Effect of auxiliary males on territory ownership in the oribi and the attributes of multimale groups

ANIMAL BEHAVIOUR, 1999, 57, 61–71 Article No. anbe.1998.0962, available online at http://www.idealibrary.com on

Effect of auxiliary males on territory ownership in the oribi and the attributes of multimale groups PETER ARCESE

Department of Wildlife Ecology, University of Wisconsin (Received 18 August 1997; initial acceptance 10 November 1997; final acceptance 3 June 1998; MS. number: A8005)

I observed free-ranging oribi, Ourebia ourebi, in Serengeti National Park, Tanzania, to determine whether group formation by males provides evidence for cooperative territory defence, a behaviour that is rare among male vertebrates. Socially dominant males that shared territories with subordinate auxiliary males were replaced by rivals less often than males that defended territories without auxiliary males. Auxiliary males marked territories with preorbital glands, dung and urine, and territories defended by male groups were marked more thoroughly than those defended by single males. Fifteen of 24 (62.5%) auxiliary males whose histories were known were born on territories defended by males that probably were their fathers. But 37.5% of auxiliary males probably were unrelated to dominant males, because male groups also formed when territory owners accepted adult male immigrants as subordinates, and when owners allowed young males to remain philopatric after evicting a male likely to have been the young male’s father. All males in groups probably had some mating access to females, but dominant males may have minimized matings of auxiliary males by guarding fertile females. These results suggest that auxiliary male oribi aided dominant males in territory defence, and that dominants traded off the risks of losing matings to auxiliaries, or being overthrown by them, in exchange for a reduction in their chance of being evicted by rival neighbours or immigrants. 

ca. 2500 h of focal animal watches and casual observations to describe how male groups formed, and to compare the behaviour of dominant and subordinate group members with respect to females on their territory, to each other, and to singleton males. I now offer four hypotheses to explain why male oribi share their territories with subordinate auxiliaries, based partly on prior studies of group-living felids (Packer 1986; Caro 1994), primates (Harcourt & de Waal 1992), bovids (Wirtz 1982; Spinage 1982) and birds (Brown 1987; Stacy & Koenig 1990; Emlen 1991).

Cooperative territory defence by male vertebrates is rare, particularly in the absence of reproductive inhibition in subordinates (Packer 1986; Brown 1987; Smith 1990; Emlen 1991; Harcourt & de Waal 1992; Caro 1994). However, male oribi (Ourebia ourebi, Neotraginae) are known to defend territories either as single males (‘singletons’) or as groups comprising a dominant male (‘dominants’) and one to three subordinate, auxiliary males (‘auxiliaries’), in each case with up to six resident adult females (Arcese 1994; Arcese et al. 1995; Brashares & Arcese 1999). Flexibility in the social behaviour of oribi allowed me to study why males form territorial groups, and to ask whether their doing so provides evidence of cooperation in the bovids, where males are much better known for their pugnacity than cooperation (Leuthold 1966, 1977; Estes 1969, 1974, 1991; Jarman 1974, 1979; Gosling 1986a). To address these questions, I used data from annual censuses of over 300 individually identified oribi on 23–29 territories monitored from 1987 to 1993 in Serengeti National Park, Tanzania. I also used records of the behaviour and proximity of oribi made during

Hypotheses and Predictions Ailing male oribi may accept subordinate auxiliaries on their territory as they loose physical ability and can no longer evict persistent intruders completely. Accepting auxiliaries may be preferable to trying to evict them if, by attempting the latter, territory owners risk injury, exhaustion or alerting additional rivals to their poor condition. The ailing male hypothesis predicts that, once on territories, auxiliary males eventually will: (1) force reversals in dominance status or evict original owners; (2) become the new owner should a dominant disappear;

Correspondence: P. Arcese, Department of Wildlife Ecology, 1630 Linden Drive, University of Wisconsin, Madison, WI 53706, U.S.A. (email: [email protected]). 0003–3472/99/010061+11 $30.00/0

1999 The Association for the Study of Animal Behaviour

61



1999 The Association for the Study of Animal Behaviour

62

ANIMAL BEHAVIOUR, 57, 1

(3) challenge dominants for mating access to females; and, overall, that (4) territorial males with auxiliaries will be evicted or disappear from the study area more often than singleton males that are more able to prevent settlement by auxiliary males. An alternative to the ailing male hypothesis is the cooperative defence hypothesis. This hypothesis posits that dominant male oribi allow auxiliaries to reside on territories in exchange for contributions to defence and a consequent decline in the risk of eviction by rival neighbours or immigrants. This hypothesis assumes that auxiliary males participate in defence, and, contrary to the ailing male hypothesis, it predicts that dominant males that reside with auxiliaries will be evicted from territories or disappear from the study area less often than males that defend territories as singletons. Two additional hypotheses address the formation of male groups but make no prediction about the year-toyear status of territory owners with or without male auxiliaries. The parental facilitation hypothesis suggests that male groups form when oribi fathers allow their sons to remain philopatric as an extension of parental care. Philopatry might be expected if it facilitates the ability of sons to inherit natal territories or acquire adjacent ones. This hypothesis assumes that auxiliary males are the sons of dominant males on whose territory they reside and that auxiliaries often gain territories by inheriting them or displacing neighbours. It also predicts that an auxiliary will be evicted from a territory if its father is evicted. Finally, the female defence hypothesis suggests that male oribi form groups to increase their per capita access to females. This hypothesis assumes that all adult males on territories have some mating access to resident females, and it predicts that the female:male ratio on territories defended by male groups exceeds that on territories defended by singletons

METHODS

Study Species and Area Oribi are a small grazing antelope (ca. 18 kg; Hofman & Steward 1972) distributed across sub-Saharan grasslands that receive up to 1000 mm of rain annually. Oribi are territorial and males mark territories by leaving deposits from the preorbital glands on plant stems, and dung and urine at middens near territory boundaries (Gosling 1972; Arcese et al. 1995; Brashares & Arcese 1999a, b). My study took place in Serengeti National Park, Tanzania, on a section of Kogatende ridge, 82 km, where oribi occurred at densities of ca. 8–10 animals/km2 (Arcese et al. 1995). The habitat is composed of parallel ridges of grassland with scattered boulders and termitaria divided by riverine forest. Oribi occupied ridge tops and sides and were observable from June to December or March each year, depending on grass height. Observations were made with 10 binoculars or a 15–40 spotting scope from an off-road vehicle. Rain fell in all months of each year (ca. 1100 mm annually) with a peak in March–May (Jongejan et al. 1991).

Age, Sexual Maturity and Individual Identification I aged oribi by body size and horn characteristics (only males have horns; Jongejan et al. 1991). I defined adult males as greater than 15 months old, which was the earliest age of territory ownership by a male (unpublished data), and aged them to 4 years using horn attributes (Jongejan et al. 1991; unpublished data). I aged females by size and defined adults as greater than 8 months old, because this is when females become sexually active (Cade 1966; unpublished data). I identified oribi using coloured facial markings, horn length and shape (males only), coat and tail colour, the shape and size of the subauricular gland in relation to the eye, the pattern of marks inside the ears, and scars (Jongejan et al. 1991). I identified most oribi easily by comparing a few traits with prior records on identity cards because oribi were highly variable in appearance (e.g. Arcese 1994). A few oribi had radiocollars (16) or ear-tags (13) and resightings of these oribi confirmed the use of natural marks as reliable.

Group Composition, Ownership and Social Status I recorded 2177 sightings of identified groups and 6530 individuals from November 1987 to February 1988, June 1988 to March 1990, October 1990 to December 1990, December 1991 to January 1992, and December 1992 to February 1993. I monitored 23 territories over the entire study, and an additional six territories by expanding the study area. The sample size for groups ranged from 23 to 29, but because territory ownership could change more than once per year, the sample size for territory owners could exceed 29. Each group was resighted at least five times annually; most were resighted much more often. I tallied changes in territory ownership by identifying all oribi in groups from 15 December to 15 January each year from 1987 to 1992, except in 1990 when I left the area in late December. If an oribi appeared or disappeared during a census period, I used 31 December as the date of record. In territories with more than one adult male, I identified owners as the male that was behaviourally dominant. Assigning dominance in group-living male oribi is straightforward. Dominants within 1–5 m of subordinates often tipped their horns towards subordinates, which then walked or ran away. Dominants occasionally chased subordinates from females that they were courting. Subordinates sometimes responded to a dominant by extending the forelegs and lowering the chest and head towards the ground (see Estes 1991, page 61). I recorded a change in ownership if an owner disappeared from its territory and the study area (‘missing’) or was evicted from half or more of a territory occupied during the prior census (‘displaced’). The fate of missing males was rarely known. The fate of displaced males was always known. Displaced males sometimes survived locally by sequestering most or all of a neighbouring territory. Most ownership changes occurred rapidly between regular visits to territories, or during March– June, when tall grass prevented regular observation. All

ARCESE: AUXILIARY MALES IN ORIBI

males recorded as owners were resident for at least 1 month.

Territory Mapping I mapped territories by noting the location of oribi and dung middens on maps drawn from aerial photographs or using a geographical positioning system satellite receiver (1991–1992). Middens occur along shared borders at intervals of about 30–100 m, and oribi are rarely observed outside their respective borders (Arcese et al. 1995; Brashares & Arcese 1999a, b). When no middens were found, I estimated the location of borders by bisecting the locations of group sightings. Territories were arranged contiguously, most shared two or more borders with neighbours (Brashares & Arcese 1999a, b), but many bordered territories not part of my study.

Time Budgets I conducted three 1-h focal-group observations on 15 groups from December 1988 to February 1989 and August to November 1989. I conducted observations on different days, in each of three time periods (0630–1030, 1030– 1430 and 1430–1830 hours) and chose groups for observation at random or as required to complete my observations. I made observations of male groups with all males present to facilitate paired-comparison statistical tests. I observed groups larger than one with the aid of an assistant. I made all observations from a vehicle at distances of 80–150 m and commenced when no oribi’s attention was focused on the vehicle, which always occurred within 5 min of turning the vehicle’s engine off. I recorded one of 10 behaviours for each group member each minute, including: (1) feeding: grazing, browsing and chewing; (2) sitting: lying with head upright or on the ground; (3) alert: standing with gaze directed and ears perked; (4) grooming: licking or scratching while standing or sitting; (5) marking: depositing dung or urine, scuffing the ground or droppings, and marking with the preorbital gland or preparing a marking site; (6) walking: walking between feeding, resting or marking; (7) chasing: running from or at a group member; (8) courting: (male only) sniffing or licking a female’s rump or inguinal area, licking urine from the ground or the urine stream, performing flehmen or foreleg lifts directed at a female (see Estes 1991, page 22); (9) horning: (male only) stabbing or shaking vegetation with the horns; and (10) out of sight: when an animal was behind a boulder, termitaria or in tall grass. I also recorded marking with the preorbital gland ad libitum because the time taken to leave marks was brief and events lasting less than a recording interval tend to be underestimated (Tacha et al. 1985). Analyses of ‘time spent marking’ and ‘preorbital marking rate’ are both presented.

Interindividual Distance I recorded distance between group members with a hand-held rangefinder as groups were encountered and at

10-min intervals during time budgets. I pooled records for each dyad in a group for December–February and August– November, because August–November corresponded to the expected rut, about 7 months prior to a peak in births from May to June (Jongejan et al. 1991). I recorded oribi within 100 m of each other as part of the same group except when they were known to be neighbours near a common border. I assigned oribi found alone an interindividual distance of 100 m.

Statistical Analyses My analyses followed standard techniques supported by SYSTAT (Wilkinson et al. 1992). I used log-likelihood ratio tests (G tests) for two by two or three comparisons of data summarized as counts, and log-linear models to test for interactions between variables in three-way tables (Fienberg 1980). I used G tests of heterogeneity (Sokal & Rohlf 1981) to test for variation between years in the ability of dominant and singleton males to maintain territory ownership. I used a two-way ANOVA to test for effects of season and social status on the preorbital marking rates, after transforming data by log10. I used a paired t test to compare preorbital marking by males on the same territory. I also employed nonparametric tests when these were better suited to the data. I used Kruskal–Wallis ANOVA to compare interindividual distances between oribi by social status. I used Mann– Whitney U tests to compare the time budgets of singleton and dominant males, and the Wilcoxon signed-ranks test and Spearman rank correlations (rS) for comparing time spent in each of six common behaviours by males in the same group. All tests are reported with their two-tailed probability value. I scaled symbol size in graphs of percentages by the square-root of N because error estimates of percentages were expected to follow from the binomial distribution (Gilbert 1979). Thus, smaller symbols reflect percentages with larger confidence intervals.

RESULTS

Changes in Territory Ownership An average of 27.3% (SE=4.1, Nyears =4) of territorial dominant and singleton males lost most or all of their territories annually (Table 1). About half of all changes in ownership occurred when a dominant or singleton was displaced by a male not identified previously (46.7% of 30 cases; Table 2). Adults on adjacent territories accounted for a similar percentage of changes in ownership (43.3% of 30 cases; Table 2). One subordinate male traversed a territory to fight with, wound, and replace a singleton. The wounded male carried a radio and died in long grass at the periphery of its territory 4 days later. One change in ownership occurred when a subordinate replaced a missing dominant male from its group; another occurred when a subordinate displaced a dominant group member (Fig. 1). One group that annexed an adjacent territory

63

64

ANIMAL BEHAVIOUR, 57, 1

Table 1. Local survival of male oribi in relation to social and territorial status from 1988 to 1992 Territorial status Year 1988–1989 1989–1990 1990–1991 1991–1992 Pooled

Status

Resident

Evicted

Dispersed

Disappeared

N

9 5 7 9 4 13 8 4 15 7 4 13 33 17 48 102

1 0 3 0 0 1 0 0 2 0 0 1 1 0 7 8

0 2 0 0 2 0 0 2 0 0 2 0 0 8 0 9

0 7 6 0 4 5 1 5 3 2 5 5 3 21 19 41

10 14 16 9 10 19 9 11 20 9 11 19 37 46 74 160

Dominant Subordinate Singleton Dominant Subordinate Singleton Dominant Subordinate Singleton Dominant Subordinate Singleton Dominant Subordinate Singleton N

Table 2. Number of new territory owners originating as neighbours, as a subordinate from the territorial group, or from outside the main study area* Neighbour Year 1988–1989 1989–1990 1990–1991 1991–1992 Total

Dominant

Subordinate

Singleton

Subordinate group member

Immigrant to area

1 0 0 0 1

1 1 2 2 6

3 3 2 0 8

1 0 0 1 2

4 2 3 5 14

*Total new territory owners exceeds recorded losses in Table 1 because one subordinate matured and acquired a neighbouring territory between censuses.

dissociated the following year, leaving the subordinate with sole ownership of most of the annexed area (Fig. 2).

Effect of Social Status on Rate of Eviction Dominant males that resided with adult male auxiliaries maintained ownership of their territories more often than singletons in all years of study (Table 1, Fig. 3). This difference was statistically significant in 1988–1989 (G1 =6.22, P<0.01), 1989–1990 (G1 =5.40, P<0.025; Table 1) and for all years pooled (G1 =8.25, P<0.005; Table 1). However, the relative advantage of dominant versus singleton males may have declined during the study (Fig. 3) and is considered further below. The advantage of dominant males was independent of my definition of territory loss because the rate at which dominant and singleton males were displaced versus missing was similar (G1 =0.01, NS; Table 1).

Adult Females, Auxiliary Males and Rate of Eviction The number of adult males and females on territories was negatively related in 3 of 4 years (Fig. 4), and the ratio

of adult females to males was greatest on territories defended by singletons in 3 of 4 years and for all years pooled (Table 3). However, the number of females and the adult sex ratio on territories were unrelated to the rate of eviction (Fig. 5a, b). In contrast, the number of auxiliary males was positively related to the percentage of males that maintained territory ownership from year to year, and no dominant male with three or more auxiliaries lost its territory (Fig. 5c). A three-way log-linear model revealed no significant interaction between ownership and the number of adult females and auxiliary males on territories. Thus, territories defended by multimale groups had fewer females on average, and fewer females per male, than territories defended by singletons. But, singleton males on territories without auxiliary males were often evicted.

Fates of Missing Males A similar percentage of dominant and singleton males that lost territories also disappeared (75% of 4 and 73% of 26, respectively; Table 1) from their territories. The only radiocollared owner that disappeared from its territory was a singleton that died after being displaced (see

ARCESE: AUXILIARY MALES IN ORIBI

above). Indirect evidence also suggests that owners that disappeared from their territories were older on average than their replacements, and that few of them regained territories elsewhere. Fourteen owners that immigrated to the study area were less than 4 years old, whereas about two-thirds of the 23 dominant and subordinate males that disappeared probably were older than 4 years. In contrast, 28% of 29 subordinate males that disappeared from their home territory were resighted on an adjacent territory in the study area (Table 1). Six of these males became territory owners (Table 2), and one joined a singleton as a subordinate. Two other subordinates became owners after a dominant male was displaced (Fig. 1) or disappeared, and one subordinate male that reached adult size after a December–January census dispersed to acquire a neighbouring territory before the subsequent census. About 45% of subordinates disappeared annually (SE=2.1, N=4years; Table 1), but most were under 4 years old and may have gained territories elsewhere. One radiocollared subordinate left its territory to wander for ca. 3 months at the fringe of the study area before being killed by a leopard, Panthera pardus.

(a)

(b) 15.1 15.2

15.2

15.1

15.1 16.1

16.1

16.1 16.2

16.1

16.3

15.1

16.2 16.3

(c)

(d) 15.2

15.2

15.1

15.1 16.1

16.1 16.1 16.2

16.1 15.1

16.2

15.1

16.3

(e)

(f) 15.2

15.2

15.1

15.1 16.1 16.1 16.2 16.1

15.1

16.2 16.3

Attributes of Male Groups Formation of groups Adult males formed groups most often when young males remained philopatric on their natal territory into adulthood (62.5% of 24 cases). However, 25% of 24 groups included subadult males that remained on the natal territory following a take-over by an unrelated male and also remained philopatric into adulthood. A further 12.5% of 24 groups formed when adult males immigrated onto territories where singletons accepted them as subordinates. I observed no case of two adult males being joined by a third immigrant, but eight juvenile males remained philopatric into adulthood on territories where two or more adult males were already present.

Time budgets by social and territorial status Male oribi spent about 50% of their daylight hours sitting, 18% feeding, 7% alert, 3% walking, and 2% grooming (results pooled; Table 4). Chasing, horning and courting were uncommon and did not vary by social or territorial status but did vary by season. Males courted females in August–November (67% of 18 males), slightly more often than in December–February (35% of 17 males; G1 =3.50, P<0.06), probably because a peak in births occurs from March to June and oribi gestate for ca. 7 months. However, courting increased primarily because none of the 11 subordinates was observed to court in December–February, but four of seven did so during the peak in breeding activity from August to November (G1 =8.08, P<0.005). Thus, subordinates courted most often when females were most likely to be fertile. Time spent in each of six common behaviours was largely independent of territorial status (Table 4). However, in both periods, subordinate males fed more than dominants, and in one period, subordinates marked and

Figure 1. Changes in ownership involving annexation and a reversal in dominance status. Numbers beside symbols indicate individual oribi and the territories in which each was identified. Solid lines are watercourses that served as borders, dashed lines are borders marked by middens (see Methods) and stippled areas are rock outcrops and thickets. (a) 10 December 1988–28 September 1989, males 15.1 and 15.2 behaved as a dominant–subordinate group, adjacent to a territory defended by male 16.1. (b) By 1 October, male 15.1 replaced 16.1 and laid within 1–10 m of females 16.1, 16.2 and 16.3. Male 15.1 also occupied his own territory; male 16.1 was not located. (c) On 2 October male 16.1 limped badly in part of its former territory (dotted line). On 17 November, I immobilized and examined 16.1 and found scars on top of the head and at the base of the ear, neck and shoulder, all spaced as though inflicted by horn tips. 1 October–17 November, I observed the following: (1) 6 October–6 November, male 15.2 alone five times inside former territory 15; (2) 1–12 October, male 15.1 established middens and preorbital marks on a new border with male 16.1, and 15.1 observed with females 16.1, 16.2, 15.1 and 16.3, which was not seen thereafter. (d) 2 November, female 16.1 and 16.2 again seen with male 16.1, which had regained use of its leg. (e) 17 November, males 15.1 and 15.2 fought by lunging at and contacting each other’s head, side and flank; a territory border was established. (f) After 11 February, male 15.1 disappeared and male 16.1 annexed its territory.

65

ANIMAL BEHAVIOUR, 57, 1

(a)

(b)

100

12.1

80

12.1

12.1

14.1

14.1

14.2

14.1

14.1

12.1

14.2

(c)

j12.2

(d)

% Males resident at year-end

66

60

40

20 11.2

14.1

12.2

14.5 14.1

12.1

12.2 14.1 12.1

(e)

0

1989

1990

1991

1992

Year Figure 3. Percentage of dominant males (◆), singletons (x) and subordinates (C) on territories at year-end December 1988–January 1992. Sample sizes in Table 2. There was significant variation in the relative ability of dominant and singleton males to hold territories over the study period (GH,1 =4.44, P<0.05). Symbol sizes scaled by square-root N (see Methods).

12.2 12.2

12.3

14.1 14.5 12.1

Figure 2. Group formation by unrelated males, territory annexation and ‘budding-off’ (Wolfenden & Fitzpatrick 1984; conventions follow Fig. 1). (a) 12 December 1988–11 August 1989, male 12.1, female 12.1 occupied territory 12. Adjacent territory 14 was occupied by males 14.1 and 14.2, and females 14.1 and 14.3. (b) By 18 August, female 12.1 immigrated to territory 14 with a 3-month-old male, 12.2. (c) Male 12.1, whose right horn was broken in half by 23 December, remained in territory 12 through December 1989 but was replaced by male 11.2 by 8 November 1990. (d) By December 1991, males 12.2 and 14.1 annex territory 12. (e) By December 1992, a border was established between territory 12 and 14. Male 14.1 seen with females 14.5 and 12.1; male 12.2 seen with immigrant females 12.2 and 12.3.

walked less. Singletons fed more than dominant males in one period. In six of 12 comparisons, the behaviours of dominant and subordinate males in groups were positively correlated (Table 4).

Role of auxiliary males in territory defence The net effect of preorbital marking by auxiliary males was that the sum of marking rates on territories defended by two males exceeded that on territories defended by singletons by 6 marks/h in December–February and

5.9 marks/h in August–November (Fig. 6). Tertiary males in groups added an additional 1.8 marks/h on average. Subordinate auxiliaries marked with their preorbital gland at 6–112% of the rate of the dominant male with which they resided (meanSE=49.50.1, Ngroups =14) and up to 17/h overall. Subordinates marked at lower rates than the dominants with which they resided (paired t test: t13 =3.95, P<0.01; Fig. 6), but the preorbital marking rates of male group members were positively correlated (December–February: rS =0.89, N=8, P<0.005; August– November: rS =0.77, N=6, P=0.05; dominants paired once per group with the highest-ranked subordinate). Marking rates varied seasonally (Fig. 6), being highest in August–November when breeding activity also peaked (see above). However, there was no overall effect of territorial status on marking rate (Fig. 6) and no statistical interaction between marking rate, season and social status. Dominant and singleton males were equally likely to deposit urine and dung on the dung of resident females given that each was within 100 m of a defecating female during a time budget (66% of 44 and 70% of 40 cases, dominant and subordinate males, respectively). Intrusions onto territories by rival males were rare. I saw only one that elicited a chase by a dominant male when a subordinate was present. In this case, the dominant chased a neighbour for ca. 2 min as two subordinates alternately stood alert or performed a ‘rocking canter’ characteristic of territorial antelope (Estes 1991, page 18). I also observed an auxiliary male fighting with a neighbour at a known border for over 15 min. The two males made frequent contact with their horns, but I

ARCESE: AUXILIARY MALES IN ORIBI

females courted by dominants, as expected if dominants used their proximity to females to guard against mating by subordinates. The median distance between females and dominant males was less than that between females and subordinates or females and singletons (Fig. 7). All males were closer to females during the peak of breeding activity in August–November (Fig. 7).

5

Females in group

4

1991

3

DISCUSSION

2

1990 1989

1

1988 0

1

2 3 Males in group

4

5

Figure 4. The number of adult male versus female oribi in groups over four censuses, December 1988–January 1992 (N=26, 28, 29, 28). Lines depict the least-squares fit to the annual data; points were omitted for clarity. Numbers indicate the year of the study. 1988: rS = −0.56, P<0.005; 1989: rS = −0.15, NS; 1990: rS = −0.21, NS; 1991: rS =0.45, P<0.01; for data from December 1988 to January 1992, respectively. The relationship for the 1991 census became nonsignificant with the omission of an outlier, recorded for a large group that underwent fission to form three groups the following year.

could not determine which male was repelling the other.

Access by auxiliary males to females The adult sex ratio in oribi groups suggests a range of mating systems from monogamy to multimale polygyny (Arcese et al. 1995), and behavioural observations support that, in at least some groups, more than one male mated. In ‘dunging rituals’ (Estes 1991, page 61), dominant and subordinate males in groups often preformed flehmen as females urinated, several subordinates courted females, and two subordinates mounted females without interruption by nearby dominants. However, dominant males also rebuffed the approach of subordinates towards

A remarkable feature of oribi social behaviour in the Serengeti is that about 32% of territories are defended by more than one adult male and some are defended by as many as four adult males (Arcese et al. 1995). The results of this study allow me to reject three hypotheses to explain the formation of multimale oribi groups. I cannot reject the cooperative defence hypothesis, however, and so I discuss this hypothesis further before comparing oribi with other antelope and group-territorial mammals and birds. Contrary to the ailing male hypothesis, only one auxiliary male oribi displaced a dominant group member (Fig. 1, Table 2). Moreover, dominant male oribi that resided with subordinates lost territories less often than singletons (Table 1, Figs 3, 4). In contrast, some observations of waterbuck (Kobus ellipsiprymnus) support this hypothesis. Territorial male waterbuck living at very high density (ca. 90/km2) sometimes tolerate ‘satellites’ that mate with transient females (Wirtz 1982). Satellites are most common on territories in the 2 months prior to a dominant male’s disappearance, and they often replace missing territory owners (Wirtz 1982). Territories with satellites also have many more adult male trespassers than territories without them (Wirtz 1982). This suggests that territorial waterbuck sometimes become unable to maintain exclusive territories and, thereafter, are overrun by challengers advancing from local dominance to territory ownership (see also Spinage 1982, page 253). The parental facilitation hypothesis was also not supported well for oribi in the Serengeti. About 40% of subordinant male oribi probably were unrelated to the male with which they resided, and young males often remained philopatric after a male that was likely to have been its father was evicted. Six subordinate males gained territories by displacing neighbours (Table 2), but at least three of these probably were unrelated to the dominant male with which they resided prior to gaining a territory.

Table 3. Mean (±SE) ratio of adult females to males on territories in 4 years from 1988 to 1992 Year Group type Singleton Multimale t statistic df P

1988

1989

1990

1991

Pooled

1.69±0.70 0.37±0.46 5.80 23.9 <0.001

1.53±0.96 0.56±0.35 3.75 24.9 <0.01

2.10±1.29 0.76±0.46 4.10 26.2 <0.01

1.84±1.17 1.36±0.45 1.57 25.5 NS

1.78±0.56 0.76±0.56 6.71 108.8 <0.001

*Sample sizes from Table 1.

67

ANIMAL BEHAVIOUR, 57, 1

100 (a) % Males resident at year-end

68

(b)

(c)

90

80

70

60

50

0

1 2 Females in group

MF Male:female ratio

3+

0

1 2+ Auxiliary males

Figure 5. Effect of (a) females, (b) male to female ratio, and (c) auxiliary males on the percentage of males still resident on their territory at year-end December 1988–1992. (a) G3 =1.65, Ngroups =12, 38, 38, 23, NS; (b) G2 =1.46, Ngroups =50, 35, 26, NS; (c) G2 =10.81, Ngroups =74, 30, 7, P<0.005). Symbol sizes scaled by square-root N (see Methods). Table 4. Median proportion of time that dominant, subordinate and singleton males engaged in six common behaviours Behaviour Period December–February

August–November

Male status

N

Sit

Feed

Alert

Groom

Mark

Walk

Dominant Subordinate rS Singleton Dominant Subordinate rS Singleton

8 11 11 6 6 7 7 7

0.48 0.63 0.52 0.48 0.37 0.38 0.70* 0.66

0.16 0.18* 0.71** 0.27* 0.21 0.34* 0.29 0.13

0.05 0.06 0.59* 0.08 0.08 0.12 −0.08 0.09

0.01 0.02 0.20 0.01 0.01 0.01 −0.12 0.02

0.02 0.01 0.65* 0.03 0.02 0.01* 0.76* 0.01

0.08 0.03* 0.18 0.03 0.06 0.05 0.79** 0.03

Males in the same group compared with Wilcoxon tests; dominant and singleton males compared with U tests. rS denotes the Spearman rank correlation coefficient for dominants and subordinates in groups. *P<0.05, **P<0.025.

I observed only one possible case of territory inheritance. Oribi that remain philopatric in the presence of their fathers may survive or reproduce better than others, but testing this requires a direct measurement of relatedness and the fitness of dispersing males. Contrary to the main prediction of the female defence hypothesis, the number of adult females and males on territories was negatively related in 3 years (Fig. 5), and the ratio of females to males on territories defended by male groups was lower than on territories with singletons in all years (Table 3). This suggests that male oribi in groups had relatively less access to females than did singleton males.

Cooperation and Male Group Formation in Oribi My results support the hypothesis that dominant male oribi tolerated subordinate auxiliaries in exchange for contributions to defence and a reduction in the rate of eviction. Dominant males lost their territories less often

than singletons (Table 1, Fig. 3) and no dominant with two or more auxiliaries lost its territory between censuses (Fig. 4). Only one of 31 replacement males probably gained a territory by evicting a dominant group member (Fig. 1), whereas at least 55% of replacement males originated on adjacent territories (Table 2). Some of the 14 replacement males that immigrated to the study area also probably came from territories that were adjacent to the study area but had not been monitored.

Scent marking by auxiliary males as an aid in defence Much work shows that scent marking functions in territory defence in mammals (reviews in Gosling 1982, 1986b, 1990). Thus, it is plausible that the superior ability of dominant males to retain territories was due partly to scent marking by subordinate auxiliaries. Auxiliaries contributed about half the preorbital marks on territories (Fig. 6) and, because singletons and dominants marked at similar rates, territories with male groups were marked

ARCESE: AUXILIARY MALES IN ORIBI

10

Median male–female distance (m)

20

Preorbital marks/h

15

10

5

0

Dominant

Subordinate

Singleton

Figure 6. Mean (±SE) preorbital marking rates by males in December–February, 1988–1989 (◆) and August–November, 1989 (e). Sample sizes for dominant, subordinate and singleton males in each period were 8, 8, 8 and 6, 7, 8, respectively. Marking rate was higher in August–November than December–February (F1,41 =7.60, P<0.01), but did not vary by status (two-way ANOVA, data transformed by log10: F2,41 =1.25, NS).

more often overall (Brashares & Arcese 1999a; this study). Auxiliaries also contributed dung to middens on territory boundaries and overmarked female dung (Brashares & Arcese 1999b; this study). If potential rivals avoid challenging owners of well-marked as opposed to poorly marked territories, marking by subordinates may enhance the ability of dominants to retain their territories. Work on oribi also suggests that scent marking plays an important role in territory defence. Oribi place preorbital marks mainly along borders shared with neighbours, and do so more often along borders adjacent to male groups as opposed to singletons (Brashares & Arcese 1999a). The rate of preorbital marking is related positively to the number adult male neighbours but not to territory size, the length of shared borders, or the number of females on the home or adjacent territory (Brashares & Arcese 1999a). Thus, because most rivals for territory ownership were neighbours (Table 2), oribi marked most along borders adjacent to the highest number of potential rivals. Male oribi also limit the volume of dung at each defecation, as expected if it is a valuable marking resource (Brashares & Arcese 1999b). This allows males to mark middens and overmark female dung more often, and the added marks of subordinates translates into higher overall feacal marking rates on territories defended by groups as opposed to singletons (Brashares & Arcese 1999b). Overmarking female dung and maintaining middens at borders may alert rivals to the presence of territorial males (Gosling 1986b). In oribi, feacal marking also may discourage challenges by potential rivals by advertising the

8

6

4

2

0

Dominant

Subordinate

Singleton

Figure 7. Median distance (m) between adult females and dominant, subordinate and singleton males in December–February, 1988–1989 (◆) and August–November 1989 (e). Males were closer to females in August–November than December–February, and dominants were closer than subordinates in each period (P<0.05 in all cases, U tests; N=17, 6, 6 and 22, 13, 13 for singleton, dominant and subordinate males in December–February and August– November, respectively).

presence of more than one adult male (Brashares & Arcese 1999b).

Potential costs of auxiliary males Group living might also entail costs for dominant males if auxiliaries fertilize females or reduce their representation in groups (Fig. 5, Table 3). Some of my observations suggest that dominant males minimized matings by auxiliaries by monitoring females and guarding those that were fertile (Fig. 7). However, paternity studies are required to determine whether auxiliary males mate despite the behaviour of dominants. The presence of auxiliary males also may reduce the number of females on territories (Fig. 5, Table 3). This might result if predation risk or competition for food were related positively to group size, such that the presence of more males in groups resulted in more emigration or less immigration and recruitment by females. Other work in the Serengeti shows that oribi prefer groups of three to four adults in size over a wide range of densities (2–30/km2, Arcese et al. 1995; unpublished data) and that larger or smaller groups approach the preferred size over time (unpublished data). Thus, male territory owners may be more willing to tolerate auxiliaries when groups are small (cf. Schaffner & French 1997) and factors other than defence may affect whether auxiliaries are tolerated.

Benefits to auxiliary males Auxiliary males often acquired territories of their own during the study (Table 2, Fig. 2), and some that

69

70

ANIMAL BEHAVIOUR, 57, 1

disappeared probably gained territories elsewhere. Auxiliary males that were related to dominants probably enhanced their fathers’ ability to hold territories and may have gained indirect fitness benefits as a result. All auxiliaries may have experienced a reduction in predation risk, or increased access to food, mates or information about vulnerable neighbours. However, estimating the magnitude of any potential benefit requires more work.

Oribi Versus Other Territorial Antelope Group formation by male oribi raises questions about why males tolerate auxiliaries, and why auxiliaries accept subordinate status (cf. Brown 1987; Emlen 1991). Some of these questions are addressed above, but comparative studies of oribi populations and other small antelope may be the best way to pursue answers in future. Oribi in eight of nine populations studied included groups with more than one adult male, and the percentage of groups with multimale groups varied from 0–40% (Arcese et al. 1995). The development and stability of oribi groups has not been studied outside the Serengeti, but ample variation in sex ratio, group composition, population density and rainfall exists to test whether these factors affect the frequency of group living in oribi. Predictions for the effect of demographic and environmental factors are available from competing models for the evolution of group living and cooperative breeding (e.g. Brown 1987; Smith 1990; Emlen 1991; Harcourt & de Waal 1992; Jennions & Macdonald 1994). Preliminary analyses for oribi suggest that multimale groups occur more often where groups are large, polygyny is common and large carnivores are abundant (Arcese et al. 1995). Aside from oribi, the blue duiker, Cephalophus monticola, is the only antelope wherein delayed dispersal is known to result in groups comprising up to four adults (Dubost 1980). More work on individually identified duikers and other small antelope seems warranted. Males that are excluded from territories in other antelope usually occupy bachelor or mixed-sex herds of nonbreeders, particularly in larger species (Estes 1974, 1991; Jarman 1974; Leuthold 1977; Gosling 1986a). In many small species, however, including klipspringer, Oreotragus oreotragus, steenbok, Raphicerus campestris, dik-diks, Modoqua spp., and duikers, Sylvicarpa and Cephalophus spp., adult males are permanently territorial and bachelor herds do not occur. However, little is known about the life history of subadult males or the acquisition of territories in these species (Estes 1991). Male waterbuck sometimes tolerate satellites (Wirtz 1982) but only at high density (Spinage 1982; Estes 1991), and this is probably a result of rival males progressively overtaking ailing territory owners (Spinage 1982; Estes 1991). Male reedbucks, Redunca spp., show signs of group living, but reports suggest that subordinate males are restricted to the interstices of territories (Jungius 1971; Irby 1979; Estes 1991). In all other territorial antelope, owners exclude rivals, particularly when fertile females are present (Estes 1974, 1991; Jarman 1974, 1979; Leuthold 1977; Spinage 1982; Gosling 1986a).

Oribi Versus Other Cooperatively Territorial Species More data are needed to draw clear comparisons between oribi and other group-territorial species, but initial results suggest that oribi resemble cooperative primates and birds more than felids. Unlike lions, Panthera leo (Schaller 1972; Packer 1986) and cheetah, Acionyx jubatus (Caro 1994), dominance among male oribi appears to affect access to females. Associations between adult male oribi also last less than 2 years on average (Arcese et al. 1995), compared with often lifelong coalitions in cooperative felids (Packer et al. 1988; Caro 1994). Long-term associations also occur in some large primates (Pan troglodytes, Goodall 1986), but male gorillas, Gorilla gorilla, have more fluid social relations that resemble those of oribi in surprising ways (Yamagiwa 1987; Robbins 1995). The presence of hierarchies, instability in group membership and year-round territoriality all are traits shared by oribi, some group-living callitrichids (Dawson 1987; Goldizen 1989; Goldizen & Terbourgh 1989; Baker et al. 1993; Schaffner & French 1997) and many cooperatively breeding birds (reviews in Smith 1990; Emlen 1991). It seems reasonable, therefore, that factors favouring group formation and territory defence by primates and birds may also favour these behaviours in the oribi. Acknowledgments The Directors General of the Serengeti Wildlife Research Institute and Tanzania National Parks facilitated my work in the Serengeti. G. Jongejan kindly allowed me to use data she helped to collect. T. Sinclair, S. Mduma, M. Borner, T. Corfield, M. Patterson and B. and K. Gillis gave much help, and J. Brashares and J. Ginsburg provided constructive comments. My work was supported by postdoctoral fellowships from the Natural Sciences and Engineering Research Council (Canada) and NATO-NSF (U.S.A.), and by the Committee for Research and Exploration, National Geographic Society (U.S.A.), NSERC, Friends of Conservation (U.S.A/U.K.), Frankfurt Zoological Society (G.D.R.), Wildlife Conservation Society (U.S.A.), USDA (Hatch grant N684), University of Wisconsin, Max McGraw Wildlife Foundation (U.S.A.), and a U.S. National Science Foundation Young Investigator Award (IBN-945812). References Arcese, P. 1994. Harem size and horn symmetry in oribi. Animal Behaviour, 48, 1485–1488. Arcese, P., Jongejan, G. & Sinclair, A. R. E. 1995. Behavioural flexibility in a small African antelope: group size and composition in the oribi (Ourebia ourebi, Bovidae). Ethology, 99, 1–23. Baker, A. J., Dietz, J. M. & Kleiman, D. G. 1993. Behavioural evidence for monopolization of paternity in multi-male groups of golden lion tamarins. Animal Behaviour, 46, 1091–1103. Brashares, J. S. & Arcese, P. 1999a. Scent marking in a small African antelope: I. The maintenance of borders between territorial male oribi. Animal Behaviour, 57, 1–10.

ARCESE: AUXILIARY MALES IN ORIBI

Brashares, J. S. & Arcese, P. 1999b. Scent marking in a small African antelope: II. The economics of marking with feces. Animal Behaviour, 57, 11–17. Brown, J. L. 1987. Helping and Communal Breeding in Birds: Ecology and Evolution. Princeton, New Jersey: Princeton University Press. Cade, C. E. 1966. A note on the behaviour of the Kenya oribi Ourebia ourebi in captivity. International Zoological Yearbook, 6, 205. Caro, T. M. 1994. Cheetahs of the Serengeti Plains: Group Living in an Asocial Species. Chicago: University of Chicago Press. Dawson, G. A. 1987. Composition and stability of social groups of the tamarin, Sanguinus oedipus geoffroyi, in Panama: ecological and behavioral implications. In: The Biology and Conservation of the Callitrichidae (Ed. by D. G. Kleiman), pp. 23–37. Washington, D.C.: National Zoological Park. Dubost, G. 1980. L’ecologie et la vie sociale du cephalophe bleu (Cephalophus monticola Thunberg), petit ruminant forestier africain. Zeitschrift fu¨r Tierpsychologie, 54, 205–266. Emlen, S. T. 1991. Evolution of cooperative breeding in birds and mammals. In: Behavioural Ecology: an Evolutionary Approach (Ed. by J. R. Krebs & N. B. Davies), pp. 301–357. Oxford: Blackwell Scientific. Estes, R. D. 1969. Territorial behavior of the wildebeest (Connochaetes taurinus Burchell, 1823). Zeitscrhrift fu¨r Tierpsychologie, 26, 284–370. Estes, R. D. 1974. Social Organization of the African Bovidae. In: The Behavior of Ungulates and Its Relation to Management (Ed. by V. Geist & F. R. Walther), pp. 166–205. Morges, Switzerland: IUCN. Estes, P. D. 1991. The Behavior Guide to African Mammals. Los Angeles: University of California Press. Fienberg, S. E. 1980. The Analysis of Cross-classified Categorical Data. Cambridge, Massachusetts: MIT Press. Gilbert, N. 1979. Biometrical Interpretation. Oxford: Oxford University Press. Goldizen, A. W. 1989. Social relationships in a cooperatively polyandrous groups of tamarins Sanquinus fuscicollis. Behavioral Ecology and Sociobiology, 24, 79–89. Goldizen, A. W. & Terborgh, J. 1989. Demography and dispersal patterns of a tamarin population: possible causes of delayed breeding. American Natualist, 134, 208–224. Goodall, J. 1986. The Chimpanzees of Gombe. Cambridge, Massachusetts: Harvard University Press. Gosling, M. L. 1972. The construction of antorbital gland marking sites by male oribi (Ourebia ourebi, Zimmerman, 1783). Zeitschrift fu¨r Tierpsychologie, 30, 271–276. Gosling, M. L. 1982. A reassessment of the function of scentmarking in territories. Zeitschrift fu¨r Tierpsychologie, 60, 89–119. Gosling, M. L. 1986a. The evolution of mating strategies in male antelope. In: Ecological Aspects of Social Evolution: Birds and Mammals (Ed. by D. I. Rubenstein & R. W. Wrangham), pp. 244–281. Princeton, New Jersey: Princeton University Press. Gosling, M. L. 1986b. Economic consequences of scent-marking in mammalian territoriality. In: Chemical Signals in Vertebrates. Vol. 4 (Ed. by D. Duvall, D. Schwarze & R. M. Silverstein), pp. 385–395. New York: Plenum Press. Gosling, M. L. 1990. Scent-marking by resource holders: alternative mechanisms for advertising costs of competition. In: Chemical Signals in Vertebrates. Vol. 5 (Ed. by D. MacDonald, D. Mu¨llerSchwarze & S. E. Natynczuk), pp. 315–328. Oxford: Oxford University Press. Harcourt, A. H. & de Waal, F. B. M. (Eds) 1992. Coalitions and Alliances in Humans and Other Animals. Oxford: Oxford University Press.

Hofman, R. R. & Stewart, D. R. M. 1972. Grazer or browser: a classification based on the stomach structure and feeding habits of East African ruminants. Mammalia, 36, 226–240. Irby, L. R. 1979. Reproduction in mountain reedbuck (Redunca fulvorufula). Mammalia, 43, 190–213. Jarman, P. J. 1974. The social organization of antelopes in relation to their ecology. Behaviour, 48, 215–267. Jarman, M. V. 1979. Impala social behavior. Zeitschrift fu¨r Tierpsychologie, 21, 1–92. Jennions, M. D. & Macdonald, D. W. 1994. Cooperative breeding in mammals. Trends in Ecology and Evolution, 9, 89–93. Jongejan, G., Arcese, P. & Sinclair, A. R. E. 1991. Growth, size and the timing of births in an individually identified population of oribi. African Journal of Ecology, 29, 340–352. Jungius, H. 1971. The Biology and Behavior of the Reedbuck in the Kruger National Park. Mammalia Depicta. Hamburg: P. Parey. Leuthold, W. 1966. Variations in territorial behaviour of Uganda kob, Adenota kob thomasi (Neumann 1896). Behaviour, 27, 214– 257. Leuthold, W. 1977. African Ungulates. Berlin: Springer Verlag. Packer, C. 1986. The ecology of sociality in felids. In: Ecological Aspects of Social Evolution: Birds and Mammals (Ed. by D. I. Rubenstein & R. W. Wrangham), pp. 429–451. Princeton, New Jersey: Princeton University Press. Packer, C., Herbst, L., Pusey, A. E., Bygott, J. D., Hanby, J. P., Cairns, S. J. & Mulder, M. B. 1988. Reproductive success of lions. In: Reproductive Success: Studies of Individual Variation in Contrasting Breeding Systems (Ed. By T. H. Clutton-Brock), pp. 363–383. Chicago: University of Chicago Press. Robbins, M. M. 1995. A demographic analysis of male life history and social structure of mountain gorillas. Behaviour, 132, 21–47. Schaffner, C. M. & French, J. A. 1997. Group size and aggression: ‘recruitment incentives’ in a cooperatively breeding primate. Animal Behaviour, 54, 171–180. Schaller, G. B. 1972. The Serengeti Lion. Chicago: University of Chicago Press. Smith, J. N. M. 1990. Summary. In: Cooperative Breeding in Birds: Long-term Studies of Ecology and Behaviour (Ed. by P. B. Stacey & W. D. Koenig), pp. 593–612. Cambridge: Cambridge University Press. Sokal, R. R. & Rohlf, F. J. 1981. Biometry. San Francisco: W. H. Freeman. Spinage, C. A. 1982. A Territorial Antelope: The Uganda Waterbuck. London: Academic Press. Stacey, P. B. & Koenig, W. D. (Eds) 1990. Cooperative Breeding in Birds: Long-term Studies in Ecology and Behavior. Cambridge: Cambridge University Press. Tacha, T. C., Vohs, P. A. & Iverson, G. C. 1985. A comparison of interval and continuous sampling methods for behavioral observations. Journal of Field Ornithology, 56, 258–264. Wilkinson, L., Hill, M. & Vang, E. 1992. SYSTAT for Windows. Version 5. Evanston, Illinois: SYSTAT. Wirtz, P. 1982. Territory holders, satellite males and bachelor males in a high density population of waterbuck (Kobus ellipsiprymnus) and their associations with conspecifics. Zeitschrift fu¨r Tierpsychologie, 58, 277–300. Wolfenden, G. E. & Fitzpatrick, J. W. 1984. The Florida scrub jay: demography of a cooperative breeding bird. Monographs in Population Biology, 20, 1–406. Yamagiwa, J. 1987. Inta- and intergroup interactions of an all-male group of Virunga mountain gorillas (Gorilla gorilla beringei). Primates, 28, 1–30.

71