Current Science,January20, 1985, Vol. 54, No.2
57
COLONY FISSION IN A SOCIAL WASP
RAGHAVENDRA GADAGKAR and N.V. JOSHI*
I
Centrefor Ecological Sciencesand * Centrefor Theoretical Studies,
:
Indian Institute of Science,Bangalore560012, India.
..ABSTRACT
A colony of Ropalidia cyathiformis with elevenadult femalesexperienceda period of
high levelsof aggressionand continuouslydecreasingbrood culminating in about half the
adults leaving the parent nestto found a new colony nearby.The group that stayedback
at the parent nest which we call the Loyalists and the group that left which we call the
Rebelswere rather evenlymatched in numbersand agedistribution. The Loyalists were
relatively more dominant compared to the Rebels and the latter appeared to have
behavedin a highly coordinated mannerevenbeforethe actual colony fission. The fission
resulted in an immediate increasein brood production in both the parent and the new
nests thereby increasing the genetic fitness of both groups. We argue first that this
phenomenonis a reflection of reproductive competition and second that it provides
strong circumstantial evidencefor kin recognition and communication.
INTRODUCTION
P RIMITIVELYeusocial insects are of special
interest from the point of view of social
evolution because,while they exhibit sterility and
the extremedegreesof altruism characteristic of
the highly eusocialantsand honeybees,theyalso
showa remarkabledegreeof flexibility. The role
of an individual in a colony is not fixed but
appearsto be determined by the socialenvironment it finds itself in. Groups of females(foundresses)come together, relegate reproduction to
the most dominant individual and the rest
remain subordinate and sterile, working to feed
and care for the offspring of the dominant
female1-3. If the dominant female dies or is
removed, however, one of the subordinate femalesceasesto work, and beginsto lay eggsand
therebybecomesthe queen4-7.
Sinceat leastsome
of the subordinate femalesare potentially capable of becomingqueenssucha situation might
be expectedto foster a great deal of reproductive
competition among nest mates.
Ropalidia,a genusrepresentingsucha primitive level of social evolution is abundantly distributed in India 8-12.Here there appears to be
scope for even greater levels of reproductive
competition. On the one hand colonies are
perennialand long lived and musttherefore often
outlive their queensproviding opportunities for
queen replacements. On the other hand new
coloniesare initiated throughout the year12.As a
result, femalesemerging at anytime must have a
finite chanceof becoming queens.
During observations of two species,R. marginata and R. cyathiformis we noticed several
behaviour patterns suggestive of reproductive
competition. On one occasion, an individual
becamevery aggressiveand challengedthe queen
by biting, nibbling and chasingher. In this case
both the queenand her challengerwere missing
the following day and the next most dominant
female beganto lay eggs4.We imagine however
that suchchallengersmust occasionallysucceed
in themselvesbecomingreplacementqueens.On
severaloccasionswe have seensingle femalesgo
off on their own to start newnests.Most of such
nestsinitiated by a lone founderesshowever fail
after a short while. It appears therefore that
emergingfemaleshavetwo reproductive options
before them, (i) to leave the parent nest and
initiate newnestson their own so that immediate
egglaying is possible although the risk of failure
is probably high, (ii) to stay on at the parent nest
and challengethe queen;here the probability of
succeedingin becomingthe next queenis prob-
58
Current Science,January
20, 1985, Vol. 54, No.2-..~
ably small but there is little chanceof immediate
nestfailure; a certain amount of inclusivefitness
on account of caring for siblings and other close
relatives must therefore often be guaranteed.
Jeanne13has classified the sub-family polistinae into 2 behaviouralcategoriesnamelythose
where new colonies are founded by one or more
queenswithout the aid of a swarmof workersand
those where new colonies are founded by a
swarm of workers accompanying one or more
queens.R. cyathiformis is believedto belong to
the first group13. During longterm observations
of a colony of R. cyathiformis however,we once
sawa female behave in a manner most remark-of
able for a non-swarming species.She left her
parental nest to initiate a nest but did so along
with about half the adults on the parent nestwho
continued to remain subordinate and work for
the new nest.The new nest subsequentlybecame
as large as the parent nestand produced several
progeny. In this paper we analysethe behaviour
of thesewaspsbefore and after the colony fission
and interpret their behaviour in the context of
reproductive competition, in an attempt to
understand the causesand consequencesof the
colony fissionas well as its probable mechanism.
MATERIALS AND METHODS
b = probability that animal i is on the
nest and animal j is away from the
nest
c = probability that animal i is away
from the nestwhile animalj is on the
nest
d = probability that both animals i andj
are away from the nest
was chosenbecauseit givesa value ranging from
-I to + I irrespectiveof the actualtime budgets.
A value of + I therefore means that the two
animals synchronised their behaviour to the
maximum extent possible within the constraints
their time-activity budgetswhile -I indicates
the maximum possible avoidance. A value of
zero is equivalent to i andj choosingto stay on
the nest or be away from it independentof each
other.
The frequencies of the dominant and subordinate behaviours were computed from the
'all occurrencesof rare behaviours' data. A nest
map showing the location of brood in different
stagesof developmentwas maintained on every
other day. Observationswere made between 19
February and 21 June 1980for a total period of
225 hr during which 282 instantaneousscanning
sessionsin addition to ad libitum observations
were performed between5 a.m and 5 p.m. The
data were analysed on the DEC 1090computer
The study was conducted on a colony of
R. cyathiformis that had built a neston a metallic
pole in the grounds of the Indian Institute of
Science,Bangalore. All adults weremarked with
spots of coloured paint for individual recognition. Behavioural data were recorded using,
the methods of instantaneousscanningand' All
occurrencesof rare behaviours'14.(For further
details seeGadagkar and Joshi15,16).
The data fr?m th~ scanswer~usedto ass~ssthe
e~tent to WhIChd~ffe~entarnm.alscoordmated
with eachother theIr tl,mesOf~l~g on and av:-ay
from the ?est. :ule ~ ~SSOClatloncoefficl~~t
betweenarnmalsI and) gIven by the formula.
Y j"= ~
!/ ad+bc
facility at the Indian Institute of Science.
History of the colony and social organisation
beforefission
A colony of R. cyathiformis that had built a
nest of 3 combs on a metallic pole at the Indian
Institute of Sciencehas beenthe focus of intensive study. Having shown that the three combs
really belongedto a single colony18,we made a
detailed study of the social organization of this
species16,
19.Although the latter study included 8
colonies most of the animals studiedbelongedto
the same colony with three combs which was
describedearlier18and is also the focus of the
where a = probability that both animals i and j
are on the nest
presentstudy.The behaviourof individual wasps
comprising this colony with referenceto social
RESULTS
Current Science,January 20, 1985, Vol. 54, No.2
59
organization has been described at length 16.
Briefly, we have constructed time activity budso .
gets
lor
d.. d
m
IVI
.
.d
II
ua
y
I
fi
entl
d
e
.
I
amma
d
s
an
d
wit
raise
d
antennae
an
'
raise
d
.'
WIngs,
Loyalist/
codeNO.16 Rebel
being
away
.'.
from
the
nest
we
have
shown
.17
that the ammals belonged to three distinct behavioural clusters namely Sitters, Fighters and
Foragers. Sitters spend a large proportion of
their time sitting and grooming, Fighters spend a
large proportion of their time sitting with raised
antennae as well as show a high frequency of
dominance behaviour while Foragers are characterised by a large proportion of time being spent
away from the nest. A detailed discussion of the
description and interpretation of such a behavioural caste differentiation is given elsewherel6.
colony
the
caste
of fission
72
oya1st
Loyalist
Loyalist
Loyalist
23
24
Loyalist
Loyalist
Sitter
Sitter
Rebel
Forager
58
Rebel
Rebel
Forager
Fighter
37
19
7
12
19
22
I"
Rebel
Rebel
da ys
at
6
16
21
time
58
24
10
5
3
Forager
Forager
17
8
col!Jny fission. After the colony fission however,
the brood on both the new nest as well as the
parent nest increased quite rapidly (figure I). We
believe that the colony fission itself was the cause
for the sudden increase in brood levels both
because of the dramatic change in brood levels
immediately following the colony fission and
also
The
In
Behavioural
F1"ghter
Forager
Forager
Sitter
L
(iv) walking, (v) being in the cells of the nest and
(vi)
"
ge
-Animal
"
sIttIng
A
b
su
jected these time activity budgets to multivariate
statistical analyses such as principal components
I .
d h
h. I I
I .
U .2
ana YSlSan lerarc Ica custer ana ysls. sIng
data on proportions of time spent in (i) sitting
and
( "' ) grooming
.,
. h' (ii). sitting with raised antennae
III
Table1 Casteandagecompositionofloyalistsandrebels
because
other
factors
such
as
a
change
in
fission
This colony had II adult female wasps and no
males on 29 May 1980. On 30 May 1980,5 female
wasps left the parent nest to initiate a new nest
about a meter away. Henceforth we will refer to
~
~ 6
the group that stayed back at the parent nest as
Loyalists. It is significant that the two groups
~
~' 5
were rather evenly matched in numbers and age
composition. The 6 Loyalists were 72, 58, 24, I 0,
~
~ I,
5 and 3 days old respectively at the time of fission
while the 5 Rebels were 58, 37, 19, 17 and 8 days
:g
§ 3
respectively (table I). The Loyalists consisted of
one Fighter, 2 Foragers and 3 Sitters while the
Rebels had one Fighter and 4 Foragers (table 1).
The Fighter among the Loyalists (animal No.2)
was the queen before the colony fission and
continued to be the queen after the fission. The
8
~ 2
;i
'0 10
Fighter among the Rebels (animal No. 17)
became the queen on the newly initiated colony.
70
t-
0 jM
Y
J
1 ~UN
1 JUL
TIME
Figure 1. Total brood (eggs+ larvae+ pupae) on the
threecombs of the parent nestuntil 30 May, the day of
Brood levels
on
colony fission (shown by arrow). After the fission,
total brood on both parent nest and new nest put
the parent nest had declined over time and
reached its lowest point on 30 May, the day of
together (solid line) and brood on new nest alone
(broken line) are shown.
The
total
brood
(Eggs + larvae + pupae)
60
CurrentScience.January20.1985.Vol.54, No.2
..
environmentalconditions can easilybe ruled out.
Severalother colonies of R. cyathiformis in the
vicinity did not show any such pattern of decrease..AFTER
m brood levels till 30 May 1980followed
by a suddenincrease. It appearstherefore that
the colony fission increasedthe geneticfitnessof
both the queens. The inclusive fitness of the
remaining animals in both groupsalso musthave
increasedas long as they were at leastmarginally
related to the queens.
.BEFORE
~
~
DominanceBehaviour
;;i
D
1.0
~
Rebels were subordinate to all the Loyalists.
Nevertheless,mean frequenciesper animal per
hourofdominanceandsubordinatebehaviour1b
within
each group reveal interesting patterns
~ 0
~
[ 0,"
(Figure 2). Among both Loyalists and Rebels,
dominanceas well as subordinate behaviourwas
relatively high before and declined after colony
fission. Before colony fission the Loyalists
showed a greater frequency of dominance behaviour than the Rebelswhile with subordinate
behaviour the pattern was reversed.When these
0,2
domInance
behaviour
dependIng
..y
on
frequencies
towards
whom
are
they
are
L
roken
directed
(figure 3), we seethat a high level of dominance
behaviour was shown by the Loyalists towards
the Rebels and 'others' while low frequencies
"--:,ere
s~own by th~ Rebelstow~rds Loyalists and
others. No domm~ncebehaviourwas recorded
betweenone Loyailst and another or between
one Rebel and another. Here 'others' r~fers to
wasps.that were presentbefore colony ~ssl0nb~t
~ad disappearedby th.eda~ of .the spilt ~rmrttlng therefore ~o classificationInto Loyailsts.or
Rebels.One m~ghtthus argue that the Loyailsts
were t~e domInant group and the Rebels the
subordInate group.
C
d.
d
oor mate
behaVlour
.greater
patterns
As an index of coordination within eachgroup
before the split, we calculatedYule's association
coefficientsbetweenpairs of wasps.We askedthe
question whether wasps within a group were
.
~
I
[
0
REBELS
L
L
DOM.lNANCESUBORDINATE
BEHAVIOUR BEHAVIOUR
Figure2. MeanFrequencies
of dominancebehaviour
and
up,
I
1
i
.J
I
LOYALISTS
~
~
b
FISSION
2.0
There was no clear cut dominancehierarchy3
therefore
on this colony
not possible
at about
tothe
saytime
for example
of the split.
if allItthe
is
...
FISSION
subordinate
b L
I
.
behaviour
oya lsts ( upper panne
before
(shaded
bars) and
per
animal
per
hour
shown
I) an d R e bel sower
(I
panne I)
after (open bars) the colony
fission.
l:oordinated amongthemselves;for instance,did
they synchronisetheir trips away from the nest?
The Rebelshad a meanassociationcoefficient
of 0.69 with 5 pairs of wasps having a value of
+ I. Whenthe pairs werechosensuchthat waspi
was a Loyalist and waspj was a Rebelthe mean
associationcoefficientwas -0.26 with 4 pairs of
waspshaving a value of -I. The mean association coefficientof 0.69 amongRebelswasfound
by Monte Carlo calculations to be significantly
than zero (P < 0.01) (standard devi..
atlon = 0,10) suggesting that the Rebels synchronisedamongthemselvestheir times of being
on or away from the nest. Similarly the value of
-0.26 for Rebel Loyalist pairs was found to be
significantly less than zero (P < 0.05) (standard
Current Science,January 20, 1985, Vol. 54, No.2
0.61/
61
ANIMAl. /HOUR
avoided
1
'I
ally
~
~
members
of the other
left the parent
an act that increased
the
,::
i'
\
Rebels
group
nest to initiate
and
the inclusive
the
and eventu-
one of its own,
fitness
Loyalists.
of both
Strassmann2O
studied
a North
American
Social wasp Po/isles
exc/amans
where subordinate
workers
had the
choice of staying
with the queens'
main nests or
f
joining
worker-initiated
showed
that
~
In our study
"V~
matched
"V~
and
be
by classified
the day of
intothe Loyalists
fissIon and
or
Rebels,
= O.ll}
suggesting
avoided
each
that
the
Loyalists
other,
the
following
lines.
sion
of
First
to
we interpret
seen before
reproductive
mates.
Second,
aggression
the low
level
high
levels
efficiency
situation
Rebel
remained
high
we believe
that
highly
level
the
of aggres-
the
high
we suggest
competition
decreasing
rearing
.led to
evidence
In Polistes
wasps,
have demonstrated
is no
nest-mates
There
evidence
on the other
hand,
components
of the recognitio~
relationship.
from
bee workers
and
in the
the
largely
coordinated
for
that a
led
thereby
nest.
to
the
Such
a
emergence
of a
subordinate
.but
within
allowed
to
however
for
bees
genetic
sy::em
Smith
have
for ability
sisters.
itself,
same
hive
nition
kind
It is thus
once
have
may have been more
the eleven wasps
because
several
eggs
allowed
fission
on differential
to believe
than one genetic
had
colony
to
of
R.
be of a
habitu-
that
there
line among
of the fission.
simultaneously
in the
can
in the
If kin recogit must
at the time
in this
been
colony
reasons
females
bees
sisters
$tudied,
that is not dependent
We do have
before
full and
if the
of sisters.
in the
that we have
be e was
I
sisters
half
full
However,
between
clear
that
between
full
and
classes
is involved
ation,
laying
not
full
they
to both
cyathiformis
its
to distinguish
between
habituate
to
been
showed
can distinguish
habituate
non
In honey
the environmental"and
Getz
h
f
w
b
d
theIr
of genetic
separated.
newly
cues
ymeanso
b
t
or
I
t
dIstinguIsh
distinguish
of
recog-
'
insects.
group21
recognition
half
nest
level
an im-
fission.
of individual
nest-mates.
testing
as a reflection
was responsible
In short
was
along
speculative)
among
that
reproductive
of brood
results
fission
the fission
of aggression,
group
our
competition
before
of
the
we suggest
level of brood.
high
RebelsI
played
of colony
.
(though
the colony
have
and half sisters raised in the same hive.
.
h
'
h
10 t ese experIments
eac
expenmenta
interpret
interesting,
a
'
later
well
DISCUSSION
like
is growing
th
they
honey
would
the
that in a process that resembles imprinting
emerged,
wasps
learn
certain
chemical
rom
,
'
that were present before colony fissIon
therefore
but had dIsappeared
could not
must
nition
in social
Gamboa
and his
of dominance
behaviourjanimaljhr
shown by Loyalists and Rebels
towards e~ch other and towards 'others', Here 'oth~rs'
We
of individuals
Besides
in the process
elrna
up of the frequency
f
3. Break
Rebels
evenly
h
role
There
deviation
of
IC
communication
portant
refer to anImals
worker
fitness
appear to have behaved
in, a highly coordinat~d
manner even before the fissIon. These two facts 10
our opinion
suggest that individual
recognition
and
Figure
andI
rather
number
age composition.
nes
r
i
were
roo
'5oc.
~
of the
study
inclusive
the two groups
in terms
their
nests
as in the case of our
behaviour
patterns
increased
both the queens and workers.
0;-"
and
satellite
This
is
been
past 16,18,We
62
CurrentScience,
January
20, 1985,Vol.54,No.2
-
therefore believe that our results are highly
suggestiveof recognition of differentgeneticlines
within a colony.
Communication by chemicalmeanshas been
establishedby Jeanne23in a situation somewhat
similar to our own. Polybiine waspsstudied by
Jeanne23initiate newnestsby swarming,that is,
t e massemigration 0 one or more queensan a
number of workers to found a new colony.
Swarming is also seenwhen a nestis abandoned
and all the wasps move on to a new nest site.
Jeanne23 showed that swarming wasps of
Polybia sericea apply a chemical substanceby
dragging their gastersover twigs and leavesand
that these scent marks are later used by the
swarm to orient itself. Although R. cyathiformis
is classifiedamong the non-swarming Vespids13
h
work was supported in part by a grant from the
Department of Science and Technology,
Government of India (Principal Investigator,
R. Gadagkar).
11 December 1984
1 P d. L Ph .
2. West,M. J., Science,1967,157,1584.
3. Gadagkar,R., CurroSci.,1980;49, m.
4. Gadagkar,R.; (Unpublishedobservations.)
5. Litte, M., Beh.Ecol.Sociobiol.,1977,2,229.
6. Litte, M., Z. Tierpsychol.,
1979,50, 282.
7. Litte, M., Sm.iths.
Contr.Zool., 1981,No.327.
8. Vecht,J.Van Der.,Zool. Verhand.
Leiden.,1962,
57, 1.
9. Gupta,V. K. and Das,B. P., Entomon.,1977,~,
209. ..
'.
f
.
t
'
IS
I
. bl
pOSSI
th
t
e
th
a.
I
d
5
e
R
b
e
1
e
h
w~sps
~a~
Z
I
1948
.ari,.,YSIO.OO.,
10.
ave
Gadgll,
43,482.
21
1
".
M.
and
Mahabal,
A.
S.,
Curro
SCI.,
1974,
used a comm~mcatory. mec~amsm similar to
that of swarming Polybla serlcea.
The conclusionsdrawn in this paperregarding
the causes(reproductive competitionand aggression) and mechanisms(kin recognitionand communication) of colony fission are admittedly
speculative.This is especiallybecauseonly one
instance of colony fission has beenstudied. But
colony fissionis an unpredictable eventand even
when a colony under study splits the departing
dlt
t 1
b ' ld
t
b It
a u smayno a ways UI anes near y. may
on the other hand be possible that colonies
...e
11. Gadagkar,R., Gadgil,M. and Mahabal,A. S.,
Proc.Symp.Ecol.Anim.Pop.Zool. Surv. India,
1982,Pt. 4, p. 49.
12. Gadagkar,R., Gadgil, M., Joshi, N. V. and
Mahabal,A. S., Proc. IndianAcad. Sci. (Anim.
Sci.),1982,91,539.
13. Jeanne,
R. L., Ann. Rev.Entomol.,1980,25,371.
14. Altmann,J., Behaviour,
1974,49, 227.
15. Gadagkar,R. and Joshi,N. V., Anim.Beh.,1983,
31,26.
..
16. Gadagkar,R.andJoshi,N.V.,Z.Tlerpsychol.,
198464 15
17 D G'h tt' V . J I Q t .t t .
e, .., n uanlalvee thoogy
I
(e.
d ) P.
closure. Ropalidia is an ideal systemto exploit
colony fission towards answering questions regarding evolution of sociality as well as communication and individual recognition. Its
primitive level of sociality coupled with the
occurrence of small and open, yet sometimes
polygynous nestsand the presenceof suchcom. as co1ony fi sslon
. wh.ICh Increases
.Boulder
P1ex strategies
18. Gadagkar,R. and Joshi,N. V., J. Zool., 1982,
198,27.
19. Gadagkar,R. andJoshi,N. V., In: TheBiologyof
the Social insects.(eds)M. D. Breed,C. D.
MichenerandH. E. Evans.Proc.IX Congressof
the InternationalUnion for the study of Social
Insects, Colorado, USA, Westview Press,
' 1982,187....
grown In large enclosures may undergo fission
and produce daughter colonies within the en-
th
fit
e
f
ness
0
b
th
0
.
groups
IS
h
per
W. Colgan John Wiley and sons New York
1978, p. 115.
"
20.
aps
unpar-
alleled.
ACKNOWLEDGEMENTS
We thank Madhav Gadgil and V. Nanjundiah
for reviewing the manuscript and H. Sharat
Chandra, C. D. Michener and Mary Jane West
Eberhard for many helpful discussions. This
Strassmann,
J..
E.,
In.
Natural
selection
and
social
beh aVlour,
.
(ed s) R ..exan
D AI
d er and D .. W
Tinkle, Chiron Press,1981,p. 45.
21. Pfennig,D. W., Gamboa,G. J., Reeve,H. K.,
Reeve,J. S. and Ferguson,I. D., Beh. Ecol.
Sociobiol.,1983,13,299.
22. Getz,W. W. and Smith,K. B., Nature(London),
1983,302,147.
23. Jeanne,R. L., Anim.Beh.,1981,29,102.