Comprehensive Description
provided by Smithsonian Contributions to Zoology
Tenrec ecaudatus (Lacepède, 1799)
Physical Description
As an adult, Tenrec ecaudatus is one of the largest living insectivores. It has virtually no tail and, when compared with the preceding genera of hedgehog tenrecs, its spinescence is vastly reduced. The dental formula is variable, being giving a total of 38 to 40 teeth. In contradistinction to all the preceding genera, this genus exhibits a profound difference in morphology when the juvenile and adult are compared. For this reason, the two age classes will be described separately.
In adult animals, the male is considerably larger than the female. His head is broad with a rather pronounced distance between the eyes which is markedly different in appearance from the female when the two are viewed frontally. The male possesses large masseter muscles which contribute to the broad appearance of his head. In addition, the canines of the male are very enlarged. The animals have a head and body length ranging from 265 to 390 mm. Captive weights range from 1600 to 2400 grams. These weights are somewhat excessive since the animals have a tendency to become rather obese in captivity. Indeed, one individual reached a maximum of slightly in excess of 3000 grams. The color of the pelage varies with the geographic origin but in general is a reddish, agouti brown. The face is a light tan and the ventrum is a light yellow. The dorsum is covered with coarse hairs some of which exhibit an almost spinescent character, especially on the crown and nape where one may speak of true spines. In addition to the dorsal hair covering, the dorsum is adorned with long hairs which protrude beyond the coat and are black in color. These are apparently involved in the perception of tactile stimuli. The hairs are exceptionally long in the middorsal area. The vibrissae are well developed in the classical pattern as outlined for Setifer and Echinops. The pentadactyl paws and hind feet are adorned with stout claws (see Figure 42).
The young juvenile Tenrec ranges in size from 85 to 160 mm in head-and-body length. The ventrum is a light yellow brown and the dorsum is colored contrastingly with dark brown longitudinal bands separated by five longitudinal bands of white spines. In the middorsal region, the row of white spines is doubled and, by means of a specialized dermal musculature, these spines may be vibrated together in the middorsal region to produce a sound (see Gould, 1965). The contrasting pattern of dark brown and white is lost at the molt to the subadult pelage. At this time the white spines are replaced by hairs. Molt begins at approximately 36 days of age and is generally completed atan age of 2 months (Figure 1). This pelage change is accomplished within the total-length-size class of 160 to 260 mm. Some middorsal spines may persist in the subadult animal and still produce sounds but these are gradually lost without replacement as the animal matures (see Figure 43).
Distribution and Habitat
Tenrec ecaudatus is widely distributed over the whole island of Madagascar. It occurs in a variety of habitats, generally characterized by some brush or undergrowth for cover and near some source of free water. Thus, the animals may be found in the rainforests of the east and in the gallery forests bordering the river systems of the west. In the vicinity of paddy fields, the animals are to be found in abundance. The animal seems equally adapted to the plateau situation and the coastal, humid rainforests.
Burrows
The structure of the burrow varies depending on the season of the year and the age and sex classes inhabiting it. Rand (1935) describes two burrows excavated in the vicinity of Ivohibe during the austral winter. These hibernating burrows were rather long and deep extending for over two meters at a depth of one meter.
During the breeding season and period of maximum activity, the females tend to construct deeper more complicated burrows, whereas the males may inhabit rather shallow systems. The description of a burrow excavated in the vicinity of Ranomafana follows: This burrow was located in natural forest at the edge of a cultivated field. The entrance lay some 7 feet above a stream. It was situated between two large stones and extended to a depth exceeding 4 feet. One of the large rocks was of sufficient size to make it impossible to reach the nest chamber although it could be touched with a stick. Another burrow system which had been excavated by the townspeople near Ranomafana was noted also to be in the vicinity of a stream.
A burrow was excavated in the vicinity of Perinet and had the following structure: It descended for a distance of about 3 feet to a blind leaf-filled chamber; however, it had more than one entrance. One entrance was situated at the base of a stump and the other which branched in a Y-pattern from the original tunnel exited on the other side of the stump near a root system. It would appear from Rand's work that the hibernating burrow of Tenrec often exhibits a single entrance plugged with earth; however, an active breeding burrow may have a Y-shaped entrance with two possible exit points connecting to form a tunnel ending in a blind nest chamber.
In order to study the microclimate of a burrow system, two burrows were selected in the vicinity of Perinet. One of these burrows proved to be empty. It was situated near the bank of a stream in primary rainforest. The ambient temperature over a 24-hour period ranged from 18° to 22.5° C. Over the same ambient range the temperature range in the burrow was 19.7° to 21.1° C, at a depth of 84 mm. A second set of 24-hour measurements was made with a burrow occupied by a female and four young. The ambient range of the shade temperature at this location was from 18° to 29° C.; over the same ambient range the burrow temperature was 22.5° to 25.5° C. This measurement was made at a depth of approximately 60 mm. Such measurements indicate that during the austral summer the burrow definitely serves to buffer the extremes of temperature to which a Tenrec may be subjected.
Activity Patterns
A family of Tenrec including a mother and four babies were allowed liberty in the large observation arena at Perinet (see Figure 4). Within this enclosure the animals constructed a burrow and we were able to monitor their activity using a photo cell device coupled to an event recorder (see Appendix F), which monitored the amount of activity at the feeding station. Such a measure, of course, does not indicate the amount of activity in other parts of the living space, but it does reflect the overall activity in the living space, since when the animals are out they make frequent stops at the feeding locus. The feeding locus was established by placing a dish of meat, to which the animals were conditioned to come, and, in addition, scattering numerous earthworms in the soil immediately adjacent to the feeding dish. Thus, in order for the animals to forage, they spent considerable time digging in the soil immediately adjacent to the dish. During such a period of time they interrupted the photo cell device.
Activity was monitored over a 48-hour period and both 24-hour intervals were remarkably similar. A bi-modal peak of activity was shown confined to the hours of darkness. Activity reached an initial maximum at 2000 hours and reached a second maximum between 0100 and 0200 hours. The first peak of activity extended from 1930 to 2130 hours; the second activity peak which was not so pronounced had a longer interval extending from 0030 to 0500 hours. Such observations are in remarkable agreement with field observations. There were ten field sightings of Tenrec, two which occurred in the late afternoon when there was still some daylight. The eight others occurred between 1800 and 2100 and 0100 to 0500 hours (see Figure 44).
The Annual Cycle, Thermoregulation, and Reproduction
During 1966 our captive colony showed a decline in feeding tendency beginning in early June and extending until the latter part of August when the animals again exhibited increased activity and feeding, reaching maximum weights in February and March, 1967. In May of 1967, we restricted their diet and induced a dramatic decline in weight persisting until late July when the animals were fed sufficient quantities to restore them to a more normal field weight.
During the first season in captivity, the decline in weight was paralleled by an absence of thermoregulation. Over an ambient range of 20° to 27.5°C, the colony exhibited cloacal temperatures varying less than half a degree from the ambient. These captive observations support the field observations that in most areas of Madagascar, especially on the high plateau, Tenrec ecaudatus enters a period of torpor in the austral autumn which persists on through the austral winter ending in the austral spring around August or September.
The gestation period is 56 to 64 days. Based on a rather extensive sample of specimens collected at Perinet, we found during the interval of 17 February to 3 March 1967, a size range in juveniles from 130 to 260 mm in total length. The majority of individuals had an average length of 180 to 190 mm. From this sample of 40 juvenile animals, we can conclude that breeding at Perinet occurs from the middle of October to early November. This inference is based on our knowledge of the growth curve and the gestation period of approximately 2 months. We can infer from the distribution of our size classes in the juveniles that the oldest juveniles caught were approximately 2 months of age. This would imply conception in the female at 4 months preceding the capture date of the juveniles.
To summarize, in the vicinity of Perinet as well as in the vicinity of Manandroy, Tenrec ecaudatus exhibits a seasonal torpor during the austral winter with breeding commencing in the austral spring during the month of October and early November. Young are then born in December and January and complete their growth and molt to subadult pelage by March and April. It would appear that unless circumstances are extremely favorable only one litter is produced on the average; however, the presence of rather small young during March does not preclude the possibility of two litters a year especially if the first litter is lost immediately in December.
Feeding Habits
Tenrec ecaudatus feeds on a variety of invertebrates, including earthworms, grubs, and orthopterans. It takes raw ground meat and mice. In general its prey-catching response strongly suggests that the animal may take other smaller vertebrates such as frogs or snakes. Its adaptability to novel foodstuffs suggests that the animal is rather omnivorous.
Ethological Studies
GENERAL MAINTENANCE BEHAVIOR
Locomotion.—Tenrec ecaudatus typically moves with a diagonal limb coordination. It will climb using the diagonal pattern, but it does not habitually take to trees. We have observed them climbing rather steep rock faces and climbing on chicken-wire fences. Tenrec has been observed to swim in rice paddies and, when startled, will take readily to water and move from one bank of a pond to another.
Exploration and utilization of living space.—When moving about in a novel environment, the animal will pause exhibiting an elongate posture and frequently lift its head and bob it slightly while inhaling and exhaling. This has been termed “testing the air.” One forepaw may be raised during the tense investigation of a novel environment while the other limbs remain planted firmly. Upon settling to an investigation, the animal may move forward slowly, pausing from time to time to insert its nose into cracks, under logs, and in the soil. Upon perceiving a potential food item, it may dig in order to retrieve its prey. When foraging unspecifically, the animal will move along with its nose only a few millimeters from the substrate or actually inserted in loose earth. Upon being startled, it will again freeze and test the air.
In a novel environment the animal will frequendy emit a piff sound which apparendy has some communicatory significance to conspecifics. If, for example, another Tenrec is moving in the vicinity and the piff sound is given, it would appear to serve as an identifier and is immediately responded to by the perceiving animal. When exploring in a novel environment, the animal soon sets up stereotyped routes and such stereotypy has an obvious adaptive advantage for, if the animal is frightened, it can automatically seek a route to its nest location.
Rest and sleep.—In the nest chamber, the animal typically sleeps in a curled posture with the head tucked ventrad. It may rest on its side in the curled position and, at higher temperatures, the animal may lay ont full length.
Marking.—When exploring a novel environment or during an encounter, the animal frequendy exhibits movements which may serve to distribute chemical signals in the environment. A commonly observed pattern includes the “perineal drag” where the animal depresses the cloacal region while locomoting forward thus dragging the cloaca against the substrate. The animal has also been noted to rub its sides against objects in the environment using a typical extension and flexion of the body while leaning against an object such as a log. Urine and feces are deposited in a locus specific fashion (see following section).
Care of body surface and comfort movements.— Tenrec ecaudatus exhibits the typical vertebrate patterns of yawning, stretching, and shaking. During the stretch when the body axis is elongated and the epaxial muscles slightly contracted, the crest may be erected. Crest erection may also occur in association with the yawn. The typical grooming pattern involves the hind foot and is termed scratching. Snout, ear, head, shoulders, under the armpit and the middle of the back may be reached by the flexible hind foot. The cloacal region and ventrum may be licked with the tongue and the toes may be nibbled after scratching. The forepaws are not involved in a face wash such as we have noted for Setifer, Echinops, and Microgale.
Urination and defecation.—Feces may be deposited near the entrance to the burrow but, in general, urination and defecation take place at a specific locus in the environment and involve a specific set of movements. Typically the animal moves forward sniffing the substrate until it encounters the location of fecal deposition. It then begins to dig with its forepaws excavating a small hole. It turns and backs into this hole and positioning its hind feet, tenses its body while it defecates and frequently urinates. It then pauses, dips its perineum, wiping the cloacal region, and kicks back to cover the feces and urine. This movement is employed by the individual when alone in its own environment and may also be exhibited as a group activity by the female and her young. Upon leaving the burrow, the female will go to her defecation spot near the entrance, dig, deposit feces, and then kick back. Generally, during this time, if the young are accompanying her, they will participate in the same reaction adjacent to her and the deposition of feces and urine can be almost synchronous once initiated by the mother.
Nest building and burrowing.—As described on p. 67, these animals do construct a rather long but simple tunnel system. The animal carries in its mouth the leaves and grass used to construct the nest. Nestbuilding behavior is shown by both sexes and, once the young have become old enough to locomote, at about the age of 3 weeks, they will also exhibit the transport of nesting materials to the burrow site.
Prey-catching behavior and foraging.—Foraging by Tenrec generally consists of probing with the nose in cracks and interstices between logs and earth. It will dig in order to excavate prey and, if it is a small prey object such as a worm or insect larvae, the prey is seized in its mouth with a slight shake of its head and immediately chewed and eaten. On the other hand, Tenrec may take rather large prey such as mice. Some experiments were conducted in order to analyze the mode of prey catching. Here are two examples:
An adult male in his home cage was offered a white mouse. The animal approached the mouse in an elongate posture and, after several hesitant attempts, rushed at the mouse and bit it once. The mouse was tossed to one side, whereupon it ran ahead a few paces and fell to one side twitching for perhaps some 90 seconds before dying. The mouse was examined and had two puncture wounds from the canines anterior and posterior to the shoulder. This prey-catching response demonstrates that the mouth alone may serve as a prey killing and capture organ. Furthermore, with large prey such as a mouse, there is a tendency to bite and toss or drop the prey only to return after it has ceased moving.
———
In contrast with this procedure, is the following protocol. A mouse was placed in a cage containing six young Tenrecs and their mother. As the mouse moved across the floor, one of the juveniles approached, evidently attracted by the sounds, and touched it with its nose. It immediately bit at the anterior part of the mouse, biting it through the head, and then shaking laterally back and forth with a slight upward and downward movement It continued to hold the mouse by the head, then pinned it down with the forepaws, and bit repeatedly over the mouse's anterior end. When the mouse was still, it began to chew and eat the animal from the head end.
———
A second mouse was introduced into a cage containing four juveniles and their mother. In this case, the mouse was approached, touched with the nose, seized with the mouth at the midpart of its body and repeatedly bitten while moving the head laterally. Then the prey was pinned with one forepaw while the animal delivered a series of rapid bites to the head and finally began chewing the mouse at the anterior end while it dangled loosely from the mouth of the Tenrec.
———
From this description we can conclude that the mouth is the primary organ of capture and killing. If the prey is large and struggles, it may be shaken from side to side and tossed or alternatively it can be pinned with the forepaws while bites are delivered at the head end of the prey.
Offensive and defensive behavior.—When startled, the Tenrec will generally run. For its size it can develop a rapid speed over a short distance. Three individuals tested in an arena were able to average 3.4 and 4.4 feet per second. The startle reaction generally involves pilo-erection. The crest and hairs in the middorsal line are prominent at this time.
Response to a predator odor is marked including pilo-erection, stamping of the forepaws, hissing and puffing, and, if brushed on the tactile hairs, the animal will deliver a slashing bite and even rush at the predator. If disturbed in the nest, the animal will attempt to attack, hiss, stamp its feet with erect crest, and exhibit a gape reaction with the mouth being held half to wide open. If contacted on the sensory hairs of the face or body, it will turn and deliver a slashing bite. When a female accompanied by juveniles is presented with predator odor, the whole group will orientate toward the source of stimulation exhibiting pilo-erection, stamping, hissing, and gaping. If they are persistently teased with a piece of cotton soaked in predator urine on the end of a stick, the whole group can be induced to rush and bite collectively at the offending object This type of reaction with a group can be induced only if there is no possibility of flight (for example, in a testing arena). The first impulse of the animal when startled in a field situation is of course to freeze and then, if further perturbed, to flee rapidly.
SOCIAL BEHAVIOR
Communication.—Auditory communication involves a variety of sounds which appear to be homologous with similar sounds noted for the preceding genera. During offensive and defensive behaviors, the animal may stamp alternately with its forefeet producing a sound. With half open mouth, the animal may hiss and, if cornered and teased, the animal will exhibit a crunching sound perhaps by grinding its teeth. If seized, the animal can emit a grunting sound which at high intensity approximates a squeak or chirp. As noted before under exploration of the living space, the animal when mildly disturbed will emit a piff sound which appears to serve as an identifying signal. (See Table 7.)
Chemical communication is implicated in the locus-specific deposition of feces and urine; we have noted under marking (p. 70), the side rub and the perineal drag.
Tactile communication is employed in a variety of contact postures including nose to rump, nose to side, nose to nose, nose to cloacal and inguinal region, nose to ear, and a grooming reaction which involves nibbling or licking the hip or nape of the partner. The possibilities for visual communication appear to be limited due to the small eye possessed by the animal and its nocturnal habits. As was noted under the description of the juvenile, young Tenrec ecaudatus have the ability to produce sounds by rubbing the dorsally situated spines together. The function of this stridulation will be discussed in a separate section.
The encounter.—Staged encounters in both a neutral arena and the home cage of a resident were run between males and females. In a neutral arena, avoidance is initially shown. After an exchange of piff signals, the animals will approach and initiate contact. In the case of a male-male encounter or a female-female encounter, contact is brief and the animals generally separate to explore alone. During the breeding season adult males will fight if placed together, and males and females will attempt to interact and show varying degrees of sexual behavior. A sexually active male will attempt to mount all females which he encounters; even pregnant females may be mounted and intromission can take place if the female is not too resistant. It is convenient to consider two phases of interaction: the contact promoting behaviors and the sexual behavior itself.
Contact and sexual behavior.—Contact involves nose to rump, nose to side, nose to nose which may, in a slightly aggressive context, involve what we have termed nose-fencing. This occurs when the animals stand together with mouths half open and push each other's nose to one side alternately. At times the nose of the partner may be grasped in the mouth without biting. In addition to the preceding, a nose to cloaca posture may be shown and a nose to ear. Highly aggressive animals may gape at one another.
More advanced contact promoting behavior involves licking by the male, especially in the area of the female's hip and her nape. The licking and nibbling of the fur on the hip may be exhibited as a gentle bite on the part of the male. The male will attempt to mount a female who does not move away after the initial investigatory phase. The male mounts by gripping with his forelegs, posterior to the female's forelimbs. At this time the male continues to lick and nip at the crest of the female. If the female is unreceptive, she may utter a “nyah” sound in bursts of three or four and this appears to be homologous to the chirp of slightly unreceptive female Echinops and Setifer. If the female is receptive, she is generally quiet after an initial vocal phase when the male effects intromission. The mount of the male is extended and may last from 5 to 12 minutes. During mount with intromission and thrusting, the male continues to lick and nibble the crest of the female and may grunt while clapping his jaws. This is a very characteristic sound and is repeated throughout the better part of the mount sequence. At the conclusion of a mount, the male on dismounting will sit upright while licking his genitalia.
For purposes of comparison we offer the following descriptions.
———
Simultaneous introduction of male and female to new cage, 14 November 1966: Male gapes and hisses, then he smells the female; licks his lips and sniffs her crown. They separate and explore. Male attempts contact but female runs off after turning to him and gaping. In about 5 minutes the male no longer makes any attempt to mount on contacting the female. The female moves to nest box; male follows and mounts. Male makes clapping sound while licking her crest but the female is so small he is unable to obtain intromission. After about 3 minutes, the female withdraws leaving the male in the box.
———
Simultaneous introduction, 14 October 1966: Female exploring; encounters male; nose to nose. Male nose to crown and nose to her ear and again nose to crown. Male follows female as she moves off. Male yawns. Female and male face one another nose to nose and “nose-fence.” Male turns to one side; mounts and commences thrusting. Male moves head from side to side, opening and closing mouth while grunting. Mount duration is 5 minutes 1 second. Male stops grunting, dismounts, and washes genitalia. The male initiates nose to crown of female, nose to side, nose to crown. Female moves away; male follows. Female yawns. Male mounts again near corner of the cage after about 7 minutes of following.
———
Male-female interaction.—As with the preceding genera, we may note that males will initiate sexual behavior regardless of the female's initial receptivity. The female who is initially unreceptive may respond with a series of stereotyped vocalizations. This may not in itself deter the male. The mount duration of the male is prolonged. The neck grip manifest in Echinops, Setifer, and Micro gale is modified in Tenrec ecaudatus to a ritualized biting and clapping of the mouth while the head is moved from side to side. Grooming in Tenrec ecaudatus, which was manifest in Microgale but absent in Echinops and Setifer, appears to be confined to nibbling the fur in the region of the nuchal crest and the side. Such grooming may grade into nips or be displayed as a nip initially (see Petter and Petter-Rousseaux, 1963; Gould and Eisenberg, 1966).
Agonistic behavior.—Fighting behavior typically involves pushing with the nose, and standing side by side while pivoting on the forefeet, thus causing the rump to smash into the side of the opponent. The shoulder may be used to push an opponent aside or the animals may stand side by side with body axes oriented in opposite directions and initiate a bite to the hip. This may be simultaneous with the animals circling and tumbling about while biting one another. A losing animal typically flees and is chased for a short distance. Fights most frequently occur between males.
PARENTAL CARE
As indicated on p. 67, the pregnant female seeks out a burrow and generally enlarges it. Nest building increases before parturition and a compact globus nest is constructed in the terminal chamber of the tunnel system. When the female has young, her tendency to defend the nest is increased and includes hissing, foot stamping with erect crest, and biting and slashing at an intruder. The female will position herself over the young while they nurse. She will lick the young especially in the area of the genitalia. Licking, in Tenrec ecaudatus, persists until the young are approximately 30 days of age. The female will retrieve the young when they are displaced from the nest by picking them up in her mouth and dragging them back to the central chamber. This begins to wane when the young are approximately 3 weeks old. As the young mature, the female is prone to lie on her side to permit nursing rather than attempt to huddle over them. The nursing response begins to wane when the young are approximately 4 weeks of age.
ONTOGENY OF BEHAVIOR
After a gestation period of some 58 to 64 days, the young are born in an altricial state but more advanced in development than most insectivores. Total length at birth ranges from 84 to 92 mm and weight from 22.8 to 27.4 grams (sample of four). The dorsal longitudinal tracts of pale hairs are visible at birth, and are approximately 5 mm long. On the ventrum, two rows of abdominal teats may be observed in the young animal. The claws are well formed. The head and ventrum are lighter in color than the dark longitudinal stripes. The eyes and auditory meatus are closed.
The animal can produce a small piff sound and squeak. It can locomote using the crossed extension pattern with some hind limb coordination. At approximately 7 days of age, the spines are quite visible in the longitudinal white stripes of the body. Especially prominent are the spines in the middorsal stripe which will become the stridulating organ. At 11 days of age, the scratch reflex is observed. It shows greater balance during turning, and the hind limbs are better coordinated with the forelimbs during forward progression. It can now stand higher on its legs when locomoting.
Eye opening begins at 9 days and is completed at 14 days. Solid food begins to be eaten at approximately 25 days of age. The molt to the reddish brown body hair begins at approximately 36 days of age and the longitudinal bands of spines and the striping effect is not discernible at 60 days of age. A growth curve is portrayed in Figure 45.
At approximately 3 weeks of age, the young begin to accompany the mother on her nightly foraging trips. The typical pattern of movement is linear, with the youngsters forming a line behind the female, although this linear pattern may vary depending on the speed of the mother. Sometimes the young may move two or three abreast behind the female. If the female stops, the young will cluster around her. They will rest when she rests, and when she forages they will cluster around the area of her activity. If the female is disturbed by some alien stimulus in the environment, she will orientate toward it and the young will do likewise.
To summarize, the response of the young to the mother: suckling and huddling in the nest occupy the first 2 weeks of life. Linear following and clustering when the female stops lasts from about 3 weeks to 35 days of age. During the interval from 35 days of age to 60 days, the young begin to forage as juvenile units and continuous contact with the female is reduced. The molt to the subadult pelage and the loss of the juvenile pelage is complete at approximately 60 days of age.
STRIDULATION AND COMMUNICATION
Gould (1965) first noted that the juveniles of Tenrec ecaudatus can produce a sound by rubbing together the quills of the middorsal region. The sound is pulsed having a broad energy distribution between 12 and 15 KHz.
In order to study the possible function of stridulation in Tenrec ecaudatus, a number of juveniles were subjected to motivational analyses. By a motivation analysis, we hoped to discern the circumstances of stridulation and the activity generally preceding and following its occurrence. We defined the motivational state of the animal in terms of the action that generally followed the stimulus. Thus, a group of behavior patterns or movements that could be associated with a subsequent tendency to flee were referred to as indicators of a flight motivational tendency. On the other hand, a set of movements and patterns associated with subsequent attack were termed indicators of an attack motivational tendency.
Juveniles were placed in a box and stimulated with a variety of foreign objects, light, touch, predator odor; and the occurrence of stridulation or nonstridulation was noted in conjunction with other movement components. From a series of some 30-odd trials, it was possible to conclude that stridulation in Tenrec ecaudatus occurs over a narrowly defined range of motivational states in contradistinction to stridulation in Hemicentetes (see p. 102). Stridulation occurs in conjunction with crest erection and erection of the center quills. Generally the subject stands high on its legs and is oriented to a potential enemy. The half open mouth and hiss very frequently accompany stridulation. Stridulation does not occur when the animal is crouched even if the center quills be erected. If the animal exhibits a strong flight tendency or a tendency to flatten or burrow into the substrate, stridulation will not occur. It would appear that stridulation accompanies a tendency to attack coupled with a strong antagonistic tendency to withhold (Figure 46).
It should be noted, however, that the threshold for stridulation in wild animals is quite different from the stridulation threshold shown by hand-raised animals. A hand-raised specimen habituated to a human observer may exhibit stridulation when excited by the appearance of the handler. Stridulation during mild excitation on the part of the hand-raised T. ecaudatus would indicate a considerable lowering in threshold for the appearance of stridulation than that threshold displayed by wild-caught young. The exact significance of this cannot be interpreted.
In order to elucidate the possible function of this stridulating, two types of playback tests were conducted with Tenrec ecaudatus (see Appendix H).
Playback test number 1: In this situation, a female with her litter of seven was established in a large arena cage (4′ × 4′ ×4′) with a nest box and a feeding location. A loudspeaker was placed alternately in several locations in the cage. Control stimuli, consisting of leaf rustling and background noise of the tape recording, were alternated with the sounds of (a) real stridulation produced by an aroused young and (b) artificial stridulation produced by stroking the dorsal quills of the young animal. The playbacks were run to the whole group under a variety of situations; however, we waited until the group was moving about in a relaxed fashion with no trace of defensive or offensive behaviors and then, on a signal, a given playback was offered whereupon a second observer noted the reactions of the group. Tests were run until the animals ceased to respond to any of the playback stimuli.
It was found that the animals would habituate to any given signal if it were presented four times in succession. Response to stridulation either real or artificial was more pronounced than response to the controls. The waxing and waning of responses to the various stimulus orders presented are included in Figure 47. When the responses during the first ten minutes of testing are considered, the following conclusions can be drawn. Control stimuli elicited no discernible response other than slight crest erection and shifting in seven animals; however, two fled. Eight tests with artificial stridulation resulted in four no discernible responses and four fleeing from the stimulus source. Out of 16 tests with real stridulation, 6 involved only crest erection and 10 involved flight from the direction of the speaker. From this it is possible to conclude that stridulation serves as a warning signal to members of the group resulting in arousal and attention. It may also serve as an indicator of identity and position of a juvenile that has been startled; however, as indicated, an identifier sound appears to involve the production of the “piff” sound. It is possible that the stridulation promotes location of the young by the female and/or location of young by young but this experiment is not decisive in answering these additional questions.
Playback test number 2: The second playback experiment involved the use of a neutral arena where a single juvenile was released at one end and, as it walked past a loudspeaker, a signal of stridulation or control sound was played back to it. Responses to this playback were scored. With the exception of two subjects, each animal was subjected to two stridulatior and two control playbacks. As a control the “huff”sound was utilized; 28 passes of stridulation were presented with 24 passes of “huff” giving a total of 52 playbacks. Thirteen subjects were employed, including one adult and 12 juveniles. In 34 of the 52 tests, there was no discernible response; 13 of these were to stridu lation and 21 to the “huff” sound. There were 18 responses; 11 of the 18 responses were approach and were directed toward stridulation; 6 responses involved moving away and 4 of the 6 were to stridulation. Out of the 18 responses, 15 were to stridulation.
This second experiment would indicate that in the novel arena, where the animal is already aroused, stridulation does not necessarily promote flight but may actually promote approach; thus, the nature of the response to stridulation is very much'a function of the testing situation. If the animals are in a group and subjected to a sudden sound of stridulation, they become aroused and move away or avoid. If they are already somewhat aroused in a novel situation and stridulation is played to them, they may approach and investigate. Thus, we are not in a position to decide on the exact function of stridulation in Tenrec ecaudatus. It may be that it promotes arousal and serves to warn of a potentially dangerous situation which may lead to breaking, running, or scattering of the group. On the other hand, in already aroused animals, it may promote location of young by the female and/or location of the young by other young.
SOCIAL ORGANIZATION
From Rand's (1935) evidence and our own field studies, we may conclude that, with the exception of the mother-young group, adult Tenrec ecaudatus forage and hibernate alone. Pairing and the pair association must be brief and probably take place in the austral spring. After a 2-month gestation period, the youngsters develop in the burrow system and begin to accompany the mother on her foraging expeditions when they are approximately 3 weeks of age. Great cohesion is shown by the mother and her group of littermates. Linear following, aggregation to the female on being startled, and foraging with the female are all manifest over a rather prolonged period of time. Our observations would indicate that the female-young foraging unit may persist for 2 to 3 1/2 weeks. During this time routes from the nesting area to the feeding ground and back are learned by the young. Selection of foodstuffs is potentiated by the association of the young with the female while she feeds. Coordination among the female and her young is ensured by the linear following tendency. The piff sound and perhaps stridulation are involved as identifiers and indicators of mood within the group. The littermates themselves may continue to associate in their foraging when about 2 months of age, although by this time the integrity of the family unit has broken down. Apparently the animals begin to take up residence alone at the time of hibernation only to re-emerge and initiate the cycle again in the austral spring.
The Ecology of Hemicentetes (Mivart, 1871)
INTRODUCTION
The genus Hemicentetes contains two species, H. semispinosus and H. nigriceps. These animals show a high tooth number, with a dental formula of for a total of 40. The teeth, however, show a reduction in size and, when compared with that of the other genera, the skull is markedly elongate and delicate. Hemicentetes possesses quills scattered on its dorsum especially concentrated on the crown and in the light colored stripes extending the length of the body. The quills are barbed and detachable with the exception of the modified group of quills in the central posterior region of the dorsum. This group of specialized quills, termed the stridulating organ, is described in the publications by Petter and Petter-Rousseaux (1963), Gould (1965), and Gould and Eisenberg (1966). These quills are enlarged and do not possess barbs and are less easily detached than the other quills on the body (see p. 102).
The animal has a boldly marked color pattern consisting of three main longitudinal stripes contrasting strongly with the black dorsal color. In the case of Hemicentetes semispinosus, the stripes and crown as well as a median stripe on the forehead are yellow. In the case of Hemicentetes nigriceps, there is no median central stripe on the forehead while the crown and body stripes are white. The venter of both species is almost free from spines. The claws on the forefeet are rather stout and some modification for burrowing is indicated (see Figures 2, 48 and 49). As was noted with the other genera, in addition to spines there is a soft underfur far more prominent in H. nigriceps than in H. semispinosus. In addition to fur interspersed with spines, long sensory hairs are distributed over the dorsum. The standard sensory hairs of the facial region are prominent including the mystacial, genal, super-orbital, and mental vibrissae.
The two species are allopatric and indicate slightly different environmental adaptations. H. nigriceps occurs from the vicinity of Manandroy south to Fianarantsoa. It appears to be confined to the central plateau edge. As one descends from the plateau into the true rainforest of the eastern escarpment, one finds Hemicentetes semispinosus. This form has been recorded from Ivohibe to Maroantsetra. We can conclude that H. semispinosus is confined to the rainforest areas below the high plateau extending its distribution to the northern rainforests and south to an undetermined locality below Ivohibe. For convenience we propose to discuss the characteristics and ecology of these species under separate headings.
H. NIGRICEPS—GENERAL ECOLOGY
Physical Description
This plateau-dwelling species is quite similar to the rainforest species with the exception of the difference in color and the fact that the pelage tends to be less spinescent. The underfur is dense and soft and the quills protrude from it so that, overall, the animal has a more woolly appearance. The quills comprising the stridulating organ are less in number on the average than is the case for H. semispinosus (Figure 50). The number of stridulating quills from a sample of 73 H. nigriceps ranges from a low of seven to a maximum of 17; the modal value was 11 quills. There is a difference when age classes are compared since younger animals tend to have more quills than older ones. Although the stridulating quills are replaced if lost, nevertheless, there seems to be a correlation between increasing age and a smaller number of quills (Figure 51).
There are no conspicuous differences between the sexes and the age classes may be defined on the basis of total length. The infant age class ranges from 58.5 to 60 mm for a minimum and 100 to 110 mm in total length at approximately 4 weeks of age. Females may breed at approximately 30 days of age when their total length lies somewhere between 120 and 130 mm. Maximum total length for adult specimens in the field was 180 mm. Field weights for adults range from 80 to 150 grams.
Distribution and Habitat
Hemicentetes nigriceps was taken in the plateau area between Manandroy and Fianarantsoa. Its habitat may be characterized as the plateau edge in the vicinity of the transition between eastern rainforest and plateau savannah. Burrows with animals may be found in the vicinity of brush and are generally never far from free water. The cultivation of rice has apparently opened suitable habitats for these animals and they are often found in the vicinity of paddy fields. When the animals utilize cultivated areas for foraging, they may adopt rather atypical habitats for burrow sites; indeed, in one area (i.e., Manandroy) almost all burrow sites were placed within an introduced eucalyptus forest but their foraging took place in the vicinity of the paddy fields and areas of cultivated manioc (see Figure 52).
Burrows and Microhabitat
From an examination of over 60 burrows, the following generalization can be made. Let us consider a typical case at Manandroy in the eucalyptus forest: Although located in a second growth habitat of eucalyptus, tree ferns and other primary elements are still to be found on the fringes of this introduced forest. The ground was covered with rotten logs, branches, and leaves, forming a mat of damp decaying vegetation. The following characteristics of the burrow were noted. It was located in the vicinity of a cultivated field under a rotten log. There was only one entrance to the burrow. The burrow was shallow, approximately 75 mm at the deepest point and around 450 mm long. The burrow entrance was plugged with leaves. Fresh feces were also noted there (Figure 53).
Coleopteran larvae, potential food items, were found in the rotten logs under which the animals tunneled. In addition, worms were found outside the eucalyptus mat in the red-brown soil near the paddy fields. The burrow provided a rather stable microhabitat for the tenrec At Manandroy, burrow temperatures were measured for 21 locations. The ambient range was from 20.4° to 30.5° C. Burrow temperatures themselves ranged from 20.5° to 26.5° C. at a distance of approximately 150 to 200 mm from the burrow entrance.
Activity and Thermoregulation
During the course of our field studies, observations were made from 0800 hours through 2300 hours in areas where Hemicentetes nigriceps occurred. Thirteen sightings were made between the hours of 1830 and 2300 hours. One additional animal was trapped sometime between the hours of 0200 and 0600. The field evidence would indicate that in this particular area H. nigriceps is almost entirely nocturnal with the peak activity in the early part of the evening. These observations were amply confirmed in captivity; however, under captive conditions the animals could be conditioned to feed during the daylight hours.
Studies of diel variations in thermoregulation were made in Madagascar both in capitivity and in the field. Between the hours of 0900 and 1000, a captive group of five H. nigriceps showed a range of cloacal temperatures from 25.0° to 27.4° C. with an average of 26.3° C. over an ambient range from 23° to 24.2° C. A second series of morning readings was made over an ambient range from 16.8° to 20° C. H. nigriceps showed a cloacal temperature range from 20.8° to 26.4° C. with an average of 23.6° C. All of these measurements were made during the month of February when the animals are not torpid for long periods. A series of late afternoon temperature readings were made between 1500 and 1800 hours, when the cloacal temperature should be showing an increase. Over an ambient range from 21.3° to 24.3° C, a sample of 14 H. nigriceps showed a range from 26.0° to 31.5° C. with an average of 28.6° C. To compare with this series, several field measurements of body temperatures were made during the early part of April preceding the entry of the population into their seasonal torpid period. Over an ambient range from 22° to 26.8° C, a total of 12 individuals showed a range from 25° to 35° C. with an average of 30.3° C.
Our data would indicate that the animal shows a diel fluctuation in body temperture with the cloacal temperature rising in the late afternoon preceding activity. The burrow undoubtedly serves to aid in conservation of heat loss and to ameliorate the more drastic changes in the ambient temperature.
The Annual Cycle and Reproduction
Annual variations in weight and activity were studied in the field and in the laboratory. From a period of 31 January to 3 February 1966, weights were determined for adult animals in the field. Eleven adults ranged in weight from 90 to 150 grams. Four H. nigriceps were returned to the laboratories at the National Zoological Park for further study. During the month of April our captive group of four attained weights from 135 to 187 grams. From the middle of April until July, this group ceased to feed, declined in weight, and did not significantly thermoregulate with the cloacal temperature remaining near the ambient. Minimum weights of 90 to 145 grams were achieved at the end of this period. Beginning in late July and early August, the animals increased their activity, began to feed, and new maximum weights were achieved in December of 1966, exceeding the previous maximum.
With no special manipulation of food or ambient conditions in the laboratories, the captive group began to show weight declines as early as December 1966 for one individual and as late as March 1967 for two others. These declines persisted until July of 1967 when the animals began to arouse, thermoregulate, and feed again. These captive data support previously recorded field observations. In the vicinity of Manandroy and Fianarantsoa, the population of H. nigriceps is generally torpid during June and July.
It would appear that the seasonal torpidity shown by H. nigriceps is in part under endogenous control, since the rhythm persisted through two seasons in captivity with no special attempt to manipulate environmental conditions. Thus, two types of thermoregulation are exhibited. During the breeding season extending from September on into February, the animals show a diel rhythm in thermoregulation but arouse during each 24-hour period to feed. In early May the animals, having reached their full weight, enter a profound torpor which parallels the austral winter.
Population of Hemicentetes nigriceps were sampled in February of 1966 and 1967. The gestation for nigriceps is similar to that displayed by H. semispinosus which is about 58 days. H. nigriceps gestation exceeds 55 days and is less than 58 days; thus it seems reasonable to assume that it is approximately the same for the two species. The litter size for H. nigriceps in wild-caught individuals excavated from their burrows during the month of February 1966 ranged from one to three with an average of 1.4 for a sample of 14 litters. During the same season in 1967, the litter size ranged from one to five with an average of three for a total of six litters. Litter size determined from captive born animals for a sample of five ranged from two to four with an average of 2.8.
The age classes for Hemicentetes nigriceps have been determined from extensive captive studies. For the purposes of our discussion, we have referred to animals in the 60 to 90 mm total length size class as infants. Juveniles include the group from 100 to 130 mm total length and adults are all those animals exceeding 130 mm total length. Although the breeding for our population of H. nigriceps was not completely synchronized, the evidence strongly suggests that breeding begins in early September and most of the first born females in the population probably become pregnant from the period of December to early January. We infer this from a knowledge of the growth curve and of the gestation period.
For example, from 28 January to 7 February 1966, collections of individuals from Manandroy were marked and released. Of the 56 individuals measured, 10 fell in the infant class, 15 were juveniles, and 31 were adults. A similar sample of 33 taken at Manandroy from 6 to 7 April 1966 indicated 8 individuals in the juvenile age class and the remaining 25 as adults. In the following year from 4 to 5 February 1967, the Manandroy population was again sampled and a total of 48 individuals were measured. Of these, 9 were infants, 9 were juveniles, and the remaining 30 adults. This would appear to substantiate our contention that some breeding takes place in December; however, a consideration of the animals in the 130 to 150 mm age class (on the basis of pelage color and wear) suggests that some births may take place in early November. Thus, there must be a first breeding period in early September or the individuals in this size range have wintered over from the previous year having been bom late in April of the preceding season. The latter hypothesis is doubtful in view of our knowledge of the growth curve. As indicated previously the Manandroy population was always marked and released. Collections for captive studies were made at Alakamisy Ambohimaha. We know from our captive records that animals may live to an age exceeding 2 years and 6 months. We know further from our sampling in 1966 and 1967 that marked individuals may survive to a minimum age of 15 months. Since females become sexually mature within 4 to 5 weeks of their birth, it is entirely possible that a given female will reproduce in two consecutive breeding seasons.
Feeding and Food Intake
In captivity, Hemicentetes nigriceps readily accepted earthworms. To a lesser extent, H. nigriceps would kill and eat the larger coleopteran larvae found in the rotten logs under which it nested. Orthopterans were not taken. H. nigriceps could be induced to take some raw chopped meat in captivity, but the animals could be maintained in good condition only if considerable numbers of worms were fed. Two H. nigriceps were collected at night while feeding. Their stomach contents consisted almost exclusively of earthworms although an arachnid abdomen was identified. Some earth, evidently ingested with the worms, was present in both stomachs. Since earthworms appear to account for a large portion of the animals' diet, it seems desirable to review some of the characteristics of earthworm ecology and behavior.
EARTHWORM BEHAVIOR AND ECOLOGY
Lumbricus terrestris shows a minimum of activity between 0700 and 1100 hours. High rates of activity are shown before 0700 and after 1100 hours. Peak activity is shown in the early mid-afternoon lasting through early morning (Laverach, 1963). Favorable conditions for earthworms in the tropics include an undisturbed soil which has a regular and adequate water supply. Generally a fine soil texture is required but this is a concomitant of the physical availability of water since water will rise to a greater height in fine soils. In addition to these soil requirements, earthworms require a regular and adequate supply of organic material. Even though the layer of humus in tropical soils is very thin, earthworms can and do occur there and may be found even in the red sandy soils in the vicinity of paddy fields. Indeed, light and medium loams appear to carry higher proportions of worms than the heavier clay type soils or more gravelly sand and alluvial soil types (Guild, 1948). Local abundance of earthworms may be quite high. Biomass of earthworms may range from 6 kilograms per acre in a maize field to 2,339 Kg/acre in an artificial forest (El-Duweini and Ghabbour, 1965).
We sampled the earthworm population in the soils composing the floor of the eucalyptus forest at Manandroy and found earthworms to be practically absent. On the other hand, earthworms did occur in the vicinity of the paddy fields and other cultivated fields. This has led us to conclude that the eucalyptus forest served as a sleeping area rather than as a primary foraging area and furthermore the animals would need to move out each night into the vicinity of the fields in order to provide themselves with sufficient earthworms.
FOOD CONSUMPTION
Consumption of earthworms by captive individuals was studied in some detail. It was found that the average intake during a 15- to 20-minute feeding period to satiation for three specimens of Hemicentetes nigriceps ranged from 13.0 to 4.2 grams. The maximum intake was by an adult and the minimum by an infant. Yet the animals feed several times a day, and in order to determine average food intake for a 24-hour period, a sample of six Hemicentetes nigriceps were fed and weighed four times in each 24-hour period. This procedure was carried out for four days. The average increase in weight plus average consumption of worms for a 24-hour period could be calculated. Food intake for adults and juveniles averaged approximately 100 grams of earthworms in each 24-hour period, that is, the animal ingests approximately its own weight in worms daily. A young juvenile would tend to exhibit approximately 1.7 grams net gain in weight for each 24-hour period. This would indicate approximately 4.1 percent of the wet weight food intake was converted to net gain in weight for the animal feeding (see Figure 54).
The nutritional value of earthworms is rather high. The average dry weight of earthworms is about 15 to 20 percent of the wet weight. Protein accounts for the largest fraction of the dry weight and has been estimated between 53.5 and 71.5 percent of the total dry weight of Lumbricus terrestris. Measurements of L. rubellus and Eisenia rosea have a similar proportion of dry weight (16.38% which consists of 16.3% protein, 17% carbohydrate, 4.5% fat, with only 15% ash residue) (Laverach, 1963). The nutritional quality of earthworms is certainly adequate for rapid growth. In captivity, our H. nigriceps grew rapidly and, indeed, some question may be raised concerning the validity of our captive growth studies since we may have been feeding a diet with nutritional qualities far exceeding those in the wild. For example, total length could increase in captivity over a 30-day period by an increment of approximately 40 mm.
As we have previously stated, we had established a marked population of animals at Manandroy which were sampled in February and again in April 1966. By recaptures, we could estimate growth in the field and found that over a 2-month period a juvenile in the 104 mm age class could increase in total length some 31 to 36 mm. Larger animals first captured in the 141 to 148 mm total lengm class would increase in total length over the same 60-day interval by approximately 14 to 15 mm. This indicates that although we may have slightly accelerated growth in captivity, we certainly did not distort too far from the field situation since increase in total length is certainly very rapid, even under field foraging conditions.
H. SEMISPINOSUS—GENERAL ECOLOGY
Physical Description
As noted previously, H. semispinosus differs from H. nigriceps by the possession of a more spinescent pelage. Even the ventrum is sparsely covered with spines and it is only on the ventrum that true hairs persist in the adult. An underfur of nonspinescent hairs is present in juveniles on the dorsum but virtually disappears with continued growth. The stridulating quills are more numerous in H. semispinosus. A sample of 13 gave a range of 14 to 18 in number with a mode of 14 (see Figure 51). Older animals may show some wear with respect to the stridulating quills and even a loss of the same.
There are no conspicuous differences between the sexes; males may be identified by manual expression of the phallus from the cloaca. The teats are very conspicuous in both males and females being surrounded by a darkly pigmented area. Age classes may be arbitrarily defined on the basis of total length. Infants range in size at birth from 60 to 66 mm total length. A juvenile is defined as an animal able to forage about and in the stage of being weaned. At this time it is approximately 90 mm in total length. A female can conceive when approximately 120 to 130 mm in total length; hence, the adult age class has been defined as any total length exceeding 125 mm. Linear growth begins to taper off at approximately 3 months of age or at a total length of 140 to 150 mm. The maximum size of a wild caught adult male was 172 mm in total length. Adult weight shortly after capture and in good condition ranged from 125 to a maximum of 280 grams. Adult animals during torpor may fall to a weight of 70 to 90 grams.
Distribution and Habitat
Hemicentetes semispinosus characteristically inhibits rainforest areas. Its denning sites may be either in primary or second growth forest. Den sites were examined at Ambitolah, Ranomafana, and Perinet. These sites may be near cultivated fields or rice paddies. On the other hand, in more mature forests, the burrow is generally located near a stream or other body of water (see Figure 55). Feeding appears to be executed in natural clearings such as areas where a landslide has been overgrown with forbs and grasses or, in one case, at the site of an abandoned village.
Burrows and Microhabitat
It is convenient to consider two types of burrows: (1) a burrow inhabited by a single individual or small family unit and (2) an extended burrow system inhabited by a large family group or colony. The following description is of a burrow uncovered at Ranomafana on 27 January 1966: A family of seven animals including an adult male, an adult female, and five young was removed from the nest. The burrow was characterized by a single entrance to a tunnel approximately 18 inches long. The burrow was very shallow averaging 2 to 3 inches in depth. It was located above a stream on a rather steep slope of approximately 45°. The burrow contour followed in part a tree root. The forest floor in the immediate vicinity of the burrow was covered with leaf litter and the following invertebrate types were noted in the soil: Millipedes, Nasutotermes, a large flightless cricket, ants, and the shell of a land snail. The soil included a thin layer of humus overlying a reddish earth.
An example of a burrow system inhabited by a colony of some 19 individuals was as follows: The burrows were located on a south facing, steep bank with naturally formed terraces of humus and leaves (see Figure 56). At the foot of the bank was a fairly level area of moist humus covered with grass. The burrow complex included three subcomponents, the first being a burrow running some 3 meters in length following the contours of the bank and ending in a chamber approximately 2.25 meters from the entrance. An extended tunnel system proceeded beyond the nest chamber for approximately 71 centimeters. Not directly connected with the major burrow was a second tunnel about 40 centimeters away. The tunnel was some 69 centimeters in length ending blindly in a nest chamber. Some 19 meters from this second burrow was a burrow system approximately 1.8 meters long ending blindly in a nest chamber. Although these were not interconnected, individual animals were seen to move from one tunnel to another, and we may conclude that this colony involved considerable interchange of the members actually found in the three component burrows. A total of two adult males, two adult females, and 14 juveniles were removed from this colony.
To summarize, a total of five burrow sites in the vicinity of Ranomafana exhibited the following characteristics: The burrow was generally located on a bank above a stream or damp area. Although a burrow may range from 1 to 6 feet in length, it generally has one entrance with a nest formed of leaves lying some 1 to 2 feet beyond the entrance. The burrow tends to follow the contour of the slope upon which it is located or it may run along a root or under a stump. If rocks are present, the burrow may be quite intricate and pass between the larger stones. An entrance of an active burrow is generally plugged with leaves. Feces may be deposited near the entrance. The following commensals have been found within the burrow system: land snail, millipede, spittle bugs, and a small frog. The potential food supply in the immediate vicinity includes wingless crickets, millipedes, earthworms, and perhaps land leeches.
Activity and Thermoregulation
At Perinet a family group consisting of a female and five babies was installed in a large observation enclosure (see Figure 4). A photocell device with an operation recorder was installed at the entrance to the nest (see Appendix F), and for 48 hours the activity of the family group was monitored by counting the number of departures and returns to the nest. The following activity pattern was discernible: From 0900 to 1300, there is sporadic activity with a major peak directly after 1200 hours. (In Herter's publication (1963b), he portrays an activity diagram by one of three subjects which also showed a peak of activity around 1200 hours.) Sporadic activity began again between the hours of 1700 and 2100 with the highest peak of activity recorded during a given 24-hour period falling between 1800 and 1900 hours. Throughout the night other peaks were noted between 2200 and 2300 hours, at 0300 hours, and an extended peak from 0400 to 0600 hours.
Field observations which included two nights at a burrow near Ranomafana yielded eight sightings of tenrecs. Five sightings were made between the hours of 1300 and 1400 and three sightings between 1800 and 2015 hours. Our field observations coupled with our seminatural activity recording suggest that Hemicentetes semispinosus does feed during the day as well as in the evening and early morning hours. This activity pattern conforms fairly well to the known activity patterns of eardiworms and, since earthworms are probably its major food resource, this may be an overall adaptation to their prey. Although H. nigriceps does not show such a diurnal activity peak, this may be related to the more exposed nature of its habitat on the high plateau which renders such diurnal foraging more dangerous. As with H. nigriceps, H. semispinosus can be trained in captivity to feed at various times during the day.
Herter (1962a) in his Figure 9 showed a double peak in body temperature for a captive H. semispinosus with a high of 30° C at noon falling to 26.8° G at 1800 hours, only to rise to almost 30° C between 2400 and 0200 hours followed by a decline to 28.2° C at 0800 hours. Although his activity data for this individual did not parallel the body temperature fluctuation, this thermoregulation pattern parallels our field activity data.
The Annual Cycle and Reproduction
Some controversy exists concerning the seasonal variation in activity and the seasonal torpor of H. semispinosus (see Gould and Eisenberg, 1966). Unfortunately, we studied the thermoregulation of H. semispinosus very little. In our captive group, held at Madagascar during 1966, the cloacal temperatures of eight animals were measured during February yielding a range from 28.5° to 31.2° C with an average of 29.7° C. These measurements were made in the morning during the theoretical low of the diel activity cycle. During April a sample of three taken at 1515 to 1530 hours ranged from 29.5° to 31.5° C at an ambient of 22° C.
In general, during the austral spring and summer, the animals maintain a reasonably stable temperature but in captivity a tendency for seasonal torpor is shown. Torpor was exhibited in the colony at the National Zoological Park during their first captive season in 1966. From late April to late July, four individuals from a sample of six showed a decreased tendency to feed and a loss in weight; however, this trend was not synchronized since two other individuals during the same period showed weight gains. During 1967 the situation was even less synchronized. Therefore, we conclude that, although individual Hemicentetes semispinosus can and do exhibit torpor, their tendency to pass into torpor is in part a function of the state of their fat reserves as well as ambient conditions. When animals become torpid, their body temperature falls very close to the ambient, there is decreased activity, and virtually no feeding.
For example, during the first season of captivity while in a period of torpor, one male passed from a 160 gram high in April to 130 grams in late July. An early July reading of cloacal temperatures showed a range from 22.1° to 33.8° C for our captive group of six over an ambient range of 21.0° to 27.5° C. Torpid animals invariably showed a cloacal temperature less than 1° C above that of the ambient. The evidence suggests that, depending upon local food abundance and temperature condition, H. semispinosus shows a facultative ability to exhibit torpor.
Gestation for H. semispinosus is 58 days. The litter size is variable. At Ambitolah three families were collected showing a litter size range from five to eight with an average of 6.3. In the vicinity of Perinet (Anevoka), four families were collected showing a litter size range from five to eight with an average of 6.2. Litters bom in captivity in our Madagascar colony showed a litter size range from seven to eleven with an average of 8.8 (sample size of four families). Age classes were defined as for H. nigriceps; 60 to 90 mm total length, infant; 100 to 130 mm total length, juvenile; 130 mm and up, adults. Females can conreive at an age of approximately 35 days. Based on this knowledge, we can make some generalizations concerning reproduction at Perinet.
During 1967 in the months of February and March, a total of 42 Hemicentetes semispinosus were collected of which 18 were infants, 13 juveniles, and 11 adults. These data suggest that the population at Perinet initiates breeding in October and that conceptions occur through December. Reproductive activity is depressed in Perinet during the month of July; however, local populations at lower elevations (e.g., Rogez) may be active and breeding (see Gould and Eisenberg, 1966). Again, we are forced to conclude that the timing of reproduction in H. semispinosus is a function of local conditions including temperature and abundance of prey.
Feeding and Food Intake
Feeding tests in captivity indicated that Hemicentetes semispinosus readily takes worms, coleopteran larvae, and some ground meat. It would appear that earthworms are its primary prey in the wild. We have already considered the activity and local abundance of earthworms in the section for H. nigriceps and these points will be utilized in the following discussion for H. semispinosus.
In the vicinity of Perinet, earth samples at four localities were taken in order to determine the local abundance of earthworms. The earth sample sizes were 2,500 square centimeters in area and 20 centimeters in depth. H. semispinosus does not forage much more than a centimeter below the surface but we took a deep sample of earth in order to obtain the worms which may be available in a reasonable volume of soil since we assumed that most of the worms in such a soil sample would move to the surface for feeding at least some time during their 24-hour activity period. Four such samples yielded 21.4 grams of earthworms. We may extrapolate then and assume that, if our sample is fair, this area could have 83.8 kilograms of earthworms per acre. This is well within the published density range for earthworms (El-Duweini and Ghabbour, 1965).
Food intake for captive H. semispinosus was similar to those results cited for H. nigriceps (Figure 54). One juvenile was observed over a period of three days and consumed an average of 104.4 grams of earthworms per day. An adult sampled over a similar period averaged 85.5 grams of earthworms per 24-hour period. This is in reasonable agreement with the results that Herter published (1936b) where his one subject averaged 99.4 grams of food over a 24-hour period. If we assume that the previous estimate of worm density is nearly correct and that our captive food consumption approximates the normal food consumption in the field, then we can assume that a H. semispinosus would consume up to 1 kilogram of earthworms in a 10-day period. Given these assumptions, if the earthworm population in an area were not replacing itself, a prime area of one acre would support 10 animals for 83 days. This is a reasonable estimate, since a family group is somewhat in the neighborhood of seven to ten animals, and it would appear that they range over an area of approximately an acre.
Growth in captivity for H. semispinosus was similar to that described for H. nigriceps and is in reasonable agreement with the growth curve published by Herter (1963b; see also pp. 97–99 on Ontogeny).
The Ethology of Hemicentetes
Because the behavior patterns of Hemicentetes semispinosus and H. nigriceps are so similar in their component parts, the general description of the behavior patterns for both species will be undertaken as a single unit. Some information concerning the behavior patterns of H. semispinosus has already been published by Herter (1963b) and by Gould and Eisenberg (1966).
GENERAL MAINTENANCE BEHAVIOR
Locomotion.—Hemicentetes locomotes by employing the diagonal coordination pattern of limb movement. In running, the heel may be lifted off the ground, but in general the animal is plantigrade. The animal seems to have little ability to jump but during the offensive and defensive behavior patterns it is able to buck and pivot on its hind legs giving the appearance of a slight hopping movement. At times a short bounce as a concomitant of pivoting or bucking may occur but this movement is in no way comparable to a springing jump as defined for quadrupedally ricochetting mammals.
The animal is capable of climbing by employing a crossed extension pattern but there is little ability to grasp with the individual digits. It certainly does not climb often and could in no way be considered as arboreal as Setifer and Echinops.
Swimming, employing a diagonal limb coordination pattern, has been described by Herter (1963b) and we have observed an animal, startled in the vicinity of a paddy field, swim across a small body of water. The animal swims rapidly with the nose and head held above the surface.
Exploration and utilization of living space.—When placed in a novel environment, the animal will generally show some hesitancy and begin to move about slowly, testing the air by lifting its head and wriggling its mobile nose. The animal will then proceed to move forward in an elongate posture pausing to sniff the substrate and test the air. The role of “tongue clicks” as a possible means of echolocating has been experimentally verified by Gould (1965).
In a novel situation, the slightest change in background stimulation will cause the animal to show spinal erection; especially prominent in the nuchal crest erection (see p. 93). The nose is utilized to probe in the substrate; apparently earthworms are recognized by tactile and olfactory stimuli. In the large observation pens, the animals have been observed to explore extensively before settling down to a set pattern of space usage. Generally a defecation point is selected in the environment which is used over and over again. A burrow site is selected and subsequently a nest is built of leaves and grass.
Foraging activity appears to be a function of the proximate environmental conditions. In general the greatest foraging activity could be elicited in areas of the pens where the earth was soft, moist, and to some extent shaded. Sunlit areas were not necessarily avoided if sufficient leaves lay on the ground to permit some shade for small invertebrates.
Our observations on both H. semispinosus and H. nigriceps indicate some shifting of nest sites. In our capture-mark-and-release program, burrows were opened and, on subsequent days of sampling in the same area, animals could be found that had moved. Of course, their movement could be attributed to the fact that the burrows had been disturbed. Nevertheless, in those cases where burrows of H. semispinosus were not disturbed and the animals were observed by inspection of the tunnel from the outside, we found some shifting at the end of a 10 to 15 day period. Turning to H. nigriceps, we have found with our marked population at Manandroy that, with a year interval and given six recaptures of animals marked during the previous year, den site shifting ranged from 50 to 800 feet. It would appear that an individual nigriceps will settle in an area and utilize den sites in an opportunistic fashion. Breeding females set up a deeper more permanent burrow. Neighbors may move in and nest with each other under certain conditions.
Within any given foraging area, den sites appear to be limited. For H. nigriceps the principal denning sites were in the eucalyptus forests bordering the cultivated fields. Within the eucalyptus forests definite preferred loci for denning could be discerned. These were generally areas containing a great number of fallen logs which provided a ready shelter under which the burrows were constructed (see Figures 53 and 58). As noted earlier for H. semispinosus (p. 85), a given denning area may be utilized for some period of time and an actual colony may be formed. Although a denning area may be established and a small colony set up, within the denning area actual burrow sites may be utilized off and on depending on the density of ectoparasites in the nest and other physical factors.
Rest and sleep.—Hemicentetes will sleep by curling into a ball and tucking its head ventrad beneath the abdomen resting most of its weight on the head and hind legs. On the other hand, the animal may rest lying on its side and lactating females habitually lie in this position when the young are suckling. The animals have also been observed to sleep on their backs but, when torpid, they generally curl in a semicircle lying on one side within the nest.
Marking.—Urine and feces may be deposited at the burrow entrance. An animal does not venture very far from the burrow when defecating during the day, and at this time the feces may be deposited within 1 or 2 inches of the entrance itself.
Hemicentetes nigriceps has been noted, during encounters with conspecifics, to drag its perineal region on the substrate. Once an animal was noted to rub its ventrum by extending and flexing the body in the substrate. Hemicentetes semispinosus has been noted to extend and flex its body while lying on its side in the soil—this action is termed a “side rub.” The perineal drag is also exhibited during an encounter situation or when exploring a novel environment. All of these scent depositing movements together with the locus specific urination and defecation are of potential significance in chemical communication.
Care of the body surface and comfort movements.— In addition to the patterns of yawning, stretching, and shaking, Hemicentetes exhibits scratching as a primary means of dressing its pelage and spines. The hind foot may reach the head and a good portion of the anterior part of the dorsum and ventrum. After a bout of scratching, the animal will often turn and nibble at its toenails. The face-wash exhibited by Setifer, Echinops, and the Microgales, is present only in an abbreviated form in Hemicentetes. One forefoot may occasionally be used to wipe the side of the face, but the animal never sits upright or exhibits a stereotyped face-washing pattern employing both forepaws simultaneously. The teeth and tongue are utilized as a cleaning organ by licking parts of the ventrum and the cloacal region.
Urination and defecation.—As noted previously, Hemicentetes tends to defecate near the burrow entrance. Defecation generally involves a freezing movement while partially extending the hind legs. At the conclusion of defecation, the animal frequently kicks back. Hemicentetes nigriceps will, on occasion, bury its feces with movements similar to those described for Tenrec ecaudatus. This pattern consists of moving somewhere away from the burrow entrance to a preferred locus, probing in the substrate, and digging with the forepaws to create a small depression. After turning around and orientating its posterior over the hole, it defecates and kicks back, partially covering the feces with earth. Such a ritualized kicking back and covering movement has not been noted for H. semispinosus although all the basic behavioral elements are present for its complete expression.
Nest building and burrowing.—Hemicentetes is a ramer good burrower possessing well-developed claws on its broad forepaws. Burrowing movements typically consist of selecting a site, generally under a log or near a rock, and commencing to dig at some intersection between the log and the substrate. Alternate movements of the forepaws are employed and the accumulated earth is kicked back with the hind feet. Simple burrows are constructed and, when a tunnel about a foot long has been excavated, the animal begins to transport nesting material. Leaves and blades of grass are seized in the mouth and carried to the nest where they are deposited. If blades of grass are being selected, the animal may grip them quite tightly with its teeth and shake its head from side to side. It is customary when Hemicentetes enters its burrow, at the conclusion of a foraging bout, to re-emerge and seize leaves lying in the vicinity of the entrance, pull them in, and deposit them to one side of its body. The net result is that the burrow entrance is effectively plugged when the animal retires to the inside.
Foraging and prey capture.—Since the principal food of Hemicentetes is earthworms, we will attempt to describe the location and capture of worms for both species. As indicated in the section concerning utilization of the living space, the animals tend to be selective in the areas in which they exhibit foraging behavior. The extent to which odor is involved in the detection of earthworms is difficult to determine; however, soil that has been impregnated with earthworm scent generally stimulates activity at that spot.
The animal, when foraging on a grassy or leaf-littered ground cover, will insert its nose at the roots of the grasses or under leaves. Upon detecting a worm, it will hesitate and then attempt to bite at the worm with its mouth while simultaneously loosening the earth to either side of the worm by scratching movements. These may be directed backwards or slightly to one side. Once the worm has been seized, it is frequently shaken and seized again. The animal exerts a strong pull while bracing its “forefeet and pulling back with its neck muscles. It will then proceed to relax the tension while shaking its head from side to side, and then initiate a new bite while stroking downward on either side of the worm with its forepaws in a very characteristic patting motion. It is interesting to note that this pattern of seizure with the mouth, shake of the head, and patting with the forepaws is displayed toward other foodstuffs such as raw meat even when these movements are no longer necessary to aid in prey capture. The sequence is quite stereotyped and is displayed whenever initially seizing any foodstuff, regardless of its nature. (See also Solenodon Eisenberg and Gould, 1966).
Some observations on the capture and consumption of coleopteran larvae are useful at this point since these large grubs have pincers and can afford some discomfort to Hemicentetes when they are being eaten. Since the grubs are quite large, juveniles have special difficulty in killing them. For example, a juvenile sniffed and attempted to pick up a grub in its mouth up to nine times. During each attempt it was unable to effect a grip and finally left the grub and walked off. As another example, an adult approached a grub, bit it, and immediately exhibited the head shake and stamping movement thereby succeeding in tearing the grub open. It then commenced to chew, and consumed the grub completely except for the head and pincers. Some adult Hemicentetes would not eat grubs. This could in part be attributed either to a lack of experience and an avoidance of a novel prey object or to the possibility that the animal had been pinched before and was deliberately avoiding the grub.
Offensive and defensive behavior.—When startled in the field, Hemicentetes generally will exhibit spinal and crest erection followed by either flight or offensive behavior patterns. The animals are capable of running quite rapidly and during our speed tests we clocked six individuals of semispinosus over a range of 1.6 to 2.7 feet per second. One individual H. nigriceps achieved a speed of 1.5 feet per second. As will be discussed later, any marked change in the background environment induces offensive or defensive behavior on the part of the animal, but the odor of predators such as Galidia elegans or Fossa fossa is especially effective in inducing an offensive reaction.
Defensive behavior consists of erecting the quills especially those on the head (nuchal crest). The degree of crest erection is in part a function of the degree of stimulus contrast7 and may range from partial erection to full forward where the spines of the head are directed anteriorly forming a circlet around the head of the animal (see Figure 59). During defensive responses the animal utters typical vocalizations: Mild arousal is accompanied by the putt-putt sound; stronger arousal involves the crunch sound plus bucking. Bucking consists of contracting the epaxial muscle plus extending the forelimbs causing the head to lift. Alternate contraction of the neck muscles causes the head to bob up either independently or in conjunction with the foreleg movements. This movement effectively serves to drive crest spines into the nose or paws of a predator. The animal is quite capable of pivoting on its hind legs and will orientate immediately toward any disturbing stimulus and continue bucking. The barbed detachable quills make this defense quite effective.
If aroused by predator odor, the animal will not only erect quills and buck but will also run towards any disturbing object bucking all the while. When disturbed in the nest, a nontorpid animal generally responds with the full-blown defense pattern and, in the case of colonies such as those formed by Hemicentetes semispinosus, all members will be aroused and give a concerted attack on any intruder in their nest or tunnel system.
Hemicentetes rarely bites but may do so when disturbed in the nest. The extreme reduction in tooth size makes biting a less adequate defense than defense involving spines.
Other offensive and defensive patterns are displayed toward conspecifics and are discussed on pages 94–95 under Social Behavior.
SOCIAL BEHAVIOR
Communication.—Auditory communication is prominent in Hemicentetes. Nonvocal communication includes the general sounds which accompany the animal's foraging activity including the stamping of the feet as an earthworm is pulled from the ground and the chewing sounds. In addition sounds are produced when the animal stridulates (Gould, 1965; see also p. 102 ff). During offensive and defensive patterns of behavior, characteristic vocalizations are produced.
The “crunch” sound which is homologous to the sound produced by Setifer and Echinops generally accompanies nuchal crest erection, body spine erection, head bucking, and/or body bucking. It may be produced with half to full crest erection prior to a startle response leading to flight. The “putt-putt” sound is generally displayed when the animal shows full-body spine erection. The nuchal crest need not be fully erected. This sound is generally produced when the animal is disturbed in the nest box or during the initial phase of an encounter. A sharp inhalation and exhalation may be produced when the animal is exhibiting bucking. A “piff” sound similar in tonal quality to that produced by Tenrec is shown but the circumstances cannot be specified. There is a graded series of squeaks including a grunt sound, a squeak, and a twitter which are produced during social contact. These are the only sounds that show a true harmonic structure. During courtship as the male approaches the female, he may lift his nose and bend the terminal tip dorsad as he exhibits a nose to ear. At this time he may utter a prolonged “hiss” which is strongly associated with courtship (see Table 8).
Chemical communication is implicated by the marking movements and the locus-specific deposition of feres and urine.
Tactile communication is involved in all contact-promoting behaviors such as naso-anal, naso-nasal, nose to ear, nose to body, crawling over, and nose to nape. During offensive and defensive behavior toward conspecifics, the animal may rump another, buck, or bite.
The encounter.—The behavior patterns employed during encounters with conspecifics are similar to those described for the other genera of Tenrecinae.
Offensive and defensive behavior toward conspecifics: (1) striking with the nose; that is, suddenly swinging the head to one side and pushing at a conspecific with the snout; (2) rumping; that is pivoting on the forelegs and pushing the rump into a partner; and (3) while standing quadrupedally with the heel off the substrate, darting the head at a partner while attempting to bite. The nuchal crest may be raised during this latter movement.
While fighting, males will attempt to bite one another in the flank or shoulder. If the body axes are oriented in opposite directions when they are standing side by side, they may mutually bite each other in the flank and roll over. In seeking to grasp one another they may bind together in a grapple and tumble about for some time. On rare occasions the agonistic behavior may involve raising the crest and actually attempting to buck and drive the crest quills into a partner.
Contact promoting behaviors include nose to anal or inguinal region, nose to nose contact, nose to ear, nose to the side of the body, nose to nape, and crawling over and crawling under. Sexual behavior includes mounting by the male and lordosis by the female. The mount is prolonged in Hemicentetes and may exceed 20 minutes in duration. The gape reaction is absent in Hemicentetes. This loss correlates with the decreased tendency to bite and the reduction in tooth size.
*Sound types similar for H. semispinosus and H. nigriceps.
The form of the encounter may be specified in part by a knowledge of the age and sex classes engaged in the interaction:
Male-male encounters are characterized by an initial contact-promoting session involving nose to anus, nose to ear, and nose to body; generally followed by moving away and/or offensive and defensive behaviors including bucking. Young males maturing in a family group may be tolerated by older males and no agonistic interaction will occur, but if two adult males who are unknown to each other are introduced in the presence of females, some severe fighting may ensue. In addition, groups that have been kept together through the torpid period during the austral winter will upon emerging often exhibit considerable agonistic behavior in the austral spring (see Gould and Eisenberg, 1966). As described previously, male fights may involve biting and locking and grappling.
Female-female encounters generally involve preliminary contact of nose to crest, nose to nose, nose to ear, nose to body, followed by moving away and avoidance.
The nature of the male-female encounters depends upon the female's sexual receptivity. In general the male will always initiate a great deal of contact promoting behavior followed by an attempt to mount. The most frequent contact-promoting behaviors are nose to nose and nose to body. Crawling under, and head-over head-under, have been exhibited only rarely. The females generally show less contact-promoting behavior but do engage in all of the patterns noted for the male with the exception of attempted mounts (Figure 60). If the females are unreceptive, they will move away. If the male persists in his mounting attempts, females will buck and show crest erection. Eventually the male will desist under these circumstances; however, moving away and bucking on the part of the female do not necessarily indicate that she will not receive the male, and the male may persist through two or three bouts of bucking and rebuttal from a female before he finally mounts her. The following protocols give some idea of the interaction sequence.
———
Example 1—Hemicentetes nigriceps: Male and female come together nose to nose. Male exhibits a nose to her crown or ear; female stands and then exhibits a nose to the side of the male and a nose to his crown. Male demonstrates lip curl and hisses in her ear. Female exhibits nose to side. Male exhibits nose to side, nose to crown, and nose to rump of female, and initiates following as the female moves away. He attempts to mount. The male will also drag his perineum and ventrum in the substrate between encounter attempts.
Example 2—Hemicentetes semispinosus: Male approaches; female stands. Male exhibits nose to ear followed by a nose to ear by the female. Male shows lip curl and hisses in her ear again. Female shows nose to ear of male. Break. Male approaches from the rear, exhibits nose to cloacal region of female; female stands. Male attempts mount and exhibits nose to crown. Male achieves mount and grasps the quills of her nape in his mouth.
———
Because H. semispinosus frequently dwells in colonies during the reproduction season, several tests were run by introducing adult males to females with half grown young. Two different males were introduced separately to a group composed of two females and eight juveniles. During a 44-minute encounter period, Male A showed the following patterns toward the adult females: Nose to crown, nose to nose, nose to cloaca, nose to ear with hissing, nose to body, attempted mounts, and finally two separate mounts. Mount durations were 10½ minutes with Female No. 9 followed by 21 minutes with Female No. 10. The mount included intromission by the male and thrusting. Adult females showed moderate contact responses to the male including nose to cloaca, nose to nose, and nose to ear. Initially the females showed moving away and some bouts of bucking but were quite docile once they had encountered the male at least twice. The second male (Male B), upon introduction, was unsuccessful in his attempts to mate with the adult females and confined most of his activities to the juveniles. Contact promoting behaviors were shown to juveniles and three times he attempted to mount juvenile animals. The juveniles reciprocated only with nose to nose contact and upon attempts to mount they bucked or squealed with an erect crest.
During Male A's mating with the adult females, the juveniles would congregate around the coupling pair in a circle or semicircle. The number of young concentrated around the mating pair varied throughout the 44-minute encounter time. During the first minute, there were two young. The number of young increased to a maximum of eight young after 11 minutes and then waned to an average of three young for the remaining 16 minutes of the encounter.
We may note the following characteristics of the mating pattern of Hemicentetes: Contact promoting behaviors are very similar to those described for the other genera of tenrecs. The neck grip confined to the nuchal crest area is still present in Hemicentetes males. The mount duration is protracted. Receptive females initially show a period of agonistic behavior toward the male followed shortly by quiescence. Figure 60 illustrates the mating ritual of Hemicentetes nigriceps and this may be compared with that portrayed for Microgale in Figure 29. The great similarities in the tactile and olfactory exchange configurations is immediately apparent.
PARENTAL CARE BEHAVIOR
Some days prior to parturition and on through the day of parturition, female Hemicentetes exhibit an increased tendency to build a nest. They assiduously collect leaves in the vicinity of the nest and transport them there in their mouths. Nest defense increases markedly during the initial period of rearing the young. “Putt-putt” sounds may be made upon initial disturbance. Further disturbance of the nest results in a crunch sound, full forward crest, bucking while stamping the feet, and even rushing. H. nigriceps will bite occasionally as will H. semispinosus; however, biting appears to be of much rarer occurrence in H. semispinosus. Female Hemicentetes lick and clean their young utilizing the tongue until the young are 10 days of age. Retrieval of young to the nest by the female persists until at least 10 days of age. From 10 to 15 days of age, the female will still exhibit attentive behavior to the young but by the time the young are 2 weeks old attentive behavior has begun to wane. Until the infants are 15 days of age, the female will extend toward a youngster, place her nose over it, and draw it under her body (see Gould and Eisenberg, 1966). The female will attempt to stand over the litter until they are approximately 15 days of age. Suckling begins to wane from 18 to 22 days.
ONTOGENY OF BEHAVIOR
Hemicentetes nigriceps.—The young of H. nigriceps mature very rapidly (Figure 61). Until approximately 14 days of age, they are very much oriented to the nest. At birth the young can both squeak and suckle and will attempt to locate the teat by switching the head from side to side. The “crunch” sound is generally shown at approximately 4 days of age. Neonates can crawl using a crossed extension pattern with some hind leg coordination but the ventrum drags until approximately 10 days when they begin to lift the ventrum from the substrate. Fully coordinated locomotion is well developed by 20 days of age. The righting response or the capacity to turn over when placed on their backs, is present by the second day. The meatus is closed until the 7th day of age. The eyes generally remain closed until the 8th day. Teeth are detectable at approximately 5 to 6 days of age. The animals will exhibit a bucking response at approximately 1 day of age. When displaced from the nest, a shiver reflex is demonstrated until approximately 9 days of age when it wanes. The nipples are visible in the young from approximately 3 days on. Hair and spines begin to show after 24 hours and develop together reaching a maximum development at approximately 20 days of age. The stridulating quills appear from 2½ to 5 days.
The young first emerge from the den at approximately 9 days of age only to return, never proceeding very far from the entrance. The young begin to accompany the mother from 12 days on and from the age of approximately 17 days they may be found foraging quite far from one another in the field. The scratch reflex begins to appear at 7 days but at this age they tumble to one side while scratching and only exhibit fully developed scratching from 15 days on. The young begin to feed on solid food at approximately 13 days of age and depend more heavily on solid food from about 17 days on. Spontaneous defecation is shown at 4 days of age. Young nigriceps will begin to dig a hole, defecate, and cover the feces at approximately 16 days of age.
From approximately 13 days to 22 days of age, the young will follow the mother loosely and cluster around her when she stops. Following begins to wane at about 20 to 30 days. First estrus is shown by females at roughly 32 days of age. Nest building is first shown by the young from 16 to 20 days only to wane and then reappear again at approximately 30 days of age. Clustering around the female is dependent on the nursing association and as lactation wanes so does clustering. The period of family group foraging may be as short as 3 to 5 days.
The young at birth are roughly 60 mm in total length. At approximately 30 days of age, they show a total length of from 110 mm to 130 mm. Growth begins to plateau at about 40 days of age. The animals are dependent on milk from birth until approximately 22 days of age when they may be weaned on solid food (see Figures 62 and 63).
Hemicentetes semispinosus.—The description of behavior ontogeny for H. nigriceps fairly well describes the situation of H. semispinosus. At birth H. semispinosus is 60 to 70 mm in total length and, at the conversion to solid food from 18 to 22 days, the animals are approximately 90 to 100 mm long. From this point on, growth is extremely rapid beginning to plateau at approximately 40 days of age when the animals may be anywhere from 140 to 150 mm in total length.
At birth the meatus and the eyes are closed. The meatus opens at approximately 6 days of age; eyes are open at roughly 7 to 8 days of age. Spines and hair are visible shortly after birth and begin to become especially prominent around 5 to 15 days of age. Stridulation is loud and clearly detectable at 16 days of age.
The youngsters begin to emerge from the den and cluster around its entrance at approximately 9 days of age and initiate following of the female from about 12 to 22 days of age. The following response is most pronounced from day 16 to 18. First estrus is estimated to occur at between 35 and 40 days of age.
SOCIAL ORGANIZATION
Hemicentetes nigriceps.—Our marked population of H. nigriceps at Manandroy provided us with much information concerning grouping tendencies in the field. In February 1966, 28 burrows were examined and the range in group size per burrow was one to four with an average of 2.05. The April 1966 series included 22 burrows with a range of one to seven animals and an average of 1.6 animals per burrow. The February 1967 samples for 21 burrows indicated a range from one to eight with an average group size of 2.7 animals per burrow. The most consistent grouping pattern was the female with juveniles or infants. This grouping included 14 of the total burrow systems examined. Solitary males occurred 10 times; solitary females occurred 15 times. A male with adult females occurred 13 times; these females may or may not have been accompanied by juveniles. Male-male and female-female associations were very rare, including two instances each.
In captivity tests were conducted by introducing alien males to groups composed of a female and her young. In general after the young were 19 days of age, the adult female showed little defense of the nest. As noted in experimental encounters, the female will respond by bucking if a male attempts to mount unless she is quite receptive. The young animals generally cluster around an alien but the cluster formation is temporary. In one of the encounter tests, the intruder definitely avoided the nest box of the female and young, denning separately, but in three other encounters, the intruder denned with the female and juveniles from the first night. The juveniles are definitely more prone to become aroused and show a forward crest at first contact with a strange male. Adult females are far more tolerant in their arousal patterns.
From these observations and from studies of the marked populations, we conclude that within a given community of Hemicentetes nigriceps adult females tend to den alone prior to parturition. After parturition they show active nest defense, tolerating no aliens. Males, however, may enter when the young are approximately 2 weeks old, and perhaps mate again with the female in the following week. In one case we noted a male persisting in association with a female through the birth of her young. Juveniles apparently are tolerated indiscriminately in the den or burrow with the female and infants. Adult males do not associate habitually with one another nor do adult females. From our observations in captivity, we have the definite impression that adult males are aggressive to one another and this probably accounts for some spacing. Juvenile males on occasion will form small bachelor groups occupying the same burrow system.
Hemicentetes semispinosus.—A total of eleven dens were examined in 1966 and 1967 in both the area of Ambitolah-Ranomafana and Perinet. As described previously (p. 86) one of these groups consisted of an extended colony of 18 individuals including 2 adult males, 2 adult females, and 14 juveniles. If we discount this maximum colony size of 18, then the groups fall accordingly to the following descriptions: (1) Solitary individuals included two instances of juveniles which were sheltering in what appeared to be temporary burrows and one pregnant female in a well-constructed burrow. (2) Two instances of two individuals denning together were noted; in one, two females both pregnant, in the other a male and a pregnant female. (3) The other five burrows noted included one group of seven including a female and her young; one group of seven consisting of two females, one with infants and the other pregnant; one group of eight consisting of a male and a female with their offspring; one group of seven consisting of a male and a female and five offspring; one group of nine consisting of a male and a female with seven infants. The larger groups of H. semispinosus when compared with H. nignceps in part result from the larger litters produced. There does appear to be a marked tendency for males to associate with females and their juveniles even when the adult female is pregnant. Our evidence from an examination of the colony of 18 indicates that some male-male spacing must occur since the number of adult males was rather disproportionate compared to the number of juvenile females available for impregnating.
Experiments in captivity indicate that a female with young of approximately 14 days of age, will readily allow an alien male to enter her den after an initial rebuff. Further observations on colonies of H. semispinosus indicate that females approaching parturition do tend to shift to a more isolated portion of the burrow system or to a separate den nearby. After the first week or so, females are rather tolerant with respect to the entrance of conspecifics but it would seem essential that the female isolate herself somewhat with her litter in order to maintain some continuity in parental care.
Potentiality for colony formation does exist in H. semispinosus and, in part, is a function of the increased litter size and a tolerant attitude on the part of the female with young of approximately 2 weeks of age. It is evident that for both H. nigriceps and H. semispinosus adult males may associate with the female for shorter or longer periods during pregnancy and nursing.
Stridulation in Hemicentetes
THE PHENOMENON OF STRIDULATION
Introduction
As noted in the preceding sections, sounds may be produced by the specialized quills in the posterior middorsal region of both species of Hemicentetes. A description of the quill structure is included in Petter and Petter-Rousseaux, 1963. Rand (1935) first noted that this group of specialized quills could be moved independently of the other body quills. Gould (1965) first recognized that the quills produce ultrasonics and described the physical characteristics of a stridulation pulse series. Frequency analysis was obtained by feeding stridulation signals from a tape loop on the Precision Instrument tape recorder into a Hewlett Packard Wave Analyser (No. 310A) and then to a Bruel and Kjaer Level Recorder (No. 2305). A stridulation sound shows little harmonic structure. It is organized into a train of pulses and, within each pulse, the energy is broad band noise from about 2 KHz to 200 KHz. Three to five peaks of energy usually occurred in groups, probably a result of three to five primary loci of contact among the three rows of interdigitating stridulating quills. Energy peaks clearly show up at 200 KHz. There was little difference between energy peaks between 50 KHz and 150 KHz. There was an abrupt drop at 150 to 200 KHz. The study by Wever and Herman (1968) indicates that Hemicentetes can hear stridulation within the lower range of the stridulation energy distribution (i.e., up to 18 to 20 KHz). The sensitivity of the ear is sufficient to permit coordinated responses to stridulation at a distance of about 4 meters.
The signal varies in its physical characteristics with respect to duration, amplitude, and the repetition rate of pulses. Some of the variations in pulse characteristics may be correlated with the mood of the stridulating animal, the individual animal, or the species of Hemicentetes which is stridulating. Young individuals of Hemicentetes generally produce a lower intensity sound when they are between the ages of 14 to 17 days. By about 17 days of age the intensity of stridulation is very near adult level. Intensity of 3 adult Hemicentetes semispinosus (two females and one male) ranged between 61 and 63 db re 2 × 10−4 microbars; intensities of two juveniles 29 and 30 days old were 60 and 61 db. Distance from the microphone to the stridulating organ was approximately 3 inches. Tape-recorded stridulation sounds were displayed on an oscilloscope and the signals were photographed using a 35 mm oscilloscope camera; these photographs were analyzed for changes in duration, pulse repetition rate and relative differences in intensity. Depending on the degree of arousal displayed by the animal, two signal types may be discerned: (1) slow intermittent stridulation and (2) fast, loud stridulation. A random sample of stridulation pulses of two H. semispinosus and two H. nigriceps while exposed to different stimuli such as dim and bright light and teasing, revealed a much higher rate of shorter duration pulses in H. nigriceps (Table 9). Range of durations was the same for both species. Probably the fewer number of stridulation quills (Figure 51) on Hemicentetes nigriceps accounts for their shorter duration and lower intensity.
The Motivational Basis for Stridulation
Stridulation is accomplished by erecting the stridulating quills and moving them back and forth so that they strike against one another and thereby set up a complex wave system. The erection of stridulation quills is generally accompanied by some partial erection of the other body quills. This immediately suggested to us that there could be some correlation between the degree of spinal erection shown by the animal and the degree or quality of stridulation. As we have outlined before, the crest is composed of specialized quills for driving into an enemy. Thus, crest erection was analyzed to determine whether stridulation correlated with the degree of spinal erection.
Such an undertaking necessitated the establishment of a behavioral classification so that, as a first step, spinal erection could be correlated with the types of behavior that the animal was engaged in. As we have outlined in the preceding sections, the behavior of each species was examined in detail so that a complete inventory of its behavioral repertoire was available to us. This permitted us to establish a functional classification; that is, a classification of behavior which grouped together those behavior patterns which appeared to be involved in the same adaptive process (Hinde, 1966, p. 12). Thus, we could accept as a starting point the sleeping animal exhibiting little arousal and no spinal erection. From this we could consider a class of activities termed maintenance activities which would involve subcategories such as: simple locomotion, care of the body surface, urination and defecation, searching for foodstuffs, ingestion and such related activities as digging and nest building.
A second major functional category, overlapping somewhat with the preceding, included those behavior patterns associated with exploration and foraging in the environment. It is within this class of activities that we may note the tendency either to approach or withdraw on the part of a given animal when encountering an unfamiliar object. Approach could lead to a variety of other actions falling under the first class of maintenance activities, such as feeding, or it could lead to new functional categories of behavior such as contact-promoting and sexual behavior toward conspecifics. Pure withdrawal involves, ultimately, the manifestation of flight behavior. If the object encountered were not a prey object or a conspecific, offensive and defensive behavior or flight behaviors could be shown depending on the quality of the stimuli received. Which course of action the animal took would in part depend on its internal physiological state and the quality of environmental stimuli impinging on it.
For the moment, let us consider the degree of crest erection as a manifestation of the degree of defensive spinal erection since the erection of the crest never occurs unless some body spines are also erected. Further, let us briefly survey those categories of behavior that may be occurring when the crest is erected. After a long period of observation, we established that no crest erection was associated with activity during relaxed locomotion or during flight. A relaxed crest with partial erection of body quills often occurred while the animal was foraging or exploring a slightly unfamiliar environment. Crest half erect with body quills erect occurred during agonistic contexts with respect to conspecifics or during exploration in a novel environment. Full crest erection with partial erection of body quills could occur during exploration or when moving into the sun from shade. Either full crest erection or forward crest (which involves a rolling forward of the brow accompanied with butting or bucking and the crunch sound) was generally exhibited during extreme offensive or defensive activities, seldom with conspecifics but more often with potentially dangerous stimulation. The behavior associated with active offense including bucking has been described previously (see p. 93). By noting the degree of crest erection and the associated circumstances, we were led to attempt a method of quantifying the degree of crest erection shown and correlate this with a known stimulus input.
Light Intensity and Crest Erection
Since crest erection would be shown to a change in the intensity of illumination, we proposed to investigate the relationship between change in incident illumination and the degree of crest erection shown by the animal. Four arousal states were defined in terms of the behavior of the animal. These were crest half erect, crest fully erect, crest forward, and crest forward plus bucking (see Figure 59).
The experiments were run as follows: The subject was placed in a box open at the front and having a background of ruled paper. The animal would then be photographed during exposure to a defined amount of light. By examining the printed photographs after the experiments, one could with the aid of the ruled background determine the angle formed between an arbitrary line drawn from the tip of the nose through the eye and the leading edge of the crest. Thus, the degree of crest erection could be expressed in terms of a measurable angle and this degree of crest erection could then be correlated with the change in light intensity.
The first series of measurements involved placing the animal in the observation box with 10 seconds exposure to four known stimuli. The first stimulus or condition consisted of low illumination (<5 foot candles), the second to dim light (32 foot candles), the third to bright light (500 foot candles), and the fourth to bright light plus mechanical stimuli (i.e., touching the animal's facial vibrissae with a stick). After each of the 10-second exposures to the defined preceding four stimuli, four pictures were taken at four frames per second. Later one could measure the degree of spinal erection as a function of four increasing stimulus intensities set arbitrarily at 1 to 4. Both H. semispinosus and H. nigriceps were tested using eight subjects of the former and six of the latter. The data are portrayed in Figure 64.
Our conclusions based on this study were: (1) crest erection is in part independent of body-quill erection; (2) Hemicentetes nigriceps is more sensitive to change in background light than H. semispinosus; and (3) the degree of crest erection is directly proportional to the increase in stimulus intensity.
A qualitative inspection of our data indicated that under the first condition, that is low illumination or less than 5 foot candles of light, the animals were prone to exhibit movement and investigatory behavior with some body-quill erection and either no nuchal crest erection or half erection. Under the second condition at 32 foot candles, the animals still moved and explored with body quills erect but there was more of a tendency to show from half to full nuchal crest erection. Under the third condition of 500 foot candles of light, there was still some gross movement with body quills erected but there was a marked tendency to show a full forward nuchal crest. With the addition of tactile stimuli to the bright light, the nuchal crest was maintained at the forward position and the animals showed a tendency to buck and/or crouch low to the ground. H. nigriceps displayed a pronounced tendency under the third condition to attempt to dig or burrow into the floor of the test box. To express the data another way, it was possible to say that the erection of body quills has the lowest threshold and there is an increasing threshold as one considers crest half erect, crest fully erect, or crest forward. The highest threshold of all appeared to involve bucking and crunching. With higher thresholds for actions, an increasing strength of the stimulus is necessary to elicit the response (Table 10).
Arousal and Stridulation
The next set of experiments endeavored to clarify whether there were a relationship between the stridulation produced by the animal, the degree of crest erection, and the change in background stimulation. The subject was placed in an arena and the observer sat in front of the arena describing its appearance and behavior by speaking softly into a microphone. The stridulation was recorded simultaneously utilizing the second channel on the tape. In order to arouse the animal, the 1,000 foot candle lamp was switched on during the observation period; thus, the subject and his condition could be verbally recorded while the stridulation rate was recorded and then the change in the background light intensity could be correlated with changes in the stridulation rate and changes in crest erection and behavior. The data from these recordings and observations are portrayed in Figure 65.
The general conclusions were as follows: (1) The position of the crest is not necessarily an indicator of stridulation rate. (2) Low intermittent stridulation can occur during moderate arousal8 even though the animal is immobile, if the animal is engaged in some activity such as exploring the substrate with its nose, or testing the air. (3) Low intermittent stridulation can occur when a highly aroused animal is moving in a jerky hesitant fashion. (4) Loud and rapid stridulation occurs when the animal is highly aroused but not frozen into immobility. (5) A sudden stimulus contrast such as a change in the ambient lighting may cause a sudden increase in stridulation rate and intensity followed by a quiet interval as the animal apparently chooses a course of action. (6) The animal does not stridulate while defecating, chewing, or when “frozen” into immobility following an initial massive change in stimulus contrast. The animal stridulates almost continuously when moving and active. Figure 66 shows actual stridulation rates with associated crest position.
Sound Intensity and Crest Erection
In order to investigate the relationship between the degree of crest erection and spinal erection shown by the animal and the change in background noise, the following observational procedure was established: The observer sat in front of an arena with a light-proof hood covering the arena and the observer's body. The animal was observed and notes were taken by means of a tape recorder while a given sound was played back to the animal. The degree of crest erection shown by the animal could be noted before and after the sound input. Thus a known sound input was given to an animal in a known state of arousal as evinced by its degree of crest erection. The sound was a recording of stridulation by a Hemicentetes. This test was necessary as a precursor to our playback experiments, since it gave us some idea of the range of variability in response by an animal in a known motivational state to a known playback input. Four Hemicentetes nigriceps were selected as subjects. The total duration of playback was greater than 3 seconds but less than 15 seconds. This was necessary to avoid a habituation effect. Analysis of the data indicated that, given the same input, it would appear that if the degree of crest erection is known beforehand, the animal after perceiving the known input will exhibit the next stage of crest erection and arousal as an initial effect. The data by which we reached this conclusion are portrayed in Table 11.
Stridulation and Circumstances of Occurrence
The animals could be observed in the various arenas and their activity monitored by means of the ultrasonic microphone (see Appendix H). By observing and monitoring with earphones, it was possible to make the following correlations: The maximum probability of stridulation occurs when the animal is half aroused or at the stage 2 level of arousal (i.e., crest half erect, center quills erect, no crouching, and active movement of the head), but it can occur during full erect and forward crest if the animal is engaged in offensive behavior. Stridulation occurs while feeding, during social contact, during courtship behavior and when the male is mounted on the female, during exploration activities, and during flight activities, or when moving away. When an animal is actually eating, it rarely stridulates, but will stridulate when it extends to pick up a worm, when it shakes a worm, or when it is pulling a worm from another animal. Subjectively speaking, the stridulation is a low intensity single burst or 2 or 3 brief pulses. If an animal is startled and flees, it stridulates loudly with a continuous train of pulses when it finally initiates flight. If an animal is grasped, the pulses are rapid, loud, and repetitive.
Each line of data begins at the start of the new light condition. The order of administering the stimuli was: dim light, bright light, no light, teasing and bright light.
The upper set of H. semispinosus data represents the stridulation rate of the same animal. The lower set of data pertains to a second animal. All the data under H. nigriceps are from the same animal.
The general conclusions from the light experiments and the sound playback experiments were as follows: (1) An unspecific change in the background stimulation, such as change in light intensity or change in sound input even if it be stridulation, will produce a general arousal in the animal. This arousal is manifest by the degree of spinal and crest erection. (2) One cannot predict what the subject will do next, but one can say that all playback experiments would be profoundly influenced by the degree of arousal exhibited by the receiving animal. (3) Hence, any standardization in playback experiments would necessitate a standardization of the degree of arousal displayed on the part of the receiving animal. (4) One of the better methods of determining arousal was to note the degree of crest erection shown by the receiving animal prior to playback.
Stridulation occurs during a wide range of activities. Low intermittent stridulation generally occurs when the animal is active and doing a variety of activities but not exhibiting strong offensive-defensive or escape reactions. Upon being presented with a sudden change in background stimulation, the animal will begin to show offensive and defensive behaviors with the preliminaries being strong crest and body spine erection. At this time, stridulation may pass through a period of being loud and rapid, then quiet, and then loud and rapid, and finally waning as the animal's degree of arousal wanes. Hence, there is only a restricted correlation between the degree of arousal shown by the animal and the quality of the stridulation. The degree of defensive arousal shown by the animal is in part a function of rather unspecific stimulus inputs. An exception to the preceding generalization is the profound arousal and active offensive behavior demonstrated by an animal when presented with the odor of a predator such as Galidia or Fossa (see p. 94).
The answer to the original question: “Is there a correlation between degree of spinal erection and stridulation?” can best be answered by saying that Hemicentetes stridulates over a wide range of motivational states. The degree of crest erection is not necessarily an indicator of the quality of stridulation being produced. Stridulation may be produced whether or not the crest is erect, whether or not the animal is crouched or high on its forelegs, whether or not the animal is exhibiting active offensive or defensive behaviors. The only necessary correlation between stridulation and spinal erection is that the middorsal line of quills be erect when stridulation is occurring. Further, the form of the stridulation whether it is soft and intermittent or loud and rapid is in part a function of arousal but not absolutely correlated with any degree of crest erection. Rather it is correlated with the degree of stimulus contrast that the animal has just received. Full forward crest position was most frequently associated with a halt in stridulation when Hemicentetes responded to the bright light stimulus. “Lights out” also halted stridulation but the crest usually relaxed from its former position. Stimulus contrast which leads to an elevation in the arousal state but not an elevation to the tertiary or quaternary states (i.e., defensive-offensive) is liable to involve the production of loud stridulation. Low intermittent stridulation is more liable to occur when the animal has not experienced any drastic change in background stimulation but is going about its activities with no profound alteration in autonomic activity.
THE PLAYBACK EXPERIMENTS9
Introduction
Considering the preceding information, the following points are cogent to the design of playback experiments: (1) The initial state of the receiving animal is important. Arousal to the point of exhibiting flight, avoidance, approach, or offensive and defensive systems of behavior including crest erection, is in part unspecific when the stimulus is considered, and all of these arousal forms may be shown to a wide variety of changes in the stimulus field including a shift in the intensity of the stridulation which is being played back to the animal. Thus, (2) the selection of the type of stridulation is important if we know the arousal state of the recipient; however, the state of arousal in the receiving individual could be estimated only imperfectly. (3) In nature, the stridulation signal constantly varies in output. Stridulation amplitude and repetition rate can be grossly correlated with the arousal state and activities of the presumptive sending animal. (4) During playback care must be exercised not to habituate the presumptive recipient to the signal (see Figure 67).
Preliminary Tests
The following preliminary tests were set up for an initial analysis of the behavior of Hemicentetes nigriceps and H. semispinosus to playback stridulation. A loudspeaker was placed in an arena which was inhabited by an established group. For example, two families of Hemicentetes nigriceps, a female and four babies, and a female and two babies were each established in a 4×4 foot arena. (This will be designated as the “small arena.”) A further example would be the Hemicentetes semispinosus family in a study cage 18×4 feet. In these established cages, the animals were allowed to move freely and when they were in the vicinity of the speaker or feeding near the speaker, an observer signaled a second man who played back a known signal to the animals. Tape loops were prepared of several types of stridulation including recordings of (1) an animal moving about in a sound insulated box with the microphone held within an inch of its stridulating organ, (2) an animal which was hand held in front of the microphone while producing rapid, loud stridulation, and (3) artificial stroking of the stridulating quills. Control sounds were used including (1) background noise of the recording and (2) the sound of rustling leaves. In order to sample the responses of animals according to their age, females with young were chosen for H. nigriceps as indicated in the preceding paragraphs, and for H. semispinosus a family of five adults and four juveniles were tested.
The results of our preliminary playback indicated considerable variability in response; however, if we consider only the responses to the first presentation in a given test series to a designated individual, then the following results were obtained:
Eight H. nigriceps received 24 stimulus presentations. Twenty-one presentations of loud stridulation were played back to the eight subjects. Ten animals avoided the loudspeaker or fled. In seven tests the animals stayed in the vicinity of the speaker exhibiting varying degrees of crest erection. In three cases the animals approached the speaker. Three presentations of low stridulation were offered which resulted in three responses which involved remaining in the vicinity of the speaker with little or some crest erection. Four presentations were made of the control stimulus which evoked no responses.
Five adult and four juvenile H. semispinosus were subjected to 15 playbacks. There were six presentations of loud rapid stridulation which resulted in six avoidance or flight responses to the loudspeaker. There were nine presentations of low intermittent stridulation plus the hiss sound which were made during a male courtship. These nine playbacks resulted in five approaches to the speaker, one stay in the same position, and three avoidance or flight responses.
An interpretation of the above results is somewhat difficult but they indicated the possibility that stridulation contains the following information: (1) It indicates the position of another Hemicentetes; (2) it indicates the mood of the sender, that is, either the sender is greatly excited or not greatly aroused, but active. Thus, the receiving animal is disposed to approach, ignore, or avoid, depending on its own motivational state.
The Orientation Test
An experiment was conducted with a series of infants during their phase of attachment to the female. We endeavored to test if the mother's stridulation served in any way to coordinate their movements and serve as an orientation cue, since stridulation could indicate the position of the mother to the infants. In order to test this hypothesis, we used families of Hemicentetes semispinosus.
An arena was set up measuring 7×7 feet. The floor of the arena was divided into 49 squares, 1×1 foot. Observations were carried out in darkness during the early part of the evening utilizing an infrared viewer. Either one or two speakers were involved. If we employed two speakers, they were placed in diagonally adjacent corners; when utilizing one speaker, we placed it in one of the four corners of the arena. The position of the speakers was changed after each trial. An animal was tested in the arena only once in a given evening and given only one test per recorded sound type. A variety of preliminary trials were run including tests where the animal was released in the center of the arena from a small box and permitted to wander about during 60 seconds of playback.
The young showed no special interest during playback and only occasionaly went to a speaker producing stridulation sounds. We detected some tendency to approach the speaker when the playback ceased; therefore, we established the following delayed response test: A young infant was placed in a box in the center square of the grid. By means of a piece of string, the box was lifted only after the infant had been subjected to playback of either stridulation or control sound for an interval of time varing from 15 to 30 seconds. No more than two passes of the same tape were permitted to a given individual, thus avoiding the habituation phenomenon. Then, the animal's movements could be traced on a sheet of paper by referring to the square of the grid to which it moved upon being released at the cessation of playback (see Figure 68). Only one playback speaker was involved in this test. Thus, the inference was that the animal would be attracted to the corner from which the stridulation had emanated. After each test with an individual, the position of the speaker was rotated so that on all subsequent tests, the speaker was in a location different from that of the preceding test.
The age of the subjects ranged from 9 to 19 days. Twenty-four individuals were tested in a series of 30 trials. There were four replicates of a test series or, if we consider the age of the subject to be a variable, there was one replicate at an age of 15 days (see Table 12). Of the 30 playbacks, 19 involved the playback of stridulation; the remaining 11 playbacks were controls. The control sounds were the background noise of the tape. Of the 19 playbacks of stridulation, 15 of the 19 infants made a direct approach to the corner containing the speaker. One showed direct avoidance of the speaker and in three of the trials the infant's movements were not conclusively oriented with respect to the speaker. In the control series, nine avoidances and two undetermined responses were recorded. The results of this delayed-response test conclusively demonstrated that (1) in the presence of stridulation an infant explores with little reference toward the loudspeaker aside from an occasional approach; (2) at the cessation of stridulation playback, an infant will orientate to and move towards the previous source of stridulation.
A test was run using the same arena with two loudspeakers to determine if there were any preference for the mother's stridulation over the stridulation of another female. Our results indicate that, upon cessation of stridulation, the animals would orientate to the location of the stridulation last heard. If stridulation were played back from the upper right corner in the first series followed by stridulation in the lower left corner in the second series, then when stridulation ceased the infant would move towards the source of the last heard stridulation regardless of whether it was the mother's stridulation or not.
THE ROLE OF THE MOTHER'S STRIDULATION IN NATURE
Experiments were run employing the large outdoor enclosure with the overhead platforms (see Figure 4). In this large enclosure, a female with a group of infants would be established in a nest box and fed and observed for some days. The floor of the enclosure was divided into a grid so that the exact location of the female and her young, when foraging together, could be noted. As previously indicated in the section on H. semispinosus, this animal will forage during the day; therefore, most of our observations were made between 1100 and 1400 hours to coincide with the midday foraging activity.
We know from our ethological studies and long periods of observation that linear following, so typical of Tenrec ecaudatus, is not displayed by the infants of Hemicentetes. Rather, the young remain in the vicinity of a female while she forages and drift with her in a loose formation, but young animals may be as much as 9 to 10 feet away from a female during this activity. Furthermore, our observations indicated that as the young mature the female has a decreasing influence on their position in space and loose clusters of juveniles may be formed rather than a tendency to orientate toward the mother herself. Apparently the young learn routes to the feeding area and back to the nest and learn specific feeding loci by associating with the female over a period of approximately 4 days. Thus, the learning period and the phenomenon of attention to the mother only persists in this species for some 4 days whereupon they begin to forage in a more and more independent fashion (see p. 97).
Thus, with the grid pattern in the large outdoor enclosure, it was possible to plot the position and estimate the distance of the female to the nearest young or the distance for a given young to the nearest sibling. These inter-individual distances were plotted as a function of the distance of a given individual from the nest.
The following tests were run utilizing young of an equivalent age: (1) The female with an intact stridulating organ was observed and the positions of the young with respect to her and the nest box plotted. (2) The female was caught, her stridulating organ glued or cut, and the subsequent deployment of the young with respect to the mother was noted. The results of these experiments are included in Figures 69–72. Clearly, the young stay near the female and/or each other, if the female's stridulating organ is immobilized. Thus, they depend on other cues such as the odor of the mother or the sound she makes while foraging; these secondary cues apparently preclude foraging at the distance of 9 to 10 feet as in the case when the mother's stridulating organ is intact.
A second test was run whereby the mother, during her foraging, was induced to enter a small box with worms in it and the box itself was controlled by a system of strings to the overhead platform. Thus, the door on the box could be closed and the box itself lifted and shifted overhead to a new position above the arena. The movements of the young and their deployment in space could then be studied as a function of whether the mother who was shifted in the box had an intact stridulating organ or a glued stridulating organ. The results are presented in Figure 73 and indicate that a displaced female with an intact organ will induce the young to drift gradually in the outdoor enclosure to the vicinity below her new locus.
A preliminary experiment approached the problem of sound deprivation when the infants were very young. The stridulating organ of one female Hemicentetes semipinosus was cut when her young were 3 days of age. At 10 days of age the infants were seen wandering over the entire outdoor enclosure and showing no signs of being able to locate their mother. At this age they ordinarily would be clustered near the female or each other and within a few feet of the nest box.
We conclude from the preceding experimental series with H. semispinosus that stridulation serves in the female-young unit as an identifier of the female's locus. The unanswered question is whether a shift in intensity of the female's stridulation influences the behavior of the young. We have on numerous occasions observed in the large outdoor enclosure that the female upon being frightened will hesitate, then begin to stridulate rapidly and flee. The youngsters in her vicinity will generally assume upright crests, attentiveness, and flee with her to the nest box. The flight is not uniform and may take over a minute and a half before all young have moved to the nest. We are unable to determine whether the young flee because they are alerted and frightened by the same stimulus that acted on the female or whether, indeed, they are further activated by the change in intensity of her stridulation, or they are induced to flee because of the sounds the female makes as she crashes through the underbrush racing toward the nest. Probably all these things have an effect on the general arousal of the young and potentiate the flight response and serve in some way to direct it. The experimental resolution of this problem remains incomplete.
OTHER FUNCTIONS OF STRIDULATION
There are other possible functions for stridulation, some of which we have investigated and others which we have only partially attempted to clarify. It had occurred to us that stridulation might be involved in echo-location. Gould's work in 1965 indicated that H. semispinosus can orientate in the absence of visual cues and locate objects in space. The animal can accomplish this when its spines have been clipped, therefore it would seem that stridulation is not entirely essential to echo-location and the animal seems to do rather well with tongue clicks.
Nevertheless, the possibility that stridulation might be involved in echo-location was not dismissed. A predator approaching Hemicentetes from behind might conceivably distort the sound field to the rear of the animal resulting in an increased echo return of the stridulation pulses, thus warning the animal of an object behind it. Hemicentetes was tested in an arena by stimulating it with predator odor to exhibit the full offensive and defensive reaction. Then various objects on the end of a stick were placed behind it and the turning frequency of the animal measured either with or without an object behind it. No special alteration in its turning tendency could be noted.
The possibility that stridulation serves as an antipredator signal was considered. Since the predator tests indicate that the detachable, barbed quills are a considerable deterrent to predation, although not a complete guarantee of freedom from the activities of predators, the boldly marked patterns of H. nigriceps and H. semispinosus are undoubtedly warning colors which serve as an antipredator mechanism. We conducted some playback experiments with one Galidia out of our group of three which would consistently kill Hemicentetes and found that the signal of stridulation could serve to orientate the Galidia to the Hemicentetes rather than warn it away. This does not exclude the possibility that stridulation could be a warning signal to predators, since it may be effective with those predators which have had an adverse experience with the Hemicentetes. Certainly the sound of stridulation is not a deterrent to a predator that has learned a technique to kill Hemicentetes (see Figure 13).
We considered the possibility that stridulation may be a mechanism for attracting earthworms to the surface of the ground. This would seem a bit far fetched when one considers the position of the organ and dispersion of the sound about the animal; nevertheless, we did play back the sound of stridulation to two species of Madagascar earthworms on several occasions employing a continuous playback loop. We could discern no tendency on the part of the earthworms to come to the surface.
Finally, we should consider once again the situation in Tenrec ecaudatus juveniles where a stridulating sound is produced. Our experiments with Tenrec ecaudatus were not extensive and our results were somewhat inconclusive. It may well be that stridulation in Tenrec ecaudatus serves to coordinate the movements of juveniles during the following of the female; however, our evidence indicates that stridulation is associated with high levels of excitement and not with low levels, as exemplified by peaceful foraging. This would appear to rule out a similar function analogous to that in Hemicentetes semispinosus and H. nigriceps. Therefore, we are forced to consider the possibility that stridulation in Tenrec ecaudatus juveniles may have been selected for because it mimics stridulation in Hemicentetes.
Consider the following possibility. Stridulation may have been evolved in Hemicentetes as a means of communicating the position of one adult to another and may be involved in the location of mates but is principally involved in the following and clustering activity of the infant Hemicentetes during the initial foraging phase with the mother. It may secondarily serve as a warning signal to predators. Thus, the color and spinescence of the juvenile Tenrec might be a partial mimicry of Hemicentetes and further set the stage for selection favoring stridulation on the part of Tenrec ecaudatus young when menaced by a predator. The stridulation would enhance further the mimicry between juvenile Tenrec ecaudatus and adult or infant Hemicentetes. This final question may never be completely resolved until further studies on the role of stridulation in Tenrec ecaudatus have been undertaken.
Comparisons and Extensions
EVOLUTIONARY TRENDS AMONG THE TENRECIDAE
If we assume that the ancestral tenrecid was small in size and similar to Geogale then we could imagine an animal with a hair-covered body; a naked tail approximately 75 to 100 percent of the body length; perhaps an imperfect homeotherm; cryptic in its behavior; and an insect-eater with some ability to forage on the forest floor as well as climb. While occupying an environment with little competition, speciation and adaptive radiation would take place whenever sufficient geographical isolation had been achieved. The initial modifications in morphology and behavior would include: (1) differentiation into small terrestrial surface foragers; that is, a head and body length less than 120 mm; and (2) evolution into forms that were more fossorially or aquatically adapted. Within the Oryzorictinae the upper extremes in size would then occur in those groups which had adapted to the least conspicuous mode of habitat utilization, i.e., the aquatic and fossorial forms; and even today, these are the forms that show the greatest head and body length (130 to 170 mm).
Aside from structural modification to enable the animals to utilize the extremes of the environment, there would be little modification of the basic behavioral repertoire. The greatest departure in the evolution of the tenrecs occurred when adaptive radiation began to give rise to the subfamily Tenrecinae. The general trend in the evolution of this group is toward a larger body size and, as a result, a loss of inconspicuousness and the development of rather elaborate antipredator behavior. The evolution of the subfamily Tenrecinae further involved the loss of the tail.
Two discernible foraging types evolved including (1) a terrestrial and semi-arboreal form in the deliberate climbers, Setifer and Echinops, and (2) the almost completely terrestrial genera with an increased digging ability and modification of the forepaws including Hemicentetes and Tenrec. Both Setifer and Echinops, although feeding on invertebrates, tended to become generalized omnivores as did Tenrec ecaudatus. On the other hand, Hemicentetes became much more specialized in its feeding techniques until it had specialized as an earthworm feeder with modifications in its skull and tooth structure.
The subfamily Tenrecinae, in the course of its evolution, early acquired a spinescent coat. It would appear that the acquisition of spines in Echinops, Setifer, and Hemicentetes has resulted in profound modifications of their antipredator behavior and offensive-defensive behavior syndrome. For example, Echinops and Setifer are able to roll the brow forward and buck with the head or to roll completely into an impregnable spiny ball. Hemicentetes evolved barbed detachable spines and has capitalized on rolling the crest forward thus exposing a crown of spines, and by employing a rushing, bucking, and pivoting technique its spines can be driven into a predator. Tenrec ecaudatus, although related to Hemicentetes (see Borgaonkar and Gould, 1965) and still partially spinescent, has tended to lose its spines and as an antipredator device to rely more on size and speed coupled with its ability to stand and fight. The juvenile T. ecaudatus, however, greatly resembles Hemicentetes. Indeed, the resemblance including the stridulatory spines is so remarkable that one is tempted to believe that in a way Hemicentetes is a specialized, neotenic Tenrec ecaudatus.
In the course of Hemicentetes' evolution the striped pattern has passed from a protective color pattern, as would appear to be the case in the young Tenrec ecaudatus, to a warning color which is part of its antipredator syndrome. The black and white pattern of H. nigriceps is a very effective nocturnal warning coloration whereas the yellow and black is a more versatile warning coloration for the rainforest adapted H. semispinosus which is also to some extent diurnal. With the exception of the aquatic Limnogale, the genus Hemicentetes is the most specialized in its feeding habits and in its antipredator behavior.
A COMPARISON OF BEHAVIOR CATEGORIES
It is instructive to consider the functional categories of behavior and to trace the evolutionary trends within each category for all the species of tenrecs included in this study.
The categories of “comfort movements” and “marking movements” exhibit considerable uniformity when the family is surveyed as a whole. The face-washing pattern is present in Microgale and is obviously a conservative pattern. It is retained in the specialized Tenrecinae within the genera Echinops and Setifer, but in the genera Hemicentetes and Tenrec this pattern is no longer present. This loss may in part be related to the fact that it is awkward for these tailless forms to sit upright in a crouch, but it is more probable that this loss is related to the specialization of the hand in Tenrec and Hemicentetes as a digging organ. The forepaw is wide, bearing stout claws, which reduces its use as a cleaning organ. Nevertheless, the propensity to use the forepaw to wipe at the face is still present in a modified form in both these genera. The ritualized wiping pattern and simultaneous stroking while sitting upright, so prominent in the other genera, has disappeared as a complex complete pattern in Tenrec and Hemicentetes.
The perineal (= cloacal) drag as a marking movement is present in all species examined. Defecation near the burrow entrance is prominent in the subfamily Tenrecinae. Digging in the substrate, then backing into it, depositing the feces, and covering them with fresh earth is unique to the genera Hemicentetes and Tenrec. In the latter genus, this behavior pattern is carried to its most ritualized form.
Foraging behavior in the Tenrecidae is rather similar. The long flexible nose is.inserted in crevices and cracks and in the substrate. If an invertebrate is located, it is seized with the mouth. The use of the forepaws in prey capture is minimal. Forepaws may be used to brace or may be used to hold the prey down while it is torn apart, but aside from pinning, the forepaws are not involved in a primary prey-catching movement. In Tenrec ecaudatus, the forepaws may be used to pin a prey object before the bite is delivered but the mouth itself is often the primary capture organ and the forepaws are involved only secondarily. This is also true of Hemicentetes when it forages for earthworms.
Offensive and defensive behavior patterns, as stated previously, have undergone great modification within the family. Biting is universal but has tended to pass out as an active pattern in Hemicentetes, especially with respect to interspecific defense, although it is still retained in intraspecific fighting behavior. The gape reaction which probably evolved from an intention movement to bite is present throughout all genera of the family with the exception of Hemicentetes where it has waned and is not shown as a display pattern.
Head bucking, pivoting on the hind legs, and stamping with the forepaws have arisen as a form of defensive behavior in the subfamily Tenrecinae. This is especially effective in the very spinescent genera where the quills on the head may be jabbed into an enemy. Such a bucking pattern is still present in Tenrec ecaudatus even with the lack of formidable spines on the head. Here, however, the buck is combined with the open mouth and slashing bite. Rolling into a spiny impregnable ball is a correlate of the completely spinescent dorsum with nondetachable spines which one finds in the specialized genera Setifer and Echinops.
Turning to the patterns involved in social interaction, we have seen that the contact-promoting behaviors displayed during initial, amiable encounters involve placing the nose or mouth in glandular areas on the body. These tactile configurations are quite uniform for the whole family from Microgale through Tenrec ecaudatus, and only minor variations are shown. In sexual behavior, licking the fur of the partner or nipping at the partner are common and shown throughout all genera. The neck grip is employed by all males while mounting the female. The mount in the family Tenrecidae is prolonged especially when one compares mount duration with those demonstrated by males of the genera Suncus and Blarina in the insectivore family Soricidae. The brief mounts of soricids are reminiscent of the brief mountings of many of the common cricetine and murine rodents. The long mount in tenrecids is reminiscent of the long mount in the genus Dipodomys of the heteromyid rodents (Eisenberg, 1963).
Nest-building patterns are very similar for all species of the tenrecids. The nest defense syndrome involves elements of the defensive behaviors originally discussed in the preceding paragraphs. The Microgales typically gape and attempt to bite. Setifer and Echinops produce “putt-putt” sounds and attempt to buck and drive spines into the offender's body. Hemicentetes exhibits a similar pattern of “putt-putt” sounds, crunching, erect crest, bucking, and stamping. Tenrec ecaudatus hisses, foot stamps, and with an erect crest and half open mouth attempts to bite and slash at an intruder.
The vocalizations, when one compares all species of the family Tenrecidae, exhibit remarkable uniformity. There is, first of all, a class of noisy sounds showing little harmonic structure. One set of sounds is related to rapid inhalation and exhalation and may be referred to as the hiss and puff. This hissing is to be found in all genera studied. The very rapid “putt-putt” sound with high overtones is characteristic of Hemicentetes, Setifer, and Echinops. The buzz or crunch sound which again exhibits little harmonic structure is to be found in all genera studied.
The second class of sounds exhibits some harmonic structure with clearcut overtones. These would include the soft squeaks and the chirps or a repeated series of brief squeaks which we term a twitter. The note form in these vocalization types may be linear or may approximate chevron form as defined by Andrew (1964). In addition to these rather harmonic clear sounds, another voiced sound, showing overtones but nevertheless a great deal of noise, can be termed the grunt This sound has been noted for Tenrec ecaudatus, Hemicentetes, and Setifer. At high-intensity stimulation the chirp itself may show blurred harmonic structure with a great deal of noise being introduced into the sound.
The female's response to her young is similar for all the species studied, involving licking and cleaning the young; retrieving the young by mouth transport if they stray from the nest; and drawing the young under her body employing the nose. The young are huddled over and suckled—these are of course basic mammalian patterns. The response of the young to the mother appears to be similar for all genera studied in that the young follow the mother for the first few nights on their foraging excursions. The genus Hemicentetes shows the most specialized mode of communication between the female and young and, in conjunction with this, the most specialized feeding habits, thus necessitating perhaps increased guidance of the young in their initial foraging trips (see Figure 74).
When the basic data concerning the reproduction of the Tenrecidae are compared in tabular form (see Table 13), we again find remarkable uniformity in duration of gestation, rapidity of development by the young, and lactation time.
In summary, then, the Tenrecidae exhibit interesting variations on a common behavioral repertoire. As stated in the introductory paragraphs, the more prominent variations are concerned with behavioral specialization for feeding and antipredator mechanisms.
THE INSECTIVORA: A CONSIDERATION OF EVOLUTIONARY LIMITS
The isolated condition of the Madagascan landmass permitted an extended adaptive radiation by at least four Orders of mammals which reached the island independently and enjoyed a lack of conventional competitors. The original invaders of Madagascar were conservative in their body plan, especially the insectivores and primates. The muscle systems, dentition, brains, and sense organs reflected an unspecialized morphology.
If we consider the Insectivora, the primitive ancestral form showed a triconodont dentition; undifferentiated muscle slips in the head region and arm; a lissencephalic brain; audition and olfaction as the primary sensory modalities; and a reduced eye which could process little complex, visual data. Complexity and specialization in morphology reached its climax with the evolution of Hemicentetes and Tenrec. Complexity and specialization of behavior patterns are reflected in the communication system of Hemicentetes and its more complex social structure. H. semispinosus exhibits colony formation. Colony formation generally results from the selection of a favorable denning site with respect to a convenient feeding locus. Colony formation leads to the utilization of the same denning space by several females, probably related by descent, and a male. Such a related colony may not persist through the adverse winter season and/or a prolonged period of torpor.
The communication system evolved by Hemicentetes permits indication of position by an individual (generally the mother to her infant). This system evolved from simple spinal erection and the rubbing together produced as a concomitant of autonomic arousal and was enhanced by selection favoring spinescence rather than hairs alone. Originally, it was probably a warning signal to a predator and secondarily it had the possibility of being a warning signal to associated individuals. It was potentially an indicator of arousal in response to a change in the ambient conditions. Later, as selection for enhancement of the signal promoted the evolution of stridulating quills, a concomitant selection favored the lowering of the threshold for stridulation; and, instead of being a warning signal alone, it became a signal indicating the position of an adult, generally the mother. She produced the sound whenever she exhibited general arousal or activity. The primary selective advantage for the enhancement of this signal and the lowering of the threshold probably was increased survivorship of the offspring of a female which produced a more audible signal with a lower threshold. The young probably survived better because they were able to learn the loci of foodstuffs and learn easily the route to the feeding area and back to the nest. The information conveyed in the signal is unspecific; nevertheless, it serves as an indicator to the young of the position of the female.
The time when stridulation is of crucial importance in the life of the animal is short, since the female-young feeding systems may last as little as 4 days; however, complex female-young interaction systems may be quite intricate over an equally short space of time in other insectivores. For example, the reader is referred to the references by Herter (1957) and Dryden (1968) concerning the caravan formation in Crocidura and Suncus. Here a female, upon emiting a specific twittering sound, causes the young to grasp her tail or one another's tail to form a continuous chain so that the mother, as she flees from the nest, can guide them to a new nesting site. Furthermore, the following response exhibited by a number of primitive mammals involves specific selection for capacities on the part of the young to respond to specific inputs from the female's body itself. This following response has evolved repeatedly and must be of extreme importance to the survivorship of the young even though it persists only a brief time.
The simple following response which one sees in Tenrec ecaudatus and to a lesser extent in Setifer setosus or Echinops telfairi is not enough to insure efficient foraging in the case of Hemicentetes when one considers its increased specialization in feeding habits. In Hemicentetes, eye size reduction reaches its maximum and the snout is in the ground as the animal probes for worms. Hence, if the young are to be effective in feeding, they cannot smell the mother nor can they see her; thus, the sound produced by the mother provides far greater efficiency in permitting the young to maintain contact with her while still foraging on their own. The communication system and colony formation exemplified by Hemicentetes semispinosus are a current limit in social complexity and communication for the Order Insectivora.
It is instructive to look at the evolutionary limits reached by other Orders of mammals which have undergone adaptive radiation on Madagascar. As in the case of the tenrecs, morphological diversity has been achieved in other Madagascar mammals which started from a conservative morphological framework. For example, in the Lemuridae, the original progenitor of the lemurs on Madagascar was probably a small nocturnal form; probably capable of estivation, and exhibiting a slightly unstable thermoregulation pattern. Through adaptive radiation diverse forms were developed including large terrestrial browsing primates, the Megadapidae, diurnal leafeaters such as Indri, and diurnal frugivores such as Lemur and Propithecus. Diurnality and a departure from an insect diet to a fruit or leaf diet was concomitant with the evolution of complex social organizations that included adults of both sexes in the case of the genus Lemur, or family groups as in the case of Indri (see Petter, 1962, and Jolly, 1967). Specialization in foraging habits and concomitant specialization in social structure characterizes the evolution of primates in the Old and New World. Thus, as a given line evolves away from the conservative stem form and by adaptive radiation develops new feeding mechanisms and diurnality, the trend leads inevitably to specialization in the form of social organization and the communication systems.
The Camivora arrived in several independent invasions on Madagascar and have undergone less dramatic radiation but one can see again the same specializations for occupancy of feeding niches, so that a large arboreal cat-like form has evolved from primitive viverrid stock and resulted in the fossa, Cryptoprocta ferox, of today. In a similar manner, a nocturnal digitigrade fox-like species, Fossa fossa, has developed on Madagascar, in marked contrast to the arboreal and diurnal forms such as Galidia elegans and Mungotictus lineatus.
If a mammalian order on Madagascar has had a long enough evolutionary history and its evolution was initiated with an unspecialized stem form possessing a conservative body plan; then, through natural selection, adaptive radiation has resulted in a rather uniform occupancy of the feeding niches available. Indeed, when the Madagascan forms are compared with continental mammals, equivalent feeding niches are filled, thus typifying complementarity as outlined by Darlington (1957). The adaptive radiation in the absence of conventional competition has allowed a rather conservative order, the Insectivora, to achieve surprising levels of social complexity.
CONFIGURATIONS IN SPACE AND TIME
The study of the tenrecs brings up several points of interest to students of mammalian behavior and evolution. When considering the evolution of mammalian behavior, one must consider the sense organs and the “inner world” of the subject. For the tenrecs, the sense of smell, the sense of sound perception, and the sense of touch are of primary importance. We have demonstrated how, in the absence of discrete visual perception, one genus of tenrec, Hemicentetes, has elaborated auditory communication as a means of group integration.
To appreciate the significance of many behavior patterns, one must consider again the tempo of the life cycle. Tenrecs are small mammals with a very rapid growth which achieve sexual maturity within a short time after birth. They can pass through a life cycle in a matter of months with generations renewing themselves on an annual basis. To the student of large mammals, it is a serious handicap to have to adjust one's time sense to the scale imposed by such small mammals as tenrecs. As outlined in the previous section, a 4-day period where the female is guiding the young to and from the feeding areas may be of vital importance in providing the selective impetus for the evolution of a complex communication system such as the stridulating organ of Hemicentetes.
In a similar manner, configurations of small mammals in space may be overlooked as trivial because their time of occurrence is brief. When, for example, a novel stimulus, such as an alien individual, is introduced to a colony of Hemicentetes, the animals orientate toward the stranger and form a semicircle or circle about him; thus, giving a configuration that is very similar to the orientation response of herds of ungulates to, for example, a newcomer. At other times, if the stimulus object is of a sufficiently strong valence and a mother tenrec is accompanied by her young, the mother may extend toward the stimulus while the young hang back. This again is analogous to the type of configuration one sees in an encounter situation with many of the social ungulates. For example, with elephants, an old female will move toward an alien stimulus such as a man, and the younger females and infants hang back (see Figure 75). Furthermore, as pointed out under the discussion of antipredator behavior in Hemicentetes and Tenrec ecaudatus, if a predator odor is wafted toward a group consisting of a mother and her juveniles, they will orientate toward the source and may exhibit attack behavior in unison.
Again, consider the similar behavior patterns that have been noted on the part of elephants and large ungulates with respect to predators. These configurations in space displayed by the tenrecs bear a remarkable similarity to those shown by larger mammals and yet, because they occur briefly and are seldom observed, they tend to be neglected. While much is made of the mobbing behavior or concerted group attacks on predators by primates, such behavior on the part of tenrecs would be brushed over: (1) because it is not known to occur or (2) because it occurs for only a few minutes and then is not repeated.
Let us consider again the configuration shown when a female forages with her infants. In Tenrec ecaudatus the female leads a long chain of youngsters following behind her in regular order. Compare this spatial configuration with that shown by an elephant herd on the march (see Figures 76 and 77) or consider the subtle clustering of the young Hemicentetes remaining approximately 9 feet from the female within hearing range of her stridulating organ. The configuration here indicates a cohesion which would never be guessed at if the existence of the stridulating signal were not known.
Given the sensory limitations of the Tenrecidae and the minor role of visual displays in their interactions, the basic behavioral repertoire of the tenrecs is common to many other mammalian Orders and, indeed, forms the fundament of behavior patterns from which most mammalian patterns are derived. The behavior patterns of more highly evolved or specialized mammals do not differ markedly from the behavior patterns of primitive or morphologically conservative mammals such as the tenrecs. Rather, more highly evolved mammals are larger, have larger brains, take a longer time to mature, and their life cycle and life timing is more similar to our own. As a result, the larger forms of mammalian life are able to assimilate a greater variety of data from their environment, store it in their nervous systems, and retrieve it for use later on. As a consequence of the prolonged association of the young with these larger mammals, the young have a longer time to acquire and process information handed down to them through the activities of their elders. It is this protocultural transmission of information; it is this degree of differentiation by the more highly evolved sense organs; it is the ability to conserve and transmit more complex information that separates higher mammals from such conservative forms as the tenrecs; but, with respect to the basic configurations in space and time and the basic behavioral repertoires, the tenrecs exhibit the fundament which is common to essentially all terrestrial Mammalia.
- bibliographic citation
- Eisenberg, John F. and Gould, Edwin. 1970. "The Tenrecs: a study in mammalian behavior and evolution." Smithsonian Contributions to Zoology. 1-138. https://doi.org/10.5479/si.00810282.27
