THE SCLERAL OSSICLES OF OPISTHOCOMUS AND THEIR PHYLOGENETIC SIGNIFICANCE

KEVIN DE QUEIROZ AND DAVID A. GOOD Museumof VertebrateZoology and Department of Zoology,University of California, Berkeley,California 94720 USA

ABSTRACT.--Driedscleral rings of Opisthocomus,, Cuculiformes, and other were examined to determine the pattern of ossicleoverlap in Opisthocomusand its bearing on the phylogeneticrelationships of this taxon. Although Opisthocomusshares a derived numberof 12scleral ossicles with cuculidCuculiformes, the patternof ossicleoverlap differs. Nevertheless,fewer modificationsare required to derive the number and pattern of ossicles in Opisthocomusfrom the conditionsin cuculidor musophagidCuculiformes than from the conditionsin any galliform.Our findingsalso indicate that the scleralrings of birds,unlike thoseof lizards, often do not conserveoverlap relationsbetween adjacentossicles during phylogeny.Received 2 March 1987,accepted 23 July1987.

THE phylogenetic relationships of the Hoa- lap in the scleralrings of larger samplesof Opis- tzin (Opisthocomushoazin) have been the subject thocomus,Cuculiformes, Galliformes, and other of a long-standingcontroversy in systematicor- relevant avian taxa. Our findings have impli- nithology. An enigmatic inhabitant of the cations both for the relationships of Opis- flooded forestsof Amazonia, Opisthocomushas thocomus and for the manner in which scleral been regardedas most closelyrelated to many rings evolve. avian taxa, but most commonly to either Cu- culiformes or Galliformes (Olson 1985; re- MATERIALS AND METHODS viewed by Sibley and Ahlquist 1972,1973). One characterthat hasbeen usedto supporta close Twenty-four scleralrings from 14 Opisthocomusin- relationshipbetween the Opisthocomusand Cu- dividuals were comparedwith rings from 42 galli- form and 36 cuculiformspecies representing all of cullformesis the morphology of the sclerotic the recognizedfamilies, subfamilies,and tribes (Mo- ring, a ring of small bones that developswithin tony et al. 1975) in those groups (Table 1). To un- the sclerain the cornealhemisphere of the eye derstand variation in scleral ossicleoverlap patterns in birds and many other vertebrates(Edinger for a given number of ossicles,we also examined 1929). Lemmrich (1931) found that the scleral representativesof other avian groupswith the same ring of Opisthocomusconsists of 12 ossiclesper number of ossiclesas Opisthocomus(Spheniscidae, Su- eye, which is lower than the 13-16 ossicleshe lidae, Psittacidae,and Todidae). found in 10 speciesof cracidand phasianidGal- We examinedonly the dried, detachedscleral rings liformes (all of which showed a modal number from museumskeletal specimens; no rings were ex- of 14) but is identicalto that found in a single amined in situ.In many casesscleral ossiclepatterns couldbe distinguishedin theserings without further cuculiform, Cuculus canorus. preparation, but in others adherent connectivetissue In addition to ossiclenumber, the pattern in was removed.This was accomplishedby immersing which the scleralossicles overlap varies among the rings in water or ethanol to soften the adherent avian taxa (Lemmrich 1931, Curtis and Miller tissue,which wasthen removedwith forceps.Because 1938).Because the samenumber of ossiclesmay we used only detachedscleral rings, the side of the havebeen derived independently, simple counts headfrom which eachring camehad to be inferred. can be misleading. The pattern of ossicleover- We used the patternsin the sameor closelyrelated lap providesevidence about ossicle homologies taxaobserved by Lemmrich(1931) in conjunctionwith and hence about cases in which the same num- the fact that the widestossicles occur at the tempo- ber of ossicleshas been derived through the rodorsaledge of the ring (best seen in Fig. 1D) to loss of different individual ossicles. determinewhether a given ring was from the right or left side. Although the pattern of scleral ossicleover- We followed Lemmrich's (1931) conventions for lap is known for many birds, it has been un- numberingossicles and describingpatterns of ossicle known for Opisthocomus.We compared the overlap.Lemmrich recognized two basicpatterns of numbers of ossiclesand their patterns of over- ossicleoverlap, which he designatedTypes A and B. 29 The Auk 105:29-35. January1988 30 DEQUEIROZ AND GOOD [Auk,Vol. 105

A 7+

9+

D E F

Fig. 1. Modal patternsof scleralrings in Opisthocomus,Galliformes, and Cuculiformes.(A) Opisthocomus hoazin(), (B) Alecturalathami (Megapodiidae), (C) Aburriaaburri (), (D) Numidameleagris (),(E) Crotophagaani (Cuculidae), (F) Tauracohartlaubi (Musophagidae). All ringsare from the right side, seen in corneal view. Scale equals 5 mm.

In Type A two ossiclesin eachring overlap, and two mined by outgroupcomparison (e.g. Maddisonet al. others are overlapped by, the immediately adjacent 1984). ossicles;all othershave one edge aboveand one be- neath the adjacentossicles (Fig. 1A, C-F). Lemmrich RESULTS pointed out that the most dorsal ossiclein Type A rings alwaysoverlaps both its neighbors,and he re- The distribution of scleral ossicle number and ferred to suchoverlapping ossicles as "+" elements. patternin Galliformes,Cuculiformes, and Opis- Approximatelyopposite this dorsal "+" element is thocomusis listed in Table 1. Opisthocomus(Fig. another "+" element that he designatedossicle num- 1A) showed a modal number of 12, with a range ber 1. He counted ossiclesaround the ring postero- from 11 to 13. Galliformes(Fig. 1B-D) showed dorsallyfrom this ventral "+" element, clockwiseon a modal number of 14, with only about 15% of the right ring and counter-clockwiseon the left. He then recordedthe positionsof the "+" elementsand the specimenshaving 13 or 15,often on a single thoseoverlapped by both of their neighbors,the "-" side. No specimenswere observedto have 16 elements,by listingthe positionsof the "+" elements ossicles,a number seen rarely in Gallusgallus separatedby commas,then a semicolon,followed by (Lemmrich 1931). Lemmrich examined only a the positionsof the "-" elementsseparated by com- single cuculiform, Cuculuscanorus, with 12 os- mas. The ossiclesoverlapped by one neighbor and sicles;all 25 cuculid specieswe examinedagreed overlappingthe other ("imbricating" ossicles)are ex- with this finding (Fig. 1E). All musophagids cluded in this notation. In this way, the ring in Fig. examined (Fig. 1F) had 13 ossicles. 1A would be recordedas 1,7;4,9.In Lemmrich's Type Opisthocomus,Cuculiformes, and cracid and B pattern, only a single ventral "+" element and a phasianidGalliformes all characteristicallypos- single dorsal "-" element are present. Hence, the pattern in Fig. lB is designated1;7. sessLemmrich's (1931) Type A patternof ossicle Hypothesesof phylogeneticrelationship were based overlap.The modal ossicledistribution pattern on the tenet that only sharedderived characterscon- in Opisthocomus(Fig. 1A) was 1,7;4,9,although tain useful information aboutrecency of commonan- five other patternswere observed(Table 1). cestry(Hennig 1966).Character polarity was deter- Within Galliformes, three major patterns were January1988] ScleralOssicles of Opisthocomus 31

TABLE1. Numbersand patternsof scleralossicles in representativespecies of Galliformes,Cuculiformes, and Opisthocomus.Pattern designationsare describedin the Methods. Numbers in parenthesesindicate subsamplesize. Modal configurationsfor variable speciesare in boldface.

Specimens/rings Taxon examined No. of ossicles Pattern

Megapodiidae Aepypodiusarfakianus 1/2 14 1;7 Alectura lathami 1/2 14 1;7 Macrocephalonrnaleo 1/2 14 1;7 Megapodiusfreycinet 6/10 14 1;6 (1) 14 1;7 (9) Cracidae Aburria aburri 2/4 14 1,9;7,10 A. pipile 4/6 14 1,9;7,10 Charnaepetesgoudotii 2/3 14 1,9;7,10 (2) 15 1,9;7,11 (1) Crax daubentoni 1/2 15 1,9;7,10 C. mitu 1/2 14 1,9;7,10 C. pauxi 2/3 14 1,9;7,10 15 1,9;7,10 15 1,10;8,11 C. rubra 2/4 13 1,8;6,9 (2) 14 1,9;7,10 (1) 15 1,9;7,10 (1) Nothocrax urumutum 1/2 14 1,9;7,10 Ortalis canicollis 3/3 14 1,9;7,10 O. cinereiceps 1/2 14 1,9;7,10 O. garrula 1 / 1 14 1,9;7,10 O. motmot 2/4 14 1,9;7,10 O. vetula 3/6 14 1,9;7,10 Penelopealbipennis 1/2 14 1,9;7,10 P. jacquacu 6/10 14 1,9;7,10 P. purpurascens 2/4 14 1,9;7,10 (3) 15 1,9;7,11 (1) Phasi.anidae Meleagridinae Agriocharis ocellata 1/2 14 1,9;6,10 Tetraoninae Bonasa urnbellus 1/2 14 1,9;6,10 Dendragapusobscurus 3/4 14 1,9;6,10 Lagopuslagopus 1/1 14 1,9;7,10 Tympanuchuscupido 1/1 15 1,10;8,11 T. phasianellus 1/2 14 1,9;7,10 15 1,10;8,11 Odontophorinae cristatus 1/1 14 1,9;6,10 C. virginianus 1/2 14 1,9;6,10 Dactylortyxthoracicus 2/3 14 1,9;6,10 Callipeplacalifornicus 16/29 14 1,8;6,10 (2) 14 1,9;6,10 (23) 14 1,9;6,11 (1) 15 1,9;6,10 (1) 15 1,9;7,11(2) Philortyxfasciatus 1/2 14 1,9;6,10 Perdicini Coturnix coturnix 1/1 14 1,9;6,10 Perdix perdix 1/1 14 1,9;6,10 Phasianini Gallusgallus 1/1 14 1,9;6,10 32 DE QUEIROZ AND GOOD [Auk, Vol. 105

TABLE 1. Continued.

Specimens/rings Taxon examined No. of ossicles Pattern G. sonneratii 1/ 1 14 1,9;6,10 Lophuraswinhoei 1/2 14 1,9;6,10 Phasianuscolchicus 1/ 1 14 1,9;6,10 Numidinae Acrylliumvulturinum 4/6 13 1,9;6,10(1) 14 1,9;6,10 (4) 15 1,9;6,11 (1) Gutteraplumifera 3/4 14 1,9;6,10(3) 14 1;10(1) G. pucherani 1/2 14 1,9;6,10 15 1,9;6,10 Numidameleagris 5/9 14 1,9;6,10(8) 14 1,9;7,10 (1) N. mirrata 2/2 14 1,9;6,10 Opisthocomidae Opisthocomushoazin 14/24 11 1,6;4,7(1) 11 1,6;4,8 (3) 11 1,7;4,8 (1) 12 1,7;4,9 (11) 12 1,7;5,9 (1) 13 1,7;4,9 (5) 13 1,8;4,10(2) Musophagidae Corythaeolacristata 2/3 13 1,7;4,9 Corythaixoidesleucogaster 1/2 13 1,6;4,9 13 1,7;4,8 Criniferpiscator 1/2 13 1,7;4,9 14 1,7;4,10 C. zonurus 1/1 13 1,7;4,9 Musophagarossae 2/4 13 1,7;4,9 Tauracocorythaix 4/7 13 1,7;4,9 T. erythrolophus 2/3 12 1,7;4,9(1) 13 1,7;4,9 (2) T. hartlaubi 6/7 13 1,7;4,9(6) 14 1,7;4,9 (1) T. persa 1/2 13 1,7;4,9 T. porphyreolophus 1/ 1 13 1,7;4,9 T. schalowi 2/4 11 1,6;4,8(1) 12 1,7;4,9 (1) 13 1,7;4,9 (2) Cuculidae Cuculinae Cacomantispyrrhophanus 1/ 1 12 1,8;5,10 Chalcitesbasalis 2/2 12 1,8;5,10 Clamatorjacobinus 3/6 12 1,8;5,10 Cuculuspallidus 1/2 12 1,8;5,10 C. solitarius 1/1 12 1,8;5,10 Misocaliusosculans 1/2 12 1,8;5,10 Phaenocophaeinae Coccyzusamericanus 3/4 12 1,8;5,10 C. melacoryphus 1/1 12 1,8;5,10 Piaya cayana 4/6 12 1,8;5,10 Crotophaginae Crotophagaani 2/4 12 1,8;5,10 C. sulcirostris 5/9 12 1,8;5,10 Guiraguira 10/15 12 1,8;5,10(13) 13 1,6;4,10(1) 14 1,8;5,10 (1) January 1988] ScleralOssicles of Opisthocomus 33

TABLE 1. Continued.

Specimens/rings Taxon examined No. of ossicles Pattern

Neomorphinae Dromococcyxpavoninus 1/2 12 1,8;5,10 D. phasianellus 1/2 12 1,8;5,10 Geococcyxcalifornianus 3/6 12 1,8;5,10 Morococcyxerythropygus 3/6 12 1,8;5,10 Neomorphusgeoffroyi 1/2 12 1,8;5,10 Couinae Coua caerulea 2/3 12 1,8;5,10 C. cristata 1 / 1 12 1,8;5,10 Centropodinae Centropusbengalensis 1/ 1 12 1,8;5,10 C. goliath 3/6 11 1,7;4,9(2) 12 1,8;5,10 (4) C. phasianius 1/ 1 12 1,8;5,10 C. senegalensis 1/ 1 12 1,8;5,10 C. sinensis 2/4 12 1,8;5,10 C. superciliosus 3/5 12 1,8;5,10

seen: cracids(Fig. 1C) had primarily a 1,9;7,10 of this monophyletic taxon (e.g. Ostrom 1976, pattern,phasianids (Fig. 1D) had mostly1,9;6,10, Gauthier 1984, Gauthier and Padian 1985). We and megapodiids (Fig. lB), the only examples know of no completescleral rings among non- in the presentstudy of Lemmrich'sType B pat- avianSaurischia, but completerings with a range , had 1;7. Most musophagidCuculiformes from 13 to 15 ossicles occur in Ornithischia (Fig. 1F) had 1,7;4,9and most cuculids(Fig. 1E) (Edinger 1929, Russell 1940, Ostrom 1961, Gal- 1,8;5,10. ton 1974). Heilmann (1926) reported 14 scleral In comparisonswithin families whose mem- ossicles in Archaeornis(=), al- bers vary widely in body (or eye) size (e.g. Co- though Wellnhofer (1974) counted "approxi- turnix vs. Agriochariswithin Phasianidae), ossi- mately 12" in a different specimen.Because Cu- cle configuration did not correlate with size. culiformes, Galliformes, and Opisthocomusare all neognaths, ratites plus constitute DISCUSSION an even closer outgroup for the relationships under investigation in the present study (Pra- As pointed out by Lemmrich (1931), cuculids ger et al. 1976,Sibley and Ahlquist 1981,Stapel are more similar to Opisthocomusthan is any et al. 1984, Cracraft 1986). Struthioand Dromaius group of Galliformes in having 12 ossiclesper have 15 ossicles(Lemmrich 1931). We observed ring; however, other groupsof birds share this 15-16 in the Crypturelluscinnamomeus number. Lemmrich found 12 ossicles in the and 15 in Tinamotispentlandii. Twelve ossicles, spheniscidCatarrhacres (= Eudyptes)chrysocome as seenin Opisthocomusand Cuculidae,is a low (1,8;5,10), the sulid Sula bassanus(1,7;4,10), and number compared with that seen in the most the 11 psittacid species he examined (1,6;4, significant outgroups and suggeststhat such a 9/1,7;4,9/1,8;5,11). There are 12 ossicles in the number is derived relative to the mode of 14 trochilid $elasphorusrufus (1,6;4,8) (Curtis and seen in Galliformes. Miller 1938). In addition, we observed 12 os- As noted above, various avian taxa possess sicles in Sula capensis(1,5;4,10), S. serrafor what appearsto be a derived number of 12 scler- (1,6,8;4,7,10),and S. abbotti(overlap patterns not al ossicles.The positions of the "+" and "-" determined), and in 1 out of 2 eyes in Todus ossicles,however, often differ among taxa with todus(1,8;5,10). the sametotal number of ossicles.For example, Birds probably are derived from within Di- other taxa with a mode of 12 ossicles have the nosauria,and thusthe closestoutgroups to birds following modal patterns:Eudyptes (Lemmrich are to be found among the nonavian members 1931)and Todus(this study) 1,8;5,10;Sula (3 spp.) 34 r•EQUEIROZ AND Goor• [Auk,Vol. 105

1,5;4,10/1,7;4,10/1,8;4,10/1,6,8;4,7,10 (Letore- ids than was the modal pattern, however, nor rich 1931, this study); Psittacidae(5 spp.) 1,6;4,9/ did any of the variants make a galliform rela- 1,7;4,9/1,8;5,11 (Lemmrich 1931, Curtis and tionship more tenable; at least three changes Miller 1938); and Trochilidae (1 sp.) 1,6;4,8 were necessaryto transform any of them to any (Curtisand Miller 1938).These differences might of the galliform patterns. be taken to suggestthat reductionto 12 ossicles The shared possessionof a derived reduced hasbeen achieved independently in eachof the numberof ossiclesper ring in conjunctionwith taxa with "+" and "-" ossiclesin different po- the minimum number of steps necessarybe- sitionsthrough the lossof different imbricating tween modal patterns suggeststhat Opisthoco- ossicles.In lizards, conservationof overlap pat- musis more closelyrelated to Cuculiformesthan tern and lossof imbricating ossiclescan account to Galliformes. As with an Opisthocomus-cucu- for most transformationsbetween scleralrings liform transition, however, only one step is with different overlappatterns (de Queiroz 1982, necessarybetween Opisthocomus and 12-ossicled Underwood 1984). In birds, however, lack of Spheniscidae,Sulidae, and Todidae, 13-ossicled conservationof overlap relationships between Coraciidae and Alcedinidae (Lemmrich 1931), adjacentossicles is supported by the occurrence or 11-ossicled (Lemmrich 1931). It of Type A and Type B rings with the sametotal is therefore just as reasonable on the basis of number of ossicleswithin a single species(e.g. scleral ring morphology to infer that Opistho- Gutteraplumifera) or in closelyrelated taxa(e.g. comusis allied with one of thosegroups as with megapodesvs. other Galliformes), and by the Cuculiformes. Further, some 12-ossicledpsit- occurrencewithin single speciesof rings con- tacidsshare an identical pattern to that of Opis- forming to the same basic type (A or B) but thocomus. having the "+" and "-" ossiclesat different We do not mean to suggestthat the positionsin the ring (e.g. Opisthocomus)(Table is closely related to , boobies,todies, 1). These differences presumably result from rollers, kingfishers,doves, or .Compar- direct shifts in overlap between adjacent ossi- isons with these taxa are provided only to il- cles, with conservation of ossicle number. lustrate that the evidence for the placement of Given the preceding considerations,we de- Opisthocomusbased on scleralring morphology termined the minimum number of - is equivocal.In general,limited conservationof ary events necessary to transform the scleral overlapsbetween adjacentossicles makes it dif- ring of one group into that of another.A cuclid- ficult to determine homologies in the avian Opisthocomustransition requires only a single scleral ring, and this limits the usefulnessof change, in which cuculid ossicle 1 becomes overlap patterns as systematic characters. Opisthocomusossicle 12 by shiftingoverlaps such Nevertheless,given the current lack of evi- that cuculid ossicle1 no longer overlaps,but is dence supporting a close relationship to some instead overlapped by cuculid ossicle2 (which other group of birds (Sibleyand Ahlquist 1972, then becomesOpisthocomus ossicle 1). Similarly, 1973), one might assumethat Opisthocomusis a musophagid-Opisthocomustransition requires allied either with Galliformes or Cuculiformes. only a single step, with an ossicleeither added Under this assumption,scleral ring morphol- or lost between the ..... ossicle 9 and "+" os- ogy suggestsa closer relationship to Cuculi- sicle1. On the other hand, at leastthree changes, formes. involving both lossesand shifts in overlap of ACKNOWLEDGMENTS ossicles,are necessarybetween Opisthocomus and any of the galliform patterns. We thank the following curatorsfor allowing us to Considerable individual variation was seen examine material under their care: G. F. Barrow- in ossicleoverlap in certain taxa, notably Opis- clough (AMNH), J. W. Fitzpatrick (FMNH), N. K. thocomus(Table 1). We examined this variation Johnson (MVZ), J. V. Remsen (LSU), C. G. Sibley (YPM), R. W. Storer (UMMZ), and R. L. Zusi (USNM). to determine if any of the nonmodal variants We thank D. A. Bell, J. G. Groth, N. K. Johnson, B.C. might be more similar (i.e. fewer changes)to Livezey, J. I. Smith, D. W. Steadman,and D. B. Wake any of the cuculiformor galliformpatterns than for critical attention to previousdrafts of this paper. was the modal, suggestinga possibletransfor- In addition, the senior author thanks P. Sereno for mation series.None of the Opisthocomusvariants information about literature on ornithischian scleral was any more similar to cuculidsor musophag- ossicles. January1988] ScleralOssicles ofOpisthocomus 35

LITERATURE CITED King, and K. C. Parkes, Eds.). New York, Aca- demic Press. CRACRAFT,J. 1986. The origin and early diversifi- OSTROM,J.H. 1961. Cranial morphologyof the had- cation of birds. Paleobiology12: 383-399. rosaurian of North America. Bull. Am. CURTIS,E. L., & R. C. MILLER. 1938. The sclerotic ring Mus. Nat. Hist. 122: 33-186. in North American birds. Auk 55: 225-243ß ß 1976. Archaeopteryxand the . DE QUEmOZ,K. 1982. The scleral ossiclesof scelo- Biol. J. Linn. Soc. 8: 91-182. porine iguanids:a reexaminationwith comments PRAGER,E. M., A. C. WILSON, D. T. OSUGA, & R. E. on their phylogeneticsignificance. Herpetolog- FEENEY.1976. Evolution of flightless land birds ica 38: 302-311ß on southern continents:transferrin comparison EDINGER,T. 1929. Uber knocherneScleralringe. Zool. showsmonophyletic origin of ratires.J. Mol. Evol. Jahrb. Abt. Anat. 51: 163-226. 8: 283-294. GALTON, P.M. 1974. The ornithischian RUSSELL,L.S. 1940. The sclerotic ring in the Had- Hypsilophodonfrom the Wealden of the Isle of rosauridae. Contrib. R. Ontario Mus. Paleontol. Wight. Bull. Brit. Mus. (Nat. Hist.) Geol. 25: 1- 3: 1-7, 2 plates. 152. $IBLEY,C. G., & J. E. AHLQUIST. 1972. A comparative GAUTHIER,J.A. 1984. Acladisticanalysisofthehigher study of the white proteins of non- systematiccategories of the Diapsida.Ph.D. dis- birds. Bull. PeabodyMus. Nat. Hist. 39: 1-276. sertation,Berkeley, Univ. California. ß & -- 1973. The relationships of the --, & K. PADIAN.1985. Phylogeneticßfunctionalß Hoatzin. Auk 90: 1-13. and aerodynamicanalyses of the origin of birds , & --. 1981. The phylogeny and rela- and their flight. Pp. 185-197 in The beginnings tionshipsof the ratite birdsas indicatedby DNA- of birds (M. K. Hecht, J. H. Ostrom, G. Viohl, and DNA hybridization.Pp. 301-335 in Evolutionto- P. Wellnhofer, Eds.). Eichstatt, Proc. Int. Ar- day. Proc.2nd Int. Congr. Syst.Evol. Biol. (G. G. chaeopteryxConf. E. Scudder and J. L. Revealß Eds.). Pittsburghß HEILMANN,G. 1926. The origin of birds. LondonßH. Carnegie-Mellon. F. & G. Witherby. STAPEL,$. O., J. A. g. LEUNISSEN,g. VERSTEEG,J. HENNIG,W. 1966. PhylogeneticsystematicsßUrbanaß WATTEL,•: W. W. DEJONG. 1984. Ratites as oldest Univ. Illinois Press. offshoot of avian stem--evidence from alpha- LEMMRICH,W. 1931ß DerSkleralring derV6gel. Jena. crystallin A sequences.Nature 311: 257-259. Z. Naturforsch. 65: 513-586ß UNDERWOOD,G. 1984. Scleral ossicles of lizards: an MADDISON,W. P., M. J. DONOGHUE,& D. R. MADDISON. exercisein characteranalysis. Pp. 483-502 in The 1984. Outgroup analysis and parsimony. Syst. structure,development and evolution of reptiles. Zool. 33: 83-103. Symp. Zool. Soc. London No. 52 (M. W. J. Fer- MORONY,J. J., JR.,W. J. BocK, & J. FARRANDJR. 1975. guson, Ed.). London, Academic Press. Reference list of the birds of the world. New WELLNHOFER,P. 1974. Das ftinfte Skelettexemplar York, Am. Mus. Nat. Hist. von Archaeopteryx.Palaeontographica 147: 169- OLSON,S.L. 1985. The fossil record of birds. Pp. 79- 216. 238 in Avian biologyßvol. 8 (D. S. Farner, J. R.