Anim. Behav., 1977, 25, 828-848

SOCIAL BEHAVIOUR WTTHIN GROUPS OF JUNGLE BABBLERS(TURDOIDES STRIATUS)

BY A. J. GASTONEdward Grey Institute, Department of Zoology, South Parks Road, Oxford

Abstract . Five aspects of intra-group behaviour among wild jungle babblers were analysed in relationto the age, sex and breeding status of the participants . The amount of participation in allopreening,sentinel behaviour, and movement initiation were found to be closely correlated with age and breedingstatus, resulting in a rough concordance between rankings based on these three activities . There wassome difference between the sexes in the degree of participation in allopreening and sentinel behaviourand this may be explained by differences in their strategies for breeding . Changes in play and roostingbehaviour with age are related to the establishment of dominance relations among birds of the yearand possible connections between other aspects of behaviour and social status are also discussed .

Birds in which offspring remain in their parents'territory for several years after fledging, forminga family flock which feeds together, breedsco-operatively, and takes part in joint territorialdefence, have been classed by Wilson (1975)as `advanced subsocial' in their organization .This group territorial pattern of social organi-zation was considered rare until recently, but isnow known to be widespread among birds ofwarm-temperate and tropical climates (Skutch1961 ; Lack 1968 ; Fry 1972 and several reviewsin the Proc. XVI Int. Orn. Cong.) . Despite theirwidespread occurrence, no detailed studies haveso far been published on social relations withinfree-living groups of this type .

The present study sets out to analyse socialrelations within groups of jungle babblers(Turdoides striatus), small insectivorous pas-serine birds (weight about 65 g), living intropical woodland and scrub . over most oflowland India. The species is similar in generalbehaviour to several species of group territorialbirds studied previously, such as the Floridascrub jay (Aphelocoma c . coerulescens) (Wool-fenden 1974, 1975), the Arabian babbler(Turdoides squamiceps) (Zahavi 1974), the white-winged chough (Corcorax melanorhamphus)(Rowley 1965 and in preparation) and thelong-tailed shrike (Corvinella corvina) (Grimes1975). Jungle babblers live throughout the yearin groups of 2 to 20 birds which defend acommon territory. Young birds remain with thegroup in which they were reared for at least 18months and dispersal after this is largely byfemales, many males remaining with their natalgroup for at least 4 years . A general introductionto the biology of the species is given by Andrews& Naik (1970) .

828

Dominance is widely recognized to be impor-tant in birds (Wood-Gush 1955 ; Crook 1961 ;Wilson 1975) and this concept has been fre-quently used in the interpretation of their socialbehaviour. Dominance has also been widelystudied among social mammals, particularlyprimates (Hall 1968 ; Rowell 1972; Richards1974), which live in groups similar in size andcomposition to those of jungle babblers and othergroup territorial birds . Several authors havecriticized the indiscriminate application of thedominance concept to situations other thanthose involving approach-retreat interactionshowever, (Rowell 1966 ; Bernstein & Sharpe1966 ; Bernstein 1970) and Jolly (1972) pointsout that there appear to be consistent differencesbetween primate species in the degree to whichdifferent measures of dominance are correlated .Bernstein & Sharpe suggest that analysis interms of sex and age, as well as status, can bemore meaningful than one based on the con-struction of a linear dominance hierarchy . Thispractice was followed in analysing the intra-group behaviour of jungle babblers .

Intra-group aggression among jungle babblerswas very rarely seen . Because of this it did notprove possible to describe dominance withingroups based on approach-retreat interactions .Even at artificial feeding sites, where food wasvery concentrated, no evidence of conflict overfood was observed, except among first-yearbirds or between members of different groups .In the absence of approach-retreat inter-

actions, or any form of overt aggression, thefollowing behaviour patterns were analysed inrelation to the age, sex and breeding status ofthe participants in order to determine the

GASTON: SOCIAL BEHAVIOUR IN JUNGLE BABBLERS

relationships between different group members :(1) Allopreening behaviour ; (2) Sentinelbehaviour; (3) Leadership behaviour ; (4) Play ;(5) Roosting behaviour .

General MethodsStudy Area

Jungle babblers were observed in an area ofabout 1 km 2 of dry deciduous woodland andtropical thorn scrub adjacent to New Delhi,N.W. India. The structure of the vegetationranged from low scrub, less than 1 m high, toclosed canopy woodland up to 15 m high,usually with an understory up to about 3 m .Jungle babblers were found mainly in thewoodland, spending most of their time feedingamong leaf-litter on the ground and sometimessearching the boles and larger branches of thetrees .

Observations were made between July 1971and June 1974 during an intensive study of theecology of the species . Most observations onintra-group behaviour were made during thewinter months (November to February) sincework in the summer was concentrated onbreeding ecology . More than 3000 hours werespent in the field, of which about half weredevoted to watching jungle babblers . Sixty-twobirds were trapped as post juveniles and markedwith coloured plastic leg-rings and a furtherseventy-seven were ringed as nestlings .

Collection of DataObservations were made through 10 x 50

binoculars at ranges from 10 to 50 m, whilefollowing groups as they moved through thestudy area. Individual groups were followed forperiods of up to 6 hr and observations duringthe winter spanned the whole daylight periodalthough more were made in the morning thanthe afternoon. Every effort was made to mini-mize disturbance by the observer but some wasinevitable . Groups which were followed regularlybecame habituated to the presence of theobserver and sometimes allowed approach towithin a few metres but catching one of the groupin a mist net immediately destroyed this habitu-ation .

All behaviour falling into the five categoriesspecified above was noted, along with the timeof day, the identity of the participants, thereactions of other group members and thepresence of any relevant extra-group stimuli .The position of the group under observationwas marked on a map of the study area at J -hr

829

intervals and the positions of all importantinter- or intra-group interactions were alsorecorded .

During periods when groups were underobservation it was usually possible to keep somebirds in view at all times, but rarely possible tosee all members of the group at once . Whilefeeding, birds were often out of sight in groundvegetation, but behaviour performed by birdswhich were perched in trees could usually bemonitored continuously because the woodlandcanopy was fairly open, particularly in winter .Most allopreening and sentinel behaviour whichtook place during a period of observations wasprobably noted, although the identity of theparticipants could not always be seen. Leaders,however, could only be identified in a smallproportion of group movements . Play behaviourwas probably frequently overlooked .

Three groups were chosen for intensive study,and these were referred to by letters ; L, P andM. The number of birds in each group in eachof the three winters of the study is given inTable I, which also shows their age composition .Specific methods for recording different typesof behaviour are dealt with separately wheneach behaviour is introduced, but beforedealing with these the vocalizations of thejungle babbler need to be described so thatreference can be made to them, where appro-priate.

VocalizationsAndrews & Naik (1970) list eight vocalizationsof the jungle babbler and describe the situations

Table I. The Size and Composition of Three Groups ofJungle Babblers Studied intensively

Abbreviations used in Tables and Figures: 1Y, bird lessthan 1 year old ; 2Y, bird more than 1 year and less than2 years old ; A, adult bird more than 2 years old ; B,breeder ; BB, breeding pair ; NB, non-breeding adult.

Group P Group L Group M

1971-72 IY 51 12 21 931

102Y and Adult 7 7 77

1972-73 lY 1 5 22Y 4 10 2 13 217Adult 5 6 3

1973-74 2Y 109

35115 2 9

21

Adult 8 7 5

830

in which they are used . A list made indepen-dently in the present study includes 11, of whichthree intergrade . Two of those described byAndrews & Naik are included here under oneheading, while two which they do not dis-tinguish are separated . Vocalizations distin-guished in the present study are listed in TableII ; names used by Andrews & Naik are given inparentheses .

The shriek call of the jungle babbler is similarto the calls given in similar circumstances by thecommon babbler Turdoides caudatus and thebulbuls Pycnonotus spp. which occupied thesame habitat. The cackle calls, although dis-tinct from the other vocalizations described,formed a continuum . The chack call wassimilar in form to the cuk call, but these twodid not intergrade in volume, the chack callalways being audible up to at least 50 m, thecuk call not usually being audible beyond 15 m .The latter may have been used more frequentlythan was apparent since only a few groupsallowed prolonged approach to within 15 m . Allthe other vocalizations listed appeared to

Table IL Vocalizations Recorded Among Jungle Bubblers ; the Names Used by Andrews & Nalk (1970) are Given in Brackets

Name

(1) Shriek call(danger call)

(2) Cackle call, type (a)(exciting call)

type (b)(2(b) and 2(c) includedunder mobbing calltype (c)

(3) Chack call(contact call)

(4) Cuk call(feeding call)

(5) Cu-cu-cu call

(6) Kya call(distress call)

(7) Gurgle call

(8) Scheer call

(9)(bBegging call

egging call)

Description

ANIMAL BEHAVIOUR, 25, 4

Short, high-pitched shriek

Low volume, intermittentwheezing

Loud, wheezy cackle

As 2(b), but including a rattling,guttural noise, rising and fallingslightly in pitch

Short 'chack' given at irregularintervals

Short, low 'cuk', givenintermittently

Low, guttural `cu-cu-cu', usuallyrepeated several times

Loud, ringing 'kya-kya-kya'

Low gurgling noise

Low 'scheer'

Rattling squawk

constitute relatively discrete signals, with littlevariation in volume or duration, although callsseven and eight were heard only a few times .

Vocalizations two and three were the mostfrequently recorded. The intergradation of thethree types of cackle call, all used quite fre-quently, probably allows greater flexibility incommunication than would be the case if theywere discrete (Struhsaker 1966), allowing quan-titative as well as qualitative information to betransmitted (Konishi 1963). Frequent inter-gradation is characteristic of the calls of primates,compared with those of birds (Mulligan &Olsen 1969) and it is also true of the vocalizationsof hyaenas Crocuta crocuta (Kruuk 1972) andlions Felis leo (Schaller 1972), both highly socialspecies .

Allopreening and ClumpingMethods Used to Record Allopreening BehaviourAn allopreening interaction was recorded

whenever one bird was judged to be attemptingto allopreen another, bringing its bill into firmcontact with the feathers . When one bird pecked

Situations in which call was used

Given by any bird at the unexpected appearance of apredator . The group immediately seeks cover and mayfreeze .Given by sentinel birds in the presence of potentialdanger, as yet remote. Often given at the approach ofthe observer. Usually no visible reaction by the group.Given by part or all of the group in the presence of aground predator or perched raptor .

Given by the entire group during inter-group con-frontations .

Given by birds detached from their group, or in replyto other chack calls. Sometimes given while sitting onthe nest.Given by any bird while feeding, but not often heard.

Given by adult birds just prior to moving to a newfeeding site . Frequently given while moving to roost,and when fledglings were being encouraged to fly .Given when trapped. This call usually summoned therest of the group at once . Tape recordings of this callwere the most effective in attracting jungle babblers .Given by males just prior to an attempted copulation,wings waving loosely at the same time.Given by females while approaching males of anothergroup. Also given by passive birds during `rough andtumble' (see below) .Given by juveniles while begging. May be used oc-casionally up to 6 months old.

another lightly on the head or neck the gesturewas interpreted as an invitation to allopreenbecause it was often followed by other solicitingmovements (see below) . For all interactionsobserved the identity of the participants wasrecorded and also the part of the body that waspreened. For the latter purpose the body wasdivided into four areas, as illustrated in Fig . 1 .

Birds performing allopreening on others weredescribed as the active participants and thosereceiving allopreening were the passive par-ticipants. Figures given for frequency ofinvolvement include all interactions in which aparticular bird took part, whether as the activeor passive partner . On many occasions morethan two birds were participating simultaneouslyin allopreening . Any period of time during whichtwo or more birds were continuously involved inallopreening was known as a `bout' and withineach bout each pair allopreening constituted aninteraction .

Priority was given to establishing the identityof birds engaged in allopreening and as a result,when more than two birds were taking part in about, it was not possible to record the durationof individual interactions. An allopreeninginteraction was considered complete when thetwo birds separated or when no allopreeningtook place for 1 min. If one bird moved toanother perch and was followed by the other,which then resumed allopreening, the resump-tion was treated as a new interaction, eventhough less than 1 min had elapsed betweenthe cessation and resumption of allopreening .

GASTON: SOCIAL BEHAVIOUR OF JUNGLE BABBLERS

831

Fig. 1 . Parts of a jungle babblers scored separately inrecording allopreening behaviour . Hn = head and neck,Br = breast and flanks, Ba = back, mantle, rump andwings, Ut = belly and under-tail coverts .

Each area of the body preened was scored onlyonce in each interaction .

Description of Allopreening BehaviourAllopreening took place between birds per-

ched in bushes or trees, and only very rarely onthe ground. In most cases the two birds involvedperched flank-to-flank on a branch with theirheads pointing in the same direction . Thepassive participant usually crouched low on thebranch, with the legs bent, raising its head andstretching its neck, while the active bird stoodup on straight legs . Some typical poses, copiedfrom field sketches, are illustrated in Fig. 2 .Occasionally the active bird perched below thepassive, on a lower branch, and reached upwardto preen from below. This happened mostfrequently when the passive bird was already

4000-AFig. 2 . Poses adopted by jungle babblers during allo-preening (redrawn by E. K. Dunn from field sketches).

832

ANIMAL BEHAVIOUR, 25, 4

being allopreened by another bird perchedbeside it .

Andrews & Naik (1970) briefly describe thetypical posture of a bird being allopreened, buta fuller description seems warranted. Duringallopreening of the head and neck the feathersof this region are erected and the head is tiltedaway from the preener to expose the chin, whileat the same time the eyes are narrowed or closed .The passive bird usually remains completelystill, apart from occasional shifts to exposedifferent areas to the preener. During allo-preening of the back, breast, or flanks, thefeathers of these regions are also erected,particularly those around the preen gland, butthe active bird does not move its bill deliberatelyfrom the preen gland to other parts of thefeathers in the way that might be anticipated ifit were oiling the feathers . Active birds give theimpression of searching for something amongthe feathers and short jabs of the bill are oftenused, presumably to remove particles of dirt, orperhaps ectoparasites. Fletcher & Inglis (1924)mention that a bird soiled with bird lime usedin its capture was allopreened particularly onthe soiled area, suggesting that allopreeningmay have a real value in feather hygiene .Running the feathers through the bill, describedby Sparks (1964) for the estrildine Amandavaamandava was not observed among junglebabblers.

Allopreening was actively solicited in somecases, not all of which were followed by preeningand the actions involved were similar to thosedescribed by Morris (1956) for estrildine finches .The solicitor approached the prospective preenerby sidling along the branch on which it wasperched until their flanks were in contact . It thenadopted the posture used during allopreening ofthe head and neck regions, with the feathers ofthese areas erected and the neck stretched awayfrom the bird being solicited . Feathers of otherparts of the body were never raised prior to thecommencement of allopreening and this may bedue to the fact that the body feathers are raisedduring inter-group agonistic diplays .

ClumpingClumping, where two or more birds perched

side by side in contact with one another, wascommon during periods of rest. This situationwas sometimes followed by the initiation ofallopreening and birds would also autopreenwhile clumped. A bird approaching another in

order to clump would usually approach a birdperched and at rest, or autopreening, but noparticular feather posture appeared to inviteclumping, unlike the estrildines described byMorris (1956). Occasionally birds attempted toclump with a bird performing sentinel behaviour(see below) and giving a low intensity cackle call .

Babblers approaching an isolated bird did notperch directly beside it but landed 5 to 10 cmaway and sidled up to clump as described. Birdsjoining a pre-existing clump of two or more,however, often flew to perch right beside theclumped birds or even perched on their backs .The vigour with which birds attempted to join aclump seemed to be positively correlated withthe number of birds in the clump, and clumpedbirds appeared to exert a stimulus for others tojoin them. Groups invariably clumped whileroosting.

Practically all instances of clumping or allo-preening involved birds belonging to the samegroup, despite the fact that in winter groupsoften mixed together for several hours at a timewhile feeding . The only instance of inter-groupallopreening recorded, apart from some associ-ated with attempts to form new groups, involvedthe breeding male of one group and a juvenilebird from another. Juveniles were allopreenedvery frequently by all group members .

Seasonal Variations in Allopreening FrequencyThe seasonal distribution of allopreening bouts

and the number of birds involved per bout areshown in Fig. 3, which demonstrates thatallopreening is more frequent during the winterthan during the summer. Breeding takes placefrom March to October. The observed dis-tribution cannot be related to the amount oftime available for allopreening because inwinter there is little time for any activities otherthan feeding. Instead the decline in the frequencyof allopreening interactions during the summerappears to reflect a general loss of sociabilitywithin babbler groups during the breedingseason, perhaps due to a higher level of aggres-sion between adult birds during the period whenthey are presumably in competition for thechance to breed .

Allopreening was observed at all times of day.In the winter the frequency of bouts rose duringthe day to a peak in mid-afternoon, but duringthe rest of the year no definite pattern wasapparent (Fig . 4). Many allopreening boutsoccurred immediately after the group hadsuffered some kind of disturbance: an inter-

GASTON : SOCIAL BEHAVIOUR IN JUNGLE BABBLERS

group confrontation, a predator alarm, ordisturbance by the observer, but this could nothave accounted for the observed distribution ofallopreening bouts because group interactionsand disturbance from predators both showedpeaks in early morning and late evening through-out the year. The peak of allopreening frequencyin mid-afternoon during the winter was probablydue to a slackening of feeding efforts at thattime of day. During the morning more than95 % of the time was devoted to feeding orsentinel behaviour .

Allopreening in Relation to Age, Sex and BreedingStatus

Four measures of allopreening involvementwere used to analyse differences between birdsof different age, sex and breeding status classes .(1) The total number of times that a bird tookpart in allopreening interactions, whether as theactive or passive participant . (2) The ratio of thenumber of times that an individual was the activeparticipant in an allopreening interaction to thenumber of times it was the passive participant ;

0J F M A M/J J A S 0 N D

Hours 68 83 164102 11 54 48 33 27 37 134

Fig. 3 . Changes in the frequency of allopreening inter-actions with time of year . `Hours' indicates number ofhours of observation in each month . + mean number ofbirds involved per allopreening bout ; • mean number ofbirds allopreening per hour ; p mean number of allo-preening bouts per hour.

this is called the `preen ratio' . (3) The relativefrequency of allopreening interactions betweenmembers of particular classes . (4) The relativefrequency with which allopreening was directedat different parts of the plumage.

Data were collected in all 3 years of the study,but the age categories considered (first year,second year and adult) were only identifiable inthe 1972 to 1973 and 1973 to 1974 seasons, andonly figures for these seasons are used in thisanalysis, which is confined to the three groupsspecified above (P, L and M) .

Table III shows the total number of involve-ments recorded for each member of these threegroups. Some observations were included inwhich only one of the two participants in theinteraction could be identified, but in most casesboth birds were known. During the 1972 to 1973season L group included two unringed birds,one of which was the breeding male and theother a non-breeding adult male . Scores forunringed birds are divided equally between thesetwo. Only birds that remained with their groupthroughout the entire winter are included .

Individuals showed little consistency in therelative frequency with which they were involvedin allopreening . One non-breeding adult memberof P group (PLG) allopreened markedly less

Nov. - Feb.March-Oct.

1 . 5

L

OL

a1. 0N

0

0)CCX0 .5a

833

05 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Time (I .S .T.)

Fig. 4. Changes in the frequency of allopreening inter-actions with time of day. Winter (November to February)and summer (March to October) are shown separately .

Ages in the left-hand column refer to 1972-73, add one year for 1973-74.

than the other two non-breeding adults duringthe 1972 to 1973 season, but in the 1973 to 1974season the position was reversed. The sameapplied to LBB, a non-breeding adult in L group .In both years, however, there was a fairlyconstant tendency for older birds to be involvedin more interactions than younger birds . Figure5 shows the percentage involvement of differentclasses in all allopreening interactions within thethree groups . For classes that contained more thanone bird the figures are expressed as mean per-centages per bird . In all cases one of the breedingpair was involved in the highest percentageof interactions, and in four cases the breedingpair were first and second . The mean rankingsfor the five classes shown in Fig. 5 are as follows(1) breeding male, (2) breeding female, (3) non-breeding adults, (4) second years, (5) first years .

Table IV shows the number of interactionsbetween birds of different classes and theirfrequency relative to the number of birds of eachclass available within the group . The data usedare the same as for Table III, except that onlyinteractions in which both participants could beidentified are included . Interactions betweenmembers of the breeding pair were the mostfrequent combination, followed by interactionsbetween the breeders and birds of any other class .Apart from these, non-breeding adults wereinvolved in more interactions with other non-breeding adults and second years had more inter-actions with second years, than any othercombinations . The least frequent combinationswere non-breeding adults with first years,

BdB9NB2Ylv

second years with first years and allopreeningbetween first years .

Table V arranges the data in Table IV toshow the active and passive roles of each classand the areas of the body that were preened .The preen ratios for each combination of classesare also given. Breeding adults were more oftenthe active partner in interactions with all otherclasses. Otherwise birds tended to be morefrequently the active partner in interactionswith birds younger than themselves and moreoften passive with birds older than themselves .In all interactions between different age classesthe older birds took the active role significantlymore often than the younger (X2 values and

0

20

0 20

20

1

-1I I 1

HI

Ch

rn

P gp.

L gp.

M gp.

Fig. 5 . Mean percentage of all group allopreening inter-actions in which individuals of five classes were involved .B = breeding adult, NB = non-breeding adult, 2Y =second-year, 1Y = first year . Percentages for each classare expressed `per bird'.

834 ANIMAL BEHAVIOUR, 25, 4

Table III. Frequency of Involvement of Different Individuals in Allopreening Interactions

P group 1972-73 1973-74 L group 1972-73 1973-74 M group 1972-73 1973-74

B d PDB 96 34 UN 101 49 MB 22 13B ~ PM 95 39 LBO 124 25 MDB 52 1NB PW 49 13 LBDB 73 9 UN 15 11

PDG 43 31 LBB 28 14PLG 22 33 LBW 77 9

UN 1002Y PY 33 19 LBY 38 6 MO 20 1

PO 37 18 LBM 75 32 MP 20 11PB 12 9YB 27

lY PR 18 23 LGR 16 36 MM 19 5LGP 36 20 MLB 43 9LGW 11 16DGLG 13 9DGDB 46

1Y (1973 cohort) LGY 4 UN (2) 7LBLB 5DGM 0

GASTON : SOCIAL BEHAVIOUR IN JUNGLE BABBLERS

835

probabilities given in Table V) .The percentage of allopreening directed at

different parts of the body in interactions bet-ween all classes is shown in Fig . 6. Breedingadults preened other birds more frequently onparts of the body other than the head and neckthan was the case when roles were reversed .Birds taking the active role with those older thanthemselves directed a larger proportion of theirallopreening at the head and neck region thanthey did when preening those younger thanthemselves . A significant inverse correlation wasfound between the preen ratios between differentclasses and the ratios of the proportion ofallopreening directed at the head and neck(C/D in Table V, rs = -0 .79, N = 7, P<O .05) .This indicates that, where two birds of differentclasses were engaged in allopreening, the birdthat was more frequently the passive partnertended to direct a greater proportion of itsallopreening to the head and neck region whenit was the active partner than did the individualthat was more frequently the active partner .

The ratio of the number of times that allo-preening was directed at the head and neckregions to the number of times that it was direc-ted at other parts of the plumage was tested by2 x 2 x2 tests for the combinations of classeslisted in Table V . Proportions differed sig-

nificantly in the cases of breeders/non-breedingadults and breeders/second years . Sample sizeswere inadequate for testing first years againstany other class . A comparison of breeders withnon-breeding adults showed that breedersallopreened first and second year birds sig-nificantly more often on parts of the body otherthan the head and neck region. When allo-preening interactions within classes were com-pared for the relative amount of preeningdirected at the head and neck no significantdifferences were found except between breedersand second year birds, where the breeding pairwere found to allopreen one another less oftenon the head and neck.

Soliciting of AllopreeningAlthough many instances of probable solicit-

ing for allopreening were observed only a fewwere considered to qualify unequivocally assoliciting. In many cases allopreening appearedto be initiated by the act of clumping, but as thiswas not always the case, and as clumping mightoccur for other reasons (e .g. roosting) these werenot counted. In other instances where theapparent solicitor raised the feathers of the neckprior to the initiation of allopreening, it wasoften impossible to see whether this was theresult of some intention movement on the part

Table IV. Frequency of Allopreening Interactions Between Birds of Different Classes

Comb. = the number of possible pairing combinations between the classes concerned . N = number of interactionsobserved.

P group L group M group

Comb. N N/C Comb . N N/C Comb . N N/C

1972-3BB : BB 1 37 37 1 36 36 1 9 9BB : NB 6 54 9 8 105 13 . 1 2 6 3BB : 2Y 10 60 6 4 20 5 4 25 6 .2BB : lY 2 7 3 . 5 10 73 7 .3 4 26 6 .5NB : 2Y 15 27 1 . 8 8 43 5 .4 2 5 2.5NB : 1Y 3 14 4 . 7 20 46 2 .3 2 6 32Y : 1Y 5 2 0 .4 10 17 1 . 7 4 18 4.5lY : lY 0 10 3 0 .3 1 6 62Y : 2Y 10 18 1 . 8 1 7 7 1 0 0NB : NB 3 20 6 .7 6 36 6 0 7 . 2

1973-4BB : BB 1 19 19 1 16 16 1 0BB : NB 12 25 2 . 1 10 16 1 . 6 6 16 0 . 7BB : 2Y 2 10 5 10 26 2 . 6 4 0 1 . 8BB : 1Y 0 6 0 4 3 0.2NB : 2Y 6 8 1 .3 25 38 1 . 5 6 11 0.5NB : 1Y 0 15 5 0 .3 6 12Y : 1Y 0 15 6 0 .4 4 2 21Y : lY 0 3 2 0 .7 1 0 0 .72Y : 2Y 0 10 12 1 .2 1 2NB : NB 15 41 2 . 7 10 6 0 .6 3 2

8 3 6

ANIMAL BEHAVIOUR, 25, 4

of the bird which was apparently being solicited .Only instances where the solicitor adopted theusual passive allopreening posture without anytrace of response from the bird being solicitedwere included in this analysis . The typicalsoliciting posture is shown in Fig . 7. Because ofthe relatively small sample size, figures for allthree groups and both years are combined inTable VI which shows the frequency with whichdifferent classes were seen to solicit or besolicited and the frequencies in relation to thenumber of bird-years involved .

First year birds accounted for 50 % of allsoliciting behaviour although they comprisedonly 21 % of the birds under observation, while69 % of all soliciting was directed at breedingadults which comprised only 19% of the birds .In general, the amount of soliciting behaviour

Table V. Preen Ratios and Relative Frequency of Preening Different Parts of the Body During Allopreening InteractionsBetween Different Age and B

Status Classes; Totals for P, L, and M Groups, 1972-73 and 1973-74 Seasons . Thective Participant is on the Left In the First Column

exhibited was inversely correlated with thenumber of allopreening interactions in which aclass was involved .

Discussion of Allopreening BehaviourHarrison (1965), in a review of allopreening

behaviour, noted that the active bird in allo-preening interactions was usually the dominant,in situations where dominance could be estab-lished by other means. Among jungle babblersbreeders, which can be assumed to be dominantto other group members on the ground that theyhave priority of access to a scarce resource(i.e. reproduction), and first year birds, form theopposite ends of a continuum of allopreeninginvolvement . This also suggests that the directionof allopreening reflects dominance relationswithin the group .

Comparisons of the proportion of allopreening directed at the head and neck region (Hn) in interactions between differentclasses.

x2 P X2 PBB :2Y / NB :2Y 5 .8 <0 .025 BB : BB / 1Y : IY 0 . 03 <0 .8-0 •9BB : lY / NB : lY 5 .8 <0 .025 NB : NB / 2Y : 2Y 1 . 7 <0 .1-0 •2BB : BB / NB : NB 2 .4 <0 .1-0-2 NB : NB / 1Y : lY 0.3 <0 .5-0 •7BB : BB / 2Y : 2Y 6 .3 <0 .025 2Y : 2Y / 1Y : 1Y 1 . 8 <0 . 1-0 •2

Areas allopreened

TotalsPreenratio x2 P

Hn asPercentageof total C/D X2 PFin Br Ba Ut

B CT : B Y 23 14 11 4 52 1 .73 5 .4 <0 .02 44(C) 1 . 1 0 . 1 0 . 7B 9 : B 6 12 7 6 5 30 40(D)BB : NB 55 29 26 23 133 2 .14 13 .4 <0 .001 41(C) 0 . 6 14 . 8 <0 . 001NB : BB 44 10 5 3 62 71(D)BB : 2Y 35 25 23 20 103 2 .78 16 . 5 <0 .001 34(C) 0 .5 10 .6 <0 .012Y : BB 24 5 3 5 37 65(D)BB : lY 25 16 18 12 71 35 .50 42 .0 <0 .001 35(C) 0 .3 3 . 5 <0 . 1IY : BB 2 0 0 0 2 100(D)NB : 2Y 40 18 10 9 77 2 .40 9 .7 <0 .01 52(C) 0 . 8 2 . 6 <0 .22Y : NB 22 6 1 3 32 69(D)NB : 1Y 23 6 7 3 39 6 . 50 14 . 0 <0 . 001 59(C) 0 .7 1 . 3 <0.3IY : NB 5 0 1 0 6 83(D)2Y : 1Y 21 9 7 3 40 13 .30 19 . 5 <0 .001 52(C) 0 . 5 2 . 5 <0 .21Y : 2Y 3 0 0 0 3 100(D)

Areas allopreened Hn as

NB : NBTotals

84

percentageof total

54Fin46

Br

Ba Ut718

132Y : 2Y 25 5

4 3 37 68IY : lY 5 3

3 0 11 46

410 171

138 100 819

Feather erection in the head and neck regionis described by Hardy (1974) as inviting allo-preening and also forming the typical submissiondisplay among neotropical jays of the genus

Hn Br Ba Ut

Hn Br

8a U1

% 501

50

SO

50

L

% 5 ,

1

% 50

% 50

(a)

Bd B9 (52)

B9 NB (133)

B9 Bd (30)

1 1)BB :2Y (103)

III,

BB

1Y (71)

I1

I1

% 50 1

2Y :1Y (40)

NB : NB (64)

GASTON : SOCIAL BEHAVIOUR IN JUNGLE BABBLERS

837

100 I

1

100

I

NB 88 (62)

2Y B6 (37)

BB

(2)

2Y N9 (321

1Y : NB

(6)

1Y : 2Y

(3)

2Y 2Y (37)

1Y : 1Y

(11)

II

Fig. 6. Percentage of allopreening directed at differentparts of the body by birds in different age/status classes .The first-named bird is the active participant . Figures inbrackets are sample sizes .

Cissilopha. His pictures show that the posesadopted by this genus are similar to those seenin Turdoides and his description of their allo-preening technique also sounds similar . In thejays however, feather erection is only sometimesfollowed by allopreening . In the timaliinespecies Garrulax leucolophus, which is closelyrelated to Turdoides, observations made oncaptive birds also showed that feather erectionin the head and neck region served both tosolicit allopreening and to express submissionduring interactions over food (personal obser-vation). These observations all demonstrate aconnection between allopreening and the ex-pression of dominance. Among estrildine finchesno consistent direction of allopreening can beobserved (Kunkel 1974), but in this case mostallopreening is between members of mated pairs .

Allogrooming is a frequent activity amongsocial primates (Marler 1965 ; Sparks 1967 ;Jolly 1972) and the frequency of involvementand proportionate participation in active andpassive roles have been found to be related tosocial status in a number of cases (Hall &DeVore 1965; Hall 1967; Kaufmann 1967 ;Kummer 1968). The tendency seen in birds andsome rodents (Armitage 1962 ; Barnett 1963) fordominant animals to be the active participantmore often than subordinates is reversed in thecase of primates, with subordinates usuallytaking the active role . Jolly (1972) gives anexample of active participation expressingdominance in the thicktailed bushbaby (Galagocrassicaudatus) however, and primates may varyin this respect.

Fig. 7 . Soliciting of allopreening in the jungle babbler ; (a) the bird on theleft solicits ; (b) the other bird responds (redrawn by E . K. Dunn from fieldsketches) .

(b)

8 3 8

ANIMAL BEHAVIOUR, 25, 4

The use of allopreening soliciting postures toexpress submission in some birds has ledHarrison (1965) to suggest that allopreeningbehaviour evolved originally to reduce aggressionby redirecting the behaviour of the potentialaggressor. Sparks (1967) also supported thisidea. The active soliciting of allopreening byjungle babblers in situations where they approachthe potential preener suggests that in this speciesallopreening behaviour has evolved beyond thepoint where it serves to reduce aggression in aspecific situation, to one where soliciting andreceiving allopreening helps to maintain anindividual's position within its group . If this iscorrect, then allopreening can be consideredas a factor in promoting group cohesion, as wellas a means of expressing social relationships .Similar ideas have been put forward to explainthe importance of allogrooming in primates(Sparks 1967 ; Jolly 1972). The fact that allo-preening involvements increase with age and thatallopreening between first-year birds is therarest category of allopreening interaction,suggest that allopreening in jungle babblers isnot initially involved in the creation of socialhierarchies, but simply expresses relationshipsalready established by other means .

Both allopreening and clumping behaviourhave generally been discussed from the point ofview of the social relations of the participants .The possibility also exists that allopreening isintended partly as a means of communicatingthe status of the birds involved to other groupmembers. If the group is viewed as an intenselycompetitive arena, as suggested by Zahavi(1974), and if coalitions are an important factorin determining individual status, as has beendemonstrated for some primates (Hall &DeVore 1965; Jolly 1972 ; Kaufmann 1967 ;Southwick & Siddiqi 1967), lions (Schaller1972) and the turkey (Meleagris gallopavo)(Watts & Stokes 1971), then it may be of

Table VI . Frequency of Soliciting and Being Solicited forAllopreening in Relation to Age and Breeding Status ;P, L, and M Groups, 1972-73 and 1973-74 Seasons

advantage for subordinate animals to demon-strate their relationship with other groupmembers, particularly the dominant(s), byallowing themselves to be allopreened/groomed .In the case of the jungle babbler, where mostsubordinates are probably offspring of thebreeding pair, competition among them woulddamage the interests of the breeders. If thedominant protects low-ranking non-breedersagainst aggression by higher-ranking birds thenit is in the interest of the low ranking birds todemonstrate to other group members the degreeto which they are accepted by the dominant(s)and this may be achieved by clumping and allo-preening with the dominant . This hypothesiswould explain why most soliciting is directed atthe dominant and why clumping exerts astimulus for other birds to clump .

Sentinel BehaviourThe pattern of behaviour in which one memberof a babbler group remains perched severalmetres above the ground, while the rest of thegroup feed below, has been described previouslyby Dharmakumarasinhji (1954) and Andrews& Naik (1970), the latter authors employing theterm `sentinel' to describe the bird remainingperched. This pattern of behaviour was seencommonly among jungle babbler groups, beingparticularly noticeable in winter .

MethodsDuring the winter of 1971 to 1972 the identity

of all sentinels that could be recognized wasrecorded and also the timing of change overs .In the following winters the identity of thesentinel was recorded every 5 min but noattempt was made to time exchanges . Sentinelbehaviour was easiest to observe in winter when,particularly in the early morning, there wasnever more than one member of the groupperched at a time. The situation in summer wasusually less clear because occasions on whichthe whole group was feeding simultaneouslywere less frequent than in winter. For muchof the time more than one bird remained perchedin trees and although it was sometimes apparentthat one bird was keeping more alert than othersthe selection of a sentinel on this basis appearedopen to subjective bias . Accordingly only recordsof sentinel behaviour collected during the winterwere used in the following analysis .Results

A sentinel was recorded as present during 82of the time spent observing jungle babbler

Solicitor SolicitedBirdyears(C)

A/C B/C(A) (B)

B d` 19 6 3 . 2B ? 27 6 4 . 5NBA 18 16 22 0 . 8 0 . 72Y 16 4 16 1 .0 0.2lY 33 1 13 2 .5 0 . 1

groups in December and January, but thisrepresents a minimum figure since sentinelsmust sometimes have been overlooked . Birdsexhibiting sentinel behaviour usually perched inthe lower branches of the woodland canopywhere they could be easily seen by the rest of thegroup. Figure 8 shows the distribution ofheights at which they perched against a typicalvegetation profile.

The mean duration of periods of sentinelactivity timed during December 1971 andJanuary 1972, in relation to time of day, areshown in Fig. 9 which demonstrates that meanduration increased from dawn to mid-afternoon,possibly decreasing again before dusk . Change-overs between sentinels were effected in threedifferent ways : (1) The sentinel flew down tojoin the group and another bird then moved upto replace it (not necessarily on the same perch) .(2) A new bird flew up and perched near thesentinel which then descended . (3) A new birdflew up to perch near the sentinel which thenbecame less alert and often began to autopreen,usually descending after a few minutes .

The lack of a fixed pattern of events at changeover suggests a flexible response to differentsituations. In winter, particularly in the earlymorning, the sentinel often began to give a lowintensity cackle call in the minute before it leftits perch and this was sometimes followed byanother bird flying up to take its place . In theseinstances the call appeared to constitute anannouncement that the sentinel was about todescend.

GASTON : SOCIAL BEHAVIOUR IN JUNGLE BABBLERS

839

4 0

80

Fig. 8. Height above the ground of birds exhibitingsentinel behaviour in relation to the average vegetationprofile .

Table VII shows the frequency with whichmembers of L, P and M groups were recordedas sentinels during the winters of 1971 to 1972and 1972 to 1973 . Figures for L group in 1971 to1972 include only January and February as onenon-breeding adult was not ringed until Decem-ber. For groups including more than oneunringed bird records were apportioned equallybetween the number of unringed birds known tohave been in each age class.

In five out of six cases the bird observed mostfrequently as sentinel was one of the followingseason's breeding pair. In M group in bothyears this was the female and in P group in bothyears, the male. In L group, where in 1971 to1972 a non-breeding male (LBB) was apparentlythe most frequent sentinel, the breeding malewas unringed and this bird might actually haveperformed more sentinel behaviour than LBB .In 1972 to 1973, when there were only twounringed birds in L group, the breeding malewas identifiable at close range by plumagecharacteristics and 66% of records where theunringed sentinel could be identified referred tothe breeding male. Table VIII shows the meanamount of sentinel activity performed by

Mins

12

8

4

008

10

12

14

16

18Time HST)

N

15 03 85 57 30 20 .

3 23 26 13

Fig. 9. Changes in the duration of sentinel behaviourwith time of day during December 1971 and January1972. Vertical bars are standard errors . (N = samplesize) .

840

ANIMAL BEHAVIOUR, 25, 4

different age classes, expressed as percentages .In general the breeding pair performed about50 % of sentinel activity and the rest was sharedbetween non-breeding adults and second-winterbirds, first-year birds rarely taking part .

Among non-breeding adults and second-yearbirds there seemed to be a tendency for males toperform more sentinel behaviour than females .Among second-winter birds in P group in1972 to 1973 the two males performed more thaneither of the two females and in M group themale second-year performed more than thefemale. In L group a non-breeding male per-formed more than the breeding female in bothyears, but in P group in both years, among non-breeding adults a female performed mostsentinel activity . In 1971 to 1972 this was BRand in 1972 to 1973, PDG and in both cases theyperformed more sentinel activity than the breed-ing female (PM) . These three females were

Table VU. Frequency of Sentinel Behaviour Among theMembers of Three Groups of Jungle Babblers . Recordswere made at 5-min Intervals

apparently rivals for the position of breeder .PDG was breeding female in the spring of 1972,being succeeded by PM in the summer, whileBR was driven out of the group by PDG afterdestroying the eggs in PDG's first nest in 1972 .Since these three were all apparently potentialbreeders in P group, it seems likely that all wereimmigrants from outside the group . In babblersfemales do not normally breed in their natalgroup (Gaston 1976) .

Changes in the mean duration of sentinelwatches during the course of the day suggestthat the amount of time spent in sentinel activitymay be determined, at least in part, by the abilityof the birds involved to refrain from feeding. Ifso the amount of sentinel activity performedmay reflect the relative food-finding abilities ofdifferent group members and this should applyparticularly at times of year when food is leastabundant and feeding time most limited. Underthis hypothesis, if the breeding pair were superiorin food-finding to other group members, theymight be expected to take a larger share ofsentinel activity during December and January,the months of lowest temperatures and shortestdays, than during the rest of the winter (Novem-ber and February).

Table IX compares the amount of time spentin sentinel activity by the breeding pair and bynon-breeders in different months . For 1971 to1972 December and January are compared withFebruary and for 1972 to 1973 December iscompared with November and February,because no data were obtained in November1971 and January 1973. Figures for P group in

Table VIII . Mean Amount of Sentinel Behaviour Per-formed by Individuals of Each Class Expressed as Per-centages of the Total Records for the Group. Only Ringed

Birds are Included

Numbers in brackets show the numbers of birds involved .

P group L group M group

1971-72B d 39 - 18B 9 - 17 33NB and 2Y 18(2) 12(3) 13(3)lY 1 (1) 1(2) 1(3)

1972-73B d 25 - 24B 9 9 12 39NB 8(4) 11(3) 13(l)2Y 6(5) 5(2) 12(2)lY 1 (1) 1(5) 0(2)

P group

L group

M group

1971-72B d

PDB 51 UN 42 MB 64B ?

UN 5 LBO 49 MDB 11 8

NB and 2Y PLG d 9 LBB d 60 MLG d 42BR Y 38 LBDB d 22 MW ? 49UN(2) 9 LBW ? 22 MR? 45

UN(2) 84 UN(2) 32

lY PO CT 1 LBY d 1 MO d 2UN(4) 18 LBM d 3 MP 9 1

MY? 2

131 283 355

1972-73B d PDB 61 UN 42 MB 18B ? PM 23 LBO 23 MDB 29

NB PLG d 18 LBB 8- 28 UN 10PDG 30 LBDB d 17PW ? 13 LBW Y 15UN? 21 2UN d 41

2Y PO d 24 LBY d 5 MO d 16PB S 18 LBM 64 13 MP ? 2Py 9 12YB 9 4PLB? 17

1Y PR d 2 LGW 2 two 0four 0

243 186 75

GASTON: SOCIAL BEHAVIOUR IN JUNGLE BABBLERS

841

1971 to 1972 are omitted because of the largenumber of unringed birds in the group thatwinter. In four out of the five cases the breedersundertook a higher proportion of sentinelactivity during December and January thanduring November and February, but this wasnot true for M group in 1972 to 1973. When allfive sets of data were pooled the difference wasfound to be significant (P<0 . 01) .

Discussion of Sentinel BehaviourEvidence from the change in the duration of

sentinel watches during the day and changes inthe proportion of sentinel behaviour performedby the breeding pair both suggest that the amountof time devoted to sentinel behaviour may berelated to birds' ability to find food . If thisability is also related to social status within thegroup then sentinel behaviour may provide aclue to status . Since sentinel birds usually perchin clear view of other group members and oftengive vocalizations at the same time it seemspossible that the role of sentinel may be adoptedpartly to advertise the status of the bird con-cerned. This hypothesis provides an explanationfor the observation that females in their natalgroup performed less sentinel behaviour thanmales, since there is no apparent evolutionaryreason why they should be concerned to displaytheir social status in a group in which they cannotbecome a breeder. Among males, however,competition can be anticipated for the privilegeof succeeding the breeding male and hence areason exists to express their social status byperforming sentinel behaviour. The same hypo-thesis might also explain why immigrantfemales, competing for the position of breeder,should perform sentinel behaviour frequently,as was observed in the case of PM, PDG and BR .

Another hypothesis that might explain dif-ferences in the frequency of sentinel behaviouris that breeders have more at stake in the sur-vival of the group than non-breeders, since ahigh proportion of the group consists of theiroffspring. The difference in frequency betweennative male and female non-breeders could thenbe accounted for by the idea that males have agreater interest in maintaining the strength ofthe group so that the size and quality of theterritory to which they succeed may be preserved .Under this second hypothesis, however, it ishard to explain the high frequency with whichfemales unrelated to the group performedsentinel behaviour and it is necessary to assume

that this was the result of an accident of samp-ling.

LeadershipFrom the beginning of the study an attempt wasmade to discover whether group movementswere initiated or directed by particular birds,but unequivocal evidence was rare . Groupscrossing open spaces between trees did so insingle file and it was sometimes possible toidentify the leading bird. This did not imply thatthe leader had initiated the movement, or thatthe other birds were following it, but simplythat the group was moving from A to B and thata certain bird was more eager, or less cautious,than the rest . Where some outside stimulus wasapparent, such as the challenge of a nearbygroup, no implication of leadership was requiredto explain group movements .

In a few instances it was possible to assign theintitiation and orientation of a movement to aparticular bird . The criteria used to identify suchan initiation were: that the initiating birdmoved before the rest of the group showedsigns of doing so, that it gave the `cu-cu-cu' call,that the whole group followed, that the move-ment covered more than 40 m, and that therewas no apparent outside stimulus orienting themovement. The last condition was difficult toverify because the birds' ability to detect calls oftheir own species was probably greater than thatof the observer, but in most cases where birdsmoved in response to audible calls from neigh-bouring groups subsequent events would havemade the cause of the movement clear . Recordsof the initiation of movements which fulfilledthese criteria are given in Table X .

In groups where the breeding male was ringed,this bird initiated 48 % of all movements, and ingroups where the breeding female was ringedthis initiated 33 % . Nearly all other records wereof non-breeding adults, with a few records ofsecond-year birds. The relative contribution ofthe breeding pair is probably biassed by thelarge sample of observations for P group, wherethe breeding male initiated movements sig-nificantly more often than the breeding female .In M group in the 1971 to 1972 and 1972 to1973 seasons the breeding female initiated moreoften than the male and in the latter season thisdifference was significant (P = <0 .0l). Althoughthe breeding male of L group was notidentifiable in the 1971 to 1972 and 1972 to 1973seasons it is likely from the number of initiationsmade by unringed birds that the contributions

842

ANIMAL BEHAVIOUR, 25, 4

of the breeding pair were approximately equal .This small sample of pairs suggests that onaverage there is probably little difference betweenthe sexes in the frequency with which membersof the breeding pair initiate movements, but thatthere may be significant differences betweenmembers of a particular pair, depending onindividual qualities .

Order of MovementWhen a group was moving in single file it was

sometimes possible to observe the order oftravel . In most cases the birds following im-mediately behind the leader were adults andthose last in line were invariably first-year birds .This pattern was most noticeable when onegroup was moving to have a territorial con-frontation with another. A number of tests weremade by replaying tape recordings of groupconfrontation calls about 60 m away from agroup which was feeding. The first birds toarrive in 10 out of 11 tests outside the breedingseason were members of the breeding pair,followed by non-breeding adults . The last toarrive, sometimes not appearing at all, werefirst-year birds . Tests made near nests during theincubation and nestling periods did not producethe same pattern . In two trials a member of thebreeding pair was first to arrive but in two othersnon-breeding adults arrived first, and in theremaining two, second-year birds arrived first,followed by non-breeding adults . These resultsmay reflect the fact that groups are oftenscattered widely when feeding during thebreeding season . Tests along these lines werehampered by the fact that jungle babblerslearned to ignore tape recordings after only oneor two trials .

Table IX. Comparisons of the Proportion of SentinelBehaviour Performed by the Breeding Pair In Mid-winter(December to January) and in November and February

Pooled data for all groups in both years, x2 = 6P <0 . 01 .

Comparison of Ranking by Allopreening, SentinelBehaviour and Movement Initiation

Hierarchies constructed on the basis of involve-ment in allopreening, sentinel behaviour and theinitiation of movements are similar to the extentthat all three activities increase with age and allare most frequently performed by members ofthe breeding pair . Apart from this the threehierarchies are not very similar, although someof the difference could be the result of samplingvariation. Table XI ranks the second-year andolder members of L, M and P groups accordingto observations made in the winter of 1972 to1973 and also gives the mean ranks based on allthree hierarchies . In mean rankings all but oneof the second-winter birds are equal to, or lowerin rank than, adult members of their group .This suggests that, at least up until the secondwinter, social status is closely correlated withage .In P and M groups second-winter males

ranked above second-winter females in thesentinel hierarchy and lower in the allopreeninghierarchy . This was also true for four second-winter birds in L group in the 1973 to 1974

Table X. Number of Observations of Birds InitiatingMovements of More Than 40 m

Overall percentage of initiations by each categoryB 1 81/166 = 48 %, B 9 65/194 = 33 %, NB 42/218= 19 %, 2Y 5/136 = 4*+ d ; 9, P <0 .001 .

* d ; ~, P <0 .01 .

P group L group M group

1971-72B S PDB 20 MB 8B ? LBO 9 MDB 15

BR 2NB or 2Y LBB 7 MR 4PLG I LBW 1 MW I

UnringedMLG I

Bn{, 2NB 1 B d, 2NB 10 2NB 2

1972-73B d PDB 42** MB 2*B ? PM 14 LBO 11 MDB 11NB PLG 1 LBW 2

PW 3PDG 11UN 4

2Y PB 1 MO 1PLB 3

Unringed B d, NB 12

1973-74B d PDB 5 UN 2 MB 2B ? PM 0 LBO 5 MDB 0NB PDG 2 LBDB 1 UN 1

Dec. and Jan . February x2 P

1971-72L group 35 (N = 205) 26 (N = 74) 2.02M group 52 (N = 293) 51 (N = 57) 0 . 03

1972-73 December Nov. and Feb.

P group 58 (N = 103) 38 (N = 63) 6 . 35 <0 .02L group 47 (N = 85) 31 (N = 81) 4 . 57 <0-05M group 57 (N = 44) 71 (N = 31) 2 . 53

GASTON : SOCIAL BEHAVIOUR IN JUNGLE BABBLERS

843

seasons, where the male DGLG was the mostfrequently involved in sentinel behaviour, butthe least often involved in allopreening. Theother second-winter male, LGW, was alsoinvolved in allopreening less than the females,but did not participate in sentinel behavioureither.

The following hypothesis might account forthe observed discrepancy between the allo-preening and sentinel hierarchies . Non-breedingmales are anxious to advertise their social statuswithin the group since their chance of succeedingto the breeding territory, or of pairing with animmigrant female, may depend on this . Con-sequently they spend as much time as possible

Table XL Comparisons of Ranking Within JungleBabbler Groups Based on Sentinel Behaviour, Initiationof Group Movements, and Preen Ratios. Only BirdsMare Than 1 Year Old are Included . Winter of 1972 to

1973

Kendall coefficients of concordance (W) for the threerankings :

P group, W = 0 .73, x2 = 19 .8, df 9, P <0 . 02 .L group, W = 0 . 46, x2 = 7 . 0, df 5, P <0.3 .M group, W = 0 . 82, x2 = 9 . 8, df 4, P <0 .05 .

performing the public display of competencewhich sentinel behaviour represents . Femalesstill in their natal group are less concerned toadvertise their status because they will leavebefore breeding . In the course of allopreening,however, which they perform in order toconsolidate their position within the group,many of these females rank higher than malesof the same age and hence occupy higherpositions in a hierarchy based on allopreening .

PlayTwo patterns of behaviour seen regularly amongjungle babblers seemed to come under theheading of play. These were named respectively`rough and tumble' and `mad flight' .

Rough and tumble behaviour consisted of twoor more birds engaging in a mock fight in whichsome lay on the ground more or less passively,while others rolled on top of them, or peckedthem deliberately but gently . At least four birdsusually took part and although it was alwaysdifficult to identify the participants, most werefirst-year birds, while breeding adults were neverobserved to take part. Ten instances of thisbehaviour were recorded among the studygroups ; three in August, two in December, andone each in September, February, March,April and May. Similar behaviour was alsoobserved on several occasions among groups oflarge grey babblers (Turdoides malcolmi), and forthis species the behaviour had also been notedpreviously by Hutson (1954) .

Mad flights consisted of one or several birdsflying rapidly and apparently aimlessly amongthe branches of a tree, twisting and turning inacrobatic manoeuvres . This display was neverseen in the open and was thought at first to be aleaf-bathing exercise because it was observedafter rain. Subsequent observations in dryweather refuted this hypothesis. Like rough andtumble behaviour, most of the birds identifiedperforming mad flights were less than 1 yearold and adults were never seen to indulge .Some mad flights occurred after aggressive

interactions between members of the samegroup. In other cases a mad flight by one birdwas followed by other birds performing thesame behaviour and in one case a mad flight bytwo birds together gave the appearance of amock chase .

The eight instances of mad flights recordedamong the study groups all occurred betweenJuly and November, with five in August andSeptember . Ten out of the 14 birds involved that

Sentinel Initiation Preening Mean

P groupB d PDB 1 1 1 1B y PM 4 2 2 2

NB PLG d 5 . 5 6 . 5 3 4PDG Y 2 3 4 3PW 9 8 4 . 5 6 5 . 5

2Y PO d 3 9 8 7PB 4 5 .5 6 .5 9 8Py ? 9 9 10 10YB 10 9 5 9PLB? 7 4 . 5 7 5 . 5

L groupB? LBO 2 1 1 1

NB LBB 0- 1 4 . 5 6 3LBDB d 3 4 .5 5 4 . 5LBW ? 4 2 2 2

2Y LBY 3 6 4 . 5 4 6LBM 6 5 4 . 5 3 4 . 5

M groupBd MB 2 2 2 2B 9 MDB 1 1 1 1

NB UN d 4 4 . 5 4 4 .5

2Y MO d 3 3 5 3MP o 5 4 . 5 3 4 . 5

844

could be identified were juveniles more than 1month old and it seems possible that thisbehaviour is related to the establishment ofdominance relations among these young birds .Most instances of overt intra-group aggressioninvolved juveniles between 1 and 3 months old .Some of these were over food being solicitedfrom an older bird, but instances of apparentlyspontaneous aggression were also observedseveral times among birds of this age .

RoostingJungle babblers in the study area roosted in treesor bushes from 1 to 5 m above the ground, butusually at least 3 m up . Roosts were often indense foliage or twigs, usually on a branch2 to 5 cm in diameter, with the innermost birdclose to where it joined a larger branch .

Groups invariably clumped while roosting.Normally the entire roost lined up side by sideon the branch, all facing in the same direction,but sometimes large groups occupied twoadjacent branches, or a few birds perched onside branches at right angles to the rest .

Behaviour on Entering the RoostIn some cases all birds entered the roost

within a few seconds, flying directly to theroosting branch and settling immediately side byside. In other cases birds moved restlessly frombranch to branch, settling briefly, and thenmoving again before all the group had assembled .Perches on which the group partially assembledand then left to roost elsewhere were named`false roosts' and sometimes as many as 15 ofthese were visited, in different parts of the corearea, before the group finally settled. Birdsentering the roost often clambered over thebacks of those already perched and tried toforce themselves between birds already clumped .Such struggles sometimes continued until it wastoo dark for observation, but in many cases theydid not occur at all, entry into the roost beingquite orderly.

Figure 10 shows the mean time in relation tosunset that groups entered their final roost ineach month and also the proportion of nightson which false roosts were visited before finallysettling . In general groups went to roost earlier,in relation to sunset, in winter than during thebreeding season. The timing did not seem to beinfluenced by the occurrence of false roosts,which were most frequent in June (small sample)and in November and December.

ANIMAL BEHAVIOUR, 25, 4

By the time that birds were entering the roostit was usually too dark to identify colour rings,but in the few cases where this was possible it wasone of the breeding pair which entered first andusually non-breeding adults or second-yearbirds which were last. In some cases adultswhich arrived near the roost after other birdshad entered sat and waited for several minutesbefore joining onto the end of the roost .

Order of Perching in the RoostAn attempt was made to investigate the order

of perching in roosts by taking colour photo-graphs of the roost by flash . This was onlypossible at roosts where vegetation did notobscure the birds and where it was possible toapproach without disturbing them . Photo-graphs were taken just before dawn, so that ifthe birds left the roost as a result of the dis-turbance they had only a short while to waitbefore commencing feeding. Altogether 30roosts were photographed, of which 17 providedpictures in which one or more birds could beidentified . The results of these identificationsare listed in Table XII.Photographs taken of M group on 4 days

during October and November 1972 all showedthe same order of roosting with the breedingmale on the inside, the breeding female beside

o L 9P.

+ P 9P-

x M gp.

Fig. 10. Time of entering the roost for three junglebabbler groups in relation to the time of sunset (lowergraph), and percentage of all roosts preceded by falseroosting behaviour (upper graph), in relation to time ofyear .

GASTON: SOCIAL BEHAVIOUR IN JUNGLE BABBLERS

it, then a second-year male, two first-year birds,a second-year female, and on the outside a non-breeding adult male . The pattern in the series ofphotographs of L group taken between Novem-ber 1972 and February 1973 was less rigid,perhaps because more birds were involved, butcertain features were fairly constant . A non-breeding adult male was the outermost bird infive out of six cases . The innermost was an adultin all four cases where it could be identified andthere was a significant tendency for first-yearbirds to be concentrated in the outer third of theroost. Table XIII shows the distribution offirst-year birds in the roost during December1972 and February 1973 .

Only a few birds were recognizable in the threephotographs of P group, but in each case theoutermost bird was a non-breeding second-yearor adult male and in two cases juveniles weretogether in the centre of the roost. The positionof the juveniles in the middle of the row wasalso observed in several roosts which were notphotographed. In these instances the juveniles,between 2 and 6 weeks old, entered the roostlast by squeezing into the centre of the group .The two pictures of L group which includedjuveniles (11 November 1972 and 3 September1973) also showed a tendency for them toconcentrate in the centre or towards the inside .

845

From the evidence of the photographs it seemsreasonable to conclude that the position of birdsin a roost is determined, at least in part, by theirage. Juveniles first perch in the centre of theroost, moving to the outer third, though notthe extreme edge, at between 2 and 3 months ofage, and then moving to the edge or the innerthird as the next generation of juveniles arrives .The outermost position is nearly always occupiedby a non-breeding second year or adult male . Thebreeding pair are usually on or near the inside .The change in the position of the first-year birdsat 2 or 3 months of age may be connected withthe generally unsettled behaviour, includingfalse roosting, which was observed at roosts inNovember and to some extent in December,corresponding to broods fledged in August andSeptember.

The most typical configuration of roosts, withadult birds at either end of the row and juvenilesin the centre, seems to confer protection andthermal insulation on those members of thegroup which should be most in need of it .Unsettled behaviour in the roost seems to beassociated with lack of discipline among youngerbirds and this in turn, like play behaviour, maybe connected with the establishment of domi-nance relations among the year class .

Table XII. Positions of Birds Identified in Photographs of Jungle Babbles Roosts. Numbers 1-x Refer to the Order ofRoosting, the First Being Closest to the Trunk of the Tree . Abbreviations : J, Juvenile ; IS, First Summer ; 2S, Second Summer ;

2/A, Second Year or Adult ; A, Adult ; Others, Usual Conventions

1

2 3 4 5 6 7 8 9

10 11

12 13

14 15

16

L group1972 Nov. 11 ?

? ? ? 1Y IY 2/A IY A ? lY ? NB d B Y NB (TDec. 12 ?

? ? ? ? ? ? ? IY A ? ? 2/A ? NB dDec. 27 A 1Y 2Y T A IY A B 9 1Y IY NB d 1Y NB a NB ?

1973 Feb. 3 ?

? ? 2Y a ? lY ? ? lY ? ? ? ?

?Feb. 5 A A NB Y 2Y a NB a A A lY 1Y 1Y IY A lY NB aFeb. 7 A A 2/A 2Y d B Y lY 2/A A A lY IY lY lY NB aFeb . 9 A d NB d A a 2Y S IY 1Y A d 2Y S B Y IY IY lY NB 9 NB dAug. 17 ? ? ? ? ? ? ? ? NB d IS A cl? is

? is

?Sep . 3 B 9 1 S J J J ? 2S S A d NB d? ? ? ?

? ?

Mgroup1972 Oct . 3 B d B Y 2Y d 1Y 1Y 2Y ? NB d

Oct. 26 B d ? 2Y d lY ? ?

?

a

Nov. 9 B d B~ 2Y d IY 1Y 2Y Y NB o-Nov. 11 B S B Y ? IY IY 2Y Y ?

P group1973 Apr . 12 ?

? ? ? J J

?

? ? ? 2Y dJul . 17 ?

? ? ? 1S 2S d NB aSep. 14 ? ? J J ? 2S o-

846

ANIMAL BEHAVIOUR, 25,4

General DiscussionIntra-group behaviour among wild birds hasbeen studied in a number of species but none inwhich groups persist for such a long time asthose of the jungle babbler . The closest approachto the situation observed in babblers is providedby winter flocks of black-capped chickadeesParus atricapillus documented by Glase (1973)and Smith (1976), although these break upduring the breeding season . In these flocksunilateral dominance hierarchies based onapproach-retreat interactions and aggressivebehaviour can be discerned and overt aggressionappears to be much more common than withingroups of jungle babblers. Dominance hier-archies have also been observed in wild flocksof house finches Carpodacus mexicanus (Thomp-son 1960 ; Kalinoski 1975) and Harris' sparrowZonotrichia querula (Rohwer 1975), but in somepasserines which regularly form flocks noconsistent dominance hierarchies can be ob-served (Ellis 1966 ; Kalinoski 1975). Somepasserines also show a tendency towardssite-related dominance, as in Steller's jayCyanocitta stelleri (Brown 1963) .Jungle babblers differ from all the above-

mentioned species in the almost total absenceof aggressive behaviour within groups, makingthe presence of a peck-right or approach-retreathierarchy difficult to determine . Posture displaysin intra-group situations were unusual, the onlycommon one being the soliciting posture forallopreening. Although a wide variety ofvocalizations could be employed, most of themwere rarely used and the only ones heardfrequently in intra-group situations were thecackle call (2) and the chack call (5) .

Despite their lack of overt aggressive behavi-our jungle babblers showed complete dominancein relation to reproduction, since only one pairin each group bred . Observations on junglebabblers and the related common babblerTurdoides caudatus suggested that new breedingmales were recruited from within the group,while new breeding females were immigrants .Competition can therefore be anticipated among

Table XIII. Distribution of First-year Birds WithinL Group Roost, December and February only

non-breeding males for the opportunity toachieve breeding status . In the few cases inwhich succession was observed the transitionwas quite orderly and without overt aggression,suggesting that the issue had been decided inadvance, and this points to the presence of acryptic dominance hierarchy, at least amongmales. The situation among females is moredifficult to interpret but since they take part tosome extent in all the same social interactionsas males the presence of a hierarchy amongfemales is also possible.

The intra-group behaviour of the junglebabbler is perhaps better compared to that ofsocial mammals rather than other birds whichhave been studied so far. In this context theabsence of clear-cut approach-retreat behaviourwithin groups of jungle babblers is surprisingbecause most groups of primates which havebeen studied show these to a certain extent, atleast among males (Jolly 1972) and so do wolvesCanis lupus, which resemble babblers par-ticularly in that only one pair within the packprobably breeds (Mech 1970) . Dominance inter-actions are relatively rare among lions (Schaller1972) and hyaenas (Kruuk 1972), although somehierarchy can be observed, while amonghunting dogs Lycaon pictus no hierarchy can bediscerned at all (Kuhme 1965) . Despite theirlack of overt dominance interactions huntingdogs resemble babblers in that reproduction islimited to a small proportion of the adultpopulation .

Some of the difference between species whichshow clear dominance hierarchies based onapproach-retreat or peck-right behaviour andthose in which this type of dominance cannot beobserved, may be the result of differences in thelength of time that species have been adaptingto a group situation and the degree of flexibilityin social structure required by their ecology.Hunting dogs, like jungle babblers, can beconsidered obligate group-living animals, sincethey always occur in packs and their system ofhunting depends on this (Kruuk 1972). Semi-terrestrial primates on which most studies ofprimate behaviour in the field have been carriedout, tend to be relatively flexible in the type ofsocial configuration that they adopt. Speciessuch as geladas Theropithecus gelada may occurin one-male or multi-male groups according toenvironmental factors (Crook 1966). Hamadryasbaboons Papio hamadryas maintain one-malegroups during the day but join into large groupsto sleep (Kummer 1968). Langurs Presbytis

Inner third Middle third Outer third18 11 9 Adult and 2Y2 11 13 lY

X2 = 11 .7, df 2, P <0.01 .

GASTON: SOCIAL BEHAVIOUR IN JUNGLE BABBLERS

847

entellus occur in heterosexual and all-malegroups in the same area (Jay 1965) . The price ofmaintaining this flexibility may be the need tomaintain overt dominance interactions as anormal part of the social repertoire .Wolves belong to a genus which otherwise

adopts pair-territorial behaviour (Wilson 1975)and their pack-hunting habits may have evolvedrelatively recently . The same is true of lions,except that the genus is otherwise largelysolitary. By contrast hunting dogs stand rela-tively isolated, taxonomically, except for thegenus Cuon which apparently shares their habits(Kleiman & Eisenberg 1973) and they haveprobably been adapting to group-living for along time. This almost certainly applies to thegenus Turdoides in which all of the 20-oddspecies appear to live in groups of three or morebirds except the somewhat aberrant T. nipalensis(Ali & Ripley 1971 and personal observation) .

The differences in ordering between therankings based on allopreening, sentinel be-haviour and movement initiation suggest thatthese rankings may be based on independentbehavioural mechanisms . A similar conclusionregarding different behavioural hierarchies ob-served within groups of primates has beenreached by Rowell (1966) and Bernstein (1970) .In the case of the jungle babbler these differencesare likely to be related to differences in thestrategies of native and immigrant females andamong native birds to differences in strategybetween the sexes regarding the maintenance ofthe group and its common territory .

The development of socialization in the junglebabbler seems to follow a predictable pattern .During the immediate post-fledging period theyoung birds spend most of their time obtainingfood from other group members and avoidingdetection by predators . This phase lasts about1 month, after which they begin to move withtheir group and do most of their foraging forthemselves. At this stage they begin to indulgein play activities and to move outwards from thecentre of the roost. The next few monthsconstitute an `adolescent phase', during whichthey show signs of social indiscipline such aschasing adults with food, pecking one anotheraggressively, and disrupting the roost . Indiscip-line declines during the first winter and by theirfirst summer their behaviour is similar to thatof older group members, although they occupythe lower end of the rankings of sentinel behav-iour, allopreening and movement initiation .

The play activities observed among junglebabblers come under the heading of social play(Loizos 1967) because they usually involve morethan one bird, although mad flights weresometimes performed alone . Social play iscomparatively rare in birds, but is exhibited bymost mammals. Among primates play is mostintense during the weaning and sub-adultperiods (Loizos 1967 ; Chance & Jolly 1970)which correspond, in terms of social status, withthe immediate post-independence period in thejungle babbler . It seems likely that this `adoles-cent phase', during which play activities aremost frequent, is the period when social relationsare established between peers and young birdsare integrated into the social structure of theirgroup. If this is the case then play may be amechanism through which social status withinthe group is determined. Carpenter (1964) con-sidered that this was one of the functions ofsocial play in primates, and Mech (1970) alsosuggested this for wolves .

AcknowledgmentsFor financial support during the field-work onwhich this paper is based I would like to thankthe Leverhulme Trustees, the Royal Societyand the British Ornithologist's Union . Forfacilities made available to me during my stayin Delhi I should like to thank Professor C . M .Dass of the University of Delhi . I am verygrateful to Dave Dawson, Euan Dunn, JohnKrebs and Christopher Perrins for criticizingearlier drafts of the paper, and I should alsolike to thank Amotz Zahavi for fruitful dis-cussions on the general behaviour of groupterritorial birds.

REFERENCESAli, S . & Ripley, S . D . 1971 .Handbook of the Birds of India

and Pakistan, Vol . 6. London and Bombay : O .U.P.Andrews, M. I . & Naik, R . M. 1970 . The biology of the

Jungle Babbler. Pavo, 8, 1-34.Armitage, K . B. 1962. Social behaviour of a colony of

the yellow-bellied marmot (Marmota fiaviventris).Anim. Behav., 10, 319-331 .

Barnett, S. A. (1963). A Study in Behaviour. London :Methuen.

Bernstein, I . S . 1970. Primate status hierarchies. In :Primate Behaviour (Ed. by L. A. Rosenblum),p. 71. New York : Academic Press .

Bernstein, I. S. & Sharpe, L. G. 1966. Social roles in arhesus monkey group . Behaviour, 26, 91-104 .

Brown, J. L. 1963 . Aggressiveness, dominance and socialo

tion in the Steller's Jay . Condor 65,

Carpenter, C. R. 1964. Naturalistic Behaviour of Non-human Primates. Pennsylvania State UniversityPress.

Chance, M. R . A. & Jolly, C. J. 1970. Social Groups ofMonkeys, Apes and Men . London : Jonathan Cape .

848

Crook, J. H. 1961 . The basis of flock organisation inbirds. In : Current Problems in Animal Behaviour(Ed. by W. H. Thorpe & C . L . Zangwill), pp .125-149 . Cambridge University Press .

Crook, J. H. 1966. Gelada baboon herd structure andmovement : a comparative report . Symp. zool .Soc. Lond., 18, 237-258 .

Dharmakumarsinhji, R. S . 1954. Birds of Saurashtra .Bombay: Private .

Ellis, C. R. 1966. Agonistic behaviour in the maleStarling . Wilson Bull., 78, 208-224.

Fletcher, T. B. & Inglis, C. M. 1924 . Birds of an IndianGarden . Calcutta : Thacker, Spink & Co .

Fry, C. H. 1972. The social organisation of Bee-eaters(Meropidae) and co-operative breeding in hot-climate birds . Ibis, 114, 1-14 .

Gaston, A. J. 1976. Factors affecting the evolution ofgroup territories in Babblers (Turdoides) andLong-tailed Tits . D.Phil . thesis, Oxford University .

Glase, J. C. 1973 . Ecology of social organisation in theBlack-capped Chickadee. Living Bird, 12, 235-267 .

Grimes, L. 1975 . Co-operative breeding behaviour andfamily relationship within groups of the long-tailed Shrike Corvinella corvina. Emu, 74, 309-310 .

Hall, K. R. L . 1967 . Social interactions of the adult maleand adult females of a Patas monkey group . InSocial Communication among Primates (Ed. by S .Altmann), pp. 261-280. University of Chicago Press .

Hall, K. R. L . 1968. Social organisation of old worldmonkeys and apes . Symp. zool. Soc. Lond., 14,265-290.

Hall, K . R. L. & DeVore, I. 1965 . Baboon social beha-viour. In : Primate Behaviour (Ed. by 1. DeVore), pp .53-110. New York: Holt, Rinehart & Winston .

Hardy, J. W. 1974. Behaviour and its evolution inneotropical jays (Cissilopha). Bird Banding, 45,253-268 .

Harrison, C. J. O. 1965. Allopreening as agonisticbehaviour. Behaviour, 24,161-209 .

Hutson, H . P. W. 1954. The Birds about Delhi. Delhi :Delhi Bird Watching Society .

Jay, P. 1965 . The Common Langur in North India. In :Primate Behaviour (Ed . by I. DeVore), pp . 197-249,New York : Holt, Rinehart & Winston .

Jolly, Alison 1972 . The Evolution of Primate Behaviour .New York: Macmillan.

Kalinoski, R. 1975 . Intra- and interspecific aggression inHouse Finches and House Sparrows . Condor, 77,375-384.

Kaufmann, J. H. 1967. Social relations of adult males in afree-ranging band of Rhesus monkeys . In : SocialCommunication among Primates (Ed . by S.Altmann), pp . 73-98 . University of Chicago Press .

Kleiman, D. G . & Eisenberg, J. F. 1973. Comparisons ofcanid and felid social systems from an evolutionaryperspective. Anim. Behav ., 21, 637-659.

Konishi, M. 1963 . The role of auditory feedback in thevocal behaviour of the domestic fowl. Z. Tier-psychol., 20, 349-367 .

Kruuk, H. 1972. The Spotted Hyaena . Chicago : Univer-sity of Chicago Press.

Kuhme, W. 1965. Freilandstudien zur Soziologie desHyaenenhundes (Lycaon pictus lupinus Thomas1902). Z. T ierpsychol., 22, 495-541 .

Kummer, H. (1968) . Social Organisation of HamadryasBaboons. New York and Basel : S . Karger.

Kunkel, P . 1974. Mating systems of tropical birds .Z. 7Ferpsychol., 34, 265-307.

ANIMAL BEHAVIOUR, 25, 4

Lack, D . 1968. Ecological Adaptations for Breeding inBirds. London: Methuen.

Loizos, C. 1967 . Play behaviour in higher primates : areview. In : Primate Ethology (Ed. by D. Morris),pp. 176-218 . London : Weidenfeld & Nicolson.

Marler, P . 1965. Communication in monkeys and apes.In : Primate Behaviour (Ed. by I. DeVore), pp .544-584. New York: Holt, Rinehart & Winston .

Mech, L. D. 1970. The Wolf. The Ecology and Behaviourof an Endangered Species. New York : TheNatural History Press.

Morris, D. 1956 . The feather postures of birds and theproblem of the origin of social signals . Behaviour,9,75-113 .

Mulligan, J. A. & Olsen, K . C. 1969 . Communication incanary courtship calls. In : Bird Vocalisations(Ed. by R. A. Hinde), pp. 165-184. CambridgeUniversity Press .

Richards, S . M. 1974. The concept of dominance andmethods of assessment. Anim. Behav ., 22,914-930.

Rohwer, S . 1975 . The social significance of avian winterplumage variability . Evolution, 29, 593-610 .

Rowell, T . 1966. Hierarchy in the organisation of acaptive baboon group. Anim. Behav.,14, 430-143 .

Rowell, T. 1972. The Social Behaviour of Monkeys.Harmondsworth, Middx. : Penguin .

Rowley, I. 1965. White-winged Choughs . AustralianNat . Hist., 15,81-85 .

Rowley, I. (in prep).Schaller, G . B. 1972. The Serengeti Lion . Chicago :

University of Chicago Press.Skutch, A. F . 1961 . Helpers among birds. Condor, 63,

198-226.Smith, S. M. 1976. Ecological aspects of dominance

hierarchies in Black-capped Chickadees . Auk, 93,95-107 .

Southwick, C. H. & Siddiqi, M . R. 1967 . The role ofsocial tradition in the maintenance of dominancein a wild rhesus troup. Primates, 8, 341-354 .

Sparks, J. H. 1964 . Flock structure of the Red Avadavatwith particular reference to clumping and allo-preening . Anim . Behav., 12, 125-136 .

Sparks, J. H . 1967. Allogrooming in primates : a review.In : Primate Ethology (Ed. by D. Morris), pp .148-175 . London : Weidenfeld & Nicolson.

Struhsaker, T. T. 1966. Auditory communication amongVervet monkeys (Cercopithecus aethiops). In :Social Communication among Primates (Ed. byS. Altmann), pp . 281-324. Chicago : Universityof Chicago Press.

Thompson, W. L . 1960. Agonistic behaviour of the HouseFinch, Part II : Factors in aggressiveness andsociality . Condor, 62, 378-402.

Watts, C . R. & Stokes, A . W. 1971 . The social order ofTurkeys. Scient . Am., 224,112-118 .

Wilson, E . O. 1975 . Sociobiology . Cambridge, Mass . :Harvard University Press .

Wood-Gush, D. G. M. 1955. The behaviour of thedomestic chicken : a review of the literature .J. Anim. Behav., 3, 81-110 .

Woolfenden, G . E. 1974. Nesting and survival in apopulation of Florida Scrub Jays . Living Bird, 12,25-49.

Woolfenden, G. E . 1975 . Florida Scrub Jay helpers atthe nest. Auk, 92,145 .

Zahavi, A. 1974. Communal nesting by the ArabianBabbler. Ibis, 116, 84-87.Received 18 June 1976 ; revised 16 July 1976 ;

MS. number: