iVIAN FLIGHT FORMATIONS

By FRANK H. I-IEPPNER

Flight formations of birds are among the most widely observed,

yet least understood phenomena in avian biology. Two recent
ornithology texts (Welty, 1962; Pettingill, 1970) collectively de-
voted less than a page and a half to discussionof the functions of
flight formation and possible mechanismsof coordination, aero-
dynamics,and other related factors. Vine (1971), Beer (1958), and
Werth (1960) speculatedon the advantages of flocking as social
behavior,but the functionsof the varioustypes of flight formations
have not receivedmuch consideration.This paper presentsa work-
ing classificationof different types of avian flight formations, and
offers some suggestionson possiblebiological advantages of each
type.
DEFINITIONS

Beer (1958) defines a flock as "... two or more birds which asso-
ciate with each other due to innate gregarioustendencies."Emlen's
(1952) flock definition is "any aggregationof homogeneous individ-
uals regardlessof size or density." Emlen's definition presents
difficultiesfor the analysisof flight formations. For example,the
term "homogeneousindividuals" excludesgroups of mixed black-
birds, which are commonlyreferredto as flocks. I proposeto divide
Emlen's definition into two parts, and offer for considerationthat:
A FLIGHT AGGREGATION is a group of flying birds, lacking
coordination in turning, spacing, velocity, flight direction of in-
dividual birds and time of takeoff or landing, assembledin a given
area. An example would be a group of terns feeding on a schoolof
fish.
A FLIGHT FLOCK is a group of flying birds, coordinatedin one
or more of the following parameters of flight: turning, spacing,
velocity, and flight direction of individual birds, and time of takeoff
and landing. Flight flocksmay differin degreeof organization,from
the highly organizedtypes seenin dowitchersand other shorebirds
to loosely coordinated flocks of birds, such as American Robins.
Coordination may at times be a result of extrinsie factors, for
examplea groupof gullsflying into a prevailingwind.
These definitions of flock and aggregation agree with general
usage definitions to be found in Webster's 2nd International Die-
tionary.
LINE FORMATIONS

Line formations are flight flocks characterizedby one or more

groupsof birds flying in a line or queue. Seenin larger birds, such
as waterfowl and pelicans,line formationstypically show a rather
high degreeof regularity in spacingand alignment. The types of
line formations are as follows:
COLUMN (Fig. 1A). In this formation birds fly in singlefile along
the flight path, one behind the other. The column can be seen in
Brown Pelicansfollowinga shoreline, although pelicans,like many

160
Vol.
45,No.2 ArianFlight
Formations [161

F•6um•: 1. Line formations. A, column; B, front; C, echelon.

other birds, will shift from type to type of formation, dependingon

circumstances.
FRONT (Fig. lB). In this formation, birds are aligned per-
pendicularto the direction of flight in a plane parallel to the earth's
surface. I have seenthis formation only as a momentary alignment
in waterfowl as a transitional formation between other types.
ECHELONS (Fig. 1C). Birds in an echelonfly in singlefile, stag-
gered stepwisefrom the bird in the lead position in the formation.
Right and left echelonscan be seenin waterfowl, the larger flocking
waders, and pelicans, cormorants, and similar coastline-followers.
Frequently a left echelonbecomesa right echelon,and vice versa,
by the expedientof a temporary breakup of the formation, rather
than a swing from side to side.
J (Fig. 2A), and V (Fig. 2B). J and V formations are right and
left echelonsjoined at the tip of the formation. V formations have
approximately the same number of birds on each leg, whereas J
formations are noticeably unbalanced. These formations are most
commonly seen in waterfowl, although I have seen them in other
large birds such as cormorants.
INVERTED J (Fig. 2C) and INVERTED V (Fig. 2D). The J
and the V have a number of variations, usually seenlesscommonly
then the primary types. In the inverted formations, the apex, or
point of the formation is positionedat the rear. The inverted J and
V occursometimesin waterfowl, and usually representa transition
formation between one of the primary types. I have not seenthis
formation persistfor long periods.
CLOSED LINE (Fig. 3A). In this formation, which is a special
caseof line formations, there is roughly a circular movement of birds
following each other in line, differing from a simple aggregationin
that a definite rotational tendency occurs. A closedline flock may
rotate in either a clockwiseor counterclockwise direction,and some-
 Bird-Banding
162] F.H. Heppner Sp.
lng1974

Fm•5•m2. Line formations. A, J; B, V; C, inverted J; D, inverted V.

FmuaE 3. Line and compoundformations. A, closedline; B, branchedV.

times a secondclosedline formation occurswithin the first, giving

the visual effect of contra-rotating concentriccircles. Closed line
formationsmay be stackedabove each other, and may be seento
bestadvantagein gullscirclingover a dumpor similarattraction.
COMPOUND LINE FORMATIONS

BRANCHED Vs AND Js (Fig. 3B). The Branched V is an

exampleof a classof formations
(V of Js,J of Vs, J of Js)in which
there is a secondarybranchingfrom the main formationtype.
Vol.
45,No.2 Avian
Flight
Formatior;s [163
VERTICAL ALIGNMENT

Line flocks may vary along the vertical axis. Birds at the head of
the flock may be lower than, at the same level, or higher than birds
toward the rear (Fig. 4A, 4B, 4C).
CLUSTER FORMATIONS

This classof formationspresentsa different aspectthan the line or

c.o.mpound
formations.In the former,oneis impressed
by the prc-
c•smn with which relatively small numbers of large birds maintain
themselves in accurate spatial alignment and angular orientation
with their neighbors, whereas in the cluster formations attention is
drawn to the coordinationthat enableslarge numbers of small birds,
flying in close order, to wheel and turn •dthout suffering mid-air
collisions. Cluster formationstypically have a three dimensional
structure. The functional significanceof this classof formation may
be quite different from that of either line or compoundtypes.

]FIGUre:4. Vertical spacingin formations. A, ascending;B, level; C, descending.

GLOBULAR CLUSTER (Fig. 5A). I have borroweda term from

the astronomersto describethis formation, which is about as long
as it is wide, and in three dimensionalaspect,resemblesan irregular
spheroid. Birds that employ this formation will generally fly in
apparentcloseorder. It can mostreadily be seenin Starlings,black-
birds,sandpipers,and Brown-headedCowbirds.Whenflyingin this
formation,birds can be seenmaking very rapid turns.
FRONT CLUSTER (Fig. 5B). This formation is broaderthan it
is wide, with spacingand turning tendingto be very precise. The
front clusteris often seenin pigeons.
164] F. H. Heppner Bird-Banding
Spring 19741

FmUR•; 5. Cluster formations. A, globular; B, front' C, extended.

EXTENDED CLUSTER (Fig. 5C). An extended cluster is an

oblate spheroidalformation, with the long axis parallel to the direc-
tion of flight. Extended clusterstend to be rather disorganized,with
frequent breakoffs,and shifts of position. A typical examplemight
be American Robins heading toward a nightly roost. This forma-
tion, in fact, may simply be a flight aggregation,birds flying in-
dependently toward a common destination.
These categoriesdo not include all possibleflight formations, as
there are intergradesbetween most of the major types. Also, other
formations, especially cluster types, defy simple description and
do not have distinctive characteristics.

DISCUSSION

Although every ornithologist and bird observer has seen most, if

not all, of the formation types describedabove at sometime during
his career, amazingly little is known about their functional sig-
nificance, taxonomic distribution, ontogeny, or evolution. This
void is due in part to the lack of technologyneededto ask the im-
portant questions. The ornithologistuntil now could but helplessly
shrug his shoulderswhen askedby the layman, "Why do geesefly in
Vees?", or "Do blackbirdsreally all turn together, and if so, how do
they all get the turning messageat once?"
A few intriguing notes are found in the anecdotal literature.
Gerard (1943) observed that unidentified birds in a flock which
was pacing his car at 35 mph (56 kmph) apparently wheeledall at
once,or at least within 5 msecof each other, although it is not clear
how he made such a precisedetermination. Gerard did not observe
any evident leadershipin the flock. Nichols (1931) proposedthat
the bunchingand wheelingof dowitcher and plover flocks servedto
Vol.
45,No.2 ,,tvian
Flight
Formations [165
hold the flock together until faster and slower individuals had
adjusted speeds.
Most of the other available information pertinent to the question
of flight formations dealsnot with flight formations per se, but with
the adaptive significanceof flocking in general. Vine (1971), for
example,developeda model systemfor a group of terrestrial organ-
isms pursued by a predator using visual detection methods, and
proposedthat in that situation, the most probable flock shape is a
tight circle. Emlen (1952) discussedsomeof the socialforcesthat
draw birdstogetherin flocksand regulatespacingbetweenindividual
birds. Ainley (1972) described the flocking of Adelie Penguins
(Pygoscelus adeliae)in terms of their spacingunder various condi-
tions. Ainley's work may be important in analyzing flight cluster
formations,becausecoordination,spacing,and turning in penguins
may not be related to aerodynamic factors.
An aerodynamic advantage resulting from formation flying has
been discussedby several authors. Lissaman and Schollenbergcr
(1970) proposedthat in one type of line formation, the V, an aero-
dynamic advantage is gained by an individual bird by maintaining
a particular wingtip-to-wingtip distance, and by angular position-
ing relative to other birds in the formation to capture tip vortex
energyfrom the wings of neighboringbirds. Cone (1968) argued, on
grounds of aerodynamic theory, that the tip vortex arising from a
flapping wing shouldnot resemblethat flowingfrom the fixed wing
typical of aircraft. Gould (1972) reported that the mean distance
between Canada Geese (Branta canade•sis)flying in a Vee was 4.1 m
(SD = 0.88) and suggestedthat measurementsof the actual airflow
conditions around a bird's wingtips will be needed to determine if
the tip vortex travels in an appropriate direction and contains
enoughenergyto make flying in V formation advantageousfrom the
standpoint of capturing otherwiselost tip vortex energy. Tucker's
(1968) techniques of flying birds in wind tunnels might well be
adapted toward this end.
It is not necessaryto proposethat closeformationflight might be
aerodynamically advantageousonly by virtue of ease in capturing
tip vortex energy. Franzisket (1951) argued that tip vortex energy
was not significant, and that close formation flying provided an
area of turbulance-free air for flight, and closevisual communication
for flock members. Von Hoist (1952) agTeedwith Franzisket and
pointed out that (at that time) a phaserelationshipin wingbeatsof
birdsflying in linear formationshad not beendemonstrated.A phase
relationshipin wingstrokeshad been proposedby Geyr yon Schwep-
penburg (1952) as being necessaryfor an hypothetical aerodynamic
advantage. Nachtigall (1970) demonstrateda phaserelationship
between wingbeatsof neighboringbirds in flocks of geese,using
cinema analysistechniques. Gould (1972), however, found no in-
phase relationship in the •vingbeatsof Canada Geeseflying in V
formation, and suggestedthat each bird acted as an independent
oscillator. Gould also found that the angle between the "legs" of
the V ranged between 28ø and 44ø. These measurementsshould be
valuable in resolvingthe questionof an aerodynamicadvantage to
166] F. H. Heppne• Bird-Banding
Spring 1974

V formation flight, oncethe characteristicsof the airflow around the

birds are known. Eichner (1954) comparedthe turning movements
and formations of pigeonswith similar alignmentsof war planes in
combat formations.
Two major classesof questions are raised by formation flying.
The first group of questionsconcernsthe line i'ormations,especially
columns, echelons,and Vs. These formations are associatedwith
large birds. We do not seesparrowsor warblers flying in Vs, nor do
we seeblackbirdsor other small birds that routinely travel in large
groupsflying in columns. If the advantageto be gained by flying
in a line formation is aerodynamic,why is there not a strong cor-
relation between flying in linear formation and necessityfor energy
economyon long flights? The small land birds flying long distances
over water, apparently do not fly in linear formations, but cormor-
ants do. If, as Lissaman and Schollenberger (1970) argue, sub-
stantial energetic advantages are gained by flying in precise, line
formations, one would expect those flocking birds most severely
pressedfor energy to demonstrate a high degreeof linear formation
use. This doesnot seemto be the case. Clearly the questionof the
aerodynamicsof formation flying cannot be resolved,in the face of
conflictingor incompletehypotheses,until data are obtained on the
air flow conditions around birds flying in closeformation.
Alternate hypotheses for explanation of characteristicsof line
formations are possible. Hamilton (1967) proposedthat the V is
advantageousfor communication,in that the V allows a high degTee
of visual communication with neighboring birds, at the same time
leaving a clear field of view to the front. In addition, Hamilton pro-
posed that flying in a flock assistednavigation by averaging the
direction preferencesof individual birds, although Keeton's (1970)
experimentson orientation and homing in singlepigeonsand small
flocksdo not supportthis view. Another possibilityis basedon the
fact that the eyesof many birds that employ line flocks (and other
birds, as well) are relatively immobile and are positionedlaterally
on the head. In order for an image to fall on the fovea of the eye,
if the bill of a flying bird were pointed in the direction of flight, an
object would have to be ahead of, and to the left or right of the bird
m question. The V or echelon might then simply be the line con~
figuration that allowseach bird in the formation (with the exception
of the leader) the opportunity to keep the image of its neighborin
maximum resolution, while at the sametime permitting the head to
be pointed forward. The V or echelonwould then be a responseto
the necessityfor maintaining a high resolution visual image of
neighboringbirds.
J.P. Hailman (pers. comm.) has suggestedthat aerodynamic
factors related to the size of the organismmay have a bearing on the
apparentcorrelationbetweenlargebird sizeand flight in line forma-
tion. Differencesin flight speedhave significanteffectsin low speed
aerodynamicsdue to differencesin Reynolds' number, and it may be
that the speedof the birdsflying in line formationis the factor which
has been significant in the evolution of this formation type.
Vol.
45,No.2 Aviar•Flight
Formatiotas [167

A frequent observationin gooseflocksis that the leading bird will

occasionally,give up its position to another bird. The untcstcd
hypothesisoffered by hunters, and others, is that the lead bird is
cleaving a path through the air, the action presumably tiring, and
must exchangethis arduoustask with anotherbird in the formation.
Again, an alternate hypothesisis possible.The lead bird in a Vet or
echelonhas less opportunity to assessthe state of the formation
than any other bird, becauseall other birds are behind it. With
forward directed eyes, the lead bird would have to turn its head to
determine visually if the rest of the formation was still there.
Assumingthat it is desirableto insure that the flock stays together,
the act of turning the head repeatedly at an angle to a 40-50 knot
relative wind might well be tiring and call for relief. Conversely,
the stimulus fatigue involved in looking for sometime at the same
neighbor in the formation might impel a bird back in the formation
to changepositions,or even assumethe lead. Stimulus fatigue is
certainly not unknown in animals, both at a simple ncuronal level
and at the more complexlevels, such as sexual behavior (Boamet
ct al., 1969), and might well be a considerationhere.
Cluster formationspresenta different set of problems. The most
spectacularaspectof these formationsalso presentsthe most per-
plexing question. Do the birds actually turn at the sametime when
they wheel, and if so, how is the turning messagetransmitted simul-
taneouslyto all birds? A corollary questionconcernsthe initiation
of a turn. Individuals and small groupsare constantly breaking off
from the main formation, effectively eliminating the possibility that
the turn is initiated by the first bird to changedirection randomly.
The stimulus for the initiation of a turn might be a sound signal
given by an individual leader; this seemsunlikely, given the din of
massedblackbird and Starling flocks. A "follow-the-leader" princi-
ple might be used,the entire flockvisuallymiraicingthe movements
of a leader, but this would present problemsin flocks of tens of
thousandsor more. Finally, other means of communication hitherto
undiscovered,perhaps involving electromagneticfields, might be
employed (much as ultrasonic communicationin bats and marine
mammals was unknown c.nly a generationago).
It may be significantthat the greater the density of the flock, the
greateris the apparent precisionof turning and movementsHepp-
ner, pers.obs.). If the birds are flying looselytogether,as in extend-
ed or front clusters,there is little danger of collision,and a follow-
the-leadertype of directionalguidancewouldbe sufficientto explain
the movementsof the formation. As the formation becomestighter,
the dangerof collisionincreases,and coordinationof movementbe-
comesmore important. Coordinationwould be important in tight
formations not only in turning and wheeling, but also in wing flap-
ping. It would be interestingto test the reactiontimes of birds that
display a high degreeof coordinationin formation, such as plovers
and sandpipers,versus birds such as warblers, which are rarely, if
ever highly coordinatedin formation.
Study of flight formations in birds reveals many unanswered
questions that might be addressed with the methods of modern
168] F. H. Heppner Bird-Banding
Spring 1974

technology, such as telemetry and radar. There is an additional

bonus in that the flight formations of birds are among the most
aesthetically pleasingaspectsof avian life.
SUMMARY

Flight formations of birds are classifiedas flight aggregations,

line flocks,compoundflocks,and clusterflocks. These major cate-
gories are then subdivided and described. Possible functional ad-
vantages for each major formation type are discussed. Alternate
hypothesesof aerodynamicadvantage or facilitation of visual com-
munication are examined for line formations representedby the V
formation of •vaterfowl. Problems of coordination in cluster flocks
are discussed.

ACKNOWLEDGMENTS

I would like to thank my associates,particularly Bruce Hunter,

Lisa Gould, Harold Pomeroy, and David Preble, for their helpful
suggestions.Jack P. Hailman offered many useful commentson the
first draft. This study was supported in part by a University of
Rhode Island Faculty Grant-in-Aid, and is Publication No. 5 of the
Arian Research Institute.

LITERATURE CITED

A•NL•¾, D. G. 1972. Flocking in Adelie Penguins. Ibis, 114: 388-390.

B•R, J. R. 1958. The compositionof bird flocks. Flicker, 30: 78-83.
B•M•, W., B•M•NT, G., ANDM. T. CL•66. 1969. Copulatory behaviorof the
ram, Oyesari½•. I[: factors affecting copulatory satiation. A•i•. •½hav.,
17: 706-711.
CON•:, C. C. 1968. The aerodynamicsof flapping birdflight. Special Scientific
Report No. 52. Virginia Inst. Marine Sci.
Erc•F•, D. 1954. Formationsfiugregelnauch beim Flug der Haustabe.
war•e, 17: 15-18.
EML•N, J. T., J•. 1952. Flocking behavior in birds. A•k., 69: 160-170.
F•ANZ•SK•T,L. 1951. i)ber die UrsachendesFormationsfiuges.
¾o•ehvarte,
16: 48-55.
G•nAgD, G. W. 1943. Synchronyin flock wheeling. Science,97: 160-161.
G•¾g v. Scnw•rrr:NBUa6, H. 1952. Vorteile der Zuggesefligkeit. Vogelwarte,
16: 116-119.
GOULD,L. L. 1972. Formation flight in the CanadaGoose,(Branta c. canadensis).
M.S. Thesis,Univ. Rhode Island.
H•MrLTON,W. J., III. 1967. Socialaspectsof bird orientationmechanisms.
Animal orientation and navigation, (R. M. Storm, ed.) Oregon State Uni-
versity Press, Corvallis.
HOLST,]•. V. 1952. Diskussionsbemerkung. J. O•'nithol., 93: 191.
K•ToN, W. T. 1970. Comparative orientational and homing performance of
singlepigeonsand small flocks. Auk, 87: 797-799.
LrSS.•MAN, P. B. S., ANDC. A. SC•OLLF•NBr:•6F•.1970. Formation flight of birds.
Science,168: 1003-1005.
NAC•r6ALL, W. 1970. Phascnbczichungcn der Flfigclschl'•ge von Gansen
w/ihrend desVerbandfiugsin Keilformation. Z. vergl.Physiologie,67: 414-422.
NrC•OLS,J.J. 1931. Notes on the flockingof shorebirds. Auk, 48: 181-185.
Pr;TT•N6•LL,O. S., Jn. 1970. Ornithologyin laboratory and field. Minneapolis,
Burgess Publishing Co.
Vol.
45,No.2 AvianFlight
Formations [169
TUCKER,V. A. 1968. Respiratory exchangeand evaporative water lossin the
flying Budgerigar.J. Exp. Biol., 48: 67-87.
VINE, I. 1971. Risk of visualdetectionandpursuitby a predatorandthe selective
advantageof flockingbehavior. J. Theor. Biol., 30: 405-422.
WELTY,J. C. 1962. The life of birds. Philadelphia,Saunders.
WrmT•, I. 1960. The problem of flockingin birds. Proc. XII Intern. Ornithol.
Congr, 6: 744-748.

Departmentof Zoology,Universityof RhodeIsland, Kingston,R. I.

02881.

Received14 December1972,accepted6 December1973.