At the time of parturition the female opossum assumes a characteristic posture, squatting on the heels with the head lowered, licking at the pouch area. She remains with the long axis of her body in a vertical plane. As the young are born, they crawl unaided to her pouch. The female Didelphis does nothing to assist the movement of the young to the pouch other than to hold her position relatively constant. Similar postures during parturition have been noted for macropod marsupials (Sharman and Calaby, 1964). The tendency on the part of the female to keep her body axis in a constant alignment seems to be important in the initial orientation of the young as they move against gravity to the pouch, but olfactory and tactile cues seem to help the neonate locate the teat area.

During the neonatal and transitional phases of development for the young, they are associated intimately with the mother, being firmly attached to the teats during the neonatal phase. The mother licks the neonates and cleans the pouch throughout this and subsequent phases of development. Thus, at the transition period (Williams and Scott, 1953), when the sense organs become functional, the mother's body is the primary environment for the young. Although the female may have a nest into which she retires, the nest does not have a strong initial valance for the young. It is the female's body to which they are attached and to which they direct all their activity.

During the so-called socialization period of development (Williams and Scott, 1953), the young become capable of some locomotion and begin to move about and interact with the mother and with their litter mates. The young may now be transported on the mother's body.

They can detach from the teat, crawl on her body, and return to the pouch unassisted. As this period proceeds, the young may be left in the nest while the mother forages independently, although they will still attempt to get into the pouch whenever possible. Eventually the young become so large that all of the litter cannot crowd into the pouch. They will, however, continue to nurse while partly outside the pouch; and at this time the mother still serves as the primary focus of social interaction. She not only comes to nurse them, she may in fact carry some of them in her pouch when she leaves the nest or permit them to ride on her body.

During the early parts of the socialization period, the young may occasionally become detached from the teat. When they are detached and uncomfortable, the young give a highpitched, chirping cry. This cry causes the female to approach and stand near them until they climb on her body. In Didelphis there are no stereotyped retrieving movements by the female to pick up a young in the mouth and transport it to a nest because in a sense the mother is the nest. Rather, there is a very stereotyped reaction whereby the mother responds to the cry of the young, approaches while clicking, and stands near it, thus permitting it to climb on her or into the pouch. Although Didelphis mothers do not often touch the young with the forepaws, females of Marmosa, Sminthopsis, Dasyuroides, and many of Macropodidae typically draw the young under them with their forepaws. This movement may "direct" the young to the pouch area. When the youngsters are left alone in the nest, they may interact with one another and become familiar with smells, textures, and postures. Although the young may follow each other and the mother at the time of weaning, no prolonged following tendency appears to persist, and play is minimal.

The foregoing synopsis of development in Didelphis highlights the stages of development and the relationship between mother and young. The nest phase for Didelphis, which follows the pouch phase of early development, may not be expressed in those marsupial species that build no nest or have no permanent shelter area (e.g., Phascolarctos, and the larger macropods). Although true play is seldom observed in the young opossum, many dasyurid marsupials show play behavior, including Sarcophilus (Hediger, 1958). Although following the mother may be shown only weakly by subadult opossums and an early dispersal of the young is the rule, many marsupial species show the capacity to form sustained family groupings (e.g., Petaurus breviceps). The young of many macropods may remain with the mother through the birth of a subsequent young. Interaction rates, following tendencies and allogrooming between mother and young, have been analyzed by Russell (1973) for Macropus eugenii and Megaleia rufa.

A REVIEW OF INTERACTIONS AMONG ADULTS

Didelphis marsupialis has been studied by Reynolds (1952) and McManus (1967, 1970). When moving into a new environment, adults mark by dragging the cloacal region on the ground. The male (and occasionally the female) will also mark by licking selected points in the living space and then rubbing his cheek on the same spot. Rubbing and licking can alternate for several minutes.

Upon approaching a female, the male generally attempts to mount after sniffing her cloacal orifice. During his approaches to the female he will emit a click sound (see McManus, 1970), which is very stereotyped in its temporal patterning. If the female is unreceptive, she will threaten or move away. Threat in the adult consists of several components, including opening the mouth wide (gape threat), hissing, growling, and biting. Similar displays play a role in antipredator behavior, including the emission of foulsmelling secretions from the anal glands and the so-called "death feigning" reaction (Fancq, 1969). Should the female threaten the male, he may remain in front of her, but turn his muzzle away so that his cheek is opposite her face (McManus, 1967). If the female is receptive and stands, the male will attempt to mount her. He may rub his cheeks on the female's body before and during the mount. The male grips the female's hind legs with his feet while assuming the mount. While attempting intromission, the pair falls to one side and completes copulation while lying on the substrate. The male uses a neck grip to restrain the female while mounting and intermittently during copulation. While mounted the male rubs his lower mandible on the neck of the female and frequently rubs the sides of his snout and cheek on her neck. The mount with intromission may last more than twenty minutes.

In the Metatheria the deposition of chemical traces with potential value in communication often involves urine and saliva, as well as specialized marking postures. In desert-adapted species, sandbathing may serve as a means of chemical marking. Sandbathing involves both rolling over and dragging the ventrum in the same locus. Sandbathing probably serves the dual function of dressing the pelage and leaving chemical traces on the substrate. Such patterns have been noted for Sminthopsis (Ewer, 1968), Dasycercus, Dasyuroides, and Antechinomys.

Head turning by the male or facing away from a threatening female have also been described for Sminthopsis (Ewer, 1968:349). This pattern will be analyzed in the section for Sarcophilus.

In some marsupial species allogrooming may be shown during an interaction. Mutual muzzle licking may terminate an encounter in Sarcophilus, while in Dasycercus cristicaudata the female is especially likely to groom the male around his cheek, while the male is likely to groom the female behind the ear, on the cheek, in the vicinity of the muzzle, and in the cloacal region.

In arboreal forms, such as Phalanger gymnotis, courtship is conducted with extreme deliberation, much of it often taking place on tree branches. After an encounter, the male may often sit to one side and exhibit a gape display, licking his lips afterward. This animal may also hang upside-down, clinging with the tail and/or the hind feet above a social partner. During social encounters both Phalanger and Trichosurus vulpecula produce short clicklike sounds. In the former it is a repetitive puff-puff while in the latter it is a sharper click apparently involving the teeth.

Before mounting, many marsupials show a typical pattern of male pursuit of a female moving rapidly ahead. In the macropods the male may grasp at the female's tail with his forepaws while in slow pursuit. This movement may involve almost a patting motion and has been noted for Bettongia (Stodart, 1966b), Macropus, and Megaleia (Sharman and Calaby, 1964).

During mounting, the males of many didelphid and dasyurid marsupials employ a neck grip, with the mouth seizing the nape of the female. Sminthopsis crassicaudata is an apparent exception (Ewer, 1968). In Perameles nasuta the male adopts an almost vertical posture during intromission, which precludes a neck grip (Stodart, 1966a), and the larger macropods have lost the male neck grip during mating. Most marsupial matings are characterized by prolonged single mounts with continuous intromission. Ewer records mounts in excess of eleven hours for Sminthopsis, while the larger macropods show shorter mounts of less than twenty minutes. Perameles shows extremely brief mounts with intromission, but several intromissions in succession precede ejaculation. For a discussion of copulatory behavior, see Eisenberg (in press).

The Analysis of Interaction Patterns

INTRODUCTION

Interactions between adults or between mother and young often involve the adoption of characteristic postures. Stereotyped postures are often considered "displays" and are thought to be components of a communication system permitting the transfer of information through the eye of the presumptive receiver. In the following section we shall suggest a method for describing the presumed displays of quadrupedal mammals. We shall use our studies of courtship and copulation in the Tasmanian devil (Sarcophilus harrisii) as an introduction to our method of analysis (see Golani, in press; Eisenberg, Collins, and Wemmer, 1975).

Some familiarity with the social behavior of Sarcophilus in the wild is a prerequisite to understanding the behavior patterns shown in captive encounters. The Tasmanian devil apparently does not form a cohesive social grouping beyond that of the female and her offspring. In nature, adults tend to move alone with no fixed denning site unless a female is in the process of rearing young. Although adults move alone, there is considerable overlap among the home ranges of neighbors (Guiler, 1970a, 1970b). In an encounter between a pair of devils, the male after contacting the female generally attempts to mount her. The development of all courtship interaction between adults is a function of (a) the degree of familiarity the partners have with each other, and (b) the stage of the female's estrous cycle.

INTERACTION VIEWED AS THE PROCESS OF SUCCESSIVE "JOINT" FORMATION

Motor interaction sequences may be studied by describing the consequences of movement as changes of contact points on the animals' own bodies (Golani, in press). For example, instead of noting that an animal faces its partner and shifts its body axis 90° clockwise, we may note that the animal has in effect shifted its partner from its front to its left side. The smallest distance between the two animals can define imaginary contact points on the two animals' own bodies, and in the above example we have described the shift of this contact point on one animal's own body. A "contact point" is defined as such whether or not actual tactile input occurs. The idea implicit in this form of description is that during interaction both partners manipulate each other around their own bodies, shifting each other around various parts of their head, torso, and pelvis. Any change in the relationship between the two animals is described in terms of its consequences on the animals' own bodies. A female may be shifted from one imaginary contact point to another on the male's body through any combination of (1) the male's movements around the female, (2) the male freezing and waiting for the female to move around, or (3) both animals moving simultaneously or successively. In this form of description the motor patterns by which specific contact points are achieved and maintained become secondary to the establishment and maintenance of the contact points themselves.

The analysis of Sequence I of Sarcophilus precopulatory behavior (Fig. 1) can serve as an example. From the point of view of the shift of the contact point on the male's own body, the sequence starts with the formation of a contact point on the male's snout. This contact point is maintained so long as the male is approaching the female (Fig. la, and Fig. 2♂, phase 1). When the distance between the snouts of the two animals is diminished to a few inches, the presumptive contact point shifts to the male's mouth (Fig. lb, Fig. 2♂, phase 2), and immediately along the right lateral side of the face (Fig. 2♂, phase 3 and 4), to the posterior lateral side of the head, where it is steadily maintained (Fig. lc, d, e, f; Fig. 2♂, phase 5). From that point, the contact point shifts along the ventral side of the head to the male's snout (Fig. lg; Fig. 2♂, phases 6, 7, and 8), where it is again steadily maintained (Fig. lh; Fig. 2♂, phase 9).

When the same motor interaction sequence is analyzed in terms of the shift of the contact point on the female's own body, it turns out that all the contact points are concentrated around her mouth. At first the contact point is steadily maintained on the female's mouth. This steady maintenance is achieved by her freezing while the male is approaching (Fig. la; Fig. 2♀, phase 1). Then the contact point shifts to her mouth, where it is steadily maintained until the end of the interaction (Fig. lb, c, d, e, f, g, h; Fig. 2♀, phases 3, 4, 5, 6, and 7). Resulting from the series of movements and postures, the female's contact point with the male is maintained constant as if there were a joint between her snout and the male's head.

Once motor interaction sequences are described in these terms, it soon turns out that the behavior consists of continuous fast shifts of the contact points on the part of the two interacting animals from one steadily maintained position to another. When two Tasmanian devils encounter each other, they "slide" along each other's bodies, shifting each other toward particular positions on their own bodies where contact is kept steady for longer time periods, by either freezing or moving in such a way that the particular contact points on their own bodies are maintained constant. Subsequently, contact is released, and each animal slides again on the other's body to new contact points, which are again steadily maintained until the next shift.

Fig. 1. Single frame drawings extracted from a film of one precopulatory motor interaction sequence for the Tasmanian devil, Sarcophilus harrisii. The male is on the right. The figures were extracted on an arbitrary basis in order to give some notion of the movements performed during the interaction. The film was taken at a speed of 24 frames per second. Letters were placed in the same absolute location on the floor of the cage to indicate lateral movements. The joints and the trajectories that were established in the sequence above are summarized in Fig. 2.

Fig. 2. Diagrams of the shift of the contact point with the partner on the animals' heads during the motor-interaction sequence presented in Fig. 1. Arrows that start and end at the same spot indicate steady maintenance of contact point (joint).

The steadily maintained contact points can be conceived of as imaginary joints"between the two interacting animals: as long as such a joint is maintained, the two animals seem to form a "superorganism," which moves as one unit, subject only to the constraints of the joint. Whereas the actual contact point is fixed, the angle between the two animals may vary. The bounds of each such joint can be rigorously specified, and once the bounds of such a joint are reached due to the movements of one or both animals, the joint either "breaks" and the two animals shift to a new joint, or, as often happens, by exerting further "force" on the joint, one animal may twist the other animal and make it roll on its back. This is analogous to the twisting of someone's arm, which forces him to bend to his knees or lie on the ground. When the contact point is with the other animal's head, the head twists first, then the neck, and only then the torso. Once the rotation ability of these three body parts is pushed to its limit, the animal further rotates byrolling on its back as if being pinched and twisted by its snout. The exertion of force on the joint is not necessarily associated with actual contact or mechanical force and may be exerted from a distance. Once such a joint is established, the movements of the two animals become constrained between the contact points with the ground and the joint on the partner (see Figs. 1 and 3). All that the animals can do is either twist their bodies between these pivot points or break the joint with either the ground or the partner. Thus movements and postures are of secondary importance, in the sense that they ensue from the particular location of these interpartner joints and the trajectories between them.

It is the joints and the trajectories of movement to the next joint that determine the positions and the movements of the animals' heads. The rest of their bodies are carried along with their heads (except during mounting and copulation). The heads either are articulated to rigid joints or move along specified trajectories, and to a great extent they determine body configuration. The reader is referred to another paper (Golani, in press), where trajectories and joints are described in terms of the movements that establish and maintain them. The orderliness and simplicity of the patterning in a metatherian carnivore like the Tasmanian devil suggest its utility as a reference or baseline for the study of so-called displays in other marsupials as well as in eutherians.

It should be borne in mind that a joint can in principle be maintained through: (1) some mechanical connection such as a continuous grasp, bite, or sustained contact between two limb segments of two animals; and/or (2) the steady maintenance of sensory input, such as tactile input from vibrissae or even visual or olfactory input.

In the following sections we shall try to review the available literature concerning metatherian display by examining classes of so-called signals grouped according to the various sense organs that may be involved in the process of communication. Wherever possible we shall attempt to discuss each system in terms of the formation and maintenance of joints.

A Classification of Presumptive Signal Systems Based on Sensory Modalities

INTERACTION FORMS AND TACTILE INPUT

During various phases of interaction between adults and between adults and young animals, various parts of the body are sniffed or touched with the muzzle and/or vibrissae. In addition, certain key areas of the body are licked. The data for several species broken into sniffing and touching (Table 1) and licking (Table 2) are given for comparison. When maintained for relatively long time spans, these activities are involved in the formation of joints. It should be obvious that often tactile and chemical signals cannot be separated in this context.

When a series of species are compared during intraspecific encounters between adult males and females, sniffing and touching in the vicinity of the mouth or muzzle appear to be the most frequent occurrence, while sniffing in the cloacal area appears to be the second most frequent configuration. Thus, head to head joints are more frequent in dasyurid marsupials than head to cloaca joints. Licking or allogrooming appear to be less frequent in the Marsupialia and, when they do occur, these actions seem to involve mainly licking the muzzle or cheek.

Fig. 3. A mouth to mouth joint (c) leads to a mutual upright. A (♀) mouth to (♂) cheek joint (d, e) results in neck and upper torso rotation by the male, which does not terminate in a roll onto back, as in Fig. 1 f, g. A (♂) mouth to (♀) nape joint (h) precedes and occurs together with mounting (i). This joint is broken, and the contact shifts to a (♀) mouth to (♀) cheek joint when the female rolls on her back (j, k, 1). For further explanations, see Fig. 1.

Table 1

Table 2

In Sarcophilus mutual licking occurs on the muzzle below and up to the line connecting genal vibrissae and the mystacial vibrissae (see Figs. 1 and 2). The partners alternate licking while lying on their sides. Such a bout of allogrooming may last from thirty to forty seconds or longer.

Allogrooming between a female and her offspring is very frequent in marsupials. During the pouch phase of development the young marsupial has a high stimulus valance for the female. In addition to licking and cleaning the pouch area, the female licks and cleans her young. Contact and grooming relationships may persist beyond the pouch phase and into later developmental phases for the young animal. This phenomenon has been reviewed in the research on the larger macropods Macropus eugenii and Megaleia rufa, by Russell (1973).

Tactile input is, of course, generated during all of the interaction and mating patterns displayed by the various marsupial species. The assumption of the T position by Sarcophilus pairs during courtship involves an actual prolonged mechanical contact between the male's cheek and the female's snout. Similar configurations are shown during the interactions of Sminthopsis (Ewer, 1968), Antechinus, and Phalanger gymnotis, to mention a few. A very pronounced hip-slamming interaction is shown in Vombatus (Wünschmann, 1970).

INTERACTION FORMS AND THE OLFACTORY CHANNEL

The importance of olfactory signals in the coordination of mammalian behavior cannot be overemphasized and the subject has been recently reviewed (Eisenberg and Kleiman, 1972). The most extensive investigation of olfactory communication in the Marsupialia was conducted by Schultze-Westrum (1965, 1969), who studied the communication system of the sugar glider (Petaurus breviceps). This particular species tends to form small, communal groups based on a mated pair and their descendent offspring. In his investigations Schultze-Westrum pointed out that marking with glandular secretions can often be dimorphic; that is, in the male the sternal gland and frontal glands are active and are rubbed on the partner to promote a "community odor." On the other hand, the female sugar glider has glands associated with her pouch area that are important in attracting the newborn young to the pouch on their journey from the cloacal orifice. In addition to secretions from glandular areas, urine, saliva, and feces are important components of the chemical communication system. Special glands, such as the paraproctal gland, may be involved in antipredator behavior and may be equally functional in both sexes.

Ewer (1968) points out that in Sminthopsis crassicaudata saliva may be important in promoting individual recognition as well as sexual identity and that sniffing at the muzzle or corner of the mouth is a primary interaction pattern exhibited by this species and takes precedent over sniffing at the cloacal opening.

Urine is obviously of primary importance in certain species, and feces appear to be of secondary importance as sources of chemical information. Often the fecal material itself is of less importance than the glandular deposits left on the feces (Schultze-Westrum, 1965).

Secretions from specialized glandular areas may be deposited with specialized marking movements, including dragging the cloacal region on the substrate, rubbing the chin or cheek, rubbing the sterum or ventrum, and, in those species which practice sandbathing, combining a ventrum rub with side rubs to incorporate at one and the same time a marking movement with a pelage dressing movement. Glands on the forehead are known only for Petaurus. The marking movements involving glandular areas are summarized in Table 3.

Table 3

Saliva would appear to be of extreme importance, although the exact information potentially available in saliva remains to be experimentally verified. Nipping or biting at bark or the surface of twigs is a common occurrence in marsupials such as Marmosa robinsoni, Sminthopsis crassicaudata, Dasyurus viverinus, Trichosurus vulpecula, and Phalanger gymnotis. A social partner coming upon such "marks" often pauses to sniff. It is suggested that salivary residues remain at these points although the possibility that actual wounds in the plant surface are responsible for odor production cannot be ruled out. Often, however, excessive salivation and drooling occur during encounters with conspecifics, e.g., Didelphis, Macropus, and Megaleia. Furthermore, spreading of saliva on the chest and then rubbing the chest area on the substrate is commonly used by male Macropus giganteus and Megaleia rufa during threatening encounters.

It may be argued that the so-called face wash of the marsupials, which is done by licking the forepaws and then sweeping simultaneously on both sides of the face with the forepaws, is not only a cleaning movement but also a form of self-marking. In fact, similar "washing" patterns in eutherians such as rodents may well subserve a dual function: impregnating the face with saliva and removing foreign matter in the vicinity of the eyes and vibrissae.

During social contacts the partners may sniff at various parts of each other's body and/or lick and nibble at the same areas. This is secondary evidence for the transfer of secretions and hence some form of chemical communication. Table 1 summarizes the data for those species we have studied. A simple inspection will indicate that mutual muzzle sniffing or sniffing at the corners of the mouth are of primary importance in a great many marsupial species. Special attention is often given to the vicinity of the ear, the cloaca, and occasionally the pouch. As pointed out earlier, prolonged examination of these areas often results in the formation of a joint. In some species sniffing the cloacal region appears to be very infrequent. Often, however, urine is actively investigated. This is especially true for Sarcophilus harrisii, which appears to spend more time sniffing urine, and then the muzzle or cheek than sniffing any other part of the body.

The postures assumed during interaction often reflect the presence of glandular areas, but the correlation is not perfect. Joint formation between a pair of Sarcophilus may involve mouth to mouth or mouth to cheek contact. Odor perception may be involved during the maintenance of these contact configurations; however, the presence of an odor-emitting structure does not necessarily imply joint formation. Cloacal dragging is an important form of marking behavior in Sarcophilus (Eisenberg et al., 1975), and yet we have never observed joint formation between the muzzle and the cloaca.

Extensive studies with eutherian mammals have indicated that olfactory signals can carry information concerning species identity, sexual identity, reproductive condition, individual identity, and prevailing motivational tendencies. It is further known that olfactory stimuli can both release sexual behavior and prime sexual behavior. Odors are important in eliciting maternal behavior and may be indirect indicators of dominance and state of arousal. All of the foregoing aspects of olfactory communication remain to be experimentally investigated in the Marsupialia, but they have been reviewed for the eutherians several times (Eisenberg and Kleiman, 1972).

INTERACTION FORMS AND THE AUDITORY CHANNEL

Recently the genesis of sounds by various marsupial species has been surveyed in Eisenberg, Collins, and Wemmer (1975). The function of several of these vocalizations may be inferred from their context. We have found it useful to classify the sounds produced by marsupials into four basic syllable types: (1) Tonal syllables have energy organized into narrow frequency bands. (2) A noisy syllable does not exhibit discrete energy bands but has energy widely distributed. (3) A mixed syllable is a sound that appears as a superimposition of noise on a harmonic series. (4) The "click" is a syllable exhibiting little harmonic structure and lasting less than .02 second.

The intermediate intensity clear or tonal calls tend to be involved in courtship and during mother-infant interactions. Infants that fall out of the pouch frequently emit a chirping call to which the mother responds by approaching the infant and allowing it to either climb back into the marsupium or onto her body. Clear calls of a loud intensity are frequently involved in a context where an animal is moderately aroused and is attempting to avoid a stimulus situation.

Hisses, screams, and growls, all noisy sounds, are widely used by marsupials in threat contexts. Clicks or clicklike sounds are often produced during the initial phases of an encounter or by females in response to displaced young. Clicks appear to give information about the exact position of the sender. They are often shown in the initial phases of courtship in a wide variety of marsupial species.

Table 4 portrays the number of syllable forms identified for a series of marsupials studied by the authors. Further details considering the physical properties of the vocalizations and contexts may be found by referring to the paper by Eisenberg, Collins, and Wemmer (1975).

The best-studied marsupial from the standpoint of vocalizations is the Tasmanian devil (Sarcophilus harrisii). Clicklike sounds are produced by Sarcophilus when the jaws are clapped together during threat; however, a click form of sound production can be generated by snorting or huffing. These sounds are often made before a physical encounter between two partners or upon separation after an encounter. Sarcophilus also produces a bark vocalization during nest defense and in other thwarting contexts. The hiss to growl to whine (or whine-growl) and terminal shriek appear to be a graded series and to indicate roughly the state of arousal on the part of the sender from mild irritation on the one hand to protest at the other extreme. Fig. 4 portrays the relative frequency of sound production during different configurations assumed during the social interaction between a male and a female Sarcophilus. It is a first approximation in the direction of correlating vocalization with motor interaction dynamics.

It should be noted that the graded series of vocalizations produced by Sarcophilus tends to parallel in intensity the form of the interaction. There appears to be a partial correlation between the amount of sound energy produced per unit time and the distance between two animals' heads and their angular relationship. For example, mutual uprights, which are associated with baring of teeth and mouth to mouth contact, are accompanied by the loudest sounds—shrieks or high-intensity whine-growls. The breaking of contact is accompanied by the production of soft, clicklike syllables which could function as position indicators. The conclusion may be tentatively advanced that certain forms of contact and maintenance involve the production of various classes of sounds that are not contextspecific but are correlated with shifts in intensity of mood (see Fig. 5). Other interaction forms (e.g., mutual muzzle licking) are not accompanied by vocalization (see Eisenberg, Collins, and Wemmer, 1975).

INTERACTION FORMS AND THE VISUAL CHANNEL

Most of the species of Marsupialia are nocturnal. It would seem logical that visual display in the conduct of intraspecific encounters would be minimal. A variety of simple patterns may be shown interspecifically in the form of various antipredator mechanisms. In most of these cases, specific movements involved in threat are accompanied by vocalizations (hisses or growls). Given the preceding limitations, a wide variety of nocturnal marsupial species exhibit some form of upright posture, standing bipedally and often exposing a white ventrum to an opponent. Assumption of the bipedal posture may also involve the production of sounds. The gape display, where the mouth is opened widely exhibiting the teeth, is common in didelphine marsupials, the family Dasyuridae, and the family Peramelidae. In some species, such as Dasyuroides byrnei, the tail is ornamented, terminating in a black brush. This terminal brush of erectile hairs may be flashed in front of a social partner and may also coordinate pursuit during courtship chases. Aslin (1974) has analyzed this display.

Table 4

Fig. 4. Sound production and contexts for Sarcophilus.

Top row, left to right: Clicklike sounds (c) predominate during exploration; during approach growls (g) increase; during a chase more whining (w) and screaming (s) are added to growl sequences; during an upright whines and screams predominate.

Bottom row, left to right: After a mutual separation, clicklike sounds are prominent either within ten seconds (crosshatch) or during a two-minute period (white); in the T posture growls are prominent with whines; during nest box defense, growls, whines, and moans (m) are frequent; during face to face contact whines (w) and screams (s) are predominant.

Data abstracted from taped encounters; each continuous vocalization counted as one bout. Percentages calculated from total number of signals recorded in a defined context. With the exception of the upright context which accounted for only nine bouts of vocalization, all other contexts ranged from sixty-four to twenty-seven bouts. Vocalization bouts recorded totaled 367.

When the nocturnal marsupials are observed under red or dim lighting, it is clear that auditory and olfactory cues are involved in the coordination of interaction patterns. Tactile cues through actual body contact or vibrissae contact appear to assist in the integration of movements. Yet the presence of highly patterned motor behavior suggests that visual input plays some role in the coordination of interaction. It is a mistake, however, to assume that, given what appears to be a display posture, the "display" functions as a visual signal.

Fig. 5. Diagrams of the horizontal aspect of the shift of the contact point on the animals' own bodies during one precopulatory motor-interaction sequence. Arrows that start and end at the same spot indicate steady maintenance of contact (joints). Note that all joints are located on the head. Full lines describe an active shift. Broken lines indicate a shift resulting from the partner's movements. Broken lines on the female's head indicate a shift of the contact point on the male's throat resulting from rotation and rolling on back. Male symbol refers to the movement of the male's muzzle on the female's head. Female symbol (9) refers to the movement of the female's muzzle on the male's head.

In a paper concerning communication in the insectivore genera Suncus, Blarina, and Cryptotis, Gould (1969) points out that sound production figures prominently in the orientation and interaction sequences of these small mammals. Although the eye reduction of shrews may be a specialization and may not reflect a primitive condition, in many respects the shrews' behavior patterns are conservative and perhaps reflect the repertoire of early nocturnal eutherian mammals (Eisenberg, 1975). Throughout the interaction of shrews, postures are assumed similar to the postures of eutherian mammals, which possess highly developed eyes. The suggestion was made by Gould that these postures are not displays in themselves; "there is no apparent sign that these postures serve any visual function in communication during aggression or courtship." No true exchange of information could take place between interacting shrews in the dark unless the vibrissae touched or the animals came into physical contact (if one disregards the production of vocalizations). Yet the animals stand on their hind legs. They exhibit a turning movement, orienting the side of the head toward an opponent. They lift the head when confronted with an opponent, as if exposing the throat (see Gould, 1969: Fig. 7b, p. 25), and they may roll on the back holding the limbs upright while urinating.

The similarity of many turning movements and postures shown by small nocturnal rodents or insectivores, whose interaction often takes place in the dark or under very dim illumination, has led us to the conclusion that the evolution of visual display patterns in many mammals is much less predominant than the evolution of such action patterns in diurnal reptiles and birds. Rather, in many mammals the emphasis has been to refine aspects of auditory and chemical communication.

Perhaps it is the formation of joints between two interacting nocturnal mammals that determines in a large measure what the form of the subsequent movements will be. Raising the head and neck while approaching an opponent, or rotating the head, or rolling on the back may all result from attempts to control and keep steady some sensory input other than visual, as we discussed for Sarcophilus in the previous section.

In diurnal species such as the larger macropods display may be more complicated (Veselovsky, 1969; Grant, 1974) and associated with visual communication in the sense described for reptiles and birds. For example, during the interaction between two males of the species Megaleia rufa, the males will assume a quadrupedal strut posture while emitting a low cluck-cluck-cluck vocalization. Assuming an upright posture, the male will then lick the forearms and chest, breaking off to scratch the earth with his forepaws. He may develop an erection and urinate on his ventrum and chest, continuing to wipe at his chest and lick while salivating profusely. With an erect penis the male may begin to advance bipedally upon the opponent.

During approach by a male toward a female red kangaroo, the female will often crouch almost prone on the ground while wiggling her ears. The exact contribution by these various movements to the communication system between two individuals is very difficult to assess. As can be seen from the brief description for the red kangaroo, potential visual components in a display are intermingled with vocalizations and chemical signals.

Discussion

THE FORMATION AND MAINTENANCE OF CONFIGURATIONS IN MARSUPIAL INTERACTION SEQUENCES

As pointed out in previous sections of this chapter, it is the location of joints that shapes a large component of marsupial motor interaction sequences. Such joints are formed as the result of an attempt on the part of one or both interacting animals first to establish and then to maintain a specific contact point on the partner's body. Figs. 5 and 6 describe the location of joints on Tasmanian devils. It is clear that once the partners establish contact with each other, they tend to shift each other along their bodies to the head, and once head to head contact is established, both animals attempt to maintain head to head contact below an imaginary line drawn below the eye connecting the cheek and the muzzle. The male can establish muzzle contact with the nape of the female as long as he moves hismuzzle from a position below her eye to a position behind her ear and then ultimately to her nape. Only after establishing this muzzle to nape contact, generally through a neck bite, can the male establish cloaca to cloaca contact and ultimately intromission (Fig. 3h, i; Fig. 69).

Fig. 6. Diagrams of joints and trajectories of contact on a male and a female Sarcophilus during two precopulatory motor-interaction sequences. Note that all trajectories and joints are located below an imaginary line from cheek to muzzle. Shift of a contact point to the female's nape occurs characteristically by moving behind her ear.

Mutual upright, rolling on the back, hip-slamming, parallel alignment while lying on the ground, and other groupings of movements in Sarcophilus (Figs. 1 and 3) may be interpreted as resulting from an attempt on the part of the partners to either establish or maintain specific joints. The location of thesejoints and the trajectories between them is largely determined by the attempts to control the location of the mouth and the teeth of the partner at specific joints on the individual's own body.

A brief inspection of Table 5, which presents a listing of frequent contacts and interaction forms assumed during adult marsupial encounters, will indicate that the interaction forms in the Tasmanian devil involve the assumption of postures that control the position of the partners' mouths with respect to the rest of their bodies. The most frequent form of contact-promoting behavior involves mutual muzzle licking. Once more, the position of the mouth, armed with the large incisor teeth and surrounded by groups of vibrissae, is under mutual control. To an extent the didelphids and dasyurids have developed mechanisms for controlling the positions of the partner's mouth similar to those described for Sarcophilus. Indeed, turning the head away from the partner may not only reduce the potential intimidating effect of the mouth, thus acting as a form of cutoff (Chance, 1962), but it may also present a tactile receptor area, the genal vibrissae, to the social partner.

One can note from this rather simplified presentation (Table 5) that the morphology of the animals determines to a great extent how they will be constrained in the conduct of an intraspecific encounter. The family Macropodidae, with its specialization for browsing and grazing, has lost stabbing canine teeth and the mouth is not used in fighting to any appreciable extent (LaFollette, 1971; Grant, 1974). Conspicuous turning of the head during encounters is reduced and instead the interaction patterns are designed to position the animal for either delivering a forward kick or cuff with the forepaws to the partner or protecting itself from such activities by appropriate warding behavior with the limbs. The upright posture has taken on new significance in macropod interactions because the animals can kick one another in this posture. By the same token, they control their positions relative to one another by clasping with the forepaws in a mutual upright. In the case of the rat kangaroo (Bettongia) such kicking may take place even when the animals are lying on their sides clasping one another (Stodart, 1966a). The similarity of both the mutual upright and lying parallel on the ground to those postures illustrated for Sarcophilus is noteworthy.

Many marsupials are capable of considerable manual dexterity. This is especially true for phalangerids, dasyurids, and didelphids. It should not be surprising then that to reach out and grasp a partner in order to restrain it is not only physically possible but common. Even in the macropods, clutching at the tail of the female may be part of the courtship ritual, as the male follows an estrous female and prepares to mount her (Veselovsky, 1969).

Gripping the hind legs of the female by employing a prehensile grip with the hind foot is common in didelphine marsupials but lost in those forms that have evolved a hind foot adapted more to cursorial running than to grasping. Once more it is the morphology of the animals that shapes a large component of theirinteraction forms.

COMPARISONS OF THE METATHERIA WITH THE EUTHERIA

Since the analysis of marsupial interaction sequences and communication processes is still in its infancy, only a few limited conclusions can be drawn from a comparison of these two taxa. As previously stated, however, any future attempt to make generalizations concerning the evolution of eutherian behavior will have to be based on a comparison with metatherians—the only available "control" group. Any future comparisons will have to take into account the size, morphology, ecology, and behavioral capacity of the compared species. Otherwise, fallacious conclusions might be drawn. For instance, Ewer (1968) concluded that the behavioral repertoire of marsupials appeared to be simplified when compared with that of eutherians. She felt that marsupials responded very much to stimuli immediately impinging on them and seemed to be less controlled by centrallyprogrammed patterns of coordination. She based these conclusions, however, on an examination of Sminthopsis, one of the smallest members of the family Dasyuridae, and then generalized by comparing its behavior with eutherians, which were not only larger but also in some respects more morphologically specialized. Had she confined her comparisons of Sminthopsis to small eutherians adapted to a similar ecological niche, she might have arrived at different conclusions.

Table 5

Frequent contact and interaction forms during adult encounters.¹

Thus, the behavioral repertoire of Sminthopsis, based on a close comparison with the Tenrecidae (Eisenberg and Gould, 1970), seems no more simplified than that of this eutherian insectivore family. Indeed, one might suspect that extremely small nocturnal mammals, whether they are eutherian or metatherian, may give every appearance of being very much under the immediate control of current sensory input.

A comparison of metatherians and eutherians is thus very risky, since we are dealing with two highly complex groups and our observational methods are only now being refined to the point where we may be in a position to tease apart the significance of the various differences we observe. A comparison of motor interaction sequences of the Tasmanian devil and a carnivore of comparable size such as the jackal (Canis aureus) must be undertaken only after relevant anatomical differences have been elucidated.

In Sarcophilus the forelimbs are relatively longer than the hind limbs, unlike Canis (Moeller, 1968). Thus, during rapid locomotion the forelimbs of Sarcophilus appear to bear the major thrust, especially during a gallop. The hand of Sarcophilus shows a certain amount of dexterity. The Tasmanian devil can both manipulate objects (Eisenberg and Leyhausen, 1972; Ewer, 1969) and grasp the body of a partner, which is not the case for Canis. On the other hand, compared with Canis, Sarcophilus has a limited mobility of the head. The neck is relatively short and somewhat inflexible. As a result, the animal must shift direction by moving the whole forepart of the body, while Canis can shift by moving the head and neck alone. Also unlike Canis, Sarcophilus has a very limited ability to flex its torso in the horizontal plane. Thus, in encounters between two Tasmanian devils the use of the fore paw in grasping at a partner, a limited mobility of the head and torso, and certain differences in the gait—such as shifting front by shifting weight backward to the rigid tail and then using the hindquarters as an axis for the shift of direction —all significantly alter the form of interaction when compared with the highly articulated, cursorial Canis.

Another factor that has to be taken into account in a search for a comparison between motor interaction sequences of the two genera is the difference in social structure. Typically a male and female wild canine court each other for a prolonged period of time and in the process establish a pair relationship. In wild canids such as the jackal ( Canis aureus) the pair relationship persists for several years at least. On the other hand, Sarcophilus apparently shows no enduring pair bond, nor is there necessarily any evidence that the male devil provisions the female during her lactation period.

Thus, in an interaction sequence between two canines much time may be spent in establishing a synchronized relationship—an entratainment of motor behavior. In an encounter between a pair of Sarcophilus the male after contacting the female generally attempts to mount her. The quality and duration of preliminary interaction between a male and female canid is a function of the degree of familiarity the partners have with each other and the stage of sexual cycling for the male and female. In an encounter between a male and female Sarcophilus the degree of familiarity is also significant, but in general the male consistently attempts to mount the female, while she, depending on her estrous cycle, may be making attempts to control the form of the male's activity. The female controls the form of the male's input by rolling on her back, assuming a mouth to mouth joint through mutual upright, assuming a mouth to cheek joint, or running away. Jackal females may assume many other postures and movements which are partly dependent on transitory motor habits that persist for several days and then give way to new motor habits (Golani, 1973). Yet, both Sarcophilus and Canis assume mutual upright, crouching, so-called T postures, head and neck rotations which might lead to torso rotations which might eventually lead to rolling on the back, hip slamming, head lifting while approaching each other frontally, and several other postures and movements that are, as every student of canine behavior knows, widespread with variations in many mammalian carnivores (see Eisenberg, Collins, and Wemmer, 1975; Golani and Mendelssohn, 1971).

As pointed out throughout this chapter, such postures and movements could be interpreted as means to establish and maintain specific joints during a motor interaction sequence. One notable difference between the two genera is that vocalizations are prominent in the initial phases of a Sarcophilus interaction (see Table 4) but are limited to growls and low whines in the interactions of Canis.

Thus, if one considers postures and interaction forms as static entities, there appears to be little difference in the complexity of the repertoire of Canis and Sarcophilus. Yet, by focusing on the fine-grain dynamics of motor interaction sequences of these two genera, we could discern three major differences in motor behavior. Whether these features reflect genuine differences of the levels of neural organization and control of the relatively "primitive" marsupial compared to the relatively advanced eutherian, or whether they simply reflect the poor visual capacity of Sarcophilus is still an open question. These differences are:

(1) Whereas in Sarcophilus joints are maintained through actual mechanical contact, such as grasping or neck gripping, or through mild tactile contact; in Canis joints may be maintained visually from a distance. The only instances in which Sarcophilus did maintain a joint from a distance was in a context in which the partner was stationary.

(2) While maintaining a joint with a stationary partner, Sarcophilus fluctuates around the joint by performing minimal shifts of front and weight (Golani, in press). Canis may maintain such joints without fluctuating around them. Since joints are maintained through a homeostatic motor activity, these differences could reflect a difference in the level of motor control shown by a marsupial, on the one hand, and an eutherian on the other.

(3) In attempting to perform a hip-slamming motion, Sarcophilus, if he fails to establish contact with the female, may continue to rotate around his center of gravity as many as five full circleswhile still shifting weight in the direction of the female. Such a phenomenon has never been described in the interaction sequences of Canis. This peculiarity in behavior might reflect a difference in the level ofmotor control shown by the two respective genera and an inability on the part of Sarcophilus to shift immediately to a more appropriate behavior.

In summary, it seems to us that the more interesting differences between eutherians and metatherians should be looked for in the fine detail of the dynamics of motor behavior rather than in the simple comparison of static forms of mutual postures and configurations. We believe that once the dynamics of marsupial behavior are studied in some detail, their behavior could then serve as a baseline for inferences concerning the evolution of mammalian display.

References

Andrew, R. J., 1962. Evolution of intelligence and vocal mimicking. Science, 137:585-89.

Aslin, H., 1974. The behaviour of Dasyuroides byrnei (Marsupialia) in captivity. Z Tierpsychol., 35:187208.

Barnes, R. D., and Barthold, S. W., 1969. Reproduction and breeding in an experimental colony of Marmosa mitis Bangs (Didelphidae). J. Reprod. Fert., 6:477-82.

Chance, M. R. A., 1962. An interpretation of some agonistic postures: The role of "cut-off" acts and postures. Symp. Zool. Soc. Lond., 8:71-89.

Collins, L. R., 1973. Monotremes and Marsupials: A Reference for Zoological Institutions. Smithsonian Institution Publication no. 4888. Washington, D.C.: Smithsonian Press.

Eisenberg, J. F., 1975. Phylogeny, behavior and ecology in the Mammalia. In: Phylogeny of the Primates: An Interdisciplinary Approach, F. S. Szalay and W. P. Luckett, eds. New York: Plenum Press, pp.47-68.

Eisenberg, J. F., (in press). The evolution of the reproductive unit in the Class Mammalia. In: Lehrman Memorial Symposium, J. Rosenblatt and B. Komisaruk, eds. New Brunswick, N.J.: Rutgers University Press.

Eisenberg, J. F.; Collins, L. R.; and Wemmer, C.; 1975. Communication in the Tasmanian devil (Sarcophilus harrisii) and a survey of auditory communication in the Marsupialia. Z. Tierpsychol., 37:379-99.

Eisenberg, J. F., and Gould, E., 1970. The tenrecs: a study in mammalian behavior and evolution. Smithsonian Contributions to Zoology, No. 27. 137pp.

Eisenberg, J. F., and Kleiman, D. G., 1972. Olfactory communication in mammals. Ann. Rev. Ecol. fcf System., 3:1-32.

Eisenberg, J. F., and Leyhausen, P., 1972. The phylogenesis of predatory behaviour in mammals. Z Tierpsychol., 30:59-93.

Ewer, R. F., 1968. A preliminary survey of the behavior in captivity of the dasyurid marsupial, Sminthopsis crassicaudata (Gould). Z. Tierpsychol., 25:219-65.

Ewer, R. F., 1969. Some observations on the killing and eating of prey by two dasyurid marsupials: the mulgara, Dasycercus cristicaudata, and the Tasmanian devil, Sarcophilus harrisii. Z. Tierpsychol., 26:23-38.

Francq, E. N., 1969. Behavioral aspects of feigned death in the opossum, Didelphis marsupialis. Amer. Midland Nat., 81:556-68.

Frith, H. J., and Calaby, J. H., 1969. Kangaroos. London: C. Hurst; New York: Humanities Press.

Golani, I., 1973. Non-metric analysis of behavioral interaction sequences in captive jackals (Canis aureus L.). Behav., 44 (l-2):89-l 12.

Golani, I., (in press). Mechanisms of motor homeostasis in mammalian display. In: Perspectives in Ethology, II, P. Klopfer and P. P. G. Bateson, eds. New York: Plenum Press.

Golani, I., and Mendelssohn, H., 1971. Sequences of precopulatory behaviour of the jackal ( Canis aureus L.). Behav., 38:169-92.

Gould, E., 1969. Communication in three genera of shrews (Soricidae): Suncus, Blarina, and Cryptotis. Communication in Behavioral Biology, Part A, 3:11-31.

Grant, T. R., 1974. Observations of enclosed and freeranging grey kangaroos, Macropus giganteus. Z. Säugetierkunde, 39:65-78.

Guiler, E. R., 1970a. Observations on the Tasmanian devil, Sarcophilus harrisii (Marsupialia: Dasyuridae). I. Numbers, home range, movements, and food in two populations. Australian ]. Zool., 18:49-62.

Guiler, E. R., 1970b. Observations on the Tasmanian devil, Sarcophilus harrisii (Marsupialia: Dasyuridae). II. Reproduction, breeding, and growth of pouch young. Australian J. Zool., 18:63-70.

Hediger, H., 1958. Verhalten der Beuteltiere Marsupiala. Handbuch der Zoologie, 8:1-27.

Heinsohn, G. E., 1966. Ecology and reproduction of the Tasmanian bandicoots (Perameles gunni and Isoodon obesulus). Univ. Calif. Pubs. Zool., 80:1-96.

Jerison, H. J., 1973. Evolution of the Brain and Intelligence. New York: Academic Press.

Kaufman, J. H., 1974. The ecology and evolution of social organization in the kangaroo family Macropodidae. Amer. Zool., 14:50-62.

LaFollette, R. M., 1971. Agonistic behaviour and dominance in confined wallabies, Wallabia rufogrisea frutica. Anim. Behav., 19:93-101.

McManus, J. J., 1967. Observations on sexual behavior of the opossum, Didelphis marsupialis. J. Mammal., 48(3):486-87.

McManus, J.J., 1970. Behavior of captive opossums, Didelphis marsupialis virginiana. Amer. Midland Nat., 84:144-69.

Marlow, B. J., 1961. Reproductive behaviour of the marsupial mouse, Antechinus flavipes (Waterhouse) (Marsupialia) and the development of pouch young. Australian ]. Zool., 9(2):203-18.

Moeller, H., 1968. Zur Frage der Parallelerscheinungen bei Metatheria und Eutheria. Zeit, für Wissenschaft. Zool., 177:282-392.

Packer, W. C., 1969. Observations on the behavior of the marsupial Setonix brachyurus (Quoy and Gaimard) in an enclosure. J Mammal., 50:8-21.

Reynolds, H. C., 1952. Studies on reproduction in the opossum (Didelphis virginiana virginiana). Univ. Calif. Pubs. Zool., 52(3):223-84.

Ride, W. D. L., 1964. A review of Australian fossil marsupials. Royal Soc. Western Australia, 47(4):97131.

Russell, E., 1970. Observations on the behaviour of the red kangaroo (Megaleia rufa) in captivity. Z. Tierpsychol., 37(4):385-404.

Russell, E., 1973. Mother-young relations and early behavioural development in the marsupials, Macropus eugenii and Megaleia rufa. Z. Tierpsychol., 33:163-203.

Schultze-Westrum, T. G., 1965. Innerartiliche Verständigung durch Düfte beim Gleitbeutler Petaurus breviceps papuana Thomas (Marsupialia: Phalangeridae). Zeit, für vergleichende Physiologie, 50:151-220.

Schultze-Westrum, T. G., 1969. Social communication by chemical signals in flying phalangers. In: Olfaction and Taste, C. Pfaffmann, ed. New York: Rockefeller University Press, pp.268-77.

Sharman, G. B., and Calaby, J. H., 1964. Reproductive behavior in the red kangaroo, Megaleia rufa, in captivity. CSIRO Wildlife Research 9(l):58-85.

Simpson, G. G., 1945. The principles of classification and a classification of mammals. Bull. Amer. Mus. Nat. Hist., 8:1-350.

Smith, M., 1973. Petaurus breviceps. Mammalian Species, 30:1-4. Amer. Soc. Mammalogists.

Sorenson, M. W., 1970. Observations on the behavior of Dasycercus cristicaudata and Dasyuroides byrnei in captivity. J Mammal., 51:123-31.

Stodart, E., 1966a. Management and behaviour of breeding groups of the marsupial Perameles nasuta Geoffroy in captivity. Australian J. Zool., 14:61123.

Stodart, E., 1966b. Observations on the behaviour of the marsupial Bettongia lesueuri (Quoy and Gaimard) in an enclosure. CSIRO Wildlife Research, 11(1):91-101.

Tyndale-Biscoe, H., 1973. Life of Marsupials. New York: Elsevier.

Veselovsky, Z., 1969. Beitrag zur Kenntnis des Fortpflanzungsverhaltens der Känguruhs. Der Zool. Garten, 37:93-126.

Williams, E., and Scott, J. P., 1953. The development of social behavior patterns in the mouse in relation to natural periods. Behaviour, 6:35-64.

Woolley, P., 1966. Reproduction in Antechinus spp. and other dasyurid marsupials. In: Comparative Biology of Reproduction in Mammals, I. W. Rowlands, ed. London and New York: Academic Press, pp. 281-94.

Wünschmann, A., 1970. Die Plumpbeutler (Vombatidae). Wittenberg Lutherstadt: A. Ziemsen Verlag.