Scientific Papers

Natural History Museum
The University of Kansas

07 August 1998 Number 9:1—-47

A Cladistic Analysis and Classification of the
Subgenera and Genera of the Large Carpenter Bees,
Tribe Xylocopini (Hymenoptera: Apidae)’

By
Rosert L. MINCKLEY?

Division of Entomology, Natural History Museum, and Department of Entomology,
The University of Kansas, Lawrence, Kansas 66045-2109, USA

CONTENTS
ANBIS TRY CIP cases loaseltosiackl HGS hSnae Neeakt SARS PSOne OSs SRG ARUBA Sore Sn on at nn nie a ee oe ee eee 2
IRIN SIRES: COD To) UO) Ca TNO) Net sere ee acces ates chee Heenan se cccns sec eesv's ones sf Gusa sddenued saacdau sede. aidoasetveiadézsucieadvusdtaancas sees 3
PNGRINOMVIEEEXG MENUS Here teense tecr ieee ee cus Conte cee face race ce sveveas Canc (Ei Suns sueat suiutuaedees uve senduvie deucedovandbaacesvdeerdacs ccwsh ves 3
(VATE RIVALS: ANNI) JMU 01 (0) DSS a aye pe PB ate BE CCRC ee eee ee pee eee ee ee 3
SVE CTITIONT OR TANG fa nccen ce snanbocndo tanga Aga cea GHG sp ERE An REEDS Er ROG Oo EEE Peete EERE CEE: eee eae eee eB 3
SHIMB Cl ONIORICHAR ACTER Serco cence tceste ss asee esa rset soc ta cn cavenie ne caver suusices ses ives vet etsesetven lode: Heiss jecdecbnees deccuet 5
GEA GER SPATS IVAINDD CODING Here eetncce ceetrneceseaces cree corse ese creas cee ones cee ae recrec sacek ste tucc sence teeusetes suncuesevencesteses 7
SER VAMOS ENE IIGP ANNA TEV. SES et catccciccascescccovescsctess ccuvvevsscuvcast ses sss éxseuuesdseasiassinssusdvegeseescncvaviedasaceveteseacescdenes 13
LCE SUUTLITS ssscoaccasatbobinchociuasacead dec Badre Sno c EASE HC Acne ESE SES ea eee are ae oer ey eee 14
RELATIONSHIPS AMONG LARGER CLADES OF THE XYLOCOPINI .......csscccsssssccessscecesscceeesececseneeecessneeeseeeeeeane 15
RELATIONSHIPS AMONG MEMBERS OF THE MAJOR CLADES AND TAXONOMIC RECOMMENDATIONS ........0-+0++« an
REV SsOvMHE SUBGENER NOE XV O COPA rie oe cccccse heen cascccedsoteciseasderscvsshensasestodectievscesdueesnscesuscesatees Di
INE WAVY ORD IEM ATES peer anennpesente etree i etre, co Le Sac SAS Payee ex soca oudeecansdebuessSecevbass cpteecteevaresesenecasaiies Dil
RE WANVORUGION VUNIGE Siemens ees cette sameeren cor <cSaddk Cotas san dsectiecs devseasescteoswstadiestssevevveckeecvendedecaedeiecadess 28
AOSV ORIABERE MATES eet ce cies Aiea ceed ee Un eetes sae Se dh tae oua cus es ase eda vecteetueseTessedessvvieecsdesdvesvennctars <besstees 28
MOEA VOR VAIN S wee memento te rue eR EM names c are sha ssc sachsseecuavsavercsaavevedasscevissarsavetschycesbexsetssesesscsetee 29
DIAGNOSES AND DESCRIPTIONS OF SUBGENERA OF XYLOCOPA .....cccccccccesccceseeeseetereeenseeenees 31
SHDPEMHSEXW/OMMEIISSA EAUCG aNd! MOUIE aie. .-eeaccnatenceevecsccscocesocaseeasceossesecosessasssteasecsassseceztivedseosssseoetes 31
SSH DPE INISM VOCE UNA eee cunrae tence caer eae ec eee oe ee eae e ae eae sonst van uutvoastesesvacsberstsssstantesecourssesszerastessseee 31

‘Contribution No. 3198 from the Department of Entomology and Snow Entomological Division, Natural History
Museum, The University of Kansas.
Present address: Department of Entomology, 301 Funchess Hall, Auburn University, Auburn, Alabama 36849-5413, USA.

© Natural History Museum, The University of Kansas ISSN No. 1094-0782

ii)

ScleNTIFIC PAPERS, NATURAL History Museum, THE UNIVERSITY OF KANSAS

Subgenus Maaiana new Subgenus .........cecsecseceeeececresresesessesseessenenessssesssessessevensesscseeseeseenenneenenneaneny SY.
Subgenus Biliuria Maa ........sscssecsecssecseceessenseeeeneeseeseceessesscesessessecscsassssanecsecsusenssssssvesnseneasecnscanacsecsieenets 32
Subgenus Rhysoxylocopa Hurd and MOure .........csssecsecsesseseeseeseesessessesesseesesseeseessansesssneaeneessessensenes 32
Subgenus Proxylocopa Hedicke, New Status .......seseececcesesssessessesseeneessesesseneesseseenseneeneeeeecseenseneats 33
Subgenus Nyctomelitta Cockerel .........:cssessecsecsessecseessessesesseesssssesssseesecsssnssersenessssnssesseasensscsecssenseenaes 33
Subgenus Ctenoxylocopa Michene .........-.-scseesecsecsereeseseeeeseeeetesesseeseessssssnssesscseeneasentsessavensenseneeneenes 33
Subgenus Xylocopa Latreille (SenSu StriCtO) «2... cece sesecee estes eeseesessesseenseessneseeseenseneaneasenenseneeneeneans 34
Subgenus Xylocopoides Michener .......-.:-sesecsecsecsessecseseeseeseesesseseeseeseesennessssesessssersessnssasenvenenecnecncenegnents 34
Subgenus Calloxylocopa Hurd and Moure .........s:essecsessessecsecs ess essesseesecseseeesecssensessenseesecsseeseanscnecencens 34
Subgenus Copoxyla Maa .......scccscecsecsseccsecssecsseceeseeseeeseeneeniesssecsssssecssssnecsssssnsenesesscstscsnecsscssscesecnsecsueenecs 35
Subgenus Gnathoxylocopa Hurd and Moure .........-secseeeecsecesreseesessesseeseenseesesssessessenseneanensensenensens 35
Subgenus Lestis Lepeletier and Serville, new Status ...........:ssessesececeses testes eeseersensesensensesseseeseeneeneans 35
Subgenus Notoxylocopa Hurd ......cccscceceeseseesecsecseeseseeneeesssseesssnsssssessssecnsensessassssssesseesecnenvenncessneencnas 36
Subgenus Schonmnherria Lepeletier ........-..scsecseceesesecsecses res eesessesseesssecseesecseessessesessesscsseneaneanecnceeenenenas 36
Subgenus Dasyxylocopa Hurd and Moure ..........sccecsessesses ses essessesseseeeseeseeseenecsensesscsseneeneanceneaneneensancs y/
Subgenus Xylocopsis Hurd and MOure ........sececeesecsessesseseesessessesssseeseessesenseesensensessesceneaneaneceeecneenenneats 37
Subgenus Nanoxylocopa Hurd and MOure ........:scscscsessesseseeseeneereeesecsessessesssseseesneseessasecsenseeseesenscess 37
Subgenus Cirroxylocopa Hurd and MOure ........::ccssesseseeseeesessecesessecseenssnssssssesseesenssasenseseesenesensenss 37
Subgenus Xylocopoda Hurd and MOure ........:: secs eeseesessesssssseeseeseensenssssssssessesensensesssessneeneenas 38
Subgenus Stenoxylocopa Hurd and Moure .........:cccsessessecsecssessesessessessssesseesecsesassnsssssencsssessenseessanes 38
Subgenus Neoxylocopa Michener ........-.:cccscsecsecsessessesteseeneeessssesessesssssssesseesecnenvsnseasessesensecsennerceecnaanas 38
Subgenus Prosopoxylocopa Hurd and MOure «.......:sssscsessesseseesessesseeessesseeseessaseseesesseneeneaneanccnceneneenes 39
Subgenus Zonohirsuta Maa ........ccsecseceecececseesecseceeseeseeressesecseeseessssssessecseessenssnssscsssssesecseasenanscnsencenenneneas 39
Subgenus Bomboixylocopa Maa ........+..ccecseesesserseseeeeeereeressestesessessssessesseesesssssensensesseseaneanecnenncneanaeesnas i)
Subgenus Xenoxylocopa Hurd and MOure uu... sees eseseeseeseeseesessseseeessensesseseesenseesenseneanennenneens 40
Subgenus Alloxylocopa Maa .......secssssecsecseeeessesseeseeeneeseseeseestessessssseeseesecsussssarscsesssenseanccnscsccencanecaasenstns 40
Subgenus Mesotrichia Westwood ........:.scsssecsecseseeeeterterieereneesesneenecnsaneaeensenssssssesesseensnseaueasessesennecnagnas 40
Subgenus Koptortosoma GribOdo ........ccceceeseseeseseeseseeeseeseseeseeeeseeseseesessscsserssseesssessnsensnseneesenceenceneneeney 41
JOINS) RAI ELO NR EF (SI TEE os cexcegoconcscapogoscoccdeenecchesoasaLeaboGo bo qnatiacaneeosescosAnotebsbabbacosbacocodascecoocbacondoneadactr douddcnoadosacec 42
INTEXPTBINID) CG) SSS scecetoschcecenosooesossosouaucadeddeceacore;eoc0t00005edcas0 sosaacacattonoa soso TadouCddbzodcsooseededndoGuaesoocodnHgcoHTGAAAvadERTOIO 44

ABSTRACT Phylogenetic analyses of the genera and subgenera of the tribe Xylocopini (Large Carpenter
Bees) were made using morphological characters coded for 51 species of the ingroup. These species represented
all but three of the previously recognized subgenera. Twenty species of bees constituting the outgroup also were
examined to evaluate primitive and derived character states within the Xylocopini. Four different analyses were
made because of uncertainty in hypotheses of character transformations and among the relationships of the
outgroup taxa. Cladistic parsimony analyses showed that previous classifications were based on combinations of
characters and not unique synapomorphies. Three major clades were supported in most analyses presented here;
two are from the Old World, and one is from the New World. A biogeographically mixed group of subgenera are
not clearly placed in this analysis; among this group are likely to be found the basal members of the tribe.

A new generic and subgeneric classification is proposed for the Xylocopini. I conclude that recognition of
Proxylocopa Maa and Lestis Lepeletier and Serville as genera creates a paraphyletic genus of Xylocopa; these
should be recognized as subgenera of an expanded Xylocopa. A new subgenus, Maaiana, is proposed here for X.
bentoni and related species formerly recognized as members of the subgenus Nodula. The following subgeneric
synonymies also are proposed: Xylomelissa Hurd & Moure is expanded to include Epixylocopa Hurd & Moure,
Alloxylocopa Hurd & Moure, Dinoxylocopa Hurd & Moure and Acroxylocopa Hurd & Moure; Ancylocopa Maa
is synonymized with Proxylocopa Maa; Schonnherria Lepeletier is expanded to include Joxylocopa Hurd &
Moure and Xylocospila Hurd & Moure; Stenoxylocopa Hurd & Moure is expanded to include Xylocopina Hurd
& Moure; Neoxylocopa Michener is expanded to include Megaxylocopa Hurd & Moure; Bomboixylocopa Maa
is expanded to include Mimoxylocopa Hurd & Moure; Mesotrichia Westwood is expanded to include Platynopoda
Westwood, Hoplitocopa Lieftinck and Hoploxylocopa Hurd & Moure, and; Koptortosoma Gribodo is expanded

SUBGENERA AND GENERA OF CARPENTER BEES, XYLOCOPINI 3

to include Oxyxylocopa Hurd & Moure, Cyphoxylocopa Hurd & Moure, Cyaneoderes Hurd & Moure and
Afroxylocopa Hurd & Moure. A key to identify the subgenera is provided.

Key words: Insecta; Hymenoptera; Apidae, Xylocopini; Cladistic analyses; Classification.

INTRODUCTION

The large carpenter bees, tribe Xylocopini, generally are
robust, fast-flying insects, some of which are among the
largest of all the bees. As their name implies, most mem-
bers of this group place their nests in dead solid plant
material, usually various kinds of wood, but some use the
hollow culms of bamboo (subgenus Biluna; Maeta et al.,
1985), or the pithy stems of Aloe (subgenus Gnathoxylocopa;
Watmough, 1974). Females of Proxylocopa construct their
nests in the ground. Some species of Xylocopini are im-
portant pollinators of particular plants (e.g., Corbet and
Wilmer, 1980; Gottsberger et al., 1988) but none seems to
restrict its pollen collecting to one or a few plant taxa. Other
well-studied features of the biology of the group include
its social behavior (reviewed in Gerling et al., 1988; Maeta
and Sakagami, 1995; Michener, 1990), exocrine gland mor-
phology (Minckley, 1994; Vinson, 1994) and product chem-
istry (Gerling et al, 1988), and male reproductive behavior
(Alcock, 1992; Gerling et al., 1988; Minckley, 1994).

The Xylocopini is one of four tribes in the subfamily
Xylocopinae. The other three tribes—Allodapini,
Ceratinini, and Manueliini—are mostly smaller than the
Xylocopini and build their nests in broken twigs or stems
(Daly et al., 1987; Michener, 1974, 1990).

The purpose of this study is a cladistic analysis of the
phylogenetic relationships among the genera and subgen-
era of the large carpenter bees, tribe Xylocopini. A world-
wide revision of this group was last undertaken in 1963
by Hurd and Moure. They recognized 53 supraspecific taxa
in the tribe: Three genera, Lestis, Proxylocopa, and Xylocopa;
48 subgenera in Xylocopa and two subgenera in Proxylocopa.
Although their work is the most comprehensive study of
the group to date, both with respect to the number of taxa
examined and the number of characters considered, they
did not provide an explicit hypothesis of phylogenetic re-
lationships for the members of the tribe. Da Cunha (1992)
did a quantitative similarity analysis of the subgenera of
Xylocopa using the characters of Hurd and Moure (1963).
This study was largely concerned with phenetic method-
ology and was not a taxonomic revision of the Xylocopini,
so no discussion of phylogenetic relationships or classifi-

catory changes were provided. Nevertheless, it is worth
noting that the relationships generated by Da Cunha’s
study should be used with caution, because similarity
methods can result in poor estimates of phylogeny (Hillis
et al., 1994; Michener, 1970), especially when the data used
were intended to diagnose groups and not to infer phy-
logeny.

In this study, I use an exemplar approach in the cladis-
tic analysis (see Phylogenetic Analysis). I also propose
some changes in the classification based on my phyloge-
netic analyses. My approach for taxonomic decisions is to
combine terminal taxa that are monophyletic based on the
phylogenetic analyses and retain the names of the prob-
lematic taxa as they were recognized before this study.
Future analyses may result in the recognition of fewer
supraspecific groups than are accepted here.

ACKNOWLEDGMENTS

This paper is part of a dissertation presented as partial
requirement for the degree of Ph. D. to the Department of
Entomology, The University of Kansas. The curators of the
institutions listed in the section on Selection of Taxa (be-
low) are gratefully acknowledged for the loan of the ma-
terial used in this study. Dr. Charles D. Michener provided
generous support and guidance in this project from its in-
ception. Others who were influential in the completion of
this project include Drs. B. Alexander, J.S. Ashe, William J.
Bell, E. O. Wiley, S. G. Reyes, R. A. B. Leschen, A. Roig-
Alsina, D. Yanega and M. D. Greenfield. Bruce Cutler, who
allowed me access to his laboratory and gave consider-
able guidance in the analysis of some of the internal mor-
phological structures used in the study, deserves special
commendation. I dedicate this paper to the late Byron
Alexander; he substantially improved the quality of this
study and piqued my appreciation for the importance of
systematics in evolutionary biology. This work was facili-
tated by grants from the American Museum of Natural
History (Theodore Roosevelt Fund), Department of Ento-
mology at The University of Kansas (Hungerford Fund),
and by the National Science Foundation grant DEB-
9302186 (C. D. Michener, principal investigator).

MATERIALS AND METHODS

SELECTION OF TAXA
In order to examine ingroup relationships, I used ex-
emplar taxa of all supraspecific taxa of the Xylocopini rec-
ognized by Hurd and Moure (1963) for which I could ob-
tain both sexes. This meant the following subgenera of

Xylocopa were not included in the analyses: Euxylocopa,
Lieftinckella, loxylocopa, Monoxylocopa, Diaxylocopa and
Dinoxylocopa. Euxylocopa was erected by Hurd and Moure
(1963) for a species X. (Euxylocopa) fraudulenta known only

4 SCIENTIFIC PAPERS, NATURAL History Museum, THE UNIVERSITY OF KANSAS

Table 1. List of taxa examined organized by the subgenera recognized
herein. An asterisk (*) indicates type species; (!) indicates the exemplars
used in the phylogenetic analysis. The acronyms, as explained in Materi-
als and Methods, show the sources of the specimens. Subgeneric assign-
ments by previous authors, if different than those proposed here, are

Table 1 continued

Koptortosoma (continued)

shown in parentheses.

Nanoxylocopa X. ciliata Burmeister *! (KSEM)
Cirroxylocopa X. vestita Hurd and Moure *! (USNM)
Xylocopoda X. elegans Hurd and Moure *! (KSEM)
Schonnherria X. bambusae Schrottky ! (Xylocospila) (KSEM)
X. loripes Smith (KSEM)
X. macrops Lepeletier ! (KSEM)
X. micans Lepeletier ! (KSEM)
X. viridigastra Lepeletier ! (KSEM)
X. viridis Smith (KSEM)
X. chrysopoda Schrottky ! (Ioxylocopa) (USNM)
Dasyxylocopa X. bimaculata Friese *! (KSEM)
Xylocopsis X. funesta Maid] *! (KSEM)
Monoxylocopa X. abbreviata Hurd and Moure * (USNM)
Diaxylocopa X. truxali Hurd and Moure * (LACM)
Notoxylocopa X. guatemalensis Cockerell ! (KSEM)
X. tabaniformis androleuca Michener * (KSEM)
X. t. azteca Cresson (KSEM)
X. t. tabaniformis Smith ! (KSEM)
Xylocopoides X. californica arizonensis Cresson ! (KSEM)
X. cyanea Smith ! (KSEM)
X. pallidiscopa Hurd (KSEM)
X. virginica Linnaeus * (KSEM)
Calloxylocopa X. tenuata Smith *! (KSEM)
Neoxylocopa X. aeneipennis (Degeer) (KSEM)
X. b. brasilianorum (Linnaeus) * (KSEM)
X. darwini Cockerell (KSEM)
X. fimbriata Fabricius ! (Megaxylocopa) (KSEM)
X. f. frontalis (Olivier) (Megaxylocopa) (KSEM)
X. grisescens Lepeletier (KSEM)
X. mexicanorum Cockerell (KSEM)
X. varipuncta Patton ! (KSEM)
Stenoxylocopa X. Paes Smith *! (KSEM)
X.m. micheneri Hurd (KSEM)
X. ruficollis Hurd and Moure! (Xylocopina) (FMNH)
Ctenoxylocopa X. fenestrata (Fabricius) *! (KSEM)
X. sulcatipes Maa (KSEM)
Copoxyla X. iris (Christ) *! (KSEM)
Rhysoxylocopa X. cantabrita Lepeletier *! (KSEM)
X. uclesiensis Pérez (USNM)
Gnathoxylocopa _X. kobrowi Brauns (USNM)
X. sicheli Vachal *! (USNM)
Xylocopa X. valga Gerstaecker ! (KSEM)
X. violacea (Christ) *! (KSEM)
Biluna X. auripennis Lepeletier ! (KSEM)
X. tranquebarorum (Swederus) (KSEM)
Bomboixylocopa _X. chinenis Friese ! (USNM)
X. rufipes Smith ! (Mimoxylocopa) (KSEM)
Nyctomellita X. myops Ritsema (KSEM)
X. tranquebarica (Fabricius) *! (KSEM)
Nodula X. a. amethystina (Fabricius) (KSEM)
X. amethystina (Fabricius) *! (KSEM)
X. bentoni Cockerell ! (KSEM)
Zonohirsuta X. dejeanii Lepeletier ! (KSEM)
Prosopoxylocopa X. mirabilis Hurd and Moure *! (AMNH)
Xylomelissa X. absurdipes Enderlein ! (Dinoxylocopa) (USNM)
X. albifrons Lepeletier ! (USNM)
X. carinata Smith * (USNM)
X. capitata Smith ! (Acroxylocopa) (KSEM)
X. erythrina Gribodo ! (Perixylocopa) (PPI)
X. lugubris Gerstaecker ! (Apoxylocopa) (PPI)
X. rufitarsis Lepeletier ! (Epixylocopa) (KSEM)
Alloxylocopa X. appendiculata circumvolans Smith *! (KSEM)
Xenoxylocopa X. chiyakensis (Cockerell) *! (KSEM)
X. inconstans Smith (KSEM)
Mesotrichia X. acutipennis Smith! (Hoploxylocopa) (BM)
X. assimilis Ritsema ! (Hoplitocopa) (RNH)
X. flavorufa (DeGeer) (KSEM)
X. latipes (Drury)! (Platynopoda) (KSEM)
X. tenuiscapa Westwood (Platynopoda) (KSEM)
X. torrida (Westwood) *! (KSEM)
Koptortosoma X. aestuans (Linnaeus) (KSEM)

X. apicalis Smith (KSEM)
X. aruana aruana Ritsema (KSEM)
X. caerulea (Fabricius) ! (Cyaneoderes) (KSEM)
X. caffra (Linnaeus) ! (KSEM)
X. kuhni Friese ! (Cyphoxylocopa) (RNH)
X. nigrita (Fabricius) *! (Afroxylocopa) (KSEM)
X. ocularis Pérez! (Cyphoxylocopa) (RNH)
X. parva Enderlein ! (Oxyxylocopa) (BM)
X. pubescens Spinola! (TAV)
X. scioensis Gribodo (KSEM)
X. senior Vachal (KSEM)
X. sinensis Smith ! (KSEM)
X. varipes Smith ! (Oxyxylocopa) (BM)
Lestis X. aeratus Smith ! (KSEM)
X. bombylans (Fabricius) (KSEM)
Proxylocopa X. oliviert (Lepeletier) * (TAV)
X. rufa (Friese) ! (TAV)
X. altaica (Popov) (Ancylocopa) (ZIA)
X. andarabana (Hedicke) (Ancylocopa) (ZIA)
X. nitidiventris (Smith) (Ancylocopa) (ZIA)
X. nix (Maa) (Ancylocopa) (ZIA)
X. parviceps (Morawitz) (Ancylocopa) (ZIA)
X. pavlovskyi (Popov) CGneseata (ZIA)
X. przewalskyi (Morawitz) (Ancylocopa) (ZIA)

from females and later synonymized by Maa (1970) under
X. (Perixylocopa) erythrina. I was not able to examine a male
of Lieftinckella. The remaining four monotypic subgenera
were examined, but are known from only one sex—
Ioxylocopa and Dinoxylocopa from males and Monoxylocopa
and Diaxylocopa from females. For some large or seemingly
variable subgenera, additional species were included.
Character analysis was restricted to these exemplars. The
two subgenera of Proxylocopa differ only in one character,
the size of the ocelli, so I included only Proxylocopa rufa
after initial examinations. All the taxa examined and the
respective repositories of specimens that I used are listed
in Table 1. Because the taxonomic categories recognized
by Hurd and Moure (1963) were narrowly defined, these
exemplars are likely to provide a good representation of
the morphological diversity and a reasonable first estimate
of the phylogenetic relationships in the group.

Specimens used in this study were from the following
institutions (acronyms in parentheses): Field Museum of
Natural History, Chicago, A. Newton (FMNH); Los Ange-
les County Museum, Los Angeles, R. Snelling (LACM);
National Collection of Insects, Plant Protection Research
Institute, South Africa, C. Eardley (PPI); The National
Museum of Natural History, Washington, R. McGinley
(USNM); The Natural History Museum, London, G. Else
(BM); Rijksmuseum van Natuurlijke Historie, Leiden, C.
van Achterberg (RNH); Snow Entomological Museum,
University of Kansas, Lawrence, C. D. Michener (KSEM);
Tel-Aviv University, Tel-Aviv, D. Gerling (TAU); Zoologi-
cal Institute, Academy of Sciences, Leningrad, Y. Pesenko
(ZIA).

Phylogenetic relationships of the large carpenter bees
with other tribes of Xylocopinae were analyzed by

SUBGENERA AND GENERA OF CARPENTER BEES, XYLOCOPINI 5

TABLE 2. List of outgroup taxa examined.

Apidae: Xylocopinae
Ceratinini
Ceratina (Ceratina) cucurbitina (Rossi)
C. (Euceratina) cyanea Kirby
C. (Lioceratina) flavolateralis Cockerell
C. (Certinidia) japonica Cockerell
C. (Calloceratina) laeta Spinola
C. (Crewella) rupestris
C. (Pithitis) smaragdula (Fabricius)
C. (Ceratinula) zeteki Cockerell
Ctenoceratina tanganyicencis (Strand)
Megaceratina sculpturata (Smith)
Allodapini
Compsomelissa zaxantha (Cockerell)
Exonueridia hakariensis
Exonuera bicolor Smith
Halterapis nigrinervis (Cameron)
Manueliini
Manuelia gayatina (Spinola)
M. gayi (Spinola)
M. postica (Spinola)
Megachilidae
Pararhophitini
Pararhophites orobinus Morawitz
Fideliini
Parafidelia friesei Brauns
Lithurgiini
Lithurge apicalis Cresson

Sakagami and Michener (1987) and Roig-Alsina and
Michener (1993). Both studies supported Ceratinini and
Allodapini as sister groups and the terminal members of
the subfamily. The sister group to the Ceratinini +
Allodapini clade is less certain; Sakagami and Michener
(1987) concluded that this group was the Xylocopini, but
the analysis of Roig-Alsina and Michener (1993), which
used more characters, concluded that it could be either
Xylocopini or Manueliini. Roig-Alsina and Michener (1993)
favored the placement of the Manueliini as the most basal
taxon in the subfamily because Xylocopini, Allodapini and
Ceratinini share a derived shape of sternum eight in the
males. They also addressed relationships of groups out-
side the subfamily Xylocopinae. The next more basal taxon
to the Xylocopinae was the family Megachilidae (consist-
ing of the subfamilies Megachilinae, Fideliinae and
Osmiinae).

I chose components of the outgroup for this study based
largely on the analysis of Roig-Alsina and Michener (1993).
For three of the analyses I polarized ground-plan states of
characters based on a topology of Manueliini as the basal
taxon in the subfamily Xylocopinae and a sister-group re-
lationship among Xylocopini and Allodapini + Cera‘inini.
From the family Megachilidae, I used three taxa (Lithurge
apicalis, Parafidelia friesei and Pararhophites orobinus) con-
sidered basal in the family as additional outgroups. Names
of all outgroup taxa are listed in Table 2. Unless otherwise

stated, I use the term outgroup to refer to a hypothetical
ancestor that represents the inferred ground-plan state of
all characters. Ground-plan character states for the
Xylocopini were hypothesized using the optimization tech-
nique of Maddison et al. (1984). This hypothetical ances-
tor has all characters coded as zero, with a few exceptions
(Table 3).

SELECTION OF CHARACTERS

[used all characters that fulfilled the conditions of ho-
mology (Remane, 1956), including some characters of the
males that are homoplasious in other bees. A priori, | could
think of no reasonable criterion that would allow me to
decide whether one character or character system (e.g.,
male genitalia vs. external sexual characteristics of the
male) was phylogenetically more informative than any
other. This approach relies on the congruence with other
characters to determine the general “goodness of fit” of
any single character (Kluge, 1989). Furthermore, in a large
group, such as the Xylocopini, it seems reasonable to ex-
pect that some characters homoplasious within some
clades will serve as synapomorphies of other clades.

Homology assessment was problematic for some char-
acters in Proxylocopa, females of which have a well-devel-
oped basitibial plate at the base of the tibia and a large
conspicuous pygidial plate. These conditions are typical
for most other ground-nesting bees, but in the rest of the
large carpenter bees the pygidial plate consists of only one
or several spines, and the basitibial plate is reduced (some-
times only represented by its apex) and positioned on the
apical two-thirds of the tibia. The placement and develop-
ment of these structures in Proxylocopa have led some work-
ers to consider it as the most primitive taxon in the sub-
family (Xylocopinae sensu Hurd, 1958; Malyshev, 1913).
Others have argued that these structures collectively rep-
resent an adaptive complex associated with nesting in the
ground and have evolved secondarily (Daly et al., 1987).
Morphological evidence supports the latter view; in
Proxylocopa the pygidial plate and basitibial plate are swol-
len lobes and completely lack a line or carina around the
lateral edges as is typical of other bees. Furthermore, if the
Manueliini, a wood-nesting group, is the basal taxon in
the subfamily Xylocopinae, the inferred ground-plan state
for the Xylocopini would be that its ancestor nested in
wood and Proxylocopa would have secondarily reverted to
nesting in the ground. In this study, I treated the above
conditions in Proxylocopa as unique derived features. How-
ever, because no comparable structures exist elsewhere in
the ingroup (Xylocopini) and some characters are trans-
formation series with multiple states, I coded all charac-
ters associated with the basitibial plate and pygidial plate
as missing in Proxylocopa. Preliminary analyses (not shown)
made with these features coded as primitive usually pro-

ScrenTiFIC PAPERS, NATURAL History Museum, THE UNIVERSITY OF KANSAS

Table 3. Data matrix used in the phylogenetic analyses for Analysis 1, 1A, 3, and 3A. Matrices for Analy-
ses 2, 2A, 4 and 4A are explained in the text. Numbers across the top of matrix correspond to character
numbers used in the text. Taxon abbreviations used were Xylocopoides' = X. californica arizonensis,
Xylocopoides*= X. cyanea, Xylocopa' = X. violacea, Xylocopa*= X. valga, Koptortosoma’ = X. sinensis, Koptortosoma*
=X. caffra, Koptortosoma*= X. pubescens, Schonnherria' = X. viridagastra, Schonnherria? = X. micans, Schonnherria*
= X. macrops, Nodula' = X. bentoni, Nodula*= X. amethystina, Notoxylocopa' = X. tabaniformis and Notoxylocopa*
= X. guatemalensis.

Character al 2 3 4 5

Taxon 123456789 0123456789 0123456789 0123456789 0123456789 0123456
Outgroup 000700000 0000000000 00?7?7000000 0000000000 0000000070 0000000
Nanoxylocopa 001000000 0100000001 0100102000 0010002000 0101000111 1020000
Cirroxylocopa 001000000 0100000001 010010!100 0010000000 0101000110 1120000
Xylocopoda' 001001000 0100000001 010010!100 0000000000 0101000100 1020000
Dasyxylocopa 011000000 0100000001 010010!000 0000002000 0101000112 1120100
Xylocopsis 001000000 0100000011 0100102000 0000002000 0101000111 1120100
Xylocopoides' 000000000 1210000101 1100004000 0100002000 0210100100 1001000
Calloxylocopa 000000000 0100000001 0100000000 0000002000 0210100100 1000000
Neoxylocopa 101001000 0110000001 0100103100 0020000011 1101000100 1120000
Megaxylocopa 101001000 0110000001 0100103100 0020000011 1101000100 1120000
Stenoxylocopa 001111000 0110000001 0100100100 0020002010 1101000100 1120000
Xylocopina 001111000 0110000001 010010!000 0000000000 1101000100 1120000
Ctenoxylocopa 000100000 1100000111 0000002000 0120002000 1100000100 0000110
Copoxyla 000100100 1100000001 0100000000 0000002000 1000000111 0000100
Rhysoxylocopa 000000000 1101000201 0100000000 0020000000 0100000110 0020000
Gnathoxylocopa 001100000 1100000111 0100100000 0010002000 1100000111 1010100
Xylocopa 000100000 1210000101 0000001000 0000002000 0110000100 0001010
Nyctomelitta 011000001 0101000111 0100102000 0000000000 0101000111 1010000
Zonohirsuta 000000000 0210002001 0111002000 0000000000 0100000000 1020000
Prosopoxylocopa 000000000 0210000000 011100!000 0000000000 0100000101 1110000
Bomboixylocopa 000000010 0210000000 0100001000 0000010001 1000000000 0021000
Mimoxylocopa 011000010 0210000001 0100115000 0000012001 1100000000 1121000
Alloxylocopa 000001000 0210000000 0111110000 0221111000 0100000000 1021000
Xenoxylocopa 100001100 0210000001 010111!010 1020011100 0000000000 1121000
Mesotrichia 000001100 0210102000 0122115011 1021211110 0110000100 1120010
Platynopoda 000000100 0210100000 0122001001 1221211010 0000010100 0011010
Hoplitocopa 011000100 1210100000 012210!001 1221211010 0000010100 1120110
Hoploxylocopa 000001100 0210112000 012211!011 1121211000 0000000000 0020010
Afroxylocopa 110001100 0210000001 0123115100 1021211100 1000000102 1121100
Oxyxylocopa 000001100 0210010000 012311!010 1221111000 0000000101 1121000
Cyaneoderes 001001100 1210000000 012310!010 1221211100 0000000101 1121000
Cyphoxylocopa 001000000 0210002000 012311!010 1221111100 0000000101 1120000
Koptortosoma’ 101000000 0210000000 010211!010 1121211100 1000000102 1021000
Koptortosoma? 101000000 0210100000 0123111010 1121211101 1000000101 1121000
Koptortosoma* 111000000 1210000000 0123115010 1101211101 1000000102 1021000
Proxylocopa 000000001 0200000!!! !!00111000 0000000000 0!:00000110 0000000
Lestis 000000000 0101000001 1100000000 0000002000 0100000111 1000011
Schonnherria' 001000000 0100001011 0100102000 0220000000 0101000112 1120100
Schonnherria? 011000100 0100000011 0100102000 0220012000 0101000112 1120000
Schonnherria® 001100100 0100000011 0101102000 0220002000 0101000112 1120000
Xylocospila 001000000 0100000001 010010!000 0020002000 0101000112 1120000
Epixylocopa 001000000 1000000201 0100012000 0020002001,1100000111 0020000
Perixylocopa 011001000 0000001201 0100102000 0000002001 1100000111 0020000
Acroxylocopa 000101010 0000001201 0100112000 0020000001 1100000111 0020000
Biluna 000000000 0000001000 0100112100 0020000000 1100000110 0020000
Apoxylocopa 000000000 0000001201 0100102000 0020000001 1100000111 0020000
Xylomelissa 000000000 0001000201 0100102000 0020000001 1100000111 0020000
Nodula’ 001000000 0101000201 0100111000 0020000001 1100000110 0020000
Nodula* 000000000 0110000201 0111000000 0020000001 1100000110 0020000
Notoxylocopa’ 001000000 0110000001 0100110000 0000002001 0101001112 1010100
Notoxylocopa? 001100000 0210000101 0100110000 0000002000 0101001112 1010100
Xylocopa* 001100000 1210000101 0000001000 0000012000 0100000100 0001010
Xylocopoides* 001000010 1210000101 1100004000 0100002000 0210100100 1001001

SUBGENERA AND GENERA OF CARPENTER BEES, XYLOCOPINI 7

duced trees with Proxylocopa arising as the basal taxon in
the tribe.

Terminology generally follows that of Michener (1944,
1965) and Hurd and Moure (1963), except that I use me-
dial (as opposed to lateral) instead of mediolongitudinal
for structures on the tibial spine, propodeum, terga and
sterna. Terga and sterna are designated as T and S, respec-
tively, and segments are numbered from anterior to poste-
rior, e.g., Sl, $2, etc. I follow Brothers (1976) in the use of
metapostnotum instead of propodeal triangle, and
Michener (1975) and Reyes (1991) in the use of ventroapical
plate for the ventral distal projection of the male
gonocoxite. Mandibular structures are described using the
terminology of Michener and Fraser (1978).

Examination of the reservoirs of the mesosomal glands
of the males required dissection of the posterior part of
the thorax using techniques described by Minckley (1994);
this was limited to specimens for which I had permission
to dissect. For ready comparison, the characters are given
the same number in the cladistic analysis and the group
descriptions. The cladograms were produced using the
computer program PAUP (Phylogenetic Analysis Using Par-
simony) Version 3.1.1 (Swofford, 1993) and analyzed with
MacClade version 3.05 (Maddison and Maddison, 1993).

CHARACTERS AND CODING
The characters, as numbered below, correspond to the
numbers in all the subsequent tables and trees and in the
group descriptions. Numbers in parentheses are the codes
used in the data matrix and on the cladograms. Unless
noted otherwise, 0 is considered the plesiomorphic state.

1. Color of male vestiture.—(0) dark brown, black,
green, or pale; (1) tawny or ferruginous on at least dorsum
of body. State 1 is not found in the outgroup.

2. Subapical tooth or convexity on upper carina of
male mandible.—(0) Absent (Fig. 1B); (1) present (Fig. 1A,
1C). State 1 is widespread in bees and occurs in the genus
Ctenoceratina (Ceratinini) of the outgroup. Because
Ctenoceratina has a number of derived features (Daly, 1988),
it is not likely to be an ancestral group in the Ceratinini
and therefore not closely related to the Xylocopini; I infer
State 1 to be apomorphic for the Xylocopini.

3. Subapical tooth on upper carina of female man-
dible.—(0) Absent; (1) present. See preceding character.

4. Number of teeth on apical margin of female man-
dible—(0) Two (Fig. 1A); (1) three (Fig. 1B). Mandibles as
shown in Figure 1C are coded as having three teeth be-
cause the two upper teeth are considered to be fused. In
the outgroup, the Ceratinini, Allodapini and Lithurge
(Michener and Fraser, 1978) have three teeth on the apical
margin of the mandible, and Manuelia, Pararhophites and
Parafidelia have two teeth. Based on outgroup analysis the

Cc

Fig. 1. External views of left mandibles of females: A. Xylocopa
(Megaxylocopa)* fimbriata. B. X. (Gnathoxylocopa) sicheli. C. X.
(Stenoxycolopa) micheneri micheneri. Abbreviations: a= acetabular groove,
sa = subapical tooth. Scale = 1 mm. Fig. B redrawn from Hurd and Moure
(1964). *Regarded as a member of the subgenus of Neoxylocopa herein.

polarity decision is equivocal. Furthermore, in those
outgroup taxa with three mandibular teeth (Ceratinini,
Allodapini and Parafidelia), the medial tooth is longest, but
in the Xylocopini the ventral tooth (ventral tooth of the
rutellum) is longest, possibly suggesting that the additional
tooth of some Xylocopini has arisen de novo. I consider three
teeth homologous in this analysis and coded the ground-
plan state as equivocal.

5. Upper rutellar tooth and preapical tooth (i.e. apex
of pollex).—(0) Separated by an emargination (Fig. 1B);
(1) fused and forming a continuous edge (Fig. 1C; Hurd
1978, Figs. 1, 2). State 1 is not found in the outgroup. This
apomorphy is shared by Stenoxylocopa and Xylocopina and
is the same as the “inner tooth” of Hurd (1978). Two-
toothed mandibles (Fig. 1A) were coded 0.

6. Epistomal suture of female.—(0) Not raised or
broadly impunctate (Hurd and Moure, 1963, Fig. 16); (1)
continuously carinate and broadly impunctate (Hurd and
Moure, 1963, Figs. 14, 15, 17, 18). In some ingroup taxa, a
carina is present on the upper epistomal suture but it does
not extend completely down the lateral edges, a condition
I coded as 0. State 1 does not occur in the outgroup.

7. Supraocellar pits of female——(0) Punctiform or not
present; (1) deeply excavated and crater-like. In the
outgroup, supraocellar pits, if present, are punctiform.

8. Malar space of female.—(0) Shorter than width of
flagellar segment 1 (Hurd and Moure, 1963, Fig. 51); (1)
much longer than width of flagellar segment 1 (Hurd and
Moure, 1963, Fig. 16). State 1 is not found in the outgroup
exemplars used here, but it is present among other species

8 ScIENTIFIC PAPERS, NATURAL History Museum, THE UNIVERSITY OF KANSAS

\
A B Cc

Fig. 2.

. \
dl =—=its \
D E

Outer view of hind tibiae of females: A. Xylocopa (Xylocopoides) cyanea. B. X. (Ctenoxylocopa) fenestrata. C.

Lestis* bombylans. D. Proxylocopa*rufa. E. X. (Alloxylocopa) circumvolans. Abbreviation: ts = tibial spines. Scale = 1 mm.

*Regarded as a subgenus of Xylocopa herein.

in the outgroup that I did not consider (Michener, 1975;
Reyes, 1993).

9. Ocelli.—(0) Not modified; (1) enlarged. State 1 oc-
curs in the subgenera Nyctomelitta and Proxylocopa (sensu
stricto) of Xylocopa. All outgroup taxa have normal-sized
ocelli.

10. Male facial maculations.—(0) Yellow on at least
some part of face; (1) without yellow markings, i.e., en-
tirely black. This is a highly variable character in the
Apoidea but in the outgroup, males of all taxa except
Lithurge, have yellow maculations on the clypeus.

11. Number of spines on outer apex of male hind
tibia.—(0) None; (1) one; (2) two (Hurd and Moure, 1963;
Fig. 37). Outgroup taxa lack these spines. The transforma-
tion series was coded as additive.

12. Number of spines on outer apex of female hind
tibia.—(0) One (Fig. 2B, 2C, 2D, 2E); (1) two (Fig. 2A; Hurd
and Moure, 1963, Figs. 36, 38). Females of the outgroup
have one or two spines.

13. Tibial spine of female.—(0) Originating at tibial
apex and lacking medial carina (Fig. 2A, 2D, 2E); (1) well-
developed and originating above, but extending to, apex
of tibia (Fig. 2B, 2C). State 1 is found only in taxa with one
outer tibial spine. State 1 is not found in the outgroup.

14. Shape of female hind trochanter—(0) Not modi-
fied (Fig. 3B, 3C); (1) roughly triangular in ventral view
(Fig. 3A). State 1 is not found in the outgroup.

15. Mid and hind trochanter of male.—(0) Spines ab-
sent; (1) ventral surface of each with spine. State 1 does
not occur in the outgroup.

16. Ventral surface of male hind femur. —(0) Not modi-
fied; (1) with basal tubercle (Fig. 3B); (2) with spine on basal
half (Fig. 3C). In the outgroup, modifications such as these
are known only in some Ceratina (Hirashima, 1969). This
character was coded as nonadditive.

17. Anterior and posterior edges of female basitibial
plates.—(0) Weakly developed or rising little from surface
of leg; (1) with numerous tubercles on both edges (Fig. 2A,
2B); (2) with numerous tubercles only on posterior edge.
The polarity proposed here may be wrong , because those
ingroup taxa with the most strongly developed basitibial
plates (a plesiomorphic condition?) have tubercles along
both edges (e.g. Ctenoxylocopa, Xylocopa sensu stricto).
Those taxa with a single row of tubercles (State 2) on the
posterior edge completely lack a carina, or tubercles, on
the anterior edge. State 1 is unknown in the outgroup. This
transformation series was coded as additive.

18. Apex of male basitibial plate-—(0) Entire or absent;
(1) bifid. As discussed in Character 17, the polarity of this
character may be the reverse of that presented here. In all
ingroup taxa with well-developed basitibial plates, the
apex of the structure is bifid and usually strongly so. In
the outgroup the basitibial plate, when present, invariably
has an entire apex. The character was coded with an entire
apex as the plesiomorphic condition.

SUBGENERA AND GENERA OF CARPENTER BEES, XYLOCOPINI 9

Tr

Tr

MS
Cc s

Fig. 3. A. Ventral view of the hind trochanter of female Xylocopa
(Mesotrichia) latipes. B. Anterior view of the femur of the male of X.
(Ctenoxylocopa) hottentota. C. Anterior view of the femur of the male of X.
(Afroxylocopa)* nigrita. Abbreviations: F = femur, t = tubercle, Tr = tro-
chanter, s = spine. Modified from Eardley, 1983. Scale = 1 mm. *Regarded
as a member of the subgenus Koptortosoma herein.

19. Apex of female basitibial plate—(0) Entire (Fig.
2D, 2E); (1) bifid (Fig. 2A, 2B, 2C). See preceding character.
20. Female basitibial plate-——(0) Open basally or, if

closed, extending onto basal fourth of tibia; (1) foveate and
on apical 3/4 of tibia (Fig. 2A, 2C; Hurd, 1961, Fig. 9). State

1 is unique to Xylocopoides and Lestis . Other taxa in Xylocopa
have lateral edges that converge basally, but do not com-
pletely close (e.g. X. [Xylomelissa] erythrina, X.
[Notoxylocopa] guatemalensis), a condition that could be
considered as a weak fovea, but I have not considered these
to be homologous to State 1.

21. Female basitibial plate-—(0) Extending to basal
fourth of tibia (Fig. 2B); (1) reduced, not extending to base
of tibia (Fig 2A, 2C, 2E). See discussion of Character 17.

22. Beginning of posterior thoracic declivity of male.—
(0) Smoothly rounded, beginning undefined (Hurd and
Moure, 1963, Figs. 24, 27; Minckley, 1994, Figs. 1A, 1@reAy
3C, 4A, 4C, 4E); (1) abruptly rounded (Hurd and Moure,
1963, Figs. 39; Minckley, 1994, Fig. 3E); (2) sharply angled
(Hurd and Moure, 1963, Fig. 33; Minckley, 1994, Fig. 5A).
See discussion in following character.

23. Beginning of posterior thoracic declivity of fe-
male.—(0) Profile smoothly rounded, position often un-
defined (Hurd and Moure, 1963, Figs. 27, 28; Minckley,
1994, Figs. 1A, 1C, 3A, 3C, 4A, 4C, 4E, 5C, 5E); (1) abruptly
rounded (Hurd and Moure, 1963, Fig. 39; Minckley, 1994,
Fig. 3E); (2) as a transverse carina that does not overhang
metanotum (Hurd and Moure, 1963, Fig. 33; Minckley,
1994, Fig. 1E); (3) marked by transverse carina that over-
hangs entire metanotum (Hurd and Moure, 1963, Fig. 34;
Minckley, 1994, Fig. 5A). Because states 0 and 1 both occur
in the outgroup, I consider the ground-plan state to be
equivocal. States 2 and 3 are unknown in the outgroup
and are strong apomorphies within Xylocopa. This trans-
formation series was treated as additive.

24. Male metapostnotum. —(0) Present; (1) absent. The
metapostnotum in bees is highly variable in size and shape
and, when present, is distinguished from the propodeum
by a change in the sculpturing of the cuticle and/or by a
line along its edge with the propodeum. Apparently the
metapostnotum can be lost in two ways: (1) absence of
unique sculpturing and marginal lines; or (2) by an increase
in the size of the propodeum relative to the metapostnotum
until the latter is completely absent externally. In the
Xylocopini the metapostnotum invariably is delimited by
a line along the edge with the propodeum and a continu-
ous size-gradation across taxa is evident. The
metapostnotum in some taxa is well developed and in
other taxa is vestigial or absent. Furthermore, in species
the males of which have an enlarged posterior mesosoma
that accommodates the mesosomal gland reservoir, the
metapostnotum typically is absent (Minckley, 1994). These
observations suggest that loss of the metapostnotum in
the Xylocopini is by a change in the size of the propodeum.
In the Ceratinini, the presence of the metapostnotum is
usually made evident by a change in sculpturing, and no
gradation in size of the propodeum is apparent; the

10 ScIENTIFIC PAPERS, NATURAL History Museum, THE UNIVERSITY OF KANSAS

”

Fig. 4. Left tegula of the male of (A) Xylocopa varipuncta and (B) X.
(Platynopoda)* tenuiscapa. Abbreviations: S = scutum, Tg = tegula. Scale =
1 mm. *Regarded as a member of the subgenus Mesotrichia herein.

metapostnotum is either well developed or completely
absent. I infer from this pattern that in the Ceratinini the
metapostnotum is lost by a reduction in its surface differ-
entiation from the propodeum and not by a change in size
of the propodeum or the metapostnotum. Therefore, I con-
sider the presence of a defined metapostnotum to be the
ground-plan condition in the Xylocopini.

25. Female metapostnotum. —(0) Present, including
those where it is minute; (1) absent. See discussion in pre-
ceding character.

26. Reservoir of the male mesosomal gland.—(0) Ab-
sent; (1) two widely separated pouches (Minckley, 1994,
Figs. 1C, 1D, 1E, 1E, 3A, 3B); (2) single layer of tubules,
usually abutting in the middle (Minckley, 1994, Figs. 3C,
3D, 4A, 4B, 4C, 4D); (3) highly convoluted tubules, as in
Neoxylocopa (Minckley, 1994, Fig. 5E, 5F); (4) single medial
pouch with one opening (Minckley, 1994, Fig. 1A, 1B); (5)
numerous, highly convoluted tubules, as in Koptortosoma
(Minckley, 1994, Fig. 5A, 5B). The presence of a reservoir is
a unique character of most Xylocopini. The reservoir in at
least some species is associated with a gland that produces
a sex-attractant pheromone, although, in the vast majority

of large carpenter bees, the function for this structure has
not been examined. There is tremendous variation in such
features as the shape of the tubules that make up the gland
reservoir and the size of the reservoir (Minckley, 1994),
which makes homology assessment difficult for some taxa.
In these difficult taxa I have opted to preserve informa-
tion and have split the characters into a number of often
autapomorphic states rather than simply scoring the char-
acter into states of presence or absence. This character was
coded as nonadditive.

27. Beginning of the thoracic declivity of male.—(0)
Anterior to the propodeum (Hurd and Moure, 1963, Figs.
29-34; Minckley, 1994, Fig. 1A, 1C); (1) on the propodeum
(Hurd and Moure, 1963, Fig. 27; Minckley, 1994, Fig. 5C,
5E). State 1 is not found in the outgroup.

28. Propodeum at its connection with metasoma.—(0)
Not modified; (1) dorsally elongated, forming a spine. This
apomorphy is unknown in the outgroup.

29. Male tegula.—(0) Not modified (Fig. 4A); (1) elon-
gate posteriorly (Fig. 4B). State 1 is not found in the
outgroup.

30. Shape of the gradulus on T1.—(0) Gradulus of Tl con-
tinued posteriorly adjacent to lateral margin of metasoma
(Hurd and Moure, 1963, Fig. 43); (1) transverse (Hurd and
Moure, 1963, Fig. 44). State 1 is not found in the outgroup.

31. Distribution of graduli on male terga.—(0) present
on Tl, T2 and T3, or more; (1) present on T1 and T2; (2)
present on T1 only. All members of the outgroup have
graduli on at least T1—-T3, except for the ceratinine genus
Pithitis, which has a gradulus only on T1 (Hirashima, 1969).
Based on this information and the argument that a
gradulus on each tergum is generally considered to be
plesiomorphic in bees (Michener, 1944), the inferred
ground-plan state for the ingroup is graduli on all terga
posterior to T3, as well as on more anterior terga. The
character was treated as additive.

32. Distribution of graduli on female terga.—(0) T],
T2 and T3, or more; (1) T1 and T2; (2) T1 only. See discus-
sion of preceding character.

33. Modifications of anterior surface of T1 of males.—
(0) Absent; (1) a fovea (Hurd and Moure, 1963, Fig. 42).
See discussion in following character.

34. Modifications of anterior surface of T1 of female.—
(0) Absent; (1) a fovea (Hurd and Moure, 1963, Fig. 42); (2)
with an entrance to a acarinarium (Fig. 5; Eardley, 1983,
Figs. 23, 24, 27). Neither a fovea nor an invaginated mite
pouch occurs in other bees. The transformation series was
coded as additive.

35. Medial groove of T1 of female.—(0) Linear; (1)
deeply sulcate (Hurd and Moure, 1963, Fig. 41). In the
outgroup, the medial groove is linear.

SUBGENERA AND GENERA OF CARPENTER BEES, X YLOCOPINI 11

$1

Fig. 5. Anterior face of Tl of a female Xylocopa (Platynopoda)*
tenuiscapa. Abbreviations: a = entrance to the acarinarium; g = gradulus;
Tl = dorsal surface of tergum 1; $1 = sternum 1; sp = spiracle. *Regarded
as a member of the subgenus Mesotrichia herein.

36. Apical margin of SI.—(0) Entire, not produced; (1)
produced (Hurd and Moure, 1963, Fig. 44); (2) medially
emarginate (Hurd and Moure, 1964, Fig. 43). State 1 occurs
in all outgroup taxa. This character was coded as nonadditive.

37. Medial carina on S6 of male.—(0) Absent or not
developed into a keel; (1) strongly developed, forming a
keel. In Ceratina, a member of the outgroup, a carina is
present on S6 but is never developed into a strong keel.

38. Medial carina on metasomal sternum of female.—
(0) Carina not present on all sterna or not continuous; (1)
continuous and present on all sterna. A well-defined me-
dial carina on all female sterna clearly is apomorphic for
the ingroup, although in some outgroup taxa (Ceratina laeta,
C. rupestris and Megaceratina sculpturata) a weakly devel-
oped medial carina occurs on S6 and in many members of
the ingroup, a carina occurs on all sterna, although it is
weakly developed or not continuous.

39. Pregradular areas on sterna of male.—(0) Paired
setal patches absent; (1) paired setal patches present on at
least one sternum (Fig. 6A). In the ingroup, these most of-
ten are on $2-S6, but in many taxa they occur on fewer
sterna. State 1 is not found in the outgroup.

40. Distinct patch of appressed short setae anterior and
dorsal to one or more metasomal spiracles of male.—(0)
Absent; (1) present (Fig. 6B). In most taxa these are found
on T2—T5, but in some taxa they are on fewer terga. State
1 is not found in the outgroup.

41. Pygidial spine of T6 of female.—(0) Not armed with
lateral preapical spines (Fig. 7A); (1) with pair of preapical
lateral spines (Fig. 7B); (2) with many preapical lateral
spines (Fig. 7C). The pygidial plate of all wood-nesting
Xylocopa (excluding Proxylocopa), and Manuelia, is repre-
sented by a raised spine that originates before, and extends
beyond, the distal edge of T6. At its base, the spine may be

ita
TS

q aan yi

0 he
“Hl et wh ‘a

NEN

sith A My te

ve

g

Fig. 6.
of S4 showing patches of appressed setae. B. Side view of T4 with ap-
pressed setal patch; arrow indicates anterior; abbreviations: sp = spiracle,
g = gradulus. Scale = 1 mm.

Xylocopa (Neoxylocopa) varipuncta male: A. Pregradular area

flanked by one or many preapical spines (prepygidial
spines of Hurd and Moure, 1963). When there are many
such spines they delimit a triangular area that may be a
remnant of a pygidial plate. In Ceratinini and Allodapini,
a developed spine is absent although the distal edge of T6
is medially elongated, forming a point. This character was
coded as additive.

42. Gonostylus of male genitalia—(0) A slender pro-
jection fused to apex of gonocoxite (Figs. 8A, 8B, 8C, 9A,
9B, 9C, 9D, YE, YF); (1) a rounded or conical setose area at
apex of gonocoxite (Hurd and Moure, 1963, Figs. 242, 244).
Lestis (Fig. 9B) was coded as having State 0 because its coni-
cal gonostylus is different from that coded as 1. This deci-
sion is based on the two features not found in other taxa
with such gonostyli, a dense brush of medially oriented
hairs on the gonostylus and a strong carina on the apex of
the gonostylus. All outgroup taxa have a distinct slender
gonostylus similar to state 0.

43. Lateral edge of penis valve.—(0) Smooth, without
a carina (Fig. 9A, 9B, 9C, 9D, YE, 9F); (1) protuberance or
carina present (Fig. 8A, 8B, 8C). State 1 is found in the
outgroup only in some members of the Allodapini
(Michener, 1975). The ground-plan state for the ingroup is
inferred to be State 0.

44. Apex of penis valve.—(0) Not hirsute (Fig. 8A, 8B,
8C); (1) hirsute (Hurd, 1961, Figs. 1, 3, 4). This apomorphy
unites Xylocopoides and Calloxylocopa and does not occur
in the outgroup.

Fig. 7. Dorsal surface of T6 of females: A. X. (Mesotrichia) flavorufa.
B. X. (Xylocopoides) californica. C. X. (Notoxylocopa) tabaniformis (modified
from Hurd and Moure, 1963). Abbreviations: p = pygidial spine; sa =
preapical spine.

45. Penis.—(0) Entirely membranous; (1) sclerotized
basally on venter. State 0 characterizes all outgroup taxa.

46. Spatha.—(0) Reduced or absent (Fig. 8A, 8B, 8C);
(1) well developed (Hurd and Moure, 1963, Figs. 131, 134).
State 1 is an autapomorphy for Notoxylocopa. Although a
spatha is present in many groups of bees, it is not so well
developed in the outgroup.

47. Gonostylus of male genitalia—(0) Slender, elon-
gate and erect, over 1/2 as long as gonocoxite (Fig. 9D);
(1) slender and less than 1/2 as long as gonocoxite (Figs.
8A, 8B, 8C, 9A, 9B, 9C, YE, YF), or entirely fused with
gonocoxite and represented only by the rounded or coni-
cal setose area at apex of gonocoxite (Character 42-1). An
elongate, slender gonostylus with numerous setae is con-
sidered to be plesiomorphic for bees. The outgroup is vari-
able in this respect, but in each tribe some member has
this plesiomorphic condition, i.e., Pararhophites
(Pararhophitini), Parafidelia (Fideliini), Euceratina
(Ceratinini), Manuelia (Manueliini), and Compsomelissa
(Allodapini) and some Allodapula (Allodapini).

48. Gonostylus of male genitalia —(0) Lacking a dense
patch of setae on its apex (Fig. 8C); (1) with dense apical
patch of setae (Figs. 8A; 9B). State 1 is not found in the
outgroup.

49. Apex of gonostylus with medially projecting
lobe.—(0) Absent (Figs. 8B, 8C; 9C, 9D); (1) present, not
well developed (Fig. 9B); (2) present, well developed (Fig.
8A). The apex of the gonostylus is entire in all outgroup taxa.

50. Ventromedial margins of gonocoxites.—(0) Diverg-
ing continuously from at, or near, ventral gonocoxal bridge
(Fig. 9A, 9F); (1) roughly parallel from ventral gonocoxal
bridge and not diverging until well above ventral
gonocoxal bridge (Figs. 8A, 8B, 8C, 9B, 9C, 9D, 9E). This
character is variable in the outgroup. The ground-plan state
has been inferred based on State 0 being a consistent fea-
ture of the Ceratinini and Megachilidae (i.e., Pararhophites,

2 ScreNTIFIC PAPERS, NATURAL History Museum, THE UNIVERSITY OF KANSAS

Fig. 8.
fied from Hurd and Moure, 1963): A. Xylocopa (Xylocospila)* bambusae. B.
X. (Cirroxylocopa) vestita. C. X. (Neoxylocopa) varipuncta. Abbreviations: lc
= lateral carina of penis valve, ml = medial lobe of gonocoxite, v =
ventroapical plate. *Regarded as a member of the subgenus Schonnherria
herein.

Dorsal, lateral, and ventral views of male genitalia (modi-

Parafidelia and Lithurge) and one member of the Manueliini
(Manuelia gayatina).

51. Medial area of posteroventral margin of
gonocoxite.—(0) Not produced (Figs. 8B, 9A, 9C); (1) with
posterior projection (Figs. 8A, 8C, 9B, 9C). State 1 is not
found in the outgroup.

52. Ventroapical plate of gonocoxite.—(0) Absent (Fig.
9A, 9B, 9C, YE, YF); (1) present, but not carinate and not
well defined (Hurd and Moure, 1963, Figs. 135, 136); (2)
carinate (Figs. 8A, 8B, 8C, 9D). The ventroapical plate of
the Xylocopini is possibly comparable to that found in the
Allodapini (Michener, 1975; Reyes, 1991). Nonetheless, the
inferred ground-plan state of the outgroup is absence of
such a structure.

53. Projection or spine on inner basal margin of un-
der surface of gonocoxite.—(0) Absent; (1) present (Fig.
9D). Many ingroup taxa have a spine along the ventral
margin of the base of the gonocoxite that is more lateral
than the condition defined here. This condition was coded
as 0. No such projection is found in the outgroup.

SUBGENERA AND GENERA OF CARPENTER BEES, XYLOCOPINI 13

—

<4

Fig. 9. Ventral views of the right sides of male genitalia: A. X.
(Xylocopa) valga. B. Lestis* aeratus. C. X. (Prosopoxylocopa) mirabilis. D. X.
(Bomboixylocopa) rufipes. E. X. (Ctenoxylocopa) fenestrata. F. Proxylocopa*
(Proxylocopa) rufa. Abbreviations: i = internal medial projection, ma =
medial apodeme, ml= medial lobe. Scale = 1 mm. *Regarded as a subge-
nus of Xylocopa herein.

54. Posterior margin of male T7.—(0) Entire; (1) with
pair of dentiform projections. State 1 is not found in the
outgroup.

55. Penis valves.—(0) Apically expanded (Figs. 8A, 8B,
8C, 9D); (1) slender and parallel (Fig. 9A, 9C). Though some
outgroup taxa have relatively slender penis valves they
are not as slender and parallel-sided as in State 1,so I have
tentatively inferred this condition to be apomorphic in the
Xylocopini.

PHYLOGENETIC ANALYSES

Four separate sets of analyses (described below) were
made. For each analysis the matrix was first analyzed with
all characters weighted equally and a second time using
successive approximation weighting (Farris, 1988). Succes-
sive approximation is an a posteriori character weighting
procedure that weights each character based on its best
performance (= fewest steps) in the previous analysis. The
weighting scheme is applied repeatedly until weights of
all characters stabilize. Before each successive approxima-
tion analysis the characters are coded in additive binary
form to insure that the weights assigned to states of
multistate characters are correct. The alternative coding
method for such characters is with all states in one col-
umn. Coding multistate characters in the latter way can
result in assignment of an average weight of all states; if
coded in additive binary form, both characters that per-
form poorly (show high homoplasy) receive more weight
and characters that perform well (show low homoplasy)
receive a lower weight than they would otherwise. Car-
penter (1988) provided an example of how these different
coding schemes can affect tree topologies. PAUP provides
many options for calculating weights for successive ap-
proximations. For ease of comparison, I followed the
weighting scheme used by the Hennig86 parsimony pro-
gram (Farris, 1988; explained by Fitzhugh, 1989), which
assigns weights to characters from 0 to 10. Nevertheless,
my results may differ from those obtained from Hennig86,
because some characters coded as polymorphic in the
PAUP data matrix would have to be considered missing
ina Hennig86 format (Hennig86 does not accept polymor-
phic characters) and might be optimized differently or as-
signed different weights.

The large size of the matrix restricted the search for the
shortest trees to heuristic searches in the PAUP program
version 3.01 (Swofford, 1990). Because these types of
searches cannot guarantee finding the most parsimonious
tree for the data, I searched by two methods in each analy-
sis. In the first search the following settings were used:
Starting trees by stepwise addition, addition sequence
simple, hold 75 trees at each step and swap on all starting
trees including trees that were nonminimal. The second
search used the random addition sequence option, which

14 ScreNTIFIC PApers, NATURAL History Museum, THE UNIVERSITY OF KANSAS

begins searches from trees produced at random by the pro-
gram. Five hundred replications were run and set to swap
on no more than 30 trees at each step up to and including
the fewest steps obtained by the first search.

The four analyses were undertaken to investigate un-
certainties in assumptions about the evolutionary changes
or homology assessment of characters, and of the relation-
ships of the outgroups discussed above in the section on
Selection of Taxa. For brevity, the four analyses are referred
to hereafter as Analysis 1, 2, 3 and 4. Results from these
four analyses after successive approximations weighting
are followed by the letter A (e.g., Analysis 1A, 2A etc.).

Analysis 1, equal weighting.—This analysis was based
on the full matrix (Table 3). Assumptions about character
ordering were as described in the Character Analysis sec-
tion (i.e., some characters were coded as additive and oth-
ers as nonadditive), and ground-plan states for all charac-
ters were inferred by assuming Manueliini to be the sister
group to Xylocopini + (Ceratinini + Allodapini) in the sub-
family Xylocopinae.

Analysis 1A, successive approximations. —This analy-
sis used the same matrix as in Analysis 1 with successive
approximations character weighting applied.

Analysis 2, equal weighting.—This analysis was done
with Proxylocopa excluded because I had coded some char-
acters as missing for this taxon in Analysis 1. Other than
the lack of Proxylocopa, this was the same matrix and char-
acter assumptions as used in Analysis 1.

Analysis 2A, successive approximations.— This analy-

sis was done by applying successive approximations to
the reduced matrix used in Analysis 2.

Analysis 3, equal weighting.—This analysis was done
using the same matrix as in Analysis 1 but with all charac-
ters coded as unordered.

Analysis 3A, successive approximations.—This analy-
sis was done using successive approximation weighting
on the matrix used in Analysis 3. Because islands of equally
parsimonious trees were found in Analysis 3, six analyses
were run starting with each of the six islands produced in
Analysis 3.

Analysis 4, equal weighting.—This analysis was done
to investigate the role that the uncertain relationships
among the tribes of the Xylocopinae might have on the
polarity assessment of the character states used in this
study. Ground-plan states of all characters were deter-
mined with the Xylocopinias the sister group to Manueliini
+ (Ceratinini + Allodapini). This reanalysis resulted in a
change of the hypothesized groundplan state for charac-
ters 25 and 36 from 0 to polymorphic (0, 1).

Analysis 4A, successive weights.—This analysis was
done using successive approximation weighting on the
matrix used in Analysis 4.

The following abbreviations are used in describing the
trees from each analysis: C.I. = consistency index, R.I. =
retention index and T.L. = treelength. For those trees re-
sulting from successive approximations weighting, statis-
tics are given for equally weighted characters to facilitate
comparisons with the other analyses, and for weighted
characters.

RESULTS

The eight different analyses were attempts to explore
the effect that various assumptions have on tree topology.
The influence on tree topologies of coding characters as
missing (Analyses 1, 1A, 3, 3A, 4, 4A), or of deleting taxa
altogether (Analyses 2 and 2A) remains insufficiently
understood (Gauthier et al., 1988; Maddison, 1993; Platnick
et al., 1991; Weins and Reeder, 1995). Additive coding of
character states (Analyses 1,1A,2,2A,4,4A) assumes that
an a priori knowledge of the evolutionary direction of char-
acter change is correct, an approach that has been ques-
tioned unless ontological data are available (Nelson, 1985).
Successive approximation weighting has been investigated
only by Farris (1969) who used a hypothetical data set with
complete character independence and a known phylog-
eny. These considerations emphasize that the results of this
study are only hypotheses, and should be treated as such.
Because of reservations concerning the transformation or-
der for several multistate characters (discussed in the Char-
acter List and Coding section), I consider Analysis 3 that
had all taxa included, characters weighted equally and

characters coded as nonadditive (unordered) to have the
least assumptions. However, it seems reasonable that the
strongest hypotheses for relationships are those that were
repeatedly supported under different assumptions.

I discuss each analysis and show one fully resolved cla-
dogram and strict consensus trees for unique topologies.
Topologies largely redundant with those from other analy-
ses are not presented. Consensus trees are not accompa-
nied with character changes because of problems with
character optimization when polytomies are present
(Maddison and Maddison, 1992). Successive approxima-
tion weighting produced more trees than corresponding
analyses with all characters weighted equally in all analy-
ses, except Analysis 2. This result was because the weight-
ing algorithm assigns weights of zero to many characters,
effectively removing them from the analysis. Hence, some
nodes were left with no character support. Successive ap-
proximation weighting was also highly sensitive to the
topologies of the initial trees. For example, in Analysis 3,
three sets of relationships were generated even though the

SUBGENERA AND GENERA OF CARPENTER BEES, X YLOCOPINI 15

weighting iteration began with a set of equally parsimoni-
ous trees from the same data set. If this weighting approach
could reliably distinguish characters that were truly ho-
mologous from those that were homoplasious, a greater
level of congruence should be obtained in the weighted
trees than was found here. A thorough study on how mul-
tiple islands of trees influence results using successive
weighting is beyond the scope of the present paper.

In the following text and cladograms, I refer to those
subgenera represented by more than one exemplar by their
subgeneric name if the exemplars were shown to be mono-
phyletic in all analyses. Thus, in the text Schonnherria,
Xylocopoides, and Notoxylocopa are used, but each refers to
multiple exemplar species (Table 3). Monophyly was not
consistently supported for the exemplars of Nodula,
Xylocopa, and Koptortosoma. For the first two groups, I use
species names of all exemplars in the discussion and on
the cladograms. I restrict use of Koptortosoma to X. caffra
and X. pubescens, and designate X. sinensis separately; the
first two exemplars were grouped together consistently,
whereas the relationship of the latter species to the other
species of Koptortosoma varied among analyses.

The following section is a summary of the relationships
among the major groups in the tribe. This is followed by a
section discussing the relationships of the component taxa
within these larger clades.

RELATIONSHIPS AMONG LARGER CLADES OF THE XYLOCOPINI

In most analyses, three groups were shown to be mono-
phyletic; for ease of discussion, I have given each an infor-
mal name. These are: (1) the Ethiopian group consisting of
Afroxylocopa, Alloxylocopa, Bomboixylocopa, Cyaneoderes,
Cyphoxylocopa, Hoplitocopa, Hoploxylocopa, Koptortosoma,
Mesotrichia, Mimoxylocopa, Oxyxylocopa, Platynopoda,
Prosopoxylocopa, Xenoxylocopa, and Zonohirsuta; (2) the New
World group consisting of Cirroxylocopa, Dasyxylocopa,
Megaxylocopa, Nanoxylocopa, Neoxylocopa, Schonnherria,
Stenoxylocopa, Xylocopoda, Xylocopsis,and Xylocospila; and (3)
the Rhysoxylocopa group consisting of Acroxylocopa,
Apoxylocopa, Biluna, Epixylocopa, Nodula (both exemplar spe-
cies), Perixylocopa, Rhysoxylocopa (sensu stricto), and
Xylomelissa.

The Ethiopian and New World groups are large and
are further divided into subgroups. For the Ethiopian
group these are; (1) the Mesotrichia subgroup consisting of
Hoplitocopa, Hoploxylocopa, Mesotrichia, and Platynopoda;
and (2) the Koptortosoma subgroup consisting of
Afroxylocopa, Cyaneoderes, Cyphoxylocopa, Koptortosoma, and
Oxyxylocopa. In the New World group, I refer to a
Neoxylocopa subgroup which includes Megaxylocopa,
Neoxylocopa, Stenoxylocopa, and Xylocopina.

Analysis 1, equal weighting.—Manueliini is basal in
the subfamily; some characters were coded as additive and

others as nonadditive. Proxylocopa is included. (Fig. 10A).
This analysis yielded eight minimum length trees (C.I. =
0.27, R.I. = 0.67, T.L. = 288.) Differences in the topology of
the eight trees produced by this analysis were among
Cyaneoderes and Cyphoxylocopa and their relationship to
Koptortosoma, X. (Koptortosoma) sinensis and Afroxylocopa.
For all other subgenera and genera of the tribe this analysis
produced a single hypothesis of relationships. Four major
clades are shown to be monophyletic. The most basal group
consists of a biogeographically diverse set of taxa that in-
cludes the Australian Lestis as the sister group to the North
American subgenera Calloxylocopa and Xylocopoides. Together,
these taxa comprise a sister taxon to the Old World subgen-
era Copoxyla + (Ctenoxylocopa + X. (Xylocopa) violacea and X.
(Xylocopa) valga). The monophyly of all members of this clade
is supported by Character 36, the emarginate shape of Sl.
This character is highly homoplasious and appears indepen-
dently six times in the tribe, thereby suggesting little confi-
dence should be placed in this grouping.

This clade and the next more-derived branch with the
genus Proxylocopa are separated by the absence of the male
metapostnotum (Character 24-1) and presence of a male
mesosomal gland (Character 26-1). Both characters are lost
and gained repeatedly in the tribe and furthermore are
functionally correlated; presence of the mesosomal gland
is often associated with absence of the metapostnotum
(Minckley, 1994). As such, the relationship of Proxylocopa
and the taxa shown to be more basal to it is best consid-
ered as unresolved.

Those taxa that are derived relative to Proxylocopa are
defined in this analysis by possession of a carina on the
venter of the gonocoxite (Character 52-2), and elaboration
of the reservoir of the male mesosomal gland from two
small pouches to a series of tubules (Character 26-2). De-
velopment of this carina is a strong feature that arises once
in this topology, although it becomes rounded (Character
52-1) in the Rhysoxylocopa group. Three major clades are
included herein; the Riysoxylocopa group is basal and the
sister group to a clade that includes the Ethiopian group
and a clade of taxa that occur primarily in the New World
(the Neoxylocopa group, Xylocopoda, Cirroxylocopa,
Notoxylocopa, Xylocopsis, Dasyxylocopa, Nanoxylocopa,
Schonnherria, and Xylocospila) but with two Old World
members (Gnathoxylocopa and Nyctomelitta).

Few characters separate these main groups; yet several
of these features probably are unique within the Xylocopini
and show little homoplasy in my analyses. The sister group
relationship of the Ethiopian group plus the large group
of New World taxa (including Gnathoxylocopa and
Nyctomelitta in this topology) is supported by the parallel
ventromedial margin of the gonocoxites (Character 50, Fig.
9B; lost in Platynopoda, Hoploxylocopa, and Bomboixylocopa)

Screntiric Papers, NATURAL History Museum, THE UNIVERSITY OF KANSAS

16

‘sisATeue suoyeutxoidde aatssaoons ayy

Aq payeiaua saa.y 6g Jo Wessopr|D SnsUasUOd JU}S “g “1 xrpuaddy ur paquosap si uoNgiystp JajereyD ‘[enba pay sieyoereyp Te Jo s}Y3IaM pue (}Xa} ay} Ul paquiosap se) siajeIeYyD
9]e}sy[NUW AAYIPpeuou pue aayippe yIO YIM aUOp seM siskyeuy ‘punoy sex} snomounsied sour yYSIa ayy JO dex} paayosai-AT[Ny “uasoyp Ayjwopurei aug “y | sisApeuy “OL “3Iq

pByjen (edos0/Ax) ‘x
eaarjoin (edoz0j/Ax) ‘x
edosojAxouaj9
edos0/Axoy/eD
sapiodos0jAx
SSOT

edo20/AxOld
pjAxodog
edoar0j/Axowily
edooojAxioqwog
edos0/Axouex
sisuauls (BWOSsoj}oO10}doy) *x
salapoeuePAD
edosojAxoydhg
edo020/AxAxO
Pwosoyojdoy
edoaojAxoyy
edoao/Axojdoy
edoso0}!|doH
epodou/}e|q
PIYoOSaWy
edoa0/Axo]|\¥
edov0/Axodosold
BINSJIYOUOZ
edooojAxoyjeu
eun|ig

juojuag (e/NPON) °X
punsAyjawe (ejNPON) x
essijawojAX
edosojAxody
edosojAxo0Vy
PdoI0/AXlad
edoav0/Axid¥
edosoj/AxosAyy
P}I/BWOJOAN
euldoo0j/Ax
edos0/Axoua}g
edoso/Axebayy
edoso0j/Axoan
sisdoa0jAx
edoaojAxojoN
Byidsoo0jAx
Pia@yuuoyosS
edoazo0j/AxAseq
epodosojAx
edoa0jAxoulg
edoso0/AxoueNn
dnosjno

eiAxodo9a

eben (edoa0j/Ax) ‘x
paapjoin (edoa0jAx) ‘x
edoao0/AxouajD
SSOT
edoso0/Axo|eO
sapiodosojAx
edO020/AxOlg

eunig

euyjsXyjowe (e/npoN) *X
ess||QwojAX
edosojAxody
edosoj/Axo10yy
edoaojAxuad
Juojuag (e/nPON) *x
edoa0/Ax!d¥
edosojAxoshyy
edosojAxouex
edos0/AxAxO
salapoaueAg
edosojAxoydho
ewosoyoj}doy
sisuauls (2Wwosopyo}doy) *x
edos0/Axoyy
edosoy!|doH
epodoufje|d
edoszo0/Axo;doy
BIYyOWOse—y
edoas0/Axol|\v
edoo0jAxowly
edoso/Axioqwiog
edov0/Axodosold
BINSJIYOUOZ
euldoa0j/AK
edos0/Axoua}s
edosojAxebayy
edoso/Axoan.
ppodos0/Ax
PdoI0/AXOWID,
PYWOJOAN
edoa0/AxojON
edovo0/Axoyjeu5
sisdoo0j/AX
edosojAxfseq
Bjidsoa0jAx
Pia@yUUOYyIS
edosojAxoueNn
dnoi3)no

6, €8)
© @&)
a

SUBGENERA AND GENERA OF CARPENTER BEES, X YLOCOPINI 17

and possession of a process or spine on the venter of the
gonocoxite (Character 51, Fig. 9C; lost repeatedly). Mono-
phyly of the Ethiopian group is supported by the posses-
sion of two spines on the male hind tibia (Character 11-2)
and female (Character 12-1, lost in Afroxylocopa), begin-
ning of the thoracic declivity abruptly rounded in the fe-
male (Character 23-1; lost in Bomboixylocopa and
Mimoxylocopa), derived loss of a slender elongate
gonostylus (Character 47-0; regained in the Mesotrichia and
Koptortosoma groups), and sparse setae on the gonostylus
(Character 48). These characters are found in other groups,
but the congruence of characters is support for the mono-
phyly of this clade.

The clade composed of the New World group and
(Notoxylocopa + Gnathoxylocopa) + Nyctomelitta is defined
by the presence of a subapical tooth on the mandible of
the female (Character 3-1) and a lateral projection on the
penis valve (Character 43-1). The subapical mandibular
tooth is a derived loss in Gnathoxylocopa and is present in
groups outside this clade. This placement of Gnathoxylocopa
is unlikely (discussed below).

Analysis 1A, successive approximations.—Successive
approximations were applied to the matrix used in Analy-
sis 1A (Fig. 10B). This analysis yielded 89 minimum length
trees (unweighted statistics, T.L. = 331, C.I. = 0.23, R.I. =
0.65; weighted statistics, T.L. = 362, C.I. = 0.64, R.I. = 0.92).
Except for the most basal clade, this analysis recovered
the same major groups as in Analysis 1. The monophyletic
group positioned at the base of the tribe in Analysis 1 is
represented as a grade of five sequential clades along the
base of the cladogram in this analysis. Hypotheses of rela-
tionships among members within these larger clades were
more numerous than in the equally weighted analysis, and
more polytomies exist in the consensus trees resulting from
these analyses.

Analysis 2, equal weighting. —Manueliini is basal in
the subfamily; some characters were coded as additive and
others as nonadditive. Proxylocopa is excluded. This analy-
sis produced 48 trees divided among three groups, or “is-
lands,” of 16 trees each (T.L. = 282, C.I. = 0.28, R.I. = 0.70).
Strict consensus trees of each of the three islands differ
from those of Analysis 1 most notably in the relationship
of Biluna to the other members of the Xylomelissa group.
Island 3 places Rhysoxylocopa (sensu stricto) as a sister
group to the remainder of the Rhysoxylocopa group, and in
Island 2 Biluna is a more derived member of the clade.
Recognition of the major clades was the same as those pro-
duced in Analysis 1. Because of the similarity of results to
those in Analysis 1, the results of this analysis are not
shown as a cladogram.

Analysis 2A, successive approximations.—
Manueliini is basal in the subfamily, some characters were

coded as additive and others as nonadditive. Proxylocopa
is excluded; successive approximations were applied. Af-
ter five weighting iterations, character weights stabilized
and produced 25 minimum length trees (unweighted sta-
tistics T.L. = 298, C.I. = 0.25; R.I. = 0.67, weighted statistics
T.L. = 511, C.I. = 0.49; R.I. = 0.86). All islands of trees con-
verged on the same consensus tree, wherein the large clades
were much the same as those in Analyses 1 and 1A. Taxa
supported as a monophyletic group in Analysis 1 form a
basal grade of taxa in this analysis. In the resolved tree
and in the consensus tree, the Ethiopian group plus the
Rhysoxylocopa group and the New World assemblage are a
single clade. The relationships among members within these
three major groups are unresolved in the consensus tree.
Analysis 3, equal weighting —Manueliini is basal in
the subfamily; all characters were unordered. Proxylocopa
is included (Figs. 11A, 11B, 13A, 13B, 15). Coding all char-
acters unordered resulted in 408 trees distributed across 6
islands (T.L. = 297, C.I. = 0.27, R.I. = 0.65). The numbers of
trees for each island are: Island 1 = 24; Island 2 = 72; Island
3 = 144; Island 4 = 72; Island 5 = 72; Island 6 = 24. Consen-
sus trees of all trees in each of these islands generally are
not well resolved but some unique relationships were
found. The Ethiopian group was shown as monophyletic
in three of six islands (Islands 2, 4, 5; Fig. 13A, 13B). In the
three islands in which monophyly for this group was not
supported, the two basal taxa, Zonohirsuta and
Prosopoxylocopa, were not included with the remainder of
the Ethiopian group. Island 1 (Fig. 11A, 11B) placed
Zonohirsuta + Prosopoxylocopa as the sister group to the clade
of Xylocopa + (Calloxylocopa + Xylocopoides) based on the
derived loss of the metapostnotum in the male (Character
24-0) and female (Character 25-0). Together, this clade is
the sister group to the remainder of the Ethiopian clade,
minus Zonohirsuta and Prosopoxylocopa. In Island 3 (not
shown), Zonohirsuta + Prosopoxylocopa are part of a
polytomy that includes the remainder of the Ethiopian
group and other taxa. In Island 6 (Fig. 15), Zonohirsuta +
Prosopoxylocopa are a sister group to Notoxylocopa +
Proxylocopa and the remainder of the Ethiopian group.

These analyses raise the possibility that Zonohirsuta and
Prosopoxylocopa are not sister taxa to the rest the Ethiopian
clade; perhaps they are parts of a more inclusive group
that taken together are basal to the rest of the Ethiopian
group. Proxylocopa is associated with the basal lineage of
the Ethiopian group in all islands. Xylocopa, Notoxylocopa,
Xylocopoides, and Calloxylocopa are basal to the Ethiopian
groups in Islands 1,3, 5, 6. All of these taxa and the Ethio-
pian group have two spines on the hind tibia of the male
(Character 11-2) and female (Character 12-1; lost in
Calloxylocopa and X. (Xylocopoides) cyanea). The topologies
of Islands 1 (Fig. 11A), 5, (Fig. 13A) and 6 (Fig. 15) require

SCIENTIFIC PAPERS, NATURAL History Museum, THE UNIVERSITY OF KANSAS

18

‘sisATeue payysiem ATTenba jo aadq snsuasuod yo11y9 ‘g ‘Zz XIpuaddy ur uMoYs UOYNqLYsIp Jo}VIeYD /puNoy saaq] FZ ay} WO Uasoyp ATWOpue. aa] paatosar AT[NJ aU ‘VW Papper se
papod a1aM Sia} ALY 9}LISYINY ‘Papnpour siajovieyp [Te pue exe} [[e YJIM auOp seM sIsATeuY *saat] JO puLISI auO UO paseg sdiysuoyKyes DjauasoyAy J ‘| purysy] ‘¢ stsATeuy

edoa0j/Axold
edooo/Axodosold
sasapoaueAg
edosojAxoydhg
edos0/AxAxO
PWOsOLIO}doY
sisuauls (ewoOsOLoj}doy) ‘x
edoso0/Axoyy
edosojAxojdoyH}
edooo}!|doy
PpodouAje}q
BIYOLJOSAYy
edooo/Axouex
edoso0/Axo||\V
edooz0/AxOully
edoso0/Axioquiog
PINSIIYOUOZ
euljsAyjowe (e/NpON) “Xx
essijawojAK
edosoj/Axody
edoso0/Axo1oy
edOd0/AXHad

eun|ig

luojuaq (B/NPON) ‘x
edoso/AxidZ
edosojAxosAyy
pjAxodo9

ehyen (edooojAx) “x
Paae/OIn (edoa0/Ax) “x
edoso0jAxouajD
S]Se7
edosojAxoyje9
saplodoao/Ax
edoao/Axojon,
P]I[AWOJIAN
edosoj/Axoyjeuy
sisdoa0jAx
PjIdsoo0j/Ax
PLaYUUOYIS
edooojAxAseq
Puldooo0/AK
edoao0/Axoua}|sg
edosojAxebayy
edoao|/Axoan
epodos0jAx
BdOD0/AXOMID,
edoao/Axouen
dnoi6jno

edod0/AxOld
edoz0/Axojon,
sa/apoauePAD
edosojAxoydhg
edoa0/AxAxO
Pwosopojdoy
sisuauls (ewosopojdoy) ‘x
edoso0/Axoyy
edozo/Axojdoy
edooojl|doyH
PpodouA/}e}dq
PIYDUOSE-y
edooojAxouax
edoz0/Axo||\V
edos0/AxOwly
edosojAxloqwog
edooojAxodosold
BINSIJYOUOZ

eByjen (edos0j/Ax) ‘x
PadR/OIA (edooz0/Ax) “x
edosojAxoyjea
sapiodos0jAx
BuysAyjawe (e/NPON) “x
Juojuag (e/nPON) *x
eunyig

PSsijawojAKy
edosojAxody
Pdoso0/Axoloy
PdOI0/AXL8dg
edooojAxidZ
edosojAxosAyy
PjAxodo9
edosoj/Axoyjeuy
edosz0jAxouajD
SHSO7

BUI BWOJOAN
sisdoo0jAK
PIIASODO/AX
PLUQYUUOYIS
edoso/AxAseq
euldosojAx
edosojAxoua}s
edosojAxebayy
edosojAxoan
PpodosojAx
PdoI0/AXOWID
edoso0/Axouen
dnoi6jno

SUBGENERA AND GENERA OF CARPENTER BEES, XYLOCOPINI 19

that Character 11-2 be gained and lost once, whereas in
other topologies, multiple gains and losses of this feature
must be postulated. Character 12-1 also occurs in the
Neoxylocopa group and Nanoxylocopa and as such is weak-
ened. In all analyses with characters coded as additive
(Analysis 1,1A,2,2A,4,4A), there was support for a mono-
phyletic Ethiopian group, as I have informally defined it.
However, if the order of character evolution assumed in the
other analyses proves to be incorrect, the relationships of
the other taxa to the Ethiopian group shown in the analysis
with all characters coded as non-additive may be robust.

Island 5 (Fig. 13A, B) places the Rhysoxylocopa group
basal in the tribe. This analysis is the only one in which
the Rhysoxylocopa group is placed in this position.
Proxylocopa + Notoxylocopa , Xylocopoides + Calloxylocopa,
Xylocopa (both species) are a grade along the stem of the
Ethiopian group. Lestis, Copoxyla + (Ctenoxylocopa +
Gnathoxylocopa) and Nyctomelitta are placed along the stem
leading to the New World group.

All islands recovered monophyletic Riysoxylocopa and
Neoxylocopa groups; however, I consider the relationship
of these subgenera to each other to be unresolved. Each
island of trees associated the Old World taxon, Nyctomelitta,
as the sister group to the large New World group referred
to above. Basal to this New World clade plus Nyctomelitta
was a group of Old World taxa that, depending on the
analysis, included Lestis, Copoxyla and Ctenoxylocopa +
Gnathoxylocopa.

Analysis 3A, successive approximations.—
Manueliini is basal in the subfamily, all characters were
coded as unordered. Proxylocopa is included (Figs. 12, 14.)
Six analyses were run starting with each of the six islands
produced in Analysis 3. After the characters stabilized,
three of the six analyses generated unique consensus trees.
The starting islands (from Analysis 3) and the statistics are:
Topology 1 (Fig. 12) was found by starting at the trees in
Islands 1, 3, and 6 (n = 110 trees; unweighted statistics:
TL. = 344, C.I. = 0.22, R.L. = 0.64; weighted statistics: T.L. =
347, C.I. = 0.52, R.L = 0.88). Topology 2 was found by start-
ing at Island 2 and 4, (n = 700+ trees [stopped before all
trees were recovered]; unweighted statistics: T.L. = 343, C.I.
= 0.22, R.I. = 0.64; weighted statistics: T.L. = 359, C.I. = 0.50,
R.L = 0.88). Topology 3 (Fig. 14) was found by starting at
Island 5 (n =700+ trees [stopped before all trees were re-
covered ]); unweighted statistics: T.L. = 353, C.I1. = 0.22, R.I.
= 0.63; weighted statistics: T.L. = 347, C.I. = 0.57, R.I. = 0.89).

Successive approximations applied to each of the is-
lands produced three very different topologies. Topology
1 (Fig. 12) divides the tribe into two large clades. In one of
these clades Proxylocopa, Notoxylocopa, Zonohirsuta +
Prosopoxylocopa, X. (Xylocopa) violacea + (Calloxylocopa +
Xylocopoides) and X. (Xylocopa) valga are progressively more

Nanoxylocopa
Xylocopsis
Nyctomelitta
Schonnherria
Dasyxylocopa
Xylocospila
Notoxylocopa
Cirroxylocopa
Xylocopoda
Neoxylocopa
Megaxlocopa
Stenoxylocopa
Xylocopina
Rhysoxylocopa
Epixylocopa
Perixylocopa
Acroxylocopa
—— Apoxylocopa

Xylomelissa
X. (Nodula) amethystina
X. (Nodula) bentoni
Biluna
Proxylocopa
Copoxyla
Gnathoxylocopa
Lestis
Xylocopoides
Calloxylocopa
Ctenoxylocopa
X. (Xylocopa) violacea
Zonohirsuta
Prosopoxylocopa
Alloxylocopa
Xenoxylocopa
Mesotrichia

fim Platynopoda
Hoplitocopa
Hopoxylocopa
Afroxylocopa

i Oxyxlocopa

| Cyphoxylocopa

Cyaneoderes
Koptortosoma
X. (Koptortsoma) sinensis
Bomboixylocopa
Mimoxylocopa
X. (Xylocopa) valga
outgroup

Fig. 12. Analysis 3. Island 1. Strict consensus cladogram of six trees
found after successive approximations weighting.

derived taxa, respectively, along the stem to the Ethiopian
group (less Zonohirsuta and Prosopoxylocopa). The second
clade is not well resolved. Lestis, the Rhysoxylocopa group,
Copoxyla + (Ctenoxylocopa + Gnathoxylocopa), form a
polytomy that collectively is the sister to the New World
group.

Topology 2 divided the members of the more basal
clades found in Topology 1 into a number of separate
branches. The most basal branch in the tribe consists of
Ctenoxylocopa as the sister group to both species of Xylocopa
(sensu stricto) and Xylocopoides + Calloxylocopa are associ-
ated with the base of the Ethiopian group. Proxylocopa is a
basal taxon and the sister group to the Rhysoxylocopa group,
Copoxyla, Lestis, Notoxylocopa + (Gnathoxylocopa +
Nyctomelitta) and the New World group.

Topology 3 (Fig. 14) was poorly resolved with
Proxylocopa, both species of Xylocopa, and Ctenoxylocopa

Screntiric PAPERS, NATURAL History Museum, THE UNIVERSITY OF KANSAS

eulsAyjowe (BjnpoN) ‘x
essijawojAx
edoazo0/Axody
edos0/Axoloy
edoa0jAxlad
eun|ig

juojuaq (ejnpoN) ‘x
edos0jAxidZ
edoaojAxoshyy
edos0jAxojON
PdOI0/AxOld
edooojAxodosold
sajapoauerg
edosojAxoydhg
edo20/AxAxO
ewosoyioj}doy
sisuauls (ewosoyoj}doy) ‘x
edoaojAxoyy
edosojAxojdoyH
edosoj\/doH
epodou/je|d
BIYOUJOSa,y
edos0/Axouex
edos0jAxo||\V
edooojAXOWwy
edosojAxioqwog
PINSIIYOUOZ

eByjen (edoz0/Ax) ‘x
aeadRjoiA (edoz0/Ax) “x
edoaojAxoyjep
sapiodos0jAx
SI]SOT

BiAxodog
edosojAxoyjeuy
edoz0/AxouajD
PYIBWOJOAN
sisdoa0j|AK
PyIdsoo0|Ax
PLBYUUOYIS
edoso0/AxAseq
euidooo0jAx
edosojAxoue}s
edosoj/Axebayy
edosojAxoan
epodoaojAx
epodos0jAxouig
epodosojAxouen,
dno6jno

euyjsXyjawe (e/NpON) “x
essijawojAK
edoz0/{xody
edoso0/Axoloy
edoa0jAxUad
eunylg

juojuaq (ejnpoN) “x
edoz0/AxidZ
edosojAxoshyy
edosojAxojoN
edoa0/AxOld

phen (edoa0/Ax) ‘x
eaarjoin (edoooj{x) ‘x
edoa0/Axodosold
sasapoauehsg
edosojAxoydhg
edoa0/AxAxO
Pwosoyoj}doy
sisuauls (Bwosopo}doy) ‘x
edosojAxoyy
edosojAxojdoH}
edosojl|doyH
Ppodou/je|q
BIYOLOSEyy
edosz0/Axouex
edoso0jAxol|\V
edoz0/Axowly
edosojAxloquiog
PINSJIJYOUOZ
edosojAxoy|ebD
sapiodos0jAx
ejAxodo9a
edosz0/Axoyjeuy
edosojAxouajD
SI1SOT

PUQWOJOAN
sisdoo0j{x
PjIdsoa0/AK
BILIAYUUOYIS
edosojAxAseq
euldos0jAx
edoaojAxoua}g
edosoj/Axebayy
edoaoj/Axoayy
ppodosojAx
epodosojAxouig
epodoso/Axouen
\uidooo|Ax

‘sisAjeue payy3iem Aqjenba jo aasj snsuasuoD ‘g *¢ xtpueddy ur umoys UOyNqLysIp JoyeIeYD “punoy soy
ZZ dy) Woy Uasoyo ATWIOpuRs 9a.Q paafosar AT[NJ aU “W “[] ANSI UI UMOYs XLeUT Wo seer] JO puRjs! JoyjOUR UO paseq stsoyjodAy syauasoyAyg “¢ pues] ¢ sisATeuy “EL “SITY

SUBGENERA AND GENERA OF CARPENTER BEES, XYLOCOPINI 21

basal in the tribe. The Rhysoxylocopa is not monophyletic
and is placed near the base of the Ethiopian group.
Analysis 4, equal weighting.—Xylocopini is basal in
the subfamily; some characters were coded as additive and
others as nonadditive. Proxylocopa is included. Eight mini-
mum-length trees were found in this analysis (T.L. = 302,
CI. = 0.26, R.I. = 0.69). The changes in character coding
resulted in few changes among the relationships of the
ingroup. This result suggests that the problems in resolu-
tion found in the other analyses are not a result of polarity
assessment based on the use of different outgroups. Be-
cause this analysis provides little new information on re-
lationships, I do not present the cladograms or consider
the results in the following discussion.
Analysis 4A, successive weighting.—Xylocopini is basal
in the subfamily; some characters were coded as nonaddi-
tive. Proxylocopa is included. This analysis produced 123
most parsimonious trees (unweighted statistics: T.L. = 324,
CL. = 0.24, R.L = 0.66; weighted statistics: T.L. = 364, CL =
0.62, R.I. = 0.92). Results of this search are not shown for
the same reasons discussed for Analysis 4 above.

In summary, these data provide strong support for the
monophyly of some groups among the Xylocopini, but the
relationships among groups are often not strongly estab-
lished. The monophyly of the Neoxylocopa group and the
Rhysoxylocopa groups were supported in all analyses.
Monophyly of the Ethiopian clade was supported in all
but three islands of Analysis 3, where there were differ-
ences they were among the position of the basal members
of this clade. The derived members of the Ethiopian group
were associated in all analyses. A number of other sub-
genera not placed in the above mentioned informal group-
ings were placed in most analyses near the base of the tribe,
either as monophyletic groups with poor character sup-
port or as members of a grade of taxa.

RELATIONSHIPS AMONG MEMBERS OF THE MAJOR CLADES AND
TAXONOMIC RECOMMENDATIONS

Rhysoxylocopa group .—In the Rhysoxylocopa group,
linclude the following subgenera recognized by Hurd and
Moure (1964): Perixylocopa, Acroxylocopa, Apoxylocopa,
Xylomelissa, Nodula, Epixylocopa, Rhysoxylocopa, and Biluna.
A gradulus on T1 of the female (Character 32-2; lost in
Perixylocopa) was the only synapomorphy for the group in
all analyses. In the other analyses, various other charac-
ters also supported the monophyly of this group includ-
ing a row of tubercles on the edge of the basitibial plate
(Character 17; Analysis 3, all islands), the continuous di-
vergence of the inner ventral margins of the gonocoxites
(Character 50; Analysis 1A, 2, 2A, 3 [Island 3]), and the
carinate ventroapical plate of the male gonocoxite (Char-
acter 52-2). Although these are unusual features for bees,
each occurs independently in other Xylocopa.

Nanoxylocopa
Xylocopsis
Nyctomelitta
Schonnherria
Cirroxylocopa
Xylocopoda
Neoxylocopa
Megaxylocopa
Stenoxylocopa
Xylocopina
Dasyxylocopa
Xylocospila
Notoxylocopa
Rhysoxylocopa
Zonohirsuta
Prosopoxylocopa
X. (Nodula) amethystina
Alloxylocopa
Xenoxylocopa
Mesotrichia
Platynopoda
Hoplitocopa
Hoploxylocopa
Afroxylocopa
Oxyxylocopa
Cyphoxylocopa
Cyaneoderes
Koptortosoma

X. (Koptortosoma) sinensis
Mimoxylocopa
Bomboixylocopa

X. (Nodula) bentoni
Epixylocopa
Perixylocopa
Acroxylocopa
Apoxylocopa
Xylomelissa

Biluna
Xylocopoides
Calloxylocopa
Copoxyla
Gnathoxylocopa
Lestis
Ctenoxylocopa

X. (Xylocopa) violacea
X. (Xylocopa) valga
Proxylocopa
outgroup

Fig. 14. Analysis 3. Island 5. Strict consensus tree of 700 trees after
successive approximations weighting. Lengths of horizontal lines are
proportional to number of characters on each internode.

In the analyses with both nonadditive and additive
characters (Analyses 1, 1A, 2, 2A), Biluna was placed at
the base of the Rhysoxylocopa clade. In Analysis 3 (all char-
acters coded as nonadditive), the basal branch of this clade
was Rhysoxylocopa (sensu stricto). Depending on the to-
pology accepted, a derived loss must be hypothesized for
one of two characters shared by most members of this
group. With Biluna as the basal branch, the mesosomal
gland reservoir of males (Character 24) is secondarily lost
in Rhysoxylocopa and Xylocopa (Nodula) amethystina. With
Rhysoxylocopa as the basal branch, the reservoir of the
mesosomal gland originates once in this clade but the row
of tubercles on the posterior edge of the female basitibial
plate (Character 17) is lost in Biluna. A case can be made in
favor of this second scenario. For example, other reduc-
tions in the basitibial plate occur in Biluna, including com-

22 ScIENTIFIC PAPERS, NATURAL History MuseEuM, THE UNIVERSITY OF KANSAS

outgroup
Nanoxylocopa
Cirroxylocopa
Xylocopoda
Neoxylocopa
Megaxylocopa
Stenoxylocopa
Xylocopina
Dasyxylocopa
Schonnherria
Xylocospila
Xylocopsis
Nyctomelitta

Lestis
Ctenoxylocopa
Gnathoxylocopa
Copoxyla
Rhysoxylocopa
Epixylocopa
Perixylocopa
Acroxylocopa
Apoxylocopa
Xylomelissa

Biluna

X. (Nodula) bentoni
X. (Nodula) amethystina
X. (Xylocopa) violacea
X. (Xylocopa) valga
Zonohirsuta
Prosopoxylocopa
Bomboixylocopa
Mimoxylocopa
Alloxylocopa
Xenoxylocopa
Mesotrichia
Hoploxylocopa
Platynopoda
Hoplitocopa
Afroxylocopa

X. (Koptortosoma) sinensis
Koptortosoma
Oxyxylocopa
Cyphoxylocopa
Cyaneoderes
Proxylocopa
Notoxylocopa
Xylocopoides
Calloxylocopa

Fig. 15. Analysis 3. Island 6. Phylogenetic hypothesis based on another island of trees from matrix shown in Figure 11. One fully

resolved tree randomly chosen from the 24 trees; character distribution shown in Appendix 4.

SUBGENERA AND GENERA OF CARPENTER BEES, XYLOCOPINI 23

plete absence of a basitibial plate in the males and reduc-
tion or absence of a midtibial spine in both sexes. These
probably are associated with movement along the smooth
inner surface of the bamboo culms used for nesting
(Malyshev, 1931; Maeta et al., 1985). In females, the
midtibial spines are functionally replaced by an apical mat
of highly modified, thickened setae on the outer midtibia.
Thus, the loss of tubercles of the female basitibial plate
may be associated with loss of other structures on the leg.

Two characters are found only among members of the
Rhysoxylocopa group: the absence or reduction of an apex
or spine on the outer apex of the hind tibia of the male
(Character 11-0), and the presence of a row of tubercles on
the posterior edge of the basitibial plate of the female (Char-
acter 17-2). Character 11-0 is found in four African taxa
Perixylocopa, Acroxylocopa, Apoxylocopa and Xylomelissa and
one Asian taxon, Xylocopa (Nodula) amethystina. Character
17-2 is shared by these same taxa plus X. (Nodula) bentoni
and Rhysoxylocopa. Reduction of the hind tibial spine of
the male is approached in Epixylocopa and Biluna. For some
specimens of Epixylocopa and Biluna, it was difficult to code
unambiguously a spine as present or absent. The hind tibia
flares at its apex and, depending on the angle in which it
is viewed, the spine can appear absent, as a rounded pro-
jection, or present. Hurd and Moure (1963) listed X.
(Nodula) amethystina as having this structure and
Epixylocopa and Biluna as lacking it. In the specimens avail-
able to me, the conditions were reversed in these groups. I
have not examined the specimens that Hurd and Moure
used in their study, so this character may vary
intraspecifically, or it was mistakenly assessed.

Despite the number of topologies found for members
of the Rhysoxylocopa group, I have proposed some synony-
mies. Hurd and Moure (1963) recognized nine subgenera
of which four were monotypic, even though they were
aware that the Riysoxylocopa group was poorly understood.
Eardley (1983), in his study of South African species of
Xylocopa, pointed out that subgeneric rank for some of these
taxa is questionable.

The monophyly of Ritysoxylocopa and Biluna is well sup-
ported. Rhysoxylocopa all have extremely small propodeal
spiracles in the males. The males of Biluna lack spines on
the hind tibia, and females possess a dense mat of hairs on
the mid-tibia. Xylomelissa and its near relatives are more
problematic. In all analyses, Xylomelissa and Apoxylocopa
+ (Perixylocopa + Acroxylocopa) are either a monophyletic
group or closely allied. Support for their monophyly is
based on the absence of an angle or spine on the hind tibia
of the male (Character 11-0). This derived loss also occurs
in Xylocopa (Nodula) amethystina, and in some analyses
(Analyses 1A, 2, 2A), this species was grouped with the
other taxa that lack this structure. Inclusion of X. (Nodula)
amethystina within Xylomelissa would entail a greatly ex-

panded definition of Xylomelissa. Males of X. (Nodula)
amethystina completely lack a reservoir for the mesosomal
gland (Character 26), as do males of Riysoxylocopa, whereas
in Xylomelissa, Perixylocopa, Acroxylocopa, Apoxylocopa, and
Epixylocopa, the male mesosomal gland reservoirs consist
of a single layer of tubules approximately as deep as they
are wide in cross section. Therefore, I define Xylomelissa
on the basis of structure of the gland reservoir, and include
one species, X. (Xylomelissa) ruficollis, with a well-defined
spine on the apex of the male hind tibia; all other species
lack or have reduced forms of this structure. Recognition
of this species as the subgenus Epixylocopa was based on
the absence of yellow maculation on the face of the males
(Character 10-1) by Hurd and Moure (1963). They clearly
stated that other than this character Epixylocopa was close
to Xylomelissa. Because yellow maculation varies within
other subgenera and intraspecifically in X. (Xylocopoides)
californica arizonensis (Minckley, pers. obs.), this feature does
not warrant recognition as a subgeneric character.

None of my analyses grouped the two species of Nodula
as a monophyletic group. As the above discussion indi-
cates, the two exemplars I used, Xylocopa (Nodula)
amethystina and X. (Nodula) bentoni, differ notably. The
outer apex of the hind tibia of the female (Character 12)
has one spine in X. amethystina and two spines in X. bentoni;
the reservoir for the male mesosomal gland (Character 26)
is not present in X. amethystina but is present in X. bentoni;
and the beginning of the posterior thoracic declivity is
abrupt in X. amethystina and smoothly rounded in X.
bentont. The male genitalia are similar, but I have found no
discrete characters that serve as synapomorphies to unite
these species. Maa (1938) recognized two groups in his
original description of Nodula, which are represented by
the two exemplars I used here. Thus, I have retained this
subgenus to include the group typified by X. (Nodula)
amethystina and propose that the group represented by X.
(Nodula) bentoni be recognized as a new subgenus.

Ethiopian group.—As discussed above, support for
the monophyly of the Ethiopian group varied. All females
of the Ethiopian group, except Bomboixylocopa and
Mimoxylocopa, have the beginning of the thoracic declivity
abruptly rounded in the female (Character 23-1) and few
setae on the male gonostylus (Character 48-1).

Prosopoxylocopa appears as a basal member of the Ethio-
pian clade in most analyses, either as the sister group to
Zonohirsuta (Analyses 1, 1A, 3 [Islands 1, 6]) or as the sis-
ter group to Zonohirsuta plus the rest of the Ethiopian clade
(Analyses 2, 2A, 3 [Islands 2, 3, 4, 5]). Prosopoxylocopa is
distinct from Zonohirsuta based on interpretation of the
development of the ventroapical plate of the male
gonocoxite. In Prosopoxylocopa, the ventroapical plate is
present, but lacks a carina, whereas in Zonohirsuta and the
rest of the Ethiopian clade (except Platynopoda) the

24 Screntiric Papers, NaturAL History Museum, THE UNIVERSITY OF KANSAS

ventroapical plate has a distinct carina (Characters 52-1,
52-2). Because the carina of the ventroapical plate seems
to be a strong character in the Xylocopini, I favor the rec-
ognition of Prosopoxylocopa and Zonohirsuta as separate
subgenera. If further studies find Prosopoxylocopa is a mem-
ber of Zonohirsuta, separate subgeneric status for both taxa
will not be warranted.

The next most derived lineages within the Ethiopian
group, in all of the analyses except Analysis 2A, includes
Bomboixylocopa + Mimoxylocopa, Xenoxylocopa, Alloxylocopa,
the Mesotrichia group and the Koptortosoma group. As stated
above, all have a distinct carina on the ventroapical plate
of the male genitalia (Char. 52-2), a medial sulcus on T1
(Char. 35) and a projection on the inner basal margin of
the venter of the gonocoxite (Char. 53-1).

Although superficially dissimilar, many features unite
Bomboixylocopa and Mimoxylocopa. These include the long
malar space in the female (Char. 8-1), a rounded thorax in
the female (a derived loss, Char. 23-0), paired setal patches
on the pregradular areas of the male (Char. 39), and patches
of appressed setae near the metasomal spiracles (Char. 40).
X. rufipes is the sole species of the subgenus Mimoxylocopa.
Maa (1940) originally placed it in the subgenus Zonolursuta.
Hurd and Moure (1963) considered Bomboixylocopa most
closely related to Nyctomelitta. All my analyses, however,
support a sister group relationship to Mimoxylocopa.

The monophyly of Xenoxylocopa, Alloxylocopa, the
Mesotrichia group, and the Koptortosoma group is supported
by 4-6 characters in all analyses, including one feature that
is uniquely derived—the produced apical edge of S1 (Char-
acter 36-1). The placement of the two basal taxa of this clade
differed between analyses. In Analysis 1 (Fig. 10A, B), 1A,
2 (all islands) and 2A, Alloxylocopa is the sister group to
the Mesotrichia and Koptortosoma groups, whereas in all is-
lands of Analysis 3 (Figs. 11A, B; 13 A, B; 15) and 3A (Fig.
12) Xenoxylocopa is their sister group. The best supported
topology depends on judgment of the strength of three
characters, all of which are unique to the Xylocopini. Char-
acter 30, the shape of the gradulus on TI, is unreversed if
Xenoxylocopa is considered to be the sister group to the
Mesotrichia and Koptortosoma groups. In contrast, if
Alloxylocopa is placed as the sister group to the Mesotrichia
and Koptortosoma groups, two characters are unreversed—
presence of a fovea on the anterior surface of T1 in the
male (Character 33) and either a fovea or the entrance to
an acarinarium on the anterior surface of T1 in the female
(Character 34). Because characters 33 and 34 could be cor-
related (similar character states in both sexes), neither to-
pology may be favored numerically.

Most analyses showed a sister group relationship be-
tween the Mesotrichia and Koptortosoma groups. The excep-
tion was in Islands 1 (Fig. 12), 2, 3, and 6 of Analysis 3A

(successive weighting analysis applied to data matrix with
all characters coded as nonadditive), which places Xylocopa
sinensis as the sister group to the remainder of the
Koptortosoma group and the Mesotrichia group. This prob-
ably is a spurious result because of the high weight as-
signed to the carinate beginning of thoracic declivity (Char-
acter 23-2, see below). Females of X. sinensis have a rounded
thoracic shape, but other characters place them with cer-
tainty into Koptortosoma. I consider this character to be a
derived loss in X. (Koptortosoma) sinensis. The perpendicu-
lar angle of the beginning of the posterior thoracic decliv-
ity in both sexes (Characters 22-2, 23-2) provides the stron-
gest support for the monophyly of the clade consisting of
the Mesotrichia and Koptortosoma groups. The Mesotrichia
group (Hoploxylocopa, Mesotrichia, Hoplitocopa, and
Platynopoda) is supported by seven inferred
synapomorphies; three of these are strong. An elongate
tegula in the male (Character 29) is unique and unreversed;
the triangular shape of the hind trochanter in the female
(Character 14-1) is strong, but found in some species of
Koptortosoma (sensu Hurd and Moure, 1963); and the slen-
der penis valves of the male genitalia (Character 55) are
found elsewhere in the ingroup only in Xylocopa (sensu
stricto) and Xylocopoides.

Mesotrichia and Platynopoda + Hoplitocopa are supported
as a monophyletic group by the reduced length of the male
gonostylus (Character 47). Hoploxylocopa retains the elon-
gate gonostylus and is thus inferred to be the basal mem-
ber of this clade, although this condition is approached by
some species of Mesotrichia (sensu Hurd and Moure, 1964
[see LeVeque, 1928]). Platynopoda and Hoplitocopa are shown
to havea sister relationship based on six synapomorphies;
the strongest one is the possession of a sclerotized base of
the penis (Character 45). Despite this evidence, there are
similarities of the male genitalia of Hoplitocopa and
Mesotrichia (e. g., the broadly fused gonostylus and large
projection from the posteroventral margin of the
gonocoxite) that suggest an examination of the other spe-
cies of Mesotrichia and Platynopoda might support a differ-
ent topology than that shown here.

The monophyly of the Koptortosoma group is supported
by the presence of a transverse carina on the scutellum
that in most females overhangs the metanotum (Charac-
ter 23-3). Relationships among the taxa of this group
(Oxyxylocopa, Cyphoxylocopa, Cyaneoderes, Afroxylocopa, and
Xylocopa [Koptortosoma] sinensis, and Koptortosoma) were
resolved by characters that display considerable ho-
moplasy or are continuously distributed. The most con-
stant of these characters, the possession of a medial keel
on S6 of the male (Character 37), supports a clade of
Cyphoxylocopa, Cyaneoderes, Afroxylocena, and Koptortosoma
(sensu Hurd and Moure, 1963), but leaves Oxyxylocopa

SUBGENERA AND GENERA OF CARPENTER BEES, XYLOCOPINI DS

defined by plesiomorphies. Furthermore, the median keel
on S6 is lost in males of some members of Koptortosoma,
and occurs in Mesotrichia (sensu Hurd and Moure, 1963)
and Xenoxylocopa. | found no derived features that unite
Oxyxylocopa, Cyphoxylocopa, and Cyaneoderes into a single
clade. The apomorphies that Hurd and Moure (1963) used
to differentiate these taxa, such as the shape of the male
face and eyes, differences in the color of the vestiture, and
the presence or absence of a mite pouch in T1 of the fe-
males, are highly variable. I conclude that these subgen-
era should be united with the subgenus Koptortosoma.

The four taxa in the Koptortosoma group with the great-
est number of derived characters (Afroxylocopa, Xylocopa
caffra, X. pubescens, and X. sinensis) are defined by four in-
ferred synapomorphies, three of which display consider-
able homoplasy. A relatively strong character is the well-
developed medial lobe at the apex of the gonostylus (Char-
acter 49-2). This character is similar to the condition found
in Schonnherria and Dasyxylocopa. However, the gonostylus
at its attachment to the gonocoxite is thick in Schonnherria
(sensu lato) and Dasyxylocopa, and slender in Koptortosoma
(sensu Hurd and Moure, 1963), thereby suggesting inde-
pendent origins in these taxa. In my analysis, the vestiture
on the dorsum of the male entirely yellow (Character 1), is
a strong synapomorphy, but because it is not found in all
species of Koptortosoma (sensu Hurd and Moure, 1963)
(pers. obs.; also see Eardley, 1983), it provides weak sup-
port for the monophyly of the group. Koptortosoma (and
its allied taxa as discussed herein) is the most species-rich
taxon in the Xylocopini, and relationships among the con-
stituent taxa would require more detailed study than I have
undertaken. The shape of the medial lobe of the gonostylus
(Character 49-2; Hurd and Moure, 1963, Fig. 241) of
Afroxylocopa, X. pubescens and X. sinensis may prove phy-
logenetically useful and should be considered in any fu-
ture revision.

Neoxylocopa and Schonnherria clades.—Analysis 3,
with all characters coded as nonadditive and weighted
equally, indicated a monophyletic New World group. In
other analyses (Analyses 1, 1A, 2[Island 3]), Notoxylocopa,
Nyctomelitta, and Gnathoxylocopa were variously placed
between the clades that include Schonnherria and
Neoxylocopa. These taxa, Notoxylocopa, and the Old World
subgenus Nyctomelitta, share three characters: a carina or
protuberance on the outer lateral edges of the penis valves
(Character 43), the presence of a tooth on the upper carina
of the female mandible (Character 3), and the absence of a
metapostnotum in the males (Character 24). Character 43
is unique to this group. Characters 3 and 24 are highly
homoplasious.

The relationships among the taxa associated with the
Neoxylocopa group and Schonnherria (i.e., Nanoxylocopa,

Dasyxylocopa, Xylocopsis, Notoxylocopa, and Nyctomelitta)
differ greatly among the analyses that included additive
characters (Analyses 1, 2, 4). Many characters serving as
synapomorphies in this group are highly homoplasious,
and few characters support these branches of the cla-
dogram. One strong feature in Schonnherria (including
Xylocospila), Dasyxylocopa, Xylocopsis, Nanoxylocopa and the
Neoxylocopa group is a sharply-angled ventroapical plate
of the male genitalia. | did not include this as a character
in the data matrix, because the ventroapical plate varies
continuously across taxa from a rounded condition with
no evidence of a carina, to a rounded condition with a weak
carina (Hurd and Moure, 1963, Fig. 135), toa flattened plate
with a sharply-angled carina (Fig. 8A, B, C). Character 52,
which concerns the ventroapical plate, was coded accord-
ing to the presence or absence of this carina. Notoxylocopa
and Nyctomelitta do not possess these features, and they
have other features in common with taxa that do not have
the sharply-angled ventroapical plate. In Notoxylocopa, such
features are the absence of a reservoir of the male
mesosomal gland (Char. 26-0), tubercles on both sides of
the basitibial plate of the female (Character 17-1, in Xylocopa
guatemalensis), and the thickly hirsute and broadly attached
gonostylus (Character 48). In Nyctomelitta, these features
are tubercles on both lateral edges of the basitibial plate of
the female (Character 17-1) and the thickly hirsute and
broadly attached gonostylus (Character 48). Notoxylocopa
and Nyctomelitta may be sister groups to the New World
group of taxa, but many characters cited above occur in
the Old World clades. In the case of Nyctomelitta, its bioge-
ography, like that of the South African endemic
Gnathoxylocopa, strongly suggests their relationship with
other Old World taxa.

Some characters suggest groupings within the clade that
includes all of the New World taxa except Xylocopoides,
Calloxylocopa and Notoxylocopa. These are: (1) Character 27-
1, the beginning of the thoracic declivity on the scutellum
of the male (Cirroxylocopa, Stenoxylocopa, Xylocopoda, and
Xylocopina), or on the propodeum (Neoxylocopa and
Megaxylocopa); (2) Character 48-0 (a loss), a slender, elon-
gate gonostylus with few setae (Megaxylocopa, Neoxylocopa,
Stenoxylocopa, Xylocopina, and Xylocopoda); and (3) Char-
acter 6-1, the presence of a carina or smooth zone along
the epistomal suture (Megaxylocopa, Neoxylocopa,
Stenoxylocopa, Xylocopina, and Xylocopoda).

Stenoxylocopa (sensu lato, see below) and Neoxylocopa
(sensu lato, see below) were closely grouped in ali analy-
ses. They share three synapomorphies; the strongest is the
presence of two spines on the outer apex of the female hind
tibia (Character 12). Stenoxylocopa is defined here to include
Xylocopa ruficollis, because it shares with Stenoxylocopa
(sensu Hurd and Moure, 1963) fusion of the upper rutellar

26 Screntiric Papers, NATURAL History Museum, THE UNIVERSITY OF KANSAS

tooth with the preapical tooth (Fig. 1C). Xylocopa ruficollis
was previously placed in the monotypic subgenus
Xylocopina by Hurd and Moure (1963); this assignment was
based on four characters: Two are plesiomorphies—the
shape of the apex of S1 and the yellow markings restricted
to the lower part of the face—and two are autapomorphies—
the short clypeus and the shape of the tegula. Neither of
these types of characters is useful in establishing relation-
ships (Wiley, 1978); without including Xylocopina,
Stenoxylocopa is paraphyletic, or Xylocopina is defined en-
tirely by plesiomorphies.

The subgenera Neoxylocopa and Megaxylocopa are united
by features of the males—yellow vestiture (Character 1-
1), patches of appressed setae on the pregradular area
(Character 39), and the distinctive shape of the tubules and
large size of the male mesosomal gland (Character 26-3).
Character 26-3 and a modification of the scutellum fora-
men that accommodates the tubules of the mesosomal
gland (Minckley, 1994, Fig. 5E, F) are unique
synapomorphies of Neoxylocopa and Megaxylocopa. Hurd
and Moure (1963) limited the subgenus Megaxylocopa to
those species having a conspicuous carina over each of the
lateral ocelli in the female. Although I agree that the sub-
genus Megaxylocopa is distinct, its recognition clearly ren-
ders Neoxylocopa paraphyletic and ignores the strong simi-
larity between Neoxylocopa and Megaxylocopa.

The clade consisting of Dasyxylocopa, Nanoxylocopa,
Schonnherria, Xylocopsis, and Xylocospila can be defined by
the presence of a well defined ventroapical plate of the
male genitalia (Character 52-2) and the morphology of the
male mesosomal gland (Character 26-2; Minckley, 1994,
Fig. 4 B). The monotypic subgenus Nanoxylocopa was
placed as a basal taxon of this group (Analysis 3A [Islands
1, 5), as the basal taxon to the entire New World group
(Analysis 3 [Islands 1, 2, 3, 5, 6]), or the sister group to
Schonnherria (sensu lato) (Analyses 1, 2 [Islands 1, 3).

Schonnherria, Xylocopsis, Dasyxylocopa, and Xylocospila
share a well-developed medial lobe on the apex of the
gonostylus (Character 49-2). This character occurs else-
where only in Koptortosoma. The male genitalia of
Xylocospila share all features found in Schonnherria and
warrant inclusion of this monotypic subgenus as a mem-
ber of that group. Like Nanoxylocopa and Xylocopsis,
Dasyxylocopa is monotypic, and was defined by Hurd and
Moure (1963) largely by the lack of features found in
Schonnherria (sensu Hurd and Moure, 1963). These four
taxa share no strong features with other groups in the
Xylocopini, and may be plesiomorphic members of the
Schonnherria lineage. Features of the mesosomal gland are
coded as missing for some of these taxa and might pro-
vide insight into the relationships of these subgenera to
the lineages associated with Schonnherria, Neoxylocopa, and
their allies. Until the relationships among these taxa are

further clarified (see further discussion below), I leave these
taxa as separate subgenera.

Miscellaneous other clades.—The remaining taxa
(Xylocopa [sensu stricto], Ctenoxylocopa, Copoxyla,
Gnathoxylocopa, Xylocopoides + Calloxylocopa, Proxylocopa,
and Lestis) have distinct characters, some of which may be
synapomorphies, but they also share a number of
plesiomorphic features. Most relationships among these
taxa were not well resolved by my analyses.

The relationship of Calloxylocopa and Xylocopoides is
strongly supported; a row of preapical spines from the
pygidial spine (Char. 41-2) and the hirsute outer apex of
the penis valves (Char. 44) are features known in no other
bees.

Other relationships reveal disturbing character con-
flicts among these groups such as, among Xylocopoides,
Calloxylocopa and Xylocopa s. str. Three strong apomorphies
found elsewhere only in the Ethiopian clade are found in
Xylocopoides and Xylocopa s. str.: two spines on the outer
apex of the hind tibia in both sexes (Char. 11-2 and 12),
and an angle or projection on the base of the ventral side
of the gonocoxite (Char. 53). As previously discussed, these
characters may indicate relationship of this group to the
Ethiopian group. In addition to character 53, Xylocopoides
and X. (Xylocopa) violacea completely lack gonostyli and
have a foveate basitibial plate (known elsewhere only in
Lestis and X. (Xylocopa) convexior Hedicke, according to the
description of Maa, 1954). Xylocopa s. str. and Xylocopoides
are also the two most northerly distributed groups in the
tribe (Popov and Ponomareva, 1961; Hurd, 1954); most
other Xylocopini are tropical. However, if these are sister
taxa this relationship forces two characters, not found in
other bees, to evolve convergently in Calloxylocopa and
Xylocopoides; a series of spines along the lateral edges of
the pygidial plate (Char. 41-2) and setae on the outer apex
of the penis valves (Char. 44). The relationship proposed
thus seems unlikely. I consider it more probable that
Calloxylocopa is a derived member of the subgenus
Xylocopoides, and that X. (Xylocopa) violacea represents the
sister group to Xylocopoides.

This hypothesized relationship is complicated by the
inclusion of X. valga in the subgenus Xylocopa, as it requires
the gain of a well-developed gonostylus in X. valga from
the fully fused condition (Char. 42) found in X. (Xylocopa)
violacea. Gonostyli are commonly lost in bees, but it seems
unlikely that once lost they would be subsequently re-
gained. This might be evidence that Xylocopa s. str., as de-
fined by Hurd and Moure (1963), is polyphyletic or
paraphyletic. If so, X. valga is probably more closely re-
lated to Ctenoxylocopa because both share many features
and have gonostyli that are slender and largely without
setae (the latter feature was not coded in this analysis).

SUBGENERA AND GENERA OF CARPENTER BEES, XYLOCOPINI Dil

The ground-nesting group, Proxylocopa, is thought by
some to be the most ancestral taxon in the tribe (see Intro-
duction). Its relationship to the other Xylocopini is largely
unresolved in this analysis, in part, because of my inabil-
ity to homologize features of the basitibial plate and py-
gidial plate. Nonetheless, those structures I could homolo-
gize suggest that it is a plesiomorphic taxon. The male geni-
talia of Proxylocopa are most similar to those of the
Rhysoxylocopa clade and/or Copoxyla. The gonostylus of
these taxa is thickly hirsute apically and broadly attached
to the gonocoxite (Char. 48) and the gonocoxites diverge
continuously from the ventral gonocoxal bridge (Char. 52).
The absence of a defined metapostnotum in the female
(Char. 25) and one tibial spine in the male (Char. 11) are
features that also occur among some of members of the
Xylomelissa group and not in Copoxyla, and may be further
evidence of close relationship. Copoxyla and Proxylocopa
lack the ventroapical plate of the gonocoxite.

I conclude that among these taxa (Xylocopa s. str.,
Ctenoxylocopa, Copoxyla, Gnathoxylocopa, Xylocopoides +
Calloxylocopa, Proxylocopa and Lestis) the most
plesiomorphic members of the tribe are likely to be found.
The relationships require further analysis before any sub-
stantiated statements of relationships can be made. Pref-

erably future analyses should include more of the species
in these groups, which should reveal new characters or
redefine some of those used in this study. I have retained
all these groups as subgenera of Xylocopa.

These analyses show that many of the characters his-
torically used to define the subgenera and genera of the
Xylocopini arise repeatedly in the tribe and many groups
are defined by unique combinations of characters rather
than by characters that do not occur in any other group. A
possible problem with these analyses is the use of outgroup
comparison for structures that might be drastically influ-
enced by a tremendous increase in size. In general, the mem-
bers of the Xylocopini are many times larger than most mem-
bers of the outgroup. Structures that are readily influenced
by the large size may make determination of the groundplan
states for these characters difficult or impossible. An example
might be the basitibial plate as it is known to readily respond
to changes in body size in other bees (Reyes, 1993).

Except for Table 1, up to this point the taxa recognized
by Hurd and Moure (1963) have been used. From here on
in the paper, the relationships indicated in the preceding
sections are accepted, as are the nomenclatural results. For
example, Lestis and Proxylocopa are hereafter treated as
subgenera of Xylocopa.

KEYS TO THE SUBGENERA OF XYLOCOPA

New Wokrtb FEMALES
1. Pygidial plate with lateral margins marked by numer-
ous spines and strongly divergent anteriorly on disc of
RIE A/C) cas cSt cake crs sdsspc cn cwttncbosstnosaseustesessaseyeassaseetée< 2

Pygidial plate with lateral margins not marked by nu-
merous spines and not strongly divergent anteriorly,
sometimes represented by a spine that often has a
shorter preapical spine on each side of its base and thus
not or scarcely represented on disc of T6 (Fig. 7B) .... 3

2 (1). Basitibial plate foveate, closed basally (Fig. 2A); outer
apex of hind tibia with two spines; median tubercle of
labrum broad, rather globose, sometimes cleft apically
[southeastern Canada and northwestern United States
EosouthernMexd Go) precsneeeeeecs cere: Xylocopoides

Basitibial plate open at the base, not foveate; outer apex
of hind tibia with long curved spine; median tubercle
of labrum narrow (body elongate, subcylindrical) [Costa
Rica to southern Mexico] .......:.ccceeeeeeee Calloxylocopa

3 (1). Medial longitudinal carina on all metasomal sterna;
clypeus usually bounded by continuous impunctate
THUG IES ert eececepare nae iAceACCHAC A CI0O/ CSO ECOL EGE GEE CESE PEC OPPEEE 4

Medial longitudinal carina is not on all metasomal
sterna or entirely absent; clypeus bounded by discon-
tinuous impunctate ridge, or ridge is entirely absent .. 5

4 (3). Upper tooth of mandible much wider than lower
tooth (Fig. 1C) [southern Arizona in the United States
to southern Argentina] .........0:cceee Stenoxylocopa
Upper tooth of mandible as wide as or narrower than
lower tooth (Fig. 1 A) [southern South America to south-
western United States, introduced on Hawaii and

(GiGi Nail | Toekeceerenscace recor ence cocoon ornorcaoT Neoxylocopa
5 (3). Propodeal triangle present ...........:ccceeteeeeeeeeteee 6
Propodeal triangle absent [western United States to Co-
Vou ial] (ire scececsacocsce-sotccecsrsrsessezavescortars ecteoee? Notoxylocopa
6 (5). Labrum with 4 or fewer tubercles ...........:ese 7
Labrum with 5 tubercles (abundant long setae) [cen-
tral Argentina to southern Brazil] ................. Xylocopsis
7. (6) perontall (carinalel On gate yeccc-ccxsecsesesesssesseaseceseeesesrae™ 8

Frontal carina short, transverse, tuberculiform and not
extending between antennal sockets [central Argentina
CO) LoywieloerriOn |BN VAN cccrcececncococencpnnnnoten romeo Nanoxylocopa

8 (7). Clypeus at least half as long as broad; median clypeal
length longer than the distance between clypeus and
AMLEC TOI OCE Ills pereesee estes sseessesssnnesecureasn eesvansasssotsaecezearete 9
Clypeus less than half as long as broad; median clypeal
length shorter than the distance between clypeus and
BUMUHeTAIONe OLMIS) Sees cnosececo-peoc tec aonnbe nonccncocEECoOuE EE Kaas -BCCEH 10

28 ScIENTIFIC PAPERS, NATURAL History Museum, THE UNIVERSITY OF KANSAS

9 (8). Apex of S1 entire, not produced [south-central Bra-
Zilito paras uraiys| ee tees secerscssneeesceeeserce ees Cirroxylocopa
Apex of S1 distinctly produced [extreme southern
Brazil) fesse cos Neesssssssctecacececsvscseusecsvaceesesestavssess Dasyxylocopa

10 (8). Eyes more convergent above than below; body in-
tegument often brilliantly metallic [southern United
States toy Ace em tinal] eenesewtasecrsesesseeensencaseere Schonnherria
Eyes at least equally convergent above and below; body
integument at most feebly metallic in bright light .. 11

1

—

(10). Mandible with three teeth on apical margin;
parocular area with a large elongate depression adja-
cent to lateral sides of antennal sockets [south-central
Sail] eee cee ere ee See Son aad Rees Gone se ieee Diaxylocopa

Mandible with two teeth on apical margin; parocular
ANCA O tIMOGIedhaSTADOVE csrsseee-cesecetesesereereeeteeeeee ree 12

12 (11). Thickly plumose hairs on the disc of the first
metasomal tergum; eyes equally convergent above and
below [central Argentina to southern Brazil] ................
Sh aie Pi APE MER SS Roh it Pe mee TS BE Xylocopoda

Sparsely plumose hairs on the disc of first metasomal
tergum; eyes more convergent below than above [south-
Gentralubrazill|peecee eee eens Monoxylocopa

New Wor_b MALeEs
. Graduli on Tl, T2 or more terga; integument usually
non-metallic, sometimes very weakly metallic.......... 2

—

Graduli absent on T2-T5, integument usually brilliant
metallic (often with apex of basitibial plate symmetri-
cally or asymmetrically bifid) [southern United States
fOvATe emtihial| eee cee Schonnherria (in part)

2 (1). Apex of basitibial plate asymmetrically or symmetri-
calllyabiticl(HisyS AY IB AG) iene ctets: sete crete aercer. 3}

Apex of basitibial plate entire (Fig. 3D, E) .................. 4

3 (2). Vestiture chiefly black, integument on discs of T1-
T5 black and covered with dark setae [central Argen-
{AU YE) 1K) SLOVO nVSSeIAN IB FPEVATD | ececeeeconocecoccnesenenetaconece Xylocopsis

Vestiture chiefly brownish, integument bare on discs
of TI-T5 and largely eburneus [Argentina] ..................
seshu sessaus soy'esapvhatotaes odbinvessveséasensvespaists Schonnherria (in part)
4 (3). T7 without dentiform projections on apex ........... 5
T7 with dentiform projections on apex [western United
SHIH) HO) (ClolkKovam onl crceescecrcnoncecnnoneoncaneccncnee Notoxylocopa
5 (4). Anterior half of T3 broadly depressed on each side
of basally carinate median line anterior margin of each
depression bearing transverse row of very long, poste-
riorly directed plumose hairs which overlie the depres-
sion [Costa Rica to southern Mexico] ...... Calloxylocopa
INS} invore wanvoyshunrerel GIS BOONE care cocsreccnesen conoamoncreanensenenseeoS 6
6 (5). Hind tibia with spines or high, thin lamella on ven-
ital) surface ct i. ..Sccz, scesten-seastesotesetvouneveceeras ctietesat. Wterers 7

Jesbinnyel (done THX} e Ton (OLS NUGLETE | 55. ancoosoctooaceesonnoceacoareeccoserecktaas: 8

7 (6). Hind tibia with high, thin lamella beginning near
base of hind tibial spur and extending obliquely for-
ward across ventral surface of tibia [central Argentina
(HO) SLOG MESIAL JET) rocncenenosoococonaconsesocenocncceNe Xylocopoda
Hind tibia without lamella; inner hind tibial spur ab-
sent; apex of tibia beneath with large, flattened, highly
polished tubercle [south central Brazil to Paraguay]...
HS Sieh be lac setesd tone ives terete erste eeten once Cirroxylocopa

8 (6). One spine on outer apex of hind tibia ................... gy

Two spines on the outer apex of the hind tibia [south-
eastern Canada and northwestern United States to
southern Mexdco) eee. ane-mrenereveeeee corer Xylocopoides

9 (8). Disc of scutellum entirely subhorizontal; posterior
declivity of thorax beginning on the metanotum or
PYOPOU OUI vcs. cssitiscecaseesseserssesosetseseeste sree eteeeee 10

Disc of scutellum not entirely subhorizontal, posterior
declivity of thorax undefined (antennal scape yellow,
tegula punctate throughout or thickly pubescent
circummarginally) [central Argentina to southern Bra-
ZAll) eases. cas vs tao Ra eee Nanoxylocopa

10 (9). Integument mostly ferruginous or yellowish with
ferruginous or fulvous pubescence; tegula ferruginous;
scape entirely ferruginous or with longitudinal yellow
frontal stripe [southern South America to southwest-
ern United States, introduced on Hawaii and Guam]..
Leedisipiosvesdstessueeseistesetssvsendessssereresscosttetedecsseatrece Neoxylocopa

Integument mostly black at least on dorsum with black-
ish, whitish, or griseous pubescence; tegula black; scape
largely or entirely black, without a longitudinal yellow
frontal Stripe 2 sc.cc:decsscdeisetsiseeresesssscesshasssvesooseesneeeeeeeeee 11

11 (10). Face with yellow maculations extending to vertex
of head [southern Arizona in the United States to south-
enn Ane emtinall i. ticccscarctceccrsecoesctocnceseteecee: Stenoxylocopa

Face with yellow maculations extending to just above
antennal sockets [extreme southern Brazil] ..................
asaslsasvancavesnveossbetsentiessee store cetncasdrersre casemeaertere? Dasyxylocopa

OLD Wor.LD FEMALES (EXCLUDING LIEFTINCKELLA)

1. Beginning of thoracic declivity represented by a trans-
verse carina that divides the scutellum into a dorsal and
verticallisurfaces! si.5 4k ee ee 2.

Beginning of thoracic declivity represented by a gently
omabnupthysroumG ecllam ele rereccetescrcceeteest stents 3

2 (1). Body length generally greater than 25 mm [Africa,
Indian subcontinent to Indonesia and north to China]
BEEBE ECS EERO EERE EEIIC EEG POSE CEE RPE ISRC CEL BU TIO SACRE Mesotrichia

Body length generally less than 25 mm [Africa, east to
the Pacific, southern Australia to northern China] ......
Koptortosoma

SUBGENERA AND GENERA OF CARPENTER BEES, X YLOCOPINI 29

3 (1). Midtibia with a spine on outer apeX ....... ee 4

Midtibia lacking spine on outer apex (anterior ocellus
bounded on either side by an impunctate, crescent-
shaped swelling) [Taiwan and southern China, to In-
donesia and west to Indian subcontinent)......... Biluna
4 (3). Pygidial plate with lateral margins not strongly di-
vergent anteriorly, represented by a spine that often has
a shorter preapical spine on each side of its base and thus
not or scarcely represented on disc of T6 (Fig. 7B, C) .. 5
Pygidial plate with lateral margins strongly divergent
anteriorly on disc of T6, more than seven times wider
anteriorly than posteriorly [Asia Minor to Turkestan

© DENTAL lee boca eae eeseeneeeeereer oreeaatoeee a ecco Proxylocopa
5 (4). Outer apex of the hind tibia with one spine or con-
Wee? ENS, 18} (C,, 1D), 18) oases cecconrogecosoncecacoascuscasseonececssoneKe 6
Outer apex of the hind tibia with two spines or con-
BERSSCBE Es (ESPN) aceon ces ones coor acreen ecarsars rarer encase venccoters 12
6 (5). Ocelli normal, diameter much less than twice the
WWALCIN GENES SEZ SD eco cerccpnnecocccosooepascooecacoce coc oceecoceooo ASSEN 7

Ocelli enlarged, diameter greater than twice the width
of the scape (basitibial plate bifid; integument reddish-
brown to black with ferruginous pubescence) [Indian
SUDCOMEMENt CO Pavel | ecceeece-cceszestecesnestecesssens Nyctomelitta

7 (6). Pygidial spine with a preapical spine on each side
Flt SENSO (CENS TAQ) cares seen besscennncocecorecotnscee coop oocanae caper Bos0e 8

Pygidial spine without a preapical spine on each side
at base (Fig. 7B) [Europe to central Asia] ....... Copoxyla

8 (7). Mandible with two teeth on apical margin (Fig. 1B) .. 9
Mandible with three teeth on the apical margin (Fig. 1A) 11

9 (8). Scutellum modified, forming a conspicuous poste-
riorly-directed lobe that overhangs entire metanotum
[Sri Lanka and peninsular India] ..............00+ Nodula

Scutellum not forming a conspicuous posteriorly-di-
TREE RETG NCO) SY oon ane cee et conn sos ooo ao sae aos E IGS oO SATE EOS 10

10 (9). Parocular area without tubercle; maxillary palpus
six-segmented [northwestern India to southern
Martie tern eee ae ie ee OS Ro Maaiana

Parocular area with tubercle below antennal socket;
maxillary palpus four-segmented [northern and east-
RICANS Erallitaill |. cerns seas oe ea aes een osc cease Lestis

11 (8). Mandible with hook-like process on the lower edge
(Fig. 1B) [Africa south of the Sahara] .. Gnathoxylocopa
Mandible without hook-like process (basitibial plate
with tubercles along both lateral edges (Fig. 2B)) [Af-
rica south of the Sahara, Turkestan region to the Indian
SUL COMELINGMAE oes sae esses cots evcee eens Ctenoxylocopa

12 (5). Minimum length of malar space less than length of
second flagellar segment; one or both lateral edges of
the basitibial plate crenulate (Fig. 2B) ...........::cce 13

Minimum length of malar space more than length of
second flagellar segment; lateral edges of the basitibial
plate entire (Fig. 2C, E) [northeastern China and Ko-
Teal | RO eee ee ee ee ee ee Bomboixylocopa

13 (12). Beginning of thoracic declivity represented as an
abrupt rounded angle in scutellar profile................. 14
Beginning of thoracic declivity broadly rounded, i.e. not
defined, thoracic shape being evenly rounded in pro-

fit ere vane aestisctein Hate Beekess cease reeset shes RE ee WZ,
14 (13). Pygidial spine with a preapical spine on each side
atibase:(Eie87 ]))aiiscsrcssesessecsrsvscsncrecvieteseceecveuseadesesnertees 15

Pygidial spine without a preapical spine on each side
at base (Fig. 7B) (lateral line of T1 transverse; vertical
fold of anterior face of T1 broad and deep) [Middle Hast
(HO) OWI NSreIM (CUMIAEII| coco emncececesensencconcnonaredynocce Alloxylocopa

15 (14). Anterior face of Tl with medial, vertical sulcus
that is broad and deep; lateral line of T1 transverse [Af-
rica south of the Sahara] ............:::se0 Xenoxylocopa

Anterior face of T1 lacking a vertical sulcus altogether
or, if present, sulcus not as well developed as above;
posterior part of the lateral line of T1 curved posteriorly,
more or less parallel to lateral margin of metasoma .. 16

16 (15). Apex of basitibial plate bifid, sometimes feebly so
(Fig. 2A, B, C) [Tibet through the tropical Orient to
Stal esislGe] Meese ersrnnsearsssnsearasesteratscereassearee Zonohirsuta

Apex of basitibial plate simple (Fig. 2D, E) (supraclypeal
area strongly swollen, bigibbose) [Madagascar] ..........
Eocene ER Eree POTTER ea ere Prosopoxylocopa

17 (13). Length of metanotum greatest medially .......... 19

Length of metanotum subequal from side to side [Medi-
terranean Basin to central Asia] ..............:00+. Xylocopa

18 (17). Discal hairs of T1 simple or sparsely plumose; first
flagellar segment shorter than combined lengths of suc-
ceeding three segments; medial groove of propodeum
distinct [Attri calle iccsssstsvtcsere eee eeseoee Xylomelissa

Discal hairs of T1 thickly plumose; first flagellar seg-
ment equal to or longer than the succeeding three seg-
ments; medial groove of propodeum indistinct [north-
ern Africa, southern Europe into Asia Minor) ..............
Bae siaed aeeidnees ie cde tvaoks ioe ec tiniens Sveti ssasss estat Saree Rhysoxylocopa

Otp Wortp Mates (EXCLUDING LIEFINCKELLA)
1 Basitibial plate absent, or its apex entire (Fig. 2E)...... 2
Basitibial plate present, apex asymmetrically or sym-
metnicalllybitidl (Bic AWB AG) cesrsrsesceresssasceeessres:sertereeee 3

2 (1). Dorsal surface of Tl angulately (or subangulately
rounded in Alloxylocopa) onto anterior surface; lateral
Nimetot Mil tran SVCUSC lec scereeetesearcs ss secesccerescsronsenarscteenertentte 5

Dorsal surface of T1 sloping or rounding onto declivous
anterior surface; lateral line of T1 curved posteriorly,

30 SCIENTIFIC PAPERS, NATURAL History MuseuM, THE UNIVERSITY OF KANSAS

more or less parallel to lateral margin of metasoma . 8

3 (2). Ocellus normal, diameter less than twice the width
Of tNEISCAP CEs ece cess cocevessssssustecas csssesssiseetonectorstestesesrseness seo: 4
Ocellus enlarged, diameter greater than twice the width
of the scape [Indian subcontinent to Java] .. Nyctomelitta

4 (3). Beginning of posterior thoracic declivity on scutel-
lum (posterolateral lobe of pronotum prolonged poste-
riorly so as to extend well onto mesepisternum and
nearly attaining preepisternal groove) [Africa south of
the Sahara, Turkestan region to the Indian subcontinent]

Ctenoxylocopa

Beginning of posterior thoracic declivity on metanotum
[Africa south of the Sahara] ................... Gnathoxylocopa

5 (2). Tegula normal, outer margin convex, rounded be-
hind (Fig. 4A), posterolateral third may or may not be
impunctate, glabrous and Shining .........cccceeree 6
Tegula elongate, outer margin at least slightly concave

(Fig. 4B), posterolateral third impunctate, glabrous, and
shining [Africa, Indian subcontinent to Indonesia and

TANGY AY HOM Gl qlitin all | saoneqnete ceccconccoee eet eacae are aeeroCaane Mesotrichia
6 (5). Dorsal surface of T1 nearly horizontal, abruptly and
angulately divided from anterior surface ...............0+. TU

Dorsal surface of T1 sloping forward, subangulately
rounding onto anterior surface (tegula punctate
throughout) [Middle East to southern China] ..............
Bee Pea eC Ooo nD NOT NOD NERC EE COLO GEOEE Alloxylocopa
7 (6). Anterior surface of T1 without a fovea [Africa south
Oi HN SENN NCEII| ccoseccacocnnccescoocaoceetocconoconoaecoac6e Xenoxylocopa

Anterior surface of T1 with a fovea [Africa, east to the
Pacific, Australia to northern China]........ Koptortosoma

8 (2). Outer apex of mid tibia with a spine ................... g

Outer apex of mid tibia without a spine (anterior ocel-
lus bounded on either side by impunctate, crescent-
shaped swelling) [Taiwan and southern China, to In-
donesia and west to Indian subcontinent]......... Biluna

9 (8). Outer apex of the hind tibia with one spine (Fig. 2B,
C7 MD PAE) ose ccsssreus sts ciusthsceuesstoarsataceeweesvers iin aosescst ses 10

Outer apex of the hind tibia with two spines (Fig. 2A)... 20

10 (9). Minimum malar length shorter than second flagel-
lar segment; face entirely black or with yellow macula-
tions only on areas below the lateral ocelli............... 11

Minimum malar length as long as length of second
flagellar segment; face with two parallel stripes run-
ning along the inner margins of the eyes from the ver-
tex to the bases of the mandibles (entire metanotum
exposed dorsally and base of propodeum horizontal,
ali Iisecie leverelllly)) OUI cecoscrscecconeerncennonsonnocosoce sacotorcce const
sessbbuonten Homan Bomboixylocopa, in part (X. rufipes)
11 (10). Flagellum cylindrical; maxillary palpus five or six
segmented; second submarginal cell of forewing much

longer posteriorly than anteriorly ...........:.:0cceeee 12
Flagellum crenulate on inferior surface; maxillary pal-
pus four-segmented; second submarginal cell of forew-
ing nearly in form of parallelogram (integument

brightly metallic) [Australia] .............:ceeeeeees Lestis
12,0); Wabrum apically entire’ -cc-ncaieecscssceccscrseeee eee 13
Labrum apically emarginate .............cccsssesceteeseeeeneees 14

13 (12). Eyes closest medially, inner orbits convex, bowed
towards antennal sockets; frontal carina sulciform, not
elevated at distal apex; beginning of thoracic declivity
represented as abruptly rounded angle in scutellar pro-
tale) [Miacla'eas cat] Peessersessesesescseenee ser aeeeerees Prosopoxylocopa

Inner orbits of eyes concave, curved inward towards
central eye axis; frontal carina cariniform; scutellum
rounded in profile [northeastern China and Korea] ....
Span cbbbbo Sena aoencacane ACASSEsOTOSaSOSESOSNCEEDS Bomboixylocopa (in part)

TES (CUD), Oly SVSTWIS LO EVES cocrcnccontensecmnccconocacatotexSrocncesccoscs5so%0° 15
Clypeus with yellow maculations ...........::ccceee 17
TINS) (CAD), Nore (OSE T/A ec crcooecconcocncocececocoso cosets soon se 3020.90" 16
Apex of T7 with pair of dentiform spines [Europe to
céntral Asiall.e..2.. 200 ee eee Copoxyla

16 (15). Propodeal spiracles small and asymmetrical; mini-
mum interorbital distance equal to or greater than
length of antennal scape [northern Africa, southern
Europe into Asia Minor] ..........:.00cee Rhysoxylocopa

Propodeal spiracles normal, not modified; minimum
interorbital distance shorter than length of antennal
Scape) [Attia call exces eeeee see sacens Xylomelissa (in part)

17 (14). Beginning of thoracic declivity indicated by an
abruptly rounded angle in scutellar profile ............. 18

Beginning of thoracic declivity not defined, thorax
evenlyaround clini toil Cleecnrssensessrssneerssatereeeeseeeee 19

18 (17). Posterolateral third of tegula impunctate, glabrous
and shining (northwestern India to southern Turkestan)
seuivileasdes ahaa ease rent neuer Save aeons onaeaee eee Maaiana

Tegula punctate throughout [Tibet through the tropi-
cal Orient to Sulawesi] ..........:.:ccccecseeeeeeeeees Zonohirsuta

19 (17). Hind femur with tubercle on ventral surface; lat-
eral ocellus not bounded laterally by conspicuous trans-
verse ridge (Africa) ...........:0 Xylomelissa (in part)

Hind femur without tubercle; lateral ocellus often
bounded laterally by conspicuous transverse ridge (Sri
Lanka and peninsular India) «........:::ceceeee Nodula

20 (9). Propodeal triangle defined [Mediterranean Basin
and central Europe to central Asia] ...........cceeeeeeeeee
nates diorcvattasvartessirt tenercarte cTotoreere eee Xylocopa (sensu stricto)

Propodeal triangle absent [Asia Minor to Turkestan re-
LOM Meee seeetecece crt aceestanoeeesnncrerseattettereecenererste Proxylocopa

SUBGENERA AND GENERA OF CARPENTER BEES, XYLOCOPINI 3]

DIAGNOSES AND DESCRIPTIONS OF SUBGENERA OF XYLOCOPA

The following descriptions are presented in the order
used for characters in the Character List and Coding sec-
tion; numbers are the same as those used there. Unique
apomorphies are in bold face type.

Subgenus Xylomelissa Hurd and Moure

Xylomelissa Hurd and Moure, 1963: 219. Type species: Xylocopa carinata
Smith, 1874, by original designation.

Epixylocopa Hurd and Moure, 1963: 223. Type species: Xylocopa rufitarsis
Lepeletier, 1841, by original designation. New synonymy.

Apoxylocopa Hurd and Moure, 1963: 226. Type species: Xylocopa lugubris
Gerstaecker, 1857, by original designation. New synonymy.

Dinoxylocopa Hurd and Moure, 1963: 230. Type species: Xylocopa absurdipes
Enderlein, 1903, by original designation. New synonymy.

Euxylocopa Hurd and Moure, 1963: 232. Type species: Xylocopa fraudulenta
Gribodo, 1894 by original designation. Synonymy with Xylomelissa
by Maa, 1968: 106.

Perixylocopa Hurd and Moure, 1963: 232. Type species: Xylocopa erythrina
Gribodo, 1894: 232, by original designation. Synonymy with
Xylomelissa in part, by Maa, 1968:106.

Acroxylocopa Hurd and Moure, 1963: 236. Type species: Xylocopa capitata
Smith, 1854, by original designation. New synonymy.

This subgenus is closely related to Nodula, Maaiana, and
Rhysoxylocopa. Males can be distinguished from all other
members of the tribe by the combination of the complete
lack of spines on the outer apex of the hind tibia, (although
this is flared in some taxa) and the presence of a basitibial
plate. Females are similar to those of Rhysoxylocopa but can
be distinguished by the presence of a distinct medial groove
on the propodeum. All species are African. In some analy-
ses presented here the Oriental subgenus Biluna was placed
in this group, although, as discussed below, Biluna is prob-
ably not a member of the clade. Xylocopa absurdipes, the
only species placed in Dinoxylocopa and known only from
males, shares all of the apomorphies of this group and so
should be included in Xylomelissa. Maa (1968: 106) effec-
tively synonymized Perixylocopa with Xylomelissa in his
species citation, even though in the discussion he stated
that he left the subgenus Perixylocopa intact.

Description—(1) Color of vestiture on male dorsum
predominately dark brown or black. (2) Upper carina of
male mandible without a tooth, except in Xylocopa erythrina.
(3) Upper carina of female mandible without a tooth, ex-
ceptin X. erythrina and X. rufitarsis. (4) Two teeth on apical
margin of female mandible except for X. capitata which
has three teeth. (5) Upper rutellar tooth and preapical tooth
of female mandible separated by emargination. (6) Female
epistomal suture smooth, without a ridge, except in X.
capitata and X. erythrina. (7) Supraocellar pits of female
punctiform. (8) Female malar space shorter than width of
first flagellar segment, except in X. capitata. (9) Ocelli not
enlarged. (10) Male with yellow facial maculations, except
in X. rufitarsis. (11) Outer apex of male hind tibia without
spines. (12) Outer apex of female hind tibia with one spine.
(13) Tibial spine of female originating at tibial apex and

without a medial carina, except in X. albifrons. (14) Hind
trochanter of male subcylindrical, not triangular. (15) Male
mid- and hind trochanter not modified. (16) Male hind fe-
mur with basal tubercle, except for X. rufitarsis and X.
albifrons. (17) Posterior edge of female basitibial plate with
a row of tubercles. (18) Apex of male basitibial plate en-
tire. (19) Apex of female basitibial plate bifid. (20) Female
basitibial plate open basally. (21) Female basitibial plate
reduced, not extending to near base of tibia. (22) Thoracic
profile of male smoothly rounded, beginning of thoracic
declivity undefined. (23) Thoracic profile of female
smoothly rounded, beginning of thoracic declivity unde-
fined. (24) Male metapostnotum absent, except in X.
rufitarsis. (25) Female metapostnotum present except in X.
capitata and X. rufitarsis. (26) Mesosomal gland reservoir a
single layer of parallel running tubules. (27) Beginning of
thoracic declivity of male on scutellum. (28) Propodeum
at its connection with metasoma not modified. (29) Male
tegula not elongated posteriorly. (30) Gradulus on T1 con-
tinued posteriorly adjacent to lateral margin of the
metasoma. (31) Graduli present on T1 to T5 of male. (32)
Gradulus only on T1 of female except in X. erythrina which
has graduli on T1-T5. (33) Anterior surface of T1 of male
without a fovea. (34) Anterior surface of T1 of female with-
out a foveal pit or entrance to acarinarium. (35) Medial
groove of T1 of female linear. (36) Apical margin of S1 en-
tire, not produced except in X. rufitarsis and X. erythrina,
in which it is medially emarginate. (37) Medial carina re-
stricted to posterior half or absent on S6 of male. (38) Me-
dial carina on sternum of females weakly developed. (39)
Pregradular areas on S3-S5 of males with paired patches
of long setae. (40) Patches of appressed setae present near
spiracles on T2-T6 of males. (41) Pygidial spine of female
with pair of preapical lateral spines. (42) Gonostylus of
male genitalia distinct although fused to gonocoxite. (43)
Outer edge of penis valve not modified, without lateral
projection. (44) Penis valve apex not hirsute. (45) Penis
entirely membranous. (46) Spatha absent. (47) Gonostylus
broadly attached to gonocoxite. (48) Gonostylus broadly
attached to gonocoxite and apically with dense patch of
setae. (49) Gonostylus apex with weakly developed, me-
dially projecting lobe. (50) Medial ventral margin of
gonocoxites diverging continuously from dorsal gonocoxal
bridge or nearby. (51) Gonocoxite lacking a projection or
spine from posterior ventral margin. (52) Ventroapical plate
of gonocoxite rounded, with a carina. (53) Medial projec-
tion or spine at base of ventral gonocoxite absent. (54) Apex
of T7 of male entire. (55) Penis valves apically expanded.

Subgenus Nodula Maa

Nodula Maa, 1938: 290. Type species: Apis amethystina Fabricius, 1793, by
original designation.

bo
NO

Maa (1938) noted that this subgenus is represented by
two “natural groups;” exemplars of both were included in
my analysis. I found no characters that define members of
this subgenus as a monophyletic group, although the male
genitalia are phenetically similar. The evidence suggests
that Nodula, as presently recognized, is either a paraphyletic
or polyphyletic assemblage and I have recognized the
group typified by X. bentoni as a separate subgenus
Maaiana. This leaves Nodula consisting of the following
species: X. amethystina, X. madurensis, X. meyeri, X.
nigrotarsata, X. phanerocephala, X. prashadi, X. ramakrishnat
and X. remota. A key to them and those in the subgenus
Maaiana, modified from Maa (1938 and 1954), was given
in Hurd and Moure (1963). Nodula is closely related to
Maaiana, Xylomelissa (s. lat.) and Rhysoxylocopa. The males
of Nodula can be distinguished from those of Xylomelissa
by the presence of a spine on the outer apex of the hind
tibia of the male and from those of Rhysoxylocopa by the
unmodified propodeal spiracles. The females of Nodula can
be distinguished from those of the subgenus Rhiysoxylocopa
by carinae dorsal to the lateral ocelli and a sharply rounded
scutellum.

Description.— As in Xylomelissa, except as follows: (11)
Outer apex of male hind tibia with one spine. (12) Outer
apex of female hind tibia with two spines. (16) Male hind
femur without basal tubercle or spine. (22) Beginning of
posterior thoracic declivity of male abruptly rounded. (23)
Beginning of posterior thoracic declivity of female abruptly
rounded. (24) Male metapostnotum evident. (26)
Mesosomal gland reservoir absent. (49) Gonostylus apex
entire, without medially projecting lobe.

Subgenus Maaiana new subgenus
Type species.—Xylocopa bentoni Cockerell 1919: 172.

The relationship of this group to Nodula and allied sub-
genera is discussed above under Nodula. As defined herein,
Maaiana consists of the following species; X. angulosa, X.
bentoni, X. bicristata, X. punctigena and X. punctilabris. A
key to the species, as the punctigena group of Nodula, was
presented by Maa (1954) and in modified form by Hurd
and Moure (1963). Males of Maaiana can be distinguished
from those of Nodula by the abrupt beginning of the poste-
rior thoracic declivity on the scutellum, from Xylomelissa
by the presence of a spine on the outer apex of the hind
tibia, and from Rhysoxylocopa by the unmodified propodeal
spiracles. Females of Maaiana can be distinguished from
those of the subgenus Rhysoxylocopa by the presence of a
carina dorsal to each lateral ocellus and a sharply rounded
scutellum.

Etymology.—The name Maaiana is in recognition of T.
C. Maa for his important work on Old World Xylocopini.

ScIENTIFIC PAPERS, NATURAL History Museum, THE UNIVERSITY OF KANSAS

Description.—As in Xylomelissa, except as follows: (11)
Outer apex of male hind tibia with one spine. (13) Hind
tibial spine of female strongly developed. (16) Male hind
femur without basal tubercle or spine. (26) Mesosomal
gland reservoir present. (49) Gonostylus apex entire, with-
out medially projecting lobe.

Subgenus Biluna Maa
Biluna Maa, 1938: 270, 276. Type species: Xylocopa nasalis Westwood, 1842,
by original designation.

This subgenus is a group of large, elongate bees that, as
far as known, nest solely in bamboo. Associated with the
very smooth inner surface of this nesting substrate are
morphological modifications of the legs and mandibles.
Males can be distinguished from those of all other taxa by
the combination of lacking a basitibial plate and of spines
on the outer apex of the tibia. Females can be distinguished
by two features of the mid-tibia: a dense mat of short, stout
setae and absence of spines on the outer apex. Hurd and
Moure (1963) considered Biluna to be closely related to the
subgenus Xylocopa. In my analyses, this subgenus is clearly
allied with a clade consisting of Xylomelissa, Rhysoxylocopa,
and Nodula. In most analyses, Biluna is shown as the most
basal taxon of this clade.

Description.—As in Xylomelissa, except as follows: (17)
Basitibial plate of female with edges slightly raised from
surface of tibia. (18) Basitibial plate of male absent. (19)
Apex of female basitibial plate entire. (25) Metapostnotum
of female absent. (27) Beginning of thoracic declivity of
male on propodeum in X. tranquebarorum, metanotum in
X. nasalis and X. auripennis. (39) Pregradular areas of male
sterna without lateral setal patches. (49) Gonostylus en-
tire, without a medial lobe.

Subgenus Rhysoxylocopa Hurd and Moure
Rhysoxylocopa Hurd and Moure, 1963: 178. Type species: Xylocopa cantabrita

Lepeletier, 1841, by original designation.

This group is closely related to Xylomelissa, Nodula,
Maaiana and Biluna. Females of Rliysoxylocopa can be dis-
tinguished from those taxa by the absence of a medial
groove on the propodeal triangle. Males are easily distin-
guished from all other taxa by the unusual propodeal spi-
racles which Hurd and Moure (1963) described as “small,
asymmetrical, lower extremity broadened, anterior mar-
gin, as determined by deeply impressed line, forming an
elongate triangular flap whose lower anterior margin
projects strongly forward.”

Description.—As in Xylomelissa, except as follows: (10)
Male face entirely black. (11) Outer apex of male hind tibia
with one spine. (13) Female hind tibia with tibial spine
strongly developed, originating well before tibial apex and
with medial carina. (26) Male without a mesosomal gland

SUBGENERA AND GENERA OF CARPENTER BEES, XYLOCOPINI 33

reservoir. (39) Pregradular area of male without lateral setal
patches. (40) Male without patches of short setae near
metasomal spiracles. (49) Gonostylus apex entire, without
medially projecting lobe.

Subgenus Proxylocopa Hedicke, New status
Xylocopa (Proxylocopa) Hedicke, 1938: 192. Type species: “X. oliviert Bril.”

(=Xylocopa olivieri Lepeletier, 1841), by original designation.
Proxylocopa (Proxylocopa) Maa, 1954: 116: 190, 193.

Ancylocopa Maa, 1954: 190, 198. Type species: Xylocopa nitidiventris Smith,

1878, by original designation. New synonymy.

This subgenus is notable in having a distinctively long
face in females as compared to other members of the tribe;
this feature probably is associated with loss of the strong
mandibular musculature for gnawing in wood. Females
are easily distinguished by the well-developed basitibial
plate located at the base of the hind tibia and by the well-
developed pygidial plate. Males can be recognized by the
set of characters including short, ovate, parapsidal lines,
no more that three times as long as wide, and two spines
on the outer apex of the hind tibia. The recognition of
Proxylocopa as a genus requires that it be the sister group
to the rest of the genus Xylocopa. If Proxylocopa is more
closely related to any of the derived subgenera of Xylocopa,
the recognition of Proxylocopa as a genus makes Xylocopa
paraphyletic. As discussed above, the distinctiveness of
Proxylocopa lies primarily with the structures associated
with soil nesting, and there is morphological and phylo-
genetic evidence that these structures have evolved sec-
ondarily. Those characters of Proxylocopa that I can homolo-
gize with structures of other Xylocopini are not generi-
cally distinct from the other taxa in the tribe. The male
genitalia suggest that this group is most closely related to
Copoxyla and/or the clade of Rhysoxylocopa, Biluna, Nodula,
Maaiana, and Xylomelissa. I favor emphasizing the homo-
geneity of the tribe, and create an unequivocally mono-
phyletic genus Xylocopa by assigning subgeneric rank to
Proxylocopa. Furthermore, it is obvious that there is no clear
distinction between the two presently recognized subgen-
era, Proxylocopa (sensu stricto) and Ancylocopa. Ocellar di-
ameter and length of the malar space were the primary
characters used by Maa (1954) to distinguish these groups.
Based on my examination of P. rufa and P. olivieri of
Proxylocopa (sensu stricto) and all species of Ancylocopa,
the ocellar diameters and the malar space grade continu-
ously together. Therefore, I synonymize Ancylocopa with
the subgenus Proxylocopa.

Description.—As in Xylomelissa, except as follows: (9)
Ocelli enlarged in some species, diameter greater than that
of antennal socket. (11) Outer apex of male hind tibia with
two spines. (16) Male hind femur without basal tubercle
or spine. (25) Female metapostnotum absent. (26)
Mesosomal gland reservoir consisting of two small, widely

separated pouches. (32) Graduli on T1-T5 of female. (39)
Pregradular areas of male sterna without setal patches. (40)
Male without patches of short appressed setae near
metasomal spiracles. (49) Gonostylar apex entire, without
medially projecting lobe. (52) Ventroapical plate of
gonocoxite absent.

Subgenus Nyctomelitta Cockerell
Nyctomelitta Cockerell, 1929: 303. Type species: Bombus tranquebarica Fab-
ricius, 1804, by original designation.

Both sexes of the species of Nyctomelitta can be distin-
guished from all other members of the genus Xylocopa by
their large size and greatly enlarged ocelli. The phyloge-
netic relationship of this group is not well established by
this study. Although seemingly allied to a clade consisting
of the majority of New World species, some structural fea-
tures and its biogeography suggest that this group is closely
related to Rhysoxylocopa and its allies. Examination of other
species in the subgenus should help to clarify these matters.

Description.—As in Xylomelissa, except as follows: (2)
Upper carina on male mandible with subapical tooth. (3)
Upper carina of female mandible with subapical tooth. (9)
Ocelli enlarged, diameter greater than that of antennal
socket. (11) Outer apex of male hind tibia with one spine.
(16) Male hind femur without basal spine or tubercle. (18)
Apex of male basitibial plate bifid. (26) Mesosomal reser-
voir of male present. (32) Female with graduli on T1-T5.
(39) Pregradular areas of male sterna without setal patches.
(40) Male without patches of short appressed setae near
metasomal spiracles. (43) Penis valve with lateral projec-
tion. (50) Medial ventral margins of gonocoxites roughly
parallel distad from ventral gonocoxal bridge. (52)
Ventroapical plate of gonocoxite weakly developed.

Subgenus Cfenoxylocopa Michener

Ctenopoda Maa, 1938: 270, 285 (preoccupied by Ctenopoda McAtee and
Malloch, 1933). Type species: Apis fenestrata Fabricius, 1798, by origi-
nal designation.

Ctenoxylocopa Michener, 1942: 282, new name for Ctenopoda Maa, 1938.

Baana Sandhouse 1943: 530, new name for Ctenopoda Maa, 1938.

Males of this subgenus are distinctive; the posterolat-
eral lobes of the pronotum are prolonged posteriorly so
that they are nearly in line with the preepisternal groove
on the mesepisternum, and the spiracles on T3 bear an el-
evated, posteriorly directed, scalelike process. Females can
be distinguished from all other groups by the combination
of: (1) the well-developed basitibial plate with a row of tu-
bercles along each lateral edge, (2) mandible with two teeth
on its apical edge, and (3) hind tibia with one spine on its
outer apex. Ctenoxylocopa is apparently most closely related
to some members of the subgenus Xylocopa (sensu stricto).

Description.—As in Xylomelissa, except as follows: (4)
Apical edge of female mandible tridentate. (10) Male face

34 SCIENTIFIC PAPERS, NATURAL History Museum, THE UNIVERSITY OF KANSAS

entirely black. (11) Male hind tibia with one outer apical
spine. (16) Male hind femur without basal tubercle or spine.
(17) Anterior and posterior edges of female basitibial plate
each with row of tubercles. (18) Apex of male basitibial
plate bifid. (21) Female basitibial plate extending to near
base of tibia. (24) Male metapostnotum present. (31) T1
and T2 of male with graduli. (36) Apical margin of S1 me-
dially emarginate. (39) Pregradular areas of male sterna
without paired setal patches. (48) Gonostylus slender and
digitiform, with very few apical setae. (49) Gonostylar apex
without medially projecting lobe. (52) Ventroapical plate
of male gonocoxite absent. (54) Male T7 with pair of apical
dentiform processes. (55) Penis valves slender.

Subgenus Xylocopa Latreille (sensu stricto)

Xilocopa Latreille, 1802:431. Rejected by Internatl. Comm. Zool. Nomencl.,
Opinion 743, 1965.

Xylocopa Latreille, 1802:379. Emended name placed on Official List of
Generic Names in Zoology by Internatl. Comm. Zool. Nomencl.,
Opinion 743, 1965. Type species: Apis violaceae Linnaeus, 1758, by sub-
sequent designation of Westwood, 1840:86.

The subgenus Xylocopa is in need of further study be-
fore its phylogenetic affinities can be clarified. The two
species used in this study suggest that this subgenus is
paraphyletic or polyphyletic; X. violacea shows clear affini-
ties with the New World subgenus Xylocopoides, whereas
X. valga probably is most closely related to the Old World
subgenus Ctenoxylocopa. Females of both species can be
distinguished from all other members of the tribe by the
combination of the two spines on the outer apex of the
hind tibia and the tridentate apical edge of the mandible.
Males can be distinguished from all other Old World
groups by the combination of face without yellow macu-
lations, the rounded profile of the posterior part of the tho-
rax, and the two spines on the outer apex of the hind tibia.

Description.—As in Xylomelissa, except as follows:
(3) Upper carina of female mandible without a tooth, ex-
cept in X. valga. (4) Apical edge of mandible of female tri-
dentate. (10) Male face entirely black. (11) Outer apex of
male hind tibia with two spines. (12) Outer apex of female
hind tibia with two spines. (16) Male hind femur without
basal tubercle or spine. (17) Lateral edges of female
basitibial plate each with a row of tubercles. (18) Apex of
male basitibial plate bifid. (21) Female basitibial plate ex-
tending to near base of tibia. (24) Male metapostnotum
present. (26) Mesosomal gland present. (32) Graduli
present on T1—-T5 of female. (35) Medial groove linear on
T1 of female except in X. valga. (36) Apical margin of S1
medially emarginate. (39) Pregradular areas on male sterna
without paired setal patches. (40) Metasomal spiracle with-
out associated setal patch. (42) Gonostylus of male genita-
lia completely fused to gonocoxite in X. violacea; fused but
recognizable, long, slender, and with few setae in X. valga.
(48) Gonostylus slender and digitiform, with very few se-

tae at apex. (49) Gonostylar apex without medially pro-
jecting lobe. (52) Ventroapical plate of gonocoxite absent.
(53) Mesal margin of venter of gonocoxite with tooth or
projection. (55) Penis valves slender.

Subgenus Xylocopoides Michener
Xylocopoides Michener, 1954: 105. Type species: Apis virginica Linnaeus,

1771, by original designation.

Females of the subgenus Xylocopoides uniquely possess
two rows of numerous preapical spines delineating the
pygidial plate on the surface of T6 and a foveate basitibial
plate. Males can be distinguished from those of all other
Xylocopini, except X. violacea, by the two spines on the
outer apex of the tibia, the dark blue metallic coloration,
and the lack of a distinct gonostylus. Although I have left
Xylocopoides and Xylocopa (sensu stricto) as separate groups,
both share several highly derived synapomorphies and
Xylocopa (sensu stricto) probably is paraphyletic with re-
spect to Xylocopoides.

Description.—As in Xylomelissa, except as follows: (3)
Upper carina of female mandible without tooth except in
X. cyanea. (8) Malar space of female shorter than width of
flagellar segment 1 except in X. cyanea. (10) Male with yel-
low facial maculations except most specimens of X.
californica arizonensis. (11) Outer apex of male hind tibia
with two spines. (12) Outer apex of female hind tibia with
two spines. (16) Male hind femur without basal tubercle
or spine. (17) Lateral edges of female basitibial plate with-
out tubercles. (18) Apex of male basitibial plate bifid. (20)
Female basitibial plate closed basally. (21) Female basitibial
plate reduced, not extending to base of tibia. (24) Male
metapostnotum present. (26) Male mesosomal gland
present. (31) Graduli on T1 and T2 of male. (32) Gradulus
present on T1-T5 of female. (36) Apical margin of S1 me-
dially emarginate. (39) Pregradular areas on sterna of males
without paired setal patches. (40) Setal patches absent near
metasomal spiracles of male. (41) Pygidial plate margined
by two rows of many preapical lateral spines converging
at base of pygidial spine. (42) Gonostylus completely fused
to, and indistinct from, gonocoxite. (44) Penis valve apex
hirsute. (48) Gonostylus slender and digitiform, with very
few apical setae. (49) Gonostylar apex without medially
projecting lobe. (50) Medial ventral margins of gonocoxites
roughly parallel beyond ventral gonocoxal bridge. (53)
Ventral surface of gonocoxite with basal angle or spine.
(52) Ventroapical plate of male gonocoxite absent. (55) Pe-
nis valves slender.

Subgenus Calloxylocopa Hurd and Moure
Calloxylocopa Hurd and Moure, 1963: 142. Type species: Xylocopa tenuata
Smith 1874, by original designation.

This is a monotypic subgenus and probably derived
from Xylocopoides. Females share with the subgenus

SUBGENERA AND GENERA OF CARPENTER BEES, XYLOCOPINI 35

Xylocopoides the numerous preapical spines margining the
pygidial plate, but can be distinguished from that group
by the basitibial plate, which is open basally and barely
rises above the leg surface. Males are easily distinguished
by the third sternum, which is depressed on the anterior
half, forming two deep depressions, and covered by long
hairs that originate on its anterior margin.

Description.—As in Xylomelissa, except as follows: (10)
Face of male entirely black. (11) Outer apex of male hind
tibia with one apical spine. (16) Male hind femur without
basal tubercle or spine. (17) Both lateral edges of basitibial
plate barely rising above tibial surface, without tubercles.
(18) Apex of male basitibial plate bifid. (21) Female
basitibial plate extending nearly to base of tibia. (24) Male
metapostnotum present. (26) Reservoir of mesosomal
gland of male absent. (31) Graduli on T1-T3 of male. (32)
Graduli on T1-T5 of female. (36) Apical margin of S1 me-
dially emarginate. (39) Pregradular areas on sterna of males
without paired setal patches. (40) Male metasomal spiracles
without nearby setal patches. (41) Pygidial plate margined
by two rows of numerous preapical lateral spines converg-
ing at base of pygidial spine. (42) Gonostylus of male geni-
talia fused completely to, and indistinct from, gonocoxite.
(44) Penis valve apex hirsute. (48) Gonostylus slender and
digitiform, with very few apical setae. (49) Gonostylar apex
without medially projecting lobe. (50) Medial ventral mar-
gins of gonocoxite roughly parallel beyond ventral
gonocoxal bridge. (52) Ventroapical plate of male
gonocoxite absent. (55) Penis valves slender.

Subgenus Copoxyla Maa
Copoxyla Maa, 1954: 191, 211. Type species: Apis iris Christ, 1791, by origi-
nal designation.

The species of the subgenus Copoxyla are among the
smallest of the Xylocopini and are closely related to
Ctenoxylocopa. Females can be distinguished by the com-
bination of a smoothly rounded thorax, the absence of
preapical spines beside the pygidial spine, and the pres-
ence of three teeth on the apical margin of the mandible.
Males of Copoxyla can be distinguished from Ctenoxylocopa
by their small size, and by the presence of a defined
metapostnotum. These features and two others—the en-
tirely black face of the male, and the single spine on the
outer apex of the hind tibia—will differentiate males of
this group from all others.

Description.—As in Xylomelissa, except as follows: (4)
Apical edge of female mandible tridentate. (7) Supraocellar
pits deep. (10) Male face entirely black. (11) Outer apex of
male hind tibia with one apical spine. (16) Male hind fe-
mur without basal tubercle or spine. (17) Anterior and
posterior edge of basitibial plate of female each with row
of tubercles. (18) Apex of male basitibial plate bifid. (21)
Female basitibial plate extending nearly to base of tibia.

(24) Male metapostnotum present. (31) Graduli on T1 and
T2 of male. (32) Graduli present on T1-T5 of female. (36)
Apical margin of S1 medially emarginate. (39) Pregradular
areas on male sterna without paired setal patches. (41)
Preapical spines of female pygidial spine absent. (49)
Gonostylus of male with a weakly developed medially pro-
jecting lobe. (52) Ventroapical plate of male gonocoxite ab-
sent. (54) Male T7 with pair of apical dentiform processes.

Subgenus Gnathoxylocopa Hurd and Moure
Gnathoxylocopa Hurd and Moure 1963: 182. Type species: Xylocopa sicheli

Vachal, 1898, by original designation.

Females of Gnathoxylocopa are easily distinguished by
the long, curved hook that arises from the ventral margin
of the mandible. Males can be distinguished by the length-
ened thoracic dorsum (i.e., the posterior thoracic declivity
begins on the posterior half of the metanotum), and by the
bifid apex of the basitibial plate. This group seems to be
relatively plesiomorphic and most closely related to
Ctenoxylocopa and Copoxyla. Nevertheless, until more spe-
cies of Ctenoxylocopa, Xylocopa, and other allied groups are
examined, relationships among these taxa should be con-
sidered as tentative.

Description.—As in Xylomelissa, except as follows: (3)
Female mandible with tooth on upper carina. (4) Apical
edge of female mandible tridentate. (10) Male face entirely
black. (11) Outer apex of male hind tibia with one apical
spine. (16) Male hind femur without basal tubercle or spine.
(17) Anterior and posterior edges of female basitibial plate
each with row of tubercles. (18) Apex of male basitibial
plate bifid. (26) Reservoir of mesosomal gland present,
well-developed with numerous convoluted tubules. (31)
Graduli present on T1-T3 of male. (36) Apical margin of
S1 medially emarginate. (39) Pregradular areas on male
sterna without paired setal patches. (50) Medial ventral
margins of gonocoxites roughly parallel beyond ventral
gonocoxal bridge. (52) Ventroapical plate of gonocoxite
absent. (54) Male T7 with pair of apical dentiform pro-
cesses. (55) Penis valves slender.

Subgenus Lestis Lepeletier and Serville, New status

Lestis Lepeletier and Serville, 1828: 795, 799, 800. Type species: Apis
bombylans Fabricius, 1779 (Opinion 657, Bull. Zool. Nomencl., 20:181—
182,1963).

This group of two Australian species is distinguished
by the four-segmented maxillary palpi, the second sub-
marginal cell of the wing which is very nearly a parallelo-
gram and not attenuate on the posterior edge as in all other
Xylocopa, and strong metallic integument of both sexes.
Females are unique in having a large, swollen tubercle on
the parocular area below the antennal socket. The
gonostylus of the male genitalia is almost completely fused
to the gonocoxite and has a spinelike medial lobe at the

36 ScreNTIFIC PAPERS, NATURAL History Museum, THE UNIVERSITY OF KANSAS

apex, a unique feature in the tribe. This group, recognized
previously as a genus, seems to be most closely related to
the other groups allied to Xylocopa (sensu stricto), although
its affinities within this group are uncertain. It is clearly a
relict group, but, if recognized as a genus, would make
the genus Xylocopa paraphyletic. Therefore, I consider Lestis
to be a subgenus of Xylocopa.

Description.—As in Xylomelissa, except as follows: (11)
Outer apex of male hind tibia with one spine. (13) Tibial
spine of female strong, originating before tibial apex and
with a longitudinal medial carina. (17) Basitibial plate not
margined with tubercles. (20) Female basitibial plate fove-
ate, closed basally. (24) Male metapostnotum present. (26)
Mesosomal gland reservoir present. (32) Gradulus present
on T1-T5 of female. (36) Apical margin of S1 medially
emarginate. (39) Pregradular areas on male sterna with-
out paired setal patches. (40) Setal patches near male
metasomal spiracles not present. (48) Gonostylar apex of
male genitalia with dense setal patch. (49) Gonostylar apex
with medially projecting lobe, but not bifid. (50) Medial
ventral margins of gonocoxites roughly parallel beyond
ventral gonocoxal bridge. (52) Ventroapical plate of
gonocoxite absent. (55) Penis valves apically slender.

Subgenus Notoxylocopa Hurd
Notoxylocopa Hurd, 1956: 2. Type species: Xylocopa tabaniformis Smith, 1854,
by original designation.

This New World subgenus is defined by the posses-
sion in males of an elongation of the dorsal bridge of the
penis valve, or spatha, a feature shared with no other mem-
ber of the tribe. Females can be distinguished from all other
New World Xylocopa by the absence of a defined
metapostnotum. Hurd and Moure (1963) considered
Notoxylocopa to be most closely related to the Old World
subgenus Ritysoxylocopa. My analysis does not place this
group firmly with any other subgenus, although, based
on three male genital characters (strongly developed me-
dial lobe of the gonostylus, thickly hirsute apex of the
gonostylus, and especially the outward projection on the
penis valves), it could be most closely related to the New
World group that includes Schonnherria. However, the
weakly developed ventroapical plate of the gonocoxite
suggests this relationship is not close; hence I consider it
likely that Notoxylocopa is the sister group to a group that
includes the rest of the New World Xylocopa, except
Xylocopoides and Calloxylocopa.

Description.—As in Xylomelissa, except as follows: (3)
Tooth on upper carina of female mandible. (11) Outer apex
of male hind tibia with one spine in X. tabaniformis, two
spines in X. guatemalensis. (12) Outer apex of female hind
tibia with two spines. (17) Anterior and posterior edges of
female basitibial plate weakly developed or rising little
from surface of tibia in X. tabaniformis, with tubercles along

both lateral edges in X. guatemalensis. (26) Mesosomal gland
reservoir absent. (31) Graduli present on T1-T4 of male.
(32) Graduli present on T1-T4 of female. (36) Apical mar-
gin of S1 medially emarginate. (39) Pregradular areas on
male sterna without paired setal patches, although setal
patches occur in some subspecies of X. tabaniformis. (40)
Patches of setae near male metasomal spiracles absent. (43)
Conspicuous carina on outer edge of penis valve. (46)
Spatha present. (49) Medially projecting lobe of gonostylar
apex well developed. (50) Medial ventral margins of
gonocoxites roughly parallel beyond ventral gonocoxal
bridge. (52) Ventroapical plate of male gonocoxite weakly
developed. (54) Apex of T7 of male with pair of dentiform
projections.

Subgenus Schonnherria Lepeletier

Schonherria Lepeletier, 1841: 207. Type species: Xylocopa micans Lepeletier,
1841 by designation of Sandhouse, 1943: 598.

Schonherria Lepeletier, 1841: 207.

Schornherria Ashmead, 1899: 71. Misspelling of Schonherria Lepeletier,
1841.

Xylocospila Hurd and Moure, 1963:109. Type species: Xylocopa bambusae
Schrottky, 1902, by original designation. New synonymy.

Ioxylocopa Hurd and Moure, 1963: 116. Type species: Xylocopa chrysopoda

Schrottky, 1902, by original designation. New synonymy.

This is a speciose New World subgenus. Most species
are brightly metallic and fast-flying, but some species are
decidedly drab and hairy. They can be distinguished from
all other New World groups by the presence of graduli
only on T1 in females and by the bifid apex of the
gonostylus and a large spine on the ventroposterior mar-
gin of the gonocoxite in males. As defined here, this group
is expanded to include Xylocospila and Ioxylocopa. Both of
these taxa possess the two strongest synapomorphies that
define Schonnherria (sensu Hurd and Moure, 1963), i.e., a
strongly bifid apex of the gonostylus and the spine on the
ventral apex of the gonocoxites. Three other monotypic
subgenera (Nanoxylocopa, Dasyxylocopa, and Xylocopsis)
erected by Hurd and Moure (1963), differ from Schonnherria
primarily in the number of graduli on the terga of the fe-
male and may eventually be shown to be plesiomorphic
members of the Schonnherria clade. However, until the re-
lationships between Notoxylocopa, the Neoxylocopa group
and the Schonnherria group are more firmly established, I
have retained Nanoxylocopa, Dasyxylocopa, and Xylocopsis
as subgenera.

Description.—As in Xylomelissa, except as follows: (2)
Subapical tooth on upper carina of male mandible absent
except in X. micans. (3) Tooth on upper carina of female
mandible. (4) Two teeth on apical margin of female man-
dible except in X. viridigastra which has three. (7)
Supraocellar pits of female punctiform or deeply excavated
in X. macrops and X. micans. (11) Outer apex of male hind
tibia with one spine. (16) Male hind femur without basal
tubercle except in X. viridigastra. (17) Anterior and poste-

SUBGENERA AND GENERA OF CARPENTER BEES, XYLOCOPINI 37

rior edges of female basitibial plate weakly developed, or
rising little from surface of tibia. (18) Male basitibial plate
bifid at apex. (23) Beginning of female posterior thoracic
declivity smoothly rounded, or abruptly rounded in X.
macrops. (26) Mesosomal gland reservoir present as a single
layer of parallel-running tubules. (31) Gradulus on T1 of
male; T1 and T2 in X. bambusae. (32) Gradulus on T1 of
female. (35) Medial groove of female T1 linear; deeply
impressed in X. micans. (36) Apical margin of S1 medially
emarginate; entire in X. viridigastra. (39) Setal patches ab-
sent on pregradular areas of male sterna. (40) Patches of
appressed setae near spiracles of male metasoma absent.
(43) Strong carina present on lateral edge of penis valve.
(49) Medially projecting lobe on gonostylar apex elongate,
forming a bifid apex. (50) Medial ventral margins of
gonocoxites roughly parallel beyond ventral gonocoxal
bridge. (51) Gonocoxite with a conspicuous spine on pos-
terior ventral margin. (54) Apex of male T7 without pair
of dentiform projections, except in X. viridigastra.

Subgenus Dasyxylocopa Hurd and Moure
Dasyxylocopa Hurd and Moure 1963: 113. Type species: Xylocopa bimaculata

Friese, 1903, by original designation.

As noted by Hurd and Moure (1963), the sexes of this
species are distinctive because of the thick pubescence. This
subgenus is similar to Schonnherria but can be distinguished
from males of that group by the greater number of graduli
in Dasyxylocopa (T1-T5) than on Schonnherria (T1 only). The
phylogenetic position of this group is discussed under
Schonnherria.

Description.—As in Xylomelissa, except as follows: (2)
Upper carina of male mandible with tooth. (3) Upper ca-
rina of female mandible with tooth. (11) Outer apex of male
hind tibia with one spine. (16) Male hind femur without
basal tubercle or spine. (17) Anterior and posterior edges
of female basitibial plates weakly developed, rising little
from surface of tibia. (26) Mesosomal gland reservoir not
examined. (32) Graduli present on T1-T4 of female. (36)
Apical margin of S1 medially emarginate. (39) Setal patches
of pregradular areas of male sterna absent. (40) Patches of
appressed setae near spiracles of male metasoma absent.
(43) Lateral edge of penis valve with strong carina. (49)
Medially projecting lobe on gonostylus apex weakly de-
veloped. (50) Medial ventral margins of gonocoxites roughly
parallel beyond ventral gonocoxal bridge. (51) Gonocoxite
with conspicuous spine on posterior ventral margin. (54)
Apex of male T7 with pair of dentiform projections.

Subgenus Xylocopsis Hurd and Moure
Xylocopsis Hurd and Moure 1963: 124. Type species: Xylocopsis funesta
Maid], 1912, by original designation.

Females of Xylocopsis can be distinguished from all other
New World taxa by the possession of five tubercles on the

labrum. Males can be distinguished by the combination of
a distinctive yellow stripe on the scape, an abruptly
rounded beginning of the thoracic declivity and two spines
on the apex of T7. The phylogenetic position of this group
is discussed under Schonnherria.

Description.—As in Xylomelissa, except as follows: (3)
Tooth on upper carina of female mandible. (4) Two teeth
on apical margin of female mandible. (7) Supraocellar pits
of female punctiform. (11) Outer apex of male hind tibia
with one spine. (16) Male hind femur without basal tu-
bercle. (17) Anterior and posterior edges of female basitibial
plates weakly developed, rising little from surface of tibia.
(18) Apex of male basitibial plate bifid. (26) Mesosomal
gland reservoir present. (31) T1 of male with gradulus. (32)
T1 and T2 of female with graduli. (36) Apical margin of S1
medially emarginate. (39) Setal patches of pregradular ar-
eas of male sterna absent. (40) Patches of appressed setae
near spiracles of male metasoma absent. (43) Lateral edge
of penis valve with strong carina. (50) Medial ventral mar-
gins of gonocoxites roughly parallel beyond ventral
gonocoxal bridge. (51) Gonocoxite with a conspicuous
spine at posterior ventral margin. (54) Apex of male T7
with pair of dentiform processes.

Subgenus Nanoxylocopa Hurd and Moure

Nanoxylocopa Hurd and Moure, 1963: 99. Type species: Xylocopa ciliata
Burmeister, 1976, by original designation.

This monotypic subgenus can be distinguished from
the other New World groups by the evenly punctate tegula
in males and by the “transversely tuberculiform” (Hurd
and Moure, 1963) shape of the frontal carina in females.
The phylogenetic position of this group is discussed un-
der Schonnherria.

Description.—As in Xylomelissa, except as follows: (3)
Upper carina of female mandible with tooth. (11) Outer
apex of male hind tibia with one spine. (16) Male hind fe-
mur without basal tubercle. (17) Anterior and posterior
edges of female basitibial plates weakly developed, rising
little from surface of tibia. (26) Mesosomal gland reservoir
present. (36) Apical margin of S1 medially emarginate. (39)
Setal patches of pregradular areas of male sterna absent.
(40) Patches of appressed setae near spiracles of male
metasoma absent. (43) Lateral edge of penis valve with
strong carina. (50) Medial ventral margins of gonocoxites
roughly parallel beyond ventral gonocoxal bridge. (51)
Gonocoxite with conspicuous spine on posterior ventral

margin.
Subgenus Cirroxylocopa Hurd and Moure
Cirroxylocopa Hurd and Moure, 1963: 102. Type species: Xylocopa vestita
Hurd and Moure, 1963, by original designation.
Hurd and Moure (1963) suggested that Cirroxylocopa
was a basal taxon, near Nanoxylocopa, of the lineage that

38 Screntiric Papers, NATURAL History Museum, THE UNIVERSITY OF KANSAS

gave rise to the Schonnherria and Neoxylocopa clades. My
analysis suggests that it is the basal member of the clade
leading to the Neoxylocopa lineage.

Description.—As in Xylomelissa except as follows: (3)
Upper carina of female mandible with tooth. (11) Outer
apex of male hind tibia with one spine. (16) Male hind fe-
mur without basal tubercle. (17) Anterior and posterior
margins of female basitibial plate weakly developed, ris-
ing little from surface of tibia. (26) Mesosomal gland res-
ervoir not examined. (27) Beginning of thoracic declivity
on male propodeum. (32) T1 and T2 of female with graduli.
(40) Patches of appressed setae near spiracles of male
metasoma absent. (43) Lateral edge of penis valve with
strong carina. (49) Apex of gonostylus without medial lobe.
(50) Medial ventral margins of gonocoxites roughly paral-
lel beyond ventral gonocoxal bridge. (51) Gonocoxite with
a conspicuous spine on posterior ventral margin.

Subgenus Xylocopoda Hurd and Moure
Xylocopoda Hurd and Moure, 1963: 105. Type species: Xylocopa elegans

Hurd and Moure, 1963, by original designation.

This is a monotypic subgenus. Males can be distin-
guished uniquely by the presence of a high, thin lamella
that begins near the base of the hind tibial spur and ex-
tends obliquely forward and across the ventral surface of
the tibia. Females also are easily identified by the parocular
areas, which are abruptly raised near the lower margins
of the eye. In my analysis, Xylocopoda is the sister group to
Stenoxylocopa and Neoxylocopa.

Description.—As in Xylomelissa, except as follows: (3)
Tooth on upper carina of female mandible. (6) Female
epistomal suture continuously carinate and broadly
impunctate. (11) Outer apex of male hind tibia with one
spine. (16) Male hind femur without basal tubercle or spine.
(17) Anterior and posterior edges of female basitibial plate
weakly developed, rising little from surface of tibia. (26)
Mesosomal gland reservoir not examined. (27) Beginning
of thoracic declivity of male on propodeum. (32) T1-T5 of
female with graduli. (36) Apical margin of S1 medially
emarginate. (39) Setal patches of pregradular areas on male
sterna absent. (40) Patches of appressed setae near spiracles
of male metasoma absent. (43) Lateral edge of penis valve
with strong carina. (48) Gonostylus with few setae, not with
a dense brush at its apex. (49) Medial lobe of gonostylar
apex absent. (50) Medial ventral margins of gonocoxites
roughly parallel beyond ventral gonocoxal bridge. (51)
Gonocoxite with a conspicuous spine on posterior ventral
margin.

Subgenus Stenoxylocopa Hurd and Moure

Stenoxylocopa Hurd and Moure, 1960: 809. Type species: Xylocopa artifex
Smith, 1874a, by original designation.

Xylocopina Hurd and Moure, 1963: 160. Type species: Xylocopa ruficollis
Hurd and Moure, by original designation. New synonymy.

This group is exclusively New World, and most spe-
cies are Neotropical. An enlarged upper mandibular tooth,
as wide or wider than the lower tooth, distinguishes fe-
males of this group from all others in the tribe. Males can
be distinguished by the combination of two spines on the
outer apex of the hind tibia and the tegula being impunctate
on its posterolateral third. Except for Xylocopina ruficollis,
males of all species have two parallel yellow stripes run-
ning from the vertex of the head to the base of the man-
dibles. Xylocopina ruficollis was considered to be
subgenerically distinct from Stenoxylocopa by Hurd and
Moure (1963) based on two plesiomorphies—the entire,
but not produced, apex of S1 and yellow markings re-
stricted to the lower part of the face—and two
autapomorphies, the short clypeus and the shape of the
tegula.

Description.—As in Xylomelissa, except as follows: (3)
Upper carina of female mandible with tooth. (4) Female
mandible with three teeth on apical margin. (5) Upper
rutellar tooth and preapical tooth (i.e., apex of pollex)
fused and forming a continuous edge. (6) Female
epistomal suture continuously carinate and broadly
impunctate. (11) Outer apex of male hind tibia with one
spine. (12) Outer apex of female hind tibia with two spines.
(16) Male hind femur without basal tubercle. (17) Anterior
and posterior edges of female basitibial plate weakly de-
veloped, rising little from surface of tibia. (26) Mesosomal
gland reservoir present, consisting of several layers of disc-
shaped tubules. (27) Beginning of thoracic declivity of male
on propodeum, except in Xylocopina ruficollis. (35) Medial
groove of T1 of female linear. (36) Apical margin of S1
medially emarginate; entire and not produced in X.
ruficollis. (38) Medial carina on female sternum continu-
ous across all segments; present but not continuous in X.
ruficollis. (39) Pregradular areas on male sterna with setal
patches, except in X. ruficollis. (40) Setal patches present
near spiracles of male metasoma. (43) Lateral edge of pe-
nis valve with strong carina. (48) Gonostylus with few se-
tae, without dense brush of setae at its apex. (49) Medial
lobe on apex of gonostylus absent. (50) Medial ventral
margins of gonocoxites roughly parallel beyond ventral
gonocoxal bridge. (51) Gonocoxite with conspicuous spine
on posterior ventral margin.

Subgenus Neoxylocopa Michener

Neoxylocopa Michener, 1954:155, 157-158. Type species: Apis brasilianorum

Linnaeus, 1767, by original designation.

Megaxylocopa Hurd and Moure, 1963: 151. Type species: Apis frontalis

Olivier, 1789, by original designation. New synonymy.

This subgenus is the most speciose and widely distrib-
uted of the New World subgenera. Females can be distin-
guished from all other members of the tribe by the combi-
nation of two teeth on the apical margin of the mandible,
a continuous impunctate ridge along the epistomal suture,

SUBGENERA AND GENERA OF CARPENTER BEES, XYLOCOPINI 39

and a strong medial carina along the entire sternum of the
metasoma. Males are distinguished by their yellow
vestiture (most species) and the beginning of the posterior
thoracic declivity on the propodeum. Hurd and Moure
(1963) based the subgenus Megaxylocopa on their extremely
large size and conspicuous carinae over the lateral ocelli.
Recognition of Megaxylocopa makes the subgenus
Neoxylocopa paraphyletic.

Description.—As in Xylomelissa, except as follows: (1)
Male vestiture on dorsum of body entirely tawny or fer-
ruginous (most species). (3) Tooth on upper carina of fe-
male mandible. (6) Female epistomal suture a continuous
ridge and broadly impunctate. (11) Outer apex of male hind
tibia with one spine. (12) Outer apex of female hind tibia
with two spines. (16) Male hind femur without basal tu-
bercle. (17) Anterior and posterior edges of female basitibial
plates weakly developed, rising little from surface of tibia.
(26) Reservoir of male mesosomal gland extremely large
and consisting of many layers of highly convoluted tu-
bules, sometimes visible through cuticle of metanotum.
(27) Beginning of thoracic declivity of male on propodeum.
(36) Apical margin of S1 medially emarginate. (38) Medial
carina on sternum of female continuous across all seg-
ments. (39) Pregradular areas of male sterna without paired
setal patches. (40) Setal patches present near spiracles of
male metasoma. (43) Lateral edge of penis valve with
strong carina. (48) Gonostylus with few setae, without
dense brush of setae at its apex. (49) Medial lobe on
gonostylar apex absent. (50) Medial ventral margins of
gonocoxites roughly parallel beyond ventral gonocoxal
bridge. (51) Gonocoxite with a conspicuous spine on pos-
terior ventral margin.

Subgenus Prosopoxylocopa Hurd and Moure
Prosopoxylocopa Hurd and Moure, 1963: 215. Type species: Xylocopa
mirabilis Hurd and Moure, by original designation.

This subgenus is likely the sister group to the remain-
der of the Ethiopian clade. The strongly produced projec-
tion of the posteriormost ventral part of the gonocoxite
(Fig. 9C) and the continuous yellow maculations from the
labrum to the supraclypeal area of the face of the male
distinguish it from all other large carpenter bees. Females
can be distinguished by the abruptly rounded beginning
of the thoracic declivity and by the basitibial plate, which
is represented by a single, very strongly projecting carina.

Description.—As in the Xylomelissa, except as follows:
(11) Outer apex of male hind tibia with two spines. (12)
Outer apex of female hind tibia with two spines. (16) Male
hind femur without tubercle or spine. (17) Posterior edge
of female basitibial plate with carina arising perpendicu-
larly from leg, without row of tubercles. (19) Female
basitibial plate apex entire. (22) Beginning of posterior tho-
racic declivity of male abruptly rounded. (23) Beginning

of posterior thoracic declivity of female abruptly rounded.
(24) Male metapostnotum present. (26) Mesosomal gland
reservoir not examined. (32) Graduli present on T1I-T5 of
female. (39) Pregradular areas of male sterna without
paired setal patches. (40) Setal patches near male
metasomal spiracles absent. (50) Medial ventral margins
of gonocoxites roughly parallel from ventral gonocoxal
bridge. (51) Gonocoxite with a bifid spine originating at
the ventroposterior margin (Fig. 9C). (52) Ventroapical
plate of gonocoxite weakly developed.

Subgenus Zonohirsuta Maa

Zonohirsuta Maa, 1938: 270, 300. Type species: Xylocopa collaris Lepeletier,
1841, not Olivier, 1789 (= X. dejeanii Lepeletier, 1841), by original des-
ignation.

The subgenus Zonohirsuta is similar to Prosopoxylocopa.
They differ by the basally open basitibial plate in females
of Zonohirsuta and by the single spine on the outer apex of
the hind tibia of males. An additional apomorphy of
Zonohirsuta is that both of the lateral edges of the female
basitibial plate are strongly carinate and project perpen-
dicularly from the tibial surface.

Description.—As in Xylomelissa, except as follows: (11)
Outer apex of male hind tibia with one spine. (12) Outer
apex of female hind tibia with two spines. (16) Male hind
femur armed with basal spine. (17) Posterior edge of fe-
male basitibial plate with carina arising perpendicularly
from leg, without row of tubercles. (22) Beginning of pos-
terior thoracic declivity of male abruptly rounded. (23)
Beginning of posterior thoracic declivity of female abruptly
rounded. (24) Male metapostnotum present. (26)
Mesosomal gland reservoir present. (32) Graduli present
on T1-T5 of female. (39) Pregradular areas on male sterna
without paired setal patches. (40) Setal patches near male
metasomal spiracles absent. (47) Gonostylar apex without
dense setal patch. (48) Gonostylus slender. (49) Gonostylar
apex without medially projecting lobe. (50) Medial ven-
tral margins of gonocoxites roughly parallel beyond ven-
tral gonocoxal bridge. (52) Ventroapical plate of gonocoxite
present, with a carina.

Subgenus Bomboixylocopa Maa
Bomboixylocopa Maa, 1939: 155. Type species: Xylocopa bomboides Smith,

1879, by original designation.

Mimoxylocopa Hurd and Moure, 1963: 203. Type species: Xylocopa rufipes

Smith, 1852, by original designation. New synonymy.

Females of this subgenus can be distinguished from all
other members of the Xylocopini by the combination of
long malar space (longer than the width of first flagellar
segment) and two spines on the outer apex of the tibia.
Males can be distinguished by the presence of appressed
setae, both on the pregradular areas of the metasomal
sterna and near the metasomal spiracles, and by two spines
on the outer apex of the tibia.

40 ScIENTIFIC PAPERS, NATURAL History Museum, THE UNIVERSITY OF KANSAS

Description.—As in Xylomelissa, except as follows: (2)
Upper carina of male mandible without tooth, except in
Xylocopa rufipes. (3) Upper carina of female mandible with-
out tooth, except in X. rufipes. (8) Malar space of female
longer than width of first flagellar segment. (11) Two spines
on outer apex of male hind tibia. (12) Two spines on outer
apex of female hind tibia. (16) Male hind femur without
basal spine or tubercle. (17) Anterior and posterior edges
of female basitibial plates rising little from surface of tibia.
(19) Female basitibial plate apex entire, or asymmetrically
bifid in X. rufipes. (24) Male metapostnotum present in X.
chinensis, absent in X. rufipes. (25) Female metapostnotum
present, extending to propodeal pit. (26) Mesosomal gland
reservoir present. (32) Graduli present on T1-T5 of female.
(35) Medial groove of female T1 deeply sulcate. (36) Api-
cal margin of S1 medially emarginate in X. rufipes, rounded
in X. chinensis. (41) Pygidial spine lacking lateral spine at
base in X. chinensis; these spines present in X. rufipes. (47)
Gonostylus slender and elongate, over half as long as
gonocoxite (Fig. 9D). (48) Gonostylus lacking dense setal
patch on apex. (49) Gonostylar apex without medially pro-
jecting lobe. (50) Medial ventral margins of gonocoxites
roughly parallel beyond ventral gonocoxal bridge, diver-
gent in X. chinensis. (51) Gonocoxite with spine on poste-
rior ventral margin. (53) Gonocoxite with projection along
inner mesal margin of base.

Subgenus Xenoxylocopa Hurd and Moure

Xenoxylocopa Hurd and Moure, 1963: 243. Type species: Mesotrichia
chiyakensis Cockerell, 1908, by original designation.

Females can be distinguished from all other members
of the tribe by the abruptly rounded beginning of the pos-
terior thoracic declivity and by the presence of a pair of
preapical spines at the base of the pygidial spine. Males
can be distinguished by the combination of sharp begin-
ning of the declivity of T1 (as occurs in Koptortosoma and
Mesotrichia), lack of a fovea on the anterior surface of T1,
and the nearly entirely punctate tegula. Xenoxylocopa and/
or Alloxylocopa are/is the sister group to Mesotrichia and
Koptortosoma. (See section on Relationships among Members
of the Major Clades and Taxonomic Recommendations.)

Description.—As in Xylomelissa, except as follows: (1)
Male vestiture tawny or ferruginous over entire dorsum
of body. (6) Female epistomal suture with a continuous
smooth ridge. (7) Supraocellar pits of female deeply exca-
vated and craterlike. (11) Two spines on outer apex of male
hind tibia. (12) Two spines on outer apex of female hind
tibia. (16) Male hind femur without basal spine or tubercle.
(17) Anterior and posterior edges of female basitibial plates
rising little from surface of tibia. (23) Beginning of poste-
rior thoracic declivity of female abruptly rounded. (25)
Female metapostnotum absent. (26) Mesosomal gland res-
ervoir not examined. (28) Propodeum at its connection with

metasoma dorsally elongated, forming a spine. (30)
Gradulus of T1 nearly coincident with lateral ventral mar-
gin of metasoma. (35) Medial groove of T1 of females
deeply sulcate. (36) Apical margin of S1 entire and pro-
duced. (37) Medial carina on S6 of male strongly devel-
oped, forming a keel. (39) Pregradular areas of male sterna
without setal patches. (40) Setal patches near male
metasomal spiracles absent. (41) Pygidial spine absent. (47)
Gonostylus of male slender and elongate; over half as long
as gonocoxite (Fig. 9D). (48) Gonostylar apex lacking dense
setal patch. (49) Gonostylar apex without medially pro-
jecting lobe. (50) Medial ventral margins of gonocoxites
roughly parallel beyond ventral gonocoxal bridge. (51)
Gonocoxite with a spine on posterior ventral margin. (53)
Gonocoxite with projection along inner mesal margin of base.

Subgenus Alloxylocopa Maa
Alloxylocopa Maa, 1939: 155. Type species: Xylocopa appendiculata Smith,

1852, by original designation.

The subgenus Alloxylocopa is similar to Xenoxylocopa but
differs in females by the lack of preapical spines beside
the pygidial spine and in males by the dark coloration and
presence of a fovea on the anterior face of Tl. In the phylo-
genetic analysis, this group is often placed as the sister
group to Mesotrichia and Koptortosoma.

Description.—As in Xenoxylocopa, except as follows: (1)
Male vestiture black or dark brown on sides and pale on
dorsum. (7) Supraocellar pits of female normal. (19) Apex
of female basitibial plate entire. (22) Beginning of poste-
rior thoracic declivity of male abruptly rounded. (28)
Propodeum at its connection with metasoma not elongated.
(30) Gradulus of T1 obliquely curved on anterior ventral
margin of metasoma. (31) Gradulus only on T1 of male.
(33) Anterior surface of T1 of male with fovea. (34) Ante-
rior surface of T1 of females with fovea. (41) Basal pair of
pygidial spines present. (51) Gonocoxite without a spine
on posterior ventral margin.

Subgenus Mesotrichia Westwood

Mesotrichia Westwood, 1838: 112. Type species: Mesotrichia torrida
Westwood, 1838, original description.

Platynopoda Westwood, in Jardine, 1840: 271. Type species: Apis latipes
Drury, 1773, by designation of Ashmead, 1899: 71. New synonymy.

Audinetia Lepeletier, 1841: 203. Type species: Apis latipes Drury, 1773, syn-
onym of Platynopoda by designation of Sandhouse, 1943: 529.

Platinopoda Dalla Torre, 1896: 202. Emendation of Platynopoda Westwood

Hoplitocopa Lieftinck, 1955: 9, 27. Type species: Xylocopa assimilis Ritsema,
1880, by one Ree New synonymy. c

Hoploxylocopa Hurd and Moure, 1963: 260. Type species: Xylocopa
acutipennis Smith, 1854, by original designation. New synonymy.

This subgenus is the sister group to the subgenus
Koptortosoma. Females can be distinguished from all other
large carpenter bees by the perpendicular angle of the be-
ginning of the thoracic declivity, the deeply excavated
(craterlike) supraocellar pits and triangular hind tro-

SUBGENERA AND GENERA OF CARPENTER BEES, XYLOCOPINI 4]

chanter. Males can be distinguished by the posteriorly elon-
gate tegulae (Fig. 4B). This is a structurally diverse group
as the description indicates; yet a number of characters
provide evidence for its monophyly. Mesotrichia, as recog-
nized herein, is an equivalent taxon to Koptortosoma; by
combining the four subgenera recognized by Hurd and
Moure (1963), the evolutionary relationships of the spe-
cies are emphasized.

Description.—As in Alloxylocopa, except as follows: (3)
Upper carina of male mandible without a tooth, except in
Xylocopa assimilis. (3) Upper carina of female mandible
without a tooth, except in X. assimilis. (6) Epistomal suture
of female continuously carinate and smooth in X. torrida
and X. acutipennis, but not so in X. latipes and X. assimilis.
(7) Supraocellar pits of female deeply excavated and
craterlike. (10) Male with yellow facial maculations except
in X. assimilis. (14) Female hind trochanter roughly trian-
gular in ventral view. (15) Mid- and hind trochanter of
male without spine or tubercle except in X. acutipennis,
which has a spine. (16) Male hind femur with basal spine
in X. torrida and X. acutipennis, without this projection in
X. latipes and X. assimilis. (22) Beginning of posterior tho-
racic declivity of male sharply angled. (23) Beginning of
posterior thoracic declivity of female a transverse carina
not extending posteriorly beyond metanotum. (24) Male
metapostnotum absent, except in X. latipes, in which it is
vestigial. (Among specimens of X. fenuiscapa I have exam-
ined, the metapostnotum is exposed in some individuals
and absent in others.) (25) Female metapostnotum present
in X. latipes and X. assimilis, but absent in X. torrida and X.
acutipennis. (26) Reservoir of male mesosomal gland
present and variable; two, widely separated pouches that
are not developed into tubules in X. latipes, and highly
convoluted mass of tubules in X. torrida. (28) Propodeum
at its connection with metasoma not modified in X. latipes
and X. assimilis but dorsally elongated, forming a spine in
X. torrida and X. acutipennis. (29) Male tegula elongated
posteriorly. (30) Gradulus of S1 virtually coincident with
anterior ventral margin of metasoma. (31) Distribution of
graduli on male terga variable: T1-T5 in X. torrida, T1-T3
in X. acutipennis and only T1 in X. latipes and X. assimilis.
(34) Anterior surface of Tl of female with entrance to a
mite pouch. (37) Medial carina on S6 of male present but
not strongly developed, except in X. torrida. (38) Medial
carina on sterna of female continuous and present on all
segments; present but not continuous in X. acutipennis. (41)
Pygidial spine lacking preapical lateral spines at base, ex-
cept in X. torrida. (42) Gonostylus of male genitalia slen-
der in X. tenuiscapa and X. acutipennis, completely undif-
ferentiated from gonocoxite in X. torrida, and broadly fused
but still recognizable in X. assimilis. (45) Penis entirely
membranous in X. torrida and X. acutipennis, and sclero-
tized basally beneath in X. latipes and X. assimilis. (47)

Gonostylus of male slender and less than half as long as
gonocoxite (also see Character 42 above). (48) Apex of
gonostylus setose, with few setae in X. latipes. (50) Medial
ventral margins of gonocoxites roughly parallel beyond
ventral gonocoxal bridge; divergent in X. latipes. (51)
Gonocoxite with strong spine from posterior ventral mar-
gin in X. torrida and X. assimilis but produced only slightly
as a rounded lobe in X. latipes and X. acutipennis. (52)
Ventroapical plate of gonocoxite present or absent in X.
latipes. (53) Gonocoxite without projection on inner mesal
margin except in X. torrida. (54) Apex of T7 of male entire,
with dentiform projections in X. assimilis. (55) Penis valves
slender and parallel.

Subgenus Koptortosoma Gribodo

Koptortosoma Gribodo, 1894: 271. Type species: Koptortosoma gabonica
Gribodo, 1894, by designation of Sandhouse, 1943: 561.

Koptorthosoma Dalla Torre, 1896: 202. Unjustified emendation of
Koptortosoma Gribodo.

Coptorthosoma Pérez, 1901:3. Unjustified emendation of Koptortosoma
Gribodo and Koptorthosoma Dalla Torre.

Orbitella Maa, 1938: 270, 305 (not Orbitella Douville, 1915).

Maiella Michener, 1942: 50:282 . Replacement name for Orbitella Maa, 1938.

Euryapis Sandhouse, 1943: 551 Replacement name for Orbitella Maa, 1938.

Oxyxylocopa Hurd and Moure, 1963: 275. Type species Xylocopa varipes
Smith, 1854, by original designation. New synonymy.

Cyaneoderes Ashmead, 1899: 70. Type species: C[yaneoderes] fairchildi
Ashm[ead, 1899] [= Bombus caeruleus Fabricius, 1841, by original des-
ignation. New synonymy.

Cyphoxylocopa Hurd and Moure, 1963: 283. Type species: Xylocopa ocularis
Pérez, 1901, by original designation. New synonymy.

Afroxylocopa Hurd and Moure, 1963: 264. Type species: Apis nigrita Fabri-

cius, 1775, by original designation. New synonymy.

This subgenus is the sister group to Mesotrichia (sensu
lato). The pronounced transverse carina on the scutellum
that completely overhangs the metanotum of females of
most species distinguishes these taxa from females of all
other large carpenter bees. Males share many features with
those of Mesotrichia and can be separated most easily by
the tegula, which is not modified in Koptortosoma and is
elongated posteriorly in Mesotrichia. The apomorphies
identified herein and by Hurd and Moure (1963) that de-
fine the monophyly of Oxyxylocopa, Cyphoxylocopa,
Cyaneoderes, and Afroxylocopa are all highly homoplasious
and do not justify recognizing these taxa as separate sub-
genera. (See section on Relationships among Members of
the Major Clades and Taxonomic Recommendations.)

Description.—As in Alloxylocopa, except as follows: (1)
Male vestiture blue in Xylocopa caerulea to tawny or fer-
ruginous over entire dorsum of body in X. nigrita, X.
pubescens, X. caffra and X. sinensis. (2) Male mandible with-
out tooth on upper carina, except in X. nigrita and X.
pubescens. (3) Upper carina of female mandible with tooth
except in X. nigrita, X. caerulea and X. ocularis. (6) Female
epistomal suture without continuous ridge and devoid of
setae, except in X. nigrita, X. caerulea and X. ocularis. (7)

42 SCIENTIFIC PAPERS, NATURAL History Museum, THE UNIVERSITY OF KANSAS

Supraocellar pits of female punctiform; deeply excavated
in X. nigrita and X. varipes. (10) Male with yellow facial
maculations, except in X. pubescens and X. caerulea. (12)
Outer apex of female hind tibia with one spine, except in
X. nigrita. (14) Female hind trochanter roughly cylindri-
cal, not triangular in ventral view, except in X. caffra. (15)
Mid- and hind trochanter of male without spines, except
in X. varipes. (16) Male hind femur without spine or tu-
bercle in most species, with basal spine in X. nigrita and X.
ocularis. (19) Apex of female basitibial plate entire, except
in X. nigrita. (22) Beginning of posterior thoracic declivity
of male sharply angled, except in X. sinensis. (23) Begin-
ning of posterior thoracic declivity of female a transverse
carina that overhangs entire metanotum but rounded in
X. sinensis. (25) Female metapostnotum absent or present
in X. caerulea. (26) Reservoir of male mesosomal gland a
highly convoluted, many-layered mass of tubules. (27)
Beginning of thoracic declivity of male on metanotum,
except in X. migrita, in which it is on the propodeum. (28)
Propodeum at its connection with metasoma dorsally elon-
gated forming a spine except in X. nigrita. (29) Graduli
present on male T1 and T2 or only on T1 in X. nigrita. (30)
Gradulus on S1 virtually coincident with anterior ventral

margin of metasoma. (32) Graduli present only on T1 of
female or T1 and T2 in X. pubescens. (34) Anterior surface
of T1 of female with entrance to mite pouch or with fovea
in X. varipes and X. ocularis. (35) Medial groove of T1 of
females deeply sulcate. (37) Medial carina of S6 of male
strongly developed, forming a keel, except in X. varipes.
(39) Pregradular areas of male sterna with paired setal
patches; absent in X. pubescens. (40) Setal patches present
near metasomal spiracles of males, except in X. varipes, X.
caerulea and X. ocularis. (41) Pygidial spine lacking pair of
preapical lateral spines at base. (47) Gonostylus of male
genitalia less than half as long as gonocoxite, with setae at
apex. (48) Gonostylus lacking dense setal patch on apex.
(49) Gonostylar apex with medially projecting lobe in X.
varipes, X. caerulea, X. ocularis and X. caffra, and lobe pro-
ducing a bifid point in X. sinensis, X. pubescens and X.
nigrita. (50) Medial ventral margins of gonocoxites roughly
parallel beyond ventral gonocoxal bridge or divergent in
X. chinensis. (51) Gonocoxite with spine on posterior ven-
tral margin, except in X. pubescens and X. nigrita. (53)
Gonocoxite with projection along inner mesal margin of
base, except in X. ocularis. (54) T7 of male with apex entire
or with pair of dentiform spines on X. nigrita.

LITERATURE CITED

Ashmead, W. H. 1899. Classification of the bees, or the superfamily
rai Transactions of the American Entomological Society 26:49—
100.

Brothers, D. 1976. Modification of the metapostnotum and origin of the
propodeal triangle in Hymenoptera Aculeata. Systematic Entomol-
ogy 1:177-182. ;

Carpenter, J. M. 1988. Choosing among multiple equally parsimonious
cladograms. Cladistics 4:291-296.

Christ, J. L. 1791. Naturgeschichte, Klassification und Nomenclature der
Insekten vom Bienen—Wespen—und Ameisengeschlecht. Frankfurt am
Main: Hermann, 535 pp.

Cockerell, T. D. A. 1908. Some bees collected by Dr. F.C. Wellman in West
Africa. Entomologist 41:34-36.

Cockerell, T. D. A. 1919. Bees in the collection of the United States Na-
tional Museum. 3. Proceedings of the United States National Mu-
seum 55:167—221.

Cockerell, T. D. A. 1929. Descriptions and records of bees. CXIX. Annals
and Magazine of Natural History (10)4:296—-304.

Corbet, S. A. and P. G. Wilmer. 1980. Pollination of the yellow passionfruit:
nectar, pollen and carpenter bees. Journal of Agricultural Science,
Cambridge 95:655—-666.

Cunha, R. A. D. 1992. Comparing phenetic structures: An application to
subgenera of Xylocopa Latreille, 1802 (Hymeno tera Anithophoridae):
Revista Brasiliera de Entomologia 36: 465-473.

Daly, H. V. 1988. Bees of the new genus Cfenoceratina in Africa south of
the Sahara (Hymenoptera: Apoidea). University of California Publi-
cations in Entomology 108:1-69.

Dalla Torre, C. G. de. 1896. Catalogus Hymenopterorum......Vol. 10. Leipzig:
Engleman, viii+643 pp.

Drury, D. 1773. Illustrations of Natural History. London: White, xxvii+90

Barley C. D. 1983. A taxonomic revision of the genus Xylocopa Latreille
(Hymenoptera: Anthophoridae) in southern Africa. Republic of South
Africa Department of Agriculture, Entomology Memoir 58:1-67.

Enderlein, G. 1903. Neue und weniger bekannte afrikanische Xylocopen,
besonders aus der Ausbeute des Freiherrn von Erlanger in Galla und
Somalica. Berliner Entomologische Zeitschrift 48:45-64.

Fabricius, J. C. 1775. Systema Entomologie.... Flensburg and Leipzig: Korte,
832 pp.

Fabactueiy C. 1793. Entomologia Systematica Emendata et Aucta. Vol. 2.

Hafniae: Proft, viii+519 pp.

Fabricius, J. C. 1798. Entomologiae Systematicae Supplementum.
Copenhagen, 572 PP.

Fabricius, J. C. 1804. Systema Piezatorum.... Braunschweig, xiv+[15]-
[440]+[1]-30.

Farris, J. S. 1969. A successive approximations approach to character
weighting. Systematic Zoology 18:374—385.

Farris, J. S. 1988. Hennig86, version 1.5. (Distributed by the author.) Port
Jefferson, New York.

Fitzhugh, K. 1989. Cladistics in the fast lane. Journal of the New York
Entomological Society 97:234—241.

Friese, H. 1903. Neue Arten der Bienengattung Xylocopa Latr. aus der
neotropischen und orientalischen Region (Hym.). Zeitschrift fur
systematiedle Hymenopterologe und Pe gare 3:202-208.

Gauthier, J. A., A. G. Kluge and T. Rowe. 1988. Amniote phylogeny and
the importance of fossils. Cladistics 4:105-209.

Gerstaecker, C. E. A. 1857. [Bees and wasps collected in Mozambique].
Monatsberichte, Akademie der Wissenschaften, Berlin, October 1857,
29:460—-464.

Gottsberger, G., J. M. F. Camargo, and I. S. Gottsberger. 1988. A bee-pol-
linated tropical community: The beach dune vegetation of Ilha de
Sao Luis, Maranhao, Brazil. Botanische Jahrbiicher ftir Systematik,
Pflanzengeschichte und Pflanzengeographie 109:469-500.

Gribodo, G. 1894. Note imenotterologiche, Nota II. Nuovi generi e nuove
specie di Imenotteri antofili ed isservazioni sopra alcune specie gia
conosciute. Bullettino della Societa Entomologica Italiana (Florence)
26:262-314. -

Hedicke, H. 1938. Uber einige Apiden vom Hindu-kusch. Deutsche
Entomologische Zeitschrift 1938:186—-196.

Hirashima, Y. 1969. Synopsis of the genus Pithitis Aine of the world (Hy-
menoptera: Anthophoridae). Pacific Insects 11:649-669.

Hurd, P. D., Jr. 1955. The Capen: bees of California. Bulletin of the Cali-
fornia Insect Survey 4: 37-72.

Hurd, P. D., Jr. 1956. Notes on the subgenera of the New World carpenter
bees of the genus Xylocopa. American Museum Novitates 1776:1-7.

Hurd, P. D., Jr. 1958. Observations on the nesting habits of some New
World carpenter bees with remarks on their importance in the prob-
lem of species formation. Annals of the Entomological Society of
America 51:365-375.

Hurd, P. D., Jr. 1959. Some nomenclatorial problems in the genus Xylocopa

SUBGENERA AND GENERA OF CARPENTER BEES, X YLOCOPINI 43

Latreille. Pan-Pacific Entomologist 35:135-148.

Hurd, P. D., Jr. and J. S. Moure. 1960. A New World subgenus of bamboo
nesting carpenter bees belonging to the genus Xylocopa Latreille (Hy-
menoptera: Apoidea). Annals of the Entomological Society of America
53:809-821.

Hurd, P. D., Jr. 1978. Bamboo-nesting carpenter bees (genus Xylocopa
Latreille) of the subgenus Stenoxylocopa Hurd and Moure (Hy-
menoptera: Anthophoridae). Journal of the Kansas Entomological
Society 51:46-76.

Hurd, P. D., Jr. and J. M. F. Moure. 1963. A classification of the large car-

enter bees. University of California Publications in Entomology 29:1—
365.

Kluge, A. G. 1989. A concern for evidence and a phylogenetic hypothesis
of pele bonships among Epicrates (Boidae, Serpentes). Systematic Zo-
ology 38:7-25.

Lepeletier de Saint-Fargeau, A. L. M. 1841. Histoire Naturelle des Insectes—
Hymenopteres. Vol. 2. Paris: Roret, 680 Pp:

Lepeletier de Saint-Fargeau, A. L. M., and A. Serville. 1828. [Articles in]
M. Diderot et al., Encyclopédie Méthodique, Histoire Naturelle. Insectes.
Vol. 10, P. A. Latreille (ed.). Paris, pp: 1-334, 1825.

Lieftinck, M. A. 1955. The carpenter-bees (Xylocopa Latr.) of the Lesser
Sunda Islands and Tanimbar (Hymenoptera, Apoidea).
Verhandlungen der Naturforschenden Gesellschaft in Basel 66:5-32.

Linnaeus, 1767. Systema Naturae. 12th ed. Holmiae: Laurentii Salvii 12:
901-1068.

Ma, T.-c.[= Maa] 1938. The Indian species of the genus Xylocopa Latr.
(Hymenoptera). Records of the Indian Museum 40:265-329.

Maa, T.-c. 1939. Xylocopa orientalia critica (Hymen.). 1. Subgenus
Bomboixylocopa novum. Lingnan Science Journal (Canton) 18:155-160.

Maa, T.-c. 1954. The xylocopine bees (Insecta) of Afghanistan.
Videnskabelige Meddelelser fra Dansk Naturhistorisk Forening i
Kjobenhavn 116:189-231.

Maa, T.-c. 1968. A revision of type specimens of Xylocopine bees in the
Genova Museum collection. Estratto Dagli Annali del Museo Civico
di Storia Naturale Genova 77:103-114.

Maa, T.-c. 1970. A revision of the subgenus Ctenoxylocopa (Hymenoptera:
Anthophoridae). Pacific Insects 12:723-752.

Maddison, W. P. 1993. Missing data versus missing characters in phylo-

enetic analysis. Systematic Biology 42:576-581.

Maddison, W. P., and D. R. Maddison. 1992. MacClade: Analysis of Phylog-
eny and Character Evolution. Version 3. Sinauer Associates, Sunderland,
Massachusetts, xi+398 pp.

Maddison, W. P., M. J. Donoghue, and D. R. Maddison. 1984. Outgroup
analysis and parsimony. Systematic Zoology 33:83-103.

Maeta, Y., S. F Sakagami, and M. Schiokawa. 1985. Observations on a
nest aggregation of the Taiwanese bamboo carpenter bee Xylocopa
(Gun tranquebarorum tranquebarorum (Hymenoptera,
Anthophoridae). Journal of the Kansas Entomological Society 58:36—
41

Maeta, Y., and S. F. Sakagami. 1995. Oophagy and egg replacement in
artificially induced colonies of a basically solitary bee, Ceratina
(Ceratinidia) okinawana (Hymenoptera, Anthophoridae, Xylocopinae),
with a comparison of social behavior among Ceratina, Xylocopa and
halictine bees. Japanese Journal of Entomology 63: 347-375.

Maidl, F. 1912. Die Xylocopen (Holzbienen) des Wiener Hofmuseums
Ein Beitrag zu einer Monographie dieser Gattung. Annalen K.-K.
Naturhistorischen Hofmuseums 26:249-330.

Malyshey, S. J. 1913. Life and instincts of some Ceratina-bees (Hy-
menoptera, prise) Come ay cand experimental study. Russkoe
Entomol. (Horae Societatis Entomologicae Rossicae), 40:1-28 [in Rus-
sian with English summary].

Michener, C. D. 1942. Taxonomic observations on bees with descriptions
of new panera and species. Journal of the New York Entomology
Society 50:273-282.

Michener, C. D. 1954. Bees of Panama. Bulletin of the American Museum
of Natural History 104:1-176.

Michener, C. D. 1975. A taxonomic study of African allodapine bees (Hy-
menoptera, Anthophoridae, Ceratinini). Bulletin of the American
Museum of Natural History 155:67—240.

Michener, C. D. 1990. Castes in xylocopine bees. Pp. 123-146 in W. Engels
(ed.), Social Insects: An Evolutionary Approach to Castes and Reproduc-
tion. Berlin: Springer-Verlag, 1-264 pp.

Michener, C. D., and A. Fraser. 1978. A comparative anatomical study of
the mandibular structure in bees. University of Kansas Science Bul-
letin 51:463-482.

Olivier, G. A. 1789. Abeille, in Encyclopédie Méthodique, Histoire Naturelle,
Publiée par une société de gens de lettres, de savans et d’artistes. Insectes,
Paris and Liege Vol. 4, cclxxxviii+ccclxxii+331 pp.

Pérez, J. 1901. Contribution a | étude des xylocopes. Actes de la Société
Linnéenne de Bordeaux (6)6:1—128.

Platnick, N. L, C. E. Griswold, and J. A. Coddington. 1991. On missing
entries in cladistic analysis. Cladistics 7:337-343.

Popov, V. B. 1947. Beeler character of some Palearctic species
of the bee genus Xylocopa Latr. (Hymenoptera: Apoidea) and their
interrelations with melittophilous plants. Trudy Instituta Zoologii,
Akademii Nauk SSSR 1:29—52 [In Russian].

Popov, V. B. and A. A. Ponomareva. 1961. Notes on the carpenter-bees
fauna of the Soviet Union (Hymenoptera, Apoidea).
Entomologieskoe Obozrnie 40:393—-404 [In Russian].

Reyes, S. G. 1991. Revision of the bee genus Braunsapis in the Oriental
Region. University of Kansas Science Bulletin 54:179-207.

Reyes, S. G. 1993. Revision of the genus Braunsapis in the Australian Re-
gion. University of Kansas Science Bulletin 55:97-122.

Ritsema, C. 1880. Description of three new exotic species of the Hy-
menopterous genus Xylocopa. Notes Leyden Museum 2:220-224.
Sandhouse, G. A. 1943. The type species of the genera and subgenera of

bees. Proceedings of the United States National Museum 92:519-619.

Smith, F. 1852. Descriptions of some new and apparently undescribed

species of hymenopterous insects from North China, collected by
obert Fortune, Esq. Transactions of the Entomological Society of
London 2:33-48.

Smith, F. 1854. Catalogue of the Hymenopterous Insects of the British Mu-
seum. Part 2. London: British Museum. Pp. 199-465.

Smith, F. 1874. Monograph of the genus Xylocopa Latr. Transactions of
the Entomological Society of London 1854:247-307.

Smith, F. 1878. Scientific Results of the Second Yarkland Mission: Based
on the Collections and Notes of the late Ferdinand Stoliczka: Hy-
menoptera. Calcutta: Government of India, pp. 1-22.

Smith, F. 1879. Descriptions of New Species of Hymenoptera in the Collection
of the British Museum. London: British Museum, xxi+240 pp.

Taschenberg, E. 1879. Die arten der Gattung Xylocopa Ltr. des fikile’schen
Zoologischen Museums. Zeitschrift fur de Gesammte
Naturwissschaft 52:563-599.

Vachal, J. 1898. Matériaux pour une révision synoptique des espeéces
Européennes et Africaines du Xylocopa Latr. Annales de la Société
Entomologique de France 67:92-99.

Watmough, R. H. 1974. Biology and behaviour of carpenter bees in South-
ern Africa. Journal of the Entomological Society of South Africa
37:261-281.

Weins, J. J. and T. W. Reeder. 1995. Combining data sets with different
numbers of taxa for phylogenetic analysis. Systematic Biology 44:548—
2)3%2),

Westwood, J. O. 1838. Description of a new genus of exotic bees. Transac-
tions of the Entomological Society of London 2:112-113.

Westwood, J. O. 1840. In J. Duncan, The Natural History of Bees, in, W.
Jardine (ed.), The Naturalists’ Library. Vol. 26. Entomology, Vol. 6.
Edinburg: W. H. Lizars, viii+17-301 pp, 30 pls.

Westwood, J. O. 1842. [In E. Donovan] Natural History of the Insects of
India, containing upwards of two hundred and twenty figures and descrip-
tions, a new edition, brought down to the present state of the science, with
systematic characters of each species, synonyms, indexes, and other addi-
tional matters by J. O. Westwood, 2nd ed. London: Henry G. Bohn. pp.
1-102.

44 ScieNTIFIC PAPERS, NATURAL History Museum, THE UNIVERSITY OF KANSAS

APPENDIX 1
CHARACTER CHANGES AND NODE SUPPORT FOR THE TREE SHOWN IN FiGure 10 A (ANALysIS 1, CHARACTER WEIGHTS EQUAL)

Accelerated transformation (ACCTRAN) of characters
has been assumed. Node numbers are those shown in Fig-
ure 10; numbers in parentheses indicate character states
as they are numbered in the text, and the direction of char-
acter state change. Arrows in parentheses with single lines
(—) indicate losses and gains in the character do not occur
along subsequent nodes. Arrow in parentheses with double
lines (=) indicate that reversals or gains occur among sub-
sequent nodes, M=male, F=female, B=both sexes.

NODES 95-494: 11. One spine on outer apex of hind tibia M (01); 19.
Basitibial plate apex bifid F (0—); 21. Basitibial plate reduced F (0-1);
41. Paired marginal spines of pygidial plate F (0—1); 47. Gonostylus
slender, short M (0—)).

NODES 94-487: 24. Metapostnotum not evident M (01); 26. Mesosomal

gland reservoir M (0—).

NODES 87-486: 26. Mesosomal gland reservoir M (1-2); 52.

Ventroapical plate of gonocoxite M (0=2).

NODES 86-477: 50. Ventromedial margin of gonocoxites diverge M

(01); 51. Spine on posteroventral margin of gonocoxite M (0=>1).

NODES 77-61: 3. Subapical tooth on mandible F (01); 43. Lateral

edge of penis valve modified M (0>)).

NODES 61-55: 36. Apical margin of Sl B (0=2); 49. Medially project-

ing lobe on apex of gonostylus M (0=1).

NODES 55-50: 32. Number of tergal graduli F (01).

NODES 50-49: 32. Graduli on T1-3 F (0=2); 49. Medially projecting

lobe on apex of gonostylus M (1=>2).

NODES 55-54: 54. Spines on posterior margin T7 M (0>1).

NODES 54-453: 18. Basitibial plate apex bifid M (0=>)).

NODES 53-52: 17. Modified margins of basitibial plate F (0—1); 51.

Spine on posteroventral margin of gonocoxite M (1=0); 52.

Ventroapical plate of gonocoxite M (2>)).

NODES 52-451: 26. Mesosomal gland reservoir lost M (2-0).

NODES 61-60: 26. Mesosomal gland reservoir M (2—3); 27. Begin-

ning of thoracic declivity M (0=1).

NODES 60-59: 6. Epistomal suture F (0=1); 48. Setae on apex of

gonostylus M (1=0).

NODES 59-458: 4. Apex of mandible tridentate F (0-1); 5. Upper rutellar

tooth and preapical tooth F (01); 12. Number of spines on outer

apex of hind tibia F (01); 40. Appressed setae near metasomal spi-
racles M (0>)).

NODES 58-457: 32. Graduli on TI-3 F (0=2); 38. Medial carina of

metasomal sterna F (0=1).

NODES 57-56: 1. Vestiture M (0=1); 4. Apex of mandible bidentate F
(1-0); 5. Upper rutellar tooth and preapical tooth F (10); 39.
Pregradular areas of sterna M (0=>1).

NODES 77—976: 11. Two spines on outer apex of hind tibia M (1=2); 12.
Number of spines on outer apex of hind tibia F (0—1); 23. Beginning
of thoracic declivity F (0—1); 47. Length of gonostylus M (10); 48.
Setae on apex of gonostylus M (1-0).

NODES 76-62: 22. Beginning of thoracic declivity M (0=1); 24.
Metapostnotum M (1=0).

NODES 76-475: 25. Metapostnotum F (0—]); 26. Mesosomal gland res-
ervoir M (2=5); 35. Medial groove of T1 F (0—1); 41. Marginal spines
of pygidial plate F (10); 53. Inner basal margin of venter of
gonocoxite M (0).

NODES 75—63: 8. Malar space F (0—); 23. Beginning of thoracic de-

clivity F (10); 39. Pregradular areas of sterna M (0-1); 40. Ap-

pressed setae near metasomal spiracles M (0=]).

NODES 75-474: 6. Epistomal suture F (01); 7. Supraocellar pit F (01);

28. Propodeal tooth B (0—1); 30. Shape of gradulus on T1 F (OI); 32.

Graduli on TI-3 F (02); 36. Apical margin of S1 B (0=1).

NODES 74-473: 19. Basitibial plate apex F (1=0); 22. Beginning of tho-

racic declivity M (0=1); 31. Distribution of tergal graduli M (0=2);

33. Anterior face of T1 M (0=1); 34. Anterior face of Tl F (0=)).

NODES 73-472: 22. Beginning of thoracic declivity M (1=2); 23. Begin-

ning of thoracic declivity F 1=2); 34. Anterior face of Tl F (12); 47.

Length of gonostylus M (0=)).

NODES 72—>66: 14. Triangular hind trochanter F (0>1); 29. Tegula shape

M (01); 38. Medial carina of metasomal sterna F (0—1); 53. Inner

basal margin of venter of gonocoxite M (10); 55. Penis valves M

(01).

NODES 66-64: 16. Spine on ventral surface of hind femur M (0=2); 31.

Distribution of tergal graduli M (2>]).

NODES 66-65: 6. Epistomal suture not raised F (1=0); 25.

Metapostnotum distinct F (10); 26. Mesosomal gland reservoir M

(5—1); 28. Propodeal tooth lost B (10); 45. Penis M (0>)).

NODES 72-471: 23. Beginning of thoracic declivity F (23); 49. Medi-

ally projecting lobe on apex of gonostylus M (0—).

NODES 71—470: 3. Subapical tooth on upper carina of mandible F (01);

37. Medial carina of S6 M (0).

NODES 70-69: 6. Epistomal suture F (10); 7. Supraocellar pit not

deeply excavated F (10).

NODES 69-68: 1. Vestiture M (0=1); 2. Subapical tooth on mandible M

(0—)); 31. Distribution of tergal graduli M (21); 40. Appressed se-

tae near metasomal spiracles M (0=]1); 49. Medially projecting lobe

on apex of gonostylus M (1=2).

NODES 68-67: 51. Spine lost on posteroventral margin of gonocoxite

M (1=30).

NODES 86-85: 32. Graduli on TI-3 F (0=2); 40. Appressed setae near

metasomal spiracles M (0-1).

NODES 85-84: 17. Tubercles on posterior margin of basitibial plate F

(02); 24. Metapostnotum distinct M (1-0).

NODES 84-383: 39. Pregradular areas of sterna with setal areas M (0—)).

NODES 83-478: 3. Subapical tooth on upper carina of mandible F (0>1);
25. Metapostnotum not defined F (0—)).

NODES 83-82: 11. No spines on outer apex of hind tibia M (10).

NODES 82-481: 24. Metapostnotum not defined M (0—1); 49. Medially

projecting lobe on apex of gonostylus M (0>1).

NODES 81-80: 16. Tubercle on ventral surface of hind femur M (0>1).

NODES 80-479: 6. Epistomal suture raised F (01).

NODES 94-493: 36. Apical margin of Sl B (0-2).

NODES 93-4839: 20. Basitibial plate foveate F (0—1); 50. Ventromedial

margins of gonocoxites diverge M (0=>)).

NODES 89-88: 41. Many spines along pygidial plate margin F (1=2);

42. Gonostylus M (0=91); 44. Bifid apex of penis valves M (01); 48.

Setae on apex of gonostylus M (1=0).

NODES 93-492: 4. Apex of mandible tridentate F (01); 1 0. Facial macu-
lations absent M (0=1); 40. Appressed setae near metasomal spiracles
M (0-1); 54. Spines on posterior margin T7 M (0-1).

NODES 92-491: 17. Tubercles along margins of basitibial plate F (0>1);
21. Extent of basitibial plate F (1=0); 26. Mesosomal gland reservoir
M (0); 48. Setae on apex of gonostylus M (1=>0); 55. Penis valves
M (0-1).

NODES 91— 90: 11. Two spines on outer apex of hind tibia M (1=2); 12.
Two spines on outer apex of hind tibia F (0=1); 40. Appressed setae
near metasomal spiracles M (1—0); 53. Inner basal margin of venter
of gonocoxite M (0=>1); 54. Posterior margin T7 M (1-0).

SUBGENERA AND GENERA OF CARPENTER BEES, XYLOCOPINI 45

APPENDIX 2
CHARACTER CHANGES AND NODE SUPPORT FOR TREE SHOWN IN Fic. 11 A (ANALYsIs 3. ISLAND 1, CHARACTER WEIGHTS EQUAL)

Abbreviations and conventions as in Appendix 1.

NODES 95-94: 11. One spine on outer apex of hind tibia M (0-1); 19.
Basitibial plate apex bifid F (01); 21. Extent of basitibial plate F
(0—1); 41. Paired marginal spines of pygidial plate F (0-1); 47.
Gonostylus slender, short M (0-1).

NODES 72-63: 26. Mesosomal gland reservoir M (0-2); 36. Apical
margin Sl emarginate B (02); 49. Medially projecting lobe on apex
of gonostylus M (051).

NODES 63-60: 13. Tibial spine F (01); 50. Ventromedial margin of
gonocoxites diverge M (0>1).

NODES 60-59: 3. Subapical tooth on upper carina of mandible F (0>1);
24. Metapostnotum riot evident M (0=1); 43. Lateral edge of penis
valve modified M (0=1); 52. Ventroapical plate of gonocoxite M
(0-1).

NODES 59-58: 13. Tibial spine F (10); 52. Ventroapical plate of
gonocoxite M (1—2).

NODES 58-57: 51. Spine on posteroventral margin of gonocoxite M
(0=>)).

NODES 57-56: 49. Medially projecting lobe on apex of gonostylus M
(1-0).

NODES 56—53: 26. Mesosomal gland reservoir M (2-3); 27. Beginning
of thoracic declivity M (0=>1); 36. Apical margin of Sl entire B (250).

NODES 53-52: 6. Epistomal suture F (01); 48. Setae on apex of
gonostylus M (1=0).

NODES 52-951: 4. Apex of mandible tridentate F (01); 5. Upper rutellar
tooth and preapical tooth fused F (0—l); 12. Two spines on outer
apex of hind tibia F (01); 40. Appressed setae near metasomal spi-
racles M (0—)).

NODES 51-50: 32. Graduli on Tl-3 F (0=2); 38. Medial carina of
metasomal sterna F (0=1).

NODES 50—>49: 1. Vestiture M (0=1); 4. Apex of mandible bidentate F
(1-0); 5. Upper rutellar tooth and preapical tooth (10); 39.
Pregradular areas of sterna without setal areas M (0>)).

NODES 56—>55: 2. Subapical tooth on mandible M (0—); 49. Medially
projecting lobe on apex of gonostylus M (0-2).

NODES 55-54: 32. Graduli on T1-3 F (0=2).

NODES 63-62: 4. Apex of mandible tridentate F (0=1); 10. Facial macu-
lations absent M (0=1); 40. Appressed setae near metasomal spiracles
M (0=91); 54. Spines on posterior margin T7 M (0=>)).

NODES 62-61: 17. Margins of basitibial plate F (01); 18. Basitibial plate
apex bifid M (01); 32. Number of tergal graduli F (0—)).

NODES 72-71: 17. Margins of basitibial plate F (02); 32. Graduli on
TI-3 F (0=2); 52. Ventroapical plate of gonocoxite M (0=>2).

NODES 71-970: 39. Pregradular areas of sterna M (0=1); 40. Appressed
setae near metasomal spiracles M (0>]).

NODES 70-69: 24. Metapostnotum not defined M (0>]1); 25.
Metapostnotum not defined F (0=1); 26. Mesosomal gland reservoir
M (0).

NODES 69-68: 26. Mesosomal gland reservoir M (1—2).

NODES 68-67: 49. Medially projecting lobe on apex of gonostylus M
(0=)).

NODES 67— 66: 11. No spines on outer apex of hind tibia M (10); 25.
Metapostnotum F (10).

NODES 66—>65: 16. Tubercle on ventral surface of male hind femur (0=)).

NODES 65-64: 6. Epistomal suture raised F (0=>)).

NODES 94-93: 11. Two spines on outer apex of hind tibia M (1—2); 24.
Metapostnotum not defined M (0=1); 25. Metapostnotum not defined
F (0=)).

NODES 93-92: 12. Two spines on outer apex of hind tibia F (O=]); 36.
Apical margin of SI B (0—2); 50. Ventromedial margin of gonocoxites
diverge M (0=91); 52. Ventroapical plate of gonocoxite M (0—)).

NODES 92-491: 48. Setae on apex of gonostylus M (1=0); 52.
Ventroapical plate of gonocoxite M (1—2); 53. Inner basal margin of
venter of gonocoxite M (0-1).

NODES 91-477: 24. Metapostnotum defined M (1=0); 25.
Metapostnotum defined F (1=0).

NODES 77-975: 10. Facial maculations absent M (0—1); 17. Margins of
basitibial plate F (0—1); 42. Gonostylus M (0—1); 52. Ventroapical
plate of gonocoxite M (2-0).

NODES 75-73: 12. Two spines on outer apex of hind tibia F (1=0); 41.
Marginal spines of pygidial plate F (1=2); 44. Apex of penis valves
M (0>)).

NODES 75-74: 21. Extent of basitibial plate F (10); 26. Mesosomal
gland reservoir M (0=1); 50. Divergence of ventromedial margin of
gonocoxites M (1=>0); 55. Penis valves M (0>1).

NODES 77-976: 22. Beginning of thoracic declivity M (01); 23. Begin-
ning of thoracic declivity F (01); 26. Mesosomal gland reservoir M
(0-2); 36. Apical margin of Sl B (20). 53. Inner basal margin of
venter of gonocoxite M (10).

NODES 91-490: 19. Basitibial plate apex F (10); 26. Mesosomal gland
reservoir M (0=5); 35. Medial groove of Tl F (0=1); 47. Length of
gonostylus M (1=0).

NODES 90-978: 8. Malar space F (01); 39. Pregradular areas of sterna
M (01); 40. Appressed setae near metasomal spiracles M (01).
NODES 90-389: 6. Epistomal suture F (0=>1); 23. Beginning of thoracic
declivity F (0=1); 32. Graduli on TI-3 F (0=2); 33. Anterior face of Tl

M (0-1); 36. Apical margin of SI B(2=1).

NODES 89-388: 7. Supraocellar pit F (0=1); 28. Propodeal tooth B (0O=1);
30. Shape of gradulus on TI (01); 37. Medial carina of S6 M (0O>1);
41. Marginal spines of pygidial plate F (1=0); 51. Spine on
posteroventral margin of gonocoxite M (0>]).

NODES 88-387: 22. Beginning of thoracic declivity M (0=2); 23. Begin-
ning of thoracic declivity F (12); 34. Anterior face of Tl F (0=2); 47.
Length of gonostylus M (0=)).

NODES 87-81: 14. Triangular hind trochanter F (0>1); 16. Spine on ven-
tral surface of hind femur M (0-2); 29. Shape of tegula M (0>1); 38.
Medial carina of metasomal sterna F (0-1); 53. Inner basal margin of
venter of gonocoxite (1-30); 55. Penis valves M (0=1)).

NODES 81-980: 26. Mesosomal gland reservoir M (5-1); 31. Distribu-
tion of tergal graduli M (0-1); 37. Medial carina of S6 M (1=0); 50.
Divergence of ventromedial margin of gonocoxites M (1-0); 51. Spine
on posteroventral margin of gonocoxite M (1-0).

NODES 80-379: 6. Epistomal suture not raised F (10); 16. Unmodified
ventral surface of hind femur M (2-90); 25. Metapostnotum defined
F (10); 28. Propodeal tooth absent B (10); 31. Distribution of ter-
gal graduli M (1-2); 45. Penis M (0-1).

NODES 87-986: 3. Subapical tooth on upper carina of mandible F (0-1);
23. Beginning of thoracic declivity F (23); 49. Medially projecting
lobe on apex of gonostylus M (0—I).

NODES 86-383: 1. Vestiture M (0=>1); 2. Subapical tooth on mandible M
(0-1); 40. Appressed setae near metasomal spiracles M (0=1); 49.
Medially projecting lobe on apex of gonostylus M (1-2).

NODES 83-482: 6. Epistomal suture F (10); 7. Supraocellar pit F 1=0);
31. Distribution of tergal graduli M (01); 51. Spine on posteroventral
margin of gonocoxite M (1=0).

NODES 86-85: 31. Distribution of tergal graduli M (0=2).

NODES 85-984: 34. Anterior face of Tl F (2>1).

46 ScIENTIFIC PAPERS, NATURAL History Museum, THE UNIVERSITY OF KANSAS

APPENDIX 3
CHARACTER CHANGES AND NODE SUPPORT FOR TREE SHOWN IN FiGureE 13 A (ANALYSIS 3, ISLAND 5, CHARACTER WEIGHTS EQUAL).

Abbreviations and conventions as in Appendix 1.

NODES 95-494: 11. One spine on outer apex of hind tibia M (0-1); 19.
Basitibial plate apex bifid F (0-1); 21. Extent of basitibial plate F
(0-1); 41. Paired marginal spines of pygidial plate F (0-1); 47.
Gonostylus slender, short M (0-1).

NODES 94-85: 36. Apical margin of S1 (0=2).

NODES 85-363: 26. Mesosomal gland reservoir M (0-2); 49. Medially

projecting lobe on apex of gonostylus M (0=1).

NODES 63-360: 13. Two outer tibial spines F (0-1); 50. Ventromedial

margin of gonocoxites diverge M (01).

NODES 60-359: 3. Subapical tooth on upper carina of mandible F (01);

24. Metapostnotum not defined M (0=1); 43. Lateral edge of penis

valve modified M (0=1); 52. Ventroapical plate of gonocoxite M (0-1).

NODES 59-58: 13. One outer tibial spine F (10); 52. Ventroapical plate

of gonocoxite M (1-2).

NODES 58-57: 51. Spine on ventroposterior margin of gonocoxite M

(0=)).

NODES 57-56: 49. Medially projecting lobe on apex of gonostylus M

(10).

NODES 56-53: 26. Mesosomal gland reservoir M (2-3); 27. Beginning

of thoracic declivity M (0=1); 36. Apical margin of SI] B (20).

NODES 53-452: 6. Epistomal suture F (0=1); 48. Setae on apex of
gonostylus M (10). NODES 52-951: 12. One spine on outer apex of
hind tibia F (0=1); 32. Number of tergal graduli F (02); 38. Medial
carina of metasomal sterna F (0-1); 40. Appressed setae near
metasomal spiracles M (0=1).

NODES 51-49: 1. Vestiture M (01); 39. Pregradular areas of sterna M

(0>)).

NODES 51-50: 4. Number of apical teeth of mandible F (01); 5. Up-

per rutellar and preapical tooth fused F (01); 26. Mesosomal gland

reservoir M (3-8).

NODES 56-55: 2. Subapical tooth on mandible M (0-1); 49. Medially

projecting lobe on apex of gonostylus M (0-2).

NODES 55-254: 32. Graduli on T1-3 F (0=2).

NODES 63-362: 4. Apex of mandible tridentate F (01); 10. Facial macu-

lations absent M (0=1); 40. Appressed setae near metasomal spiracles

M (0=51); 54. Spines on posterior margin T7 M (01).

NODES 62-61: 17. Margins of basitibial plate F (01); 18. Basitibial plate

apex bifid M (0=1); 32. Number of tergal graduli F (0-1).

NODES 85-84: 11. Two spines on outer apex of hind tibia M (1-2); 12.

Two spines on outer apex of hind tibia F (0-1).

NODES 84-82: 48. Setae on apex of gonostylus M (1=0).

NODES 82-80: 50. Ventromedial margin of gonocoxites diverge M (01).

NODES 80-64: 12. One spine on outer apex of hind tibia F (1-0); 41.

Marginal spines of pygidial plate F (12); 42. gonostylus M (0=1);

44. Apex of penis valves M (0=1).

NODES 80-379: 19. Basitibial plate apex entire F (10); 22. Beginning of

thoracic declivity M (0-1); 23. Beginning of thoracic declivity F (0=1);

26. Mesosomal gland reservoir M (0-2); 36. Apical margin of S1 B

(2-40); 52. Ventroapical plate of gonocoxite M (0-1).

NODES 79-478: 47. Length of gonostylus M (1=0); 52. Ventroapical plate

of gonocoxite M (12).

NODES 78-77: 22. Beginning of thoracic declivity M (1-0); 24.
Metapostnotum defined M (0-—]); 25. Metapostnotum defined F (0-1);
26. Mesosomal gland reservoir M (2-45); 35. Medial groove of Tl F
(0=1); 53. Inner basal margin of venter of gonocoxite M (0=1).

DO

NODES 77-65: 8. Long malar space F (0=1); 23. Beginning of thoracic
declivity F (1=0); 39. Pregradular areas of sterna M (0=1); 40. Ap-
pressed setae near metasomal spiracles M (0=>)).

NODES 77-376: 6. Epistomal suture F (01); 32. Graduli on TI-3 F (02);
33. Anterior face of Tl M (0-1); 36. Apical margin of SI B (01).
NODES 76-75: 7.Supraocellar pit F (01); 28. Propodeal tooth B (0S);
30. Shape of gradulus on Tl F (01); 37. Medial carina of S6 M (0=));
41. Marginal spines of pygidial plate F (1=0); 51. Spine on

ventroposterior margin of gonocoxite M (0=1).

NODES 75-974: 22. Beginning of thoracic declivity M (02); 23. Begin-
ning of thoracic declivity F (12); 34. Anterior face of Tl F (0=2); 47.
Length of gonostylus M (01).

NODES 74-68: 14. Triangular hind trochanter F (0=1); 16. Spine on ven-
tral surface of hind femur M (0-2); 29. Shape of tegula M (0=1); 38.
Medial carina of metasomal sterna F (0-1); 53. Inner basal margin of
venter of gonocoxite M (1-0); 55. Penis valves M (0=1).

NODES 68-967: 26. Mesosomal gland reservoir M (5-1); 31. Distribu-
tion of tergal graduli M (0-1); 37. Medial carina of S6 M (1=30); 50.
Divergence of ventromedial margin of gonocoxites M (10); 51. Spine
on ventroposterior margin of gonocoxite M (10).

NODES 67-66: 6. Epistomal suture F (1=0); 16. Unmodified ventral
surface of hind femur M (2-0); 25. Metapostnotum defined F (10);
28. Propodeal tooth B (10); 31. Distribution of tergal graduli M (1
=2); 45. Penis M (0>1).

NODES 74-973: 3. Subapical tooth on upper carina of mandible F (0-1);
23. Beginning of thoracic declivity F (23); 49. Medially projecting
lobe on apex of gonostylus M (0-1).

NODES 73-970: 1. Vestiture M (0=1); 2. Subapical tooth on mandible M
(0-1); 40. Appressed setae near metasomal spiracles M (0=1); 49.
Medially projecting lobe on apex of gonostylus M (1-2).

NODES 70-69: 6. Epistomal suture F (10); 7. Supraocellar pit F 10);

31. Distribution of tergal graduli M (01); 51. Spine on posteroventral

margin of gonocoxite M (1=0).

NODES 73-972: 31. Distribution of tergal graduli M (0-2).

NODES 72-971: 34. Anterior face of Tl F (21).

NODES 82-981: 10. Facial maculations M (0=1); 17. Margins of basitibial
plate F (0=1); 21. Extent of basitibial plate F (10); 26. Mesosomal
gland reservoir M (01); 53. Inner basal margin of venter of gonocoxite
M (01); 55. Penis valves M (0=)).

NODES 84-83: 24. Metapostnotum not defined M (0=1); 25.

Metapostnotum not defined F (0=>1).

NODES 94-493: 17. Margins of basitibial plate F (02); 32. Graduli on

TI-3 F (0=2); 52. Ventroapical plate of gonocoxite M (0=2).

NODES 93-492: 11. One spine on outer apex of hind tibia M (1-0); 39.

Pregradular areas of sterna with setal patches M (0-1); 40. Appressed

setae near metasomal spiracles M (0=)).

NODES 92-991: 24. Metapostnotum not defined M (01); 26. Mesosomal

gland reservoir M (0=2).

NODES 91-987: 11. One spine on outer apex of hind tibia M (0-1); 25.

Metapostnotum not defined F (0=>)).

NODES 87-86: 3. Subapical tooth on upper carina of mandible F (0=>1).

NODES 91-90: 49. Medially projecting lobe on apex of gonostylus M
(01).

NODES 90-489: 16. Tubercle on ventral surface of hind femur M (0>1).

NODES 89-88: 6. Epistomal suture F (0=1).

SUBGENERA AND GENERA OF CARPENTER BEES, X YLOCOPINI 47

APPENDIX 4
CHARACTER CHANGES AND NODE SUPPORT FOR TREE SHOWN IN FiGure 15 (ANALYsIS 3. ISLAND 6, CHARACTER WEIGHTS EQUAL).

Abbreviations and conventions as in Appendix 1.

NODES 95-494: 11. One spine on outer apex of hind tibia M (O01); 19.
Basitibial plate apex bifid F (01); 21. Extent of basitibial plate F (0);
36. Apical margin of S1 (0-2); 41. Paired marginal spines of pygidial
plate F (0-1); 47. Gonostylus slender, short M (01).

NODES 92-972: 48. Setae on apex of gonostylus M (0=>1).

NODES 72-363: 26. Mesosomal gland reservoir M (0-2); 49. Medially
projecting lobe on apex of gonostylus M (0-1).

NODES 63-60: 13. Two spines on outer apex of hind tibia F (0-1); 50.
Ventromedial margin of gonocoxites diverge M (01).

NODES 60-59: 3. Subapical tooth on upper carina of mandible F (O>1);
24. Metapostnotum not defined M (0=1); 43. Lateral edge of penis
valve modified M (01); 52. Ventroapical plate of gonocoxite M (0).

NODES 59-58: 13. One spine on outer apex of hind tibia F (10); 52.
Ventroapical plate of gonocoxite M (1-2).

NODES 58-57: 51. Spine on ventroposterior margin of gonocoxite M
(O>).

NODES 57-56: 49. Medially projecting lobe on apex of gonostylus M
(130).

NODES 56-53: 26. Mesosomal gland reservoir M (2-3); 27. Beginning
of thoracic declivity M (01); 36. Apical margin of S1 M (20).

NODES 53-352: 6. Epistomal suture F (0=1); 48. Setae on apex of
gonostylus M (10).

NODES 52-351: 4. Apex of mandible tridentate F (0-1); 5. Upper rutellar
tooth and preapical tooth fused F (01); 12. Two spines on outer apex
of hind tibia F (01); 40. Appressed setae near metasomal spiracles
M (0>1).

NODES 51-50: 32. Graduli on T1-3 F (0=>2); 38. Medial carina of
metasomal sterna F (0>1).

NODES 50-49: 1. Vestiture M (0=>1); 4. Apex of mandible bidentate F
(130); 5. Upper rutellar tooth and preapical tooth separate F (10);
39. Pregradular areas of sterna M (0=)).

NODES 56-555: 2. Subapical tooth on mandible M (0-1); 49. Medially
projecting lobe on apex of gonostylus M (0-2).

NODES 55-354: 32. Graduli on TI-3 F (0=2).

NODES 63-462: 4. Apex of mandible tridentate F (01); 1 0. Facial macu-
lations absent M (0=1); 40. Appressed setae near metasomal spiracles
M (0=51); 54. Spines on posterior margin T7 M (0=1).

NODES 62-361: 17. Margins of basitibial plate F (0=1); 18. Basitibial plate
apex bifid M (0=1); 32. Number of tergal graduli F (0-1).

NODES 72-371: 17. Margins of basitibial plate F (0=2); 32. Graduli on
T1-3 F (0=2); 36. Apical margin of S1 B (2-30); 52. Ventroapical plate
of gonocoxite M (0-2).

NODES 71-470: 39. Pregradular areas of sterna M (0=1); 40. Appressed
setae near metasomal spiracles M (0).

NODES 70-369: 24. Metapostnotum not defined M (0=1); 25.
Metapostnotum not defined F (01); 26. Mesosomal gland reservoir
M (OI).

NODES 69-368: 26. Mesosomal gland reservoir M (1-2).

NODES 68-367: 49. Medially projecting lobe on apex of gonostylus M
(0O>)).

NODES 67-366: 11. Two spines on outer apex of hind tibia M (1=0); 25.
Metapostnotum defined F (10).

NODES 66-365: 16. Tubercle on ventral surface of hind femur M (0>1).

NODES 65-64: 6. Epistomal suture F (01).

NODES 92-991: 11. Two spines on outer apex of hind tibia M (1-2); 12.
One spine on outer apex of hind tibia F (0-1); 26. Mesosomal gland
reservoir M (0—l).

NODES 91-973: 10. Facial maculations absent M (0=1); 17. Margins of
basitibial plate F (0=>1); 21. Extent of basitibial plate F (10); 53. In-
ner basal margin of venter of gonocoxite M (0=1); 55. Penis valves M
(0O=)).

NODES 91-490: 36. Apical margin of SI B (20); 50. Ventromedial mar-

gin of gonocoxites diverge M (0=1); 52. Ventroapical plate of

gonocoxite M (0-1).

NODES 90-974: 22. Beginning of thoracic declivity M (0=1); 23. Begin-

ning of thoracic declivity F (01); 26. Mesosomal gland reservoir M

(12).

NODES 90-489: 24. Metapostnotum not defined M (01); 25.

Metapostnotum not defined F (0>)).

NODES 89-487: 19. Basitibial plate apex F (10); 26. Mesosomal gland
reservoir M (1-95); 35. Medial groove of T1 F (0=>1); 47. Length of
gonostylus M (1=>0); 52. Ventroapical plate of gonocoxite M (12);
53. Inner basal margin of venter of gonocoxite M (0=)).

NODES 87-975: 8. Long malar space F (0=1); 39. Pregradular areas of
sterna M (0=1); 40. Appressed setae near metasomal spiracles M
(0=)).

NODES 87-86: 6. Epistomal suture F (0=1); 23. Beginning of thoracic
declivity F (0=1); 32. Graduli on T1-3 F (0-2); 33. Anterior face of Tl
M (01); 36. Apical margin of SI B (01).

NODES 86-85: 7. Supraocellar pit F (051); 28. Propodeal tooth B (01);
30. Shape of gradulus on TI F (0-1); 37. Medial carina of S6 M (0-1);
41. Marginal spines of pygidial plate F (1>0); 51. Spine on
ventroposterior margin of gonocoxite M (0=)).

NODES 85-84: 22. Beginning of thoracic declivity M (0=2); 23. Begin-
ning of thoracic declivity F (12); 34. Anterior face of T1 F (0=2); 47.
Length of gonostylus M (0).

NODES 84-378: 14. Triangular hind trochanter F (0=1); 29. Shape of
tegula M (01); 37. Medial carina of S6 M (1-0); 38. Medial carina of
metasomal sterna F (0-1); 53. Inner basal margin of venter of
gonocoxite M (1-90); 55. Penis valves M (01).

NODES 78-476: 16. Spine on ventral surface of hind femur M (0=2).

NODES 78-477: 6. Epistomal suture F (1=0); 25. Metapostnotum de-
fined F (10); 26. Mesosomal gland reservoir M (5-1); 28. Propodeal
tooth M (1=30); 31. Distribution of tergal graduli M (0=2); 45. Penis
M (0-1).

NODES 84-383: 3. Subapical tooth on upper carina of mandible F (0-1);

23. Beginning of thoracic declivity F (23); 49. Medially projecting

lobe on apex of gonostylus M (0-1).

NODES 83-380: 1. Vestiture M (0=>1); 2. Subapical tooth on mandible M

(0-1); 40. Appressed setae near metasomal spiracles M (01); 49.

Medially projecting lobe on apex of gonostylus M (1-2).

NODES 80-379: 6. Epistomal suture F (10); 7. Supraocellar pit F 1=0);

31. Distribution of tergal graduli M (0=1); 51. Spine on

ventroposterior margin of gonocoxite M (1=0).

NODES 83-482: 31. Distribution of tergal graduli M (0=2).

NODES 82-81: 34. Anterior face of Tl F (21).

NODES 89-388: 48. Setae on apex of gonostylus M (0=>1).

NODES 94-493: 41. Marginal spines of pygidial plate F (12); 42.
Gonostylus M (0=51); 44. Apex of penis valves M (0-1); 50. Ventro-
medial margin of gonocoxites diverge M (01).

PUBLICATIONS OF THE
NATURAL HISTORY MUSEUM, THE UNIVERSITY OF KANSAS

The University of Kansas Publications, Museum of Natural History, beginning
with Volume 1 in 1946, was discontinued with Volume 20 in 1971. Shorter research
papers formerly published in the above series were published as The University of
Kansas Natural History Museum Occasional Papers until Number 180 in Decem-
ber 1996. The Miscellaneous Publications of The University of Kansas Natural His-
tory Museum began with Number 1 in 1946 and ended with Number 68 in Febru-
ary 1996. Monographs of The University of Kansas Natural History Museum were
initiated in 1970 and discontinued with Number 8 in 1992. The University of Kan-
sas Science Bulletin, beginning with Volume 1 in 1902, was discontinued with Vol-
ume 55 in 1996. The foregoing publication series are now combined in a new series
entitled Scientific Papers, Natural History Museum, The University of Kansas, be-
gun with Number 1 in 1997. Special Publications began in 1976 and continue as an
outlet for longer contributions and are available by purchase only. All manuscripts
are subject to critical review by intra- and extramural specialists; final acceptance is
at the discretion of the editor.

The publication is printed on acid-free paper. Publications are composed us-
ing Microsoft Word® and Adobe PageMaker® on a Macintosh computer and are
printed by The University of Kansas Printing Services.

Institutional libraries interested in exchanging publications may obtain the Sci-
entific Papers, Natural History Museum, The University of Kansas, by addressing
the Exchange Librarian, The University of Kansas Libraries, Lawrence, Kansas 66045-
2800, USA. Available back issues of The University of Kansas Science Bulletin may
be purchased from the Library Sales Section, Retrieval Services Department, The
University of Kansas Libraries, Lawrence, Kansas 66045-2800, USA. Available is-
sues of former publication series, Scientific Papers, and Special Publications of the
Natural History Museum can be purchased from the Office of Publications, Natu-
ral History Museum, The University of Kansas, Lawrence, Kansas 66045-2454, USA.
Purchasing information can be obtained by calling (913) 864-4450, fax (913) 864-
5335, or e-mail (kunhm@ukans.edu). VISA and MasterCard accepted; include ex-
piration date.

Series Epitor: William E. Duellman

PRINTED BY
THE UNIVERSITY OF KANSAS PRINTING SERVICES

LAWRENCE, KANSAS