24Aug2002 14:12 AR AR175-EN48-13.tex AR175-EN48-13.SGM LaTeX2e(2002/01/18) P1:GCE ARREVIEWSINADVANCE10.1146/annurev.ento.48.091801.112611 (Somecorrectionsmayoccurbeforefinalpublicationonlineandinprint) E V I E R W S E Annu.Rev.Entomol.2003.48:283–306 I C doi:10.1146/annurev.ento.48.091801.112611 N N Copyright(cid:1)c 2003byAnnualReviews.Allrightsreserved A D V A First published online as a Review in Advance on August 28, 2002 C S B B T OMPARATIVE OCIAL IOLOGY OF ASAL AXA A T OF NTS AND ERMITES Barbara L. Thorne DepartmentofEntomology,UniversityofMaryland,CollegePark,Maryland20742; e-mail:[email protected] James F. A. Traniello DepartmentofBiology,BostonUniversity,Boston,Massachusetts02215; e-mail:[email protected] KeyWords Formicidae,Isoptera,eusocialevolution,colonystructure,divisionof labor (cid:2) Abstract Lackingacomprehensivefossilrecord,solitaryrepresentativesofthe taxa, and/or a definitive phylogeny of closely related insects, comparison of the life historyandsocialbiologyofbasal,livinggroupsisoneofthefewavailableoptions fordevelopinginferencesregardingtheearlyeusocialevolutionofantsandtermites. Comparisonsofaselectgroupofbasalformicidandisopterantaxasuggestthatthe reproductiveorganizationofcoloniesandtheirpatternsofdivisionoflaborwerepar- ticularlyinfluenced,inbothgroups,bynestingandfeedingecology.Opportunitiesfor serial inheritance of the nest structure and colony population by kin may have been significant in the evolution of multiple reproductive forms and options. Disease has beenasignificantfactorintheevolutionofsocialorganizationinantsandtermites, buttheadaptivemechanismsofinfectioncontroldiffer.Evaluationsoftheconvergent anddivergentsocialbiologyofthetwotaxacangeneratenoveldomainsofresearch andtestablehypotheses. CONTENTS INTRODUCTION .....................................................284 ORIGINOFANTSANDBASALANTSYSTEMATICS ......................285 ORIGINOFISOPTERAANDINTERFAMILIALRELATIONSHIPS ...........286 ANCESTRALECOLOGYOFTERMITES: FOCUSONTERMOPSIDAE ...........................................288 INFERENCESREGARDINGANCESTRALLIFEHISTORIES ................288 REPRODUCTIVEPLASTICITYINANTSANDTERMITES .................289 ColonyStructure,Gynes,andReplacementReproductives ...................289 ReproductiveConflict ................................................292 DIVISIONOFLABOR .................................................292 FORAGINGBIOLOGY ................................................295 EVOLUTIONARYPATHOBIOLOGY .....................................296 0066-4170/03/0107-0283$14.00 283 24Aug2002 14:12 AR AR175-EN48-13.tex AR175-EN48-13.SGM LaTeX2e(2002/01/18) P1:GCE ARREVIEWSINADVANCE10.1146/annurev.ento.48.091801.112611 284 THORNE (cid:1) TRANIELLO DISCUSSION ........................................................297 ReproductiveStructureandDivisionofLabor .............................298 SerialReproductiveInheritancebyKin ..................................299 DiseaseRisk,theEvolutionofResistance, andSocialOrganization ..............................................300 ConclusionandProspects .............................................300 INTRODUCTION Thestudyoftheevolutionofcomplextraitsisgreatlyfacilitatedbytheexistence ofphylogeneticintermediatesthatexpressgradualtransitionsincharacterstates, and the comparative analysis of these transitions often provides the most com- pelling data for revealing patterns and developing robust hypotheses regarding selectivefactorsthatinfluenceevolutionarychange.Suchsteppingstoneinterme- diates,however,arerarelypresentasarelativelycompletefossilorlivingseries, reflectingincompletepreservationanddiscoveryoffossils,speciesextinction,the evanescenceofannectantforms,andthefactthatevolutionisnotalwaysagradual process.Eusocialityisahighlycomplextraitofprofoundevolutionaryinterestbe- causeoftheexistenceofsubfertileorsterilecolonymembers.Comparativestudies ofsistergroupsandbasaltaxahavebeeninsightfulinexaminingtheevolutionof eusocialityincladesofbeesandwaspsbecausemodernspeciesshowaclineoflife historiesthatrangefromsolitarytoeusocial.Similarcladisticanalysesareabsent inantsandtermitesbecausealltheroughly10,200speciesoflivingantsandover 2600speciesofextanttermitesareeusocial,andsolitaryancestorsaresufficiently distanttoobscurethelinkagesamongselectiveregimes. ThestructuralelementsofsocialorganizationintheHymenopteraandIsoptera arehighlyconvergent.UnlikethehaplodiploidHymenoptera,however,bothsexes of termites are diploid, rendering explanations for eusocial evolution based on asymmetriesingeneticrelatednessgeneratedbymeiosisandfertilizationinappli- cabletotermites.Nevertheless,thesimilaritiesanddifferencesinthepreadaptive characteristicsofeachgroupandecologicalforcesthatimpelledtheevolutionof socialorganizationmayoffersignificantsociobiologicalinsight.Lackingtheop- portunityforcomparativestudywithintaxa,hereweexplorecommonalitiesand contrastsinthelifehistory,colonystructure,reproductivedynamics,andsocioecol- ogyofthemostprimitivelivinglineagesofantsandtermites.Althoughthesephy- logeneticallydivergentinsectsdifferinfundamentalways(suchasholometaboly inHymenopteraandhemimetabolyinIsoptera),eusocialityisbasedontheelabo- rationoffamilyunitsinbothgroups,andeusocialevolutionmaybeconstructively discerned through focused comparative assessments. Observations on the biol- ogyofextanttaxacannotbeusedtodefinitivelyreconstructancestralstatesprior to the evolution of worker subfertility or sterility and thus cannot appropriately be used to test hypotheses or predictions regarding the evolution of eusociality. Onceprotoantsorprototermitescrossedthethresholdofeusociality,lifehistory constraints, especially those related to reproductive division of labor, may have 24Aug2002 14:12 AR AR175-EN48-13.tex AR175-EN48-13.SGM LaTeX2e(2002/01/18) P1:GCE ARREVIEWSINADVANCE10.1146/annurev.ento.48.091801.112611 BASALANTSANDTERMITES 285 beenessentiallyirreversible.Datathatallowspecificcomparisonsarenotalways available,soacollateralgoalofthispaperistoidentifydomainsofresearchthat wouldfurtheradvancesuchanapproach.Webeginwithanoverviewofthephylo- geneticoriginofantsandtermites,identifythebasaltaxaconsideredinthereview, andjustifytheinclusionsofthosegroupscentraltoourcomparativeanalysis. ORIGINOFANTSANDBASALANTSYSTEMATICS Theoriginsoftheformicidthemeofsocialorganizationhavebeensoughtinthe vespoid wasps. There is a void in social behavior between basal ants and their closestvespoidrelatives,althoughthefossilrecordofferssomeevidenceofhow and when ants attained their morphological distinctiveness and suggests a basic timeline for the emergence of the socially advanced groups. The hypothetical ancestral vespoid wasps are thought to be linked to ants through the subfamily Sphecomyrminae,thepleisiomorphicsistergrouptoallants,withitsextinctCre- taceousfossilgeneraSphecomyrmaandCretomyrma(124).Sphecomyrmafreyi, dating from New Jersey amber of the late middle Cretaceous, exhibits a con- stellation of nonsocial wasp and ant traits: short bidentate mandibles, a reduced andwinglessthorax,apetiolarconstriction,andsignificantly,whatappearstobe ametapleuralgland(33).Itisconsideredthe“nearlyperfectlinkbetweensome ofthemodernantsandthenonsocialaculeatewasps”(33,p.23).Kyromyrmaneffi, thefirstspecimenofanextantantsubfamily(theFormicinae),alsocollectedfrom the New Jersey amber (c. 92 million years ago, has an acidopore and is 50 mya olderthanSphecomyrma(23).Thissuggestsadivergenceofthebasallineagesof ants from the Sphecomyrminae approximately 105–110 mya. Further details of thefossilrecordandadaptiveradiationofantsaregiveninHo¨lldobler&Wilson (33)andCrozieretal.(14),whichprovidemoleculardatadatingtheoriginofants totheJurassic. Thebasaldivisionofthe17antsubfamilies(5)separatestheMyrmicinae,Pseu- domyrmecinae,Nothomyrmeciinae,Myrmeciinae,Formicinae,andtheDolicho- derinae from the remaining subfamilies. The Nothomyrmeciinae, Myrmeciinae, andPonerinaeincludegeneraconsideredpivotalinantsocialevolutionbecauseof theircomparativelyprimitivemorphologyandsocialorganization(33).Thefinal basalgroupofantsistheAneuretinae,aformicoidcomplexsubfamilyonceglobal indistributionbuttodayrepresentedbyasinglespecies,Aneuretussimoni,found inlimitedareasofSriLanka(41). ThesubfamilyNothomyrmeciinaeismonotypic,knownonlyfromthesingle extantandelusiveNothomyrmeciamacropsClarkfromAustralia.Itsrediscovery in 1978 (104) was somewhat akin to finding the “Holy Grail” of myrmecology, andthecollectionandobservationofqueenrightcoloniesmadepossibledetailed accountsofthesocialorganizationofthisrelictspecies(31,104).Thebasalchar- acteristicsofthisantincludeawasp-likemorphology(104),anexceptionallyhigh levelofinactivity,andlowlevelsofsocialexchangesamongworkersintheirsmall colonies(40).Queensdonotreceivefoodorotherpreferentialtreatment;indeed, 24Aug2002 14:12 AR AR175-EN48-13.tex AR175-EN48-13.SGM LaTeX2e(2002/01/18) P1:GCE ARREVIEWSINADVANCE10.1146/annurev.ento.48.091801.112611 286 THORNE (cid:1) TRANIELLO queensandworkersrarelyinteract.Remarkably,queenslivinginintactlaboratory colonies collect and feed on insect prey on their own. The subfamily Myrmici- inaeisrepresentedbytheAustralianbulldogantsofthegenusMyrmecia,which formscoloniesof600–900workers(25).Coloniescanbefoundedindependently bysinglequeensorpolygynouslybygroupsoffemales(12). ThePonerinaeisalargeanddiversesubfamilywhoserepresentativesdisplaya mixtureofbasalandderivedmorphologiesandsocialcharacters.Primitiveponer- ines include Amblyopone, an ant that exhibits morphologically and behaviorally primitive traits, although other genera in the tribe Amblyoponini show highly derivativecharacteristics.Becauseofthegreatdiversityofponerineants,includ- ing numerous species with clearly derived traits (81,82), we concentrate on the moreprimitiveforms,usingA.pallipesasamodel,whilenotingthatotherspecies ofAmblyoponemayvarywidelyintheirbiologyandevenincludequeenlessforms (37).InA.pallipes,alatequeensdiscardtheirwingstoestablishnewcoloniesand forageduringthecolonyfoundationstage(33).Nests,whichhousesmallcolonies averaging roughly a dozen workers, are composed of simple chambers and gal- leriesinsoilanddecayedwood.PopulationsofA.pallipesarepatchybutcanbe locallyabundant(112). ORIGINOFISOPTERAANDINTERFAMILIAL RELATIONSHIPS The higher-level phylogeny of termites has received considerable interest in the pastdecade.AlthoughmonophylyoftheDictyopteraisaccepted(28,50),relation- shipsamongthedictyopteranordersBlattaria,Mantodea,andIsopteraarenotfully resolved(6,17,42,44–47,51,57,59,108).Despitethetopologicaluncertaintyre- gardingwhethercockroachesorsomelineage(s)ofaparaphyleticcockroachclade arethesistergrouptotermites,thereisconsensusthatstudyofthelifehistoryand social organization of the relict wood roach genus Cryptocercus provides con- structivecomparisonandpotentialinsightsintothebiologyofprototermitesand potentialselectiveforcesfavoringtheevolutionofeusociality(10,67,68,107).To date,thefossilrecordexposesnomissinglinksthatindicateintermediatestages betweentheorders,soidentifyingthemostimmediateancestorsofIsoptera,and gleaningthehintsthattheymightrevealregardingthetransitionfromsolitaryto eusociallifehistories,hasbeenimpossible. TheearlyevolutionandintrafamilialrelationshipsofIsopteraalsoarenotfully understood,butseverallinesofevidenceidentifythemostbasallineagesandpro- videincreasingdefinitionoftheirphylogeny.Theearliestknownfossiltermitesare fromtheCretaceousandarerepresentativesoftheHodotermitidae,Termopsidae, and possibly Mastotermitidae. These Mesozoic termites are distinctly primitive butreasonablydiversified,suggestinganoriginoftheorderintheUpperJurassic (109).Hodotermitidae,representedinmodernfaunabythreegenera(19species)of 24Aug2002 14:12 AR AR175-EN48-13.tex AR175-EN48-13.SGM LaTeX2e(2002/01/18) P1:GCE ARREVIEWSINADVANCE10.1146/annurev.ento.48.091801.112611 BASALANTSANDTERMITES 287 highlyspecialized“harvester”termites,hastheoldestdescribedfossil(∼130mya) (52)andatotalofsixgenera(sevenspecies)intheCretaceous.Althoughthefor- agingbehaviorsandcolonyorganizationofextanthodotermitidsarederived,they retainpleisiomorphicmorphologicalcharacters(18). Termopsidae,thesistergroupofHodotermitidae[(16,18,73);butsee(105)],is representedbyatleastfourknowngenera(fivespecies)intheCretaceousandfive moderngenera(20species,the“dampwood”termites).Termopsids,especiallythe relictHimalayanArchotermopsiswroughtoniDesneux,areconsideredbymanyto bethemostprimitivelivingtermiteswithrespecttocolonysize,socialorganization, nestingbiology,andcastepolyphenism(35,71,101,105,107,108). Mastotermitidae,apparentlyrepresentedintheCretaceousbytwogeneraand radiatingbroadlybytheTertiary(4generawithmorethan20speciesknownfrom Australia,Europe,NorthandSouthAmerica,andtheCaribbean)(109),nowexists asonlyasinglespecies,MastotermesdarwiniensisFroggatt,withanaturaldistri- butioninmoist,tropicalregionsofNorthernAustralia.Mastotermitidaeisviewed uncontroversiallyasthemostbasallivinglineagewithinIsopteraandasthesister taxontoallotherlivingtermites(16,17,43,47,48).M.darwiniensishasdistinct pleisiomorphic characters, but it also features a number of highly derived char- acteristics.Forexample,M.darwiniensishasanearlyandapparentlyirreversible splitindevelopmentofnondispersiveforms,soldierssecreteadefensivechemical, malereproductiveshaveauniquetypeofmultiflagellatesperm,colonypopulation sizescanbelarge(severalmillionindividuals),galleryconstructionoccurswithin nests,andextensiveforagingtunnelsconnectfoodsourceslocatedawayfromthe nest(21,53,105,108).M.darwiniensisthusexemplifiesacommonevolutionary pattern:Itretainssomeprimitivefeaturesbutalsohasapomorphicanatomicaland lifehistoryelements. Alongwiththesethreeconfirmedancientfamilies,someclassic(48,75)andone recentfamily-levelphylogeny(43)placeKalotermitidae,includingthe“drywood” termitesalongwithsomedampwoodspecies(53),asamongthemostbasalclades. KalotermitidsdonotappearinthefossilrecorduntilthePaleocene;thereare446 modernspeciesin21genera(11).Currenthypothesesofrelationshipamongthe four basal termite families Mastotermitidae, Hodotermitidae, Termopsidae, and KalotermitidaedifferonlyinthepositionofKalotermitidae.Taxonomicsampling issuesandlackofintegrationbetweenmorphologicalandmolecularstudieshave impededresolutionoffamily-levelphylogenies,butforthepurposesofthispaper weassumethatMastotermitidae,Hodotermitidae,andTermopsidaecomprisethe most basal living termite clades. This is in accordance with recent phylogenetic analysesthatdifferintopology,butincludethesamethreefamiliesasmostbasal [((((((T,R),S),K),(Tp,H)),M),B)1(105);(((((((T,R),S),K),Tp),H),M),B)] [Donovanetal.(16)andEggleton’s(17)“majorityconsensusrule”phylogeny]. 1B,Blattaria;M,Mastotermitidae;H,Hodotermitidae;Tp,Termopsidae;K,Kalotermitidae; S,Serritermitidae;R,Rhinotermitidae;T,Termitidae. 24Aug2002 14:12 AR AR175-EN48-13.tex AR175-EN48-13.SGM LaTeX2e(2002/01/18) P1:GCE ARREVIEWSINADVANCE10.1146/annurev.ento.48.091801.112611 288 THORNE (cid:1) TRANIELLO ANCESTRALECOLOGYOFTERMITES: FOCUSONTERMOPSIDAE BecausesocialorganizationoflivingmembersoftheMastotermitidaeandHodoter- mitidae appears to be derived, we typically draw inferences regarding ancestral socioecologyfromlifehistorypatternsofmodernTermopsidae.Thisinterpreta- tion has been broadly held among termitologists (18,72,76,101,107), although thereissomecontroversyregardingwhetherthedevelopmentalflexibilitytypical ofTermopsidaeisanancestralorderivedcharacteristicoftermites.Thetraditional view is that “true workers,” i.e., individuals that diverge early and irreversibly from the imaginal line, are a derived feature in termites (1,70,72,76). Accord- ing to this view, the worker caste developed at least three times independently becausetrueworkersoccurinMastotermitidae,Hodotermitidae,Serritermitidae, Rhinotermitidae, and Termitidae (3,29,66,76,77,85,86,107). In Termopsidae, Kalotermitidae,andthemostprimitiveRhinotermitidae,helpershavemarkedde- velopmentalflexibilitythroughouttheirlives;allindividualsexceptsoldiersmay differentiateintoreproductives(70,76,87,101,107)orundergoregressivemolts torevertfromthenymphallineinto“pseudergates”(76). Based on hypotheses of interfamilial phylogenetic relationships, however, Thompsonetal.(105)followWatson&Sewell(118,119)insupportingirreversible worker differentiation as an ancestral element of termite social evolution rather thanaderived,phylogeneticstate.Itisdifficulttoevaluatethispostulate,however, becausetherearesofewlivingrepresentativesoftaxakeytothisinterpretation, i.e.,thefamiliesMastotermitidaeandHodotermitidae,andthosespeciesthatexist arehighlyderivedinothersocialattributes(18,48,101,107).Nestingandfeeding habitsmaydrivetheevolutionofsocialbehaviors,obscuringphylogeneticanaly- sesbasedonpresumedhomologoustraits.Forexample,trueworkersareinvariably foundinspeciesthatforageawayfromthenestexploitingmultipleresources,and helperswithlifelongflexibledevelopmentaloptionsoccurin“one-piecenesting” groupsthatconsumeonlythewoodinwhichtheyliveandthereforefaceeventual resourcelimitationandinstability(1,3,29,53).Thiscorrelationsuggestsbiolog- icalsignificancebetweentermitenestingbiologyandpresenceorabsenceoftrue workersinmodernspecies.Theancestralworkerhypothesis(105,118,119)thus carrieslinkedimplications,suggestingforexamplethatorganizedforagingaway from the nest is an ancestral trait and that the one-piece life type with minimal nestarchitectureandforagingrestrictedtothenestwoodissecondarilyderived. Eggleton(17)rationallyadvocatesresolutionofphylogeniesbeforeattemptingto mapsocial,behavioral,developmental,orbiogeographiccharacters. INFERENCESREGARDINGANCESTRALLIFEHISTORIES Considering extant basal ants and termites as “windows” into ancestral life his- tories, it is apparent that individual species in either taxon rarely provide an en- tirelycrediblemodelreflectingthebiologyofthegroupearlyaftertheevolutionof 24Aug2002 14:12 AR AR175-EN48-13.tex AR175-EN48-13.SGM LaTeX2e(2002/01/18) P1:GCE ARREVIEWSINADVANCE10.1146/annurev.ento.48.091801.112611 BASALANTSANDTERMITES 289 eusociality.Modernspeciesbelongingtoeventhemostbasallineageshaveblended assemblagesofprimitiveandderivedtraits,thusconfoundinginterpretations.DNA sequenceshelpresolvethisissueforphylogeneticanalyses,butnomethodological safetynetexistsforevaluatingsocialevolutionbecausehomologyofbehavioral traits can be difficult to verify and may be influenced by ecology or other de- rivedlifehistoryattributes.Wearethuslefttodrawinferencesbasedonsuitesof characters considered to be primitive, compiled from a number of living taxa to yield a composite of likely traits and ancestral ecology of extinct lineages rela- tivelyclosetothecuspofeusocialevolution.Ourcomparisonsoflikelyancestral charactercomplexesfromantsandtermitesideallywillyieldproductiveinsights regarding both commonalities and differences, and therefore potentially signifi- cant influences, favoring the evolution of eusociality in these insects. We focus on four broad and interrelated areas: reproductive plasticity, division of labor, foragingbiology,andevolutionarypathobiology.Wethenconcludewithadiscus- sion of potential commonalities influencing the evolution of eusociality in these groups. REPRODUCTIVEPLASTICITYINANTSANDTERMITES ColonyStructure,Gynes,andReplacementReproductives Developing a conceptual framework for the comparative analysis of reproduc- tive variability in ants and termites has historically been impeded by the num- ber and complexity of fertile and sterile forms, the existence of anatomical and physiologicalintermediates,nomenclaturedifferences,andsemanticcontroversy. To facilitate comparison and clarify our discussion we catalog the types of re- productivesfoundineachgroupusingcurrentlyrecognizedterminology(33,88, 106). Inants,reproductivedivisionoflaborpresentsitselfinthetypicaldimorphic queen and worker castes: The queen is derived from the dispersing alate form, establishesanewcolony,andisdistinguishedfromherdaughtersbysize,theex- tent of ovarian development and behavior. Among basal ant species, the wings may be reduced [as in the case of Nothomyrmecia (31)] and size differences maybelimitedtoasomewhatbroaderthoraxbearingthescarsofthewingsthat are discarded following the dispersal flight. Some basal ants have fertile forms (ergatogynes)thataremorphologicallyintermediatebetweenindependent,dealate colony-foundingqueensandworkersandinseminatedworkers(gamergates).Er- gatoid queens, which are found in some species of Myrmecia (13,26), have a greater number of ovarioles than workers do, a filled spermatheca, and may re- place a typical queen. Some ponerine ants, including Amblyopone (37,83), are queenless.Reproductionbygamergates,whichpossessafunctionalspermatheca andareinseminated,occursinthesespecies.Colonieshavinggamergatesoccupy stressful environments, have reduced dispersal, mate within or nearby the nest, andreproducebyfission,asmaycolonieswithergatoidqueens(83). 24Aug2002 14:12 AR AR175-EN48-13.tex AR175-EN48-13.SGM LaTeX2e(2002/01/18) P1:GCE ARREVIEWSINADVANCE10.1146/annurev.ento.48.091801.112611 290 THORNE (cid:1) TRANIELLO Recentresearch(96,97)hasbeguntouncoverunexpectedandexcitingdetailsof thereproductiveandgeneticorganizationofcoloniesofNothomyrmeciamacrops. Thisbasalantisfacultativelypolyandrous;sampledqueensweresinglyormultiply mated to unrelated males, with an overall average of 1.37 matings per queen (96). Worker nestmates are related by b=0.61 +/− 0.03. Workers appear to be incapable of laying eggs. The mechanism of queen replacement in colonies of N. macrops is rare among ants and bears some resemblance to the pattern of colonyinheritanceexhibitedbysomebasaltermites.Althoughnewlyinseminated queens found N. macrops colonies monogynously, comparisons of worker and queen genotypes in some sampled colonies contained resident queens that were thesistersratherthanthemothersofworkers.Furthermore,larvaeweregenetically identified as the queen’s progeny and not the offspring of reproductive workers. Thelikelyexplanationforthisgeneticstructureisthattheoriginalcolony-founding queenhaddiedandbeenreplacedbyoneofherdaughters. ThecolonylifecycleofN.macropshasbeenreconstructedasfollows(96,97): New queens, one of which may inherit the parental colony, are produced from overwintering larvae that can develop from eggs laid in the autumn into gynes duringthefollowingyear,evenintheeventofdeathofthequeenmother.Are- placementqueencanproducesexualoffspringinherfirstyear.Overall,N.macrops illustrates a low level of serial polygyny; primarily daughters, but at least occa- sionallyunrelatedqueens,areadoptedbyorphanedcolonies.Underthecondition ofcolonyinheritancebydaughters,inclusivefitnessbenefitsextendtotheorigi- nalcolony-foundingqueen(throughtherearingofgrand-offspringfollowingher death)aswellastoworkers(throughtheproductionofniecesandnephews).The brachypteryofnewqueensmayreflectlimiteddispersalandareproductivestrategy designedtofavorreplacementofthemotherqueenbyherdaughters.Ecological constraintssuchashabitatpatchiness,nestsitelimitation,andtherisk-pronefor- agingbehaviorofthepartiallyclaustralfoundingqueensmayhavefavoredcolony inheritanceinN.macrops. Doesthepresenceofsuchreproductiveflexibilityinoneofthemostprimitive extantantsaccuratelyreflectanancestralcondition?Althoughithasbeenargued that the brachypterous queens of N. macrops favored the evolution of daugh- ter replacement (97), it is also possible that brachyptery evolved concomitantly withdaughteradoptionundertheselectivepressureofdispersal-relatedmortality. Again,wenotethedifficultiesinherentinanalyzingtheevolutionofsocialtraits in basal species whose biology may be a constellation of primitive and derived characters. Severaltypesofreproductivesexistintermites.Thetermskingandqueentyp- icallyrefertothecolony-foundingmaleandfemale.Theseprimaryreproductives are imagoes (alates) that drop their wings after pairing. Founding pairs in basal groupsarenearlyalwaysmonogamous,althoughtherearesomerecordsofasso- ciatedgroupsofprimaryreproductives(27).Neotenicsaretermitereproductives thatarenotderivedfromalates,butdifferentiatewithintheirnatalcolony,breeding withaparent,sibling,orotherinbredrelative.Neotenicdifferentiationtypically 24Aug2002 14:12 AR AR175-EN48-13.tex AR175-EN48-13.SGM LaTeX2e(2002/01/18) P1:GCE ARREVIEWSINADVANCE10.1146/annurev.ento.48.091801.112611 BASALANTSANDTERMITES 291 occursupondeathorsenescenceofthefoundingreproductiveofthesamesex(55). Multipleneotenicsofeachsexdevelop,persistingastypicallyconsanguineousre- productive groups in most basal termites (21) but surviving as only one pair in Kalotermitidae (72). Neoteny, literally meaning reproduction as an immature, is relatedtohemimetaboly,requiringoneortwomoltsthatmodifymorphologyand producefunctionalsexorgans(72,106).InMastotermes,neotenicsmaydevelop from workers (119); in termopsids neotenics may form from any individual in instarfourorabove(exceptsoldiersorimagoes)(71),althoughnotrueneotenics are known in Archotermopsis (35,89). Soldier neotenics occur in six species of termopsids (64,107). In Archotermopsis, the gonads of all soldiers are as well developedasinalates(35). Alloffspringhelpersintermopsidfamilies(exceptsoldiers)retainthecapacity to differentiate into fertile reproductives (in the case of termites, either alates or neotenics).Theyarethuspoisedtopotentiallyinherittheirparents’resourcesofa nest,food,andestablishedfamily(65,107).Insuchasystemofserialreproductive inheritancebykin,asinthecasesofergatogyneandgamergateantsreproducing in their natal nest, all colony members gain inclusive fitness benefits and some individualsattaindirectfitnessadvantages.Thesecumulativefitnesscomponents maywellexceedaverageindividualfitnessprospectsofdispersing,fertileoffspring inasimilar,solitaryspecies,thusfavoringhelpersthatremainintheirnatalcolony. InZootermopsis,numerouscoloniesmaybeinitiatedinthesamelog,eventually resultinginintercolonyinteractions,whichcanleadtodeathofreproductivesand opportunitiesforreplacementbyneotenics(107). Inancestralgroups,thereproductiveskewbetweenreproductivesandhelpers may have been less discrete. Imms (35) reported that worker-like individuals of Archotermopsis wroughtoni have extensive gonad development and a fat body equivalent to alates. He observed a captive worker-like A. wroughtoni lay seven eggs. The eggs did not develop normally, but whether due to sterility, lack of fertilization,orlaboratoryconditionsisunknown.Eusocialityitselfisviewedas a continuously varying categorization depending on the portion of progeny that reducesorforegoesreproduction(101,103). Thepossibilityofmergedorindistinctcoloniesfunctioningwithinsinglepieces of wood has been raised several times. Concerning Archotermopsis, Imms (35, p.126)observed,“Ihave,onseveraloccasions,comeacrossthreeorfourqueens with a single large colony of ova and larvae, which probably represent several colonies which have become confluent.” Fused colonies or colony complexes have also been suggested in Stolotermes (20,63,110) and Zootermopsis (B.L. Thorne, personal observation). These observations and their generality, context, and implications are difficult to evaluate; identification of discrete but adjacent colonies within a log is often impossible. Sufficient descriptive evidence exists, however, to encourage genetic examination of these circumstances, especially relationships among the reproductives found with the possibly merged groups. Recent work on M. darwiniensis (21) suggests that although neotenics within a colony are often inbred, they sometimes originate from more than two genetic 24Aug2002 14:12 AR AR175-EN48-13.tex AR175-EN48-13.SGM LaTeX2e(2002/01/18) P1:GCE ARREVIEWSINADVANCE10.1146/annurev.ento.48.091801.112611 292 THORNE (cid:1) TRANIELLO lineages,ashasbeenindicatedinsomemorederivedtermites(8).Extensivestudy oncolonygeneticstructureandthepossibilityofmergingorintroductionofforeign reproductivesisrequiredforbothbasalantsandtermites. ReproductiveConflict Conflict among nestmates whose fitness interests are incongruent is common in socialinsects.Theapparentrarityofnuptialflightsinsomebasalantgenera(13), thepresenceoffertilehelpersinbothantsandtermites,andmultiplereplacement reproductives raise the possibility of intracolony reproductive conflict. Repro- ductive conflict is manifest in oophagy, the existence of inhibitory pheromones, dominance structures, mutilation, and policing behaviors (37,62). In basal ants, larvalhemolymphfeedingbyqueenshasbeendescribedinPrionopeltaandAm- byloponesilvestrii(38,60)andhasbeeninterpretedasaformofqueennutrition, althoughthebehaviorcouldalsorepresentamechanismtoregulatereproductive capability. Because of the monogynous and monoandrous organization of basal termite societies,conflictssimilartothoseobservedinantswouldnotbeexpecteduntil colony members approach a state of reproductive competence; then policing or otherrelatedmechanismsofreproductivecompetitionmightbeevidentinspecies withflexibledevelopmentbecausenearlyallindividualshavethepotentialtodiffer- entiateintoreproductiveforms.Roisin(85)citesreportsofintracolonymutilation in termopsids, kalotermitids, and some rhinotermitids and proposes that compe- tition among late instar helpers, including nymphs attempting to become alates might explain such behaviors. He suggests that siblings bite wing pads, which causessomeindividualstodeflectfromalatedevelopment,creating“lowerstatus” helperswithreducedchancesoffuturedispersal.Subsequentwingbudregenera- tionandformationofanormalalateispossible,butwithdelayandadditionalmolts (107,118).Roisin(85)proposesthatthemutilated“losers”inintracolonyconflicts formed the original helpers in termites. The contexts under which primitive ter- miteslosewingbudsneedtobebetterunderstoodbeforethishypothesis,orthe implicationsofmutilationbehaviorsintermites,canberigorouslyevaluated(107). Forexample,wingbudscarsintermopsidsareoftenduetoself-inducedabscis- sion rather than mutilation by colony members (35,107). Research on complete colonies of Zootermopsis in the laboratory suggests that self-abscission occurs when there are opportunities to become a replacement reproductive, perhaps in- ducingpre-alatestoshedwingpadsanddifferentiateintoaneotenicinthenatal colony(107). DIVISIONOFLABOR Theprimaryaxisofdivisionoflaborinbasalantandtermitespeciesisreproduc- tive, but colonies theoretically may partition tasks according to the size and age ofsubfertileorsterileindividuals.Pheromones,temperature,andnutritiondirect