Social ontogeny and behavioural diversity

J. Zool., Lond. (1979) 188, 251-266

Social ontogeny and behavioural diversity :consequences for Bighorn sheep Ovis canadensis inhabiting desert and mountain environments JOELBERGER* Ethology Group: Department of Environmental, Population, and Organismic Biology, University of Colorado Boulder, Colorado 80309, U.S.A. (Accepted 10 October 1978)

(With 5 figures in the text) The development of social behaviour was studied in three natural populations of Bighorn sheep (Ovis canadensis). Individuals in all populations exhibited the same behavioural repertoires, but the utilization of specific behaviours among both infants and adults differed between populations. It was suggested that differences in the adult utilization of behaviour patterns result in part from behaviours used during infancy. In turn, the social and physical environments affected the development of subsequent behaviour patterns in sheep older than Iambs.

Desert sheep (in southern California) used mounting behaviours often, presumably because they matured sexually at an early age. Sheep in a more northern environment (British Columbia) used different behaviour patterns more frequently perhaps as a consequence of interactions experienced in larger bands. Additionally, playful interactions were reduced due to physical hazards in thedesert, but they wereat least nine timesas frequent in the northern study population. Play in large groups resulted in the utilization of more different kinds of behaviour patterns. These behavioural findings are interpreted ecologically as consequences of inhabiting environments that differ socially and physically.

Contents Introduction.. . . . . . . . . . . . .

Methods . . . . . . . . .. . . . .

Study populations . . . . .. .. ..

Data collection . . . . . . . . . .

Analyses . . . . . . . . .. ..

Ethogram . . . . . . .. . . . .

Results . . . . . . . . .. .. ..

Sex and age differences in behavioural development Behavioural development in different environments Behavioural diversity . . . . . . . . . .

Discussion . . . . . . . . .. .. ..

Development of motor patterns . . . . ..

Social ontogeny in different environments . . . .

Population differences in behavioural ontogenies Behavioural development and diversity . . ..

References . . . . .. . . . . . . ..

..

..

.. ..

.. .. ..

.. ..

..

..

..

.. .. .. .. ..

..

.. .. .. .. ..

.. .. .. .. .. .. ..

.. .. .. .. .. .. .. .. .. ..

.. .. ..

..

.. ..

.. ..

.. .. .. .. .. .. .. ..

..

..

..

..

..

..

.. ..

Page 252 252 252 254 254 254 255 255 255 259 260 260 261 261 262 264

* Present address: Conservation and Research Center, National Zoological Park, Smithsonian Institution, Front Royal, Virginia 22630, U.S.A. 251

+

0022-5460/79/06025 1 16 S02.00/0

Q 1979 The Zoological Society of London

252

J. BERGER

Introduction Many factors (e.g. body size, external morphology, habitat complexity, food availability and predictability, demographic properties) influence the social organization of a species.

Numerous studies of mammalian organizational systems (Brown & Orians, 1970; CluttonBrock & Harvey, 1977; Ejsenberg, 1966; Fisler, 1969; McBride, 1964)have considered adult spacing patterns as “end points” of a developmental process. However, such studies have not concentrated on the process itself, namely the ontogeny of social behaviour (Bekoff, 1972; Poirier, 1977). Studies of social development are only now beginning to bridge these gaps between social development and later adult social organization (e.g. Barash, 1974).

Recently, Bekoff (1977) and Kleiman & Brady (1978) studied the social development of infant canids and considered the ways that interactions between infants lead to later behaviours, such as dispersal.

Nevertheless, only rarely has social development been compared quantitatively between the same mammalian species in allopatric populations inhabiting different environments (see Baldwin & Baldwin, 1971; Berger, in prep. a ; Kummer, 1971), although adult social behaviour has been studied intraspecifically in ungulates (Geist, 1971; Lent, 1969; Nievergelt, 1966; Shackleton, 1973; Sinclair, 1977; Schaller, 1977), rodents (Barash, 1974; Fisler, 1965) and primates (Baldwin & Baldwin, 1973; Clutton-Brock & Harvey, 1977; Richards, 1974; Crook, 1970; Eisenberg, Muckenhirn & Rudran, 1972). In Bighorn sheep (Ovis canadensis) I found that the distribution of food affects band size(s) differently in desert and northern environments, and that subsequent spatial associations in lambs were influenced by the size of the groups within which they are found (Berger, in press).

The idea that the distribution and predictability of food resources exerts strong selective pressures on the evolution of social systems is not new (Brown, 1975; Graul, Derrickson & Mock, 1977; Orians, 1969; Wilson, 1975). For Bighorn sheep, then it would prove most constructive to understand how behaviour develops in different habitats, what factors are responsible for any observed differences, and the adaptive significance for differential social ontogenies. Additionally, since group sizes (the social environments) and habitats (the physical environments) of Bighorn sheep differ in complexity (Berger, in prep. a, b), one might hypothesize that differences in social development lead to differences in the frequencies of utilization of adult behaviour patterns. In this paper the relationships between social development, ecological conditions, and adult behaviours in three populations of Bighorn sheep are explored.

Methods Study populations California Bighorn sheep (0. canadensis culiforniunu) were studied in the Chilcotin-Cariboo region of the interior of British Columbia from May through November, 1976. A second population of California Bighorns was studied on Hart Mountain, Great Basin Desert, eastern Oregon from May to August, 1977. These sheep were transplanted from the Chilcotin to Hart Mountain in 1954 where they were extirpated at the turn of the century. At the time of the transplant they numbered 18; the current population size is estimated at 200 (Kornet, 1978). The third population studied was located in the Santa Rosa Mountains, Colorado Desert, southern California. These sheep are commonly called desert or peninsular Bighorn (0.canadensis cremnobates). The sheep population in these arid insular mountains is estimated at 250 (Merrit, 1974). They were studied from January through April, 1977 (see Fig. 1 for locations). Further details of the habitats, climate, and study populations are found in Berger (in prep. b).

SOCIAL O N T O G E N Y A N D B E H A V I O U R A L DIVERSITY

253

254

J . BERGER

Data collection Data on sheep behaviours were collected by stalking and observing sheep in the field on almost a daily basis. I observed sheep in the Chilcotin for 896 hours, those at Hart Mountain for 293 hours, and in the desert for 454 hours. Behaviours and social interactions (see ethogram) were recorded into a casette tape deck and later transcribed or taken directly on note cards.

Adult male and female sheep may be sexed visually on the basis of horn and body size differences (Geist, 1968, 1971). However, at a young age when no dimorphism exists, lambs may be sexed by one of three methods : (1) Male lambs urinate from a standing position while females squat. This method of sexual determination was used most frequently as other methods were not always as accurate. (2) Testes may be visible in males. However, if testes are not visible, it does not necessarily mean that a lamb is not a male. (3) Once horns develop male horns grow faster and are thicker and flare more widely than those of females. I utilized the classification scheme proposed by Geist (1968, 1971), i.e. Class 1 rams, ewes, male yearlings, etc. The data presented in this paper represent only those situations in which I was certain of sexual identities.

Ana Iyses All statistical analyses of the differences between percentages (probabilities) were performed using the Brandt and Snedecor method when there were two or more samples (see Snedecor, 1956; section 9.9) and the arcsin transformation method for testing the equality of two percentages (suggested by Sokal & Rohlf, 1969:607). This method generates a test statistic, t,, which may be compared with a normal deviate (are under the normal curve).

Ethogram For Bighorn sheep, 17 behaviour patterns have been categorized and described by Geist (1971 :134-143). This ethogram may be enlarged to include at least six more behaviours. Below are listed and described the behaviours that were quantified during social interactions.

Sexual patterns Mount. (seeGeist, 1971) Mounting is the only overtly sexual pattern used by sheep. Flehmans or other sexual patterns that occur primarily during the breeding season were not quantified.

Mounting is common year-round.

Contact patterns Head butt. (see Geist, 1971) A sheep that butts another with its head.

Clash. (see Geist, 1971) Two sheep that rear up on their hind legs and follow this motion by slamming their heads into one another.

Touch. Two sheep that lower their heads and place them in contact with one another. No pushing occurs and their heads remain in contact for at least one second.

Pushopponent.Two or more sheep that push with their heads the rumps or sides of other sheep.

Threat patterns Threat jump. “This is an intention movement to clash” (Geist, 1971:143). Threat jumps are often sufficient to discourage an opponent from fighting and no further contact ensues.

Horn threat. “. . . this is an intention movement to butt and as such is a true weapon threat” (Geist, 1971:142-143). Sheep lower their heads in a position to butt an opponent.

Head threat. Young sheep that have not yet developed their horns will lower and orient their heads in the same way as when horn threats are performed.

The above three patterns are indeed threats, as Geist (1971) correctly pointed out, because: (1)

SOCIAL O N T O G E N Y A N D BEHAVIOURAL DIVERSITY

255

when dominant individuals threaten subordinates they retreat and little or no contact occurs; and (2) when threats are insufficient to discourage the approaches of other sheep, contact patterns follow.

Display patterns Low stretch. (see Geist, 1971) A display of horns performed by lowering the head.

Twist. (see Geist, 1971) A display of horns performed by lowering and twisting the head.

Present. (see Geist, 1971) A horn display performed by raising the head and lifting the horns.

Front leg kick. (see Geist, 1971) “. . . is linked to the display threats of sheep, such as the low stretch, present, and twist; it is virtually a contact display threat” (Geist, 1971:139). Since front leg kicks are not contact patterns such as butting, and probably serve to convey information about body size (as do horn displays for horn size) they were categorized as display patterns.

Walther (1974) recognized that threat and dominance displays in horned ungulates may often be similar but argues that they merit distinction because basic differences also exist. For example, in Bighorn sheep threats indicate readiness for fighting by directly utilizing the horns and placing them in a position for contact. In contrast, displays are not actively conveying a threat (although a subordinate individual may feel threatened by the presence of a dominant sheep) and displaying individuals are not oriented in a fighting position.

Rotational patterns Neck twists. This pattern is more exaggerated than a head shake (Geist, 1971) and it is often used in playful interactions. Neck twists are common to lambs, ewes, and yearlings.

Gambol. Gambols entail jumping with all four legs off of the ground in a somewhat vertical direction. The body axis is rotated and twisted in midair. Both ewes and lambs have gamboled so high that they fell over backwards. The oldest ram observed gamboling was about six years old.

Heel kick. Heel kicks entail jumping off of the ground and kicking simultaneously both rear legs or heels. Byers (1977) described an apparently identical motor act in the play of Ibex kids.

All of the above mentioned behaviour patterns except displays occur during play. Rotational movements function as play signals (Berger, in prep. a).

Results Sex and age diflerences in behavioural development Males generally use more contact, threat, sexual, and display patterns than do females (Fig. 2). There were no significant differences (arcsin proportions test) in the proportional use made of patterns by male and female lambs. These data indicated a trend in which male lambs were more likely to engage in all patterns other than rotational movements (see Discussion) than female lambs. Displays were infrequently observed in females of all ages and for males they developed only after two months of age (Fig. 2).

Behavioural development in diflerent environments The week (age) of first appearance of different motor actions in lambs is given in Table I. Generally, behaviours appeared at similar ages in all populations except that two display patterns occurred about three months earlier in desert sheep than they did in Chilcotin sheep. The most probable reason displays were not observed in some populations was because I departed from those study sites before lambs developed those patterns (i.e. they were too young).

J . BERGER

256 Threat

Contact

Display

Rotational

‘ 0

Males

=./

/*-•

t0

*

.’.\

*

*

*

*

.-¤

--0

Females

L

Y

n.1669

2:l

A

945

L

Y

7 36 116

A

371

L

Y

A

2342 88

A

261

I

I

L

Y

L

A 128

16 8 46

Y

0 0 55

/?

A 350

E

G

FIG.2. Sex and age differences in the development of behaviour patterns in Bighorn sheep. L=lambs (two data points are represented for lambs; those occurring before two months of age and those after this age); Y=yearlings; A=adults (i.e. ewes and Class 1 rams); asterisks indicate areas of statistical significance (arcsin proportions test) at P < 0.05 level ; circles = males, squares = females

TABLE I Age of appearance of motor actions in lambs. A minimum offive observations of an action was necessary for inclusion Locality

r

Pattern of behaviour

Motor act

7

desert

transplant

Chilcotin

one week one week one week one week one week one week two weeks one week one week one week No No No No

...... ......

____________~

Contact

Threat Rotational

Sexual Display

head butt clash touch push opponent head threat threat jump neck twist gambol heel kick mount low stretch front leg kick twist present

~~

No= not observed.

......

......

..

......

3 months 24 months

No No

...... ...... ...... ...... ......

......

...... ......

6 months 6 months 7 months

No

257

S O C I A L O N T O G E N Y A N D BEHAVIOURAL. D l V E R S I T Y 60

24

12

Threat

Contact

Rotation01

Sexual

FIG.3. Histogram illustrating the per cent relative frequency of behaviour patterns used by three populations of Bighorn sheep lambs. For any given population, per cent relative frequency equals the number of motor acts per behaviour pattern divided by the total number of motor acts for all behaviour patterns. D=desert (n=572); T= transplant (n= 579); Chilcotin (n= 1144).

Female yeorlings

Ewes -

70 56 t

G

42

a"

28

?

T

14

Male yearlings

-[ Rams -- -

70

5

a"

I

42 28 14

con

ihr

rot

k sex

il dis

dis

con

thr

rot

sex

FIG. 4. Histogram illustrating the per cent relative frequency of behaviour patterns used by various sex and age classes of Bighorn sheep in three populations. Within any given population, the per cent relative frequency of occurrence of behaviours in a sex or age class is calculated in the same manner as it was in Fig. 3. Rams equal Class 1 rams only. D=desert (n=1120); T=transplant (n=465); C=Chilcotin (n= 1026).

J . BERGER

25 8

The frequencies of utilization of motor patterns for sheep in different environments was analysed by using two methods. The first compared the proportion of actions representing all behaviour patterns within a population to the per cent utilization of the same behaviour patterns in other populations. These percentages are uncorrected for group sizes but they represent the proportion of acts occurring within the entire behavioural repertoire (the second method corrects for group size differences).

The observed relative frequencies of motor acts for lambs are shown in Fig. 3 ; for yearlings, ewes, and Class 1 rams in Fig. 4. Desert lambs engaged in significantly more contact and sexual patterns than did Chilcotin lambs (see Table I1 for a summary of the statistical analyses for Figs 3 and 4). Conversely, lambs from the Chilcotin utilized significantly more rotational patterns and threat patterns. Differences in behaviour pattern utilization were not significant for other categories of behaviour except rotational patterns.

These data also were analysed by comparing the rates of interactions directly between populations by compensating for differences in group sizes. The number of behaviour patterns/hour/individual was calculated so that I could determine whether the differences in percentages (Figs 3 and 4)were real (see Table 111).That is, I found that those behaviour patterns that comprised a greater per cent component within populations also occurred at a higher rate when compared between populations.

Generally, the results of Figs 3 and 4 and Tables I1 and I11 showed that threat, contact, and sexual patterns occurred more frequently in desert yearlings, ewes, and Class I rams than they did in Chilcotin sheep. More specifically, the major significant differences for males may be summarized as follows : T A B L EI1 Sutntiiary of tests of equality of two percentages for relative frequencies of behaviour patterns observed it2 lambs, yearlings, ewes, arid Class I rams

Lambs

Yearlings females

males

Ewes

Class I Rams

Contact

Threat

Rotational

Sexual

Des vs Chi1 Des vs Trans Chil vs Trans

* * * 4.48 NS 1.27 * * * 4.90

** 3.13 * * * 3.68 NS 0.93

* * * 4.75 * * * 5.42 * * * 5.80

* * * 5.62 NS 0.86 * * * 4.76

Des vs Chi1 Des vs Trans Chil vs Trans Des vs Chi1 Des vs Trans Chi1 vs Trans Des vs Chi1 Des YS Trans Chi1 vs Trans Des vs Chi1 Des vs Trans Chi1 vs Trans

* 2.19 NS 1.80

*

*** 4.90 NS 1.75

-

-

-

-

NS 1.08 * 1.90

* * 2.72 NS 0.96 NS 1.74 NS 1.68 * 1.90 * 1.85 * * * 5.76 * * * 5.43 NS 1.23

Class I Rams

*

2-08

NS 0.48 * 2.20 * 2.29 * * * 6.00 NS 0.86 * * * 6.19 ** 2.95 * * * 3.39 ** 3.26

+ Male Yearlings

*** 6.72

2.31 * * 2.61 *** 5.07 ** 2.96 * * * 3.90 * 2.49 * * 2.91 * 2.02 * 2.27 NS 0.49 NS 1.08

* * * 5.61

NS 0.36 NS 1.44 NS 1.68 * * * 6.31 * 2.29 * * 3.14 NS 0.38 NS 0.79 NS 0.60 Des vs Chil Des vs Trans Chil vs Trans

2.16

-

-

-

-

*

Display

* j : * 4.70 NS 0.77 * * 3.10 * * * 4.72 * * 3.12 NS 1.73

Des=desert; Trans= transplant, Chil=Chilcotin. *=P<O.O5; **=P<O.Ol; ***=P<O.OOl; numbers=& (test statistic-see statistical analyses). NS=no significance (see also Figs 3, 4)

SOCIAL O N T O G E N Y A N D B E H A V I O U R A L DIVERSITY

259

TABLE I11 Rates of interactions of different sex and age classes of Bighorn sheep in three populations

Pattern

Contact Threat Rotational Sexual Display

Lambs Des Tra Chi 0.203 0.030 0.083 0.065 -

0.191 0.035 0.018 0.022

0.120 0.054 0.123 0.021

-

-

Female yearlings Des Tra Chi

Male yearlings Des Tra Chi

Des

0.076 0.055 0.025 0.009 0.132 0.043 0.067 0.024 0.002 0 0,007

0.060 0.021 0.006 0.008 0.059 0.018 0.020 0.016 0.033 0.002 0.033 0.013 0.005 0.003 0.018 0.005 0.003 0.018 0 0 0 0.003 0.035 0.025

Ewes Tra Chi

Class I rams Des Tra Chi

0.028 0.026 1.23 0.004 0.013 0.49 0.007 0.035 0.13 0 0 0.31 0 0.004 0.28

0.12 0.07 0.01 0.06 0.17

0.30 0.10 0.10 0.02 0.36

Rates= behaviour patterns/hour/individual. Sample sizes are given in Figs. 3 & 4. Des= desert; Tra= transplant; Chi= Chilcotin

(1) desert males (=yearlings+ Class 1 rams) mounted more often than did Chilcotin males ( =yearlings + Class 1 rams) ; (2) desert Class 1 rams mounted more often than Chilcotin equivalents, although the differences from male yearlings from each environment was not significant ; (3) desert Class 1 rams and male yearlings engaged in more threat and contact patterns than did Chilcotin equivalents; (4) Chilcotin Class 1 rams and male yearlings displayed more often than did their desert equivalents.

For females the major significant differences may be summarized as follows: (1) Chilcotin females (both ewes and yearlings) used more rotational patterns than did desert females; (2) desert females utilized more threat and contact patterns than did Chilcotin females.

Also, females in all populations utilized more rotational and less sexual and display patterns than males (Figs 2, 3, 4).

Behavioural diversity Various sized groups engaged in social play (see Berger, in prep. a for a categorization of play). Play was characterized by the use of motor patterns from different contexts (e.g. chasing, jumping, headbutting, etc.). In addition to those patterns listed in the ethogram, rubbing and nuzzling (see Geist, 1971) occurred during play. Sequences of play began when three or more individuals engaged in some form of exaggerated locomotor-rotational or contact activity and it terminated when less than three individuals persisted in these actions. A direct relationship was found (r = 0.52; P < 0.01 ; n = 37) between group size and the number of different behaviour patterns in lambs that occurred during play (Fig. 5).

Also a significant correlation existed (r = 0.33 ; P < 0.05 ; n = 37) between the number of players and length (as measured by the number of motor actions) within a sequence. That is, the more players within a group the more acts that occurred.

Measures of behavioural diversity ( H ) were calculated for all populations by using Shannon-Wiener information theory and utilizing the formula :

J . BERGER

260

. = 0.52 P = 0.01 r

n

I1 = "

I

I

= 37

I

I

r-033 PC

0 05

n = 37

6

12

18

24

Group size

FIG.5 . The relationships between group size and (a) the number of different behaviour patterns, and (b) the number of acts (length) of play sequences of Bighorn lambs in the Chilcotin. Regressions are: (top) Y=0,33X+ 4.54 and (bottom) Y=1.77X+ 13.27.

where p i equals the proportion of motor actions observed within a population and represented by any given sex and age class. This measure of diversity was used to determine within which population it would be easier to predict the occurrence of motor actions. In all sex and age classes (i.e. lambs, female yearlings, male yearlings, ewes, and Class 1 rams), Chilcotin sheep were always more diverse behaviourally than their desert equivalents. The specific H values for these sex and age classes are: Chilcotin sheep; 0.612, 0.586, 0.552, 0.921, 0.609; desert sheep; 0.508, 0.436, 0.377, 0.666, 0.577; transplant; 0.523, 0.472, 0.356, 0.624, 09601.

Discussion Development of nzotor patterns Male and female iambs showed no significant differences in the frequency of motor

SOCIAL O NT O G E NY A N D B E H A V I O U R A L DIVERSITY

261

pattern development possibly due to small sample sizes (as a result of observational difficulty in correctly identifying lamb sexes). Males tended to use more contact, threat and sexual patterns. Studies of primates (Baldwin & Baldwin, 1971; Kummer, 1971; LeResche, 1976), ungulates (Sachs, unpublished; Byers, in prep.), and rodents (Poole & Fish, 1976) also indicated that sexual dimorphism in behaviour exists at early ages. For American sheep (Bighorns and Thinhorns, 0. dalli), Blood et al. (1970) and Bunnel & Olsen (1976) showed that males gain weight faster than females. My data illustrate that behavioural dimorphism coincides with differences in growth rates.

Social ontogeny in different environments Geist (1971) stated that Bighorn lambs are capable of butting, clashing, threat jumping, and mounting soon after birth. In each of my three study populations one week-old lambs used these patterns (Table I). However, population differences existed in the development of displays by males. Desert male lambs used the low stretch and front leg kick at a younger age (about three months) than did Chilcotin male lambs. Other than displays, behavioural ontogenies followed a similar timecourse for populations in all environments.

On the other hand, the frequency of utilization of behaviour patterns was very different between environments (see Figs 3 and 4). Desert lambs and males (=yearlings + Class 1 rams) mounted more often than did their peers in the Chilcotin. Also, desert lambs and older sheep (yearlings, ewes, Class 1 rams) used less rotational patterns than Chilcotin sheep. Since desert sheep engaged in less social and locomotor play than did Chilcotin sheep, and rotational movements communicate play intention (Berger, in prep. a), it is not surprising that these movements also occurred less frequently in the desert.

Two more important population differences existed. First, desert lambs used more contact patterns than mountain lambs, yet the former threatened less often. Contact patterns represented more than 50 % of the patterns used by desert lambs. Desert lambs were less diverse behaviourally than Chilcotin lambs. In other words, desert lambs (and other desert sheep too) had the same behavioural repertoires as Chilcotin sheep, but individuals performed fewer different patterns. Second, desert yearlings (of both sexes), ewes, and Class 1 rams used more contact patterns than Chilcotin sheep of equal sex and age classes.

Thus, again as with lambs, sheep from all three environments were characterized by the same ethograms, but desert sheep were behaviourally less diverse. They used fewer different patterns but in greater frequencies, whereas Chilcotin sheep utilized all patterns but not as often.

Population diflerences in behavioural ontogenfes At least four major questions may now be asked regarding these population differences in social development. The first two questions concern sexual behavjour and social play. They are: (1) Why was sexual behaviour more prominent in desert lambs and older males? and, (2) Why did desert sheep play less?

Desert sheep grow faster and mature sexually (at about 1h years of age) approximately one year earlier than sheep from northern environments. These regional differences appear to be genetically mediated (Berger, in prep. b). Since natural selection has most likely favoured early puberty in desert sheep, it is not surprising that more mounting occurs at early ages in desert sheep. Turner (in prep.) found that desert males are born more

262

J . BERGER

precocial than mountain lambs and, in the former, testicular descent occurs at about two to three weeks of age. In mountain lambs testes descend at about three months of age.

Although Chilcotin sheep matured sexually at a later age than sheep in the Santa Rosa Mountains, they played more. Geist (1971), Shackleton (1973), and Horejsi (1976) found that more playful lambs occurred in high quality populations which were characterized by well fed and early maturing (defined by these authors as independence) individuals than those from low quality populations. Berger (in prep. b) provided data that suggested sheep from the Santa Rosa Mountains were better nourished than those from the Chilcotin (although in the same paper data was presented incompatible with the concept of population quality). Why, then, did desert sheep play less than Chilcotin sheep, thus contradicting Geist’s predictions for population quality ?

It was suggested that play occurred less often in the Santa Rosa study population due to lack of “playgrounds”, hazards of the physical environment, and a less complex social environment (e.g. smaller group sizes) (Berger, in prep. a). Furthermore, it was suggested that greater food dispersion in the desert resulted in smaller and more widely dispersed foraging bands (Berger, in press). Consequently, lambs spent more time alone (time not with their peers or adults) and were independent at an earlier age than were transplant or Chilcotin lambs.

Group size(s) may indeed be an important variable in the development of social patterns.

Figure 5 shows that when lambs played in larger groups more different behaviour patterns occur than when play was observed in smaller groups. Even if the number of distinctly different behaviour patterns observed was a function of more individuals, the fact remains that yearlings, ewes, and Class 1 rams from the Chilcotin still made use of a greater behavioural repertoire than their conspecifics from environments with smaller group sizes.

Also, more acts per sequence were incorporated into play when groups were large.

Since group sizes in the desert are small (Berger, in press; Simmons, 1969) and lambs may be born at any time of the year (Hansen, 1965), those individuals born in a desert environment may not be exposed to nearly as complex a social environment (one with many incoming stimulii; i.e. many lambs) as those born in mountainous or more northern habitats (where group sizes are larger and parturition more synchronized).

In desert environments, Bighorn sheep lambs are not only faced with few peers but also those of different sizes and ages. During contact play, different aged lambs play infrequently (Berger, in prep. a). In Death Valley, California, probably the most arid environment inhabited by sheep in North America, Welles & Welles (1961) reported small band size and no nursery bands. Social play may be very infrequent in such an extreme and arid habitat.

Baldwin & Baldwin (1971, 1977) found that squirrel monkeys (Saimiri) in large groups played more extensively than those in small groups. Other field and laboratory studies of primates have confirmed that individuals raised in enriched social environments tend to be behaviourally less retarded than those reared in socially impoverished environments (see Hinde & Spencer-Booth, 1967; Hinde, 1974; Mason, 1961a, b, 1962). The social environment, as has been amply demonstrated, plays a large role in the behavioural development of infants.

Behavioural development and diversity Two additional major questions that may be asked about the observed differences in social ontogenies between populations, are : (3) What factors caused behavioural diversity

SOCIAL O N T O G E N Y A N D B E H A V I O U R A L DIVERSITY

263

to be greater in the Chilcotin? and, (4) What benefits may be accrued by behaviourally more diverse individuals? Before proposing explanations to these questions, the Neotenization Hypothesis of Geist (1971) will first be considered as a possible explanation for the differences in social behaviour in my study populations.

The Neotenization Hypothesis of Geist (1971) predicts that sheep occupying more recently colonized habitats act more aggressively (more contact and threat patterns than displays) than those colonized at an earlier time. In essence, the individuals within these populations will be characterized by larger horn size and will act in a neotenic manner or as if they were juvenile. Furthermore, Geist (1971) and others (see references therein) suggested that ‘deserts were colonized by sheep later than other regions in western North America (for instance, British Columbia, Washington, Oregon and California). Although it could be argued that the present data tend to support the Neotenization Hypothesis (because desert sheep did in fact act in a more juvenile manner than sheep from the Chilcotin, these are interpreted entirely differently (see below)).

Geist studied and compared different species of sheep (0.dalli and 0. canadensis) when he proposed the Neotenization Hypothesis. Certainly genetic differences do exist between different species. (The taxonomy of sheep is far from clear: see Cowan, 1940; Geist, 1971 ; Schaller, 1977.) In the Neotenization Hypothesis, Geist (1971 :332) implied that a genetic component exists for differences in behaviour in habitats colonized at various times in history. Or, he felt that behavioural differences in populations could be attributed to food differences (e.g. quality) at the time of colonization, but that the differences in behaviour still persist today. For instance, habitats recently colonized were (are) below carrying capacity, thus more food (and less intraspecific competition) occurred. Individuals within such habitats, then, were capable of growing faster and more vigorously, and delivering more forceful blows during combat. However, Geist (1971 :308), stated that the differences between his study populations of two species “. . . cannot be explained in terms of genetic differences; they are plausible only in terms of quality differences”.

From Geist’s discussion of the Neotenization Hypothesis it is not obvious why desert sheep would inherently behave more aggressively than other populations of sheep based simply on their past dispersal histories, unless genes for behavioural differentiation have become fixed within specific populations. Are more neotenic sheep (e.g. desert populations) genetically predisposed to act more aggressively than less neotenic (e.g. more northern) populations? If so, why?

I interpret differences in behavioural ontogenies of Bighorn sheep populations more parsimoniously; in terms of ecological factors rather than previous dispersal histories (or genetic differences). Thus, with regard to question (3) it appears that desert sheep were behaviourally less diverse than other sheep populations due to a number of factors that included small group size, less social facilitation, and less play. These last factors most likely were a result of a precarious physical environment while the first factor was due to the distribution of food resources.

In contrast, sheep in the Chilcotin utilized the full range of their behavioural repertoire.

There are several factors that may be responsible for increased behavioural diversity in these sheep. First, they matured sexually at a later age than desert sheep. In many mammals, prolonged sexual maturation allows for greater learning and social experience(s) (Hinde, 1974; Wilson, 1975). Secondly, large groups (above a minimum size) provide more complex social environments (Anderson & Mason, 1974; Goy & Goldfoot, 1973). Thirdly, Chilcotin

164

J. BERGER

sheep inhabited physical and social environments that were more conducive for play.

Although it has never been demonstrated experimentally in non-human mammals that additional play leads to increased utilization of different behaviour patterns, play nevertheless most probably allows individuals to refine motor skills (Rasa, 1973; Eisenberg & Leyhausen, 1972; see also Bekoff, 1976; Fagen, 1976).

With reference to question (4), it seems that natural selection would favour those individuals inhabiting complex social environments (e.g. larger groups) to develop more sophisticated communication systems (i.e. visual, vocal, etc.) in order to signal more efficiently. Marler (1975, 1976) found that primates inhabiting larger and more constant groups possessed more graded signals. For Chilcotin sheep, too, it would also appear adaptive for individuals to communicate by utilizing as many behaviours as possible. The four display patterns for which I provided data are distinctly different from one another and, although gradations do occur for each pattern, the patterns do not grade into one another (except for the low approach and twist; see Geist, 1971). Certainly selection will favour those individuals that are more adept at communicating regardless of group size.

But, in large groups, individuals who can communicate an intended message with the most precision will experience less ambiguity in reception.

In conclusion, the dfferences between behavioural ontogenies in different environments are illustrative of “open behavioral programs” (Mayr, 1974). The terms “adaptation” and “function” should be scrutinized most carefully before being used to explain biological (or behavioural) phenomena (Hinde, 1975; Tinbergen, 1965; Williams, 1966). It thus appears that although subspecies of Bighorn sheep possess a species typical behavioural repertoire (Geist, 1974), environmental conditions influence greatly and modify the timecourse(s) that social development follows. Furthermore, this study illustrates clearly that behaviour is indeed labile and it varies in different ecological settings. Before generalizations about the nature of species specific social development are valid, different populations under contrasting ecological conditions must be studied in the field.

I wish to express my thanks to Marc Bekoff for his time, support, guidance, and most of all his patience and friendship throughout my education. John Byers, Peter Stacey, and Melody Serena assisted in ways too numerous to elaborate upon. The following people assisted me in logistics while I was in the field: Vernon Bleich, Barbara Cromer, Eldon McLaury, Harold Mitchell, Stephen Walker, and John and Ann Walsh. Also, I appreciate the comments of David Armstrong, Ruth Bernstein, Carl Bock, Richard Jones, and Philip Thompson on a previous manuscript.

Kathy Walker was a constant source of inspiration during the project’s final stages. This study was aided by: the British Columbia Branch of Fish and Wildlife, Boyd Deep Canyon Research Station, California Department of Fish and Game, United States Fish and Wildlife Service, American Museum of Natural History, Sigma Xi, Council for the Preservation of Desert Bighorn Sheep, the Department of Environmental, Population, and Organismic Biology, and the Graduate School of the University of Colorado. Lastly, I wish to thank my parents and grandfather for the “care” packages they sent me and my brother, Neal, for interrupting my solitude in the field with his visits.

REFERENCES Anderson, C . 0. & Mason, W. A. (1974). Early experience and complexity of social organisation in groups of young rhesus monkeys (Macaca mularta). J . romp. physiol. Psyckol. 87: 681-690.

Baldwin, J. D. & Baldwin, J. I. (1971). Squirrel monkeys (Saimiri) in natural habitats in Panama, Colombia, Brazil. and Peru. Primates 12: 45-61.

SOCIAL ONTOGENY A N D BEHAVIOURAL DIVERSITY

265

Baldwin, J. D. & Baldwin, J. I. (1973). The role of play in social organization: comparative observations of squirrel monkeys (Saimiri). Primates 14: 369-381.

Baldwin, J. D. & Baldwin, J. I. (1977). The role of learning phenomena in the ontogeny of exploration and play. In Primate bio-social development; biological, social and ecological determinants: 343406. Chevalier-Skolnikoff, S . & Poirier, E. (Eds) New York: Garland Press.

Barash, D. P. (1974). The evolution of marmot societies: A general theory. Science, Wash. 185: 415420.

Bekoff, M. (1972). The development of social interaction, play and metacommunication in mammals : an ethological perspective. Q. Rev. Biol. 47: 412434.

Bekoff, M. (1976). Animal play: problems and perspectives. Pers. Ethol. 2 : 165-188.

Bekoff, M. (1977). Mammalian dispersal and the ontogeny of individual behavioural phenotypes. Am. Nut. 111: 715-732.

Berger, J . (in press). Weaning, social environments, and the ontogeny of spatial associations in Bighorn sheep.

Biol. Behav.

Berger, J. (in prep. a ) Social play in bighorn sheep: structure, function and ecology.

Berger, J. (in prep. b ) Evolution of breeding age in female American sheep.

Blood, D. A., Flook, D. R. & Wishart, W. D. (1970). Weights and growth of Rocky Mountain Bighorn sheep in western Alberta. J. Wildl. Mgmt 34: 451455.

Brown, J. L. (1975). The evolution of behavior. New York: Norton and Co. Brown, J. L. & Orians, G. H. (1970). Spacing patterns in mobile animals. A . Rev. Syst. Ecol. 1: 239-262.

Bunnel, F. L. & Olsen, N. A. (1976). Weights and growth of Dall sheep in Kluane National Park Reserve, Yukon Territory, Can. Fld Nut. 90: 157-162.

Byers, J. A. (1977). Terrain preferences in the play behavior of Siberian ibex kids (Capra ibex siberica). Z . Tierpsychol.

45: 199-209.

Byers, J. A. (in prep.). Maternal facilitation ofplay in ibex kids (Cupra ibex).

Clutton-Brock, T. H. & Harvey, P. H. (1977). Primate ecology and social organization. J. Zool., Lond. 183: 1-39.

Cowan, I. McT. (1940). Distribution and variation in the native sheep of North America. Am. Midl. Nut. 24: 505-580.

Crook, J. H. (1970). Social behavior in birds and mammals. New York: Academic Press.

Eisenberg, J. F. (1966). The social organization of mammals. Handbuch Zool. (Berl.) 8 (10): 1-92.

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

Eisenberg, J. F., Muckenhirn, N. & Rudran, R. (1972). The relation between ecology and social structure in primates.

Science, Wash. 176: 863-874.

Fagen, R. (1976). Exercise, play and physical training in animals. Pers. Ethol. 2 : 189-219.

Fisler, G. F. (1965). Adaptation and speciation in harvest mice of the marshes of San Francisco Bay. Univ. C a l g PublsZ001.77: 1-108.

Fisler, G. F. (1969). Mammalian organizational systems. Contr. Sci. 167: 1--32.

Geist, V. (1968). On the inter-relation of external appearance, social behaviour, and social structure of mountain sheep. Z . Tierpsychol. 25: 199-215.

Geist, V. (1971). Mountain sheep. Chicago: Univ. Chicago Press.

Geist, V. (1974). On the relationship of social evolution and ecology in ungulates. Am. Zool. 14: 205-220.

Goy, R. W. & Goldfoot, D. A. (1973). Hormonal influences on sexually dimorphic behavior. Handb. Physiol. Sect.

7. 2 : 169-186.

Graul, W. P.,Derrickson,S. R. & Mock,D. W. (1977). On theevolution of avain polyandry.Am. Nut. 111: 812-816.

Hansen, C. (1965). Growth and development of desert bighorn sheep. J . Wild/. Mgmt 29: 387-391.

Hinde, R. A. (1974). The biological bases of human social behaviour. New York: McGraw-Hill.

Hinde, R. A. (1975). The concept of function. In Function andevolution in behaviour: 1-1 5. Baerends, G., Beer, C. & Manning, A. (Eds) Oxford: Clarendon Press.

Hinde, R. A. & Spencer-Booth, Y. (1967). The effect of social companions on mother-infant relations in rhesus monkeys. In Primate ethology. Morris, D. (Ed.) London : Weidenfeld and Nicolson.

Horejsi, B. L. (1976). Suckling and feeding behavior in relation to lamb survival in bighorn sheep (Ovis cunadensis cunudensis Shaw). PhD. diss. University of Calgary, Calgary.

Kleirnan, D. G. & Brady, C. A. (1978). Coyote behavior in the context of recent canid research; problems and perspectives. In Coyotes: biology, behavior andmanagement: 163-190. Bekoff, M. (Ed.). New York: Academic Press.

366

J. BERGER

Kornet, C. (1978). Status and habitat use ofbighorri sheep on Hart Moioztain, Oregon. Masters thesis, Oregon State University, Corvallis.

Kummer, H. (1971). Primate sscieties: group techniques of ecological adaptation. Chicago : Aldine-Atherton.

Lent, P. C. (1969). A preliminary study of the Okavango leche (Kobus leche leche). E. Afr. Wild. J. 7: 147-157.

LeResche. L. A. (1976). Dyadic play in Hamadryas baboons. Behaviour 17: 190-205.

Marler, P. (1975). On the origin of speech from animal sounds. In The role of speec'h in language: 1 1-37. Kavanaugh, J. F. & Cuttings, J. E. (Eds) Cambridge: M I T Press.

Marler, P. (1 976). Social organization, communication and graded signals: the chimpanzee and the gorilla. I n Growkg points in erhology: 239-280. Bateson, P. P. G. & Hinde, R. A. (Eds) Cambridge: Cambridge Univ.

Press.

Mason, W. A. (19610) The effects of social restriction on the behavior of rhesus monkeys. 2. Tests of gregariousness.

J . romp. physiol. Pswhol. 54: 287-290.

Mason, W. A. (1961b). The effects of social restriction of the behavior of rhesus monkeys. 3. Dominance tests. J. conrp. physiol. Psychol. 54: 694-699.

Mason, W. A. (1962). The effects of social restriction on the behavior of rhesus monkeys. 4. Responses to a novel environment and to alien species. J. romp. physiol. Psychol. 55: 363-365.

Mayr, E. (1974). Behavior programs and evolutionary strategies. Am. Scient. 62: 650-659.

McBride, G. (1964). A general theory of social organization and behavior. Univ. QdPap. (Vet.Sci.) l ( 2 ) : 75-110.

Merrit, M. F. (1974). Measurement of utilization of bighorn sheep habitat in the Santa Rosa Mountains. Trans.

Desert Bighorn Counc. 18: 4-17.

Nievergelt, B. (1966). Der Alpensteinbock (Capra ibex) in seinen Lebensraurn. Berlin : Paul Parey.

Orains, G . H. (1969). On the evolution of mating systems in birds and mammals. A m Not. 103: 589-603.

Poirier, F. E. (1977). Introduction. I n Primate bio-social development; biological, social arid ecological determinants: 1-42. Chevalier-Skolnikoff, S. & Poirier, F. (Eds) New York: Garland Press.

Poole, T. B. & Fish, J. (1976). An investigation of individual, age and sexual differences in the play of Rattus norvegicus (Mammalia: Rodentia). J . Zool., Lond. 179: 249-260.

Rasa, 0. A. E. (1973). Prey capture, feeding techniques, and their ontogeny in the African dwarf mongoose (Helogale undulata). 2. Tierpsychol. 32: 449-488.

Richards, A. (1974). Intraspecific variation in the social organization and ecology of Propithecus verreauxi. Folia Pritimtol. 22: 178-207.

Sachs, B. (Unpublished). Sex and age differences in the play of domestic sheep.

Schaller, G. B. (1977). Mountain monarchs: wild sheep andgoats o f t h e Himalaya. Chicago: Univ. Chicago Press.

Shackleton, D. M. (1 973). Population quality and bighorn sheep. PhD. diss. University of Calgary, Calgary.

Simmons, N. M. ( I 969). The social organization, behavior and environment of the desert bighorn sheep on the Cabeza Prieta Game Range, Arizona. PhD. diss. University of Arizona, Tucson.

Sinclair, A. R . E. (1977). The African buffalo; a study ofresource limitation ofpopulations. Chicago: Univ. Chicago Press.

Snedecor, G . W. (1956). Statisticnl methods. Ames: Iowa State College Press.

Sokal, R. R. & Rohlf, F. J. (1969). Biometry. San Francisco: Freeman and Co. Tinbergen, N. (1965). Behavior and natural selection. Znt. congr. Zool. 16 (6): 519-542.

Walther, F. (1974). Some reflections on expressive behavior in combat and courtship of certain horned ungulates. In Behavior ofungulafesand its relrrtion to mangement: 56-106. Geist, V. & Walther, F. (Eds) (IUCN Publ. 24.)

Morges, Switzerland.

Welles, R. E. & Welles, F. B. (1961). The bighorn ofDeath Valley. Wash D.C.: U.S. National Parks. (Fauna Ser. 6.)

Williams, G. C. (1966). Adaptation and natural selection. Princeton: Princeton Univ. Press.

Wilson. E. 0. (1975). Sociobiology; the new synthesis. Cambridge: Belnap Press.