SHORT
COMMUNICATIONS
Wing Flutteringby Mud-GatheringCliff Swallows:Avoidanceof
"Rape" Attempts?
ROBERT W. BUTLER
CanadianWildlifeService,Box340, Delta, BritishColumbiaV4K 3Y3, Canada
Beecherand Beecher(1979) suggestedthat colonially nestingmale Bank Swallows(Ripariariparia)
escortedtheir mates7 to 8 days following pair formation to guardagainst"promiscuouscopulations."
They observedthat unguardedfemaleswere frequentlychasedby males,and, althoughchasesended in copulationon only threeoccasions,
the authors
interpretedfemale-escorting
behaviorby their mates
sized flockson the wing-fluttering behavior of individual swallows, I randomly chosea bird as it arrived at the mud sourceand recordedwhether it fully
extended,partiallyextended,or foldedthe wings over
the back when it began to gathermud. I then counted or estimated the size of the flock. In addition, I
cut a small hole in the side of 27 nests and examined
the contents to determine
clutch-initiation
dates. I
found that before egg laying (28 April-4 May) and
The Cliff Swallow(Petrochelidon
pyrrhonota)alsois during egglaying (5-28 May) swallowsfully extenda colonial nester, but I saw no indication of nor can ed or partially extendedtheir flutteringwings sigI find any referenceto mate guarding. Here, I present nificantly more often in flocksof greater than two
data that suggestthat Cliff Swallows flutter their birds than in smallergroups(X2 = 9.2, P • 0.001).
wings above their backswhile gatheringmud for During incubationand the nestlingstage(29May-17
their nestsprimarily to reducerapeattemptsand sec- June), however, there was no significantdifference
ondarily to reducethe theft of mud pellets used in in wing flutteringbehaviorwith increasingflocksize
nest construction.
(1-2 birds, 3-5 birds, >5 birds). Furthermore, swalMy studywas conductedApril-July 1980-1981at lows flutteredrather than folded their wings signifthe CrestonValley Wildlife InterpretationCentre, 10 icantlymoreoftenduring egglayingthan eitherbekm west of Creston, British Columbia (49ø10'N, fore or after egglaying (X2 = 15.76,P • 0.001)(Fig.
116ø36'W).The Cliff Swallowcolonyhad been estab- 1). Clearly, wing fluttering increasedin frequency
lished in the past 6 yr. A maximumof about 200 with flock size and the onsetof egglaying.
completed nests was counted in 1980 and 1981.
Cliff Swallowsthat folded their wings appearedto
Another Cliff Swallow colony numbering over 500 be attackedby other Cliff Swallowsmore often than
nests was locatedabout 1 km away, although those thosethat flutteredtheir wings, althoughthe attacks
birds were not observedgatheringnestmaterialwith happenedtoo quickly to be quantified.To test the
the swallowsin my study. The colonyin my study hypothesisthat wing flutteringreducedthe frequenbuilt their nests on floor beams supported about 2 cy of attacks,I recordedthe number of attacksdim abovewater by verticalpilings.During my study rected at a stuffed, female Cliff Swallow when her
as "mate guarding."
about 40 Barn Swallows (Hirundo rustica)also nested
wingswerewired motionlessin the foldedand flut-
in the Cliff Swallowcolony.The colonyis locatedon teringpositions.On 19 May 1980I placedthe model
a 616-ha marshthat is dominatedby Equisetumsp., with its wings in the folded positionat a mud-gathCarexsp., and Myriophyllumsp. The Cliff Swallows ering site where swallowsgatheredin flocksof 10gatheredmud from the exposedbanksof a dykelo- 15 individuals. The model was vigorouslyattacked
cated within
18 times in 300 s of observation.
50 m of their nests.
Attacks resembled
In 1981 Cliff Swallowswere first seen at the study
area on 13 April. Nest building began by about 25
the copulation attempts describedfor many other
birds, includingswallows,althoughsemenwas nev-
April, and clutcheswere initiated 5-28 May. Cliff
er depositedon the model.Typicallythe attacking
Swallowsgathermud by alighting on the ground in bird alighted on the model's back and graspedthe
flocksof up to 20 or more birds, raisingthe tail about nape with its mandibles.The attackerthen moved
15øabovethe horizontaland flutteringthe extended its tail in a lateral motion, as if searchingfor the
wingsabovethebackwhilepeckingmudintoa pel- model'scloaca,while it vigorouslyflutteredits wings
let. That pellet is then flown back to the nest, and above its back. Those attacks lasted about 5-15 s,
thepartnersexchange
places(Emlen1954,pers.obs.). althoughin four casesthe attackerremainedmountBrown (1910)said that wing fluttering prevented ed for 45 s until it drooped its wings on either side
Cliff Swallowsfrom stickingin the mud and soiling of the model and the tips touchedthe ground. Within a few seconds a second bird attacked the first at-
their feathers. I noticed, however, that individual
Cliff Swallowsflutteredtheir wings more vigorously tackerand both birds departedaftera brief skirmish.
when other Cliff Swallows were present than when Someswallowsignoredthe modelat the mud source
they were alone.To quantify the effectof various- and gatheredmud aroundand underit. Othersig758
October1982]
ShortCommunications
759
lOO
Z
LU
½r
•-
40
2o
0
Fig. 1. Percentageof Cliff Swallowsin various-sizedflocksthat folded, partially fluttered, and fully
flutteredtheir wingsbefore(I, n = 78), during(II, n = 103),and afteregglaying(III, n = 216).
nored attackerson the model and continued gathering mud. In attacksinvolving live swallows the
victim always struggledvigorouslyand usually escaped within a few seconds.
I repeatedthe above experiment on the same day
using the model, but this time I extendedthe wired
the mud source.Resultsdid not differ significantly
between the model tests and actual attacks, indicat-
ing that habituation to the model had not occurred
(Fig. 2). Thoseresultsalsosuggestthat the behavior
of live swallowstowards the model accuratelyrepresents the behavior
between live swallows. There
alsowere significantlymore attacksduring egg layvation. To test that habituation
toward the model
ing than preceding or following that stage in both
had not occurred,! repeatedthe first experimentfor tests(one-sidedProportionTest;Z pre-egg/egg
= 3.16,
210 s, which resultedin sevenattacks.Admittedly, Z egg/post-egg
= 4.47, P < 0.001).
the length of time that the model was not attacked
In a few instances,an attackingswallow stolemud
would havebeena bettermeasureof attackfrequen- from the attacked bird, while in others the attacks
cy than the number of attacks, but the results still appeared to be of a sexualnature. To test whether
indicatethat wing flutteringreducesattacksby other Cliff Swallowswere stealingmud from one another,
Cliff Swallows.
I placed a pellet between the folded-wing model's
In one instancewhen no modelwaspresent,I wit- mandibles and placed the model on the mud source
nesseda Cliff Swallow attackthree mud-gathering for about300 s during the late egg-layingstage.The
Cliff Swallowsin about15 s. It appearedthat the sole model was mounted several times, but no attempt
intention of the attackerwas to pounceon the three was made to stealthe mud. That suggeststhat the
swallows,becauseit never gatheredmud, although motive for the attackswas not to steal mud pellets,
there was plenty of room to do so. Eachtime, the althoughmud stealingmight be moreprevalentdurvictim departedfollowing a brief skirmish.I alsosaw ing the early nest-buildingstage.
a Cliff Swallow attack a live Tree Swallow and a dead
The reasonfor wing fluttering by mud-gathering
Cliff Swallows remains uncertain. ! have shown that
Barn Swallow at the mud sourcein May 1980.
In 1981,I usedthe folded-wing modelto examine wing flutteringis mostfrequentduringthe egg-laythe seasonalfrequencyof attacksfrom 28 April-16 ing stage.In addition, the attackson the model reJune and comparedthose resultswith the number of sembledcopulationsand were mostfrequentduring
attackson live swallowsfrom 2 May-15 June, all at the egg-layingstage.I alsowitnessedCliff Swallows
wings. There were no attacksduring 355 s of obser-
760
ShortCommunications
[Auk,Vol.99
MODEL
1NATUR
o
I
]I
•I
I
]I
Fig.2. Thenumberof attacks
perminutedirected
by CliffSwallows
towarda stuffed,folded-wing
model
andtowardlive Cliff Swallowsbefore(I), during(II), and afteregglaying(III).
stealing mud from one another at the mud source,
nerableto extra-paircopulations
whileguardingthe
however. To confuse the issue further, Cliff Swal-
nest. I made few observationsof pairs at the nest.
Emlen (1954), however, mentionedforced-copula-
lows begin egg laying before the nest is complete,
so that a mud-gatheringswallowcouldbe building
a nest, laying eggs, or both.
One possibleexplanationfor the evolutionof wingfluttering behavior of mud-gatheringCliff Swallows
is as follows.Cliff Swallowsinvestlargeamountsof
time and energyin building the nest. Withers (1977)
showed that the daily energy expenditureof Cliff
Swallows was greater during nest constructionthan
during incubation or nestling periods. At the end of
the 1981 nestingseason,only 36% (n = 139) of the
Cliff Swallownestsin this study couldsupporteggs,
interpretationis that the energeticcostsof building
the nest require sharing the duties between both
membersof the pair.
Cliff Swallowspresumablyare very susceptibleto
rape attemptswhile gatheringmud in a tight flock.
Hooglandand Sherman(1976)foundthat extra-pair
copulations
in BankSwallowswereattemptedon the
ground.Furthermore,the Cliff Swallowtail-up posture may solicitrape attempts,becauseit resembles
the female'scopulatorypostureat the nest. There is
so most swallows
strong competition within the flock for sites at which
must build
nests at the start of a
tion attempts toward females in their nests. Another
new season.That implies that theremight be strong
competition for old nests and nest material. I found
that nestsfrom previousyearswere occupiedabout
10 days before the swallowsbegan constructionof
to gather mud, and alighting Cliff Swallowscould
presumablysettlemore easilyon the backof a swallow with foldedwings than on one with fluttering
wings. Barlowet al. (1963)found that sunningCliff
new nests. In addition, Cliff Swallows steal mud from
Swallowsraised their wings to prevent othersfrom
one another'snests(Emlen 1954)and from one another settlingon their backs.Any swallow that settledon
(this study). The lossof a nestby a pair of Cliff Swal- the back of anothercouldstealits mud or usurpits
lows would mean a lossof a largeinvestmentof time spot at the mud sourceregardlessof sex. The same
and energy and would favor the evolution of nest would hold true whethera femalealightedon a male
guarding. Cliff Swallowsappearto guardtheir nests or a female.If a maleswallowalightedon a female,
(Emlen 1954,Mayhew 1958).Emlen (1954)has shown he couldcopulatewith her, stealher mud pellet, or
that one member of a mated pair remains in the nest usurpher spotat the mud source.Any swallowthat
while the other is at the mud source.After collecting flutteredits wings, however,would reducethe chance
a pellet of mud, the swallow returns to the nest and of being settledupon. All of this suggests
that natexchangesplaceswith its mate. Nest guarding pre- ural selectionhas favored nest guarding over mate
cludesmate guarding and thus freesthe male to seek
opportunities for extra-pair copulations with unguarded females.That implies that femalesare vul-
guarding in Cliff Swallows. Males could be cuckolded, but they alsohave opportunitiesto cuckoldother
males.I never witnesseda rape that appearedto be
October 1982]
ShortCommunications
761
LITERATURE CITED
successful, however, and I believe that successful
rapesendingin spermtransferarerare. I neverfound
semen deposition on the Cliff Swallow model, although similar studiesby other investigatorshave
found semennear the cloacalregionof modelBank
Swallows (Hoogland and Sherman 1976) and Savannah Sparrows (Passerculus
sandwichensis)
(Weather-
BARLOW,J. C., E. E. KLAAS, & J. L. LENZ. 1963.
head and Robertson 1980).
BROWN, F. A.
Femalesmight increaseor decrease
their fitnessby
participatingin extra-paircopulationswith malesof
unknown fitness. Presumably,it is to the female's
advantageto copulateconservativelywith males of
known fitness (i.e. their mates) and not to risk de-
creasedfitnessby matingwith othermales.
J.P. Goossenassistedin the field, J.P. Savard and
G. E. J. Smith provided statisticaladvice, R. J. Can-
nings from the VertebrateMuseumat the University
of British Columbia supplied some Cliff Swallow
models, and M.D. Beecher, D. R. Flook, K. Vermeer,
and P. Weatherheadcommentedon the manuscript.
I thank all of you.
Sunning of Bank Swallowsand Cliff Swallows.
Condor
65: 438-448.
BEECHER, M.D.,
& I. M. BEECHER. 1979.
Socio-
biologyof BankSwallows:reproductivestrategy
of the male. Science 205: 1282-1284.
1910. Cliff Swallows. Bird-Lore 12:
137-138.
EMLEN,J.T., JR. 1954. Territory,nestbuilding, and
pair formationin the Cliff Swallow.Auk 71: 1635.
MOOGLAND,J. L., & P. W. SHERMAN. 1976. Advan-
tagesand disadvantages
of Bank Swallow (Ripariariparia)coloniality.Ecol.Monogr.46: 33-58.
MAYHEW,W. W. 1958. The biology of the Cliff
Swallow in California.
Condor 60: 7-37.
WEATHERHEAD,P. J., & R. J. ROBERTSON. 1980.
Sexualrecognitionand anticuckoldrybehavior
in savannasparrows.Can. J. Zool. 58: 991-996.
WITHERS,P. 1977. Energeticaspectsof reproduction by Cliff Swallows.Auk 94: 718-725.
Received
19November1981,accepted
11February1982.
Effect of Intrusion Pressureon Territory Size in Black-chinnedHummingbirds
(Archilochus alexandri)
MARY E. NORTON,1'2PETERARCESE,
1 AND PAUL W. EWALDa
lWashington
StateMuseum,Universityof Washington,
Seattle,Washington
98195USA, and
2'aMuseum
of ZoologyandMichiganSocietyof Fellows,Universityof Michigan,
Ann Arbor,Michigan48109USA
Within many avian species,territory size is inverselycorrelatedwith foodabundance(Pitelkaet al.
1955,Gasset al., 1976,Myers et al. 1979).Two hypotheseshave been proposedto explain these correlations:(1) animalsmayassess
resourceavailability
directly and defend areas that contain sufficient
amountsof food, or (2) animals may adjust territory
size in responseto an intermediatevariable, intru-
mainedconstantor increased,territorysizedoubled.
This increasein territory size was attributed to the
decreasedintrusion pressure.Similarly, Ewald et al.
(1980) documented an inverse correlation between
intrusion pressureand territory size in a colony of
Western Gulls (Larusoccidentalis).Becausegulls do
not forage on their territories,variationsin food
abundance
could not have caused the variation
in
sion pressure.By the secondhypothesis,areasof territory size in this colony.
Thesetwo hypotheses,however,need not be mugreaterfood abundanceare morecostlyto defend
because
theyattractmorecompetitors,
theresultbeing tuallyexclusive.Hixon (1980)developeda modelof
smaller territories.
optimal territory size in which he dealt with both
food availabilityand intruder density.Integralto his
model is the assumptionthat thesetwo factorscan
vary independentlyof eachother,but he pointsout
teractionof prey density and intrusion pressurewas that disproportionatelyhigh food production in a
controlledstatistically,the effectof food densityon given territoryin comparisonwith surroundingterterritorysizewasno longersignificant,yet the cor- ritories might causea concurrentincreasein intrurelationbetween intrusionpressureand territory size sion pressure.Thus, theoretically,thesetwo factors
washighlysignificant.Furthermore,
whenintrusion canacttogetherand havethe sameeffecton territory
pressuredeclinedseasonallywhile prey densitiesre- size.
Ewald et al. (1980), Myers et al. (1979) and Hixon
(1980)all suggestthat further studiesare needed on
2 Presentaddress:2022 Franklin Avenue East, Seattle,Washington
98102 USA.
speciesin which intrusionpressureand food abun-
Myerset al. (1979)testedthesehypotheses
in studies of winteringSanderlings(Calidrisalba).Their resuitssupportedthe secondhypothesis:when the in-