The Auk 110(2):304-324, 1993 PHYLOGENY, BIOGEOGRAPHY, AND EVOLUTION OF THE BROADBILLS (EURYLAIMIDAE) AND ASITIES (PHILEPITTIDAE) BASED ON MORPHOLOGY RICHARD O. ?RUM Museumof Natural Historyand Departmentof Systematicsand Ecology, Universityof Kansas,Lawrence, Kansas 66045, USA ABsTRACT.--Phylogeneticanalysis of syringealmorphology and two osteologicalcharacters indicatesthat the broadbills (Eurylaimidae)are not monophyletic,but consistof four clades with successivelycloser relationships to the Madagascanasities (Philepittidae). An analysis of thesedata combined with hindlimb myologycharacters described by Raikow(1987) yields the sameresult. The sistergroup to Philepittaand Neodrepanisis the African broadbill Pseu- docalyptomena.The sistergroup to this cladeincludes all of the Asian broadbills,except the monophyleticgenus Calyptomena. The African genusSmithornis is the sistergroup to all other broadbillsand asities.A biogeographicanalysis indicates that the Madagascanendemics share a most-recentbiogeographic connection with the central African genusPseudocalyptomena. Phylogeneticassociations between transitions in bill morphologyand diet indicatethat bill morphologieshave evolved both in associationwith evolution of frugivory and nectarivory, and in apparentresponse to intrinsicfactors within the contextof frugivorousand insectiv- orousdiets. A phylogeneticclassification of the broadbillsand asitiesis proposedin which all broadbillsand asitiesare placed in five subfamiliesof the Eurylaimidae,and the separate family Philepittidae is abandoned.Received 27 January1992, accepted 23 November1992. THE BROADBILLS(Eurylaimidae), asities (Phi- Cymbirhynchus,Serilophus, Psari$omus, and Calyp- lepittidae), and pittas (Pittidae) form a clade of tomena)and two genera are found in Africa Old World suboscinepasserines (Sibley et al. (Smithornisand Pseudocalyptomena).The asities 1982, Raikow 1987). They are broadly distrib- include four speciesin two genera (Philepitta uted in tropical Africa, Madagascar,Asia, and and Neodrepanis)that are restricted to Mada- the Australo-Papuanregion. Although they are gascar.The 23 to 31 speciesof pittasare usually much less diverse than the New World subos- placed in the single genus Pitta (Mayr 1979, cines, the Old World suboscines include an eco- Sibley and Monroe 1990).They range through- logically diversegroup of terrestrialand arbo- out the Old World tropics, but are most diverse real insectivores,frugivores, and nectarivores. in southern Asia. As one of the major basalpasserine clades, the The taxonomichistory of the Old World su- broadbills, asities,and pittas are an interesting boscinesand their position within the passer- group with which to investigate the biogeo- ines has been thoroughly reviewed by Raikow graphic history of the Old World tropical avi- (1987). Of particular interest here are the Af- fauna. In particular, the phylogenetic relation- rican and Madagascangenera that were origi- shipsof the asitiesmay provide insight into the nally classifiedas oscinesand, subsequently, biogeographichistory of Madagascar.In this were recognized to be suboscineson the basis paper, I present a phylogenetic analysisof the of syringeal morphology. Smithornisand Pseu- syringealmorphology of the broadbills(Eury- docalyptomenawere placed in the oscine fly- laimidae)and asities(Philepittidae), and usethis catcher family Muscicapidae (Sharpe 1901, phylogenetic hypothesis as a comparative Rothschild 1909). Subsequent anatomical ob- framework in investigations of the biogeo- servationsby Bates(1915) and Lowe (1924,1931) graphichistory and evolutionary ecology of the demonstrated that they were not oscines,but group. rather African broadbills. In the 1800s,Philepitta The broadbills include 14 currently recog- was associated with the Sturnidae, Paradisaei- nized biologicalspecies in eight genera(Peters dae, and Nectariniidae, but was confirmed by 1951, Sibley and Monroe 1990). Six genera are Forbes (1880b) to be a suboscine based on sy- restrictedto southernAsia (Corydon,Eurylaimus, ringealmorphology. Neodrepanis, the Madagas- 3O4 April 1993] Broadbilland Asity Phylogeny 305 can genusof sunbird-asities,was placedin the METHODS oscinefamily Nectariniidaeuntil Amadon (1951) I observed46 syringealspecimens of 13 speciesof madesyringeal observations demonstrating that broadbillsand asities.The sample included all but it was suboscine,and placed it in the Philepit- five speciesof broadbillsand asitiesin 9 of 10 genera. tidae. An additional 23 specimensof 10 speciesof Pittawere Although syringealmorphology was impor- examinedfor outgroupcomparison. The samplein- tant in the initial recognition of the Old World cluded 28 specimens that were cleared and double suboscines (Forbes 1880a, b, Bates 1915, Lowe stainedfor cartilageand bone (Dingerkusand Uhler 1924, 1931, K6ditz 1925, Amadon 1951), it has 1977, Cannell 1988) by Wesley E. Lanyon. The re- not been usedexplicitly to reconstructthe phy- maining unclearedsyringeal specimens were treated logeneticinterrelationships of thesebirds. Ames with reversible iodine stain (Bock and Shear 1972) for (1971)described the syringesof a broadsample resolutionof musclefibers. Complete descriptions of the syringeal morphology of broadbills and asities, of Old World suboscinesand recognizedthat and list of the specimensexamined are presentedin the asitiesand somebroadbills were quite sim- the Appendix. The syringesof Pitta, Smithornis,Ca- ilar in syringealmorphology. However, in the lyptomena,Psarisomus, and Cymbirhynchusare illustrat- eclecticspirit of the day, he concludedthat these ed in Figure 1, and those of Eurylaimus,Serilophus, similarities were primitive but still sufficient to Pseudocalyptomena,Philepitta, and Neodrepanisare supportplacement of the two familiestaxonom- shown in Figure 2. I alsoexamined suboscine skeletal ically near to one another. specimens from the American Museum of Natural Sibley et al. (1982) proposedthat broadbills History and the Field Museum of Natural History. and pittasform a cladethat is the sistergroup Syringes and skeletonswere observed under a Wild to the New World suboscines based on DNA- M5A binocular dissectingmicroscope and illustra- tions were prepared with a camera lucida. DNA hybridization. However, Sibley et al. For the phylogeneticanalysis, I acceptthe mono- (1982) lacked DNA samplesof the asities,and phyly of the cladeincluding broadbills and asities no molecularhypothesis for their phylogenetic basedon Raikow's(1987) observationof three shared relationshipshas been published. derived myological characters (original character Most recently, Raikow (1987) performed a numbersfrom Raikow 1987):(13.1) attenuateM. gas- thorough phylogenetic analysis of the hind- trocnemius pars medialis with a concavecranial mar- limb myology of the Old World suboscinesand gin; (14.1) restrictedinsertion of M. gastrocnemius; produceda well-resolvedhypothesis of phy- and (18.1) plantar viniculum present,subsequently logeny for the group using 23 hindlimb myo- reversed in Neodrepanis.(These three charactersare not included in the analysisto simplify the calcula- logical and 5 other morphologicalcharacters. Raikow concluded that: the Old World subos- tions.) For outgroupcomparison, I acceptthe mono- phyly of the Old World suboscineclade--including cinesexcluding the Acanthisittidaeform a clade; the Philepittidae, Eurylaimidae, and Pittidae--with the pitta,broadbill, and asity families are mono- the New World suboscines(including Furnarioidea phyletic; and the asitiesand broadbills are sister and Tyrannoidea)as their sistergroup (Feduccia 1974, groups.He alsopresented the first phylogenetic 1975,1976, Sibley et al. 1982,Raikow 1987,Sibley and hypothesis for the interrelationshipsof the Ahlquist 1990).The Acanthisittidaeare groupedby broadbills,placing the African genus Pseudo- some authors with the other Old World suboscines, calyptomenaand the Asian Calyptomenaas the but I acceptRaikow's (1987) hypothesis that the Acan- thisittidaeare the sistergroup to the oscinepasser- first and secondsister groups, respectively, to ines. Variationswithin the ingroup--the broadbills the rest of the genera in the family. and asities--werepolarized by outgroupcomparison Here, ! describethe syringealmorphology of to the pittas, the New World suboscines,and the os- the broadbills (Eurylaimidae) and asities (Phi- cines (Wiley 1981, Maddison et al. 1984). The New lepittidae),and presenta phylogeneticanalysis World suboscines and the oscines were not included of syringeal charactersand two osteological in the computercalculations because they were con- characters.These data are analyzed in combi- cluded to be primitive for all of the charactersana- nation with hindlimb myologyand other mor- lyzed (seecharacter descriptions). Syringeal variation was coded as 19 characters. phological characters described by Raikow Complex multistate characterswere coded as either (1987),and then this phylogenetichypothesis orderedor unordered,depending upon whether the is used as a historical framework for investi- ingroup variation represented a coherent series of gatingthe biogeographyand evolutionaryecol- states. Complex characterswere ordered if the de- ogy of the broadbills and asities. rived states implied a hierarchical transition series 306 RICHARD O. PRUM [Auk, Vol. 110 A B C T A1 B1 BI D E s B1 Fig. 1. Left dorsolateralview of syrinx of: (A) Pitta versicolor(AMNH 4378);(B) Smithornisrufolateralis (AMNH 2232);(C) Calyptomenaviridis (AMNH 7999);(D) Psarisomusdalhousiae (AMNH 2998);and