EVIDENCE FOR A POLYPHYLETIC ORIGIN OF THE PICIFORMES STORRS L. OLSON NationalMuseum of Natural History, SmithsonianInstitution, Washington,D.C. 20560 USA ABSTRACT.--Despitetwo recent anatomical studies to the contrary, the order Piciformes appears to be polyphyletic. The structure of the zygodactyl foot in the Galbulae is very distinct from that in the Pici, and no unique sharedderived charactersof the tarsometatarsus have been demonstratedfor thesetwo taxa. The supposedlythree-headed origin of M. flexor hallucis longus sharedby the Galbulae and Pici is doubtfully homologousbetween the two groups, leaving only the Type VI deep flexor tendonsas defining the order Piciformes.This condition is probably a convergentsimilarity. Evidence is presentedsupporting a close relationship between the Galbulae and the suborder Coracii and between the Pici and the Passeriformes.There are fewer characterconflicts with this hypothesisthan with the hy- pothesis that the Piciformes are monophyletic. Problems concerning fossil taxa are also addressed.Received 24 September1981, accepted15 May 1982. A MONOPHYLETICorigin of the Piciformes with outsidegroups in a manner indicating that appears to have gained support from the si- the zygodactyl condition in cuckoos,parrots, multaneous appearance of two cladistic, ana- and Piciformes had arisen independently, tomicalpapers (Swierczewski and Raikow 1981, through convergence. Simpson and Cracraft 1981) that concur in the Although I certainly do not advocate a traditional conceptof the order--a conceptthat monophyleticorigin of zygodactylbirds, the has prevailed at least since the time of Gadow argumentsthat Simpson and Cracraft(1981) and (1893). I depart from this view in considering Swierczewskiand Raikow (1981)present against each of the two major subdivisionsof the Pic- such a hypothesis do not meet the require- iformes, the Galbulae (Bucconidae, Galbuli- ments of their cladisticmethodology. Simpson dae) and the Pici (Capitonidae, Ramphastidae, and Cracraft(1981: 484) concludeonly that "the Indicatoridae, Picidae), to be more closely re- relationships of cuckoos and parrots remain lated to another group than to each other. My among the most enigmaticwithin ornithology purpose here is (1) to show that the evidence .... "although "there is a generalacceptance for monophyly is weak, uncorroborated,and among avian systematists. that piciforms has in part been misrepresentedby Simpson are most closely related to coraciiformsor to and Cracraft (1981), and (2) to make prelimi- passeriformsand that cuckoosand parrots are nary suggestionsas to the probableclosest rel- not." They concedethat "this hypothesishas atives of the Galbulae and the Pici. yet to be tested cladistically "Swier- czewski and Raikow (1981: 469) state that: "The THE WEAKNESS OF THE EVIDENCE FOR muscular component of the foot mechanism is PICIFORM MONOPHYLY quite different in [the Cuculidae,Psittacidae, and Piciformes] ... which supports the con- Zygodactyly.--Obligate zygodactyly, the tention (Bock and Miller 1959: 30) that those condition in which the fourth toe is perma- groups became zygodactyl independently." nently reversedand has an enlargedaccessory Both setsof authors have thus tacitly accepted articulatingprocess (the "sehnenhalter"),oc- differences between taxa as evidence of non- curs in cuckoos (Cuculidae, Cuculiformes), relationship,a procedureof which Cracrafthas parrots (Psittacidae,Psittaciformes), and in the been outspokenlycritical (see Olson 1982). Piciformes. This is obviously a derived con- The most complete and original work on the dition in birds that could be used to define nature of the zygodactyl foot is that of Stein- these taxa as a monophyletic group in a cla- bacher (1935), whose resultshave seldombeen distic sense, unless it were shown that each of accurately represented (a notable exception thesezygodactyl taxa sharesderived characters being Sibley and Ahlquist 1972). Steinbacher 126 The Auk 100: 126-133. January 1983 January1983] PolyphyleticOrigin of thePiciformes 127 A B C D Fig. 1. Posterior(top row) and lateral (bottom row) views of the distal end of the tarsometatarsusin the four groups of birds with obligate zygodactyly(illustrations from Steinbacher1935). A, cuckoo,Centropus ateralbus(Cuculidae, Cuculiformes);B, jacamar, Galbularuficauda (Galbulidae, Galbulae, "Piciformes"); C, toucan, Ramphastostoco (Ramphastidae,Pici, Piciformes);D, parrot, Amazonaochrocephala (Psittacidae, Psittaciformes).Abbreviations: Sh = sehnenhalter,C II = trochleafor digit II, C III= trochleafor digit III, C IV = trochleafor digit IV, G IV = articulatingsurface for digit IV, Gsh = articulatingsurface of sehnen- halter, M 16 = depressionfor origin of M. extensorbrevis digiti IV, R = groovefor tendon of M. extensor brevisdigiti IV, M 4 = depressionfor tendonof M. flexorperforatus digiti IV. It canbe seenthat the form of zygodactylyin the Galbulae is completelydifferent from that in the Pici, and the two can in no way be regardedas homologous. The bestinterpretation of the evidencefrom the tarsometatarsusis that zygodactyly evolved independently in all four of these groups. showed that there were four distinct types of viding evidence "that osteological... char- morphologyof the tarsometatarsusin birds with acteristicsof zygodactyly are distinct for the obligatezygodactyly, with that in the Galbulae piciformsand different from cuckoosand par- being as different from that in the Pici as either rots," Simpsonand Cracraft(1981: 484) have of thesetwo is from parrotsor cuckoos(Fig. 1). dearly misrepresentedthe factsand Steinbach- In eachof thesefour groupsthere is a sehnen- er's interpretation of them. halter. Steinbacher (1935: 234) even identified Nowhere do Simpson and Cracraft, nor any a sehnenhalterin owls (Strigiformes),which other authors(e.g. Bockand Miller 1959),show are facultatively zygodactyl. Thus, the state- that there are derived characters of the tarso- ment by Simpsonand Cracraft(1981: 485) that metatarsusthat will distinguish the Galbulae "zygodactyly and the presenceof a sehnen- and Pici from parrotsand cuckoosand that will halter can be interpreted as derived characters establish the Piciformes as a monophyletic definingthe piciforms as monophyletic" is dis- group. In fact, the apparentlyless modified ingenuous.In citing Steinbacher(1935) as pro- trochlea IV and sehnenhalter in the Galbulae 128 $To•s L. O•,soN [Auk, Vol. 100 are actually more similar to the condition in expanded so that fibers also originated from cuckoos,whereas the larger,more discrete,and the proximal end of the fibula and from the distally projecting sehnenhalterin the Pici is median raphe of the adjacentM. flexor perfor- more like that in parrots(Fig. 1). There simply atus digiti III. [Incidentally, Fig. 6d in Simpson is no evidence in the structure of the tarso- and Cracraft (1981) is mislabelled--"Fpp3" metatarsus that will demonstrate a close rela- shouldread "Fp3."] In Trachyphonusthe fibers tionship between the Galbulae and the Pici. originatingon the fibula, and thoseoriginating Swierczewski and Raikow (1981) present six on the raphe of M. flexorperforatus digiti III, myological charactersto define their Clade B are slightly separatedfrom the main belly of (= Galbulae) and six additional myological the musclethat originatesin the popliteal fos- charactersto define their CladeG (= Pici). Thus, sa, thus making the muscle three-headed. these two taxa differ from each other in at least Clearly, M. flexor hallucis longus, in contrib- 12 myological charactersof the hind limb, as uting to the flexion of two toes in addition to well as having a completelydifferent structure the hallux, has become strengthenedby ex- of the tarsometatarsus associated with their re- panding the area of its origin to the two nearest spective forms of zygodactyly. May we not, available structures.This is directly correlated then, apply the same statement that Swier- with the Type VI arrangementof the flexorten- czewski and Raikow used against monophyly donsand is part of the samecharacter complex. of all zygodactylbirds to argue againstmono- I could not detect any separateheads of or- phyly of the Piciformes,namely that "the mus- igin in the bucconid Hypnelus,however. In- cular componentof the foot systemis quite dif- deed, Swierczewski(1977: 57) statesthat in the ferent in thosegroups," a differencesupporting Galbulidae and Bucconidae the heads are the contentionthat they "becamezygodactyl "somewhatdifficult to separatefrom each oth- independently" (Swierczewski and Raikow er." He alsonotes that the "commonbelly ex- 1981: 469)? tends only about two-thirds the length of the Origin of M. flexor hallucislongus.--Swier- tibiotarsus" in the Galbulae, versus almost the czewski and Raikow state that M. flexor hal- entire lengthof the tibiotarsusin the Pici. Thus, lucis longus has three heads in the Piciformes, it seems far from certain that the nature of the which they interpretas a derivedcondition that origin of this muscle is homologousbetween supports monophyly. Simpson and Cracraft the Galbulae and the Pici or even that it can (1981:483) imply this conditionto be unique really be said to have three heads in the Gal- to the Piciformesby stating that "the muscle bulae. arises by one or two heads in other birds," Type VI flexor tendons.•adow (1893) de- whereas Swierczewski and Raikow discuss the fined the Piciformesby their possessionof the fact that M. flexorhallucis longus also has three Type VI configurationof the deep flexor ten- heads in most passerines.Because