Phylogeny of the Ammobatini and revision of the Afrotropical genera (Hymenoptera: Apidae: Nomadinae)

66 Pages·1997·21.6 MB·English

by  C D Eardley

J.HYM.RES.
Vol.6(2),1997,pp.353-418
Phylogenyofthe Ammobatini and Revision ofthe Afrotropical
Genera (Hymenoptera: Apidae: Nomadinae)
CONNALD.EaRDLEYANDDeNISJ.BROTHERS
(CDE)BiosystematicsDivision,PlantProtectionResearchInstitute,PrivateBagX134,Pretoria,
0001SouthAfrica;(DJB)DepartmentofZoologyandEntomology,UniversityofNatal
(Pietermaritzburg),PrivateBagXOl,Scottsville,3209SouthAfrica
—
Abstract. ThephylogenyofthegeneraofAmmobatiniwasstudiedusingcladisticmethods
andtheclassificationisconsequentlyrevised.Thetribeformsamonophyleticgroupthatcom-
prisesfivemonophyleticgenera:PasitesJurine,SphecodopsisBischoff,AmmohatesLatreille,Mela-
uewpisSaussureandOreopasitesCockerell,andonemonotypicgenusSpinopasitesWarncke.They
alloccurintheAfrotropicalRegionexceptOreopasites,andtheAfrotropicalspeciesarerevised.
Pasites(MorganiaSmith,OmachthesGerstaeckerandPasitoiiiachthesBischoff,newsynonymy)in-
cludes17Afrotropicalspecies,Sphecodopsis{PseudopasitesBischoffandPseudodichroaBischoff,new
synonymy)10species,andAinwobatesandMclaneinpisareeachknownfromasingleAfrotropical
species.Tennewspeciesaredescribed(allattributedtoEardleyalone):Pasitesiiilssoni,P.paiilyi,
P. Iiumectus, P.gnomus, P. iiarnihiensis, P. somalicus, Sphecodopsisvespericena, S. longipygidnitn,S.
namaquensis and Ammobates auster. Thirty-fourspecies namesarenewlysynonymized: Pasites
nigerrimusFrieseandPasitomachthesargentatusBaker=Pasitesbarkeri(Cockerell);Morganiachubhi
Cockerell,M.nigritulaBischoffandM.peratraCockerell=PasitesfrieseiFriese;Omachthesnigripes
aFOrmniadecsehM,t.hMeaolsrtigoaarlnbCiooagcuktfeotrartteiulsslCoF=crkiPeeasrseei,ltleM,soMrca.gransniuifbeafxorn(taGitesarlCesioticaskeiecsrkeeClrol)c,;keMMr..elsnltiogrardnii/tdihoCMroa.cxkeoSgrtierllavlin,ediM=C.ocPv.koeiderinecslhilsroCu=oscSPkamesiritethle;ls
jCeoncskeeuriel(lFri=eseP)a;siMteosrghaisntiraioh(iGsetrrisotatercaknesrv)a;alMenosrigsanBiisachmoafrfs,haMll.iaCloicvkaelerneslilsC=ocPkaesriteelsljoannedsiM(.Cocrkuefirtealrls)i;s
OmachthesahessinicusFriese,MorganiafiilviventrisBischoff,M.rhodesianaBischoff,M.apicalisBis-
MBCbcehioocoscrhfckgifheai,oranfeMnfiil,.aclaSlt.eu=CoraonnlcieSgskrpoieheCrenoCecscoliokcsldkeoerBpareisenlsilldcslh,=oMmfiM.fnS.ipabihprtneoeicdlvsoiistdMciuoeomlpparassgiC(asoCCncoosikcceaekkmreeiperrrleeeulrlfl,plal)uM;n(c=.CtPoaaciPstknaiaedstreieceCtsiloeslcspa)ky.eaCgproKpmecelalkeyeletsruoe=snltiloFS,(rtpCihhMoeee.sccekgo,needruonedSpelipsrlhci)aeas;cuaoadMndcaooidprdCesgoAaicaftskrnaeoirrt(auerFflrorelipus,eifcsiceMeian)l.ls;a
speciesareprovided.
Thepurposeofthisstudyistoprovide edge of the identity of the Afrotropical
a comprehensive revision of the system- speciesandtherelationshipsbetweenspe-
aticsoftheAfrotropicalAmmobatini(Ap- cies.
idae:Nomadinae).Toattainthisobjective. The Ammobatini are cleptoparasitic
threemainaspectswereinvestigated.The bees,alsoknownascuckoo-bees.Theylay
firstwastoestablishwhethertheAmmo- theireggsinthenestsofpollen-collecting
batini isamonophyletictaxonand, inso bees that provision each larval cell with
doing, establish its validity. The second sufficientfoodforthelarvatodevelopto
wastoreviewthegenericclassificationof maturity. LikeotherNomadinae, theam-
theAmmobatiniandgainanunderstand- mobatines lay each egg in a hole in the
ingoftherelationshipsamongthegenera, wallofthehost'slarvalcell.Whentheegg
Thethirdwastoacquireasoundknowl- hatches the tiny mobile first-instar larva.
354 JournalofHymenopteraResearch
whichhaselongatesharp-pointedmandi- &Michener(1993)definedtheNomadinae
bles, kills thehostlarvaoreggand then toincludeonlythosegenerathatcomprise
consumesthefoodprovidedforthehost. a monophyletic group, namely the Am-
An outstanding account of cleptoparasi- mobatini, Ammobatoidini (includingHol-
tismisgivenbyAlexander(1990). copasites Ashmead), Biastini, Caenoproso-
Pollinatingbeesareoneofthemostim- pidini, Epeolini,Neolarrini,Townsendiel-
portant groups ofbeneficial insects. It is Hni, Hexepeolini, Nomadini and Brachy-
therefore important to study thesystem- nomadini. Other groups that were
atics and biology of their cleptoparasites previously included in the Nomadinae
becausetheyaffectthepopulationdynam- (nowAnthophorinae)are:Isepeolini,Pro-
ics of pollen-collecting bees. Bees of the tepeolini, Osirini (including Epeoloidini
ammobatine genus Sphecodopsis Bischoff, and Parepeolus Ducke, Roig-Alsina 1989)
fcioerseoxfamSpcrlaep,tearreLekpenloewtinerto&pSaerravsiiltliese(Rsopze-- a(AnldexCaoenldieorxoi1d9e9s0,,nRooiwg-iAnlstihneaT&etMraipcehdeinienri
en&Michener1968).Bothofthesegenera 1993).
are endemic to southern Africa. Scrapter The Nomadinae has its greatest diver-
species pollinate indigenous plants and sity in the Nearctic and Neotropical
are potentially important in the pollina- Regions, is fairly well represented in the
tion of agricultural crops. The genus is Afrotropical and Palaearctic Regionsand
prevalent in the semi-arid regions of is poorly known from the Oriental and
southern Africa where insect-pollinated Australian Regions. In the Afrotropical
indigenousplantsformanimportantcom- Region,itcomprisesfourtribes:Nomadi-
ponentofthegroundcoverandpasture. ni, Epeolini, AmmobatoidiniandAmmo-
Cleptoparasitism among bees is a de- batini. The Nomadini, Epeolini and Am-
rived trait and it has evolved indepen- mobatoidiniarerepresentedthereonlyby
dently several timeswithinthebees (Al- theirnominategenera.TheNomadiniand
exander 1990). The features that are the Epeolini were revised by Eardley &
uniquetocleptoparasiticbees,suchasthe Schwarz (1991) and Eardley (1991), re-
loss of the scopa, are derived, although spectively. TheAmmobatoidini isknown
theymayresembletheprimitivestatefor fromtheAfrotropicalRegionfromasingle
the bees as a whole. Other features that female specimen described as Ammobato-
separatecleptoparasiticbeesfrompollen- ides hrauusi Bischoff. The Ammobatini is
collecting bees are the thickened integu- the largest and most diverse tribe ofAf-
mentand themorerobuststing (Alexan- rotropicalnomadinebees.
der 1990). There are many otherconver- The Ammobatini occur mainly in the
gent traits in the cleptoparasitic bees Nearctic, Palaearctic and Afrotropical
which, as indicated by Alexander (1990), Regions.Althoughtheyhavenotbeenre-
complicateattemptstotracetheirphylog- corded from theOriental Region, several
eny. Palaearcticspeciesareknownfromareas
In spite of the difficulties involved in thatbordertheOrientalRegionandthese
the study ofthe phylogeny ofcleptopar- species possibly extend into that Region.
asiticbees,theNomadinae,thelargestand TheydonotoccurintheNeotropicaland
most diverse lineage of cleptoparasitic Australian Regions. The Nearctic (Rozen
bees (about 1200 described species), has 1992)andPalaearctic(Warncke1983)fau-
been well studied by Rozen (1966, 1974, nashavebeenwellstudiedandtheAfro-
1977, 1991), Rozenetal. (1978),Roig-Alsi- tropicalfaunaisrevisedhere.
na (1987, 1991), Alexander (1990) and Mostofthe previousworkon thesys-
Roig-Alsina & Michener (1993). Roig-Al- tematics of the Afrotropical Ammobatini
sina(1991),Rozen(1991)andRoig-Alsina consistsofdescriptionsofnewspeciesand
Volume6,Number2,1997 355
distributionrecords.Theoriginaldescrip- thesubfamily Anthophorinaeofthefam-
tionsaregenerallyvague,withoutillustra- ily Anthophoridae. Popov's Ammobatini
tionsandinadequatefortherecognitionof contained the genera Ammobates (for
the species. Bischoff (1923) provided a which he described two new subgenera,
comprehensive revision ofthe Afrotropi- XerammobatesPopovandEuphileremusPo-
calAmmobatini,buthisworkhasseveral pov),Caesarea,ParammobatodesPopovand
shortcomings(e.g.,hedidnotstudymuch Oreopasites. HisPasitiniconsistedofPasi-
of the type material) and it has become tes, Morgania, Omachthes, Pseiidopasites,
outdated. The present study is the first Sphecodopsis, Pasitomachthes and Pseudodi-
treatmentofthesebeesthathasincluded chroa.HemadenomentionofMelanempis.
anexaminationofnearlyallthetypema- Sustera(1958),incontrast,placedtheNo-
terialandastudyofmaleandfemaleter- madini, Ammobatini and Pasitini in the
minalia. Andrenidae.
Authorship of the new species de- Baker (1971), inhisdiscussiononPasi-
scribed inthispaperisattributedtoCDE tomachthes,supportedPopov's(1951)clas-
only. sification.Rozen&McGinley(1974)found
HISTORIACMAMLORBEAVTIIENWIOFTHE eavniddepnhcyeloignetnhyeirofstthuedylaornvateheofsytshteesmeabteiecss
thatOreopasitesandPasitesarecloselyre-
The history of the classification ofthis lated, with Ammobates somewhat diver-
groupofcloselyrelatedbeesmaybeout- gentandSphecodopsisfarthestaway.
lined as follows. Anearly attempt to ar- Warncke(1983),inarevisionofthePa-
range them into a system ofhigherclas- laearcticfauna,tookacompletelydifferent
sification was by Dalla Torre (1896). He approachandplacedalmosttheentirePa-
placed allbeesinthefamilyApidaeand laearctic and Afrotropical faunas of am-
placed thegenera thatarecurrentlycon- mobatine bees {sensu Michener) into the
sidered to belong to the Ammobatini, genus Pasites, which he subdivided into
namelyAmmobatesLatreille,PasitesJurine six subgenera: Parammobatodes, Spinopasi-
and OmachthesGerstaecker,togetherwith tesWarncke,MicropasitesWarncke,Euphil-
several other genera, in the subfamily eremus,AmmobatesandPasites.Heconsid-
Coelioxyinae.Ashmead(1899)dividedthe ered Morgania, Omachthes, Pasitomachthes,
Apidae into several families and trans- PseudopasitesandSphecodop'sistobejunior
ferredtheCoelioxyinaetothefamilyStel- synonyms of Pasites (sensu stricto).
idae,whichincludedmostoftheparasitic Warncke(1983)didnotgiveadetailedex-
bees. planation for his actions and made no
Michener(1944)providedthefirstcom- mentionofOreopasitesandMelanempis,ex-
prehensive study inwhichbeeswereas- ceptthatinchangingthenamePhileremus
signed to tribes. He placed Oreopasites aterSaussure(=Melanempisatra)toPasites
Cockerell, Ammobates, Morgania Smith, madagascarensis he indicated that he con-
OmachthesandPasitesinthetribeAmmo- sidered Melanempis to be a synonym of
batini(Apidae:Anthophorinae).Michener Pasites.
(1944)alsosuggestedthatCaesareaFriese, Subsequent to Wamcke's (1983) study,
Melanempis Saussure, Parammobatodes thetribalclassificationoftheNomadinae,
Friese, Pasitomachthes Bischoff, Pseiidodi- based onadultmorphology, wasstudied
chroa Bischoff and Sphecodopsis Bischoff byRoig-Alsina(1987,1991)andAlexander
mightbelongintheAmmobatini. (1990), neither of whom adopted
Popov(1951)dividedMichener's(1944) Warncke's(1983)classification.Roig-Alsi-
Ammobatini into two distinct tribes, the na (1987), inhisdiscussiononthephylo-
Ammobatini and the Pasitini, placed in genetic relationship between the Caeno-
356 JournalofHymenopteraResearch
prosopidini,BiastiniandAmmobatini,de- MATERIALSANDMETHODS
MfotuafisibnecltehehdeeWonatNnehmoretcmhk(aeA1ed9m4'dim4ins)osa'tb.era(,iA1tb9luis8ettn3ixa)ioatnendcdialnenatrsdhtsai(hth1feo9ihs9ce'0at)ts,ideronienisdcneonrbhodeiot-ssf pafsliehlcnIyanttltheyiaedonnokfbanoytfaotnwetodmnhvpeeetArlAfutfrco8roi0otdd0trearotmopesopipnictescahctaleirlmsagpetAeenenmcsemit)roehisbecaa(mntcroldieanpnsrosiaeil-,--l
causehehadnotstudiedthegroupinsuf- available extra-African representativesof
ficient detail. Rozen (1992) discussed the the tribe (183 specimens from the Pa-
tribal characters in detail without recog- laearcticandNearcticRegions)werestud-
nisingeithertribes. ied. Alltheavailabletypematerialofthe
Roig-Alsina(1987, 1991)andAlexander Afrotropicalspecieswasexaminedduring
(1990) demonstrated that the Caenopro- thecourseofthestudy.Thetypematerial
sopidiniisthesistergroupoftheAmmo- ofeight specieswas not studiedbecause
batini.IntheAmmobatinithesixthmeta- itcouldnotbeobtained:Pasitomachthesar-
somalsternum(S6)ofthefemaleisbifur- gentatusBakerwasidentifiedfromthede-
cateorsecondarilysimpleposteriorly,and tailed description and comparison with
in the Caenoprosopidini this structure is typematerial ofotherspecies; Omachthes
bilaterallyseparated.Theinference,bythe capiensis Friese, Pseudodichroa fiimipennis
above mentioned authors, that thebifur- Bischoff and Phileremus (Melanempis) ater
cate female S6 in the Ammobatini gave Saussurewerereliablyidentifiedfromau-
riCinaseetnhtoeopCrtahoeesnobopipliradotiesrnoaiplilydiissneipaairmamptolenideospchotynhdlaitettitiohcne atVhtaorcrahitatulal,tuisMvoelrFygriadenseieta,errmoOtimunanecddihctemhpaestseBriigsaaclbh;oofnPfeasnaistniedss
gntihr,eoruhpoR.woeiTvgh-eeArl,msiownnaaosp(n1h9oy8tl7,yde1om9fo91nt)shteorrAaAmtlmeedoxbabanytdieei-r- Msbmeaoadtrsiesge.radinTailhm.aeosItstnrtlfoupyoidrcymaoanotCfoipcoerknxeetvorriean-olulAtshferlryielcmaaadbneielnttseaiorxnfamcietwrnytaepasdee
(NstcBi1itaona9rtIims9anet0atdi)-edIhn.-diiti.shCntaaahAtseeeltnt,ueohsdxpeiiasrRetNnooseedsiroegool-rtpnaAairl(xdtr1soihi9innen9nii0ao))pf,whcay(thl1slhoa9oewd8pgee7o(e)vsnAtesoymridb,meboloemfyibontttdanhhhi--eee- sldioucaanfopnbtredetmedcleha,ssia'tr.t.mceores'lViFnelOaeosqetlgrcuheiteetstonhorracatrtteheapi.ecnwuorodbaTnrlhsdmifteelasnyodthopdreeatiervldssiestarrtrladebu,iricadbesotiutrteitrfdhdimireoboaanufrmrtlrereiooecoWggmacnhilioiivvrttrteedhyeenens,-
sister group of this clade. Subsequently (1983). The acronyms for the museums
Roig-Alsina (1991), using different char- from which material was borrowed are
acters,cametothesameconclusionasAl- listed in the acknowledgements section.
exander(1990).Theabsenceofapygidial Where geographic coordinatesaregiven,
plate,asmentionedbyRoig-Alsina(1987) theyareindegreesandminutes(separat-
is not a synapomorphy of the ((Ammo- ed by a period), not decimal degrees.
batini-I-Caenoprosopidini)-I-Biastini),as Where reference is made to 'the Code',
severalspeciesofPasiteshavewelldevel- thismeanstheInternationalCodeofZoo-
opedpygidialplates.Recently,Roig-Alsi- logical Nomenclature, 3rd Edition (Inter-
na&Michener(1993)consideredthemto national Commission on Zoological No-
belongintheApidae.Thecurrentfamilial menclature1985—).
placementofthesebees,atopicthatisbe- Morphology. Theterminologymainly
yondthescopeofthestudy,hasbeenac- followsthatofMichener(1944).Sexualdi-
cepted. morphism in adult Ammobatini is slight
Volume6,Number2,1997 357
and, apart from Sphecodopsis, Oreopasites, theNeolarriniwasentirelydifferentfrom
Melanempis and Spinopasites in which that in the (Ammobatini + Caenoproso-
males have eleven flagellar segments, is pidini), and thesistergroupofthe(Neo-
largelyconfinedtotheposteriorregionof larrini + (Ammobatini + Caenoprosopi-
the metasoma. A single detailed descrip- dini)), the Townsendiellini (Roig-Alsina
tion for both sexes of each species has 1991)(representedbyTownsemiiellacalifor-
thereforebeen given, withthediagnosHc nica Michener)was used topolarize that
sex-limited characters ofeach sex explic- character (the relationships between the
aitrleyudseescdrifobredt.heThmeetaabsbormeavliatteirognasaTndansdterS- NAemomloarbraitnii,niTo+wnsCeanednioeplrloisnpiidainndi aAnmdmoth-e
na,respectively(e.g.TlandSIrefertothe batoidini were questioned by Alexander
first metasomal tergum and sternum, re- (pers. comm.) following additional re-
spectively). Vestiture generally refers to search, however). Where a possible evo-
the relatively fine hairs and where setae lutionary progression could be deter-
arespecificallymentionedthesearethick- minedbetweendifferentstatesofachar-
er hairs. The sixth metasomal tergum of acter within the in group, successivede-
the female of some species has a brush rived states (0 = primitive; 1, 2 & 3 =
posteriorly (located below the pygidial successive derived states), with nonaddi-
platewhenthisstructureispresent).This tivebinarycoding,wasused.
brush hasbeenreferredtoasthesubpy- The different states of each character
gidial brush (Fig. 6); when it has thick were incorporated in data matrices. A
hairsdorsallyandfinevestitureventrally question mark (?) was used where the
ithasbeenreferredtoasdifferentiated.In state in a species could not be studied,
certain taxa the posteromedian region of such assex-limited characters forspecies
the fifth metasomal sternum of the fe- inwhichtheappropriatesexisunknown
male,whenviewed frombehind,formsa orwasnotavailableforstudy.Inorderto
distinctfurrow. Rozen's(1968a)terminol- rootthecladogram, ahypotheticalances-
ogyhasbeenusedforthisstructurewhich torwithallcharacterscodedas wasadd-
hereferredtoasbeing'gutter-like'.Inthe ed.Thefirstmatrix(Table1)givesallthe
illustrationsofthemaleterminaliathean- relevant information on each species. In
teriorendisatthebottomandtheposte- the formation of the second data matrix
riorendatthetop. (Table2)somespeciesweregroupedinto
Cladistics:Adultsofeachincludedspe- species groups (reasons given below),
cieswerethoroughlyexaminedandeach each of which is represented by the
character for which distinct states oc- groundplanofthatgroup(derivedasex-
curredindifferentspecieswasincludedin plainedbelow).Thethirdmatrix(Table4)
thematrix.Polarizationofcharacterswas includesonlygroundplansofthegenera
based strictly on out-group comparison andwasderivedfromthesecondmatrix.
andtheputativesistergroupwastakenas CladogramsweregeneratedusingHen-
theoutgroup. Intheanalysisattempting nig86, version 1.5 (Farris 1988). The first,
todemonstratethemonophylyoftheAm- second and fourth cladograms (Figs. 1, 2
mobatini as a whole, the sister group of &4)originatefromtheanalysisofthein-
the Ammobatini, the Caenoprosopidini, formationinTables1,2&4,respectively,
wasincludedintheingroup.TheNeolar- without using character weighting (com-
rini (represented by Neolarra vigilaiis mandsm*;bb*).Thethirdcladogram(Fig.
(Cockerell)), which is the sister group of 3)resultedfromtheuseofsuccessiveap-
(Ammobatini + Caenoprosopidini)(Roig- proximationscharacterweighting(repeat-
Alsina 1991), was then taken as the out ed application of m*; bb*; xs w) in the
group. Foronecharacter(50)thestatein analysis of the information in Table 2.
358 JournalofHymenopteraResearch
Table1. Datamatrixofcharacterstatesforspecies(charactersandcodingofstatesaccordingtoAppendix2).
I^i\»"i characters
Ancestor 0000000000 0000000000 0000000000 0000000000 0000000000 00000000
SSppihll.,cvaepsipceorliacena& 2001111101 0110010001 0012110100 1000010001 1001001101 10000000
Sph.hngipygidium 2001111171 0110010001 0012110110 1000010001 1001001101 1???????
Sph.villosa 2001111101 0110010001 0012110100 170001???? 77???????? 70000000
Sph.namaquensis 2001111171 0110010001 0012110100 1000010001 1001001101 1777777?
Sph.minutissima 2001111001 0110010001 0012111100 1000010001 1001001101 10000000
Sph.aculeatn 2001111101 0110010001 0012110100 1000010001 0001001101 10000000
Sph.semirufa 2001111101 0110010001 0012110100 1000010001 0001001201 10000010
Sph.capcnsis&
Sph.fumipennis 2001111171 0110010001 0012110100 1000010001 0001002301 27777777
P.barkeri 1001100011 0110010101 0021001010 1010010000 0110010001 10000000
P.friesei 1001100011 0110010101 0021001010 1110110100 0110010001 10010000
P.nilssoni 1001100011 0110010101 0021001010 1710117777 ??77777777 70010000
P.paulyi 1001100011 0110010101 0021001000 1110110100 0210010001 10010000
P.braunsi 1001100071 0110010101 0021001010 1110110100 0210010001 1?77?77?
P.humectus 1001100011 0110010101 0021001000 1110110110 0110010001 10010000
P.bicolor 1001100011 0110010101 0021001010 1010110110 0110010001 10010000
P.tegularis 1001100011 0110010101 0021001010 1010110100 0110010001 10010000
P.gnomiis 1001100011 0111010001 0021001000 1110010000 0110010001 12010100
P.carnifex 1001100011 0110010101 0022110110 1110010000 0210010001 10000000
P.dichrous 1001100011 0110010101 0022110110 1110010000 0210010001 10010010
P.jenseni&P.histrio 0011100011 0110010101 0011001010 1110010000 0210020001 10000000
P.namibiensis&P.jonesi 0111100012 0110012101 0011001010 1110010000 0210020001 10000000
P.rufipes 2001111171 0110010101 0022110010 0111001000 0210020001 17777777
P.appletoni 1001100011 0110010101 0021001010 1110010100 0210020001 10000000
P.sonwlkus 1001100071 0110010101 0021001010 1110010100 0210020001 17777777
P.maculatus 1011100011 0101010101 0021001010 1111010000 0210020001 10000000
Spi.sphwtus 1001100071 0100010111 0021001000 1100000001 0001001201 2777777?
M.atra 2101111102 1001011111 0022110010 0000011010 0001001301 10010000
O.vanduzeei 1001100002 1000011011 0020010000 1100010001 0001001201 11000000
O.linskyi 1001100002 1000021011 0020010000 1100010001 0001001201 11000000
A.verhoeffi 1001100002 1000010111 0021000000 0010010001 0000101211 11000010
A.rostratus 0111100002 1000010111 0020001000 0110000001 0000101211 11000010
A.robustus 0111100002 1007710111 0020001000 01???77777 ????77?777 7?????77
A.teheranicus&:A.bniieri0111100072 1000010111 0020001000 0110000001 0000101211 17777777
A.hipporiensis 0111100002 1000011111 0020001000 0710007777 ?????77777 71000010
A.mavnmwiistakisi 0111100002 1000010111 0020001000 1110000001 0001101211 10000010
A.handlirschi 0111100002 1007711111 0020001000 1110007??? 7777777777 71000010
A.deprcssus 0111100072 1000010111 0020001000 0100000001 1000101211 11000020
A.punctatiis 1011100002 1000010111 0020001000 0110010101 0000101211 11000000
A.ancylae 1111100002 1007710111 0020001000 1110010001 0000101201 11000020
A.solitarius 1011100072 1000010111 0020001000 1110010001 0000101201 1???????
A.vinctus 1011100072 1000011111 0020001000 1110010001 0000101211 11000010
A.auster,A.similis,
A.niveatus&:A.si/riacusl011100002 1000010111 0020001000 1110010001 0000101201 11000010
A.nifiventris 1011100002 1000011111 0020001000 1110010001 0000101201 11000010
A.irankus 1011100002 1000011111 0020001000 1110010001 0000101201 11000020
A.dubius 1011100002 1000010111 0020001000 1110010001 0000101201 11000020
A.assimilis 1011100002 1000010111 0020001000 1110017777 77???????? 71000020
A.opacus 1011100072 1007711111 0020001000 1110010001 0000101201 1???????
A.armeiiiacus 1011100002 1007710111 0020001000 1110010001 0000101201 11000020
A.sanguineus 1011100002 1000010111 0020001000 1111010001 0000101201 11000020
A.biastoides 0001100002 1000010111 0020001000 1111010001 0000101201 11000010
Volume6,Number2,1997 359
Table1. Continued.
360 JournalofHymenopteraResearch
Table2. Datamatrixofcharacterstatesforspeciesgroups(charactersandcodingofstatesaccordingto
Appendix2).
Taxon
Voix'ME6,Number2,1997 361
maTtaibolnes3c.harWaecitgerhtweaisgshitginnegd(tmoaexaicmhucmhawreacitgehrtd=ur1i0n).ganalysisofdatainTable2aftersuccessiveapproxi-
VVei>^ht Charjcler
Weight10: 4,5,15,16,20,21,22,30,35,39,42,43,45,47,49,50,53,56,58.
Weight5 10,46,48,52,57.
Weight4: 9,25,29,44.
Weight3; 23.
Weight2 6,7,8,11,13,26,28,33,40,41,54,55,57.
Weight1 1,12,18,19,24,32.
Weight0: 2,3,14,17,27,31,34,36,37,38,51.
single species group, the matrix was al- Thesameapproachwasadoptedinthe
teredinastepwisemannerfromthemost developmentofgroundplansforthegen-
primitivetothemostderivedstate).After era (Table4), whichresultedinthemore
each alteration the resultant matrix was derivedstate(i.e.,state1)beingpreferred
re-analyzedandthetreelengthcompared fortwocharactersinAmmobates.Forchar-
withthatoftheformeranalyses.Thedata acter1states and1gavetreesofsimilar
setthatgavetheshortesttreewaschosen. length.State1 waspreferredbecauseitis
Wherethedifferentstatesprovidedclado- apparently moreprimitive forthegroup,
grams of equal length, each state was being reversed in A. rostratus. InAmmo-
studied for evidence suggesting which hatesstate1ismorecommonandthisstate
state was primitive forthegroup. In the alsooccursinSpinopasitesandOreopasites,
absenceofsuchevidencethemoreprimi- whereas state2 occurs in (Melanempis +
tivestatewaschosen.Thisprocedurewas Sphecodopsis)(Fig.2).Forcharacter31 the
repeatedforeachspeciesgroup. derived state(i.e.,state1)gavetheshort-
ForAmmobates,intheoxianusgroupthe esttree.
morederivedstateforcharacters10,11& PHYLOGENYOFTHEAMMOBATINI
19 gave theshortesttree (i.e., state 1). In
both the rostratus and punctatus groups Asindicatedabove,ananalysisincluding
the more derived state for character 31 allspeciesseparatelywasnotsuccessfulbe-
(i.e.,state1)andintherostratusgroupfor cause of missing data. The analysis of
character 52 (i.e., state 2) gave a shorter ground-planadultcharacterstates(Appen-
tree and was therefore preferred. In the dix2)ofthespeciesgroups(Table2),how-
punctatusgroupcharacter57isrepresent- ever, resulted in 48 most parsimonious
e&d2b)y, otfhrewehidcifhfesrteanttest1atgeasv(ei.et.,hestastheosrt0e,st1 cTlhaedosgtruadmys,ofeeaacchhwoifththaesleen4g8thtroefes1,5a4nsdtetphse.
tree. strictconsensustree(Fig.2)indicatedacon-
Table4. Datamatrixofcharacterstatesforgenera(charactersandcodingofstatesaccordingtoAppendix2).
I.iM.n
362 JournalofHymenopteraResearch
sistent pattern ofseven major clades. The gidial plate (state 52.1), which associates
basal branch consistently represented the Ammobatesand Oreopasitesandwasgiven
Caenoprosopidini, which was included in aweightof5intheanalysisusingsucces-
anattempttodemonstratethemonophyly sive approximations character weighting
oftheAmmobatini,whiletheothersixma- (Table 3). The mandibles posterolaterally
jorcladesmoreorlessrepresentedtheam- directedinrepose(18.0)andtheundiffer-
mobatine genera as defined by Michener entiated vestitureon theventrolateralre-
(1944). Theconsistencyin thecomposition gion of the mesepisternum (27.0), which
ofthesesixcladesledtotheirbeingconsid- group Oreopasites with (Melauempis +
eredheretoconstitutegenera{Pasites, Spi- Sphecodopsis),haveweightsof1 and re-
nopnsites, Oreopasites,Ammohates,Sphecodop- spectively (Table3). ThegroupingofOr-
sisandMelauempis,Fig.1).Mostofthetrees eopasiteswithAmimibatescanalsobemore
differedonlyintherelativepositionsofthe easily explained when considering the
speciesgroupswithineachclade. biogeographyofthesebees,theformerge-
Analysis of the data using successive nus is Neotropical and the latter mainly
approximations character weighting re- Palaearctic, where as the other ammoba-
sultedin12mostparsimonioustrees,each tinegeneraareprimarilyAfrotropical.
witharawlengthof156steps.Thediffer- The analysis of generic ground plans
encesinthelengthsofthetreesproduced (Table 4), without character weighting,
withandwithoutcharacterweightingap- gaveasinglemostparsimonioustree(Fig.
parently resulted from the differentcon- 4). The configuration of the tree differs
figurations of the species groups within fromthatproducedbytheformeranalyses
eachgenusandnotfromdifferencesinthe (Figs. 1-3)onlyintheplacementofSpino-
configuration of the genera. The final pasites.Thedifferenceissignificantbecause
weight assigned to each character in the itmakesOreopasitesthesistergroupofAm-
weightedanalysisisrecordedinTable3. mobates,whereasintheanalysisofspecies
Theonlydifferenceintherelationships andspeciesgroupsOreopasites,Ammobates
betweenthegenerainthe48mostparsi- and(Melauempis + Sphecodopsis)(Figs.2&
moniousunweightedcladogramswasthe 3)aremorecloselyrelatedtooneanother
relativepositionofOreopasites.Insomeof than to Spinopasites. The reason for the
thecladogramsitformedthesistergroup changeisthatinthegenericgroundplans
ofAnmwbates, with (Sphecodopsis + Meln- of Spinopasites, Oreopasites and Ammobates
nempis)asthesistergroupof(Ammobates thehindtibiahasthicksetae(32.1)andthe
+ Oreopasites) (Fig. 3), while intheother posteromedianregionofthefemaleS5has
cladograms it formed the sistergroup of adistinctprotuberance(48.2).InSphecodop-
(Sphecodopsis + Melauempis), withAmmo- sis and Melauempis thehind tibiahasfine
batesasthesistergroupof(Oreopasites + vestiture (32.0). Sphecodopsis has a small
(Sphecodopsis+Melauempis)).Thisresulted posteromedianprotuberanceonthefemale
in a polytomy for Oreopasites, Ammobates S5 (48.1)andMelauempishasa large, gut-
and (Sphecodopsis + Melauempis) in the ter-like protuberance (48.3), making the
consensustree(Fig.2).ThepositionofOr- derivationoftheposteromedianprotuber-
eopasites in the consensus tree, produced ance on the female S5 dichotomous. Be-
usingsuccessiveapproximationscharacter causeoftheabsenceofinformationonthe
weighting (Fig. 3), was the same as that maleofSpinopasites it ispossiblethatdis-
which occurred most frequently among coveryofthemalemayaltertheinterpre-
the trees produced without character tationoftherelahonshipbetweenSpinopas-
weighting and was accepted as themost itesanditscongeners.
probablephylogeny.Evidencesupporting Discussion of the generic relationships
thischoiceisthereductioninthemalepy- isbased mainly on the cladogram ofthe