cannibalism of subordinates’ eggs

8/6/2019 Cannibalism of subordinates’ eggs

http://slidepdf.com/reader/full/cannibalism-of-subordinates-eggs 1/4

499Naturwissenschaften 84 (1997) © Springer-Verlag 1997

Cannibalism of subordinates’ eggs

in the monogynous queenlessant Dinoponera quadriceps

Thibaud Monnin, Christian Peeters1

CNRS URA 667, Laboratoire d’Ethologie Experimentale et Comparee,Universite Paris Nord, 93430 Villetaneuse, France

Received: 15 May 1997 / Accepted: 20 August 1997

Naturwissenschaften 84, 499–502 (1997) © Springer-Verlag 1997

1Present address: CNRS URA 258, Labora-toire d’Ecologie, Universite Pierre-et-MarieCurie, 7 quai St Bernard, 75252 Paris cedex05, Francee-mail: [email protected]

About 100 species of ants in the sub-family Ponerinae have lost the mor-phologically specialized queen caste,and workers can mate and lay fertil-ized eggs. All workers have identicalphenotypes, but only a proportion of them reproduce sexually (these aretermed gamergates [1]). Either mono-gyny (one gamergate) or polygyny(several gamergates per colony) oc-curs depending on the species [2]. Re-productive division of labor in mostqueenless ants is regulated by aggres-sive interactions leading to the forma-tion of hierarchies [3], similarly as invarious social wasps lacking di-morphic queens and workers [4]. Sub-ordinate workers generally have unde-veloped ovaries, except that in manymonogynous species a few high-rank-ing subordinates can often lay eggs.Unlike wasps where many femalesmate in each colony, in many mono-gynous queenless ants only the top-ranking worker (“alpha”) can mate,and thus other egg-layers are re-

stricted to producing males. Alpha isoften able to destroy subordinates’eggs, which is distinct from the con-sumption of nonviable “trophic” eggs

common in other ants. In Diacammasp., the single gamergate eats most of the haploid eggs occasionally laid byvirgin workers [5]. Similarly in somefunctionally monogynous Polistinewasps, alpha selectively destroys theeggs which are laid by subordinatesduring the initial stage of colonyfoundation (“queen policing”) [6].

Dinoponera quadriceps is a queenlessant with small colonies (82±29 work-ers, range 39–141, n =17) collected inBahia state (Brazil). Colonies aremonogynous; dominance interactionslead to a linear hierarchy, and onlythe alpha worker copulates [7]. Ovi-position and oophagy were observeddirectly in 15 colonies kept in the lab-oratory. We report here that a fewhigh-ranking subordinates infrequentlylay unfertilized eggs, but that theseare destroyed by alpha. Oophagyseems to be regulated by an identifiedchemical signal present on the cuticleof alpha workers and on their eggs.A total of 129 ovipositions were re-

corded during 885 h, and this lowegg-laying rate is characteristic of thePonerinae [2]. Alpha workers (whocan produce eggs even before theymate) laid 111 eggs (Table 1). In ninecolonies alpha monopolized oviposi-tion, while in the six remaining colo-nies subordinate workers (having thebeta, gamma, and delta ranks in thehierarchy) laid 18 eggs (14% of total).Dissection of all workers (n =914)

confirmed that alphas (virgin or ga-mergate) have much better developedovaries than the subordinate egglayers (Fig. 1). In colonies where beta

workers were not observed to oviposittheir ovaries were undeveloped.The fate of 100 newly laid eggs wasmonitored for about 30 min to deter-mine the natural pattern of oophagy.New eggs have a pure white colorwhich darkens after a few hours; dark eggs were never eaten. New eggs aregenerally carried by the mothers forseveral minutes before being depos-ited in the egg pile. All but one egglaid by alpha survived, while in con-trast most of those laid by subordi-nates were destroyed (13/18, i.e.,

72.2%; Table 1). Alpha was responsi-ble for 12 of 14 observed cases of oo-phagy (85.7%); virgin alphas behavedsimilarly to gamergates. Oophagy oc-curred 22±7 min (n =9) after oviposi-tion, but exceptionally an egg waseaten after 65 min. Ten eggs werestolen directly from the mandibles of mothers. Five eggs escaped destruc-tion because alpha did not comeacross them (in one case she wasbusy ovipositing). There is good evi-dence that alpha recognizes new eggswhich are not her own and cannibal-izes them (“queen policing”). Whensubordinates carried eggs newly laidby alpha, they were not destroyed if alpha encountered them again.Furthermore, in one case alpha re-moved and ate a new egg recentlyadded to the egg pile 37 min after ithad been laid by a subordinate.Using solid-phase microextraction(SPME) we previously showed thatalpha differs from other colony mem-bers by the quantity and relative pro-portion of a long-chain cuticular hy-

drocarbon (9-C31: 1 or 9-hentriacon-tene) [8]. To investigate whether thereis more 9-C31 on alphas’ eggs than onsubordinates’, we measured 13 newlylaid eggs with SPME, which is a non-destructive technique (no solvent isused). Ovipositing workers were re-moved from their colonies while anegg was already half outside of thegaster. Due to excitement mothers lettheir egg fall to the ground, and thus



there was no direct contact betweenmandibles and egg. The latter was im-mediately taken with forceps andsampled with a Supelco 7-lm polydi-methylsiloxane fiber designed to ex-tract high-weight hydrocarbons. Theglass fiber was rubbed on each sideof an egg for 1.5 min, thus totaling3 min of rubbing (eggs are large:

3.3±0.14 mm long and 0.8 mm wide,n = 34). Afterwards the fiber was in-serted in the injector port of a gaschromatograph for desorbtion (columntemperature increased from 708C to300 8C at 5 8C/min, then isothermic at300 8C for 20 min).The cuticle of egg-layers was alsomeasured on the day of the oviposi-tion or sometimes 1 day later, follow-ing the method described previously

[8]. Since oviposition by beta is rare,we removed alpha (virgin or gamer-gate) in two colonies and used thefirst eggs laid by the replacement al-pha. Beta almost always accedes tothe top rank, and both the percentageand amount of 9-C31 increase after 2weeks [9]. Thus eggs initially laid bythe replacement alpha are considered

to be equivalent to eggs laid by beta.Fourteen major peaks exceeding anarbitrary threshold were used to calcu-late the relative proportion of 9-C31

[8]. As seen in Fig. 2, there was apositive correlation between percent-age of 9-C31 present on ovipositingworkers and that on their eggs. Alphayielded more 9-C31 than the replace-ment alpha (7.7±2.7 and 3.2±0.4%,n = 9 and 4 respectively, Mann-Whit-

ney U test: P<0.01; larger samples de-scribed in [9]), and alpha-laid eggsyielded more 9-C31 than eggs laid byreplacement alphas (16.2±4.9 and6.1±0.9%, n = 9 and 4 respectively,Mann-Whitney U test: P<0.01). Oncean egg appears at the tip of the abdo-men, 3.6±0.4 min (n = 25) elapse be-fore it is fully laid, and 9-C31 maypassively diffuse from the cuticle of the mother to her egg. It is not clearwhy the proportion of 9-C31 shouldbe higher on the eggs than on the cu-ticle (Fig. 2), but this may be relatedto surface properties of the chorion.Sixteen eggs laid by alphas from fivecolonies, and thus presumably yield-ing a high percentage of 9-C31, were

immediately removed and transferredinto foreign colonies (n =4) to test thehypothesis that alpha recognizes hereggs on the basis of a high percentageof 9-C31. Indeed, she is the onlyworker of the colony with this cuticu-lar profile, and thus eggs rich in 9-C31 are always her eggs under naturalconditions. Only 4 of these 16 intro-duced eggs were destroyed by the re-sident alphas. This contrasts with our

500 Naturwissenschaften 84 (1997) © Springer-Verlag 1997

Table 1. Oviposition and oophagy observed during 885 h in 15 colonies of D. quadriceps. Thebeta worker laid eggs in six colonies, gamma laid in three colonies, and delta in two colonies

Status of egglayer

Eggs laid Eggs removed forchemical analysis

Eggs left in the nest to check for oophagywithin 30 min of oviposition

or introduction toforeign nest Not

destroyedDestroyedby alpha

Destroyed bysubordinates

Gamergate 49 9 39 0 1Virgin alpha 62 20 42 0 0Beta 11 4 7 0Gamma 4 1 3 0Delta 3 0 2 1Total 129 29 86 12 2

Fig. 1. Pattern of ovarian activity based on the dissection of 914 workers from 15 colonies.Subordinate workers had developed ovaries in six of these colonies, while in others only alphahad developed ovaries. The mean number of both mature oocytes (as large as an egg) and

opaque yolky oocytes differed significantly between alpha (virgin or mated) and subordinateegg layers (Mann-Whitney U test, P<0.05; identical letters refer to values which are not differ-ent). All other workers had undeveloped ovaries

Fig. 2. Relative amount of 9-hentriacontenepresent both on the cuticle of mothers and ontheir eggs. *, Replacement alpha (equivalentto beta); v, virgin alpha; *, gamergate. Thepercentages of 9-C31 extracted from workersand their eggs are correlated by a linear re-gression ( y =1.68 x+2.41, r 2= 0.687, Pearsonproduct-moment correlation: P<0.05)



previous observations that 13 of 18eggs laid by subordinates had beendestroyed by nestmate alphas. Thus,transferred alpha eggs were less likelyto be eaten ( v2 test: P<0.01), althoughoophagy occurred more often (4/16)than in the normal situation (1/82).This may be due to disturbance while

introducing the eggs in a tested colo-ny or to the colony odor of the intro-duced eggs which are perceived asalien by a resident alpha. Indeed,workers of Ectatomma tuberculatumdiscriminate between larvae and co-coons from same or different colonies[10], and queens of the polygynous

Leptothorax acervorum preferentiallyeat nonnestmates’ eggs rather thannestmates’ [11].In some Polistine wasps lacking mor-phologically specialized queens, oo-phagy could also be regulated by an

olfactory signal present on the eggs.In Polistes dominulus, alpha destroysnew foreign eggs which are experi-mentally introduced, while introduced1-day-old eggs are rarely destroyed,and 30-h-old eggs not at all [12]. InP. dominulus alpha has also a specificcuticular hydrocarbon profile which isrelated to ovarian activity [13], andwe hypothesize that this odor diffusesto the eggs during oviposition and in-hibits oophagy, as in D. quadriceps.This is supported by the observationthat at the beginning of the egg-layingperiod in P. dominulus, alpha hasslightly developed ovaries [14], andsometimes destroyed her own eggs[12]. Since female wasps lacking fullydeveloped ovaries do not have thecharacteristic hydrocarbon profile of dominant reproductives [13], theireggs may not carry the signal inhibit-ing oophagy. This also appears to bethe case in D. quadriceps, where sub-ordinate egg-layers yield less 9-C31

than alpha because their ovaries areless developed. Therefore alpha can

discriminate their eggs and destroythem.In honey bees, where the castes arehighly dimorphic, very few workersoviposit in queenright colonies, andtheir eggs are destroyed by otherworkers [15]. These policing workerscan discriminate queen-laid eggs fromworker-laid eggs because the formerare marked with pheromones appar-ently originating from the Dufour’s

gland of the queen [16]. In Polistes fuscatus, eggs experimentally rubbedwith Dufour’s extract from subordi-nates were destroyed by alpha, whileeggs rubbed with secretions from al-pha were not destroyed by subordi-nates [17]. However, it remains un-clear whether alpha’s secretions really

inhibit oophagy because they wereonly tested with subordinates, andthese normally do not destroy eggs[6, 12]. Additional factors may inhibitoophagy in Polistes, because eggs arelaid in individual cells, unlike antswhere they are generally gathered inpiles. Thus the alpha wasp may mem-orize the cells where she recently ovi-posited and destroy new eggs encoun-tered in other cells [12]. Egg markinghas also been reported in Solenopsisinvicta, where ovipositing queens de-posit poison gland products onto their

eggs with the sting. These chemicalsare not involved in the regulation of oophagy, however, but apparently at-tract workers to the eggs and protectthem from fungus and bacteria [18].In colonies of D. quadriceps that arenot recently created by fission of amother colony, all workers are daugh-ters of the gamergate. Since she matesonly once [7], workers are full-sistersand are more related on average tosubordinates’ eggs developing innephews (0.375) than to the gamer-gate’s unfertilized eggs developing inbrothers (0.25). Thus natural selectionis unlikely to favor the destruction of subordinates’ eggs by sterile workers(“worker policing” [19]). In contrast,most workers in honey bee coloniesare half-sisters (the queen mates mul-tiply) and are more related to thequeen’s haploid eggs than to nestmateworkers’ eggs. Thus worker policingleads to the destruction of worker-laideggs [15]. Furthermore, “queen polic-ing” would not be possible in thehuge Apis colonies, unlike in Dinopo-

nera where there are only very fewsubordinate egg-layers to watch.The regulation of selective oophagyby an olfactory cue present on theeggs and related to the ovarian activ-ity of mothers may only be possiblein species with significant inequalitiesamong egg layers (i.e., alpha and betain D. quadriceps), such as the queen-less ant Pachycondyla sp. from WestJava [20] and some Polistes wasps

[6]. In contrast, such regulation seemsimpossible in species where severalindividuals have fully developed ova-ries and lay eggs equally. All the egglayers would yield the putative“dominant” odor and mark their eggs,preventing differential oophagy. Thereis indeed no oophagy in the polygy-

nous queenless ants Pachycondylaberthoudi [21], Leptogenys schwabi[22], and Harpegnathos saltator (J.Liebig and C. Peeters in preparation),while oophagy is indiscriminate in thequeenright ant Leptothorax acervorum[11] and the Polistine wasp Ropalidiarufoplagiata [23], both of which arepolygynous.Oviposition by a few high-rankingsubordinates appears to be a stablecharacteristic of reproductive systemsregulated by a linear dominance hier-archy. Skew models predict that alpha

yields some reproduction to subordi-nates as an inducement to remain inthe society, i.e., “staying incentives”[24]. In D. quadriceps, although alphahas an efficient mechanism to recog-nize and destroy eggs which are nother own, it is adaptive for a few sub-ordinates to oviposit since having de-veloped ovaries gives them an advan-tage over other nestmates when anopportunity to become alpha arises.Furthermore, several eggs did surviveand should develop into males whichcan copulate in foreign colonies(mating opportunities occur through-out the year [7]). From the colony’sperspective, subordinate ovipositionshould not be counterselected becauseit allows for the quick replacement of an existing reproductive, followingeither the decline of an old gamergateor the advent of colony fission (whichoccurs instead of independent founda-tion in permanently queenless ants).

We are grateful to C. Malosse (INRAVersailles) for help with the chemical

analyses.

1. Peeters C, Crewe R (1984) Inseminationcontrols the reproductive division of la-bour in a ponerine ant. Naturwis-senschaften 71:50–51

2. Peeters C (1993) Monogyny and poly-gyny in ponerine ants with or withoutqueens. In: Keller L (ed) Queen numberand sociality in insects. Oxford Univer-sity Press, Oxford, pp 234–261

501Naturwissenschaften 84 (1997) © Springer-Verlag 1997



3. Heinze J, Holldobler B, Peeters C (1994)Conflict and cooperation in ant societies.Naturwissenschaften 81:489–497

4. Roseler P-F (1991) Reproductive com-petition during colony establishment. In:Ross KG, Matthews RW (eds) The so-cial biology of wasps. Cornell Univer-sity Press, Ithaca, pp 309–335

5. Nakata K, Tsuji K (1996) The effect of

colony size on conflict over male-pro-duction between gamergate and domi-nant workers in the ponerine ant Dia-

camma sp. Ethol Ecol Evol 8:147–156

6. Reeve HK (1991) Polistes. In: RossKG, Matthews RW (eds) The socialbiology of wasps. Cornell UniversityPress, Ithaca, pp 99–148

7. Monnin T, Peeters C (1998) Monogynyand regulation of worker mating in thequeenless ant Dinoponera quadriceps.Anim Behav (in press)

8. Monnin T, Malosse C, Peeters C (1998)Solid phase microextraction and cuticu-lar hydrocarbon differences related to re-

productive activity in the queenless ant Dinoponera quadriceps. J Chem Ecol(in press)

9. Monnin T (1997) Regulation de la re-production chez la fourmi sans reine

Dinoponera quadriceps: structures hier-archiques et communication olfactive.Unpublished Ph.D. thesis, UniversiteParis Nord, 139 p

10. Feneron R, Jaisson P (1995) Ontogenyof nestmate brood recognition in a pri-mitive ant, Ectatomma tuberculatum Oli-vier (Ponerinae). Anim Behav 50:9–14

11. Bourke AFG (1994) Indiscriminate eggcannibalism and reproductive skew in amultiple-queen ant. Proc R Soc Lond255:55–59

12. Gervet J (1964) Le comportementd’oophagie differentielle chez Polistesgallicus L (Hymen. Vesp.). Ins Soc11:343–382

13. Bonavita-Cougourdan A, Theraulaz G,

Bagneres A-G, Roux M, Pratte M, Pro-vost E, Clement J-L (1991) Cuticularhydrocarbons, social organization andovarian development in a polistinewasp: Polistes dominulus Christ. CompBiochem Physiol 100B:667–680

14. Turillazzi S (1980) Seasonal variation inthe size and anatomy of Polistes galli-

cus (L.) (Hymenoptera: Vespidae). Mon-it Zool Ital (NS) 14:63–75

15a. Ratnieks FLW, Visscher PK (1989)Worker policing in the honeybee. Nature342:796–797

15b. Ratnieks FLW (1993) Egg-laying, egg-removal, and ovary development by

workers in queenright honey bee colo-nies. Behav Ecol Sociobiol 32:191–19816. Ratnieks FLW (1995) Evidence for a

queen-produced egg-marking pheromoneand its use in worker policing in thehoney bee. J Apicult Res 34:31–37

17. Downing HA (1991) A role of the Du-four’s gland in the dominance interac-tions of the paper wasp, Polistes fusca-

tus (Hymenoptera: Vespidae). J InsectBehav 4:557–565

18. Vander Meer RK, Morel L (1995) Antqueens deposit pheromones and antimi-

crobial agents on eggs. Naturwissen-schaften 82:93–95

19. Ratnieks FLW (1988) Reproductive har-mony via mutual policing by workers ineusocial Hymenoptera. Am Nat 132:217–236

20. Ito F (1993) Functional monogyny anddominance hierarchy in the queenlessponerine ant Pachycondyla (=Bothro-

ponera) sp. in west Java, Indonesia.Ethology 95:126–140

21. Sledge MF, Crewe RM, Peeters CP(1996) On the relationship between ga-mergate number and reproductive outputin the queenless ponerine ant Pachycon-

dyla (= Ophthalmopone) berthoudi. Na-turwissenschaften 83:282–284

22. Davies SJ, Villet MH, Blomefield TM,Crewe RM (1994) Reproduction and di-vision of labour in Leptogenys schwabi

Forel (Hymenoptera Formicidae), apolygynous, queenless ponerine ant.Ethol Ecol Evol 6:507–517

23. Sinha A, Premnath S, Chandrashekara

K, Gadagkar R (1993) Ropalidia rufo- plagiata: a polistine wasp society prob-ably lacking permanent reproductive di-vision of labour. Ins Soc 40:69–86

24a. Reeve HK, Ratnieks FLW (1993)Queen-queen conflicts in polygynous so-cieties: mutual tolerance and reproduc-tive skew. In: Keller L (ed) Queen num-ber and sociality in insects. Oxford Uni-versity Press, Oxford, pp 45–85

24b. Keller L, Reeve HK (1994) Partitioningof reproduction in animal societies.Trends Ecol Evol 9:98–102

502 Naturwissenschaften 84 (1997) © Springer-Verlag 1997

Molecular and Morphological Data Suggesta Single Origin of the Polydnavirusesamong Braconid Wasps

James B. Whitfield

Department of Entomology, University of Arkansas, Fayetteville, AR 72701,

USA; email [email protected]

Received: 22 October 1996 / Accepted: 28 August 1997

Naturwissenschaften 84, 502–507 (1997) © Springer-Verlag 1997

Endoparasitoids exhibit some of themost intricate and intimate interspecif-ic physiological interactions recog-nized among animals [1]. These typi-

cally wasplike insects feed as larvaewithin the bodies of other insects (of-ten caterpillars), ultimately killingtheir hosts. A wide variety of endopar-

asitoid/host interactions are known [2],which in turn are mediated by diverseagents such as venoms, teratocytes,and viruslike entities. These mediatingagents, introduced into the host by thefemale parasitoid during oviposition[3, 4], misdirect or dismantle thehost’s defense response and develop-ment, thereby enabling the larva(e) todevelop freely inside [5].

Perhaps the most remarkable agentsintroduced into the hosts of some en-doparasitoid wasps (Braconidae andIchneumonidae) are viruslike entitiesknown as polydnaviruses. Polydna-viruses are unusual in two respects.First, they are integrated into thewasp’s chromosomal DNA andpassed on from parent to offspring inmendelian fashion [6–8]. Second,within the cells of the wasp’s ovarian

cannibalism of subordinates’ eggs

Documents