Sexual Behavior in the Northern Elephant Seal Mirounga Angustirostris Author(S): Burney J

Home , Northern elephant seal

Sexual Behavior in the Northern Elephant Seal Mirounga angustirostris Author(s): Burney J. Leboeuf Reviewed work(s): Source: Behaviour, Vol. 41, No. 1/2 (1972), pp. 1-26 Published by: BRILL Stable URL: http://www.jstor.org/stable/4533425 . Accessed: 19/02/2013 19:16

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

BRILL is collaborating with JSTOR to digitize, preserve and extend access to Behaviour.

http://www.jstor.org

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions SEXUAL BEHAVIOR IN THE NORTHERN ELEPHANT SEAL MIROUNGA ANGUSTIROSTRIS

BURNEY J. LEBOEUF 1) (Crown College, University of California, Santa Cruz, Calif., U.S.A.)

(With 12 Figures) (Rec. 25-I-1971)

To understand mammalian reproductive behavior, systematic data repre- senting all species is needed. Investigations must be carried out both in the laboratory, where rigid controls can be effected, and in nature, where ques- tions concerning the evolution and adaptive value of the behavior are ap- plicable. In the past, the focus of study has been overwhelming in favor of laboratory and domestic animals such as: mice, rats, guinea pigs, rabbits, cats, dogs, monkeys, chickens, pigs, sheep, cattle, and horses. With few exceptions, little is known about the behavior of these animals in their natural habitats; indeed, since most of them are domesticated, the question has little meaning. There are even greater gaps in our knowledge of reproductive behavior in wild animals that are not amenable to study in the laboratory or cannot be observed within the confines of an outdoor enclosure. This is hardly sur- prising since dugongs, sperm whales, black-footed ferrets, ant eaters, and rock hyraxes, to name a few, are difficult to locate, let alone observe at the right time and long enough to record mating behavior. However, some wild animals, such as several species of pinnipeds, do not present unusual difficulties in observation. Seals and sea lions offer several advantages for study in their natural habitats. In most species, the breeding season is short and takes place at the same time each year. Mating occurs on land or on ice in many species and these locations are traditional. There are

I) I thank MICHAELBONNELL, FINN SANDEGRENand several students for field assistance, IAN STIRLINGand RICHARDGANTT for helpful comments on the manuscript, and CECILE SCHREIBERfor preparing the resume. The work was in part supported by Grant GB-I632I from the National Science Foundation.

Behaviour XLI I

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions 2 BURNEY J. LEBOEUF

few visual obstructionson the rookeriesand since the animals don't move around much, observationscan be made from blinds thereby minimizing disturbance.Individuals can be identified by scars and in a few species, individualscan be marked.Several different species may breed in the same locationthus facilitatingcomparative studies. Males are particularlyinter- esting subjects for sexual researchbecause of the great demandsimposed on them by their social systems. The "haremmaster" is perhaps the best known example of the polygynousmale. The specieswe know most aboutare those whichare of greatestcommercial importanceto man; these include the Alaska fur seal, Callorhinusursinus (BARTHOLOMEW,I953; BARTHOLOMEW& HOEL, I953; PETERSON, I968), the Cape fur seal Arctocephalus pusillus (RAND, 1955) and the Southern elephant seal, Mirounga leonina (LAWS,I956a, I956b,; CARRICK,CSORDAS & INGHAM,I962). Despite intensive study of related phenomena, copulation has been observed rarely, if ever, in the Weddell seal, Leptonychotes wed- delli Lesson, (MANSFIELD, I958; RAY & DE CAMP, I969), the Hawaiian Monk seal, Monachus schauinslandi (KENYON & RICE, I959; WIRTZ,I968), the Grey seal, Halichoerus grypus (HEWER, 1957; CAMERON, 1969), the Harbor seal, Phoca vitulina (SCHEFFER & SLIPP, I944; VENABLES & VENABLES, I957; BISHOP, 1967), and the walrus, Odobenus rosmarus (BROIOKS,I954; LOUGHREY,I959). Investigations of the sea lion family have been limitedby difficulty in distinguishinga mount from a completecopula- tion due to the posture the pair assumes during copulation (PETERSON & BARTHOLOMEW, I967; PETERSON, I968). The purpose of this study was to describe and analyse the sexual behavior pattern in the Northern elephant seal, Mirounga angustirostris. BARTHOLO- MEW(1952) provides a particularly lucid description of the copulatory pattern in this species and LAWS(I956a; i956b) and CARRICKet al. (I962) give a good account of this behavior in the Southern elephant seal M. leonina. The aim of this study was to extend these studies by emphasizing quantitative aspects of the sexual pattern of marked males and females from day to day for periods covering three breeding seasons, relating individual performance to the social context in which it occurred, and comparing the copulatory pattern in elephant seals to other species that have been studied. Particular attention was devoted to differentiating complete copulations from incomplete attempts to mate since few statements about the frequency and time course of matingcan be made withoutthis distinction.Because the social system of elephantseals is quite different from that of sea lions, the former exhibit a social hierarchy and the latter are territorial, differences in their respective copulatory patterns might be expected.

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions SEXUAL BEHAVIOR IN NORTHERN ELEPHANT SEAL 3

METHODS

Observations were conducted at Afio Nuevo Island during three consecutive breeding seasons: I968, I969, and I970. Afio Nuevo Island, 19 miles north of Santa Cruz, California, is the northern-most colony in the breeding range. The reproductive period begins in early December when breeding males arrive and ends around 15 March when the last female departs. The first females arrive in late December and mating occurs between mid-January and mid-March. Elephant seal's breed in two locations on Afio Nuevo Island, a large sandy beach on the northeast side of the island called Area 17 and a smaller sandy cove on the southwest or seaward side of the island called Area 3 (see Figure 2 in ORR & POULTER, I967). Four blinds formed a semi-circle around the breeding beach on Area 17 where the majority of observations were made; one blind overlooked Area 3. The population of breeding males numbered 75-I 15 during the three seasons observed: approximately 5o% of these males were fully developed adults and the rest were sub-adults. The latter ranged in age and size from males just going through puberty, and about the size of females, to males that were just reaching full growth. Even the youngest and smallest sub-adults on the breeding beaches in February and March were sufficiently mature to mate. The number of females in the population ranged from 90oin I968 to 328 in I970. Several observers assisted in the study making it possible to observe the animals an average of 6 hours a day from I December to 15 March each year. Observations usually began at dawn around 0700 hours and ended at sunset around 1700 hours. Observations were made at night when there was suf- ficient moonlight to see the animals. With the aid of binoculars, spotting scopes, cameras and timers, a serial record was kept describing selected behavior patterns and the time at which they occurred. As each male arrived at the start of the season, he was given a name or number which was printed in 6 to 12-inch letters on each side of the animal's back anterior to the hindflippers with a bleaching agent (LE BOEUF & PETERSON, I969a). These marks were readily visible from the blinds and lasted throughout the breeding season until the annual moult in June and July. Numbered metal and plastic tags attached to the webbing of the hindflippers facilitated recognition and remarking in each subsequent season. Approxi- mately 20% of the females on Area 17 were marked with bleach or paint each year. Thirty females were tagged in I969 and I970. During the 3-year period, additional observations were conducted at all

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions 4 BURNEY J. LEBOEUF

other Northern elephant seal breeding colonies: the large rookeries on San Miguel Island, Isla de Guadalupe, and Islas San Benito, and the smaller rookeries on San Nicolas Island, Los Coronados, and Isla Cedros. The following are various terms used to describe components of sexual behavior in elephant seals: Flipper over back (FOB): foreflipper over back, signalling an attempt to copulate (Fig. iB). No genital contact. Mount: postural adjustment following FOB in which the genitals are apposed and the male exhibits pelvic thrusts without penetration (Fig. ic). FOB is maintained throughout. Lumbar flexure (LF): curvature of the mounter's pelvis toward the mountee which accompanies an especially deep thrust and insertion of the penis into the female's vaginal orifice (Fig. Id, Ie). FOB maintained throughout. Disrupted mount (DM): FOB and mount prevented from proceeding to LF because the male is threatened and moved by another male. Disrupted copulation (DC): withdrawal of the male after LF has occurred due to threat or attack by another male (Fig. if).

RESULTS Description of copulation. The salient features of copulation as it occurs on land can be described briefly. Copulation is always initiated by the male. He goes directly to the side of a female lying on her belly, puts a foreflipper over her back, pulls her strongly toward him and attempts to establish genital contact (Fig. Ia, ib, ic). There is no preliminary courtship; the male does not investigate the head or anogenital region of the female. If the female moves away, the male usually pursues her and pins her down by dropping the full weight of his head and neck on her back. An unreceptive female will continue to struggle and resist copulation by emitting a croaking type of vocalization, flipping sand with her foreflippers back toward the face of the male, and swinging her hindquarters back and forth laterally, which causes her hindflippers to smack hard against the male's ventrum and partially extruded penis. A persistent male will attempt to subdue the protesting female by repeatedly dropping the weight of his head and neck on her body, or he will obtain a better purchase on the female by getting a hold on the back of her neck with his teeth. Despite the male's aggressiveness, the unreceptive female eventually wriggle free. A cooperative female is passive, responding to the male's flipper over her back by lying still, elevating her perineum in the lordotic posture, and spreading her hindflippers (Fig. ib). The bull shifts his weight quickly from his belly to his side and positions himself so that his penis is in contact with

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions BEHAVIOUR XLI (I972): 4 PLATE I

F Fig. i. Copulation: The male pins the female down with the weight of his head and neck (a), then puts a foreflipper over the female's back (b) just prior to turning on his side and establishing genital contact (c). Several pelvic thrusts precede insertion which is accompanied by lumbar flexure (d, e). The bull disengages as a dominant bull advances towards him (f).

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions BEHAVIOUR XLI (1972): 4 PLATEII

Fig. 4. The large harem on area I7 at Afio Nuevo Island in early January, I968. The harem is on the right and contains approximately 45 females, some of whom have newborn pups. A few bulls can be seen in or near the aggregation of females but the majority of them are dispersed on the beach to the left of the harem. Note the distance between each male.

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions BEHAVIOUR XLI (I972): 4 PLATE III

40..~

Fig. 6. Female (arrow) leaving the rookery is pursuedby several males.

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions BEHAVIOUR XLI (I972): 4 PLATE IV

Fig. 8. A small part of the crowded rookery, Pilot Rock beach, at Isla de Guadalupe. Note the close proximity between males at the right of the picture.

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions BEHAVIOUR XLI (1972): 4 PLATE V

Fig. 9. An aerial view of the western tip of San Miguel Island showing the distribution of elephant seals (arrows) and California sea lions (the dark clusters near the water) in February. California sea lions, Steller sea lions and Alaska fur seals also breed in this same general area at other times of the year.

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions BEHAVIOUR XLI (1972): 4 PLATE VI

Fig. o1. A single bull holds a small harem of females and their pups in a small sheltered cove at Isla San Benito Oeste.

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions BEHAVIOUR XLI (I972): 4 PLATE VII

......

...... ^... .

I..>

Fig. I. The posture of copulation in the Steller sea lion shortly after insertion (a) and just prior to the termination of copulation when the female turns and bites the male's neck (b).

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions BEHAVIOUR XLI (I972): 5 PLATE VIII

A. :: . :. .: .....

4~~~~~~~~~~~~~~~~~~~~~~."e

~~~~~~~~~~~~

t*4t4V·*'~ is. .A ..

1. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0tOMM

Y......

Fig. I2. Genital investigation by the Alaska fur seal male prior to mounting (a). The copulatory posture of the male and female before insertion (b).

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions SEXUAL BEIIAVIOR IN NORTHERN ELEPHANT SEAL 5

MALE FEMALE

Direct approach ------^Avoids by moving away neck rest \ Avoids by croaking,

flipper over back sandflipping, and neck grab swinging perineum weight on ventrum back and forth horizontally

Lies still

Shifts weight.to one side Makes genital contact

Arches perineum and spread hindflippers

Pelvic thrusts

Lumbar Flexure Insertion

Ejaculation - Withdraws Moves away Fig. 2. The behavioral sequence leading to copulationin elephant seals. the female's vaginal opening (Fig. ic). Once in this position, the male probes and thrusts forcefully with only the distal end of the penis extruded. The penis may slide over and under the vaginal orifice several times before insertion occurs. Insertion is signalled by an especially deep pelvic thrust which is held, resulting in an externally visible flexure of the lumbar region of the spine (Fig. Id, ie). Full tumescence of the penis occurs with intromission. The male may pull the female toward him at approximately one half minute intervals while at the same time pushing his pelvis forward. The female may elevate her perineum a few inches and move it slowly back and forth.

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions 6 BURNEY J. LEBOEUF

Otherwise there is little activity during coitus and both animals remain quiet until disengagement. Ejaculation occurs at some time during this period. Copulation is terminated by one member of the pair withdrawing a few feet from the other, probably as detumescence occurs, or by the intrusion of another bull (Fig. If). The mating sequence is summarized in Fig. 2. Unlike in Otariids, intromission in elephant seals is easy to recognize. As described above, its presence and duration is indicated by LF. Two other changes in the configuration of the male's body indicate insertion and make it easy for an observer to confirm it from virtually any angle: (I) wrinkling of the skin on the abdomen and side into ring-like folds due to the sharp bending of the body (Fig. Id, Ie), and (2) invagination of the sacral region of the spine and rudimentary tail (Fig. Id). If these signs are present, the observer can be confident that the male is inserted and that copulation is complete if it continues without interruption. In this study, it was not possible to determine whether ejaculation is abrupt and occurs at the beginning or end of the period of insertion or whether it is continuous. On several occasions fluid was seen seeping from the vaginal orifice shortly after insertion had occurred.

Social organization, population dynamics and sexual activity. Although the entire elephant seal population on Afio Nuevo was studied and almost I,ooo copulations were observed in three breeding seasons, the analysis which follows is restricted primarily to data collected on Area 17, on the largest breeding beach on the island. Adult bulls are extremely aggressive when they arrive on the island in early December, spending much of their waking hours threatening, fighting and chasing each other. A social hierarchy is developed early in the season and bulls arriving later must compete in the established "peck-order". Aggression causes spacing out of males on the beach. If a male dominates another, he keeps him 5 to 7 meters away; likewise, a subordinate avoids one who is dominant to him by constantly moving away a safe distance. The social hierarchy, including the aggression between males that it involves, is maintained until the second week in March. When females arrive in late December, they settle on the beaches that are already frequented by bulls. Males achieve proximity to females in relation to their rank; the higher a male's social rank, the closer his contact with females. The highest ranking males remain on the breeding beach continuously until mid-March (LE BOEUF& PETERSON, I969a; LE BOEUF,I970a).

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions SEXUAL BEHAVIOR IN NORTHERN ELEPHANT SEAL 7

200-

180-

II \ \ . ....**""-...... I 160-

140- i / *./ \ - 120 \FEMALES o) JUVENILES \

80- \ LL / \ MALES /\ . /_ ' '^ \ PUPS 40- I\

NOV DEC JAN FEB MAR APR 1967 1968 Fig.season~~~~~~~~~~~~~~~~~~~~~~~.i 3. A schematicdiagram showing the numberof animalsin variousage classeson Aino Nuevo Island during the 1968 breeding season.

Unlike males, females aggregate in a tight pod or harem 1) in the center of moeta/obethe sandy beach.n2yastm.Te Their number peaksnraei between h26 Januaryubro andae Ist ofa les ratsotatte atoofmxiumnmbr"ffeaestomxiu Februarynumber~~~ofmlsicesddrn? and then gradually declines untilh allosctv of them haveer:281 departed by mid-.:i March.and4~~~~~~~.,:.,. Figure, 3 summarizes changes in numbers of animals from various age classes present on the entire island during the course of the I968 breeding season. Table I shows the number of animals that were present on Area 17 during the breeding seasons observed, the beginning and end of each mating period, and number of copulations observed, and their location and completeness. It is evident that the population on this beach is increasing at a rapid rate and reflects changes in the entire island population. The total number of females more than doubled in 2 years time. The increase in the number of males was less great so that the ratio of maximum number of females to maximum number of males increased during the 3 consecutive years: 2.8:I, 3.7:I, and 4.6: I.

I) The term, "harem"(Laws, I956a, p. 82), refers to a social unit comprisedof a few maturebulls, one of which is dominant,and a variable numberof cows from two to about 200. I do not imply that one male keeps a fixed group of females nor that he herds them.

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions 8 BURNEY J. LEBOEUF

TABLE I

Populationnumber and copulationfrequency on area I7, Ano Nuevo Island

I968 I969 I970 Total no. females I112 199 277 Maximum females per day 102 I70 256

Total no. males 7I 88 104 Maximum males per day 36 46 56 Total no. males copulatinga Io (I4) I6 (i8) 28 (27)

Date of ISt copulation 22 Jan I6 Jan 22 Jan Date of last copulation 13 Mar 13 Mar iI Mar

Total copulations 152 I89 5II Cops. in harem b 130 (86) 148 (78) 46I (90) Cops. on peripheryb I6 (io) 29 (I5) 7 (I) Cops. in waterb 6 (4) I2 (6) 43 (8)

Disrupted copulations b 30 (20) 31 (I6) 136 (27) By other malesb 25 (i6) I5 (8) II8 (23) By self b 5 (3) i6 (8) I8 (4) a Figure in parentheses is percentage of total no. males per season. b Figure in parentheses is percentage of total copulations per season.

Each year only a single harem of females formed (Fig. 4) except for late in the I970 season when the aggregation separated into two parts. During each season there was a single beachmaster or alpha male who dominated all other males and who moved freely anywhere on the beach and paired with females in any location. When the harem split late in I970, the second ranking male took possession of the smaller segment of females. Even at this time the beachmaster came over occasionally, chased this male away, and mated with receptive females. Below the alpha male, entrance into the harem was determined strictly by social rank and the size of the harem. During the beginning and end of each season when there were 40 females or less, the beachmaster was able to keep other males out. As the size of the harem increased over this number, males ranking just below him joined the females on the periphery of the harem. Although the alpha bull could move any of these bulls out at anytime, when he chased one bull out, another would enter from the opposite direction. Eventually the alpha bull settled down in the center of the female aggregation where he had access to the majority of females. Several other high ranking males could usually be found in or near the pod of females when it contained as many as 250 females. Lower ranking males surrounded the periphery or patrolled the shallow water offshore.

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions SEXUAL BEHAVIOR IN NORTHERN ELEPHANT SEAL 9

Table I shows that the percentage of males copulating in a season rose each year. This was not only due to the increase in the number of females, which made it difficult for the beachmaster to defend the harem, but also to individual differences in alpha bulls from one year to the next. Some beachmasters were more vigorous in keeping males away from females than others. This is treated more fully in a later section. Much of the increase in number of males copulating in 1970 was due to the participation of sub-adult males. With the small number of females present in I968, only fully grown males were able to breed. The number of copulations observed increased dramatically from I968 to I970. The majority of sexual activity, 78% to 90% of total copulations, took place within the harem. The remainder of the copulations observed took place on the periphery of the harem, defined as o1 meters away in any direction or in the water just offshore. Copulations were interrupted frequently while in progress; the percentage of total copulations interrupted was lowest in I968 (I6%) and highest in I970 (27%). Interruptions were due to the advance of threatening males or less frequently, were self-induced as when a male broke off copulating in order to prevent a lower ranking male from accosting a female.

Time of copulation and general activity. Mating occurred both during the day and at night and reflected closely the general activity of the animals in the rookery. Unlike some Otariids (e.g., the California sea lion, Zalophus californianus) which tend to be relatively active by day and quiescent at night, elephant seals showed no clear diurnal cycle. At any time, the action of a few individuals could spread and facilitate the activity of others. A typical example of this "domino effect" is what happened when a sub-adult male entered the harem on a warm day when the majority of females and high ranking bulls were sleeping. Usually, the third- ranking bull issued a vocal threat which sent the sub-adult scampering away, sometimes over and through a segment of females and pups. The startled females began to vocalize, alerting other animals nearby. The second ranking bull or the alpha bull returned the threat-call of the number three bull causing him to retreat and shift his position. If the latter didn't move, he was attacked by the higher ranking male who moved toward him on a direct line crashing through females and pups in his path. Aggressive encounters usually continued for several hours causing numerous shifts in the relative position of males. During this time several males usually copulated. Mating activity must be considered against this acyclic background of general arousal. It was evident that some environmental factors such as high temperatures

Behaviour XLI 2

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions 10 BURNEY J. LEBOEUF

(particularly solar radiation), wind, and possibly darkness, reduced general activity and copulation frequency but these variables were not measured closely and therefore cannot be related systematically to copulatory behavior. Observations on moonlit nights revealed that copulations took place with the same frequency as during the day. However, on dark nights when observation was impossible, the noise level of the rookery was considerably lower than that heard during the day, thus indicating less general activity. But copulating animals are not very vocal unless the female is unreceptive, so it is entirely possible that low ranking males infiltrated into the harem at night without being discovered. Until further information becomes available, this question of sexual activity on dark nights will remain unanswered. It is reasonable to assume that sexual activity occurred as readily at night as during the day and that the participants were the same. 3.00

i_0 0--.1969 igure5 shows copulationfrequency as a functionof time of day (700 to 7 8 9 10 11 12 13 14 15 16 17 18 19 TIME OF DAY(HOURS) Fig. 5. Copulationfrequency as a function of time of day.

I900 hours) for the three breeding seasons studied. The pattern from year to year was very similar; three peaks per day are evident each year. By pooling all observations it appears that copulations were most frequent in the morning between ogoo and I Ioo hours. Each year low activity was observed between IIoo and 1200 hours. A second peak of sexual activity occurred between 1200 and 1400 hours, followed by another low period between 1400 and I6oo hours. Copulatory frequency increased again towards sunset, particularly between I6oo and I700 hours. During each breeding season copulatory activity was highest between the I4th and 22nd of February. As many as 6I copulations were observed between 0800 and I800 on 14 February 1970. Behavior of male. Within the harem, copulator activityy of males was directly related to

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions SEXUAL BEHAVIOR IN NORTHERN ELEPHANT SEAL II social rank; bulls with high social rank did most of the breeding and the higher a male's rank, the greater the percentage of copulations he engaged in (LE BOEUF & PETERSON, I969a; LE BOEUF, I97oa). Table 2 shows that bulls at the top four social positions alone accounted for 87, 89, and 62% of the breeding during the three seasons studied. The relationship between social rank and percentage of total matings among the four ranks each year was virtually perfect. Focusing on the individual, it is important to determine the length of time in years that a male succeeds in breeding because this bears on the question of non-random mating and selection. An indication of the degree of cor- respondence in mating from one season to the next can be gained from an examination of the data in Table 3. Seven males accounted for 97% of the

TABLE 2 The percentage of matings per season by bulls at the four highest social ranks on Area 17 at An~oNuevo Island

Social Rank I968 1969 1970 I 40 49 28 2 27 23 i8 3 13 I3 8 4 7 4 8

Totals 87 89 62

TABLE 3 Social rank and breeding success among the 7 highest ranking males on Area 17 in I968 and their performance in 2 subsequent years

I968 I969 I970 Males Social* % Social % Social % Rank matings Rank matings Rank matings GL I 37

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions 12 BURNEY J. LEBOEUF

breeding in 1968. Five of these males continued to do 8o% of the mating in I969 and four of them were still doing 40% of the mating in I970. The top two males in I968 fell in rank and showed a greatly reduced share of the breeding during the last two years. GLS, ranked third the first year, rose to the alpha position in I969 and his breeding success increased accordingly. Although he continued to hold the alpha position for the first 6 weeks of the 1970 season, he was deposed when the females started to come into estrus and plummeted down to the ninth position; as a result of his low rank during the breeding period, he copulated infrequently. PIN improved his position the second year by competing on the other side of the island where he became the alpha bull of the small harem and accounted for Ioo% of the copulations observed. He returned to the large harem in I970 with about the same rank he had in I968, but did not breed. BO's development is most interesting since he progressed from seventh to second and finally to first; his percentage of the total matings shows a similar increase over the 3 years. Evidently, males do contribute to the population gene pool for more than one season but the exact length of time and the full genetic import of this contribution remains to be determined. Although the high ranking males did most of their breeding in the harem (Table 4) they also influenced copulatory activity outside the harem as well. The second, third- and fourth-ranking bulls accounted for 50% of the non- harem copulations in I968, 24% in 1969, and 32% in I970.

TABLE 4 Percentage of copulations in and out of the harem by the four highest ranking males Location 1968 1969 I970

In harem 91 92 95 Out of harem(Io metersor more) 9 8 5 Males with social ranks of five or less accounted for the remainder of the copulations outside the harem. They attempted to mate with females on the periphery of the harem, particularly those separated by several meters from the others, or they waited on the water's edge and intercepted departing females as they entered the water (females are still in estrus when they leave the rookery). The latter were predominately sub-adult males. If any male was successful in mating with the departing female, it was the highest ranking male that gave pursuit. High ranking males only left the harem at the beginning and end of the season when there were few females available and they could not secure females in the harem.

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions SEXUAI, BEHAVIOR IN NORTHERN ELEPHANT SEAL 13

Competition among sub-adults for departing females was particularly intense and sometimes conventional social patterns broke down. By the time a female reached the surf zone, as many as ten sub-adults might surround her (Fig. 6). The first male to reach her mounted but was usually displaced quickly by a higher ranking individual. Perhaps several sub-adults mounted her before she either escaped or was mounted finally by the highest ranking individual around. Even if he achieved insertion, the other males remained near forming a tight circle around the copulating pair. The distance between these males was closer in this context than one sees in similar situations on land. As the males pressed in close, the copulating male usually lifted hi:s head from time to time to issue a threat-call, at which time, the female often eluded him. On a few occasions one of the sub-adult spectators bit the copulating male on the back and then swam away quickly to avoid the attack of the male whose copulation he had interrupted. Duration of copulation. From the time a male put his foreflipper over the back of a receptive female it usually took him about three minutes to obtain insertion (x = 171 sec., s = I31 sec., N = 88). Intromission, the duration of LF, lasted = = about five minutes when it occurred on land (x 293 sec., s I30 sec., N = IOI). In general, copulations in the water lasted longer and were more variable than those on land (x = 370 sec., s = 221 sec., N = 30). Low ranking males tried to hold females and copulate with them several times; some of these matings lasted as long as 20 minutes. In those cases where a high ranking male pursued a departing female and copulated with her in the water, the duration of copulation did not differ from the time it took on land. The longer duration of copulations in the water may have been due to the difference in the substrate, the greater number of young bulls copulating or the fact that many males copulating in this situation had low ranks and therefore, copulated only infrequently. It is difficult to say which variable was most important. An analysis of the first copulations of 12 males, each of whom copulated no more than 3 times in a single season, revealed longer copulation durations than the overall mean (x = 403 sec., s = 84 sec.). Even the duration of initial copulations by the highest ranking bulls was longer than the overal mean given in the first paragraph of this section = = (x = 372 sec., s 92 sec., N 7). Figure 7 shows changes in the duration of copulations, with changes in the frequency of copulations by the alpha male, BO, during the 1970 season. In general, copulation duration was longest during the first and last Io-day periods, when copulation frequency was lowest. The Pearson product correla-

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions 14 BURNEY J. LEBOEUF

400 375 350 325 ,300

. 275- - 1250 225 -

200 - 175

150

w 5

25 30 I 5 10 15 20 25 1 5 10 January February March Fig. 7. Changes in the duration (upper) and frequency (lower) of copulation for BO, the alpha male during the I970 season. The upper figure shows means and standard deviations for Io-day intervals (vertical lines). tion coefficient for copulation frequency and copulation duration is -.80 (p <.05, F = 5.33)-

Termination of copulation. The manner is which copulations were terminated was recorded in over 200 cases. Eighty-eight percent of the time, the male made the first move away from the female. The remainder of the time, either the female withdrew first or both animals parted simultaneously. It appeared as though detumescence occurred and then it was simply a matter of who moved first. Usually, it was the male who simply shifted from his side to lying down on his belly.

Frequency of copulation. Some beachmasters copulated almost every day during the mating period, as Fig. 7 shows for the alpha male in I970. This male copulated as much as nine times in one day. On days in which no copulations were recorded, observations were not made or were limited in time. In 1969, one male copulated i3 times on the periphery of the harem between 0922 and 1450 hours; all copulations were with the same female! In the same year, the second ranking male, CLS, copulated io times in o1 hours. In males of several species, e.g., turkey (SCHEIN & HALE, 1965), rat

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions SEXUAI, BEHAVIOR IN NORTHERN ELEPHANT SEAL 15

(LARSSON, 1956), and ,sheep (BERMANT, CLEGG & BEAMER, I969), there is a period between successive copulations during which the male will not copulate. The interval is shortest when the male copulates with a different female each time. One would expect this interval, often called the refractory period, to be small in seals because of the demands of the harem system and the number of different females available at one time.

TABLE 5 The shortest intervals between successive copulations by three different alpha bulls in three breeding seasons (to closest minute)

I968 I969 1970

5 2 7 5 II 8 8 12 9 25 I5 o0 32 I8 10 34 27 II 36 30 I5 38 34 17

Table 5 shows the eight shortest intervals recorded between two complete copulations for the three different alpha males in each year studied. The intervals extend from termination of the first copulation to the start of the second. Alpha males were selected because they, alone, are able to copulate at will. It is evident that a second copulation within 5-30 minutes after the temination of the first was not unusual. The interval between the termination of a disrupted copulation and a subsequent copulation was even shorter; often as brief as 30 seconds, depending on the duration of insertion before the first copulation was disrupted.

Choice of "female". Only breeding males mounted during the breeding season and the recipients were adult females and less frequently, yearlings and weaned pups. Males never mounted other males nor did females mount males or other females. The cues males used to determine estrus in females could not be determined. Bulls did not investigate females prior to mounting. Yet, it was clear that males knew when a particular female was in estrus for she was mounted far more frequently than others in non-estrus. However, the latter were mounted too, by the youngest as well as the oldest breeding males. This was particularly noticeable in January before the first females came into estrus. Over Ioo

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions I6 BURNEY J. LEBOEUF

mounts were recorded between Io December and 14 January 1968, 7 days before the first copulation was observed. Mounting appeared to be indis- criminate; if a female was accessible, she was mounted. When she resisted vigorously, the male eventually gave up and moved to another female. A female separated from the others, e.g., one arriving from the sea, was a particularly attractive stimulus who was usually hounded until she joined the other females in the harem. Perhaps one reason why females aggregate is to avoid harassment from amorous males. Newly-arrived pregnant females and females that had recently given birth were mounted even though very unwilling and unreceptive. Both suffered rough treatment from courting bulls and they sought to avoid this type of contact. During the early part of the season, juveniles (predominately yearlings) were found on or near the breeding areas (Fig. 3). Later in the season, weaned pups, who are actually heavier than the yearlings, were more numerous than breeding females. Both categories of young animals were mounted by subadult and fully adult males; this was observed 50 times in I968. They resisted in the same manner as unreceptive females (Fig. 2), but the mounting male could be aggressive and tenacious of purpose. The neck grab often drew blood and in a few cases, the skulls of newly weaned pups were fractured. As in females, the isolated pup was most likely to elicit mounting. Perhaps in order to avoid being mounted, both yearlings and weaners separated from the female harem and grouped together in their own pods (females chased weaners out of the harem to prevent them from suckling). Banding together also seemed to be important for thermoregulation, for pups spaced out on warm days and crowded together when the weather was cold. Mounting among the developing young was never observed on land and only rarely in the water. This activity is common in the young of many ter- restrial mammals (e.g., BEACH, ROGERS & LE BOEUF, 1968; LE BOEUF, I97ob). All mounting by males, whether directed to receptive or nonreceptive females, to juveniles or weaned pups, was correctly oriented. Aberrant head and sidemounts, seen so frequently in laboratory rats and domestic dogs, never occurred.

Disrupted mounts (DM's) and Disrupted Copulations (D C's). Males spent most of their active hours trying to copulate or preventing others from doing so; these activities were mediated by threat and fighting. Each male attempted to prevent those subordinate to him from copulating.

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions SEXUAI BEHAVIORIN NORTHERNELEPHANT SEAL 17

As soon as a low-rankingmale mounteda female, a dominantmale in the vicinity usually threatenedhim with a stereotypedvocalization or attacked him directly (LE BOEUF& PETERSON,I969b); the subordinate animal always moved away, althoughsometimes not fast enough to avoid being bitten, and the mount was disrupted.It is worth noting that mounterswere disrupted whetherthe femalemountee was in estrus or not. Often, the subordinatewas not discovereduntil he had obtainedinsertion. In this case, threator attackforced the male to retreatand the copulationwas disrupted.Normally a threatenedmale was so effectively intimidatedthat he didn't take the time to withdrawhis penis before moving away; he rushed directly away from the threatenor,sometimes swinging the female around before the penis disengaged.Usually, the dominantmale copulatedwith the female immediately. Only a few males interferedwith the matingof others and they were the maleswith the highestsocial ranks.In I968, the top four maleswere respon- sible for all DC's and 92% of the DM's observed.Of I,ooo mountsattempted during that season, 43% were disrupted.Below the rank of six, 74% of mountsattempted were disrupted.It shouldbe clearthat althoughmany males got the chanceto mount females, few of them escapeddetection and interference, and thus few managedto copulate. The higher the rank of a male, the more frequentlyhe disruptedmounts and copulationsand the less frequently he was treated this way by other males (see Table 2 in LE BOEUF & PETERSON I969a). In I969, the alpha malewas responsiblefor six out of nine DC's; he himself,was never bothered while in copula.The most completerecord of DC's was obtainedin I970 and will be exploredin more detail. In I970, the relationshipbetween social rank and frequencyof disrupting copulationswas again high and positiveas evidencedby the data in Table 6. TABLE 6 Social rank and disruptedcopulations

Social No. copulations No. males No. copulations Males rank disrupted disrupted disrupted by BO

BO I 62 13 NOT 2.5 15 Io 9 2STP 2.5 TI 9 I3 HOL 4 4 3 II GA 5 3 3 7 RDR 6 5 5 5 OB 7 2 2 5

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions I8 BURNEY J. LEBOEUF

The social rank of an individual was determined by the number and identity of animals from which he retreated in aggressive encounters (LE BOEUF & PETERSON, I969a). Males disrupted copulations of those immediately below them in rank most frequently. The last column in Table 6 shows that the alpha bull, BO, interrupted bulls with high ranks more frequently than others. He never disrupted any male below the rank of seven more than 3 times. This relationship was in part due to the fact that bulls of near equal rank were close to each other on the beach. The males that were most frequently disrupted had social ranks of 2 or 3; they were in a position to attempt more copulations than males below them in rank, but they were doing this in close proximity to the beachmaster, who worked hard at preventing this activity. The alpha male was never interrupted by other males while copulating. The duration of DC's varied from a few seconds to as much as eight minutes. For 104 DC's the mean duration was approximately one-half as long as the mean duration of uninterrupted copulations (x = 138 sec., s 121 sec.). Eighty-six percent of the copulations were disrupted before they had gone on for three minutes.

Self-interrupted copulations (SIC's). Some males terminated their own copulations prematurely in order to prevent other males from copulating. Only the top 3 bulls did this regularly. Once again, the higher the male's rank, the more frequently he interrupted his own copulation to prevent the copulation of another. There were 12 SIC's in I969; io by the alpha male and the rest were divided equally among the number 2 male and the number 3 male. The alpha bull in I970 accounted for 9 SIC's out of I8 observed; the beta male had 5. SIC's occurred from io seconds to 4Y2 minutes after insertion took place. The mean duration of these interrupted copulations was almost identical to that of DC's (x 3= 35 sec., = 73 sec., N = 39). Males returned to the same female within a few minutes and copulated again in 22% of the cases observed. Whenever this was done, the pre-interruption copulation duration was less than one minute. SIC's were easy to distinguish from the male terminated, uninterrupted copulations mentioned in a previous section. In the former, the male usually reared up on his forequarters and threatened an intruding male vocally while still engaged in copulation. When the intruding male failed to retreat, the dominant bull withdrew his erect penis from the female and directly attacked the intruder. In male terminated copulations, the male moved away from the female as detumescence occurred in order to adjust his posture; this was usually after being inserted for about 5 minutes.

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions SEXUAL BEHAVIOR IN NORTHERN ELEPHANT SEAL i9

With so much disruption of mating occurring, one may wonder about the conditions that prevailed when a male other than the alpha male initiated copulation. In almost every case where the start of copulation was observed, males dominant to the one attempting copulation (particularly the alpha male) were busy fighting, copulating or they were asleep or facing the opposite direction. It was a rare occasion when a dominant bull "saw" a subordinate attempt copulating without attempting to prevent it.

Behavior of the female. Table 7 summarizes some of the temporal aspects of reproductive behavior in females. Eighty five percent or more of the females that landed on Afio

TABLE 7

Reproductive data collected on Mirounga females in three breeding seasons at Ano Nuevo Island I968, I969 and I970 (in days)

Mean & N Stand.Dev. median mode range

Arrival to parturition I6 6.6 7 7 2-I6 +2.8 Parturition to ISt cop. 15 23.6 24.5 24 17-30 +3.2 Parturitionto departure 21 27.5 27 27 I9-34 (nursingperiod) -+3.5 Arrival to departure 21 34.3 33 31 28-40 +3.0

Nuevo Island during the breeding season were pregnant. The majority arrived at night and were found in the harem in the morning. Those observed "hauling out" moved directly towards other females. Females delivered a single pup on about the seventh day after arrival and it was nursed daily until the female left the rookery, about 27 days after the pup was born. Most females were in estrus during the last 3-4 days they spent on land. But there are difficulties in determining the length of estrus with precision because marked females were sometimes hidden from view and it was necessary to observe these individuals continuously in order to document copulations and the length of time the female was receptive. Obviously, copulations occurred when there were no observers in the blinds and consequently the figures derived fall short of the true figures. Therefore, the figures presented below can only be considered minimal estimates.

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions 20 BURNEY J. LEBOEUF

If a female's estrous period is defined as the interval between her first and last copulation, the data collected in three seasons indicate that in one estrous period, females may: (a) be in estrus for as long as 13 days. (b) copulate at least once a day on as many as 5 separate days. (c) copulate as many as 13 times a day. (d) copulate with as many as 4 different males. (e) copulate as often as I8 times throughout the entire period. The above figures are the highest that were observed during three breeding seasons. Most marked females were receptive for about 3 days and were observed to copulate on 3 separate days. Four to five copulations in one day was very common and about 50% of the females copulated with at least two different males. Like males, some females copulated a number of times in succession with only a few minutes in between. Several copulated a second time within three minutes after the first copulation had ended. One female copulated 4 times in one hour and 26 minutes and another female copulated three times within the space of 30 minutes.

Observations on other rookeries. Copulatory behavior per se did not differ on other elephant seal rookeries despite gross differences in population density and terrain. However, the latter influenced male social behavior and harem structure. In general, breeding females grouped in the following places: 5-30 meters inland on broad sandy or coblestone beaches, near the water's edge along narrow beaches, and a few meters inland in sheltered coves. All of these breeding locations were used at Isla de Guadalupe, the largest of northern elephant seal rookeries with a population of approximately I5,000 animals, but most breeding took place on large sandy beaches. At Pilot Rock beach, where over 4,000 animals were counted during February, 1970, the population density was so high that harems were confluent and inter- individual distance between bulls was noticeably smaller than at any other rookery (Fig. 8). The frequency of fights and threat behavior seemed unusually high and females were constantly harassed by subadult and subordinate males who could not be driven out by the dominant males. DM's and DC's appeared to be more common than at Afio Nuevo Island but quantitative data was not obtained. Like on other rookeries, bulls frequently trampled pups and this undoubtedly contributed to the extensive pup mortality observed. However, pup mortality was not restricted to newborns less than a week old

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions SEXUAL BEHAVIOR IN NORTHERN ELEPHANT SEAL 21 but also included I-2 week old pups that had suckled, grown and appeared otherwise healthy. Deaths among the latter were observed rarely on less crowded rookeries. At San Miguel Island, where approximately 3,000 animals were counted in February, I968, virtually all breeding occurred on the expansive sandy beaches at the western tip of the island (Fig. 9). Space was abundant and some harems, located just above the surf line or as much as IOOmeters inland, were contiguous while others were separated by hundreds of meters. The social organization of males with respect to harem locations was similar to the situation described for the large harem at Afio Nuevo Island. Males could approach females from all directions and the degree of harem infiltra- tion was primarily a function of the number of females the harem contained. Most of the breeding at Islas San Benito took place in sheltered coves. High banks restricted access to these beaches to the sea and made it easy for a single bull to hold as many as 50 females while effectively keeping out male intruders by patrolling the shallow water at the mouth of the cove (Fig. IO). This type of terrain and harem location made it very difficult for subadult males to breed.

DISCUSSION There are obvious similarities and striking differences in the copulatory pattern of the elephant seal, a member of the family, Phocidae, and other pinnipeds, particularly sea lions in the family, Otariidae. The frequency and duration of copulation (of males) is similar as well as the location which is primarily on land. One class of differences has to do with the form of copulation and manner in which it is initiated and terminated. Otariid males mount the female more canenm (Fig. IIa, i2b); the Northern elephant seal male lies on his side like two other phocids, the Southern elephant seal (LAws, I956), and the Grey seal (HEWER, I960), and the walrus (EDWARD ASPER, personal communication). Sea lion males, such as the California sea lion (PETERSON & BARTHOLOMEW, I967), the Steller sea lion, Eumetopias jubata (GENTRY, I970; SANDEGREN, I970), the Alaska fur seal (BARTHO- LOMEW, I953), and the Cape fur seal, Arctocephalus pusillus, (RAND, I955), the Kerguelen fur seal, A. tropicalis (PAULIAN, I964), and other Antarctic fur seals, typically nuzzle the anogenital region and sometimes the head of the female prior to mounting (fig. I2a). Presumably, olfactory and gustatory cues obtained from this investigation aid in assessing the estrous condition of the female. Non-estrous females are evasive and males rarely succeed in mounting them. Elephant seals mount females without preliminaries, regard- less of their estrous condition.

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions 22 BURNEY J. LEBOEUF

Female Steller sea lions (GENTRY, I970; SANDEGREN, I970) and Alaska fur seals (PETERSON,1968) often initiate sexual encounters by approaching the male and nipping at his head and neck; Northern elephant seal females never solicit coitus, although LAWS (I956a) suggests its occurrence in the southern species. In the elephant seal, copulation is terminated by the male while in Otariid species, females usually terminate copulation. GENTRY(1970) reports that femal Stellers terminated 92% of the copulations he observed. The female raises her head, turns and bites the male on his chest and neck, and attempts to struggle free (Fig. lib). The male is dislodged or dismounts shortly after. Similar behavior has been observed in several species of fur seals, e.g., RAND (I955). The most significant difference in the behavior of elephant seals and Otariid species has to do with the response of a male witnessing another copulate. As a rule, elephant seal bulls disrupt matings and sea lion males do not. These differences emanate from their entirely different social systems. In the social hierarchy of elephant seals, competition is between individual males and is not specific to place. Social rank is all important and only the top ranking male, since he is avoided by all others, can mate without being interrupted. All other males have someone to avoid who can displace them from an estrous female at any time. Otariid males set up territories which they defend against other males. The borders are "marked" by aggressive threat displays directed toward neighbors for the purpose of keeping them out. When females come into estrus, the territorial bull copulates on his territory in full view of adjacent territorial bulls. The latter never disturb him (RAND, I967; GENTRY, 1970) although in the Alaska fur seal, they often take advantage of his lapse of territorial and harem maintenance (herding of cows) by capturing some of his females (BARTHOLOMEW,I953). The differences which exist in the copulatory behavior of female elephant seals and Otariid females may have evolved as a result of differences in the social organization of males. Female elephant seals of both species copulate several different times over a period of several days. Since many of their matings are interrupted and may be incomplete due to inter-male competition, multiple copulations may be a necessary compensation to insure insemination and pregnancy. On the other hand, the completeness of copulation in Otariids, made possible by the privacy insured by the territorial system, might be expected to require fewer copulations of the female to insure pregnancy. This appears to be the case. BARTHOLOMEW& HOEL (1953) report that Alaska fur seal cows ..."are in estrus only for a few hours or at most for a day, and ordinarily they copulate but once." In the same species, PETERSON

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions SEXUAL BEHAVIOR IN NORTHERN ELEPHANT SEAL 23

(I968) found that the mean number of copulations per female was 1.2. In 1500 observation hours covering two breeding seasons, SANDEGREN (I970) never saw a Steller female copulate more than once, even though some females solicited males after the initial mating. GENTRY(1970) reports that 85% of of the female Stellers in his study area copulated only one time; the remainder copulated twice but no more. Although it is impossible to determine the course that evolution took which resulted in the current difference in female sexual physiology and behavior, it is probable that differences in male social behavior was an important selection factor. Except for the excellent series of papers on the Southern elephant seal by LAWS (I953, I956a, I956b) and CARRICK and his colleagues (e.g., CARRICK et al, 1962), and HEWER'S observations of the Grey seal (HEWER, I957; I960), adequate descriptions of the sexual pattern in phocids do not exist, thus intra-family comparisons with the Northern elephant seal are severely limited. In most cases, this is not the fault of the investigator so much as the secretive habits and inaccessibility of the animals themselves. In general, M. angustirostris and M. leonina are remarkably similar; the differences are quantitative rather than qualitative. Most apparent is the shift in time of breeding and the shorter period between arrival and parturition, and between parturition and copulation for Southern females. The Southern elephant seal breeds in October and November, gives birth approximately 5 days after it arrives on the rookery, and copulates approximately I8 days after parturition (LAWS, I956a). Southern females may be in estrus longer than Northern females and M. leonina males do not interrupt their copulations prematurely to attack intruding males (LATS, I956a). Copulation in the Grey seal occurs around I6 days after parturition either on land or in the water with a definite preference for the latter. During copulation on land, the male lies on this side, places his foreflipper over the female's back and may exhibit a neck grab; copulation lasts from 15 to 45 minutes. Copulation between the same bull and cow may occur several times and bulls copulate with several cows. Males attempt to mount females that are not in estrus (HEWER, I957; I960). In general, these aspects of coital behavior are rather similar to corresponding behaviors in elephant seals. However, more detailed observations on Grey seals are needed for a thorough comparison of the two species. Harbor seals (SCHEFFER & SLIPP, I944), Monk seals (KENYON & RICE, 1959), and Weddell seals (RAY & DE CAMP, I969; CLINE, SINIFF & ERICK- SON, I97I) appear to copulate almost exclusively in the water and therefore little is known about their behavior.

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions 24 BURNEY J. LEBOEUF

SUMMARY

The mating behavior of Northern elephant seals, Mirounga angustirostris,was studied during the 1968, I969 and I970 breeding seasons at Afio Nuevo Island, 19 miles north of Santa Cruz, California. Copulationtakes place primarily on land, from January to March; it is initiated and terminatedby the male and lasts approximately5 minutes. A few males do most of the breeding and the higher a male's rank in the social hierarchy, the more frequently he copulates. Some bulls may maintain high rank and participate in mating for 3 breeding seasons. Low ranking males are kept out of the harem so they attempt to copulate with females on the periphery or in the water with departing females. Males prevent sub- ordinates from mounting females and disrupt copulations in progress. The higher a male's rank, the more freedom he has to copulate without interference, and the more frequently he interferes with the copulationof others. The highest ranking males interrupt their own copulations prematurely to attack other males and prevent them from copulating. Males mount weaned pups, yearlings, and non-estrous females in addition to estrous females; the latter are mounted most frequently. Females come into estrus 24 days after parturition.They are receptive for about 3 days (range I-1I3 days), during which they may copulate several different times with one or more males. Females are usually in estrus when they leave the rookery and return to sea. The greatest difference in the copulatorypattern of the elephant seal, a true seal, and membersof the sea lion family, is that the former interfere with the copulationsof other males while the latter do not. This may be due to differing social systems, a social hierarchy in elephant seals and a territorial system in sea lions. In elephant seals, where copulations are interrupted,mating may be incomplete. This is not as likely with sea lions where copulationsare never interruptedbecause males honor the boundariesof their neighbors. The differences in female behavior and sexual physiology-a long promis- cuous estrous period in elephantseals as opposedto a brief estrus with only one copulation in sea lions - may have evolved as a compensationfor the consequencesof the social behavior of males.

REFERENCES

BARTHOLOMEW,G. A. Jr. (1952). Reproductive and social behavior of the Northern Elephant seal. - Univ. Calif. Pub. Zool. 47, p. 369-47I. - (I953). Behavioral factors effecting social structure in the Alaska fur seal. - Trans. I8th North Amer. Wildl. Conf., p. 481-502. - & HOEL,P. G. (1953). Reproductivebehavior of the Alaska fur seal, Callorhinus ursinus. - J. Mammal. 34, p. 417-437. BEACH, F. A., RODGERS,C. M. & LE BOEUF, B. J. (1968). Coital behavior in dogs: Effects of estrogen on mounting by females. - J. Comp. Physiol. Psychol. 66, p. 296-307. BERMANT, G., CLEGG, M. T. & BEAMER, W. (I969). Copulatory behaviour of the ram, Ovis aries. I: A normative study. - Anim. Behav. 17, p. 700-705. BISHOP,R. H. (I967). Reproduction,age determination,and behavior of the Harbor seal, Phoca vitulina L., in the Gulf of Alaska. - M.S. dissertation,Univ. of Alaska, 12I pp. BROOKS, J. W. (1954). A contribution to the life history and ecology of the Pacific walrus. - Alaska Co-operativeWildlife Research Unit. Spec. Rep. I, 103 pp. CAMERON, A. W. (I969). The behavior of adult gray seals (Halichoerus grypus) in the early stages of the breeding season. - CanadianJ. Zool. 47, p. 229-233. CARRICK,R., CSORDAS,S. E. & INGHAM,S. E. (1962). Studies on the Southern elephant seal, Mirounga leonina (L.) IV. Breeding and development.- C.S.I.R.O. Wildlife Research 7, p. I6I-I97.

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions SEXUAL BEHAVIORIN NORTHERNELEPHANT SEAL 25

CLINE,D. R., SINIFF,D. B. & ERICKSON,A. W. (I97I). An observationof underwater copulation for the Weddell seal. J. Mammalogy (in press). GENTRY,R. L. (I970). Social behavior of the Steller sea lion. - Ph.D. dissertation, Univ. of California, Santa Cruz, 113 pp. HEWER,H. R. (I957). A Hebridean breeding colony of Grey seals, Halichoerus grypus (Fab.), with comparative notes on the Grey seals of Ramsey Island, Pembroke- shire. - Proc. Zool. Soc. Lond. 128, p. 23-68. (I960). Behaviour of the Grey seal (Halichoerus grypus Fab.) in the breeding season. - Mammalia24, p. 400-421. KENYON,K. W. & RICE,D. W. (I959). Life history of the Hawaiian monk seal. - Pacific Science 13, p. 215-252. LARSSON, K. (I956). Conditioningand Sexual Behavior in the Male Albino Rat. - Alm- quist & Wiksell, Stockholm. LAWS,R. M. (I953). The elephant seal (Mirounga leonina, Linn.). I. Growth and age.- Falkland Is. Depend. Surv. Sci. Repts 8, p. 1-62. (I956a). The elephant seal (Mirounga leonina, Linn.) II. General, social and reproductivebehaviour. - Falkland Is. Depend. Surv. Sci. Repts 13, p. 1-88. (I956b). The elephant seal (Mirounga leonina, Linn.). III. The physiology of reproduction.- Falkland Is. Depend. Surv. Sci. Repts I5, p. 1-66. LE BOEUF,B. J. (g97oa). Fighting in elephant seals. - Natural History (in press). (I97ob). Copulatory and aggressive behavior in the prepuberallycastrated dog.- Hormones and Behavior I, p. 127-I36. & PETERSON,R. S. (I969a). Social status and mating activity in elephant seals. - Science 163, p. 91-93. & - (I969b). Dialects in elephant seals. - Science I66, p. I654-I656. LOUGHREY,A. G. (I959). Preliminary investigation of the Atlantic Walrus, Odobenus rosnmarusrdsmarus (Linnaeus). - Wildlife Management Bulletin Series I, No. I4, p. 1-123. MANSFIELD,A. W. (I958). The breedingbehaviour and reproductivecycle of the Weddell seal. (Leptonychotesweddelli Lesson). - Falkland Is. Depend. Surv. Sci. Repts I8, p. I-41. ORR,R. T. & POULTER,T. C. (I967). Some observations on reproduction,growth, and social behavior in the Steller sea lion. - Proc. Calif. Acad. Sci. 35, p. 193-226. PAULIAN,P. (I964). Contributiona 1'etudede l'otarie de l'ile Amsterdam.- Mammalia 28, p. I-I46. PETERSON,R. S. (I968). Social behavior in pinnipeds with particular reference to the Northern fur seal. - In HARRISON,R. J. et al., The behavior and physiology of pinnipeds.Appleton-Century-Crofts, New York. p. 3-53. & BARTHOLOMEW,G. A. Jr. (I967). The natural history and behavior of the Cali- fornia sea lion. - Amer. Soc. Mammalogists, Spec. Publ. I, Stillwater, Oklahoma, P. I-79. RAND, R. W. (1955). Reproductionin the female Cape Fur Seal, Arctocephalus pusillus. -- Proc. Zool. Soc. Lond. 124, p. 717-740. (I967). The Cape fur seal (Arctocephalus pusillus) 3. General behaviour on land and at sea. - Investl. Rep. Div. Sea Fish. S. Afr. 60, p. 1-39. RAY,C. & DE CAMP,M. A. (I969). Watching seals at Turtle Rock. - Natural History 128, p. 26-35. SANDEGREN,F. E. (1970). Breeding and maternalbehavior of the Steller sea lion (Eume- topias jubata) in Alaska. - .S. dissertation, Univ. of Alaska, 138 pp. SCHEFFER,V. B. & SLIPP, J. W. (1944). The Harbor seal in Washington State. - Amer. Midl. Naturalist 32, p. 373-416. SCHEIN, M. & HALE, E. (I965). Stimuli eliciting sexual behav- In BEACH, F. A. (Ed.) Sex and Behavior. Wiley, New York, p. 440-482. Behaviour XLI 3

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions 26 BURNEY J. LEBOEUF

VENABLES,U. M. & VENABLES,L. S. V. (1957). Mating behaviour of the seal Phoca vitulina in Shetland. - Proc. Zool. Soc. Lond. 128, p. 387-396. WIRTZ,W. 0. (I968). Reproduction,growth and development,and juvenile mortality in the Hawaiian monk seal. - J. Mammal. 49, p. 229-238.

RPiSUME L'accouplementde l'elephant de mer du Nord, Mirounga angustirostris, a ete etudie pendant les periodes de reproductionde 1968, I969, et 1970 dans l'ile Aiio Nuevo, situee a I9 miles au nord de Santa Cruz, California. L'accouplementprend place principalementsur terre, du mois de Janvier au mois de Mars; il dure a peu pres cinq minutes, et c'est le male qui en prend l'initiative. Un petit nombre de males assume la plus grande part de l'acte reproductif, et plus un male tient un rang eleve dans la hierarchie sociale, plus il s'accouplesouvent. Certains males repro- ducteurs se maintiennent parfois a un rang eleve et prennent part a la reproduction pendanttrois periodes de reproductionsuccessives. Plus un male tient un rang eleve dans la hierarchie sociale, plus il s'accouple souvent. Parce que les males proches du bas de l'echelle sont tenus a l'ecart du harem, ils tentent de s'accoupler soit avec les femelles situees aux alentours du groupe, soit avec celles qui sont deja dans l'eau, sur le depart. Les males empechentleurs inferieurs de couvrir des femelles et interrompentles accou- plements deja en cours. Plus un male est de rang eleve, plus il est libre de s'accoupler sans empechement,et plus il intervient souvent dans les accouplementsd'autrui. Un male de plus haut rang interrompt prematurement sa propre copulation afin d'attaquer et d'empecherd'autres males. Les males couvrent les jeunes une fois sevres, les jeunes d'un an, et les fenelles non en rut aussi hien (ue celles en rut - ces derniires plus fre- quemment. Les femelles sont en rut 24 jours apres la parturition; elles sont receptives pendant a peu pres trois jours (variations de I a I3 jours), pendant lesquels il arrive souvent qu'elles s'accouplent plusieurs fois, avec differents males. Elles sont souvent en rut lorsqu'elles quittent la "rookery"et reprennentle large. La difference principale entre les habitudes d'accouplementd'un elephant de mer (un veritable phoque) et celles d'une otarie reside dans la fait que le premier intervient dans l'activite des autres males. Ceci peut provenir de la difference entre leurs organisations sociales: hierarchie chez l'elephant de mer, organisation territoriale chez l'otarie. Alors que l'accouplementest parfois non acheve (parce qu'interrompu)chez le premier, ceci est tres infrequent chez le second, car le male honore les limites entre territoires. I1 est fort possible que les differences de physiologie et de comportemententre les femelles des deux families (une longue et tres active periode de rut chez les elephants de mer, un rut ne comportant qu'un accouplement chez les otaries) se soient developpees comme mnoyen d'operer un redressement- chez les elephants de mer - des consequencesdu comporte- ment social des males.

This content downloaded on Tue, 19 Feb 2013 19:16:23 PM All use subject to JSTOR Terms and Conditions