SOCIAL BEHAVIOUR OF A COLONY OF THE YELLOW-BELLIED

MARMOT (Marmota flaviventris)

BY KENNETH B. ARMITAGEDepartment of Zoology, The

The yellow-bellied marmot has many char-acteristics that lend themselves to the study ofsocial behaviour and its role in the ecology of aspecies. These animals are diurnal and can beobserved directly, they can be trapped readilyand marked, and they live in relatively compactgroups .

This study was conducted from June 18th,through August 30th, 1955 ; July 12th throughAugust 24th, 1956; June 10th through August23rd, 1957 ; and June 14th through August 21st,1961, as project number 82 of the Jackson HoleBiological Research Station . Only adult be-haviour will be described unless otherwiseindicated .Dr. L. Floyd Clarke, Director of the Jackson

Hole Biological Research Station, providedliving and research facilities . I wish to thank themany members of the National Park Servicewhose co-operation made it possible to conductthe study inside national parks where theanimals are not molested .The study was supported by the New York

Zoological Society, the Gans Fund of BethanyCollege, West Virginia, and the National ScienceFoundation .

Study AreaThe marmot colony selected for detailed

study is located inside the southern boundary ofYellowstone National Park along a river ter-race (Fig. 1) on the east side of the Snake Riverand slightly south of the junction of the LewisRiver with the Snake River at an elevation ofabout 6,880 feet . Old bridge foundations formthe central point of colony activity. The colonyextends about 1,500 feet in a north-southdirection and about 500 feet in an east-westdirection, covering about 17 acres . The areawas mapped by taking compass sights betweenburrows and pacing off the distance . Eachburrow-system was numbered (Fig . 2) .

The flat east of the bank was covered with alow-growing sagebrush interspersed with grasses .The bank had only a few grasses and forbs(Fig. 1). The flood plain west of the bank and

University of Kansas, Lawrence

319

south of a line drawn between burrows 48 and 60(Fig. 2) was covered with a dense growth oftimothy and white clover (Trifolium repens) .

The colony is split into two sections by a smallstream of warm water flowing from hot springsnorth of the colony . The stream usually formed abarrier to movement in June, but thereafteranimals were able to cross on three small logsthat protruded from the bottom of the streamwest of B 1 (the capital B plus a number will beused throughout the paper to designate a par-ticular burrow-system).

MethodsThe marmots were observed from a tower

erected at the top of the bank northeast of B1(Fig. 2). The activity and position of eachanimal were recorded every hour on a map ofthe colony and other observations were re-corded in a notebook.

Animals were trapped with wire traps manu-factured by the National Live Trap Company ofTomahawk, Wisconsin. A variety of baits,apples, honey, peanut butter, clover and oats,was tried. Oats and clover were more successfulthan the others and were used exclusively afterthe first year. Each animal was colour markedfor field identification with Nyanzol A fur dye(Fitzwater, 1943) and was given permanentidentification by means of a size 3 self-piercingstrap tag. During the four years of the study therewere three known cases of tags being lost . Twoof these occurred in retrapped animals whosnagged the ear tag on the wire of the cage .Other animals were observed with tags hangingloose. Whenever such an animal was retrapped,it was retagged .

Organization of the ColonyBurrow SystemThere were two types of burrows in the

colony. One type I designated the home-burrow .The home-burrow is where an animal normallyspends the night, the place to which it willretreat when danger threatens, and the burrowwhere the young live. The home-burrow appears

320

ANIMAL BEHAVIOUR,

to have a strong psychologicalvalue, as a marmot will makeevery attempt to return to itshome-burrow when an alarmcall is sounded . I designatedall other burrows in the colonyas auxiliary burrows . Theyserve as places of refuge whenan animal is unable to returnto the home-burrow . There isno certain way to distinguisha home-burrow from an auxil-iary burrow except by theactivity of the animals . How-ever, home-burrows alwayshave at least • three openingsand auxiliary burrows com-monly have two and sometimesonly one opening .

The burrows are connectedby a trail system . The trailswere formed by the marmotsrepeatedly travelling the samepath so that the vegetationwas eliminated . Once formed,the trails provide a familiarand direct route of travel .Their significance is illustratedby the following experiment .Most trapped animals weremarked near a home-burrowand when released, dashedinto the nearest opening . Someanimals were trapped nearauxiliary burrows or in trailsand when released, dasheddirectly to an entrance . How-ever; when a marmot was re-leased away from a trail, theanimal bounded in any direct-ion, first to the left, then to theright, until it came across atrail, then dashed directly to aburrow. The burrow which itentered frequently was not theone closest to it, but was deter-mined by which trail the animalhappened to find . The trailseventually lead to feeding areas .Couch (1930) also reportedtrails between the burrowsand the presence of "tempor-ary burrows under isolatedrocks or excavations". These

X, 3-4

KEYO BURROW SYSTEM

---- TRAIL

ROCK OR ROCK .PILE

Q= LOG OR LOGPILE

Fig . 2 . A map of the colony showing the burrow systems, trails, major rocks andlog piles and major feeding areas east of the hot creek . North is at the top of the page,

ARMITAGE : SOCIAL BEHAVIOUR OF A COLONY OF THE YELLOW-BELLIED MARMOT

PLATE V

Fig. 1 . A view of the colony to the south from B16 under the log in the foreground . The Snake River is at the upper rightcorner. The hot creek that divides the colony into East and West sections flows from the right centre to uppermiddle of the picture and passes between the logs making up B13 on the left and the logs making up B21 on the right .The meadow to the right of the hot creek is clover and timothy . The flat above the bank on the left is covered with alow-growing sagebrush and a few scattered grasses .

Anim. Behav., 10, 3-4

ARMITAGE: SOCIAL BEHAVIOUR OF A COLONY OF THE YELLOW-BELLIED MARMOT 32 1

temporary burrows seem to be the same asthe auxiliary burrows described above. Trailsor pathways linking burrows also were reportedfor the striped ground squirrel (Evans, 1951)and for M. marmota (Bopp, 1954, 1955b, 1956) .Distributions of Adults and Litters

Fourteen of 19 litters over the four years wereconcentrated in the central part of the colony atburrows 1, 13, 14, and 21 (Fig. 2, Table I) .These burrows are centrally located and allhave either logs or large rocks under which theburrows are located. There were no logs orrocks at burrow 30, where there was one litterTable I. The Distribution of Adults, Yearlings and Litters

for Each Year of Study .

A-adults. Y-yearlings. L-Litter .

in four years . There were litters in two of thefour years at burrows 35 and 58 . These also haverocks under which the burrows are located,but they are on the periphery of the colony.The location of adults shows the same centralconcentration (Table I). Although some of theperipheral burrows were inhabited by adults,only one of these had adults all four years of the

study. Therefore, the most successful (as meas-ured by persistence of use) burrows are thosewith large objects under which the marmotsburrow and are those located in the central partof the colony. A similar grouping of burrowshas been observed in the red squirrel (Layne,1954), and the striped ground squirrel (Evans,1951) and in the alpine marmot of Europe(Bopp, 1954) .Sex Ratio

Of 33 adults whose sex was identified, twowere males. Only one male was identified in anyyear. Thus the colony consists of a male with aharem. Sex ratio of 54 young, determined in1956, 1957, and 1961 was 1 . 16a :1 y for thethree years .

Home RangeThe home ranges of eight females in 1957

(Fig. 3) show three patterns . Three home rangesare essentially separate, several have slight over-lap, and several have major overlap. Burt (1943)considered overlap to be neutral range ; however,"neutral" range does not adequately describethe complex relationships among marmots . Forexample, the home range of ?TEST was almostcompletely overlapped by the home range of?635. During June, frequent encounters be-tween these animals showed Y635 dominant toNEST, who usually left the area. By July apattern was established so that the two utilizedessentially the same home range, but occupieddifferent parts of the home range at the sametime, or visited the same area at different times .The temporal separation of dominant and sub-ordinate animals utilizing the same home rangealso was shown by other marmots .The shape of the home . range seemed to

depend primarily on the location of the feedingarea utilized by the animal . The home ranges ofthe animals east of the hot creek were roughlyflask-shaped ; the home range of 9627 was oval .

The size of the home range primarily dependedon the nearness of the feeding area (Fig . 3,Table II) . Females on the east side of the hotcreek had the largest home ranges and wentfarther to feeding areas. Females 629, F35,and 1278 had smaller home ranges and all livednear feeding areas .

However, agonistic behaviour and the densityof the population modified size and shape ofsome home ranges. Some animals travelled agreater distance to feeding areas when thepopulation was more dense as evidence by the

Number of adults and littersBurrownumber 1955 1956 1957 1961

1 IA, 1L

IA, IL IA, 1L IA, 1L

3 1A, 1L

2A IA, 1L

13 IA, 1L

IA ly

14 IA IA ]A, IL

16 1Y

ly ly

21 5A, 2L

5A, 21, 4A, 1 L 3A,2Y, I L

25 IA

1A IA

26 I A

29 ly

30 2A, I L I A

35 2A, 1L

1A IA, IL 1A

40 I A

51 ' IY 1Y

58 1A, IL IA, IL 1Y

59 ly

322

ANIMAL BEHAVIOUR, X, 3-4

largest mean size of homerange in 1955 (Table 11) . Blair(1953) reported that there wasprobably less ranging wherefood was dense but did notfind any evidence that popu-lation density influenced thesize of individual home ranges .Home ranges underwent

some seasonal changes . Someof the animals crossed from theeast to the west side of the hotcreek when the water leveldropped. During the first twoweeks of August when someanimals hibernated several ofthe remaining animals fre-quented areas that were avoidedearlier .

There were also year to yearchanges in home range thatwere correlated with changesin the individuals inhabitingnear-by burrow systems .

Behavioural PatternsDaily Activity Cycle

Although the daily activitycycle was different for eachindividual, a generalized popu-lation cycle, in its main outline,is applicable to any adultmember of the colony (Fig . 4) .The pattern is based on 26days of records in 1956, 19days in 1957, and 43 days in1961 .Emergence from the burrow

was roughly correlated with thetime when the first rays of thesun reached the colony area .However, this period of majoremergence extended over atleast an hour and some of theless active animals appearedeven later. On four cloudydays the animals appearedat about the same time as onsunny days . Because many ofthe burrows did not face thesun and were often underrocks or logs, it does not seemlikely that light was responsibledirectly for initiating activity .As many animals defecated

000

0

0

0

00

t'\0.,\

i!

I~ jit ~mi 01 m i.

li0

0 r

0 1

.I. . . jol

KEY

k0 •1 i'

0

BURROW SYSTEM~x

ROCK OR

PILEQ

ROOK P ILEI

h mI

LOG OR LOGPEILl2t

a - 635

\'~ Il:

0

TEST --'627i~ l

m12780634-..-

693/622

0

-•-• F 35

72'

Fig. 3 . The home ranges of eight adult females in July and early August of1957. Symbols for each animal are placed in areas frequently visited . The linesindicate the boundaries of the home ranges . Note that the home range ofX693 extends across the hot creek .

ARMITAGE: SOCIAL BEHAVIOUR OF A COLONY OF THE YELLOW-BELLIED MARMOT 323

Table II . The Population and Home Ranges of M. flaviventris on the Study Area . Home Range is for Adults only for theMonth of July. Home Range is Reported as the Maximum Distance in feet that an Animal Ordinarily Travelled from its

Home-burrow to its Feeding Area.

If the two 9's that wandered between B21 and an area 150 yards north-east of the colony are included . the range i* s72-1082 feet ; the mean west of the hot creek is 438 feet ; the mean of all home ranges is 40 feet .

immediately after emergence, onecould speculate that intestinal physi-ology served as the stimulus foremergence . About the first 15-30

AUGUST

minutes after emergence were spentgrooming and sunning. Gradually

i

the animals dispersed to the feedingareas ; by 8.00 a.m. this activity

v,

reached a peak that lasted for aboutm e

F1 SITTING

two hours. The remainder of the day® ACTIVE was spent in activities such as sun-

Z 7

FEEDING

-

ning and digging, or was spent inw 6 the burrow with occasional foraysa

-

JULY

-

for feeding. From about 10.00 a.m .w 5 until about 4.00 p.m. there was a

marked reduction in overall activity .4

- -

There was a second peak of activity3

concerned mainly with feeding from

2

%

last) hour ofuntil

theDuring the

_

turned to their burrows. UsuallyI

~

they sat near the burrow entrancfor a short time before entering the

6

8 9 10 n 12 1 2 3 4 5 6 7 8 burrow All the animals entered theirA

P MA M .TIME OF DAY

burrows by 15-30 minutes aftersunset. This general picture was

Fig . 4 . The pattern of daily activity of adults for July and August for highly modified by such things as1956, 1957, and 1961 . The activity data for 1955 were notrecorded in alarm calls, thunderstorms, inter -the three categories of sitting, active and fading . Totaljactivity for actions with other members of the1955 shows the same pattern as the above Active refers toan y activity ;

. e.g . running, playing, other than feeding and sitting .

colony, etc .

∎

1955 1956 1957 1961

PopulationAdult 17 12 11 8

Yearling

3 1 3 5

Young 45 18 25 12

Range of size of home range :

East of hot creek 108-720 144-468 108-468 180-432

West of hot creek 72-216 72-198 72-198 72-180*

Mean size of home range :

East of hot creek 461 336 368 306

West of hot creek 136 110 115

119

Mean of all home ranges 232 178 200

180I

324

ANIMAL BEHAVIOUR, X, 3-4

Range Mean

Centre of colony 0-2 . 2/min . 0 .8/min .

South flat feeding area 3 .5-6/min . 5 . 1/min .

The general pattern changed somewhat in

the colony (Table IV) . The behaviour of Y634August. The peak of activity was later in the

varied with the area in which she was feeding .morning and the late afternoon activity reached

When she fed at the southeast border, she sat upa peak earlier and the level was sustained over a

frequently ; she never was observed to sit uplonger period of time . This shift in activity may

when feeding near the centre of the colony .have resulted from the decreased day length of

Feeding in the last hour before sunset was pre-August and/or colder temperatures in the early

dominantly near the home-burrow .morning hours . The total mount of activity alsowas lower, primarily because many adults Table IV. Number of Times an Adult Marmot Sat Uphibernated .

While Feeding. Data from 10 Observations of Three orThere is some further indication that the

More Adults from Each Area .general pattern of activity was modified by air-temperature (Table III) . Although the data arescanty, there appears to be a general drop inactivity when the air-temperature exceeds 20°C .Feeding

Typically the animals lay down and crawledthrough the beds of clover while feeding. Thus

Locomotiona clover patch soon came to have a large numberof "crawlways" winding and criss-crossing

There were two basic patterns of locomotionthrough it. Usually an animal raised its head

which differed in the way the tail was carried .while chewing. Sometimes an animal sat up from

Adults going to the feeding grounds usuallytime to time. Animals feeding at the south flat

held the tail up in a half moon and waved itsat up more frequently than animals feeding on

from side to side . When reacting to an alarmthe west side of the hot creek near the centre of

signal, the tail was held down or pointed directlyTable III. The Per Cent. of the Adult Population Active at Various Temperatures for Selected Hours of the Day in July

of 1955 and 1956 .The number of observations is given in brackets prior to the range . Where no range is given, only one observation wasmade .

TemperatureCo.

Time

9:00 10 :00 11 :00

1 17:00

18:00

Range Av.

Range

Av. Range Av .

I Av .

Av.

14 (2) 88-100

94

18 (5) 55-100

77 100

I

100

19 (3) 68-100

80 88 100

100

20 (2) 55-66 60 77 45

21 67 (2) 11-55 27

22 (4) 37-67

50 62

23 (2) 37-55

46 (3) 22-62 35 33

67

24 (3) 25-56

40

25 18 (2) 12-31 21

26 68

27I

25

28 31

ARMITAGE: SOCIAL BEHAVIOUR OF A COLONY OF THE YELLOW-BELLIED MARMOT

posteriorly . Each of these patterns was associ-ated with dominance-submissiveness relation-ships. A dominant animal held the tail up,cocked in the "half-moon" position whereas asubmissive animal always held her tail low .Either position of the tail was associated with aslow walk or with a rapid, bounding run, agallop in which the hind feet are set before thefore-legs (Munch, 1958 ; Muller-Using, 1954) .

Those animals living in the centre of thecolony usually moved from the home-burrowto the feeding ground or to an auxiliary burrowen route to the feeding ground in one move .However, an animal living on the east side of thecolony usually moved much more cautiouslywhen going to the south flat feeding area . Forexample, X634 once spent 17 minutes movingfrom B1 to B50 and stopped and looked aroundthirty-eight times, sitting up seven of these .During the same time ?524 went from B13 tonear B33 in two minutes . Thus ?524 moved atthe rate of about 72 feet per minute while ?634was moving at the rate of about 25 feet perminute. However, Y524 did not always move asslowly as indicated by this example ; she oftenmoved at about the rate of 60 feet per minute,but always stopped more frequently . Whenmoving at this more rapid rate, the boundingmotion was used. After each stop, she pumpedher tail up and down once as she started for-ward again. This tail pumping was quite char-acteristic but has no known significance. It maybe considered an intention movement .Sunning

Early in the morning when the air was cool,the animals orientated their bodies broadsidetowards the sun. However, during the latemorning or afternoon, the marmots lay length-wise toward the sun. During the hot part of theday sunning was restricted and animals weremore likely to lie in the shade of a boulder orlog pile .Sitting Up

This position was nearly always used whenevera marmot was looking around in response to analarm, or while at a feeding ground, or whileen route to or from a feeding ground . Thisposition was not used when a marmot was ex-ploring its immediate surroundings, but onlywhen the marmot was observing a distant objector scanning the landscape . Citellus armatus(author's observation) and prairie dogs (King,1955) sit up in much the same circumstances asdo marmots .

325

GroomingGrooming was likely to occur at any time

during the day, but was most frequent shortlyafter the period of emergence during the morn-ing and prior to feeding. There were only fiveobservations of adults grooming other adultsduring the four years of study . In each case thegrooming was associated with agonistic be-haviour in which the dominant animal groomedthe submissive animal .Nest BuildingDigging and the collecting and carrying of

grass were infrequent. There were four newburrows (9, 59, 60, 61) added on the east side ofthe hot creek during the course of the study .Several animals participated in the digging .Digging was done with the forelegs, the dirtbeing thrown between the hind legs . Dirt thatpiled up at the entrance of a burrow was pushedaway by the animal lying down and shovingthe loose dirt with its chest and forelegs .Marmots lay down to gather grass . They tore

some loose with the mouth and positioned thegrass in the mouth with their forelegs, repeatingthese steps until the mouth was full . Usually onlyone trip was made to the burrow . Of 26 observa-tions of gathering of grass by adults during theperiod of study, 14 of these were in Augustabout the time of hibernation .

Social InteractionsPlayActivity designated as play was frequent

among the young, but was seen to occur be-tween adults only in 1955 . Several times adultsplayed with young. When the animals wereidentified, a female was playing with one ormore members of her own litter .Communication

Communication is used in the same sense asit is used by Scott (1958 : p. 189) ; i .e ., the main-tenance of contact between members of a group .Marmots employ sight and hearing for the mainpart, with some use of smell and touch .

The basic sound was a shrill whistle which is ahigh cry originating in the vocal cords (R . & D .Muller-Using, 1955 ; Munch, 1958) . This soundalerted other members of the colony who fre-quently responded by sitting up and lookingaround. The alarm call seemed to be the samebasic note, but given higher and sharper . Whenthis higher call was given, marmots reacted byrunning to a burrow .

3 26

ANIMAL BEHAVIOUR, X, 3-4

Frequently young marmots at play made asound resembling a shriek . This was occasion-ally given by a submissive adult when she wasapproached by a dominant male adult . A fourthcall, a growl, was heard only from animals intheir burrows .

The alarm call was elicited by moose, deer,horses, bears, fishermen, sandhill cranes, ravens,blackbirds, swallows, marsh hawks, a great blueheron and a coyote . Frequently animals causingthe alarm call were at a distance from the colony ;e.g ., across the river . Usually not all the marmotsreacted to an alarm call, but only those on theside of the colony from which the "intruder"was observed .

The reaction to birds was varied . Ospreysflying along the river near the colony did notelicit an alarm call. The reactions of marmots toother birds occurred as the result of a suddenmovement or sudden appearance of the bird inquestion. There was no indication that themarmots were reacting to a particular shapemoving in a particular direction as described forgallinaceous birds (Tinbergen, 1951 : pp. 31and 77) .

Marmots reacted to coyote calls, the bark of acow elk, and the clanging of a trap closing . Onlythe coyote call elicited a major response .No particular animal acted as a sentinel .

There always were some animals active and thealarm was given by whichever animal first re-acted to the stimulus . Sometimes the stimulusdid not produce an alarm call in some of theanimals .

The author considers the main function of thewhistle of the marmot to be an alert or alarmcall. However, German workers studying M .marmota have interpteted the call as an acous-tical territory marker (Bopp, 1955a, 1958 ;Munch, 1958) . However, Muller-Using (1955)emphasizes the warning nature of the call andpoints out that the call may function incident-ally as an acoustical territory marker . As will bediscussed later, the author does not think thatterritorial behaviour occurs in M. flaviventris .The whistle may be used in agonistic behaviour,but nearly always the animal being chased wasthe one that whistled .

Several times a particular individual wasobserved whistling long after any other animalin the colony was alerted. Such an individualalso appeared to be highly nervous . Perhaps thebasic stimulus for the call of the marmot is aninternal uneasiness . Any such sound producedas a result of the uneasiness would be heard by

other near-by animals . A prime cause of uneasi-ness would be an intruder such as a predator.Thus natural selection would establish alertand alarm calls from the vocal utterance of in-ternal distress . Uneasiness might also resultwhen another marmot entered the home rangeof an individual . The same call might be givenand might appear to be a territorial marker .Another means of communication seemed to

involve several stimuli, but primarily olfactory .This was designated the "greeting" . It was mostcommonly seen between young or between anadult and a young. The two animals approachedeach other head on, with tails arched and ap-peared to sniff at each other's cheeks (Fig . 5) .

Fig. 5 . Two young showing the "greeting position" . Noteespecially the manner in which the tails are held and thatthe animals appear to be sniffing each other's cheeks .Drawing sketched from a colour transparency.

Usually there was little if any contact . Therewas no mouth contact as described by King(1955) for prairie dogs . A similar pattern hasbeen described for M. marmota (Muller-Using,1955 ; Munch, 1958) .Sometimes only one animal was active in a

"greeting" . Frequently the active animal re-mained several feet away from the animal thatwas "greeted". Often the "greeter" appeared tosniff the "greeted" . Usually this ended whenthe "greeter" (a young) ran away from the"greeted" (an adult) . This pattern might also beone of dominance-submissiveness as it is similarto the "ready alert" described below .Agonistic Behaviour

Agonistic behaviour refers to conflict be-tween two individuals (Scott, 1956) . Its use ispreferred to the term "aggressive" behaviour asthe latter recently is used in a more restrictedsense as a "tendency to initiate a vigorous fight"(King & Gurney, 1954 : p. 326). Agonisticbehaviour is broader in its implications and en-ables one to encompass a large amount of inter-action between individuals in one category ofbehaviour. Aggressive behaviour, then, is oneaspect of agonistic behaviour.

ARMITAGE : SOCIAL BEHAVIOUR OF A COLONY OF THE YELLOW-BELLIED MARMOT 32 7

One position frequently used by adults inconflict was similar to the "greeting" except thebodies were held lower and more curved and thetails were held lower. This position was desig-nated the "ready alert" as it often was followedby one adult chasing the other. Frequently achase occurred only after the submissive animalbegan to run .There is some indication that mounting or

attempting to mount may be associated withdominance. The two adults that were observedattempting to mount other adults were clearlydominant to other adults in their areas .

Most relationships were complex . For ex-ample, in 1955, Y1287 with her litter and a largerfemale, dominant to Y1287, occupied B30, eachusing a different set of entrances . But ?1287was not completely submissive as she chasedother adults, except the male from B21 . Once,after being chased by the larger female of B30,?1287 retaliated by biting her from behind, thenfled to her burrow.

On the east side of the hot creek in 1955, Y524was clearly dominant to all other marmots, andthe yearling was submissive to all other adultseven at her own burrow . The remaining twofemales tended to be dominant at their homeburrows, with the exception that they were bothsubmissive to Y524 .

?524 occupied burrow systems 13 and 14 .In 1955, she frequented the area from B16 to B9 .But in 1956 she sometimes went north to B16,but usually she went via B13, along the hot creekand crossed over and fed near B32 and B33 .Female 634 now occupied B1 . She always wentsouth or east of B 1, never went north and nevercrossed the hot creek. There was no observedcontact between Y634 and Y524 . The latter wasaggressive to yearlings and young but did notcome into contact with other adults prior to lateAugust. At the start of the period of observationin 1956 there were two adults at B3 . Only one,?693, persisted throughout the summer . Thisanimal was clearly submissive to 9634. Duringpart of the summer ?693 frequented the areaaround BI and never abandoned going there .But Y693 avoided contact with ?634, even enter-ing the burrow at BI when Y634 approached .Thus dominance or submissiveness was notnecessarily associated with a topographic area,but was always associated with interactionsbetween animals .

In 1957, Y693 had essentially the same homerange as she had in 1956. Y634 moved to B14and spent much of her time feeding in the north

flat feeding area where she never went in 1956 .Y524 was not present. ?622 now ranged to thesouth. She was clearly dominant to ?693 andprobably was dominant to Y634. In general, ?634tended to avoid Y622 who had complete freedomof movement. The relationship between thesetwo animals illustrates the subtle nature ofaggressive-submissive actions postulated byColias (1944 : p . 83), "It is possible that aggress-iveness in a more subtle form may be expressedsimply by independence of action, and itsopposite, submissive behaviour, merely byavoidance of another animal" .

Some animals tended to be submissive where-ever they were in the colony . The reaction of2693 to young marmots typifies this behaviour .In 1957, she had a litter of young at B13 thatappeared above ground about three weeks laterthan the other young in the colony . By this timethe older young were wandering widely. When-ever one approached B13, it was vigorouslychased by Y693 . But when Y693 went into anarea where she was extending her home range,she reacted quite differently to young marmotsand elements of her basic submissiveness ap-peared :

10.29 a.m. 8th July, 1957. Young chased byT693 at B13. 10.35 a.m. ?693 went from B13to hot creek and crossed over. A young (fromB21) was at B32 and 2693 lay down on the bankwhile the young sat and watched her. She thenmoved past the young, at first holding her bodylow, tail down in slinking fashion . Then ransouth. The young ran to B32 (Fig . 6 a-e) .

Agonistic behaviour may lead to territorialbehaviour. Burt (1943 : p. 351) has specifiedterritory as "the protected part of the homerange" and Carpenter (1958 : p. 228) has em-phasized that territoriality is a "behaviouralsystem which is expressed in a spatial-temporalframe of reference" . The agonistic behaviourof marmots did not have a spatial reference norwas it territorial in the sense that it is in prairiedogs where all members of a coterie defend thearea that they inhabit against members of ad-joining coteries (King, 1955). Bopp (1954,1955b, 1956) writes of family and of colonyterritories in M. marmota but uses the wordterritory to mean home range . There was nodefence against other individuals in the colony .Munch (1958) reports one incident of agonisticbehaviour, but does not present any data forterritorial behaviour . Therefore, one must con-clude that agonistic behaviour with its attend-ant dominance-submissiveness relationships be-

328

tween individuals is characteristic of marmotsand that territorial behaviour, sensu strictu,does not occur.

Agonistic behaviour may be a characteristicfeature of sciurid behaviour. Gordon (1936)reported chases and churring in Sciurus fremontiand S. douglasii and he observed some sem-blance of chase order in Citellus lateralis andEutamias quadrivittatus . Layne (1954: p. 259)reported that the red squirrel, Tamiasciurus

ANIMAL BEHAVIOUR, X, 3-4

Fig. 6 . (a) Female 693, on her way to B32, stops as a young from B21 sits up and looks at her . (b) The young ap-proaches the adult, both animals have tail down . (c) The modified greeting occurs . The young is the "greeter" as it extendsits head toward the adult. The adult crouches lower in submissive position . (d) The young backs off and the adultassumes a more dominant position with her head up . (e) The adult moves past the young, but the tail of the adultremains low as she continues to display elements of submissive behaviour . The above sequence of activity took placein an area where 9693 was extending her home range. Near her home-burrow she would have reacted to the youngby actively chasing it . The drawings were sketched from a series of colour transparencies .

hudsonicus does " . . . not normally establishexclusive property rights over specific areaswithin the home range, and that territorialbehaviour is usually restricted to a small areaimmediately surrounding a feeding station orparticular den site" . . Many instances ofscolding, chasing, and lunging at anotheranimal were reported, indicating a general pat-tern of agonistic behaviour. Yerger (1953)observed chasing in the chipmunk in central

ARMITAGE : SOCIAL BEHAVIOUR OF A COLONY OF THE YELLOW-BELLIED MARMOT 329

New York. The peak of chasing occurred in theautumn and was associated with a high popu-lation. Much of the behaviour seemed to beagonistic, some of which was territorial. Theauthor has observed agonistic behaviour in C .armatus in which the patterns of activity ap-peared to resemble the territorial behaviour ofthe prairie dog (King, 1955) .

HibernationSome adults hibernated during the first two

weeks of August. Females that had littered andthe young did not hibernate prior to the lastweek of August when observations were dis-continued.

The most characteristic change in behaviourwas breakdown in patterns of home range .During August, 1956, females 634, 524, and 693all extended their home ranges across the hotcreek where they never had been previously .

A second striking change in behaviour was inthe level of activity. Alert calls and conflictswere less numerous (Table V) than earlier in theseason. It was often possible to approach withina few feet of animals that earlier would not allowthe observer to approach within several yards .

The third characteristic change in behaviourwas an increased wandering whereby some adultstraversed great areas of the colony and in somecases left the colony . Anthony (1923) postulatedthat seeking hibernation sites might be the causeof such wandering .

DiscussionThe patterns of agonistic behaviour indicate

that the adults which make up a colony of mar-mots must be aggressive enough to obtain aplace, but at the same time some adults must besubmissive enough to stay out of the way of orappease a more aggressive animal . Such a re-lationship implies that some animals may nothave the behavioural (and underlying physio-logical) characteristics that will enable them topersist in the colony. Perhaps, then, agonisticbehaviour has a role in the population dynamicsof the marmot .

Such a role is evident from the data on sexratios . As only one adult male is present in thecolony, this means that, in effect, half of theyoung are lost to the population because of theirsex. In addition, harassed adults and yearlingsoften emigrate.

Table V. The Number of Chases and Alert Calls per Observation Hour .Mean date of emergence of young, 19th June . Hibernation began in the 7th week in 1961 and in the 8th week in allother years. Yearlings are included with the adults .

Adult:Adult Adult:Young Alert Calls

Time 1955

1956 1957 1961 1955 1956 1957 1961 1957

1961

Pre-emergenceof young 5 7 0

•3 1

Post-emergenceof young :

1-2 weeks 1 .25 36 3 8 25 2 1 03 27 1 .25

3-4 weeks 30 10 2 7 30 03 12 37 75

5-6 weeks 95 •30 0 08 28 09 0 8 1 1 08 37

7th week 50 •26 1 1 07 25 0 22 2 1 1 1 50

8th week 0

0 0 10 50 0 5 28 2 0 28 2 5

9th week 1 1 i 0 0 0 1 1 0 0 0 0 20

Mean 38 17 24 28 12 1 1

Mean/Adult 022 012 018

Mean/Individual 0057 0030 0050(includes young and adults)

3 30

ANIMAL BEHAVIOUR, X, 3-4

Agonistic behaviour was directed particularlytoward yearlings. Although yearlings formed 21per cent . of the adult population over the foursummers, they were involved in 52 per cent . ofthe chases. For the four years, of a total of 12yearling present when observations began, fourremained at the close of observations . Three ofthese were in 1961 when the adult populationwas the lowest .

However, agonistic behaviour is not directlyrelated to population density . The rate of con-tact between adults was only slightly higher in1955 than in 1961 (Table V), although there weretwice as many adults in 1955 . Although therewere more adults in 1957, the rate of contactwas lower than in 1961 . In 1955 the contactswere scattered through the population, but in1961, ?636 was involved in 32 per cent . of all thecontacts . This female had a marked effect as anindividual as she caused two adults to leave thecolony and caused other adults to have largehome ranges in order to avoid her .

Agonistic behaviour may also act on theyoung. The number of adult :young chases/hourwas greatest in 1955, although the rate of chases/hour/total population was only slightly higherthan in 1961 (Table V) . But in 1955 the youngwandered over a greater distance within thecolony and half the young emigrated during thesummer (thus reducing the rate of contacts)whereas not more than one-third emigrated inany of the other years .

In summary, agonistic behaviour may func-tion to reduce the number of marmots in acolony when the population density is high,because the number of contacts between animalsis increased and emigration by members of allage groups results. Or similar results may occurbecause of the high level of agonistic behaviourof one or two individuals .

Evidence of predation on marmots is scarce .Wolves (Fryxell, 1926) and coyotes and variousraptors have been described as predators (Borrell& Ellis, 1934 ; Couch, 1930 ; Craighead, 1951),but weasels probably are not predators on mar-mots (Quick, 1951) . The author found marmothair in one of twenty-one sets of coyote scatscollected in the vicinity of the colony . It seemsreasonable that coyotes capture some of themarmots, especially young, that leave thecolony. Most of the animals taken in this man-ner might be marmots that would die anywayfor a lack of a proper place in which to live .Because of a lack of evidence that predation isan important factor in the population dynamics

of the yellow-bellied marmot, it is postulatedthat agonistic behaviour is the main factor thatlimits the numbers of marmots in a colony .However, data are not available on other aspectsof population dynamics, such as reproduction,mortality, and indirect physiological effects ofagonistic behaviour .

In conclusion, there seem to be two advantagesto the social organization of M. flaviventris .First, agonistic behaviour aids in regulationof the population on the colony site ; secondly,social facilitation through alarm calls, presenceof reproductive mates, etc., increases the prob-ability of survival of the resident population .These advantages of social organization seemto be contradictory in that the first limits thenumber of animals and the second prevents thereduction of animals and also acts to increasetheir numbers. These opposing functions can beinterpreted as a homeostatic mechanism formaintaining a near optimum number of animalson a colony site . The second function assuresthat there will be a sufficient number of animalsto maintain the species ; the first assures thatoverpopulation with its potential mal-effects willnot occur .

SummaryThe social behaviour of a colony of yellow-

bellied marmots was studied from June 18th toAugust 30th, 1955 ; July 12th to August 24th,1956; June 10th to August 23rd, 1957 ; andJune 14th to August 21st, 1961 . The colony islocated along a river terrace near the southentrance of Yellowstone National Park on theeast side of the Snake River slightly south of thejunction of the Lewis and Snake Rivers .The home-burrow is where the young are

reared, where an animal normally spends thenight, and where an animal goes when an alarmcall is given . Auxiliary burrows are used tempor-arily as a place of refuge . The burrows are con-nected by a trail system that directs the move-ments of the animals . The distribution of adultsand litters indicated a preference for burrowsin the central part of the colony .

Patterns of home range were of three types,those with no overlap, those with slight overlapand those with major overlap . Animals whichutilized the same home range frequently avoidedone another . The shape of the home range de-pended primarily on the nearness of the feedingarea. Agonistic behaviour modified both sizeand shape of home ranges . Some home rangeswere larger when the population was more dense

ARMITAGE : SOCIAL BEHAVIOUR OF A COLONY OF THE YELLOW-BELLIED MARMOT

or when agonistic behaviour was more frequent .Patterns of home range tended to change duringthe first two weeks of August when some animalshibernated . Changes in home range from year toyear were correlated with changes in the in-dividuals occupying home-burrows .

The animals emerged from their burrows atabout sunrise. There was a morning peak ofactivity followed by a midday low with asubsequent second peak of activity in the lateafternoon. All the animals entered their bur-rows by 30 minutes after sunset .Communication between marmots involved

the senses of sight, smell, touch, and, most im-portantly, hearing . The alarm call was given inresponse to any kind of intrusion into the life ofthe colony . No particular animal acted as a sen-tinel. Olfactory communication consisted of the"greeting" in which two animals sniffed eachother's cheeks .

One or two animals were dominant . Severaltended to be submissive to all with which theyhad contact. Females with young may be aggress-ive in the vicinity of their home-burrows to-ward other animals, but may be submissive inother parts of the colony. Other relationshipsseem best characterized by a kind of neutralityin which each of the two animals avoids theother. Dominance is characterized by inde-pendence of action, tail flagging and grooming .Submissiveness is characterized by avoidance ofother animals, slinking posture with tail downand submission to being groomed . Territorialbehaviour does not seem to occur in the yellow-bellied marmot .

Hibernation began in the second week ofAugust. Conflicts and alert calls were less num-erous and activity in general greatly decreased .Increased wandering was shown by some adults .

It is postulated that agonistic behaviour is oneof the factors that limits the number of marmotson a colony site .

REFERENCESAnthony, A. W. (1923) . Periodical emigrations of

mammals. J. Mammal., 4, 60-61 .Blair, W. F. (1953) . Population dynamics of rodents

and other small mammals. Ad. in Genet ., 5,1-41 .Bopp, P . (1954). Zur Topographie der Murmeltierter-

ritorien . Rev . suisse Zool., 61, 374-380.Bopp, P. (1955a) . Der Schrei der Murmeltiers als akus-

tische Territoriummarkierung . Saugetierkundl .Mitt ., 3, 28 .

Bopp, P. (1955b). Kolonialterritorien bei Murmeltieren .Rev. suisse Zool., 62, 295-299 .

331

Bopp. P. (1956) . Zur Topographie eines Kolonial-territoriums bei Murmeltieren. Rev . suisse Zool .,63, 255-261 .

Bopp, P. (1958). Fluchtdistanzen and territorialesVerhalten beim Murmeltier . Mitt. der Natur-forschenden Gesellschajt Schaf hausen, 26, 1-6.

Borell, A . E. & Ellis, R. (1934). Mammals of the RubyMountains region of northeastern Nevada. J.Mammal., 15, 12-44 .

Burt, W. H. (1943) . Territory and home range conceptsas applied to mammals . J. Mammal., 24, 346-352.

Carpenter, C. R . (1958). Territoriality : a review of con-cepts and problems . In Behavior and Evolution,Eds. Roe & Simpson. Yale U . Press, 557 pp .

Colias, N. (1944) . Aggressive behavior among vertebrateanimals . Physiol. Zool., 17, 83-123 .

Couch, L. K . (1930) . Notes on the pallid yellow-belliedmarmot. Murrelet, 11, 3-6.

Craighead, F. C. (1951). A biological and economicevaluation of coyote predation. Report preparedfor the New York Zoological Society and TheConservation Foundation, 23 pp .

Evans, F . C . (1951) . Notes on a population of the stripedground squirrel (Citellus tridecemlineatus) in anabandoned field in southeastern Michigan . J.Mammal., 32, 437-449 .

Fitzwater, W. D . Jr . (1943) . Color marking of mammalswith special reference to squirrels . J. Wild!. Mgt.,7, 190-192.

Fryxell, F . M. (1926). An observation on the huntingmethods of the timber wolf. J. Mammal., 7, 226-227 .

Gordon, K . (1936). Territorial behavior and socialdominance among Sciuridae . J. Mammal., 17,171-172 .

King, J . A . (1955) . Social behavior, social organization,and population dynamics in a black-tailed prairiedog town in the Black Hills of South Dakota.Contr. Lab . Vert. Biol. Univ. Mich., 67, 123 pp .

King, J . A. & Gurney, N . L . (1954) . Effect of earlysocial experiences on adult aggressive behaviorin C57BL/10 mice. J. comp. physiol . Psycho! ., 47,326-330 .

Layne, James N . (1954) . The biology of the red squirrelTamiasciurus hudsonicus (Bangs) in central NewYork . Ecol. Monogr ., 24,227-267 .

Muller-Using, D. (1954) . Beitrage zur Oekologie derMarmota m. marmota (L.) . Z. Saugetierk ., 19,166-177 .

Muller-Using, R. & D. (1955). Vom "Pfeifen" desMurmeltieres . Z. Jagdwissensch ., 1, 32-33 .

Munch, H. (1958) . Zur Oekologie and Psychologie vonMarmota m . marmota. Z . Saugetierk., 23,129-138.

Quick, H . F . (1951) . Notes on the ecology of weasels inGunnison County, Colo. J.Mammal., 32, 281-290 .

Scott, J . P . (1956) . The analysis of social organizationin animals. Ecology, 37, 213-221 .

Scott, J . P . (1958). Animal Behaviour . U . of ChicagoPress, 281 pp .

Tinbergen, N. (1951) . The study of instinct . Oxford U .Press . 228 pp .

Yerger, R. W. (1953) . Home range, territoriality, andpopulations of the chipmunk in central NewYork . J. Mammal., 34, 448-458 .

Accepted for publication 30th October, 1961