Relationships and Evolution of Flamingos (Aves: Phoenicopteridae)

Relationships and Evolutionof Flamingos

(Aves: Phoenicopteridae)

STORRS L. OLSONand

ALAN FEDUCCIA

SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY • NUMBER 316

SERIES PUBLICATIONS OF THE SMITHSONIAN INSTITUTION

Emphasis upon publication as a means of "diffusing knowledge" was expressedby the first Secretary of the Smithsonian. In his formal plan for the Institution, JosephHenry outlined a program that included the following statement: "It is proposed topublish a series of reports, giving an account of the new discoveries in science, andof the changes made from year to year in all branches of knowledge." This themeof basic research has been adhered to through the years by thousands of titles issuedin series publications under the Smithsonian imprint, commencing with SmithsonianContributions to Knowledge in 1848 and continuing with the following active series:

Smithsonian Contributions to Anthropo/ogySmithsonian Contributions to Astrophysics

Smithsonian Contributions to BotanySmithsonian Contributions to the Earth Sciences

Smithsonian Contributions to PaleobiologySmithsonian Contributions to ZoologySmithsonian Studies in Air and Space

Smithsonian Studies in History and Technology

In these series, the Institution publishes small papers and full-scale monographsthat report the research and collections of its various museums and bureaux or ofprofessional colleagues in the world cf science and scholarship. The publications aredistributed by mailing lists to libraries, universities, and similar institutions throughoutthe world.

Papers or monographs submitted for series publication are received by theSmithsonian Institution Press, subject to its own review for format and style, onlythrough departments of the various Smithsonian museums or bureaux, where themanuscripts are given substantive review. Press requirements for manuscript and artpreparation are outlined on the inside back cover.

S. Dillon RipleySecretarySmithsonian Institution

S M I T H S O N I A N C O N T R I B U T I O N S T O Z O O L O G Y • N U M B E R 3 1 6



Storrs L. Olsonand Alan Feduccia

SMITHSONIAN INSTITUTION PRESS

City of Washington

1980

ABSTRACT

Olson, Storrs L., and Alan Feduccia. Relationships and Evolution of Fla-mingos (Aves: Phoenicopteridae). Smithsonian Contributions to Zoology, number316, 73 pages, 40 figures, 2 tables, 1980.—Previous evidence supposedlyshowing a relationship between flamingos and either storks (Ciconiiformes) orducks (Anseriformes) is re-examined in light of the recent hypothesis derivingflamingos from shorebirds of the order Charadriiformes. Anatomical charac-ters used to indicate relationship between flamingos and storks are shown toconsist entirely of primitive "non-anseriform" traits found in several otherorders, including Charadriiformes. Most of the presumed anseriform charac-ters of flamingos also occur in the Charadriiformes, as do all of the charactersof flamingos that do not occur in either Ciconiiformes or Anseriformes. Thedistinctive life history and behavior of flamingos is demonstrated as being verysimilar to that of the Recurvirostridae (Charadriiformes), particularly theAustralian Banded Stilt {Cladorhynchus leucocephalus), but is unlike that of storksor ducks. The appendicular myology of Cladorhynchus is described and is foundto be quite similar to that of flamingos, whereas neither is close to storks. Thethigh muscle M. iliotibialis medialis, heretofore considered unique to flamin-gos, was discovered in Cladorhynchus but not in other Recurvirostridae. Evi-dence from osteology, natal down, oology, and internal parasites stronglysupports a charadriiform derivation of flamingos; pterylosis does not contra-dict such a relationship; and knowledge of the early evolution of flamingosand Anseriformes offers a logical explanation for their sharing similar mallo-phagan parasites. The earliest certain flamingo, from the early Middle Eoceneof Wyoming, is described herein as a new monotypic genus and species thatwas intermediate in size and morphology between the Recurvirostridae andmodern flamingos. Other aspects of paleontology of flamingos are discussed.Evolutionary steps in the development of filter feeding in birds are outlined.The structure of the feeding apparatus of flamingos is shown to be entirelydifferent from that of the Anseriformes, but is strikingly convergent towardsthat of baleen whales. Morphological and behavioral precursors for filterfeeding are shown to occur in the Charadriiformes but not in the Ciconi-iformes. Flamingos (Phoenicopteridae) clearly belong in the order Charadri-iformes, suborder Charadrii, immediately following the Recurvirostridae.

OFFICIAL PUBLICATION DATE is handstamped in a limited number of initial copies and is recordedin the Institution's annual report, Smithsonian Year. SERIES COVER DESIGN: The coral Montastreacavemosa (Linnaeus).

Library of Congress Cataloging in Publication DataOlson, Storrs LRelationships and evolution of flamingos (Aves, Phoenicopteridae)(Smithsonian contributions to zoology ; no. 316)Bibliography: p.1. Flamingos—Evolution. 2. Flamingos—Classification. 3. Birds, Fossil. 4. Birds—Evolu-

tion. 5. Birds—Classification. I. Feduccia, J. Alan, joint author. II. Title. III. Series:Smithsonian Institution. Smithsonian contributions to zoology ; no 316QL1.S54 no. 316 [QL696.C56] 591s [598'.34] 80-607058

Contents

Page

Introduction 1Acknowledgments 3

Review of Previous Anatomical Information 3Life History and Behavior 9Myology 15

Appendicular Myology of Cladorhynchus 17Pterylosis 31Natal Down 34Oology 34Parasitology 36Biochemistry 39Osteology 39

Skull 39Postcranial Skeleton 40

Paleontology 43A New Stilt-like Flamingo from the Middle Eocene of Wyoming 47

Order CHARADRIIFORMES Huxley, 1867 48Suborder CHARADRII Huxley, 1867 48

Family PHOENICOPTERIDAE Bonaparte, 1831 48Juncitarsus, new genus 48

Juncitarsus gracillimus, new species 48Aspects of Evolution of Filter Feeding 58Conclusions 66Literature Cited 69

111



Storrs L. Olsonand Alan Feduccia

Introduction

The living flamingos (Phoenicopteridae) con-stitute a small and easily recognized group of fourto six species of large, brightly colored waterbirdswith extremely long legs and neck and a uniquelyshaped bill adapted for filter-feeding. They arecolonial and characteristically inhabit highly sa-line bodies of shallow water. The existing speciesare placed in three genera that are separable intotwo groups on bill morphology: (1) Phoenicopterus,which has a more primitive filtering apparatus,and (2) Phoenicoparrus and Phoeniconaias, which aremore specialized (Jenkin, 1957). The last twogenera are separated from each other only by thepresence or absence of the hind toe, which is veryreduced in flamingos in any case, so that contin-ued recognition of the genus Phoeniconaias is prob-ably unwarrented. We have not detected char-acters of generic value in the postcranial skeletonthat will separate the two groups differing infeeding adaptations. For purposes of extrafamilialcomparisons, modern flamingos can be regardedas essentially monogeneric, although we have

Storrs L. Olson, Department of Vertebrate Zoology, National Museumof Natural History, Smithsonian Institution, Washington,DC. 20560. Alan Feduccia, Department of Zoology, Universityof North Carolina, Chapel Hill, North Carolina 27514.

made no innovations in generic usage in thepresent paper.

The proper phylogenetic position of the Phoen-icopteridae has long been a perplexing problemin avian systematics. Because of the conflictingnature of the taxonomic evidence presented sofar, the relationships of these birds have neverbeen satisfactorily resolved. Flamingos are mostoften placed with storks, herons, and ibises in theorder Ciconiiformes. This assemblage consists oflarge, long-legged waterbirds having a long neck,"desmognathous" palate, and usually altricialyoung. Olson (1979) has recently suggested thatthe order may be an entirely artificial collectionof unrelated families. The association of flamin-gos with storks and their supposed allies wasprobably made originally solely on the basis oftheir large size and general proportions. Thisallocation was later "supported" by a body ofevidence that we shall show to be altogetherfallacious.

Other workers noted that the lamellate bill,webbed feet, mallophagan parasites, and somegeneral aspects of the behavior of flamingosseemed to point toward a relationship with ducksand geese (Anseriformes). Controversy has sub-sequently centered on whether flamingos aremore closely related to the storks or to the ducks.

1

SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

Continual equivocation has resulted in some au-thors elevating the flamingos to an order of theirown (Phoenicopteriformes), an action that merelyavoids the question of relationships.

All comparisons of flamingos have traditionallybeen made with Ciconiiformes on one hand andAnseriformes on the other, but as we shall see,most of the characters cited as showing relation-ship also occur outside those two groups. Manyof the supposed ciconiiform characters of fla-mingos are really only negative characters thatare simply "non-anseriform" in nature and notreally indicative of relationship to storks. Char-acters of flamingos that are shared with bothAnseriformes and Ciconiiformes have even beencited by some authors as evidence of a commonorigin of all three groups, despite the fact thatthese characters are likewise found in birds ofseveral other orders.

The first alternative to the ciconiiform-anseri-form hypothesis of flamingo relationships wasadvanced by Feduccia (1976), who proposed thatflamingos are related to shorebirds (Charadri-iformes) and have no affinity with the Ciconi-iformes. This idea came about through studies ofthe fossil bird Presbyornis (McGrew and Feduccia,1973; Feduccia and McGrew, 1974; Feduccia,1976, 1977a, 1977b, 1978, in press). Bones ofPresbyornis occur in great concentrations in earlyEocene deposits of the western United States andthe birds were evidently highly colonial, as aremodern flamingos. For this reason and becausePresbyornis had originally been described by Wet-more (1926) as a new family of Charadriiformesrelated to the Recurvirostridae, Feduccia (1976)made comparisons with shorebirds. He foundboth Presbyornis and the Phoenicopteridae to bemost similar in their postcranial osteology to theCharadriiformes and very different from storksand herons. We now know, however, that theskull of Presbyornis is duck-like (Feduccia, 1978;Olson and Feduccia, in press) and, apart from arather close similarity in the naso-frontal area, isquite different from that of flamingos. Presbyornisprovides evidence for a charadriiform origin ofthe Anseriformes (Olson and Feduccia, in press),but similarities between Presbyornis and flamingos

may be attributable to both having been derivedfrom the Charadriiformes, rather than indicatingthat they are closely related to each other. Nev-ertheless, Presbyornis coincidentally pointed theway towards the proper association of flamingoswith the Charadriiformes.

In the course of our investigations we foundthat all the evidence relating to the systematics offlamingos could be best explained by their havinga charadriiform origin. Furthermore, it soon be-came apparent not just that flamingos were de-rived from the Charadriiformes but that theywere derived from a particular family, the Re-curvirostridae. In this connection we have maderepeated comparisons with the AustralianBanded Stilt, Cladorhynchus leucocephalus, an ex-traordinary bird that shares a number of uniquetraits with flamingos and which can in manyrespects be regarded as forming an intermediatebetween the Recurvirostridae and the Phoenicop-teridae.

In view of our findings, we take satisfaction innoting that in Willughby's Ornithology (Ray,1678), which is regarded as the foundation of thescientific study of birds, the flamingo appears inBook III, Part II, with "birds of a middle naturebetween swimmers and waders, or that do bothswim and wade" (page 312), in a separate section,the "whole-footed [webbed] long-legged birds"(page 320), that otherwise includes only the avo-cet (Recurvirostra, Charadriiformes). These twowere far removed from the "cloven-footed water-fowl" in which were placed the storks and herons.

In his Systema Naturae, Linnaeus (1758) gener-ally followed the classification of Willughby andRay for birds, "and where he departed from hismodel he seldom improved upon it" (Newton,1896:8 [intro.]). Linnaeus led off his order Grallaewith the flamingos, followed immediately byspoonbills (Threskiornithidae), certain storks (Ci-coniidae), and the species he included in Ardea,which was composed mainly of herons, storks,and cranes. These were succeeded by shorebirdsand various gruiforms. The association of flamin-gos with storks and herons, thus begun, has con-tinued up to the present and has quite successfullyobscured the true relationships of flamingos.

NUMBER 316

Sibley, Corbin, and Haavie (1969:176) pro-vided a comprehensive review of the classificationand relationships of flamingos, citing previousevidence from paleontology, anatomy, parasites,behavior, and new evidence from biochemistry,from which they concluded that the flamingosshould be "treated as a suborder, Phoenicopteri,in the Order Ciconiiformes and that, in a linearlist, the Anseriformes and Ciconiiformes beplaced adjacent to one another." Because theirstudy has come to be regarded as a standardreference on the systematics of flamingos and hasbeen cited as showing proof of ciconiiform affin-ities (e.g., Parkes, 1978), we have analyzed theevidence they presented in critical detail, aug-menting this with new data and evaluating thewhole in light of the charadriiform affinities firstproposed by Feduccia (1976).

ACKNOWLEDGMENTS.—Without the alacritouscooperation of individuals and museums in Aus-tralia, we would not have had access to some ofour most crucial and convincing evidence. Forproviding information and specimens we are es-pecially indebted to A. R. McEvey and LoreneReid, National Museum of Victoria, Melbourne;Shane Parker, South Australian Museum (SAM),Adelaide; and G. M. Storr, Western AustralianMuseum, Perth. James A. McNamara generouslyprovided us with his unpublished manuscript onthe feeding ecology of Cladorhynchus.

Other specimens were made available to usthrough the kindness of John Farrand, Jr., andMalcolm C. McKenna, American Museum ofNatural History (AMNH), New York, andEleanor H. Stickney, Peabody Museum of Nat-ural History, Yale University, New Haven. Weare grateful to Robert J. Emry and Arnold D.Lewis for their efforts in trying to locate animportant fossil flamingo locality. Pierce Brod-korb kindly permitted the examination of certainfossils in his care.

Illustrations are by Sigrid K. James, LloydLogan, Ellen Paige, and Jaquin B. Schulz. Pho-tographs of specimens are the painstaking workof Victor E. Krantz. SEM photographs of egg-shells were prepared by Mary Jacque Mann.Anna L. Datcher and Jean J. Smith patiently

typed several versions of the manuscript.For discussions, information, and help in elu-

cidating a diverse variety of topics, we are in-debted to Mary Heimerdinger Clench, K. C.Emerson, C. Lewis Gazin, James G. Mead, Har-old Mayfield, Kenneth C. Parkes, Franklin L.Pearce, and Richard L. Zusi. We are particularlygrateful to James C. Vanden Berge for his exten-sive comments on the section on myology, whichwas also criticized by Robert J. Raikow. Foroffering overall comments on the manuscript weare indebted to Ernst Mayr, David W. Steadman,and George R. Zug.

Review of Previous Anatomical Information

Sibley et al. (1969) have most recently beencited by Parkes (1978:9) in connection with thesupposed "substantial amount of evidence . . .gathered by earlier workers that indicates rela-tionship of the flamingos to the Ciconiiformes"that "would have to be disposed of satisfactorilybefore serious consideration can be given to Fed-uccia's proposed phylogeny for flamingos thatinvolves no relationship at all to Ciconiiformes."In their discussion of anatomical evidence, Sibleyet al. (1969:159, table 3) present a summary ofmost of the characters cited in the earlier litera-ture on flamingo systematics (reproduced here asTable 1). In accordance with Parkes' admoni-tions, we have analyzed each of these.

"Characters Shared with the Ciconiiformes"

1. "(partly) nidifugous": The intention isnot clear, particularly since "nidifugous" is givenlater as a character (19) shared with Anseriformes.According to the terminology of Nice (1962), allCiconiiformes except flamingos are semi-altricial(unable to leave nest, down covered, fed by par-ents); Anseriformes are type 2 precocials (leavenest first day or two, down covered, follow parentsbut feed themselves); flamingos are semi-preco-cial (stay at nest though able to walk, downcovered, fed by parents). Charadriiformes may beeither precocial (e.g., shorebirds) or semi-preco-cial (e.g., gulls). Downy flamingo chicks leave thenest within three or four days and so seem much


TABLE 1.—Summary of anatomical characters of flamingos reproduced exactly as in Sibley etal. (1969, table 3) except with characters numbered to facilitate reference to the text

Characters shared with

Ciconiiformes

Characters shared with

Anseriformes

Characters shared withboth orders

Characters shared withneither order

Developmental:1 (partly) nidifugous2 two coats of down

Integumental:3 down structure4 pterylosis5 aftershaft present

Skeletal:6 basipterygoid process

present7 palatine and vomer8 rostrum9 pelvis

10 number ribs

Muscular:11 flight muscle

attachment12 gastrocnemius

Others:13 carotid artery

arrangement14 cervical air sacs divided15 intestinal convolutions16 penis rudimentary17 abdominal air sacs large

Developmental:18 thick down on young19 nidifugous

Integumental:20 feather structure21 waterproof plumage22 webbed feet23 lamellate bill

Skeletal:24 nasal aperture25 supraorbital depression26 lachrymals27 quadrate28 mandibular angle29 pectoral girdle

Others:30 caeca31 tongue shape

Integumental:32 tufted oil gland present33 11 primaries34 diastataxic

Skeletal:35 carinate36 desmognathous37 holorhinal38 pervious nares39 no ectocondylar process40 16-25 cervical vertebrae

Muscular:41 ambiens present

Integumental:42 reduced hallux43 inverted bill44 filter apparatus

Muscular:45 flexor tendons type IV46 1 pair syrinx muscles47 small femoral-caudal48 BXY + muscle formula

Other:49 type of air cells in lung

more like the young of shorebirds or ducks thanthe helpless young of Ciconiiformes. Among theCharadrii, the young of Cladorhynchus do not leavethe nest immediately after hatching, as do thoseof all other Recurvirostridae, but the degree ofparental attendance is not yet known.

2. "two coats of down": Sibley et al. (1969)stress the fact that in young flamingos there aretwo successive coats of natal down, a conditionthey represent as characteristic of the Ciconi-iformes but absent in the Anseriformes. The sit-uation is more complicated than they indicate,and is discussed beyond (page 34), where weshow that the condition of the natal down inCladorhynchus is unique in the Charadrii formes inbeing like that of flamingos.

3. "down structure": Sibley et al. (1969)cite Reichenow (1877) as noting that flamingos

have simple, unbranched down like that of Ci-coniiformes and unlike the branched rhachidialdown of Anseriformes. The down of flamingos,however, does not differ in this respect from thenon-rhachidial down of the Charadriiformes orthat of most other precocial birds.

4. "pterylosis": This comes ultimately fromNitzsch's (1867:132) statement that the pterylosisof Phoenicopterus is "perfectly Stork-like." Thiscame about through contrasting the pterylosis offlamingos and storks with the highly divergentpterylosis of herons. There is actually little simi-larity between the pterylosis of flamingos andstorks (see page 31).

5. "aftershaft present": The aftershaft is ab-sent or rudimentary in the Anseriformes and it ispresent in flamingos and Ciconiiformes. It is alsopresent in the Charadriiformes and many other

NUMBER 316

orders of birds (Beddard, 1898).6. "basipterygoid process present [sic]": This

would be interpreted as a lapsus calami if Sibleyet al. (1969:161) had not repeated that the "pres-ence of a basi-pterygoid process" is a sharedcharacter uniting flamingos with Ciconiiformes.This statement may derive from Beddard's (1898:441) erroneous assertion that the skull in Phoeni-copterus has "basipterygoid processes, to which theanterior ends of the pterygoids are attached." Infact, the basipterygoid processes are invariablyabsent in the Ciconiiformes and are absent oronly very rudimentary in flamingos, whereas inthe Anseriformes the so-called basipterygoid pro-cesses are particularly large and well developed.These processes may well not be homologous fromone group of birds to another. The basipterygoidprocesses are quite variable in the Charadri-iformes, being absent in some groups and presentin others. This character is certainly not useful fordetermining the relationships of flamingos.

7. "palatine and vomer": See character 8below.

8. "rostrum": The supposed similarity ofthe palatine, vomer, and rostrum of flamingos tothose of storks seems to have been taken by Sibleyet al. from uncritical comments of Reichenow(1877) and Gadow (1877) and may also stem inpart from Huxley's (1867:461) erroneous and un-substantiated statement that in flamingos "thegeneral structure of the rostrum is quite similarto that found in the Storks and Herons." Quiteto the contrary, the general structure of the ros-trum of adult flamingos is not like that of anyother bird. Reichenow's (1877) paper concernsonly the Ciconiiformes, to which he had alreadyassumed the flamingos belonged, and he sawsimilarities where none existed. Although distinc-tive, the palate in flamingos is actually moresimilar to that of shorebirds (as exemplified bythe Recurvirostridae) than that of storks (pages39-40.

9. "pelvis": This seems to stem from a singlesentence in Gadow (1877:386), stating that thepelvis in flamingos is purely stork-like and hardlysimilar to that of ducks at all, particularly in thefeatures of the pubis. This is partly true, although

further comparison shows the pelvis of flamingosto be even more similar to that in the Recurvi-rostridae than to storks (page 41).

10. "number ribs": We have not located aprior reference to this character but it may derivefrom some of the statements of Gadow (1877),who later stated that "from a taxonomic point ofview Ribs are valueless" (Gadow, in Newton,1896:789). We found that flamingos may haveeither 5 or 6 ribs attached to the sternum. Thesame number may be found in the Charadri-iformes, in some Ciconiiformes and Anseriformes,and in many other orders as well (Fiirbringer,1888).

11. "flight muscle attachment": This comesfrom the observations of Weldon (1883:647).

In Storks it is well known that the pectoralis major is dividedinto two or more layers, easily separable from one another,and that its attachment to the humerus forms a tendinousarch beneath which the brachial muscles pass from thecoracoid to the arm. In Phoenicopterus, Gadow has shown thatthese features are exactly repeated; . . . I need hardly pointout that this condition is absolutely unknown among La-mellirostres [= Anseriformes].

The "pectoralis major" (= M. pectoralis) isknown to be divided into two or three layers inthe Gruidae, Cathartidae, Procellariiformes, Pel-ecaniformes, and Scopidae, in addition to storks(George and Berger, 1966:307-308). Hudson et al.(1969) describe this muscle as being divided inthe Laridae and Alcidae. We dissected specimensof Cladorhynchus Uucocephalus and Himantopus mexi-canus (Recurvirostridae) and found this muscle tobe distinctly divided into two sections, which arepartially coalesced along the dorsolateral margin(page 19). The general configuration is as de-scribed and illustrated by Weldon (1883) forPhoenicopterus and is indeed quite distinct from thecondition we observed in a specimen of duck,Anas poecilorhynchus. The division of this muscle incertain storks (Leptotilos) is actually quite differ-ent from that of flamingos (Vanden Berge, 1970).This appears to be another character that sepa-rates flamingos from the Anseriformes but that isshared with numerous other orders, including theCharadriiformes.

12. "gastrocnemius": Weldon (1883) is again


cited for noting that flamingos supposedly differfrom the Anatidae in having the gastrocnemiusarising by three heads instead of two and inhaving the accessory semitendinosus muscle (=M. flexor cruris lateralis pars accessoria) present.He did not mention that in the last characterflamingos differ just as much from storks, inwhich the accessory semitendinosus is also absent.Furthermore, the gastrocnemius in at least someAnatidae arises by three heads (Raikow, 1970:34). The three-headed gastrocnemius and theaccessory semitendinosus occur in many birds,including the Charadrii (George and Berger,1966). Actually, the gastrocnemius in flamingoshas four heads (Vanden Berge, 1970), as does thatof Cladorhynchus (Figure la).

13. "carotid artery arrangement": Sibley etal. regard Glenny's (1955) findings on the carotidarteries as only questionably indicating a rela-tionship between flamingos and the Ciconi-iformes. Flamingos have type B-2-s carotid arter-ies. Anseriformes, Charadriiformes, all Ciconiidaeand Threskiornithidae, and most Ardeidae havetype A-l carotids. A few of the Ardeidae havetype B-l carotids, Ardeola speciosa has type B-2-s,and Botaurus lentiginosus may have either type B-1or B-2-s. The B-2-s condition was evidently de-rived independently from the A-l condition atleast twice within the Ardeidae. The B-2-s con-dition in flamingos had to have been derivedindependently of that in Botaurus lentiginosus andArdeola speciosa and it could as easily have comefrom the A-l condition in shorebirds or ducks asfrom the A-l condition in Ciconiiformes, a pointlater specifically admitted by Sibley and Ahlquist(1972).

14. "cervical air sacs divided": This charac-ter, along with character 17 ("abdominal air sacslarge"), stems from Weldon's (1883) observationsthat the air sacs in flamingos are in these respectsmore like those of storks than ducks. In a modernaccount of the air sac system of birds (Duncker,1971), no mention is made of any peculiarities offlamingos. The cervical air sacs are paired in mostbirds and are unusually large in the Sphenisc-idae, Anatidae, Falconidae, and Accipitridae(Duncker, 1971:49), so it would appear that this

is but another character in which flamingos differfrom ducks but resemble many other birds. Thesame is true of the large abdominal air sacs, whichare typical of most birds. According to Duncker(1971:62) storks differ from all other birds inhaving the posterior thoracic air sac divided intomedial and posterior sacs. If any conclusion canbe drawn from this it is that the air sacs offlamingos do not seem to agree with those ofeither storks or ducks.

15. "intestinal convolutions": Beddard (1898:441) stated only that in flamingos "the intestinesare not duck-like," which carries no informationabout affinities with other groups. We note fromthe table in Gadow (1889) that the general intes-tinal pattern in flamingos is found in the Cha-radrii as well as in storks, although Gadow'sattempts to use the varying patterns of intestinalconvolutions in birds as taxonomic charactershave been largely discredited (see Sibley andAhlquist, 1972:21-22).

16. "penis rudimentary": This feature is ofimportance only in contrast to ducks, which havea well-developed penis. The penis is rudimentaryor lacking in most birds, including flamingos andshorebirds, although it might be noted that Bed-dard (1898:37) states that a penis exists in theCiconiiformes (he does not say in which forms)and in Burhinus (Charadriiformes), in addition toratites, tinamous, cracids, and ducks.

17. "abdominal air sacs large": See character14 above.

"Characters Shared with Anseriformes"

18. "thick down on young": Thick down isalso present in the young of Charadriiformes andother precocial birds.

19. "nidifugous": See character 1 above.20. "feather structure": See character 21 be-

low.21. "waterproof plumage": To Chandler

(1916) the feather structure of flamingos wasmuch more similar to that of Anseriformes thanCiconiiformes, but he did not specifically com-pare flamingos with Charadriiformes. Sibley etal. (1969:160) felt that the "close, hard, water-proof nature of the plumage as a whole, shared

NUMBER 316

by flamingos and geese, could easily be the resultof convergence." They cite Rutschke (1960) asshowing that "water birds in different orders aremore alike in feather structure than non-aquaticbirds even within the same order." Nonetheless,it is worth observing that the feathers in theRecurvirostridae are notably close, hard, anddense; photographs of Recurvirostra americana en-gaged in feeding show the plumage to be quitewaterproof as well (Tremaine, 1975:75).

22. "webbed feet": Species of birds withwebbed feet occur in a variety of orders otherthan Anseriformes, including the Charadri-iformes. Within the Recurvirostridae, Recurvirostraand Cladorhynchus have webbed feet, while Himan-topus has only vestigial webs.

23. "lamellate bill": As detailed later (pages59-60), the structure of the bill and lamellae inflamingos and ducks is fundamentally differentand is not indicative of relationship.

24. "nasal aperture": This character and thefollowing three appear to have been based mainlyon the authority of Shufeldt (1901:305), whostated that

in its external narial apertures; in the possession of supraor-bital glandular depressions; to a small degree in its lachrymalbones; . . . in its quadrates; in the possession of large recurvedprocesses at the mandibular angles—the skull of the Fla-mingo is more or less anserine in character.

The nasal region in flamingos actually differsconsiderably from that of ducks, particularly inlacking the large internal narial opening in therostrum characteristic of the Anseriformes.

25. "supraorbital depression": These depres-sions accomodate salt glands and are found in asupraorbital position in the Charadriiformes andmost other water birds, including even Hesperornis.In the Pelecaniformes the salt glands are locatedwithin the orbit, as they are in storks and herons,but not in the Threskiornithidae. The conditionof this character in flamingos is as much like thatof shorebirds as that of ducks.

26. "lachrymals": In addition to Shufeldt'scomment above, Beddard (1898:441) also madethe misleading statement that the "lachrymalsare large and rather duck-like." The lachrymalsin ducks are thin, antero-posteriorly elongated

bones that are broadly fused with the skull (ex-cept in Anseranas) and that do not nearly reachthe jugal bar. In flamingos the lachrymals areentirely different, being unfused, thickened,pneumatized bones that reach to the jugal barbut have a much narrower contact with the skull.

27. "quadrate": Despite the remarks of Shu-feldt (1901), the quadrates of flamingos are verydifferent from those of ducks in every aspect,particularly in the nature of the articulation withthe lower jaw.

28. "mandibular angle": This refers to thelong, recurved, bladelike retroarticular processesof the mandible, which are superficially verysimilar in ducks and flamingos. Nevertheless thereare differences in the development of these struc-tures between the two groups and the similaritiesprobably are the result of similar needs for in-creased attachment for M. depressor mandibulae(page 64).

29. "pectoral girdle": Evidently Shufeldt's(1901:313) observations have been used by Sibleyet al. for this character. Sibley and Ahlquist(1972:7) have remarked that Shufeldt "worked ina rather haphazard fashion, simply describingand comparing what he happened to have beforehim," and it is widely recognized that his writingson osteology and paleontology are unreliable. Wefind little similarity between flamingos and ducksin the bones of the shoulder girdle.

30. "caeca": The caeca are well-developed inboth flamingos and Anseriformes but are rudi-mentary in Ciconiiformes. Beddard (1898:336)states that in the Charadrii the caeca are "nearlyalways large." Large caeca are also present inother orders of birds, such as Gruiformes.

31. "tongue shape": Although flamingos andducks have large fleshy tongues used in filter-feeding, the morphology and placement is en-tirely different in the two groups, as is the struc-ture of the bony hyoid apparatus (page 60).

"Characters Shared with Both Orders"

32. "tufted oil gland present": A tufted oilgland occurs in the Charadriiformes and in themajority of other orders of birds.

33. "11 primaries": This character is mis-

8 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY

placed by Sibley et al. (1969) and should be listedas shared with the Ciconiiformes (storks only).Ducks, shorebirds, and Ciconiiformes exceptstorks have 10 functional primaries, whereas inflamingos, Ciconiidae, and Podicipedidae(grebes) there are 11 (W. D. Miller, 1924). Theextra primary in flamingos probably evolved witha lengthening of the wing, as the number ofsecondaries in flamingos is considerably greaterthan in storks (page 34). The taxonomic distri-bution of this character strongly suggests inde-pendent derivation in all three instances.

34. "diastataxic": The diastataxic conditionof the secondaries is found in many orders ofbirds, including all the Charadriiformes.

35. "carinate": An entirely useless characterin this instance, as all birds except the "ratites"are "carinate."

36. "desmognathous": The so-called des-mognathous condition of the palate is found inother orders and must have evolved numeroustimes from the schizognathous condition. Thefossil bird Presbyomis shows that the desmogna-thous palate of ducks must have evolved from theschizognathous palate of shorebirds (Olson andFeduccia, in press) and the same is likely true offlamingos.

37. "holorhinal": As with the preceding, theholorhinal nostril occurs in many orders, includ-ing certain Charadriiformes (e.g., Burhinus, Pluvi-anus). This character varies within the Ciconi-iformes (as presently constituted) in that theThreskiornithidae are schizorhinal, except forspoonbills (Platalea), which are sometimes second-arily holorhinal, thus demonstrating the probableease with which the holorhinal condition canarise.

38. "pervious nares": A condition found inmost birds, including Charadriiformes.

39. "no ectocondylar process": This referspresumably to the ectepicondylar prominence atthe distal end of the humerus, though it is notevident why this should have been mentioned inany discussion of flamingos. The ectepicondylarprominence is developed into a large spur inProcellariiformes and most Charadriiformes butis usually reduced in most other birds, including

flamingos. The spur, however, is absent in theJacanidae and Burhinidae, smaller in the Recur-virostridae and Rostratulidae than in most Char-adrii, and has been secondarily lost or reduced inother Charadriiformes (e.g., Scolopax minor) andtheir derivatives.

40. "16-25 cervical vertebrae": The numberof cervical vertebrae in birds may vary at almostany taxonomic level. According to Fiirbringer(1888, volume 2, table 41), flamingos have 18 or19 cervicals, the Ciconiiformes have 16 or 17,with only the Ardeidae having 18 or 19. TheCharadriiformes usually have 15, or 16 in thecase of the Burhinidae and Jacanidae. In theAnatidae, the cervicals range in number from 16to 25, with the high numbers occurring in thelong-necked swans. Obviously, this is a highlyadaptive character. The greater number of cer-vical vertebrae in flamingos than in shorebirds isprobably correlated with the evolution of a longneck associated with the exceptional feeding ap-paratus.

41. "ambiens present": The ambiens muscleis also present in the Charadriiformes and manyother orders of birds. In Ciconiiformes, however,it is present only in some ibises and some storks(Vanden Berge, 1970).

"Characters Shared with Neither Order"

42. "reduced hallux": The hallux is reducedor absent in modern flamingos. Although thoughtnot to be of taxonomic consequence, the hallux isnevertheless always present and rather well de-veloped in the Ciconiiformes and Anseriformes.In the Charadriiformes, on the other hand, thehallux has been reduced or lost repeatedly. Forexample, in the Recurvirostridae, the hallux ispresent in Recurvirostra and absent in Himantopusand Cladorhynchus. Sibley et al. (1969:160) notethat the short toes of flamingos are different fromeither the Ciconiiformes or Anseriformes. Thetoes are quite short in many Charadrii and no-tably so in the Recurvirostridae.

43. "inverted bill": See character 44.44. "filter apparatus": These characters are

unique to flamingos and provide one of the veryfew major points of separation of the Phoenicop-

NUMBER 316

teridae from other Charadriiformes.45. "flexor tendons type IV": Of the various

patterns of the deep flexor tendons (see Georgeand Berger, 1966:446-449), the Ciconiiformes andmost Charadriiformes have the type I configura-tion, and the Anseriformes have the type II con-figuration. Flamingos differ in having type IVdeep flexor tendons, a condition that occurs inseveral groups in which the hallux is reduced orlost, including the Recurvirostridae (page 30).

46. "1 pair syrinx muscles": A condition alsofound in the Charadriiformes (including Clador-hynchus, pers. obs.) and other orders (Beddard,1898).

47. "small femoral-caudal [sic]": This is thefemorocaudal of Garrod (1874), designated "A"in his thigh muscle formulae, which forms part ofthe caudo-iliofemoralis of modern usage. We donot know why Sibley et al. (1969) list this as"small" because it is actually lacking in flamingos(Vanden Berge, 1970), as the following charactersignifies.

48. "BXY+ muscle formula": This repre-sents only part of the expanded thigh muscleformulae used by some modern authors (seeGeorge and Berger, 1966:233). The formula inmost shorebirds is ABXY+, since the caudo-femoral part of M. caudo-iliofemoralis ("A") isusually present. This muscle is now known to belacking in certain genera of Charadrii (page 27),as well as in flamingos. In storks and in herons itis present in some genera and absent in others(Vanden Berge, 1970), and thus appears to beeasily lost and probably of little taxonomic use.

49. "type of air cells in lung": This evidentlyderives from a misreading of Weldon's (1883:639)statement that in flamingos "the lungs presentnothing remarkable, but the air-cells and theirassociated septa are strikingly characteristic."Further reading reveals that by "air-cells" Wel-don was referring to the air sac system (see char-acter 14 above). Nothing in his paper indicatesthat anything inside the lungs of flamingos isdistinctive.

The foregoing analysis should be sufficient to

reveal the glaring inadequacy of the anatomical"evidence" that has been used previously in at-tempting to determine the relationships of fla-mingos. The characters supposedly demonstrat-ing affinity with the Ciconiiformes we have shownto be of wide occurrence in other orders of birds,including the Charadriiformes, and are usefulonly in showing lack of affinity with the Anseri-formes. Advocates of a ciconiiform relationshipfor flamingos are now faced with the fact thatthere is not a single convincing anatomical char-acter supportive of such a hypothesis.

Anatomical evidence for a relationship be-tween flamingos and Anseriformes is equally de-ficient. Some of the characters, such as webbedfeet and large caeca, are shared with the Cha-radriiformes as well. Others, particularly featherstructure, cannot be evaluated without propercomparison with Charadriiformes. Still other ap-parent similarities to Anseriformes, such as thefeeding apparatus, are not really similarities at alland are actually completely different in the twogroups. Flamingos simply possess none of thederived characters of osteology and myology thatdistinguish the well-defined monophyletic orderAnseriformes, so their ancestry must be soughtelsewhere.

Apart from features that are unique, almostevery one of the characters of flamingos cited bySibley et al. (1969) may be found in the Cha-radriiformes, including all of those that do notoccur in either the Ciconiiformes or the Anseri-formes. Thus, even without the additional infor-mation supplied beyond, the evidence thus farpresented is much better explained by a charad-riiform origin of flamingos.

Life History and Behavior

Various aspects of general life history and be-havior have been summarized and discussed bySibley et al. (1969:163), who concluded thatmany of the distinctive characteristics of flamin-gos are probably related to their extreme gregar-iousness, which, like the feeding apparatus, is partof their adaptation to a "narrow ecological niche"in "barren isolated areas." They interpreted the

10 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGV

behavioral evidence as indicating that "approxi-mately equal numbers of behavioral elements areshared by flamingos with the Anseriformes andwith the Ciconiiformes," and they considered it"unlikely that any valid conclusions about rela-tionship can be drawn from these data" (page168). We were at first inclined to agree, but uponreading Hamilton's (1975) accounts of behaviorin North American Recurvirostridae we were im-pressed with the apparent similarities to flamin-gos. This led us to the literature on the AustralianBanded Stilt, Cladorhynchus leucocephalus, where wediscovered that the life history of this species veryclosely approximates that of flamingos.

Flamingos are highly social birds living in largeflocks in remote, barren areas of shallow salinelagoons or alkaline lakes. They feed mainly onsmall crustaceans and insect larvae, or in the caseof the more specialized species, on micro-organisms such as diatoms. Because of the isola-tion of breeding colonies, even the correct sittingposture of incubating flamingos was not knownuntil 1880 (Allen, 1956); the nest and eggs of theabundant Phoeniconaias minor were not discovereduntil 1954 (Brown, 1959),xand those of Phoenico-parrus jamesi were discovered only in 1957 (John-son, et al., 1958). For a given flock, both thetiming and site of breeding may vary irregularly.Nesting colonies are usually large and verydensely packed; territoriality is virtually absent,as is response to terrestrial predators. A single,chalky white egg, large in proportion to the bird,is usually, but not always, deposited on a raisedcone of mud. Young are covered with white orgrayish unpatterned down and remain in the nestfor several days before departing and gatheringin large groups attended by adults, at which timethey often swim. The parents feed the young forextended periods. This general pattern of lifehistory is not duplicated among any of the Cico-niiformes or Anseriformes.

We will briefly compare the above with whatlittle is known of the overall life history ofCladorhynchus leucocephalus and then compare as-pects of the behavior of flamingos with that ofsome of the better-known Recurvirostridae. Mostof the significant information relating to the life

FIGURE 1.—The 1930 nesting colony of the AustralianBanded Stilt, Cladorhynchus leucocephalus, at Lake Callabonna,South Australia: a, part of the colony showing the extremelyclose spacing of nests; b, portion of the colony with adultbirds sitting on nests; c, adults rising from nests. (Photo-graphs from McGilp and Morgan, 1931.)

NUMBER 316 11

FIGURE 2.—The 1930 nesting colony of the AustralianBanded Stilt, Cladorhynchus leucocephalus, at Lake Callabonna,South Australia: a, closer view of adults at nests; b, adultswith newly hatched young; c, young birds remaining in nestafter adults have temporarily departed. (Photographs fromMcGilp and Morgan, 1931.)

history of Cladorhynchus may be found in the pub-lications of Glauert and Jenkins (1931), McGilpand Morgan (1931), Howe and Ross (1931), Jones(1945; a summary of all previous literature on thespecies), Carnaby (1946), Beruldsen (1972), Jen-kins (1975), and Kolichis (1976). In addition, wewere fortunate to be supplied by J. A. McNamara(1976) with a copy of his unpublished paper onthe feeding ecology of Cladorhynchus.

Cladorhynchus lives in remote areas in the south-ern half of Australia, where it occurs in greatflocks that frequent very saline temporary lakesin which may be found enormous quantities ofsmall crustaceans, upon which the birds feed(Jones, 1945). McNamara (1976), who examinedgut contents of Cladorhynchus, found a variety offood items, but principally brine shrimp (Parar-temia) and brine flies (Ephydridae). In connectionwith the heavy fat deposits mentioned by Sibleyet al. (1969) as being characteristic of flamingos,it should be noted that McNamara's specimensof Cladorhynchus had extremely heavy subcuta-neous fat deposits, whereas specimens of Himan-topus and Recurvirostra taken at the same time andplace did"not.

The breeding habits of Cladorhynchus remainedunknown until 1930 (Glauert and Jenkins, 1931),a mystery not without parallel among flamingos.It was then discovered that Cladorhynchus breedsin great, densely packed colonies on islands andsandbars. McGilp and Morgan (1931) estimatednearly 27,000 pairs in a colony at Lake Calla-bonna (Figures 1, 2). Nests consist of depressionsin the ground and are regularly spaced about 30cm apart (Figure la). Nests in the Callabonnacolony had from 1 to 5 eggs, with most nestshaving 3 eggs; clutches of 2 eggs appeared to becomplete (McGilp and Morgan, 1931; Kolichis,1976). Adult and young birds in a colony arepractically fearless, with no apparent predatorresponse. The timing and place of breeding ofCladorhynchus appear to be extremely irregularand are correlated with variable local conditions.For example, the great colony found on LakeCallabonna in 1930-1931 is the only known nest-ing record for South Australia; all others are fromWestern Australia (J. A. McNamara, in litt., to


Olson, 30 May 1978).There appears to be a nearly complete lack of

aggression or territoriality in Cladorhynchns, as istrue of flamingos. McNamara (in litt. to Olson,30 May 1978) reports that

only once in about 110 hours observation over 8 months didI see anything like display (courtship or otherwise). Thistook place on a bank in the salt fields with dry powderyground; one bird knelt on the ground with wings out andthe other standing while both fenced with bills. Also at thistime one bird chased another for a short distance (10") [25cm] then both stopped and assumed an "aggressive" posturewith head down, bill pointed at [the] other and back feathersraised.

This could correspond to the bill-fighting andbowing reported in Phoenicopterus by Rooth(1965).

Furthermore, McNamara quotes Mr. TomSpence, Director, Perth Zoological Gardens, asnoting that

the display of the Banded Stilt is quite spectacular and veryreminiscent of that of the Lesser Flamingo. They parade ina tight little pack with their long mantle feathers raised (justlike a Flamingo) and synchronise their movements prettyclosely. I have also seen this in the wild.

This behavior may be the exact equivalent of themarching display of flamingos (Kahl, 1975). Rais-ing the "back" feathers is a characteristic featureof flamingo displays, and in the breeding plu-mage of Cladorhynchus the tertials are greatly elon-gated, much more so than in other Recurvirostri-dae.

The eggs of Cladorhynchus are exceptionalamong the Charadriiformes for their large sizeand white ground color (Howe and Ross, 1931),both of which recall the eggs of flamingos. Al-though most eggs of Cladorhynchus are markedwith dark lines and blotches, as typical of otherCharadriiformes, the markings are quite variable,and McGilp and Morgan (1931:45) reported thata "fair proportion" were nearly immaculate andat least one was found that was pure white (Figure12).

Most remarkably, the young of Cladorhynchusare clothed in white, unpatterned down (Figure 3),a combination not met with anywhere in theremainder of the Charadrii (Jehl, 1968), but

FIGURE 3.—Chick of the Australian Banded Stilt, Cladorhyn-chus leucocephalus, showing the dense, unpatterned, whitedown, unlike that of any other Charadrii, but similar toflamingos.

which is typical of flamingos. A most importantpoint of similarity with flamingos is that Clad-orhynchus has two coats of nestling down, a char-acter not found in any other known shorebird(page 34).

It is not known when the young leave the nestor whether the adults feed the young, but McGilpand Morgan (1931:43) report that disturbedyoung "kept scampering back towards the nests"and their photographs show many young birdsremaining in nests after the temporary departureof adults (Figure 2c). This is in contrast to NorthAmerican Recurvirostridae as observed by Ham-ilton (1975:89): "Even before chicks are dry, theybegin to toddle away from the nest, and soonafter the brood has hatched, the young can nolonger be found in the vicinity of the nest."

The accounts of Cladorhynchus emphasize therapidity of breeding and the simultaneous hatch-

NUMBER 316 13

ing of young in a colony. The young aggregateand are attended by adults, which is probablynot different from the so-called "creche" behaviorof flamingos. McGilp and Morgan (1931) men-tion "three old birds leading perhaps 60 youngones into the water" and they include a photo-graph of a flock of chicks swimming off into thelake in the manner of young flamingos. At LakeMarmion, Western Australia, Kolichis (1976:116)"saw Banded Stilts on the water in small groupsvarying from three young with two adults toabout 40 young with six to nine adults."

Cladorhynchus provides a nearly perfect inter-mediate in its life history and development be-tween flamingos and the more typical Recurvi-rostridae and thus supplies evidence not only thatflamingos are derived from the Charadrii, butmore particularly from the Recurvirostridae.There could hardly be a more rewarding en-deavor for a field ornithologist than a detailedcomparative ethological study of Cladorhynchusand flamingos.

In analyzing the ecological requirements of theRecurvirostridae, Hamilton (1975:14) found thatalthough Himantopus generally resorts to fresh wa-ter, "avocets [Recurvirostra] prefer alkaline or salinehabitats." They feed mainly upon brine shrimp,Artemia, and adults and larvae of brine-flies, Ephy-dra (Hamilton, 1975; Wetmore, 1925:14). Wehave already noted the similar food habits ofCladorhynchus. On Bonaire the "staple diet" offlamingos is the larvae and chrysalids of Ephydra(Rooth, 1965:51).

Sibley et al. (1969:164) maintained that adultflamingos do not voluntarily swim, and inter-preted this as being contrary to an anseriformderivation, thereby implying similarity to Cico-niiformes. Rooth (1965) records Phoenicopterus ruberon Bonaire as swimming while feeding. Brown(1975:42), in countering Sibley et al., stated thatPhoeniconaias minor in Africa feeds from near thesurface of the water away from the shoreline and"habitually swim; in fact they could not havereached their present numbers if they did notswim" (see Figure 4).

While swimming, flamingos may feed by tip-ping up in manner reminiscent of swans and

dabbling ducks, which has been interpreted as anindication of relationship. Rooth (1965:43) de-scribes this behavior in Phoenicopterus ruber as fol-lows:

The head, neck and the rostral part of the body are totallyimmersed, more or less straight down. The rear part of thebody then sticks up out of the water, so that the legs have tocarry out all possible movements to preserve the balance ofthis position... .This unsteady position necessitates that themovements of the feet can take place in the surface wateronly. It can be seen that the tibiotarsus is directed at anangle upwards and that only one half or three quarters ofthe tarso-metatarsus is in the water.

Avocets and Cladorhynchus likewise do a consid-erable amount of feeding while swimming andalso tip up in order to obtain food (Ross, 1924;Hamilton, 1975; McNamara, 1976; Dinsmore,1977). Compare the following description for Re-curvirostra americana as given by Hamilton (1975:52) with that of Phoenicopterus ruber above.

The feeding bird immerses its head and breast into the waterby tipping on the transverse axis from a swimming or breast-wading position and brings the bill to the bottom. Theupended position of the bird is maintained by a backwardkick of the feet; the tibiotarsi often break the surface....

Flamingos vocalize in flight and the calls ofsome are reminiscent of gooselike honkings, whichhas also been cited as favoring anseriform affini-ties. The notes of Phoeniconaias minor are higherpitched than in Phoenicopterus ruber and have beenlikened to a "yelping" (Cramp and Simmons,1977:367). Proper comparisons are wanting, butit should be noted that the Recurvirostridae arenotorious for vocalizing in flight and "yelping" isa term commonly employed to describe the soundof their nearly incessant calls when disturbed onthe breeding grounds.

The use of neck postures in the threat displaysof flamingos was likened to that in the Anatidaeby Sibley et al. (1969), whereas the raising of thescapular and the back feathers in threats was saidto be more characteristic of the Ciconiiformes.Both of these behavioral components are foundin the Recurvirostridae (Hamilton, 1975, fig. 15).Bill-dipping or displacement drinking, which isperformed by flamingos and ducks but not Ci-coniiformes (Sibley et al., 1969), is also part of the


FIGURE 4.—Lesser Flamingos, Phoeniconaias minor, at Lake Hannington, Kenya, showingnumbers of birds swimming in the middle of the lake. (Photograph by Alan Feduccia.)

normal precopulatory display in the Recurviros-tridae (Hamilton, 1975:73). The near absence ofspecialized pairing displays and the seasonal pairbonds of flamingos are likewise characteristic ofthe Recurvirostridae. In both flamingos and re-curvirostrids, in such courtship as exists, there isa similar heavy reliance on ritualized preening,bill-dipping, head shaking and splashing of water.The "bowing" display in Phoenicopterus (Rooth,1965:88, fig. 14) might be equivalent to theposture assumed in the "Group Circle" display ofRecurvirostra americana (Hamilton, 1975:68, fig.16a). In flamingos, the "attraction function ofcourtship display to other members of the species"was cited by Sibley et al. (1969:167) as resemblingthe Ciconiiformes. However, Hamilton (1975:68)observed that the ritualized special postures usedin highly social group interactions in the Recur-

virostridae "seem to attract conspecifics from nearand far."

Copulation in flamingos takes place away fromthe nest and usually in water, unlike Ciconi-iformes but similar to most Anseriformes (Sibleyet al. 1969). The same is true, however, in theRecurvirostridae. The reader is invited to notethe similarity in the copulatory posture of Phoen-icopterus (Rooth, 1965; fig. 16) and that of Recur-virostra americana (Hamilton, 1975, fig. 17*).

Although under certain circumstances flamin-gos will lay their eggs in a depression in theground, as do most shorebirds, they usually con-struct a characteristic truncated cone of mud,upon which to deposit the egg. According toRooth (1965:99-100), "the function of these highconstructions is clear at high water although theliterature often reports the flooding of colonies,

NUMBER 316 15

despite the high nests." He then adds, "It istypical that the wader Himantopus himantopus [Re-curvirostridae] which breeds along the shores ofthe salinas [on Bonaire] also makes such highnests!" The exclamation point is his; we can onlyregret in light of the present discussion that thisstatement was not elaborated upon. Nests of avo-cets and stilts usually consist of a depression inthe ground lined with any available material,including mud chips (Hamilton, 1975:82). In anumber of references cited by Hamilton (1975:83), avocets and stilts are reported as addingmaterial to their nests during floods to raise theeggs up on a platform. All 15 nests in a colony ofRecurvirostra avosetta were adjusted in this mannerduring a night of heavy rains (Makkink, 1936:35). Certainly it is not difficult to envision thatthis behavior could be modified to produce thatof flamingos, whereas the nest-building habits ofCiconiiformes, which generally construct arborealplatforms of sticks, does not seem at all appropri-ate as a precursor of the phoenicopterid method.

Although none of the other Recurvirostridaeare as highly colonial as Cladorhynchus, all aregregarious and their nesting sites are composedof loose aggregations to which the term "colony"is consistently applied in the literature. As withflamingos there is no site fidelity and coloniestend to shift location from one year to the next(Hamilton, 1975:79).

In contrast to the Ciconiiformes, with theirelaborate nest-relief ceremonies during incuba-tion, flamingos have almost none—one bird walksup and its partner walks away, sometimes indulg-ing in displacement feeding (Rooth, 1965:107).Much the same behavior is recorded for the Re-curvirostridae, although Hamilton (1975:85) in-terpreted the displacement "feeding" as displace-ment nest building, in which objects are pickedup and tossed about.

Apparently the only unique etho-ecological ad-aptations of flamingos not found in recurviros-trids are the parental feeding of the young (pos-sibly present in Cladorhynchus) and the single eggclutch. Both of these must be directly correlatedwith the highly specialized feeding apparatus ofadult flamingos. Because this structure does not

develop until late in ontogeny, some parentalfeeding of young would be a necessity. The singleegg clutch probably evolved as a result of thisincreased parental responsibility.

The foregoing is only a brief sketch, but shouldserve to illustrate the similarities in the life historyand behavior of flamingos and the Recurvirostri-dae. Not only can all the supposed ciconiiformand anseriform behavioral traits of flamingos beaccounted for by having flamingos derived fromthe Charadrii, but also the seemingly uniquecharacteristics of the Phoenicopteridae that areirreconcilable with either storks or ducks can beexplained as well.

Myology

Vanden Berge (1970) conducted a broad, com-parative study of the appendicular myology ofthe Ciconiiformes, including the Phoenicopteri-dae. We regard it as significant that he found "noconsistent pattern in the appendicular muscula-ture of the Ciconiiformes [that] contributes sig-nificantly to a taxonomic clarification of thislarge, loosely applied assemblage of long-legged,long-necked, wading type birds" (page 361). Sucha conclusion would be expected if the Ciconi-iformes were entirely an artificial group (Olson,1979). In the following comparisons we use "Ci-coniiformes" for the families traditionally in-cluded in the order, but excluding flamingos.

One of Vanden Berge's most interesting discov-eries was a previously unknown thigh muscle,unique to flamingos, to which he later appliedthe name M. iliotibialis medialis (Vanden Berge,1976). In preliminary dissections we discoveredthis muscle in Cladorhynchus, whereas it was absentin Recurvirostra and Himantopus. This highly im-portant discovery led us to examine the rest of theappendicular myology of Cladorhynchus to makefurther comparisons with flamingos. Details arepresented in the following section. The tabula-tions below summarize myological comparisonsbetween flamingos, Cladorhynchus, and Ciconi-iformes, especially storks. Naturally, not all ofthese characters are of equal value.


A. Unique Character in Which Flamingos and

Cladorhynchus Differ from All Known Birds

1. Presence of M. iliotibialis medialis.

B. Characters Shared by Flamingos andCladorhynchus That Differfrom All Ciconiiformes

1. Origin of M. latissimus dorsi pars cranialis not partlyfrom cervical vertebrae.

2. Origin of M. latissimus dorsi pars metapatagialis partlyfrom synsacrum.

3. Far caudal origin of M. serratus superficialis pars cau-dalis.

4. More distal origin of M. scapulohumeralis caudalis.5. M. pectoralis pars subcutanea thoracica and pars sub-

cutanea abdominalis present and well-developed.6. M. tensor propatagialis pars brevis with two separate

tendons.7. Greater fusion of tendons of M. extensor metacarpi

radialis.8. Distal head of M. extensor longus digiti majoris vestigial.9. Both origin and insertion of M. adductor alulae fleshy.

10. Nature of origin of M. interosseus ventralis (particularlydifferent from storks).

11. Combination of M. iliofemoralis internus being shortand stout and inserting caudal to origin of M. femoro-tibialis internus.

12. Ambiens present and inserting directly on belly of M.flexor perforatus digiti II.

13. Nature of flexor cruris lateralis complex.14. Greater proximal width of flexor cruris medialis.15. None of the digital flexors with a fleshy origin from shaft

of fibula.16. M. flexor perforans et perforatus digiti II with two heads.17. M. flexor perforans et perforatus digit III with two

heads.18. Insertion of M. popliteus distal to origin.19. Deep flexor tendons type IV.20. M. extensor hallucis absent.21. M. flexor hallucis brevis vestigial.

C. Characters Shared by Flamingos,Cladorhynchus, and One or More

Ciconiiformes but Differing from Storks

1. Propatagial slip of M. biceps brachii present and well-developed.

2. M. iliofemoralis ("B") present.3. Scapular origin of M. triceps brachii entirely fleshy.4. Origin of M. latissimus dorsi caudalis from last free

vertebra, synsacrum, and fasciae.5. M. scapulohumeralis cranialis present.6. Origin of M. coracobrachialis cranialis tendinous.7. Ventral head of M. deltoideus minor present and rea-

sonably well-developed.

8. Postacetabular portion of M. iliotibialis lateralis present.9. Pars accessoria, M. flexor cruris lateralis, present.

10. Extensive fleshy insertion of M. pubo-ischiofemoralis.11. More proximal origin of M. adductor digiti II.12. Insertion of M. iliofemoralis internus caudal to origin of

M. femorotibialis internus.

D. Characters in Which Flamingos and StorksDiffer from Other Ciconiiformesbut Also Agree with Cladorhynchus

1. M. fibularis brevis absent.2. Nature of origin of M. serratus profundus.3. More tendinous origin of M. rhomboideus superficialis

pars scapularis.4. M. iliofemoralis short and stout.

E. Characters in Which CladorhynchusDiffers from Flamingos

1. M. iliotibialis medialis inserts on patellar tendon.2. More fusion in elements of tensor propatagialis complex.3. M. flexor hallucis brevis arises more proximally.4. M. subcoracoideus with two distinct heads.5. Origin of M. humerotriceps.6. Lateral head of M. subscapularis smaller and more

dorsolaterally situated.7. Nature of M. pectoralis propatagialis.8. Nature of insertion of M. coracobrachialis cranialis.9. Belly of M. ulnometacarpalis ventralis longer.

10. Total lack of fusion in the parts of M. femorotibialis.11. Heads of M. gastrocnemius with less distal fusion.12. M. serratus superficialis pars metapatagialis from two

distinct slips.13. M. caudofemoralis ("A") present.14. Belly of M. flexor digitorum profundus notched by

insertion of M. brachialis.15. M. pronator profundus fused with M. extensor longus

digiti majoris.16. Origin of M. fibuiaris longus not fleshy.

The appendicular myology of flamingos in noway supports a relationship with the Ciconi-iformes. The preceding tabulation lists 22 char-acters in which flamingos differ from all Ciconi-iformes and 34 in which they differ from theCiconiidae in particular. In all of these charactersflamingos agree with Cladorhynchus, even to thepossession of M. iliotibialis medialis, a muscle notknown to occur in any other birds.

Differences between Cladorhynchus and flamin-gos are summarized in tabulation E. Of these,characters 1 through 3 are but modifications offlamingo-like conditions not found in any cico-niiform. Characters 4 through 8, and possibly 9,

NUMBER 316 17

probably represent a single functional complexassociated with changes in pneumatization of thehumerus and consequent modification of theshoulder. Characters 10 through 13 involve onlyfusion of muscles to derive a flamingo-like stateand may be size-related. Character 13 has notaxonomic value even at the family level. Thesignificance of characters 14 through 16 is notapparent, but all of them are variable within theCiconiiformes, as currently constituted.

The differences between flamingos and Clado-rhynchus are both fewer in number and less insignificance than those between flamingos andCiconiiformes. Appendicular myology providesstrong evidence in favor of a relationship betweenflamingos and Charadriiformes, particularlyCladorhynchus, whereas there is almost no similar-ity between flamingos and storks.

APPENDICULAR MYOLOGY OF Cladorhynchus

We dissected the appendicular musculature ina single adult female of Cladorhynchus leucocephalus(SAM B30446), making comparisons chiefly withthe descriptions of ciconiiform myology in Van-den Berge (1970). Beyond checking for the pres-ence of M. iliotibialis medialis in various Recur-virostridae, we have made no attempt to compareCladorhynchus with other shorebirds, nor have werepeated Vanden Berge's dissections of flamingosexcept to check a few specific muscles in tworather emaciated specimens of Phoeniconaias minor,the only fluid-preserved flamingos immediatelyavailable to us. Vanden Berge (1970) did notexamine this species but dissected specimens ofPhoenicopterus ruber, P. chilensis, and Phoenicoparrusjamesi. It is not our purpose to provide a detaileddescription of the myology of Cladorhynchus, butmainly to present a comparative assessment of itssimilarities and differences with regard to flamin-gos.

We have adopted the myological nomenclatureof the Nomina Anatomica Avium (Baumel et al.,1979), using the sequence of Vanden Berge's(1970) original paper for facility of comparisonand quoting his nomenclature in parentheseswhen it differs substantially from that of the NAA.

For certain muscles, Vanden Berge (1970)found little or no significant variation within theCiconiiformes. Of these, those in Cladorhynchusthat did not differ from his descriptions are listedas follows and are not discussed further: M.brachialis, M. supinator, M. flexor digitorum su-perficialis ("sublimus"), M. extensor digitorumcommunis, M. extensor metacarpi ulnaris, M.ectepicondylo-ulnaris ("anconeus"), M. flexor al-ulae ("flexor pollicis"), M. extensor brevis alulae("extensor pollicis brevis"), M. abductor digitimajoris ("abductor indicis"), M. iliotrochanteri-cus cranialis, M. iliotrochantericus medius, M.iliofibularis ("biceps femoris"), M. tibialis crani-alis, M. extensor digitorum longus, M. plantaris.

M. latissimus dorsi pars cranialis{"anterior")

This is a broad, very thin muscle arising as athin tendinous sheet from the spinous processesof the second through fourth thoracic vertebrae.The insertion is fleshy on the cranial surface ofthe proximal third of the humerus.

Vanden Berge (1970) found that this muscle inflamingos originated from the first through thirdthoracics, whereas in all other Ciconiiformes theorigin involved one or more cervical vertebrae.Despite the fact that in one skeleton of Cladorhyn-chus we examined, the first thoracic rib had nosternal component and would therefore have tobe considered as a "cervical," our evidence sug-gests that the origin of M. latissimus dorsi crani-alis is from the same series of vertebrae as inflamingos. In neither flamingos nor Cladorhynchusis the origin in part from the true cervical seriesas it is in Ciconiiformes.

M. latissimus dorsi pars caudalis^posterior'")

The origin is tendinous mainly from the cranialmargin of the ilium and also from the last freevertebra and surrounding fasciae. This is as de-scribed by Vanden Berge (1970) for flamingosand ibises and differs from that of herons (dorsalvertebrae only), Balaeniceps (last 2-3 free verte-brae and synsacrum) and storks (origin not in-volving fasciae).


M. latissimus dor si pars metapatagialis

This slender but well-developed muscle origi-nates from the cranial tip of the synsacrum andpartly from the last free thoracic vertebra. Theinsertion subtends the humeral feather tract. Fla-mingos agree with Cladorhynchus in having theorigin partly from the synsacrum whereas in theCiconiiformes it is entirely from the vertebrae.

M. rhomboideus superficialis

PARS CLAVICULARIS.—The origin is a thin ten-dinous sheet from the last cervical and first tho-racic vertebrae and the insertion is fleshy alongthe cranial portion of the dorsal (vertebral) mar-gin of the scapula extending ventrad on the ad-jacent dorsal projection of the clavicle, ventral tothe origin of M. tensor propatagialis.

PARS SCAPULARIS.—The origin is a tendinoussheet extending from the second thoracic vertebrato the last free thoracic. The insertion is fleshy onthe dorso-medial margin of the cranial three-fourths of the scapular blade, extending as farcaudally as the bend in the blade.

According to Vanden Berge (1970) a moreextensively tendinous pars scapularis differen-tiates flamingos and storks from other Ciconi-iformes; Cladorhynchus fits this pattern, though theorigin is longer than reported for those forms.

M. rhomboideus profundus

This deep muscle arises as a broad, thin, ten-dinous sheet from the spinal processes of the firstthrough the fourth thoracic vertebrae. The inser-tion is fleshy along the costal aspect of the caudalfour-fifths of the scapula. This condition agreesgenerally with that of most Ciconiiformes exceptBalaeniceps (Vanden Berge, 1970).

M. serratus profundus

This muscle consists of three distinct slips aris-ing fleshy from the last cervical and first twothoracic ribs; these insert fleshy on the costalsurface of the cranial half of the scapula.

Flamingos and storks both agree with the con-formation in Cladorhynchus. These differ, on theone hand from herons, in which the muscle arisesonly from cervical vertebrae (four fleshy slips) or

from cervical vertebrae and the first thoracic (fiveslips), and on the other hand from ibises andBalaeniceps, in which the three slips arise from thelast two cervical and first thoracic vertebrae.

M. serratus superficialis pars cranialisand pars caudahs

These are entirely separate muscles in Clador-hynchus. Pars cranialis arises fleshy from the lastcervical and first thoracic ribs, with a thin ten-dinous sheet spreading to the penultimate cervi-cal. The insertion is by a thin tendinous sheetattaching to the cranial sixth of the ventral mar-gin of the scapula. Pars caudalis originates froma thin sheet of tendon on the third through sixththoracic ribs and inserts by a thin tendinous sheeton the ventral margin of the caudal third of thescapula. In the far caudal origin of M. serratussuperficialis caudalis, flamingos differ from Ci-coniiformes and agree with Cladorhynchus. In hisone specimen of Phoenicopterus chilensis, VandenBerge (1970:303) found the origin to be on tho-racics 2 through 4, as in storks and some ibises.Phoenicoparrus jamesi and Phoenicopterus ruber wereshown to be unique in the Ciconiiformes in hav-ing the origin from thoracics 2 through 5, and inthe latter species "the origin extended posteriorlyalmost to the sixth thoracic," the condition metwith in Cladorhynchus.

M. serratus superficialispars metapatagialis

This well-developed muscle consists of two dis-tinct slips, one arising fleshy from thoracic ribs 4and 5 and the other arising tendinously from thesixth thoracic. The insertion is on the humeralfeather tract—we did not notice that it fused withM. latissimus dorsi metapatagialis or that therewas a tendinous slip to the scapula as describedfor flamingos and herons (Vanden Berge, 1970),but either condition may have been obscured byspecimen damage.

Cladorhynchus appears to differ from flamingosin that separate slips are not mentioned by Van-den Berge (1970). The more caudal origin inCladorhynchus agrees with Phoenicopterus ruber butthe origin in P. chilensis and Phoenicoparrus is re-

NUMBER 316 19

ported to be from thoracics 3 to 5 as in storks andibises (Vanden Berge, 1970).

M. scapulohumeralis cranialis

This is a very small, entirely fleshy muscle, onlya few fibers wide. The origin is from the ventralsurface of the glenoid facet of the scapula; at itsmidpoint it is partly attached to M. scapulohu-meralis caudalis. The insertion is just distal to theridge dividing the tricipital fossa, between thehumeral heads of M. triceps brachii.

Vanden Berge (1970) found this muscle onlyin herons and two species of flamingos, but not inPhoenicopterus ruber or any other Ciconiiformes. Itwas well-developed in the fossil flamingos Palae-lodus and Juncitarsus and its reduction or loss inmodern species may be correlated with increasedpneumatization of the humerus and the conse-quent loss of area for insertion.

M. scapulohumeralis caudalis

The origin is fleshy from the full length of thedorsal and lateral surfaces of the scapula exceptfor that part occupied by M. subscapularis caputlaterale ("pars externa"). The insertion is bothtendinous and fleshy on the distal end of theventral rim of the tricipital fossa. The more pos-terior origin is apparently similar to that of fla-mingos and differs from herons and storks (Van-den Berge, 1970).

M. pectoralis

The origins of this great muscle are both fleshyand tendinous from the postero-lateral process ofthe sternum and adjacent ribs, fleshy from theventral rim of the carina and from the lateralsurface of the clavicle and Membrana sternocor-acoclavicularis. The insertion is partly by a ten-dinous connection to the bicipital surface of thehumerus but with the principal tendon attachingon the ventral margin of the deltoid crest. M.pectoralis is partially divided into two distinctlayers; a larger external one with a tendinoussheet on its antero-medial surface separating itfrom a smaller medial muscle layer. These layersfuse only in the posterior fourth of the wholemuscle mass.

As discussed previously, much was made in theearlier literature of the layering of this muscle inflamingos, but as we have seen, this serves only toseparate them from ducks and will not distinguishthem from a variety of other birds, includingshorebirds such as Cladorhynchus.

M. pectoralis pars propatagialis

This subcutaneous slip of the pectoralis is onlyweakly differentiated in Cladorhynchus, being rep-resented by the separation of fibers attachingalong a raphe that also forms the origin of M.tensor propatagialis. Vanden Berge (1970) indi-cates this muscle to be best developed in heronsand flamingos within the Ciconiiformes, but suchwould not appear to be the case in Cladorhynchus.

M. pectoralis pars subcutanea thoracicaand abdominalis

Pars subcutanea thoracica and Pars subcuta-nea abdominalis are present and constitute abroad, well-developed sheet on the skin of thebreast and flank, attaching to the lateral abdom-inal wall and to the pubis. Vanden Berge (1970)found both slips to be well-developed in flamin-gos. In herons, pars subcutanea abdominalis ispresent or absent; pars subcutanea thoracica ispresent but very narrow. Both slips are absent instorks, ibises, and Balaeniceps. Thus, flamingosagree much better with Cladorhynchus in this re-spect than with any Ciconiiformes.

M. supracoracoideus

The origin is fleshy from the dorsolateral sur-face of the carina and the medial portion of thebody of the sternum; although the belly lies overthe sternocoracoclavicular membrane, it does notseem to attach to it. The muscle tapers to atendon that passes through the triosseal canal toinsert dorsally on the dorsal tubercle ("externaltuberosity") of the humerus. Vanden Berge(1970) does not specifically mention the conditionof this muscle in flamingos.

M. coracobrachialis cranialis

This is very well developed in Cladorhynchus,arising tendinously from the lateral surface of the


head of the coracoid and the glenoid joint cap-sule. The insertion is very strong both by tendonsin the ligamental furrow and to the prominencebetween the ligamental furrow and the bicipitalfurrow (impressio m. coracobrachialis cranialis)and by fleshy fibers on most of the bicipitalsurface. Distally it forms a sheet of tendon that iscontinuous with the origin of the dorsal (cora-coidal) head of M. biceps brachii. The tendinousorigin agrees with flamingos, herons, and Balaen-iceps and differs from storks and ibises, but theinsertion in Cladorhynchus is quite different fromthat described for flamingos or any of the Cicon-iiformes by Vanden Berge (1970), and is probablytypical for the Charadriiformes, to judge by theconformation of their humeri.

M. coracobrachialis caudalis

The origin is fleshy from the lateral process ofthe sternal end of the coracoid and the adjacentportion of the sternum, with a tendinous connec-tion to the tip of the craniolateral (sternocora-coidal) process of the sternum. The insertion is bya tendon to the tip of the ventral tubercle ("in-ternal tuberosity") of the humerus. The originpartly from the sternum agrees with flamingos,two species of herons, and two genera each ofstorks and ibises studied by Vanden Berge (1970),whereas in all other forms the origin was foundto be confined to the coracoid.

M. sternocoracoideus

The origin is mixed fleshy and tendinous fromthe ventral surface of the craniolateral process ofthe sternum and the first two sternal ribs. Theinsertion is fleshy in the sternocoracoidal impres-sion and process of the coracoid. We did notdetect any layering in this muscle as described byVanden Berge (1970) for flamingos and storks.

M. subcoracoideus

This is a large, bulging, rounded muscle masswith two distinct heads. The dorsal head arisesfleshy from the medial surface of the dorsal partof the clavicle, while the larger ventral head arisesfleshy entirely from the medial surface of thesternocoracoclavicular membrane. These heads

become extensively commingled with M. sub-scapularis and contribute to a common tendinousinsertion with that muscle.

M. subcoracoideus is weakly developed in her-ons and Balaeniceps and is more extensive in otherCiconiiformes and flamingos. Cladorhynchus differsfrom all of them in having two separate heads, acondition that is typical of many other birds,however (George and Berger, 1966).

M. subscapularis

This muscle arises from two distinct heads lyingon opposite sides of the scapula. The lateral (ex-ternal) head originates fleshy from the dorsolat-eral surface of the cranial fifth of the scapula,behind the glenoid. The medial (internal) headoriginates fleshy from most of the costal or ventralsurface of the scapula, with a long, narrow, ta-pering extension running the full length of theventral margin. The insertion is a common ten-don with M. subcoracoideus on the ventral tu-bercle of the humerus. In the clearly dorso-lateralposition and short length of the lateral head,Cladorhynchus differs from the condition describedfor flamingos and the Ciconiiformes, in all ofwhich except herons the external head is de-scribed as being as large as, or larger than, theinternal head (Vanden Berge, 1970).

M. tensor propatagialis

Pars longa has a partly fleshy attachment onthe proximal end of the deltoid crest and a broadtendinous attachment to the propatagial slip fromM. pectoralis. In addition, the belly extends prox-imally over the shoulder and has a common originwith pars brevis from the dorsal portion of theclavicle.

Par brevis terminates as two separate tendons(Figure 5a), the proximal tendon inserting di-rectly on dense fasciae over the tendon of originof M. extensor metacarpi radialis. The distaltendon is closely bound to that of pars longaproximally; at the level of the propatagial elasticligament, it diverges to insert on the proximalforearm as a bifurcated tendon. The distal ramusof this bifurcation is wider and receives an inelas-tic tendinous slip from the elastic ligament itself.

NUMBER 316 21

The elastic ligament of the propatagium isdirectly continuous with the tendon of pars longaand, at the same level, receives the propatagialslip from M. biceps brachii. The elastic ligamentthen trifurcates in a complicated manner, thethree rami merging distally to form a distal in-elastic tendon that is buried in a sheath in theleading edge of the propatagium. This inelastictendon encloses a tough fibrous dilation imbed-ded in the skin opposite the insertion on thecarpal joint. In addition to the inelastic tendinousslip that unites with the bifurcated tendon of parsbrevis, previously mentioned, there is also a verynarrow tendon from the leading elastic ramus; itterminates on the retinaculum that restrains thetendon of the superficial digital flexor.

Vanden Berge (1970) reports that flamingosdiffer from all Ciconiiformes in possessing twoseparate tendons of pars brevis. This, in fact,appears to be a feature characteristic of Charad-riiformes, as Fiirbringer (1902, pi. 21: figs. 218—

219) illustrates a similar condition in Larus (Lar-idae) and Parra (— Jacana, Jacanidae). In thespecimen of Phoeniconaias we examined (Figure5b), M. tensor propatagialis was similar to that ofCladorhynchus except that there was no fusion ofthe distal tendon of pars brevis with the tendonof pars longa proximally; the smaller propatagialslip of the biceps, instead of fusing to this com-bined tendon gives rise to a small, partially elastic,separate tendon that inserts far distally on thetendon of pars longa, near the distal end ofthe radius. Most of the differences exhibited byCladorhynchus could probably be accounted for byfusion of elements present in flamingos. M. tensorpropatagialis in flamingos and Cladorhynchus isvery different from that of storks (see Fiirbringer,1902, pi. 21: fig. 222).

M. deltoideus major

This arises fleshy from the dorsal portion of thescapula between the acromion and the glenoid

biceps brachii __propatagial slip

biceps brachiipropatagial slip

tensor propatagialispars brevis

tensorpropatagialis

pars brevis

FIGURE 5.—Dorsal view of right wing to show configuration of M. tensor propatagialis inCladorhynchus leucocephalus (a) and Phoeniconaias minor (b). The double tendon of pars brevisappears to be a fairly typical charadriiform character that is absent in storks. The biceps slip isalso typical of Charadriiformes and occurs in "Ciconiiformes" only in the Threskiornithidae.Blackened portions indicate elastic tendons. (Not to scale.)


facet and also from the strong coraco-scapularligament. The belly lies on the dorsal surface ofthe deltoid crest and has a fleshy attachment onthe edge of the crest and along a line taperingdistally on the shaft for one-third the length ofthe humerus, becoming tendinous towards thedistal end.

The fleshy origin of this muscle is as describedfor flamingos, Ciconia, Eudocimus, and Plegadis anddiffers from other genera of Ciconiiformes, inwhich the origin may be partly tendinous or morelateral (Vanden Berge, 1970).

M. deltoideus minor

This muscle consists of two distinct heads orig-inating in the triosseal canal and passing out overthe insertion of M. supracoracoideus. They areentirely separate, at least after emerging from thetriosseal canal, with the larger dorsal head com-pletely overlying the ventral head. The dorsalhead originates fleshy, partly from the acromionof the scapula but mostly from the procoracoidprocess of the coracoid. The insertion is fleshy onthe proximal end of the deltoid crest just distal tothe dorsal tubercle. The origin of the ventral headwas not traced, having come away with the re-moval of M. supracoracoideus. The belly becomestendinous at the level of the dorsal tubercle of thehumerus and inserts in tissue in the ligamentalfurrow on the ventral side of the humerus. Theventral head is either very weak or absent instorks and Balaeniceps, better developed in flamin-gos, ibises, and certain herons, and best developedin other herons (Vanden Berge, 1970); flamingosthus agree with Cladorhynchus and differ fromstorks in this respect.

M. triceps brachii

The scapular head (M. scapulotriceps) arisesboth fleshy and tendinous from the lateral side ofthe scapula just caudal to the glenoid facet, andalso by a short adjacent retinaculum to the edgeof the scapula slightly more caudally. There is atendinous retinaculum to the humerus in com-mon with the insertion of M. latissimus dorsi parscaudalis. The belly continues along the entireshaft of the humerus to insert on the proximal

end of the ulna.The humeral head (M. humerotriceps) arises

fleshy from two portions bisected by the insertionof M. scapulohumeralis caudalis. These two por-tions fuse on a median raphe and have fleshyorigins from the entire caudal surface of the shaftof the humerus, occupying the double, non-pneu-matic fossa characteristic of most Charadrii anddiffering from all Ciconiiformes and flamingos.The common tendon of these two parts insertsboth fleshy and tendinous on the olecranon.

Flamingos agree with Cladorhynchus and ibises,except Ajaia, in having the scapular origin en-tirely fleshy, whereas in other Ciconiiformes it isat least partly tendinous and in storks there aretwo distinct tendons of origin (Vanden Berge,1970).

M. biceps brachii

This muscle arises from two heads; the moreventral humeral head originates by a short tendonfrom the ventro-distal edge of the bicipital crest.The smaller dorsal or coracoidal head arises fromthe broad, strong, distal tendon of M. coraco-brachialis cranialis. This head gives rise to thestrong slip to M. tensor propatagialis. The twoheads give rise to entirely separate tendons, themore ventral one inserting on the proximal endof the radius, and the dorsal one inserting at theproximal end of the "brachialis" scar of the ulna.

In the Ciconiiformes, the biceps slip is well-developed only in flamingos and ibises. It is ves-tigial in herons and absent in storks and Balaeni-ceps. In the more distinct division of the belly ofthe biceps, Cladorhynchus appears to differ some-what from the forms described by Vanden Berge(1970), in all of which the belly is essentiallyundivided proximally.

M. expansor secundanorum

This is a small, diffuse muscle inserting on thebases of 3 or 4 secondaries at the proximal end ofthe ulna and from fasciae over the dorsal surfaceof the elbow joint. The very slender tendon ex-tends proximad and enters the axilla. We did nottrace the delicate origins, but encountered one ofthem that emerged from connective tissue in the

NUMBER 316 23

axilla and attached to the dorsal surface of thesternocoracoidal ligament. The belly is notweakly developed as in herons and is therefore inagreement with flamingos, storks, and ibises.

M. pronator superficialis

The origin is both tendinous and fleshy fromthe cranioventral corner of the distal end of thehumerus ("attachment of anterior articular liga-ment" of Howard, 1929). The insertion is by atendinous sheet opposite the insertion of M. su-pinator on the cranioventral edge of the proximalthird of the radius. There is an indication of adivision in the belly. The origin is apparentlyentirely tendinous in flamingos and Ciconiiformesexcept Eudocimus and Ajaia (Vanden Berge, 1970).

M. pronator profundus

This arises tendinously from the ventral epicon-dyle (entepicondyle) of the humerus and insertsfleshy along the ventral surface of the proximalhalf of the radius, including the proximal sixth orso. It fuses distally with M. extensor longus digitimajoris ("extensor indicis longus"). This is likethe condition described for ibises and most heronsand differs from flamingos, storks and Balaenicepsin which the muscle becomes tendinous distallyand does not fuse with M. extensor longus digitimajoris (Vanden Berge, 1970).

M. flexor carpi ulnaris

This is the most caudal of the ventral musclesof the ulna and arises as a strong tendon from theflexor process at the distal end of the humerus.There is a proximal division of the belly, thecombined fibers of which run the full length ofthe ulna. The insertion is tendinous on the ulnare.Vanden Berge (1970) does not mention any dis-tinctive condition in this muscle in flamingos.

M. flexor digitorum profundus

The origin is from two slips on the ventralsurface of the proximal end of the ulna that areseparated by the insertion of M. brachialis. Thetendon forms about midway along the ulna, runsover the pisiform process, under M. flexor alulae,and extends along the cranial edge of the carpo-

metacarpus to insert at the base of the distalphalanx of the major digit. The belly is alsonotched by M. brachialis in ibises, herons, andBalaeniceps but evidently not in flamingos andstorks (Vanden Berge, 1970).

M. ulnometacarpalis ventralis

This differs from flamingos and Ciconiiformesin that the belly is longer, tapering over the distalthree-fourths of the ulna.

M. extensor metacarpi radialis

This muscle is clearly divided into two bellies,the larger ventral one arising fleshy from theectepicondylar spur of the humerus and the dorsalone from a superficial tendon in common withthe insertions of M. tensor propatagialis brevis.About three-fourths distally the length of theradius they form a common tendon that runsalong the leading edge of the wing through agroove in the radius, over the carpal joint, toinsert on the extensor process of the alular meta-carpal, there partly fusing with the underlyingtendon of M. extensor longus alulae ("pollicislongus").-

In flamingos the two bellies fuse distally andproduce a common tendon, whereas in all Cicon-iiformes the bellies are separate and the tendonsdo not fuse except at the distal end of the radius.Cladorhynchus approaches the condition in fla-mingos in that the fusion of the tendons is fartherproximad, although the bellies are neverthelesscompletely separate.

M. extensor longus alulae("pollicis longus")

This is about as described for flamingos andCiconiiformes except that we could not detectthat there were two heads as reported by VandenBerge (1970).

M. extensor longus digiti majoris("indicus longus")

This muscle is as described for flamingos byVanden Berge (1970). The distal head is ex-tremely small, only about 3 to 4 mm long andless than 1 mm in width. It is better developed in


Ciconiiformes but in flamingos it occurs only asa few fleshy fibers in Phoenicopterus chilensis and itis absent in P. ruber and Phoenicoparrus jamesi (Van-den Berge, 1970).

M. ulnometacarpalis dorsahs

This muscle appears to be similar to the typicalform in flamingos and Ciconiiformes in general;there are two bellies, in contrast to the single bellyreported for Leptoptilos (Vanden Berge, 1970).

M. abductor alulae ("pollicis")

There is no suggestion of a division as reportedfor Phoenicopterus ruber but not other flamingos(Vanden Berge, 1970).

M. adductor alulae ("pollicis")

Both the origin and the insertion are fleshy; theinsertion is on the alular bone and not on feathers.The same is true for flamingos, whereas in allCiconiiformes except Balaeniceps the origin is ten-dinous distally and the insertion is mainly onfeathers, except in ibises.

M. interosseus dorsalis

The origin is as described for Ciconiiformes.The tendon bifurcates after the metacarpopha-langeal joint and sends a short tendon to thedorsal surface on the proximal half of the proxi-mal phalanx of the major digit. The other ramuscontinues distally and inserts as a broad band onthe proximo-dorsal edge of the distal phalanx,with a weaker extension going out to the tip ofthe phalanx and the bases of the distal primaries.No ramus to the proximal phalanx is mentionedby Vanden Berge (1970) for the Ciconiiformes orflamingos.

M. interosseus ventralis ("volaris")

The origin within the metacarpal space is fromthe entire margin of the major metacarpal ("II")and all but the distal fifth of the minor metacar-pal ("III"). This is as in flamingos and Balaenicepsexcept that in those forms the attachment to theminor is less extensive and is confined to theproximal half of the bone. In storks there is noattachment to the minor, whereas in herons and

ibises the origin runs the entire length of theminor metacarpal.

M. flexor digiti minons ("///")

The origin is fleshy, begins at the insertion ofM. ulnometacarpalis dorsalis, and extends alongthe remainder of the caudal margin of the minormetacarpal ("III"), with the longest fibers arisingin the distal third of the belly. This is in agree-ment with flamingos, storks, and Balaeniceps,whereas in herons and ibises the main fleshy fibersare more proximally situated.

M. iliotrochantericus caudalis

This muscle originates fleshy from almost theentire preacetabular ilium, the origin being some-what more aponeurotic posteriorly. The insertionis a broad tendon along the craniolateral surfaceof the trochanter, extending proximally deep toM. iliofemoralis externus ("glutaeus medius etminimus"). Vanden Berge (1970) notes that inthe stork Leptoptilos this muscle is covered only bythe very extensive aponeurosis of M. iliotibialislateralis, whereas in the Phoeonicopteridae, as inCladorhynchus, it is covered by the fleshy bellies ofthe iliotibialis complex.

M. iliofemoralis externus("glutaeus medius et minimus")

The origin is fleshy from the edge of the iliumabove the acetabulum, becoming aponeuroticcaudally. The insertion is by a fairly long tendonattaching in the middle of the lateral aspect ofthe proximal end of the femur. As in the Phoen-icopteridae, Ciconiidae, and Threskiornithidae,the insertion is cranial to that of M. ischiofemor-alis, in contrast to Balaeniceps and the Ardeidae,in which it is caudal (Vanden Berge, 1970).

M. iliofemoralis internus ("iliacus")

This is a short, diagonal muscle originatingfleshy from the ventral border of the preaceta-bular ilium and inserting fleshy on the medialsurface of the femur, below the head, proximaland caudal to the origin of M. femorotibialisinternus. The insertion is in agreement with bothflamingos and ibises and differs from storks, in

NUMBER 316 25

which the insertion is cranial to the origin of M.femorotibialis internus (Vanden Berge, 1970).Cladorhynchus agrees with both flamingos andstorks in that this muscle is short and stout,whereas in other Ciconiiformes it is long andnarrow.

M. ambiens

The ambiens is present in Cladorhynchus andarises from a fairly long tendon from the pectinealprocess of the pelvis. Its long, thin tendon ofinertion runs through the patellar tendon andthen deep to the insertion of M. iliofibularis toinsert mainly on the belly of M. flexor perforatusdigiti II, but also with a tenuous extension to M.flexor perforatus digiti III. The ambiens is absentin herons and Balaeniceps, and possibly in Scopusand some storks (see references cited by VandenBerge, 1970:333). In those Ciconiiformes in whichthe ambiens is present, the insertion "is directlycontinuous with a small medial head" of M.flexor perforatus digiti II, except in flamingos, inwhich the insertion is on the belly of that muscle(Vanden Berge, 1970:333). In this respect then,flamingos are more similar to Cladorhynchus thanto any of the Ciconiiformes.

M. iliotibialis lateralis

This is a broad, thin sheet originating as anaponeurosis from the dorsal crest of the ilium andcranially from the dorsal crest of the synsacrum.It covers much of the lateral surface of the thigh,including a portion of M. iliofibularis ("bicepsfemoris") caudally and Mm. iliotrochanterici cra-nially. The postacetabular and preacetabularportions are continuous and not separated prox-imally by a tendinous sheet as in flamingos, butdistally they are separated by an aponeurosis(Figure 6b) that fuses with the underlying M.femorotibialis and inserts on the patellar tendon.Flamingos differ from Cladorhynchus in having thetwo heads of this muscle "separated by a thintendinous sheet covering the II. troc. post,[caud.]" (Vanden Berge, 1970:335), but they aremore similar to Cladorhynchus than to the Ciconi-idae or Scopus, in which the postacetabular por-tion is absent and M. iliofibularis is exposed.

M. iliotibialis medialis andM. iliotibialis cranialis ("sartorius")

In Cladorhynchus these two muscles (Figure 6b)originate from a common aponeurosis on themost cranial portion of the dorsal crest of thesynsacrum. M. iliotibialis medialis is a broad(8.5 mm), thin strap underlying the M. iliotibialiscranialis, which is considerably narrower (4.5 mmat midpoint). The caudal portion of M. iliotibialismedialis is covered by M. iliotibialis lateralis.Both Mm. iliotibialis medialis and cranialis inserton the patellar tendon, the former being moremedial to the latter. Vanden Berge (1976) reportsthat M. iliotibialis medialis inserts at the base ofthe inner cnemial crest of the tibiotarsus in fla-mingos; thus in Cladorhynchus the insertion on thepatellar tendon probably represents a slightlymore primitive state. M. iliotibialis medialis waspresent and developed to the same degree on bothsides of the specimen of Cladorhynchus dissected inthis study.

We examined specimens of Recurvirostra ameri-cana (Figura 6a) and Himantopus mexicanus, in nei-ther of which was M. iliotibialis medialis present.The occurrence of this muscle in Cladorhynchusand flamingos and no other known birds, is verystrong evidence of relationship. In view of theamount of concurring data, it seems incontro-vertible.

M. femorotibialis

M. FEMOROTIBIALIS INTERNUS.—This is a sepa-rate muscle originating fleshy along the entiremedial face of the femur distal to the insertion ofM. iliofemoralis internus ("iliacus"). It insertspartly by a strong tendon on the medial side ofthe tibial head just behind the inner cnemial crestand contributes weakly to the medial portion ofthe patellar tendon, covering much of the medialsurface of the distal end of the femur. This tendondoes not cover the tendon of insertion of M.iliotibialis medialis as reported for Phoenicopterus(Vanden Berge, 1970).

M. FEMOROTIBIALIS MEDIUS.—The origin isfleshy from the caudolateral surface of the shaftof the femur, beginning at the level of the tro-


illotrochanterlcuscranialis/

iliotrochintericuscaudaHs

iliafcmoralis•xttrnus

ilietrochantericuiialii

iliofemoralisexternus

ftmorotlbialit

fcmorotibialit

FIGURE 6. Lateral view of right th igh of Recurvirostra americana (a) and Cladorhynchtis leucocepha-

lus (b) with M. iliotibialis Iateralis reflected from origin. M. iliotibialis medialis is absent inRecurvirostra but present in Cladorhynchtis. This muscle occurs elsewhere only in flamingos.

chanter. It barely overlaps the insertion of M.iliotrochantericus cranialis and continues onaround to the cranial surface of the shaft at thelevel of the distal extent of the trochanter. Theinsertion forms much of the patellar tendon,which, as mentioned, is partly fused with thesuperficial aponeurosis of M. iliotibialis Iateralis.

M. FEMOROTIBIALIS EXTERNUS.—This is also anentirely separate entity, not fusing with femoro-tibialis medius. It arises fleshy from the caudola-teral distal three-fifths of the femur. It overliesthe iliofibular ("biceps") loop and inserts as afairly broad tendon on the patellar tendon, distalto most of the insertion of femorotibialis medius.

Cladorhynchus differs from flamingos, storks, andBalaeniceps in the lack of fusion of Mm. femoro-tibialis medius and externus. In the Ciconiiformesthe external muscle was reported as most clearlydeveloped in herons and ibises.

M. caudo-iliofemoralis {"piriformis")

M. ILIOFEMORALIS.—This muscle arises mainlyfleshy (tendinous caudally) from the ventral edgeof the postero-lateral corner of the ilium and

extends as a thin triangular sheet, encompassinga small tendinous area in the middle of the dorsaledge, to insert via a short tendon (fusing withthat of M. caudofemoralis) on the caudolateralface of the shaft of the femur about one-third theway distally.

M. CAUDOFEMORALIS.—This is present in

Cladorhynchus and arises from a long, thin tendonon the ventral surface of the caudal bulb andextends as a narrow strap across the thigh deepto M. flexor cruris Iateralis and dorsal to M. flexorcruris medialis, having a common insertion withM. iliofemoralis by a fairly long tendon.

These are two of the "Garrod formula" thighmuscles that have figured so prominently in muchof the past taxonomic literature. M. caudofemor-alis ("A" of Garrod, 1874) is absent in flamingos(including Phoeniconaias, this study) but is presentin Cladorhynchus. The same variation occurs withinthe Ciconiidae and the Ardeidae, in which thismuscle is present in some genera but absent inothers (Vanden Berge, 1970). Fisher and Good-man (1955) found M. caudofemoralis in two spec-imens of Grus americana but it was absent in a

NUMBER 316 27

third. Within the Charadriiformes it has beenreported as absent in Burhinus (George and Berger,1966). The late George Hudson (unpublishedMS, fide Vanden Berge) found it to be present inBurhinus bistriatus, however, but absent in severalgenera of Scolopacidae and in Steganopus andLobipes (Phalaropodidae).

M. iliofemoralis ("B" of Garrod) is present inCladorhynchus, flamingos, and ibises, but is uni-formly absent in other Ciconiiformes.

M. flexor cruris lateralis("semitendinosus")

Both parts of the muscle present somewhat ofan anomalous condition in Cladorhynchus (Figurela). Pars pelvica arises from a broad aponeurosison the caudo-lateral corner of the ilium, beingpartly fused with the aponeurosis of origin of M.transversus cloacae. It passes distally as a narrowstrap to insert on a very short, nearly verticalraphe shared with pars accessoria, the fibers ofwhich fuse with the slender belly of M. gastro-cnemius pars intermedia to insert fleshy in thepopliteal fossa of the femur, some of the fibersfusing medially with M. pubo-ischiofemoralispars lateralis. There is no medial tendon from theraphe to the underlying M. flexor cruris medialis.

Vanden Berge (1970:338) states that in herons,Balaeniceps, and flamingos pars accessoria is

clearly separable from M. gastrocnemius parsintermedia and that in the Phoenicopteridae "thefleshy insertion extends over the distal three-fifthsof the femoral shaft." He did not examine Phoen-iconaias, however, and the two specimens we dis-sected differed greatly from his description for theother genera of flamingos. In lateral view (Figure1b), the arrangement in Phoeniconaias is quite sim-ilar to that seen in Cladorhynchus except that theraphe between pars pelvica and pars accessoriaextends as a long diagonal intersection that iscontinuous with M. gastrocnemius pars interme-dia. Therefore, pars accessoria in Phoeniconaias isnot separable and attaches to a shorter area onthe distal end of the femur, far less than three-fifths of the femoral shaft. In medial view (Figure8) the raphe is more distinct and gives off atendinous slip to the tendon of insertion of M.flexor cruris medialis, unlike Cladorhynchus.

In most birds possessing both parts of the lat-eral flexor, the fibers meet at a greater angle thanin Cladorhynchus and Phoeniconaias. It is difficult toreconcile what we observed in Phoeniconaias withVanden- Berge's description of flamingos basedon Phoenicopterus and Phoenicoparrus. The similaritybetween Cladorhynchus and Phoeniconaias is never-theless considerable and these genera differ fromthe Ciconiiformes in general and the storks inparticular, as in that group the accessory part is

gistrocnemiuspart medialis

pars inttrmcdia

FIGURE 7.—Lateral view of right hindlimb showing the disposition of the M. flexor crurislateralis complex in Cladorhynchus leucocephalus (a) and Phoeniconaias minor (b).


FIGURE 8.—Right hindlimb of Phoeniconaias minor with M.flexor cruris lateralis reflected outward to reveal its medialsurface and the connection with M. flexor cruris medialis.This connection is absent in Cladorhynchus.

only a vestigial bundle of fibers that does notinsert on the femur at all (Vanden Berge, 1970).In the condition of the raphe and the loss of theconnection with the medial flexor, Cladorhynchusseems more specialized than any of the flamingos,Ciconiiformes, or more typical shorebirds.

M. flexor cruris medialis(' 'semimembranosus'')

This muscle arises fleshy from the ventral (pu-bic) margin of the ischium, overlying the pubisbehind the obturator foramen. It is a broad ta-pering strap that inserts by a tendon on a shortridge on the medial surface of the shaft of thetibiotarsus, about 1 cm distal to the medial artic-ular facet. Vanden Berge (1970) notes that inflamingos the medial flexor is extremely wideproximally compared with other Ciconiiformes.Without seeing these forms ourselves it is difficultto evaluate the condition in Cladorhynchus, but themuscle is nevertheless quite wide proximally.

M. ischiofemoralis

This arises fleshy from most of the concavitycaudal to the acetabulum, including the mem-brane of the ilioischiatic fenestra. It inserts bothtendinous and fleshy on the caudo-lateral surface

of the trochanter, caudal to the insertion of M.iliofemoralis externus. In agreement with flamin-gos, storks, and ibises, and in contrast to heronsand Balaeniceps, it lies cranial to M. flexor crurismedialis.

M. obturatorius medialis et lateralis^'obturator internus et externus")

In Cladorhynchus, as in all the families examinedby Vanden Berge (1970) except the Ardeidae, thetwo muscles are fused and cannot be separated.The major part forms an elongate oval arising onthe medial side of the pelvis from the margins ofthe ischio-pubic fenestra. It passes through theobturator foramen and inserts fleshy on the cau-dal aspect of the trochanter.

M. pubo-ischiofemoralis("adductor longus et brevis")

Pars lateralis and pars medialis are fused prox-imally so as to be indistinguishable; they originatefrom a broad tendon on the cranio-ventral por-tion of the ischium, covering the ischio-pubicfenestra and the corresponding portion of thepubis. The insertion is fleshy on the caudomedialsurface of the shaft of the femur above the pop-liteal fossa, the fibers coalescing distally with thecombined flexor cruris lateralis pars accessoriaand gastrocnemius pars intermedia, which insertsfarther distally. The fused heads are found inflamingos and all Ciconiiformes except herons,whereas the fleshy insertion in Cladorhynchus ismost similar to that described for flamingos (Van-den Berge, 1970).

M. fibularis [peroneus] longus

This muscle arises from a broad aponeurosisover the entire tibial head and underlying mus-culature, but is most strongly attached to the rimof the inner, and tip of the outer, cnemial crests.Distally the belly is divided by a long tendinousarea. The tendon extends down the lateral surfaceof the tibial shaft, sends off an aponeurotic at-tachment to the posterior rim of the internalcondyle and continues diagonally over the lateralsurface of the intertarsal joint and lateral to thehypotarsus to insert just distal to the hypotarsuson the tendon of M. flexor perforatus digiti III.

NUMBER 316 29

Vanden Berge (1970) indicates that in flamingosand Ciconiiformes other than herons, the originfrom the tibial crest and most of the fibula isfleshy, which is not true of Cladorhynchus.

M. fibularis [peroneus] brevis

Absent in Cladorhynchus. This muscle is presentin all Ciconiiformes except storks and flamingos,a fact that has often been mentioned in theliterature (see references in Vanden Berge, 1970:343). Its absence in Cladorhynchus, and also inRecurvirostra (George and Berger, 1966), negatesone more supposed storklike character of flamin-gos. Hudson's manuscript notes (fide VandenBerge) indicate this muscle to be unilaterallypresent in Recurvirostra americana, and absent in R.novaehollandiae, Himantopus, and certain otherCharadrii.

M. gastrocnemius

In Cladorhynchus this muscle arises from fourentirely separate heads (Figure la). The commontendon of the four heads is the most superficialon the caudal surface of the shank and inserts onthe hypotarsus.

PARS MEDIALIS ("INTERNA").—This originatesfleshy from the entire medial surface of the innercnemial crest, extending over onto the patellartendon so as to obscure the insertions of M.iliotibialis cranialis and medialis.

PARS INTERMEDIA ("MEDIA").—This originatesfleshy from the popliteal fossa in common withthe fused accessory part of the lateral cruralflexor. It becomes tendinous and inserts on thedistal margin of pars medialis about three-fourthsthe length of the latter.

PARS LATERALIS ("EXTERNA").—This arisesfleshy from the lateral supracondylar crest of thefemur, being overlain by M. femorotibialis exter-nus and overlying the iliofibular loop. It fuseswith the fourth head to form a tendon that inturn fuses midway down the shank with thetendon of pars medialis.

"FOURTH HEAD. '—This slender muscle arises inthe popliteal fossa from a long, narrow tendonseparate from, and distal to, the origin of thecombined pars intermedia and accessory lateral

crural flexor. It inserts distally on pars lateralisabout 6 mm before the latter becomes tendinous.

The four-headed gastrocnemius is also reportedfrom storks, ibises, and flamingos, but does notoccur in herons or Balaeniceps (Vanden Berge,1970). In flamingos the pars medialis covers theanterior surface of the knee and the entire patellartendon and meets the proximal edge of parslateralis laterally. This is not true of Cladorhynchusbut may well be a size-related factor. We notedalso that in Phoeniconaias there was considerablymore distal fusion of the heads of the gastrocne-mius, these being less easily separated and distin-guished than in Cladorhynchus.

Digital Flexors

Vanden Berge (1970) notes that at least someof the various flexor muscles of the toes havefleshy origins from the shaft of the fibula in all ofthe Ciconiiformes but not in the Phoenicopteri-dae. In Cladorhynchus, as in flamingos, there is nofleshy attachment between any of the flexors andthe shaft of the fibula.

M. flexor perforans etperforatus digiti II

This muscle arises from two distinct heads, oneoriginating by a short tendon on the articularcapsule opposite the lateral condyle of the femurand the other head from a broad, thin aponeu-rotic sheet partly coextensive with the patellartendon and partly covering the origin of M.fibularis longus. Some of the deep fibers alongthe midline fuse with those of the underlying M.flexor perforans et perforatus digiti III. The ten-don runs through the deepest, most medial,groove of the hypotarsus and the insertion issimilar to that in flamingos and ciconiiforms.

Flamingos agree with Cladorhynchus and differfrom all Ciconiiformes in having this muscle dou-ble-headed instead of single-headed.

M. flexor perforans et perforatus digiti III

This can also be considered as having twoheads, though these are more intimately con-nected than in the preceding muscle. The caudalhead is fleshy and arises from the head of the


fibula and possibly from the lateral collateralligament between the fibula and the distal end ofthe femur. The cranial head arises partly fromthe disto-lateral surface of the patellar tendon,overlying that portion contributed mainly by M.femorotibialis externus, and is most strongly at-tached to the tip of the lateral cnemial crest.There is a weak vinculum with M. flexor perfor-atus digiti III about 1 cm above the trochlea andthe insertion is typical. Again, flamingos agreewith Cladorhynchus and differ from Ciconiiformes,in all of which this muscle has but a single head.

Mm. flexores perforati

Of the three perforated flexors, that for digitIV is the most superficial and that for digit II isthe deepest. These three muscles are fused proxi-mally and are all essentially double-headed, withthe heads diverging around the insertion of M.iliofibularis. The superficial or lateral heads of IVand II take a common origin tendinously fromthe head of the fibula. The lateral portion of IIIattaches to the belly of II. The three fused deepor medial heads take their origin fleshy from thedepths of the popliteal fossa, distal to the inser-tions of the accessory lateral crural flexor and theintermediate and fourth heads of the gastrocne-mius. The deep surface of M. flexor perforatusdigiti II fuses partly with the belly of the muchsmaller M. flexor hallucis longus. The tendon ofIV becomes thickened at the level of the trochleaand is perforated at the level of the base of thetoe; it gives rise to paired insertions at the distalends of each phalanx. The tendon of III receivesthe tendon of M. peroneus longus and continueson to receive another, but weaker, vinculum fromM. flexor perforans et perforatus digiti III about0.5 cm above the trochlea. This vinculum is alsopresent in flamingos and all ciconiiforms exceptthe Ardeidae. The tendon of III is perforated inthe middle of the basal phalanx and bifurcates toinsert on the ventral corners of the base of pha-lanx 2. The tendon of II inserts as a thickenedcapsule at the base of phalanx 1 digit II.

M. flexor hallucis longus

Although the hallux is absent, its flexor ispresent and has a partly fleshly origin medial to

the distal attachment of the iliofibular loop.There is a short tendinous attachment to the mostcaudolateral protruberance of the lateral condyleof the femur and the origin is fleshy medial tothis; the fleshy origin continues from a rapheextending distally on the deep surface of M. flexorperforatus digiti II. The belly is very short, flat,and triangular. The slender tendon runs throughthe lateral side of the tibial cartilage and througha groove in the lateral crest of the hypotarsus.About 1 cm or less distal to the hypotarsus it fuseswith the tendon of M. flexor digitorum longus.This corresponds to the type IV condition of thedeep flexor tendons, according to the terminologyof Gadow (1896). This condition is found in anumber of often unrelated birds in which thehallux is absent or greatly reduced. It occurs inflamingos but in none of the Ciconiiformes.

M. flexor digitorum longus

This is the deepest muscle on the caudal aspectof the shaft of the tibiotarsus. One head arisesfleshy from the head of the fibula and the proxi-mal third of its shaft. The other head extendsmedially on the shaft of the tibiotarsus alongsidethe origin of M. plantaris. The tendon is thedeepest one passing through the sulcus betweenthe two crests of the hypotarsus; it receives thetendon of M. flexor hallucis longus just distal tothe hypotarsus. At the distal end of the tarso-metatarsus it trifurcates and inserts on the threedigits. The significance of the deep flexor tendonsis discussed above.

M. popliteus

This small muscle arises from the head of thefibula deep to all other muscles. It runs diagonallyto attach disto-medially on a small protru-berance on the caudal face of the proximal endof the shaft of the tibiotarsus. Flamingos agreewith Cladorhynchus and differ from all Ciconi-iformes in having the insertion of M. popliteusdistal to the origin (cf. Vanden Berge, 1970).

Intrinsic Muscles of the Tarsometatarsus

Because of the extremely slender shaft and theabsence of the hallux, the muscles originating on

NUMBER 316

the tarsometatarsus in Cladorhynchus are greatlyreduced or absent (true also of flamingos),whereas most of these muscles are much betterdeveloped in all Ciconiiformes. As in the case ofthe deep flexor tendons, it would be easy to arguethat similarities in these respects between Clado-rhynchus and flamingos could be due to indepen-dent reduction of the hallux, yet it must beremembered that this has never occurred in theCiconiiformes. In some cases we have not madecomparisons when Vanden Berge (1970) has notspecifically described the condition in flamingos.

M. extensor hallucis longus

Absent in Cladorhynchus and flamingos exceptPhoenicopterus, where it is stated to be present butvestigial (Vanden Berge, 1970). It is well-devel-oped in Ciconiiformes.

Mm. extensores brevi

M. extensor brevis digiti III and M. extensorbrevis digiti IV arise from a bundle of connectivetissue just distal to the lateral cotyla on the dorsalface of the tarsometatarsus. They appear to havevery few muscle fibers associated with them. Thatfor digit IV is the most lateral but its long tendonthen runs deep to that of digit III and is closelybound to the face of the shaft of the tarsometa-tarsus; where it enters the distal foramen it iscovered by the expanded sheath of the tendon ofIII. The long tendon of M. extensor brevis digitiIII runs deep to that of M. extensor digitorumlongus; above the distal foramen it fans out intoa broad sheath covering most of trochleae III andIV and having its insertion at the bases of theirphalanges. Vanden Berge (1970) does not givemany details of these muscles in flamingos but itis evident that they are better developed in Ci-coniiformes.

M. abductor digiti II

This is a fleshy muscle on the medial aspect ofthe distal fifth of the tarsometatarsus. It insertson the base of phalanx 1 digit II.

M. flexor hallucis brevis

This tiny muscle arises as a wisp of fibersdescending from the ligamentous distal extension

of the medial crest of the hypotarsus. The belly isabout 8 mm long and the exceedingly slenderthread of a tendon extends distally on the medialmargin of the shaft of the tarsometatarsus andappears to lose itself in the belly of M. abductordigiti II. This muscle is also vestigial in flamingos,in which, however, it appears to arise fartherdistally. It is better developed in all Ciconi-iformes.

M. adductor digiti II

This arises lateral to M. flexor hallucis brevisfrom the distal end of the lateral crest of thehypotarsus and adjacent surface of the tarso-metatarsus, tapering for 13 mm before becomingtendinous. The long tendon inserts on the medialside of the base of phalanx 1 digit II. This musclealso arises proximally in flamingos and ibises,whereas in Ciconiiformes it is more distally lo-cated.

M. lumbncalis

Not detected in Cladorhynchus and quite vesti-gial in flamingos and in Ciconiiformes, exceptherons (Vanden Berge, 1970).

M. abductor digiti IV

The tendon of insertion attaches on the lateralcorner of the plantar aspect of phalanx 1 digit IVand runs the length of this short tapering muscle.The fibers are oriented obliquely from this tendonto the distal fifth of the lateral side of the plantarface of the tarsometatarsus. This is about as de-scribed for flamingos and storks and differs fromthe condition in herons (Vanden Berge, 1970).

Pterylosis

To Nitzsch (1867:132), the pterylosis of fla-mingos was "perfectly Stork-like," a statementwith which Gadow (1877) agreed and which hasbeen repeated without corroboration ever since.On the contary, we found the pterylosis of fla-mingos to be quite different from that of storks.

Nitzsch's great opus, which was published post-humously, is the original and only class-widestudy of comparative pterylosis. Unfortunately, ithas never been superceded by a more modern


and comprehensive analysis. Nitzsch first con-trasted his family Pelargi (Ciconiidae and Scopi-dae) with the Erodii (Ardeidae) and showed therather generalized pterylosis of storks to be quitedifferent from the peculiar pterylosis of herons.He then discussed the "Odontoglossae" (=Phoenicopteridae) and found the pterylosis offlamingos to be of a more usual pattern andtherefore "perfectly Stork-like" (as compared toherons, one may assume).

Mary H. Clench (in litt. to Olson, 8 December1978) states that "the only valid approach topterylography is a study of the tract patterns—how the feathers are arranged in rows, the rela-tionship of the rows to each other, the numbersand lengths of rows (to a lesser extent)—in otherwords, what one can see within a tract, not just itsgeneral outlines (and resulting apteria)." Neitherour brief inquiry into pterylosis, nor the studiesof Nitzsch for that matter, meet these require-ments. Our observations and illustrations (Figure9) are intended solely to show that there is littlereason to regard the pterylosis of flamingos asbeing storklike. We emphasize that the followingis not intended as a proper pterylographic anal-ysis.

To investigate pterylosis, we examined fluid-preserved specimens of Phoeniconaias minor, Ciconiaabdimit, and Cladorhynchus leucocephalus, from whichwe had clipped the feathers. All of these speci-mens had been partially dissected, obscuring theconfiguration of the feather tracts in certain areas,particularly the abdomen in the flamingo andstork, and the neck and ventral midline in Clado-rhynchus. Terminology follows that of Lucas andStettenheim (1972).

The great density both of contour feathers anddown in Phoeniconaias and Cladorhynchus (Figure 9)contrasts markedly with the sparse feathering andreduced down of Ciconia. In Ciconia the neck isfeathered in two lateral bands separated by wideventral and dorsal apteria; the latter is an anteriorextension of the interscapular apterium. The twocervical apteria extend two-thirds the length ofthe neck, the remainder of which appears to benearly continuously feathered all around. Phoeni-conaias and Cladorhynchus have the neck much

more densely clothed in feathers and down thanin Ciconia. There appears to be no dorsal cervicalapterium or ventral cervical apterium in Clado-rhynchus. Only the latter is present in Phoeniconaias;it is much smaller than in Ciconia and extends forless than one-fourth the length of the neck. InCladorhynchus, a small lateral cervical apteriumlies on each side of the base of the neck. Theseare absent in Phoeniconaias and Ciconia.

Ciconia has a rather wide sternal apterium; thisis narrower in Phoeniconaias and we could notdetect it in our specimen of Cladorhynchus, which,however, had previously been cut down the mid-line. The pectoral tract in Ciconia is broad andclothed with regularly spaced, sparse, weak feath-ers. Phoeniconaias and Cladorhynchus are much moredensely feathered and have several rows of closelyspaced, large, stiff-shafted feathers in the dorsalportions of the tract. Cladorhynchus differs fromPhoeniconaias in having pectoral apteria that de-limit a median tract, in which, as mentioned, thesternal apterium of our specimen was absent ordestroyed.

In Cladorhynchus and Phoeniconaias, the femoral,crural, and dorsal caudal tracts are nearly contin-uously and rather densely feathered, and severallongitudinal rows of stiffer feathers occur alongthe ventral margin of the pelvis above the crus.This differs from Ciconia, in which the dorsalcaudal tract consists of a few rows of sparsefeathers separated from the femoral tracts bylateral pelvic apteria. Ciconia also has crural ap-teria, which appear to be lacking in Cladorhynchusand Phoeniconaias. The uropygial gland is dis-tinctly tufted in Phoeniconaias and Cladorhynchusand nearly bare in Ciconia. David W. Johnston,(in litt. to Olson, 17 October 1978) found thisgland in Phoenicoparrus to be more similar to thatof Himantopus than to that of either Mycteria (Ci-coniidae) or Anas (Anatidae).

In dorsal view (Figure 9), the humeral tracts inCladorhynchus and Phoeniconaias are broad, elongatepatches of extremely densely packed and verystiff feathers; in contrast, this tract in Ciconia issmaller and consists posteriorly of a few very largefeathers and anteriorly of sparse, weak feathers.

As previously noted, there are 11 primaries in

NUMBER 316 33

Oodorhynchus

FIGURE 9.—Semidiagramatic views contrasting the pterylographic patterns of a stork (Ciconiaabdimu), flamingo {Phoemconaias minor), and the Australian Banded Stilt (Cladorhynchus leucoce-phalus). (Question marks indicate those areas in our specimens that were destroyed by previousdissections. These drawings are intended only to show basic patterns and do not accuratelyrepresent the distribution of feathers within tracts.)


storks and flamingos. Cladorhynchus has 10. Ofsecondaries (counting tertials) there are 19 inCladorhynchus, 20 in Ciconia, and 25 in Phoeniconaias.This, along with the general proportions of thewing, suggests that the entire wing in flamingoshas become elongated, which could account forthe greater number of secondaries and the addi-tion of a primary. This has coincidentally resultedin flamingos having the same number of primar-ies as storks, whereas the number of secondariesis considerably greater.

There are substantial differences in pterylosisbetween all three of the forms we examined.There is at least some similarity between Clado-rhynchus and Phoeniconaias in the density of feath-ering, particularly of the neck and pectoral andhumeral tracts. This could represent an indepen-dently evolved response to the caustic habitat inwhich each lives. We conclude only that thepterylosis of flamingos cannot be said to be stork-like.

Natal Down

The literature on natal down contains consid-erable misleading information. The subject is ofimportance here because the natal down of fla-mingos has been said to be like that of storks. Thefollowing summary is largely an outgrowth ofcorrespondence with Kenneth C. Parkes, to whomwe are indebted for supplying unpublished obser-vations.

In the young of certain birds there may appearto be two successive coats of down, i.e., two setsof downy growth emerging sequentially from thesame set of feather follicles. This is strictly trueonly of a limited number of taxa such as Procel-lariformes, Gaviiformes, and Sphenisciformes. Inmost other such birds the second coat of down isactually the very plumulaceous tips of the emerg-ing juvenal plumage.

The Anseriformes and Charadriiformes have asingle coat of natal down, whereas the Ciconi-iformes and flamingos, are thought to have two,although it is known that herons have but one(Palmer, 1962). The condition in the Scopidae

and Balaenicipitidae is unknown to us. Althoughibises (Threskiornithidae) are said to have twocoats of down (Palmer, 1962), the supposed sec-ond coat is body down that grows from differentfollicles (Parkes, in litt. to Olson, 27 November1978). Storks and flamingos have a second coatof down formed by the tips of the juvenal plu-mage. Sibley et al. (1969) gave considerable em-phasis to this character in attempting to showciconiiform affinities for flamingos.

In two fluid-preserved specimens of downychicks of Cladorhynchus, we found clear evidenceof two successive stages of down, the first of whichcould still be seen attached to the second (Figure10). Feather growth in these specimens was insuf-ficient to ascertain whether the second set ofdown was actually the tips of the juvenal plu-mage, however we assume this to be the case. Noother member of the Charadrii is known to havetwo coats of natal down. The white colorationand lack of markings in the downy young ofCladorhynchus are likewise unique in the Charadrii(page 11). In view of all the other flamingo-likeattributes of Cladorhynchus, the nature of its nataldown assumes considerable significance. The con-dition of the natal down in flamingos, therefore,is at least as similar to that of Cladorhynchus as itis to that of storks.

Oology

The eggs of flamingos are large relative to thesize of the bird and this is likewise one of thestriking features of the eggs of Cladorhynchus. Al-though Cladorhynchus leucocephalus is somewhatsmaller than Recurvirostra americana, its eggs aremuch larger (Figure 11). The weight of the egg ofC. leucocephalus taken as a percentage of bodyweight is twice that of R. americana (Table 2). Thesame is true of the flamingo Phoenicopterus ruber,versus the stork Mycteria americana; both of thesebirds are about the same size, yet the relative eggsize of the flamingo is twice that of the stork(Table 2). Naturally, because of allometric effects,the relative size of the eggs is smaller in the largerbirds (the body weight of the flamingo is ten

NUMBER 316 35

FIGURE 10.—Feather from a chick of the Australian Banded Stilt, Cladorhynchus leucocephalus,showing (a) the "two coats of down," the upper portion being true down and the lower probablybeing the plumulaceous tip of the juvenal feather, as in flamingos, and (b) further enlargedview of same to show the juncture of the first and second coat of down.


FIGURE 11.—Comparison of the pelves (top) and eggs (bot-tom) of the Australian Banded Stilt, Cladorhynchus leucoce-phalus (a), and American Avocet, Recurvirostra amencana (b),to show the disproportionately large size and light color ofthe egg of Cladorhynchus.

times that of Cladorhynchus and its egg weightobviously could not increase by a like proportion).The large egg size in Cladorhynchus and flamingosis probably correlated with the need to producelarge well-developed young that are better ableto survive in a harsh and often ephemeral habitat.

Flamingo eggs are white, unmarked, and cov-ered with a rather thick chalky coat that is absent

TABLE 2.—Relative egg weights of two recurvirostrids, aflamingo, and a stork (all weights in grams; averages of eggweights obtained from water filled egg-shells (n = 3, exceptCladorhynchus, n = 2); body weights are averages: Recurvirostra,17 females (Hamilton, 1975); Cladorhynchus, 2 unsexed(McNamara, 1976); Phoenicopterus, 3 males and 2 females(Rooth, 1965); Mycteria, 2 males (Palmer, 1962))

Species

Recurvirostra americana

Cladorhynchus leucocephalus

Phoenicopterus ntber

Mycteria americana

Body

weight

Eggweight

Egg weight

as %of

body weight

310

238

3050

3171

3044

16086

9.718.55.22.7

in storks and shorebirds. As mentioned, theground color of the eggs of Cladorhynchus is pecu-liar among the Charadrii in being white, andmarkings may occasionally be absent (Figure 12).Although McGilp and Morgan (1931:44) de-scribe the texture as "chalky," there is no powderycoating as in flamingos.

We examined the surface structure of the egg-shell of Recurvirostra, Cladorhynchus, Phoenicopterus,and Ciconia with a scanning electron microscope,first removing the powdery layer from the fla-mingo egg with a small buffing wheel. The tworecurvirostrids and the flamingo exhibit a ratheruniform granular structure that is significantlydifferent from that of Ciconia, in which the surfaceof the shell is regularly pitted with distinct pores(Figure 13), a condition long recognized as beingtypical of storks (Des Murs, 1860). We concludefrom this only that the eggs of storks bear noresemblance to those of flamingos.

Even if the distinctive attributes of the eggs ofCladorhynchus evolved independently of those ofthe Phoenicopteridae, this at least demonstratesthat a flamingo-like condition can be derivedfrom that of the Charadrii.

Parasitology

Sibley et al. (1969:162) reviewed the literatureon the parasites of flamingos but concluded onlythat this evidence was "as conflicting and difficultto interpret as is that from morphology."

NUMBER 316 37

FIGURE 12.—Eggs of the Australian Banded Stilt, Cladorhynchus leucocephalus, to show variationin markings, which may occasionally be absent. (Photograph from McGilp and Morgan, 1931.)

Concerning the cestode parasites, the remarksof Baer (1957:274) are of considerable interest:"One finds in flamingos four monotypic, endemicgenera (Amabilia, Cladogynia, Gynandrotaenia, Lep-totaenia). The last two, very specialized overall,are derived from a family of cestodes character-istic of the Charadrii" (our translation). Baer(1951, 1957) has argued that the high degree ofhost specificity of cestodes renders them useful forassessing phylogenetic relationships of hosts. Theevidence in this case is neither conflicting nordifficult to interpret, as it agrees with the hypoth-esis of charadriiform origins of flamingos.

The mallophagan ectoparasites of flamingos,however, present a different picture. Hopkins(1942:102) reported that flamingos are parasi-tized by three genera of feather lice that occurelsewhere only on the Anseriformes. He regardedthis "as complete proof that the Flamingos areAnseriformes, and even that their divergencefrom the stock that gave rise to the Anatidae tookplace at no very remote period of time." However,

K. C. Emerson (pers. comm.) has indicated to usthat the species ofAnatoecus, Anaticola, and Trinitonfound on flamingos differ from those found onthe Anatidae to a degree that would suggest theirdivergence did not occur recently.

Some authors have considered host transfer ofMallophaga to be frequent (Mayr, 1957; Strese-mann, 1959), suggesting that flamingos "haveacquired their feather lice from the Anatidaesince they have lived in the same environment. . . " (Sibley et al., 1969:162). This seems oversim-plified and were this often the case, one wouldexpect most waterbirds to have the same kinds ofMallophaga.

The evolutionary history of flamingos and theAnseriformes suggests a logical possibility fortransfer of Mallophaga that transcends merechance. Modern flamingos nest in dense coloniesin harsh, saline environments, and therefore,when breeding, are actually relatively isolatedfrom other birds. The available evidence indicatesthat the same was probably true of early flamin-


FIGURE 13.—Scanning electron micrographs of eggshell surface: a, stork, Ciconia nigra; b,flamingo, Phoenuopterus ruber; c, Australian Banded Stilt, Cladorhynchus leucocephalus; d, avocet,Recurvirostra americana. Note the distinctly pitted surface in the stork as opposed to the unpitted,granular surface in the others. (Scale = 1 mm. Photographs by Mary Jacque Mann.)

gos (page 48). As represented by Presbyornis, theancestors of the Anseriformes were likewise ex-tremely colonial shorebirds that lived in the samekinds of habitat (Olson and Feduccia, in press).For at least part of their history, flamingos were

probably contemporaneous with Presbyornis. Op-portunity for transfer of Mallophaga from outsidesources would presumably have been reduced dueto the lack of other species of birds adapted tosuch a harsh environment. On the other hand,

NUMBER 316

one may readily imagine that the often ephemeralnature of suitable breeding sites might at timeshave caused colonies of flamingos and Presbyornisto be situated adjacent to one another, or theymay even have bred in mixed colonies. Denselypacked, mixed flocks of flightless downy youngcould have resulted from such a juxtaposition,and would have provided ample opportunity fortransfer of ectoparasites. Thus, the similarity ofanseriform and phoenicopterid Mallophaga is notnecessarily due to a close relationship of the hosts,but to the fact that the ancestors of these twogroups shared similar specialized habits and hab-itat and were in intimate contact with each otheras a consequence.

Specimens of Mallophaga from Cladorhynchuswere kindly forwarded to us by Shane Parker ofthe South Australian Museum. They were iden-tified as belonging to the genera Actornithophilus,Austromenopon, and Quadraceps, which are charac-teristic of the Charadriiformes (K. C. Emerson,pers. comm.). Thus, the Mallophaga of flamingosmust have been acquired subsequent to theirdivergence from the Recurvirostridae.

A charadriiform origin of flamingos provides aconcordant explanation for the parasites theyharbor, whereas parasitology provides no evi-dence in favor of a relationship between flamingosand the Ciconiiformes.

Biochemistry

Biochemical information from several sourceshas been used in attempts to determine the rela-tionships of flamingos. Mainardi (1962) investi-gated red-cell antigens; Sibley (1960) used paperelectrophoresis to compare egg-white proteins;and Sibley et al. (1969) studied starch-gel electro-phoresis of egg-white proteins and hemoglobins,tryptic peptides of hemoglobins, and thin-layerelectrophoresis of tryptic peptides of ovalbumin.Sibley and Ahlquist (1972) reviewed the egg-white protein evidence for non-passerine birds,including flamingos. These investigations ante-dated the use of the isoelectric focussing technique(see Sibley and Frelin, 1972), which results inmore detailed patterns somewhat more suscepti-

ble to evaluation by the uninitiated. Brush (1979)has recently shown that there were serious pro-cedural difficulties inherent in Sibley's electro-phoretic methodology that render dubious theresults obtained.

Regardless, these biochemical investigationsare unfortunately uninformative for interpretingthe relationships of flamingos because all com-parisons were conducted with the traditional ideain mind that flamingos must be related either tostorks or to ducks. Sibley et al. (1969) concludedthat the biochemical evidence showed flamingosto have more in common with Ciconiiformes thanAnseriformes, but in the absence of any meaning-ful comparisons with Charadriiformes, particu-larly the Recurvirostridae, such a conclusion haslittle significance. As long as biochemistry is usedin avian systematics merely to legitimize precon-ceptions based on the incomplete anatomicalstudies of the nineteenth century, its usefulness asa taxonomic tool is bound to remain severelylimited.

Osteology

Feduccia (1976) discussed and illustrated osteo-logical features of flamingos that indicate rela-tionships to the Charadriiformes rather than Ci-coniiformes. These observations are augmentedbelow. Because even with the removal of thePhoenicopteridae from the Ciconiiformes, the or-der is still probably hopelessly unnatural (Olson,1979), and because storks in particular have mostoften been mentioned as relatives of flamingos,we have confined our comparisons to the Cicon-iidae. Terminology generally follows Howard(1929).

SKULL

Despite the great modification of the skull inflamingos, the architecture of the palate is decid-edly not like that of a stork (Figure 14). Thepalatines of flamingos are longer and narrowerand the pterygoids shorter and more expandedanteriorly than those of storks, thus resemblingshorebirds. The choanal aperture extends poste-riorly to the palatine-pterygoid articulation inflamingos and shorebirds, whereas it ends consid-


erably anterior to this in storks. The posteriormargin of the palatine curves gently into thearticulation with the pterygoid in flamingos andshorebirds, whereas in storks it is sharply truncateand the palatines become much narrower beforearticulating with the pterygoid (Figure 14). Thefact that the palatines in flamingos have fusedalong the midline ("desmognathy"), as in storks,seems of relatively minor importance comparedwith the otherwise great differences in the palatesof these birds.

In his studies of the middle ear region of theskull, Saiff (1978) found that the Phoenicopteri-dae differed significantly from other families of"Ciconiiformes," though he made no comparisonswith the Charadrii.

A character in the skull of flamingos that isprobably of greater taxonomic significance thanpreviously thought is the presence of distinctpaired occipital fontanelles above the foramenmagnum. These are characteristic of most Char-adrii and certain Alcidae; they are present inmost Anseriformes, which were in turn derivedfrom the Charadrii formes (Olson and Feduccia,in press). In the Gruiformes, occipital fontanellesoccur solely in the closely related Aramidae andGruidae, which may be among the more derivedmembers of the order. Elsewhere, these aperturesare found only in the ibises, Threskiornithidae,which are probably a transitional group between

FIGURE 14.—Semidiagramatic ventral view of palate:a, stork, Cicoma episcopus; b, flamingo, Phoemcopterus ntber;c, recurvirostrid, Cladorhynchus Uucocephalus. The flamingo, aswould be expected from the unique feeding apparatus,differs from either of the others, but is less similar to thestork. Note particularly the distinctive constriction of thepalatines (pi) before they contact the pterygoids (pi) in thestork.

Gruiformes and Charadriiformes (Olson, 1979).The presence of occipital fontanelles in flamingosis best interpreted as a derived character sharedwith the Charadrii.

POSTCRANIAL SKELETON

In the following characters of the postcranialskeleton, flamingos agree with Recurvirostridaeand differ from the Ciconiidae.

CORACOID.—(1) Shaft much shorter andstouter; (2) procoracoid larger, projecting farthermedially, perforate (except in Cladorhynchus); (3)acrocoracoid narrower and much more elevatedabove the glenoid facet; (4) furcular facet a dis-tinct, flat, rectangular surface; (5) sternal margincurved, not straight; (6) sternal facet muchlonger, not confined to medial half of sternal end;(7) sterno-coracoidal process deeper and moretruncate. In most shorebirds, the brachial tuber-osity and furcular facet are more extensive thanin flamingos (Figure 15); this area has been notedby Strauch (1978:309, fig. 21) to be variable inthe Charadriiformes. Its reduction in flamingosmay well be correlated with the changes in theproximal end of the humerus and the associatedmodification of the shoulder musculature.

SCAPULA.—(1) Acromion long and pointed, notsquare and blunt; (2) shaft narrower proximally.

B

FIGURE 15.—Ventral view of left coracoid: a, stork, Cicomaabdimu; b, flamingo, Phoemcopterus ruber; c, recurvirostrid,Recurvirostra americana.

NUMBER 316 41

FURCULA.—(1) Narrow, U-shaped, with clavi-cles parallel, not diverging dorsally; (2) scapulartuberosity longer, narrower, and more pointed;(3) hypocleidium very small, not approachingcarina (in storks the hypocleidium is very largeand articulates by a large facet at the anteriorend of the sternal carina; see Figure 16c).

STERNUM.—(1) Longer and narrower; (2) car-ina does not project anteriorly beyond the man-ubrium; (3) manubrium a deep blade, not asimple peg-like process (Figure 16; see also Fed-uccia, 1976, fig. 6).

HUMERUS.—(1) Bicipital crest long, taperinggently into shaft, not jutting out abruptly; (2)scar for attachment of M. pectoralis not lyingmainly distal to deltoid crest; (3) entepicondylarprominence poorly developed; (4) olecranal fossanarrow and distinct; (5) scar for M. latissimusdorsi posterioris lying ventral to midline of shaft,not dorsal as in storks; (6) distal end relatively

A A

FIGURE 16.—Outline of furcula in anterior view (left col-umn) and left lateral view of sternum, coracoid, and furcula(right column): a, recurvirostrid, Cladorhynchus leucocephalus\b, flamingo, Phoenicopterus ruber; c, stork, Ciconia episcopus. Notethe large hypocleidium of the furcula in the stork and itssolid articulation with the apex of the carina.

narrower.CARPOMETACARPUS.—(1) Metacarpal I more

proximally directed, less vertical; (2) ventro-medial rim of carpal trochlea less prominent;(3) proximal metacarpal symphysis much shorter;(4) metacarpal III straight, not bowed ventrally;(5) distal metacarpal symphysis relatively longer.

PELVIS.—(1) Anterior iliac shield not markedlyexpanded anteriorly; (2) postacetabular ilium notfused to synsacrum; (3) obturator foramen rela-tively smaller.

FEMUR.—(1) Proportionately shorter andstouter; (2) anterior and proximal extent of tro-chanter greater; (3) configuration of muscle scarson proximo-lateral surface similar in flamingosand recurvirostrids and quite distinct from thatof storks; (4) anterior rim of internal condyleextends farther anteriorly and proximally; (5)internal condyle in medial view more pointed;(6) rotular groove deeper. It should be noted thatthe femur in Cladorhynchus is stouter and moreflamingo-like than in Recurvirostra or Himantopus.

TIBIOTARSUS.—(1) Inner cnemial crest muchmore extensive anteriorly, appearing squared inmedial view (see Feduccia, 1976, fig. 2); (2) innercnemial crest extends proximally far above levelof outer cnemial crest, whereas in storks the crestsare at about the same level (Figure 17); (3) notch

FIGURE 17.—Proximal end of right tibiotarsus in anteriorview: a, stork, Ciconia abdimn; b, flamingo. Phoemconaias minor;c, recurvirostrid, Cladorhynchus leucocephalus. Note the muchreduced inner cnemial crest (u) in the stork.


present in distal rim of internal condyle; (4)anterior intercondylar fossa expanded proxi-mally, not almost round as in storks (Figure 18);(5) distal end more expanded, particularly me-dially, with the internal ligamental prominencelarger; (6) internal condyle in medial view longerand narrower (particularly true of Cladorhynchus,which is most similar to the extinct genus Palae-lodus in this respect); (7) posterior articular surfacedoes not extend as far proximally.

TARSOMETATARSUS.—(1) Intercotylar knobmuch broader and more rounded, not notchedlaterally (Figure 24); (2) shaft much more lat-erally compressed; (3) distal foramen smaller; (4)inner trochlea much higher and more posteriorlyrotated (Figure 24; also Feduccia 1976, figs. 3 and4); (5) scar for hallux reduced or absent.

PHALANGES.—(1) All phalanges, including un-guals, much shorter, wider, and deeper, whereasin storks the phalanges, including the well-devel-oped hallux, are long and slender.

The vertebral count and the structure of thecervical vertebrae of flamingos differ from bothshorebirds and storks, but these modifications arealmost certainly correlated with the feeding ad-aptations. The first three or four anterior thoracicvertebrae are fused into a notorium in modernflamingos, a condition absent in shorebirds andstorks but which evidently has evolved in flamin-gos since the early Eocene (page 57). The firstthoracic vertebra in flamingos bears a large he-mapophysis, which is true of the Recurvirostridaebut not storks, in which this process is absent.

There is a superficial resemblance between fla-mingos and storks in the distal end of the tibi-otarsus, in both of which there is a deep, distinctanterior intercondylar fossa and a tubercle on thetendinal bridge. The shape of this fossa, however,is nearly round in storks, whereas it is expandedproximally in flamingos, particularly on the lat-eral side, where it incises the external condyle(Figure 18). Furthermore, the tubercle on thetendinal bridge of flamingos is nearly continuouswith another tubercle directly proximal to it,

FIGURE 18.—Distal end of right tibiotarsus: a, stork, Ciconiaabdimii; b, flamingo, Phoeniconaias minor. Although in bothfamilies there is a deep intercondylar fossa, its shape, as wellas that of the rest of the element and the placement of thetubercles is entirely different.

whereas in storks the second tubercle is separateand lies entirely lateral to the first.

The sterna of flamingos and storks have buttwo posterior notches. Most Charadrii typicallyhave a four-notched sternum but only twonotches occur in the Jacanidae, Rostratulidae,Thinocoridae, and some individuals of Burhini-dae. It is likely that the loss of the medial notchesin flamingos is a secondary development.

The proportions of the hindlimb elements offlamingos are quite different from those of storksand are more similar to those in the Recurviros-tridae. As an example, the absolute length of thefemur in Phoenicopterns ruber is less than that inCiconia nigra, whereas the tarsometatarsus of theflamingo is much longer than that of the stork.

Because of the number of unique non-skeletalfeatures shared by flamingos and Cladorhynchus, itshould be noted that the osteology of Cladorhynchusdoes not differ appreciably from that of the othermembers of the Recurvirostridae (Strauch, 1978;this study) and it lacks the skeletal peculiaritiesdiagnostic of the Phoenicopteridae.

NUMBER 316

Significant osteological differences between fla-mingos and shorebirds are seen, of course, in themodification of the feeding apparatus and cervi-cal vertebrae. These specializations are unique toflamingos and hence cannot show relationships,but as discussed later, they are more easily derivedfrom the charadriiform condition than that inCiconiiformes, particularly storks.

Apart from the feeding specializations, thegreatest differences between flamingos and typi-cal Charadrii are seen in the humerus and theshoulder joint. Many of these differences areprobably due to the pneumatization of the prox-imal end of the humerus. This is almost certainlya size-related phenomenon, because in almost allbirds the size of extant flamingos the humerus ispneumatized. Comparison of the fossil frigatebirdLimnofregata (Pelcaniformes: Fregatidae) with Re-cent species in the same family, shows the greatextent to which the humerus may be modifiedwith increasing pneumatization (Olson, 1977:23).Adaptations paralleling those of flamingos, suchas the reduction of the ectepicondylar spur of thehumerus and the brachial tuberosity of the cora-coid, are also found in the charadriiform speciesScolopax minor, whereas they are absent in othermembers of the genus (Olson, 1976). The hu-merus is not pnemumatic in 5. minor, however,and the modifications of the wing probablyevolved in connection with sound production dur-ing mating displays. Nevertheless, this shows thatadaptations similar to some of those in the wingand shoulder of flamingos have evolved within asingle genus of Charadriiformes.

With the exceptions noted above, the osteologyof the Phoenicopteridae conforms very well withthat of the Charadrii. On the other hand, nothingin their skeletal structure can be used to supporta relationship with the Ciconiiformes.

Paleontology

The fossil record of flamingos has been inven-toried by Brodkorb (1963a) and briefly discussedby Sibley et al. (1969). We shall review the im-portance (or lack thereof) of the various fossils

attributed to flamingos, adding a description ofa significant new genus from the middle Eoceneof Wyoming.

A few fragments of Cretaceous bones have beensaid, on very insufficient grounds, to belong tobirds related to flamingos. The earliest of these,and one of the earliest birds known, is Gallornisstraeleni, described by Lambrecht (1931) as a newgenus of Anatidae from the Lower Cretaceous(Neocomian) of France. This was based on "theproximal portion of a femur and a scrap of hu-merus, the latter of no comparative value" (Brod-korb, 1963b:63). Using only Lambrecht's illustra-tions for comparison, Brodkorb (1963b) decidedthat Gallornis was definitely not anseriform andhe referred it to the neighborhood of flamingos,placing it in the extinct family Scaniornithidae.We regard it as impossible to determine the affin-ities of Mesozoic birds with such undiagnosticfragments; in fact, it is difficult enough and usu-ally impossible to determine the nature even ofearly Tertiary birds from such material (see Ol-son, 1977:31).

The same applies, though even more emphat-ically, to a single vertebra from the late Creta-ceous (Campanian) of Sweden that Lambrecht(1933:335) optimistically described as a flamingounder the name Parascaniomis stensioi.

The avifauna of the late Cretaceous (Maes-trichtian) Lance Formation, analyzed by Brod-korb (1963b), contains what was described as anew genus and species of flamingo, Torotix cle-mensi, based on the distal end of a humerus (nota "head" as stated by Sibley et al., 1969:158).Brodkorb placed this in a new family Torotigidae,to which he arbitrarily assigned Gallornis andParascaniomis.

Brodkorb (1963b) identified two other familiesof birds from the Lance Formation: the Cimolop-terygidae, an extinct family of Charadriiformesknown only from coracoids, and the Lonchody-tidae, a family consisting of two species supposedto belong to the Gaviiformes, and known from afairly diagonostic distal end of a tarsometatarsus,as well as a fragment of carpometacarpus that weregard as indeterminable.


After preliminary examination of much of thismaterial, we consider it quite possible that all ofthe birds so far known from the Lance Formationmay be referable to the Charadriiformes. As evi-denced by the following, the holotype tarsome-tatarsus of Lonchodytes estesi is more similar to theCharadrii than to the Gaviidae: (1) lesser lateralcompression of the shaft; (2) greater elevation ofthe outer trochlea relative to the middle trochlea;(3) lesser retraction of the inner trochlea. ProfessorBrodkorb (pers. comm.) now concurs with us atleast that L. estesi is not gaviiform; we consider itto be closest to the Charadriiformes.

The holotype of Torotix clemensi lacks the largeectepicondylar spur typical of most Charadri-iformes and thus resembles both the Phoenicop-teridae and the Presbyornithidae, a charadriiformfamily near the ancestry of ducks. This humeruscould well belong to one of the species of Cimolop-teryx, known only from coracoids, while the typeof Lonchodytes estesi may be referable to one of thelarger species of Cimolopterygidae. Without bet-ter material it probably is not possible to deter-mine whether the Lance avifauna contains fla-mingo-like or presbyornithid-like Charadri-iformes, or neither.

Scaniornis lundgreni Dames (1890), from thelower Paleocene of Sweden, is based on a hu-merus, scapula, and coracoid. Dames' rather in-different illustration shows these to be poorlypreserved, mostly as impressions in a slab. Al-though nothing in his illustration rules out thepossibility that the species was a flamingo, neitheris there anything diagnostic to be seen. Until thespecimen is reprepared and restudied, its affinitiesmust be regarded as uncertain. The family Scan-iornithidae proposed by Lambrecht (1933) is un-justified and undiagnosable on the basis of avail-able information.

The genus Telmabates from the lower Eocene ofPatagonia, was regarded by Howard (1955) asforming a new family, Telmabatidae, havingsome affinities with flamingos. Feduccia andMcGrew (1974) showed that Telmabates and Tel-mabatidae are junior synonyms of Presbyornis andPresbyornithidae, respectively, known from con-

temporaneous deposits in the western UnitedStates. Most, if not all, of the flamingo-like char-acters of these birds are actually charadriiform innature. We treat these charadriiform-anseriformmosaics in another paper dealing with the originof the Anseriformes (Olson and Feduccia, inpress).

The genus Agnopterus was founded by Milne-Edwards (1868) for his species A. launllardi, basedon the distal portion of a tibiotarsus lacking mostof the diagnostic features. His illustration doesnot serve to confirm or deny his placement of thegenus with the flamingos. Lambrecht's (1933)elevation of it to family rank cannot be justifiedon the basis of such a specimen.

Agnopterus turgaiensis Tugarinov (1940), fromthe upper Oligocene of Kazakhstan, is based ona well preserved distal end of a tibiotarsus thatseems, on the basis of the original illustrations, tobe correctly called a flamingo. That it properlybelongs to the same genus as A. laurillardi may bequestioned, particularly as it is geologically muchyounger and also nearly contemporaneous withPalaelodus and the earliest known form of Phoeni-copterus, with both of which it needs comparison.

Lydekker (1891) described a coracoid from theupper Eocene of England, referring it with aquery to Milne-Edwards' genus Agnopterus as A.hantoniensis. Harrison and Walker (1976b) referreda humerus to this species, and made it the type ofa new genus, Headonornis, of Telmabatidae, eitheroverlooking or ignoring the fact that Feducciaand McGrew (1974) had synonymized this familywith the Presbyornithidae. Harrison andWalker's diagnosis of Headonornis does not diag-nose the genus, nor do they explain why the taxonwas referred to the Telmabatidae or how it differsfrom Telmabates (= Presbyornis). Their illustrationof the humerus shows it to be quite different fromthat of flamingos, but the same may not be trueof the type coracoid. Until a proper comparativestudy is made, the significance of these specimensto the evolution of either flamingos or the Pres-byornithidae cannot be determined.

The nomenclatural and taxonomic status ofElorms littoralis and E. grandis, from the lower

NUMBER 316 45

Oliogocene of France, was reviewed by Olson(1978). The specimens on which these names werebased have been lost and the taxa are now knownonly from Milne-Edwards' (1867-1871) illustra-tions and descriptions. These do not precludephoenicopterid affinities for these taxa, but be-yond this little else can be said. Elornis anglicusLydekker, from the upper Eocene of England,was synonymized by Harrison and Walker(1976b) with Actiornis anglicus Lydekker, formerlyof the Phalacrocoracidae, which they then trans-ferred to the Threskiornithidae. Any supposedphoenicopterid affinities of this taxon are thusdoubtful.

Tiliornis senex Ameghino (1899), based on anunillustrated coracoid from the Lower Oligoceneof Argentina and assigned to the Phoenicopteri-dae, is a meaningless name until the type islocated and restudied.

Hitherto, the earliest unequivocal flamingosbased on sufficient material to have some evolu-tionary significance have come mainly from lateOligocene or early Miocene (Aquitanian) depositsin France. These consist of an extinct species in amodern genus, Phoenicopterus croizeti, and severalspecies in the extinct genus Palaelodus. Six speciesof the latter have been named, mostly distin-guished on size. Although known from abundantmaterial, no modern comparative study has everbeen made of Palaelodus and a revision would nodoubt result in the recognition of fewer species.One species, Palaelodus steinheimensis Fraas (1840),from the late Miocene of Germany, is misidenti-fied to family, being based on what appears to bethe distal end of a tibiotarsus of a large anseriform(Olson, pers. obs.). Species larger than any ofthose of Palaelodus have been named in the genusMegapaloelodus, from the early Miocene and earlyPliocene of North America (A. H. Miller, 1944;Brodkorb, 1961; Howard, 1971). We have exam-ined additional specimens of Megapaloelodus fromthe middle Miocene (Clarendonian) of Texas(AMNH 9109) and the Aquitanian of France(USNM 18497). This genus is very similar to, andmay not be separable from, Palaelodus.

Palaelodus and Megapaloelodus are usually sepa-

rated from the Phoenicopteridae as a distinctfamily, Palaelodidae. Most skeletal elements ofPalaelodus agree well with the Phoenicopteridae.The cervical vertebrae are greatly elongated as inmodern flamingos, and the humerus is decidedlyflamingo-like as well. The specimen of bill previ-ously assigned to Palaelodus by Milne-Edwards(1867-71, vol. 2), was referred by Cracraft (1973:87-88) to the gruiform species Probalaerica proble-matical the type of which is also a bill. In ouropinion, the true identity of these Aquitanianrostra, as well as the skull that forms the type ofthe supposed ibis Ibidopodia palustris Milne-Ed-wards, remains problematical. The latter needscomparison with the cranium illustrated byRothausen (1966) that was found in associationwith postcranial elements of Palaelodus. In anycase, the morphology of the bill of Palaelodus isnot certainly known.

The most obvious difference between Palaelodusand modern flamingos is the shorter tibiotarsusand the shorter and more laterally compressedtarsometatarsus of the former. These modifica-tions suggest that the species of Palaelodus mayhave occupied more of a duck-like swimmingniche than do typical flamingos. The differencesin the proportions of the hindlimb between typi-cal flamingos and Palaelodus are hardly greaterthan seen between the genera of recent Recurvi-rostridae (Figure 19). Although the osteology andsystematics of Palaelodus and Megapaloelodus re-quire further study, we believe that they areclearly flamingos and should be included in thePhoenicopteridae.

Contemporaneous with Palaelodus in the Aqui-tanian of France is a typical flamingo, Phoenicop-terus croizeti Gervais, known from a number ofspecimens, including cranial material. Harrisonand Walker (1976a) have recently proposed anew genus, Gervaisia, with P. croizeti (which theyconsistently misspell "croiseti") as the type. Theybased their diagnosis solely on fragments of billthat they happened to have on hand, evidentlywithout consulting any of the more diagnosticmaterial that is available in various other mu-seums. As happens too frequently with Harrison


FIGURE 19.—Right tarsometatarsi in medial view: a, Phoent-conaias minor, Phoenicopteridae; b, Palaelodus sp., Phoenicop-teridae; c, Himantopus mexicanus, Recurvirostridae; d, Clador-hynchus leucocephalus, Recurvirostridae. Note that the differ-ence in proportions between the two genera of flamingos isabout the same as that between the two recurvirostrids.(Scale in cm.)

and Walker, their name was preoccupied, in thiscase thrice over. Kashin (1978) noted that Ger-vaisia Bonaparte, 1854 (Aves), Gervaisia Waga,1858 (Myriapoda), and Gervaisia Robineau-Des-

FIGURE 20.—Rostrum of the Miocene flamingo Phoenicopteruscroizeti, drawn from a cast of a specimen in the Naturhisto-risches Museum, Basel. The curvature is less than in adultsof modern species of Phoenicopterus but is similar to that seenin their earlier developmental stages (see Figure 38c). (Scale= 3.5 cm.)

voidy, 1863 (Insecta), all antedate Gervaisia Har-rison and Walker, 1976a, and he therefore pro-posed the name Harrisonavis as a substitute. Thedistinguishing feature of the supposed new genuswas said by Harrison and Walker to be the lessbent bill with a broader tip. Figure 20 illustratesa nearly complete rostrum of P. croizeti and showsthe less bent nature of the bill, the shallower keel,and the broader tip, all of which occur in thedevelopmental stages of modern Phoenicopterus (seeFigure 38t). Thus P. croizeti does not possess char-acters of generic significance as compared to ex-tant taxa, the morphology of the bill being pre-cisely what would be expected in an earlier, some-what more primitive form of Phoenicopterus. Weregard Gervaisia Harrison and Walker, 1976a, andHarrisonavis Kashin, 1978, as junior synonyms ofPhoenicopterus Linnaeus, 1758.

From Australia, A. H. Miller (1963) describedtwo flamingos of roughly the same age as Palae-lodus and Phoenicopterus croizeti. One of these wasreferred to a modern genus as Phoenicopterus novae-hollandiae, and the other was described in a newgenus as Phoeniconotius eyrensis. The latter wasbased on the distal end of a tarsometatarsus andtwo phalanges and was characterized as beingmore massive and less specialized for swimmingthan modern flamingos. All fossil flamingos sub-sequent to the early Miocene (except a few spec-imens of Megapaloelodus) have been referred tomodern genera.

NUMBER 316 47

The fossil record of flamingos hitherto knowncan be summarized as follows. None of the Me-sozoic and early Tertiary fossils said to representflamingos have been identified with certainty andfor the present they may be discounted. In thelate Oligocene or early Miocene of Europe occursa species in the modern genus Phoenicopterus hav-ing a somewhat more primitive bill structure thanrecent forms. Contemporaneously, there was aradiation of shorter-legged flamingos (Palaelodusand Megapaloelodus), apparently more specializedfor swimming. At least one species of these per-sisted until the early Pliocene. None of thesefossils gives any better clue to the ancestry offlamingos than do the modern forms themselves.

To the foregoing may now be added a newgenus and species of middle Eocene flamingo.This is the earliest certain member of the Phoen-icopteridae and is based on a meaningful repre-sentation of the skeleton. It is particularly signifi-cant in showing an even closer approach to theCharadriiformes than do recent flamingos.

A New Stilt-like Flamingofrom the Middle Eocene of Wyoming

Among the many fossil birds that had beensubmitted to the late Alexander Wetmore forstudy and which had remained unidentified, wasa small series of bones from a single locality inthe middle Eocene Bridger Formation of Wyo-ming. These were collected in 1946 and 1947 byC. Lewis Gazin, Franklin L. Pearce, and G. F.Sternberg. According to Pearce's recollection, thefossil birds at this locality all came from the samelevel and from a relatively small area, althoughno specimens were found in association.

Other than a single, undiagnostic distal end ofa femur of a yet unidentified larger bird (USNM244323), apparently found somewhat apart fromthe other specimens, all of the avian fossils from

FIGURE 21.—Bones of very young individuals referred toJuncitarsus gracillimus: a, two portions of tibiotarsi: b, distalend of tarsometatarsus. These probably indicate the presenceof a breeding colony. (Scale = 1 cm.)


this site appear to be from a single species. Atleast three adults or sub-adults are represented,along with several bones, possibly from a singlejuvenile individual (Figure 21) that was almostcertainly too young to fly at the time of death,thus suggesting the possibility of a breeding ag-gregation. The individuals vary in size (Figure22), perhaps indicating that the species was sex-ually dimorphic.

This species was intermediate in size betweenthe larger extant Charadrii and the smallest ex-tant flamingos. Its osteology shows a mosaic ofcharacters of both groups, but its presumablyderived features are those of flamingos, therebyproviding the earliest definite record of thePhoenicopteridae. In the following descriptionswe have included some comparisons with Pres-byornis (Presbyornithidae) in order to facilitateidentification of specimens in the event that ad-ditional material comes to light.

Order CHARADRII FORMES Huxley, 1867

Suborder CHARADRII Huxley, 1867

Family PHOENICOPTERIDAE Bonaparte, 1831

Juncitarsus, new genus

TYPE-SPECIES.—Juncitarsus gracillimus, new spe-cies.

DIAGNOSIS.—Referable to the Phoenicopteridaeby having the humerus with a single pneumaticforamen, a distinct oval scar for M. scapulo-humeralis cranialis, and the ectepicondylar spurlacking; cervical vertebrae modified similarly tomodern Phoenicopteridae; intercotylar knob oftarsometatarsus high and broad. Tarsometatarsusextremely long and slender, most similar to Hi-mantopus (Recurvirostridae); proximal end of hy-potarsus level with cotylae; hypotarsus complex,with closed canals; scar for hallux present; distalforamen large and elongate; inner trochlea mark-edly retracted posteriorly, with distinct wing; pos-teroproximal border of middle trochlea truncate.Attachment for anterior articular ligament of

FIGURE 22.—Anterior view of proximal ends of three tarso-metatarsi of Juncitarsus gracillimus showing variation in size.(Natural size.)

humerus sloping steeply. Thoracic vertebrae notfused into a notorium.

ETYMOLOGY.—Latin juncus (a reed or rush) plustarsus, latinized from Greek tarsos (ankle), in ref-erence to the extremely slender tarsometatarsus;the generic name is masculine.

Juncitarsus gracillimus, new species

FIGURES 23,24b,25-27a, 28-31

HOLOTYPE.—Left tarsometatarsus lacking onlya portion of the anterior part of the proximalend; hypotarsus slightly abraded. Vertebrate pa-leontological collections of the National Museumof Natural History, USNM 244318 (Figures 23,25, 26/>,c); collected in 1947 by C. Lewis Gazin(original number 140-47).

TYPE-LOCALITY.—Wyoming, Sweet waterCounty, "low buttes to extreme NE of TwinButtes and within Bridger" (C. L. Gazin, 1946field catalog), in "yellow layer near base of buttes,beneath a hard grayish green sandstone" (C. L.

NUMBER 316 49

B D

FIGURE 23.—Holotype, left tarsometatarsus of Juneitarsus gra-cillimus (USNM 244318): a, anterior view; b, posterior view;c, lateral view; d, medial view. (Natural size.)

Gazin, 1947 field catalog). Although Gazin (pers.comm.) indicated that these buttes were thoseindicated in the SE corner of section 33, T15N,R109W (Buckboard Quadrangle, U.S. GeologicalSurvey 15 minute series topographic map, 1966),successive searches of this locality in 1976, 1977,and 1978 by Arnold D. Lewis, R. J. Emry, andOlson, disclosed no vertebrate fossils other thana few turtles. We later found that Gazin's 1946field catalog refers to the locality as "near 1/4section 22/27 corner," which would place itslightly farther to the north (McKinnon JunctionQuadrangle); outcrops fitting the above descrip-tions were located here in 1979 by Olson andLewis, but no additional bird remains were found.

HORIZON.—Middle Eocene, Bridger Forma-tion. The site is very low in the Bridger, hencethe specimens are earliest middle Eocene in age.

ETYMOLOGY.—Latin gracillimus (very slender).PARATYPES.—Proximal two-fifths of right tar-

sometatarsus lacking hypotarsus and not com-pletely ossified, USNM 244319 (Figure 26a);proximal two-thirds of right tarsometatarsus withassociated fragments of inner and outer trochleae,USNM 244322; distal end of right tarsometatar-sus of a very young individual, USNM 244327;left tibiotarsus of very young individual lackingtarsals and with proximal end unossified, USNM244334; distal end of right tibiotarsus of juvenilelacking tarsals, USNM 244337; abraded distalend of right femur, USNM 244320; pedal pha-lanx 1 of right digit II, USNM 244329; distaltwo-thirds of pedal phalanx, USNM 244335;worn proximal half of left humerus, USNM244325 (Figure 27a); distal end of left humerus,USNM 244330 (Figure 286); proximal fourth ofleft ulna, USNM 244326 (Figure 28a1); portion ofa shaft of an ulna, USNM 244324; left radiale,USNM 244333 (Figure 28c); anterior two-thirdsof left scapula of a juvenile, USNM 244328 (Fig-ure 28a); unfused frontal bone of a juvenile,USNM 244336 (Figure 31); anterior half of cer-vical vertebra similar to the 6th cervical of Phoen-icopterus, USNM 244338 (Figure 30a,b)\ cervicalvertebra similar to the 16th of Phoeni cop terns,USNM 244321 (Figure 29); thoracic vertebrae


possibly equivalent to the 17th and 18th verte-brae of Recurvirostra, USNM 244332 and 244331(Figure 30c). All from the same locality and ho-rizon as the holotype; C. L. Gazin field numbers40-46, 139-47, 141-47, and 142-47.

MEASUREMENTS (in mm).—Holotype: Totallength 182, depth through hypotarsus about 12,greatest proximal width about 10.5, least widthof shaft 3.8, least depth of shaft 4.0, width of shaftat midpoint 3.9, depth of shaft at midpoint 5.1,greatest width through trochleae 10.6, greatestdepth through trochleae 11.9, width of middletrochlea 3.9, depth of middle trochlea 6.0, depththrough inner trochlea 6.3, depth through outertrochlea 6.7.

Paratypes: Tarsometatarsus USNM 244322,proximal width 9.6; ulna USNM 244326, proxi-mal depth 9.7, proximal width 6.7, distance fromtip of olecranon to distal extent of attachment ofanterior articular ligament 8.3; humerus USNM244325, distance from head to distal end of pec-toralis attachment 24.5, length of scar for M.scapulohumeralis cranialis 4.0; humerus USNM244330, distal width 13.2, depth through externalcondyle 7.4, diagonal length of external condyle6.1 (humerus length estimated at 100); radialeUSNM 244333, greatest diameter 7.6; pedal pha-lanx 1 digit II USNM 244329, length 19.1; tho-racic vertebra USNM 244331, length of centrum10.3; cervical vertebra USNM 244321, lengththrough centrum 14.4.

DIAGNOSIS.—As for the genus. Smaller thanany other known species of Phoenicopteridae;larger than any member of the Recurvirostridae.

DESCRIPTION.—Tarsometatarsus: The extraor-dinary tarsometatarsus of this species is somewhatsmaller than that of the smallest living flamingo,Phoemconaias minor, but is much more slender (Fig-ure 24), its proportions being approached amongrecent birds only by the stilts {Himantopus; Recur-virostridae). This element, in fact, is so nearlyidentical in morphological details to that of Hi-mantopus, that without the associated material itwould probably have been referred to the Recur-virostridae, and within that family would havebeen regarded as a close relative of Himantopus.Only in the following details can the tarsometa-

tarsus of Juncitarsus be distinguished from that ofHimantopus: larger size; distal foramen larger andmore elongate (Figure 25d,e); middle trochleawithout the slight lateral deflection of Himantopusbut identical to Recurvirostra, distinct, small scarfor hallux (Figure 2be) (hallux lacking in Himan-topus); wing of outer trochlea narrower andslightly more produced posteriorly (Figure 25/);attachment for M. tibialis cranialis more distinctand elongate (Figure 25a); intercotylar knobmarkedly higher and broader (Figure 26a). Thelast feature is the only salient character separatingthe tarsometatarsus of Juncitarsus from that of theRecurvirostridae and the only one that shows acloser approach to a condition in modern fla-mingos.

The tarsometatarsus of Juncitarsus can be distin-guished from that of modern flamingos by thefollowing features: outer trochlea narrower, withwing in lateral view more perpendicular to theshaft, not angling distally (Figure 25/); distalforamen larger and more elongate; proximal bor-der of posterior face of middle trochlea truncate(Figure 25<?), not tapering proximally to form adistinct V as in modern flamingos, but exactlyduplicating the condition in Recurvirostra; innertrochlea rotated farther medially, giving the distalend of the bone a narrower aspect than in modernflamingos (Figure 24), and in medial view with awing distinctly set off from the body of the tro-chlea (Figure 25c), as in recurvirostrids; hypotar-sus more complex (Figures 2bb, 26b), with closedcanals matching Himantopus, not consisting of twohigh calcaneal ridges with no canals as in modernflamingos; hypotarsal block in proximal view(Figure 26b) more pedicilate and in lateral view(Figure 23c) at nearly the same level as the cotylae(also true of Palaelodus), not set well below thecotylae as in modern flamingos; base of hypotar-sus extending distally as a ridge on shaft (Figure

FIGURE 24.—Anterior view of left tarsometatarsus: (a) stilt,Himantopus mexicanus (Recurvirostridae); (b) Juncitarsus gracil-limus (Phoenicopteridae; reconstructed from two specimens);(c) flamingo, Phoeniconaias minor (Phoenicopteridae); (d) stork,Ciconia abdimii (Ciconiidae). (Scales = 1 cm.)

NUMBER 316 51

III)S; ]

ill£" v

NUMBER 316 53

FIGURE 25.—Details of holotype tarsometatarsus of Juncitarsusgracillimus (USNM 244318): a, proximal end, anterior view;b, proximal end, posterior view; c, distal end, medial view;d, distal end, anterior view; e, distal end, posterior view;/ , distal end, lateral view. (Specimen coated with ammoniumchloride; X 3.)

25b), much more pronounced than in modernflamingos but similar to Himantopus.

The tarsometatarsus of Juncitarsus is more slen-der and elongate than in either Presbyornis orPalaelodus. In Palaelodus (Figure \9b) the shaft ismuch deeper and more laterally compressed, thehypotarsus more complex, the proximal end pro-portionately wider, the intercotylar knob heavier,

FIGURE 26.—Tarsometatarsi of Juncitarsus gracillimus: a, prox-imal end in anterior view, paratype of subadult (USNM244319) to show development of intercotylar knob; b, holo-type (USNM 244318), proximal view; c, same, distal view.( X I )

the inner trochlea much deeper and morerounded, and in posterior view more elongate. InPresbyornis, the hypotarsus is much wider basally,has no closed canals, the inner calcaneal ridge isreduced and the outer one produced distally as ahook; the distal end of the tarsometatarsus differsfrom that of Juncitarsus in that the inner trochleais less retracted posteriorly and lacks a distinctwing; the middle trochlea is more elongate andnot truncate on the posterior face, and the outertrochlea is deeper. The tarsometatarsus of Junci-tarsus bears no resemblance to that of any memberof the Ciconiiformes (e.g., Figure 24).

Humerus: The two portions of humerus of Jun-citarsus, unlike the tarsometatarsus, bear no resem-blance to the humerus in recurvirostrids. Theproximal end (Figure 27a) appears to have beenpneumatic (a small portion of the border of thepneumatic foramen remains in the one specimen)and does not have the deeply excavated,nonpneumatic tricipital fossae that are character-istic of the Recurvirostridae, most other Charad-riiformes, and Presbyornis. In the distal end, thedistinct ectepicondylar spur characteristic of re-curvirostrids and most other shorebirds is absent,as in flamingos.

•ft

FIGURE 27.—Proximal ends of left humeri of fossil flamingos:a, Juncitarsus gracillimus (USNM 244325); b, Palaelodus sp.(Scales = 1 cm).


The most characteristic feature of the proximalend of the humerus in Juncitarsus is a distinctelliptical scar for M. scapulohumeralis cranialis(the "supraspinatus" scar of Howard, 1929), dis-tal to the pneumatic foramen and lying betweenthat opening and the midline of the shaft. Asimilar deep scar is characteristic also of Palaelodus(Figure 21b) and may be faintly seen in thoseindividuals of modern flamingos in which thepneumatic foramen is less extensive and does notobliterate this area. We regard the resemblanceof this scar in Juncitarsus, Palaelodus, and modernflamingos as significant, as we have not seen asimilar condition in any modern group of birds,including the Ciconiiformes, Anseriformes, orCharadriiformes. The overall configuration anddetails of the proximal end and shaft of the

humerus in Juncitarsus very closely resemble thatin Palaelodus and modern flamingos. In all of theseforms, the margin of the deltoid crest describes along, gentle curve extending well down the shaft,in contrast to that seen in Ciconiiformes, in whichthe deltoid crest is variable but always shorterand more angular, as, for example, in storks andherons. Likewise, in Juncitarsus, Palaelodus, andmodern flamingos, the scar of M. latissimus dorsipars cranialis is quite similar, being situated mid-way between the midline and the external marginof the shaft, whereas in Ciconiiformes this scar iseither on the midline or lies between it and theinternal margin of the shaft.

The single distal end of humerus available forJuncitarsus (Figure 286) has the anconal surfaceand the area of the brachial depression crushed

FIGURE 28—Paratypes of Juncitarsus graallimus: a, left scapula, dorsal view (USNM 244328);b, distal end of left humerus, palmar view (USNM 244330); c, left radiale, proximal view(USNM 244333); d, left ulna, internal view (USNM 244326). (X 3.)

NUMBER 316 55

and is therefore not particularly useful. One dis-tinctive feature is the raised internal margin andsteeply angled surface of the attachment of theanterior articular ligament. From Presbyornis itdiffers further in having the internal condyle indistal and palmar views more rounded and theexternal condyle relatively narrower. Palaelodusand modern flamingos differ from Juncitarsus inhaving the external condyle curved internally atthe proximal tip.

Ulna: The proximal end of an ulna of Junci-tarsus (Figure 28d) is generally similar to that inmodern flamingos, but the attachment of theanterior articular ligament is broadly triangularand does not extend distally along the impressionof M. brachialis. In this respect Juncitarsus morenearly resembles the Recurvirostridae and differsgreatly from Palaelodus—in which the attachmentof the anterior articular ligament is very narrowand elongate. The proximal radial depression inJuncitarsus is deeper than in the forms just men-tioned.

Radiale: A well-preserved left radiale (Figure28c) is generally similar to that of modern fla-mingos, although we have not made extensivecomparisons of this element. It agrees with fla-mingos, as opposed to recurvirostrids, in beingmore elongate and in having a distinct notch inthe dorsal surface. It is proportionately narrowerthan in modern flamingos, but otherwise is moresimilar to the Phoenicopteridae than to the Ci-coniiformes or Charadriiformes.

Scapula: The only available scapula of Junci-tarsus (Figure 28a) is the anterior portion from ajuvenile. The coracoidal articulation is perhapsslightly better developed than in Phoenicopterusbut not as distinctly produced as in Palaelodus orPresbyornis, nor a distinct ball as in storks andherons. The acromion is not as elongate andpointed as in Presbyornis, Palaelodus, or modernflamingos, and is not unlike that of Himantopus.The shaft is narrow, rather rounded in cross-section, and deflected ventrally, as in Palaelodus,Presbyornis, and modern flamingos, and quite un-like the broad flattened condition in Ciconi-iformes.

Vertebrae: We undertook comparisons of thefour vertebrae included among the material ofJuncitarsus with little anticipation of discoveringmuch of interest. We found, however, these spec-imens were of great value in assessing the affinitiesand adaptations of the bird.

The most complete specimen corresponds veryclosely with the 16th cervical of Phoenicopterus(15th of Phoeniconaias), differing mainly in thenarrower and longer prezygapophyses and thenarrower and deeper articular surfaces of thecentrum. In modern flamingos, this vertebra oc-curs at the point of transition from the greatlyelongated anterior cervicals to the shortened pre-thoracic cervicals. It is characterized by havingthe ventral surface of the centrum elongate andflat, with two distinct anterior hemapophyses,and by having a large, sloping, expanded neuralcrest. The agreement of the fossil with modernflamingos is striking (Figure 29), particularlysince we found no vertebra comparable in pro-portions and details in any of the other orders ofbirds possibly related to flamingos, including theCharadriiformes. In the Recurvirostridae theequivalent vertebra is much shorter, has no neuralcrest, and has only a single hemapophysis. In theAnatidae, the vertebrae in this area are somewhatsimilar but are not as elongate and do not possessa combination of a high neural crest and doublehemapophyses. The transitional cervical verte-brae in the families of Ciconiiformes are not aswell demarcated, except in herons. There is con-siderable difference between the families, butnone approaches the condition in flamingos. Inherons, for example, the ventral surface of themost nearly comparable vertebra bears a large,thin crest, totally unlike the double hemapo-physes and flattened and elongated centrum ofmodern flamingos and Juncitarsus.

A fragmentary anterior portion of an anteriorcervical of Juncitarsus (Figure 30a,b) indicates avery elongate and slender vertebra with a dis-tinctly oval cross-section (Figure 30a) such as seenin modern flamingos. It is most similar to the 6thcervical in Phoenicopterus, differing in having thehemal spines reduced and more widely divergent

NUMBER 316 57

FIGURE 29—Cervical vertebra of Juncitarsus gracillimus (para-type USNM 244321) compared with the 15th cervical ver-tebra of the modern flamingo Phoemconaias minor (primedletters): a, lateral view; b, dorsal view; c, ventral view;d, anterior view; e, posterior view of centrum. (Juncitarsus X2.6; Phoemconaias X 2).

and the ventral surface of the centrum forming aslight crest rather than being flattened; in the lastfeature it is like Himantopus. Only the Ardeidaehave comparably elongated vertebrae. The fossilagrees with flamingos and differs from herons inthe deeper anterior articular surface and the shal-lower depression on the ventral surface posteriorto the articulation.

Another vertebra of Juncitarsus (Figure 30c)bears a costal facet and therefore formed part of

FIGURE 30.—Paratypical vertebrae of Juncitarsus gracillimus:a, anterior cervical equivalent to 6th cervical vertebra ofPhoenicopterus, cross section (USNM 244338); b, same, lateralview of right side; c, thoracic vertebrae equivalent to the17th and 18th vertebrae of Recurvirostra, lateral view of leftside (USNM 244332, 244331). (X 3.)

the thoracic series. It bears the remains of a singlelarge hemal spine and thus came from the ante-rior portion of the series. In modern flamingosthese vertebrae are fused into a notorium and theneural spines are less prominent. The presentvertebra had a large neural spine and was un-fused. It bears no resemblance to any of thethoracic vertebrae of modern flamingos butagrees very closely with the 17th vertebra (2ndthoracic) of Recurvirostra.

A second specimen of thoracic vertebra (Figure30c) was also unfused and bears a costal facet.The centrum is laterally compressed unlike mod-ern flamingos. This vertebra also agrees well withRecurvirostra and appears to be the one succeedingthe previously mentioned thoracic.

The resemblance of the thoracic vertebrae ofJuncitarsus is closest to the Charadriiformes. Aswith the cervicals, there is great variation in thethoracic vertebrae between the families of Cico-niiformes, none of which have the centra as com-pressed as in Juncitarsus. In storks and herons noneof the thoracics bear well-developed hemal spines.In ibises some of the thoracics are fused into anotorium, but this involves a different series ofvertebrae than that of modern flamingos.

Frontal Bone: Associated with the other mate-rial of Juncitarsus is a small unfused frontal bone,the sutures of which obviously had not closed. Itis important in showing in the supraorbital areaa distinct impression for a salt gland (Figure 31),in the same position as seen in modern flamingos.

Pedal Phalanx: A toe bone appears to be pha-lanx 1 of right digit II. The element differs frommodern flamingos in being less laterally com-pressed, and from storks and herons in being lesselongate. Its resemblances are closer to the Re-curvirostridae.

Femur: An abraded distal end of a femur re-ferred to Juncitarsus preserves no useful characters.

DISCUSSION.—Juncitarsus is the earliest knownfossil that is definitely referrable to the Phoeni-copteridae. It is known as yet only from theearliest Middle Eocene. Presbyornis, a colonialcharadriiform near the ancestry of Anseriformes,is abundant in early Eocene deposits and thus


FIGURE 31.—Unfused frontal bone of juvenile specimen ofJuncitarsus gracilUmus (USNM 244336) showing depression forsalt gland (arrow). (X 3.)

must have been contemporaneous with the ances-tors of Juncitarsus.

The available evidence indicates that Juncitar-sus may have nested colonially. It had salt excret-ing glands and therefore inhibited saline environ-ments at least intermittently, if not entirely. Bothof these conditions are typical of modern flamin-gos.

Juncitarsus possessed apparently derived char-acters of the cervical vertebrae and humeruswhich ally it with the Phoenicopteridae, althoughit retains the morphology of the Charadriiformesin the thoracic vertebrae and most features of thetarsometatarsus. The last element is particularlysimilar to that of the Recurvirostridae, corrobor-ating our proposed relationship between thePhoenicopteridae and the Recurvirostridae. Thatthe cervical vertebrae of Juncitarsus are specializedlike those of modern flamingos suggests that someof the aspects of the highly modified feedingapparatus of flamingos had evolved by the Mid-dle Eocene. The thoracic vertebrae of Juncitarsusare unfused, indicating that the notorium of mod-ern flamingos evolved subsequent to the special-izations of the neck.

Juncitarsus provides a mosaic of phoenicopteridand charadriiform characters and in size bridgesthe gap between the modern members of theRecurvirostridae and the Phoenicopteridae. Itshows no approach to the morphology of eitherCiconiiformes or Anseriformes.

Aspects of Evolution of Filter Feeding

The most characteristic feature of flamingos istheir filter-feeding apparatus, which has beenlikened to that of the Anseriformes and thereforehas been cited as evidence for a relationshipbetween the two groups. Greatly detailed descrip-tions of the structure and function of the feedingapparatus are available for flamingos (Jenkin,1957) and for ducks (Goodman and Fisher, 1962;Zweers, 1974; Zweers et al., 1977). These lack,however, a more general overview of evolutionaryevents that may have led to development of suchadaptations.

Filter feeding is commonly employed in diversegroups of invertebrates and in some fishes andtadpoles. Among higher vertebrates, however,well-developed straining devices have evolvedonly in baleen whales and in three groups ofbirds: flamingos, ducks, and one genus of petrels{Pachyptila, Procellariidae). Certain penguins(Spheniscidae) and auks (Alcidae) show some ofthe structures needed for feeding on small orga-nisms but have not evolved lamellate strainers. Itis instructive to examine their adaptations first.Zusi (1974a) has provided a valuable account offeeding adaptations in penguins, and Bedard(1969) has discussed those of alcids. The plank-ton-eating forms of these groups are characterizedby an enlarged tongue, a broadened rostrum, andcornified papillae on the roof of the mouth and,in the case of penguins, on the tongue, innersurface of lower jaw, and larynx also. The en-larged tongue may be accomodated in two ways;either by a distensible gular pouch as in alcids, orby a deepening of the bones of the lower jaw, bestexemplified in penguins by species of the genusEudyptes. The penguins and alcids thus adapted,feed largely on small crustaceans which arethought to be taken individually rather than witha netlike scooping (Zusi, 1974a: 76). Prey is takenvery rapidly, however, and is held and manipu-lated toward the throat by action of the tongue.

The next steps in the evolution of filter feedingare well illustrated by the petrels of the genusPachyptila, sometimes known as whalebirds, onaccount of the baleen-like feeding apparatus in

NUMBER 316 59

some species. In Pachyptila there is a continuumof variation in bill shape, from proportions nearlytypical of other members of the family (P. belcheri)to the grotesquely broadened bill of P. vittata (seeillustrations in Murphy, 1936; Fleming, 1941).All have an enlarged tongue, which, as with theplankton-feeding alcids, is accomodated by a dis-tensible gular pouch rather than by a deep lowerjaw. Each possesses at least the rudiments oflamellae, always on the inner surface of the edgeof the upper jaw, with more pronounced devel-opment posteriorly than anteriorly. In the nar-row-billed species these "lamellae" are very smalltransverse projections extending medially lessthan a millimeter from the cutting edge of thebill. They reach their extreme in P. vittata, whichpossesses a bona fide straining device consistingof a series of plate-like lamellae extending ven-trally some 3.5 mm from a ridge inside the edgeof the bill.

Although Murphy (1936:614) held that in thenarrow-billed forms of Pachyptila "it would behard to believe that the faint striations [in theupper jaw] could have any function whatsoever,"it is more difficult to believe that such structureswould evolve without having some function.Quite probably the rudimentary lamellae in theless specialized filter-feeding petrels provide gapsfor the expulsion of water while the prey is heldin place with the tongue. With increasing spe-cialization, the lamellae enlarge and begin toassume the role of food retention while the tongueforces out water and moves the prey toward thethroat. As with penguins (Zusi, 1974a) and fla-mingos (Jenkin, 1957), increasing specializationof the feeding apparatus in Pachyptila is correlatedwith a decrease in size of prey (Watson, 1975).

The filter-feeding petrels thus illustrate severalimportant points. First, the initial adaptation forfilter-feeding is enlargement of the tongue. Thisis accompanied or followed by widening anddeepening of the bill and the development oflamellae. Secondly, the initial function of lamel-lae may be to permit the expulsion of water ratherthan to retain prey. Third, the species of Pachyptilashow progression from a primitive, nearly unmod-ified species, to a highly evolved lamellate filter

feeder within a single extant genus, suggestingthat the great broadening of the bill and thedevelopment of a fine strainer may occur rapidlyin birds. Fleming (1941) considered speciation inPachyptila to be the result of climatic changesassociated with Pleistocene glacial events, which,if true, would make the more specialized filter-feeding adaptations in the genus of very recentorigin.

Flamingos are the most specialized of avianfilter feeders. The peculiarly bent bill, used in an"upside down" attitude when feeding, separatesthe living flamingos from all known birds. Twostrikingly different modifications of the bill occurin flamingos, the least specialized being that ofPhoenicopterus, the most derived condition beingthat found in Phoenicoparrus and Phoeniconaias. Inall flamingos the lower jaw is deep and swollenand the upper jaw is narrow. In Phoenicopterus(Figure 32), the lower jaw is essentially a largecylinder housing an immense tongue equippedwith a series of spiny protuberances. Fine ridgeslying transverse to the long axis of the bill occuron the narrow dorsal surface of the two sides ofthe lower jaw. The upper jaw is rather flat, witha slightly developed median keel. Coarse hooklikelamellae are present on the outer edge of theupper jaw, inside of which is a series of diagonally-oriented, fine plate-like lamellae lying in a narrowband between the outer edge of the jaw and thebare median keel.

In Phoenicoparrus and Phoeniconaias the upperjaw is much narrower and bears a deep mediankeel (Figure 39a), with long rows of exceedinglyfine microscopic, feather-like lamellae. The coarselamellae along the edge of the jaw are much finerand less hooklike than in Phoenicopterus. In thelower jaw, the space for the tongue is greatlyreduced because the dorsal lamellate ridges seenin Phoenicopterus have expanded and are turnedinward and downward to provide a greater sur-face area for rows of fine lamellae opposing thoseof the keel on the upper jaw. Phoeniconaias feedson much smaller food particles than does Phoeni-copterus (Jenkin, 1957).

The filtering apparatus in the Anatidae is quitedifferent from that of flamingos but the essential


FIGURE 32.—Head of the flamingo Phoemcopterus ruber. Lowerjaw removed and displaced below, viewed from the externalsurface. Note that the tongue is housed in the lower jaw andthat neither the lower jaw nor the keel on the upper jaw areas deep as in Phoeniconaias (see Figure 39).

FIGURE 33.—Head of a Mallard Duck, Anas platyrhynchos.Upper jaw removed and displaced above, viewed from theinternal surface and reversed to preserve the same orienta-tion. Note that the tongue is accomodated by the upper jawand that there are two bulges in the tongue that form adouble piston suction pump, unlike the mechanism in theflamingo.

structures are all present: enlarged tongue, broadbill, and lamellae. The bill shape of ducks is toofamiliar to need description but perhaps it hasnot been realized that it presents an adaptationnot found in other filter-feeding vertebrates: theenlarged tongue is accomodated by the upper

rather than the lower jaw (Figure 33). This is afundamental difference between ducks and fla-mingos and one that has not been emphasized bythose attempting to ally the two groups. Further-more, the underlying bony structure of the tongueis very different in the two groups. In ducks, theparaglossale, the terminal bone of the hyoid ap-paratus supporting the tongue, is a distinctiveexpanded, spatulate bone, whereas in flamingosit is slender and more rodlike, as is typical of mostbirds.

In typical anatids, rows of lamellae lie alongthe inner surfaces of both the upper and lowerjaws, the latter series being divided into upperand lower rows. The morphology of these lamel-lae may vary considerably, even within a genus.In Anas, the lamellae may be coarse, hard plates,as in the Mallard Anas platyrhynchos, or they maybe prolonged into fine hairlike fringes as in thecongeneric Shoveler Anas clypeata. Again, as inPachyptila, the evolution of fine lamellate struc-tures seems to be accomplished with relative ease.In no case, however, are the lamellae in ducks likethe rows of platelets or feather-like strainers offlamingos.

The process of filtration in flamingos is quitedifferent from that of ducks. In flamingos, themouth is opened and water enters along the entiregape, which then closes, the tongue forcing waterback through the filtering devices. In the normalfeeding process of Phoeniconaias and when Phoeni-copterus feeds on mud, this process is modified sothat the outer lamellae function as excluders,keeping out larger inorganic particles (Jenkin,1957). In ducks, the bill and tongue act as asuction pressure pump consisting of two pistonsin a cylinder (Figure 33), with the water enteringat the tip of the bill and being expelled posteriorly(Zweers et al., 1977).

The advocates of ciconiiform relationships offlamingos have not hypothesized how the com-plicated filter-feeding apparatus might haveevolved from the hard, pointed bill of a stork;indeed it is difficult to envision such a transfor-mation taking place. Several authors (Reichenow,1877; Jenkin, 1957; Sibley et al., 1969) have

NUMBER 316 61

FIGURE 34.—Bills of two specimens of the African Open-bill Stork, Anastomus lamelligerus, toshow the variation in development of the bristly pads on the ramphotheca. It is obvious bytheir placement that these structures could have no filtering function.

mentioned the so-called "lamellae" of the open-bill storks, Anastomus, in connection with theevolution of straining devices in flamingos, butthe analogy is invalid. The structures on therhamphothecae of Anastomus (Figure 34) are in noway comparable to lamellae. Kahl (1971:27)much more appropriately refers to them as"leathery columnar 'pads.'" He has shown thatthe open-bill structure in Anastomus provides aforceps-like action of the bill tips that functionsin preying on mollusks, particularly large snails,which are the chief item of diet in these storks.

The bristly pads on the rhamphothecae ofAnastomus probably do not function in feeding.They occur along the edges of the gap in the billand their relative prominence appears to be cor-related with the age of the individual, as there isconsiderable variation in development (Figure34). In most birds, the edges of the upper and

lower jaws come in contact, serving to keep thehorny rhamphothecae worn down and probablysharpened as well (Tunnicliffe, 1973). Such con-tact is not possible in Anastomus, however, and therhamphothecae appear to be unable to checkeach other against continued outgrowth. Thusthe bristly pads in this genus may well be purelyadventitious structures. Regardless, Anastomusdoes not represent any stage in the evolution offilter feeding.

Because we propose a shorebird ancestry bothfor flamingos and for ducks (Olson and Feduccia,in press), there should be taxa among the Char-adriiformes having morphological precursors ofphoenicopterid and anatid feeding adaptations.Within the Recurvirostridae, Recurvirostra andCladorhynchus are adapted for feeding rapidly onlarge numbers of small prey items. Accordingly,the tongue is enlarged and is short, broad, and


fleshy (Figure 35), to facilitate movement of smallprey into the throat. The development of a large,fleshy tongue, as we have noted, is the first majorstep involved in the evolution of filter feeding. Instorks, however, the tongue is rudimentary (Gard-ner, 1925).

An unrecognized example of a shorebird show-ing the beginning stages in the development offilter feeding is the Red Phalarope, Phalaropusfulicarius (Phalaropodidae). In the other two mem-bers of the Phalaropodidae (Lobipes lobatus andSteganopus tricolor) the bill is extremely slender,pointed, and needlelike (Figure 36/>), but in Phal-aropus the bill is broad, deep, and subspatulate inshape (Figures 36a, 376, 38a), with wide, flexiblemargins. The tongue is correspondingly enlargedand from the postero-lingual rims of the lowerjaw project a few well-developed, narrow papillaethat may function as strainers (Figure 36a). Inother phalaropes the tongue is narrow and pa-pillae in the lower jaw are absent or very rudi-mentary (Figure 36£). Phalaropus fulicarius feedsmainly on small insects and crustaceans and isreported to take food at a much faster rate thanother phalaropes (Bent, 1927). Harold Mayfield(in litt. to Olson, 17 January 1978) writes that onthe breeding grounds, Phalaropus may be seen tofeed by extracting chironomid midge larvae fromfloating vegetation and often is observed withstrands of vegetation hanging from the mouth.

B

FIGURE 35.—Mouth of the American Avocet, Recurvirostraameruana, to show the broad, fleshy tongue. The acquisitionof such a tongue is the first step in the development of filter-feeding.

FIGURE 36.—Dorsal view of the lower jaw and tongue of aRed Phalarope, Phalaropus fulicarius (a), and a NorthernPhalarope, Lobipes lobatus (b). The broad bill, enlargedtongue, and papillae of Phalaropus probably indicate a ca-pacity for primitive filter-feeding.

He reports that stomachs almost always containbits of vegetation apparently taken in acciden-tally while feeding. Thus it appears that Phalaro-pus actually engages in a primitive straining typeof filter-feeding as would be predicted from itsmorphology.

The bill in most Charadriiformes is typicallyrather weak and flexible, due in large measure tothe long schizorhinal nostrils and the short pre-maxillary and mandibular symphyses. This hasproved to be a very plastic structure, and conse-quently the Charadriiformes exhibit the greatestdiversity in feeding adaptations of any order ofbirds (see examples illustrated in Zusi, 1974b).Contrast, for example, the bill of a SpoonbillSandpiper (Eurynorhynchus) with that of a skim-mer (Rynchops).

NUMBER 316 63

It is well known that the bill in young flamingosis straight and only later in ontogeny developsthe characteristic crook. This has received onlypassing comment in the literature, with mostauthors stating that the flamingo bill is "goose-like" in its early stages. The bill in flamingos isnever really gooselike but in its earliest stages canbetter be likened to that of a shorebird such asPhalaropus, being rather long and narrow withsomewhat flexible margins (Figure 37). In lateralview, the bill is deeper than usual for most Char-adrii (Figure 38b) and ontogenetic changes in-volve increasing the depth of the mandible and

the curvature of both jaws (Figure 38c,d). At onepoint in the development of Phoenicopterus ruber,the bill is shaped almost exactly as in the earlyMiocene species P. croizeti (Figures 20, 38c), andthe ontogeny of the bill in modern flamingosappears to reflect rather closely their phylogeny.To evolve such a structure from the conditiontypical of the Charadrii would mainly requirefusion by increased ossification of various separateelements, whereas to evolve the flamingo feedingapparatus from the already greatly fused cranialstructure of a stork would involve a series ofhighly improbable intermediate stages.

\

B

FIGURE 37.—Dorsal view of bill of downy chick of a flamingo, Phoenicopterus ruber (a) comparedto an adult Red Phalarope, Phalaropus fulicarius (b).


FIGURE 38.—Lateral view of bill: a, Red Phalarope, Phalar-opus fulicanus; b-d, progressive developmental stages of fla-mingos: b, downy chick of Phoemcopterus ruber less than a weekold; c, chick of P. ruber about 30 days old; d, chick ofPhoemconaias minor 7 weeks old (adapted from Kear andDuplaix-Hall, 1975, plate 48). Note the progressive deepen-ing of the mandible and bending of the bill with increasingage.

Although the filtering techniques of flamingosand ducks are entirely different from one another,both methods require repeated and rapid openingof the bill, which can be accomplished by movingeither the upper or the lower jaw, or both. Be-cause M. depressor mandibulae causes protrac-

tion of the upper jaw as well as depression of themandible (Zusi, 1967), it has a very importantrole in the feeding process in flamingos and inducks. The enlarged bladelike retroarticular pro-cesses of the mandible in both groups provideincreased area for the insertion of the depressormandibulae. The retroarticular processes and de-pressor mandibulae are well-developed in severalgroups of shorebirds, notably so in Recurvirostra(Burton, 1974), and although not as exaggeratedas in flamingos and ducks, are quite suitable asprecursors. Such retroarticular processes are lack-ing in storks and herons, however. Because of theimportance of the depressor mandibulae in filter-feeding, regardless of how such filtering is accom-plished, it is not improbable that the enlargedretroarticular processes of flamingos arose inde-pendently of those in ducks.

The astute naturalist Stejneger (1885:153)noted in passing the resemblance between thefeeding apparatus of flamingos and that of baleenwhales, but devoted only a sentence to it. Thesimilarity between the head of a flamingo andthat of the right and bowhead whales (Eubalaenaand Balaena) is quite striking. In these whales theupper jaw is quite narrow, with a ventral keelbearing long plates of baleen. The lower jaw isvery deep and swollen, accomodating a hugetongue and rising up so as to cover the baleenplates of the upper jaw when the mouth is closed.Unlike flamingos, there are no straining deviceson the lower jaw, but otherwise the general ap-pearance of the head is so like that of flamingos(Figure 39) as to be one of the more outstandingexamples of convergence in the animal kingdom.

The resemblances extend to the skeleton aswell. The rostrum in Eubalaena and Balaena, par-ticularly as compared to that in other baleenwhales such as Balaenoptera, is markedly crookedand bent downward so that it is superficially likethat of a flamingo (Figure 40). The bony structureof the lower jaw is not like that of flamingosbecause the great depth of the lower jaw inBalaena is achieved by a mass of connective tissuebuilt up on a mandible not substantially differentfrom that of other baleen whales.

NUMBER 316 65

B

FIGURE 39.—Head of a Lesser Flamingo, Phoemconaias minor (a), compared with that of a BlackRight Whale, Eubalaena glaaalis (b), to show the convergent similarities in the filter-feedingapparatus.


FIGURE 40.—Skull of a Lesser Flamingo, Phoemconaias minor(a), compared with that of a Black Right Whale, Eubalaenaglaaalis (b), to show the similarity in the shape of the long,decurved bony structure of the rostrum.

The bend in the rostrum of Balaena and in thedeep-keeled flamingos confers one obvious advan-tage: it allows the filtering devices in the middleportions of the rostrum to become much longerand more extensive. Balaena has the longest andfinest baleen of any of the filter-feeding whalesand as a consequence feeds on smaller organisms.

The bent rostrum must confer other advan-tages, however, for Phoenicopterus has no deep keel.First of all, it provides space for an enlargedtongue while still allowing the tips of the jaws tocome together, thereby permitting flamingos topick up individual food items (Rooth, 1965) andto manipulate nesting material. More impor-tantly, the bend in the bill also allows the entirelength of the gape to be opened with a minimumof movement of the jaws (Jenkin, 1957). At leastin flamingos, the bend must first have arisen forthese purposes, being followed by its use for ex-pansion of the filtering apparatus in the moreadvanced species.

The similarity of the heads of flamingos andwhales is probably a good indication that thebent structure of the flamingo bill arose in ac-cordance with the constraints of filter feedingalone, the "upside-down" feeding posture havingevolved secondarily. The more primitive baleen

whales, with their straighter and wider upperjaws, might provide useful analogs to the mor-phological stages through which flamingos mayhave passed during their evolution, although thisis demonstrated just as well in the ontogeny ofliving species (Figure 38).

Modern flamingos feed in highly saline watersand the available evidence indicates that theearliest fossil flamingo, Juncitarsus, may have oc-curred in a similar habitat. The same is true ofPresbyornis, a filter-feeding charadriiform near theancestry of ducks. Highly saline lakes provideabundant food in the form of swarms of brine-adapted arthropods and growths of algae. Thesmall size of these organisms, however, wouldnecessitate a filtering device for effective feeding.Furthermore, in such habitats there would bestrong selection for feeding methods that preventthe ingestion of water, in order to avoid the higherenergetic cost of metabolic excretion of salt. Thus,highly saline lakes may have been the originalhabitat for both flamingos and Anseriformes.There, selection would have favored the devel-opment of specialized filtering mechanisms thatmight never have evolved under less strenuousconditions.

To summarize, the filter-feeding adaptations offlamingos are seen to be entirely different fromthose of the Anseriformes and are not indicativeof a relationship between these two groups. Mor-phological precursors for such adaptations existamong the Charadriiformes but not in storks.

Conclusions

The association of the Phoenicopteridae withthe Ciconiiformes has resulted solely from histor-ical traditions and misconceptions. We haveshown that there is no substantial or convincingevidence of relationship between flamingos andeither the Ciconiiformes or Anseriformes, whereasall of the attributes of flamingos are compatiblewith a charadriiform origin, including those thatonce seemed inexplicable when flamingos wereconsidered as bewildering mosaics of ciconiiformand anseriform characters.

If flamingos were derived from Ciconiiformes

NUMBER 316 67

and converged towards Charadriiformes, onewould expect fossil flamingos to have more stork-like characters than modern flamingos (whichhave no stork-like characters anyway). This isdefinitely not the case, for the earliest knownflamingo, Juncitarsus, has no similarities to theCiconiiformes and in many respects is more sim-ilar to the Recurvirostridae than are modernflamingos. We cannot see that convergent evolu-tion is really responsible for the confusion sur-rounding the classification of flamingos; theirsimilarities to storks go no farther than the simi-larities between all birds with long legs and longnecks.

The Anseriformes likewise prove unsatisfactoryas ancestors of flamingos. Although members ofboth groups have webbed feet, swim and tip upwhile feeding, and have honking vocalizations,so, too, do members of the charadriiform familyRecurvirostridae. The fact that both flamingosand ducks are filter feeders is of no systematicconsequence because their basic structural adap-tations for filtering are completely dissimilar andevolved independently of one another. Flamingosand Anseriformes share similar mallophagan par-asites, but new knowledge of the evolutionaryhistory of these two groups provides a logicalexplanation for this that does not involve a closerelationship of the hosts: the ancestors of bothflamingos and ducks originally evolved in thesame harsh saline environments and probablyhad similar breeding habits that would havepromoted transfer of parasites between the downyyoung. No anatomical evidence exists that sup-ports an ancestral-descendant relationship forAnseriformes and flamingos.

On the other hand, in virtually all of theirfeatures that are not unique, flamingos resemblethe Charadriiformes. This is supported by osteol-ogy, myology, behavior, internal parasites, andby the existence of a fossil form intermediatebetween the Recurvirostridae and modern fla-mingos.

In their prediliction for highly saline environ-ments, their coloniality, simplified breeding dis-plays, habit of feeding on very small prey, en-

larged fleshy tongue, and large retroarticular pro-cesses of the mandible, the Recurvirostridae pos-sess the necessary ecological, behavioral, and mor-phological precursors of the more specialized fea-tures of flamingos. Furthermore, modern fla-mingos and the recurvirostrid Cladorhynchus sharecertain characters that are either derived for allbirds (M. iliotibialis medialis; breeding habits) orare derived within the Charadriiformes (two coatsof white, unpatterned nestling down; elongatedtertials). Cladorhynchus is nevertheless clearly amember of the Recurvirostridae. Thus the evi-dence suggests that flamingos evolved directlyfrom the Recurvirostridae rather than from aproto-recurvirostrid. Otherwise, it would have tobe assumed either that all the derived charactersshared by Cladorhynchus and flamingos evolvedtwice, which is implausible, or that all recurviros-trids once possessed these characters and thatRecurvirostra and Himantopus have lost them and"re-evolved" all their primitive, more typicallycharadriiform characters, which is even less likely.If flamingos evolved directly from the Recurvi-rostridae, then it must be assumed that the Re-curvirostridae were in existence before the middleEocene, when the first flamingos appear. As yet,the only fossils possibly belonging to the Recur-virostridae are a few fragmentary bones of Mio-cene age.

The principal differences between modern fla-mingos and the Recurvirostridae are in the highlymodified feeding apparatus, the increased num-ber and specialization of the cervical vertebrae,and the pneumaticity of the proximal end of thehumerus, with the associated changes in osteologyand myology of the shoulder. These characterscan then be said to define the family Phoenicop-teridae. Although the nature of the feeding ap-paratus is not known for the fossil genera Junci-tarsus and Palaelodus, both taxa possess modifiedhumeri and cervical vertebrae and therefore areclearly referable to the Phoenicopteridae.

Within the Phoenicopteridae there are threebasic generic groups: Juncitarsus, from the middleEocene; Palaelodus, from the late Oligocene orearly Miocene to early Pliocene; and "Recent"


flamingos, from the late Oligocene or early Mio-cene to the present. None of the known featuresof the long-legged Juncitarsus would preclude itfrom being very close to or on a direct line withRecent flamingos, whereas the Palaelodus grouprepresents a separate, shorter-legged swimminglin- ige. There is, as yet, no evidence to indicatewhether Palaelodus diverged from the more typicalflamingos before or after the evolution of Juncitar-sus.

The Phoenicopteridae are shorebirds with ahighly evolved feeding apparatus. They are de-rived not only from members of the suborderCharadrii, but from a particular family of thatsuborder, the Recurvirostridae. Their distinctive

peculiarities entitle them to rank as a separatefamily, but it is not possible to accord them anyhigher position, such as a suborder or order,without raising each genus or family of Charad-riiformes with a specialized feeding adaptation toan equivalent rank—an obviously undesirableprocedure. To add proper perspective, it shouldbe observed that flamingos differ no more fromthe Recurvirostridae than, for example, the An-hingidae differ from the Sulidae, yet these lasttwo families are almost always included in thesame suborder of Pelecaniformes.

Therefore, we place the Phoenicopteridae inthe order Charadriiformes, suborder Charadrii,immediately following the Recurvirostridae.

Literature Cited

Allen, Robert Porter1956. The Flamingos: Their Life History and Survival.

National Audubon Society Research Report, 5:1-285,illustrated.

Ameghino, Florentino1899. Sinopsis geologico-paleontologica: Suplemento (adiciones

y correccwnes). 13 pages. La Plata: Imprenta yEncuadernacion "La Libertad."

Baer, Jean G.1951. Ecology of Animal Parasites. 224 pages, 162 Figures,

Urbana: University of Illinois Press.1957. Repartition et endemicite des cestodes chez les

reptiles, oiseaux et mammiferes. In Premier Sym-posium sur la specificite parasitaire des parasitesdes vertebres. International Union of Biological Sciences,series B, 32:270-292.

Baumel, J. J., A. S. King, A. M. Lucas, J. E. Breazile, andH. E. Evans, editors

1979. Nomina Anatomica Avium. 637 pages, illustrated.London: Academic Press.

Bedard, Jean1969. Adaptive Radiation in the Alcidae. Ibis, 111:189-

198.Beddard, Frank E.

1898. The Structure and Classification of Birds. 524 pages,252 figures. London: Longmans, Green, and Co.

Bent, Arthur Cleveland1927. Life Histories of North American Shore Birds:

Order Limicolae (Part 1). United States NationalMuseum Bulletin, 142:1-420, 55 plates.

Beruldsen, Gordon R.1972. Unusual Feeding Method by Stilts and Avocets.

Australian Birdwatcher, 4(4): 129-130.Brodkorb, Pierce

1961. Birds from the Pliocene of Juntura, Oregon. Quar-terly Journal of the Florida Academy of Sciences, 24(3):169-184, 7 figures.

1963a. Catalogue of Fossil Birds, Part 1 (Archaeoptery-giformes through Ardeiformes). Bulletin of the Flor-ida State Museum, Biological Sciences, 7(4): 179-293.

1963b. Birds from the Upper Cretaceous of Wyoming.Proceedings, XIIHh International Ornithological Con-gress, 1:55-70, 10 figures.

Brown, Leslie H.1959. The Mystery of the Flamingos. 116 pages, illustrated.

London: County Life Limited.1975. East Africa. In J. Kear and N. Duplaix-Hall,

editors. Flamingos, pages 38-48. Berkhamsted: T.and A. D. Povser.

Brush, Alan H.1979. Comparison of Egg-White Proteins: Effect of Elec-

trophoretic Conditions. Biochemical Systematics andEcology, 7(2): 155-165, 6 figures.

Burton, P.J.K.1974. Feeding and the Feeding Apparatus in Waders: A Study

of Anatomy and Adaptations in the Charadrii. 150 pages,41 figures. London: British Museum (NationalHistory).

Carnaby, I. C.1946. A Further Nesting Record of the Banded Stilt.

Emu, 46(2): 156-158.Chandler, Asa C.

1916. A Study of the Structure of Feathers, with Refer-ence to Their Taxonomic Significance. Universityof California Publications in Zoology, 13(11):243-446,7 figures, plates 13-37.

Cracraft, Joel1973. Systematics and Evolution of the Gruiformes

(Class Aves), 3: Phylogeny of the Suborder Grues.Bulletin of the American Museum of Natural History,151(1): 1-127, 51 figures, 49 tables.

Cramp, S., and K.E.L. Simmons, editors1977. The Birds of the Western Palearctic. Volume 1, 722

pages, 108 plates. Oxford: Oxford UniversityPress.

Dames, W.1890. Ueber Vogelreste aus dem Saltholmskalk von Lim-

hamm bei Malmo. Bihang till Kongliga Svenska Ve-tenskaps-Akademiens Handlingar, 16, part 4 (1):1 11,1 plate.

Des Murs, Oeillet1860. Traite general d'oologie ornithologique au point de vue de

la classification, xix -I- 640 pages. Paris: FriedrichKlincksieck.

Dinsmore, J. J.1977. Notes on Avocets and Stilts in Tampa Bay, Flor-

ida. Florida Field Naturalist, 5(2):25-30.Duncker, Hans-Rainer

1971. The Lung Air Sac System of Birds. 171 pages, 41figures. Berlin: Springer-Verlag.

Feduccia, Alan1976. Osteological Evidence for the Shorebird Affinities

of the Flamingos. Auk, 93(3):587-6Ol, 9 figures.1977a. Neuer Stammbaum von Ente und Flamingo. Btld

der Wissenschaft, Akzent, 5(7):[1], 1 figure.1977b. Hypothetical Stages in the Evolution of Modern

Ducks and Flamingos. Journal of Theoretical Biology,67:715-721.

69


1978. Presbyomis and the Evolution of Ducks and Fla-mingos. American Scientist, 66(3):298-304, 7 figures.

In press. Evolution von Enten und Flamingo. Proceedings,XVII International Ornithological Congress, Berlin.

Feduccia, Alan, and Paul O. McGrew1974. A Flamingolike Wader from the Eocene of Wyo-

ming. Contributions to Geology, University of Wyoming,13(2):49-61, 13 figures, 1 table.

Fisher, Harvey I., and Donald C. Goodman1955. The Myology of the Whooping Crane, Grus amen-

cana. Illinois Biological Monographs, 24(2): 1-127, 40figures.

Fleming, C. A.1941. The Phylogeny of Prions. Emu, 41:134-155, 9 fig-

ures.Fraas, Oscar

1840. Die Fauna von Steinheim, mit Riicksicht auf diemiocanen Saugethier- und Vogel-reste des Stein-heimer Beckens. Jahreshefte des Veretns fur vaterldn-dische Naturkunde in Wurtlemberg, 26:145-306, 13plates.

Fiirbringer, Max1888. Untersuchungen zur Morphologie und Systematik der

Vogel, zugleich ein Beitrag zur Analomie der Stiitz- undBewegungsorgane. 2 volumes, 1751 pages, 30 plates.Amsterdam: Von Holkema. [Issued as KoninklijkZoblogisch Genootschap "Natura Artis Magistra," Bijdr-agen tot de Dierkunde (Amsterdam), 15.]

1902. Zur vergleichenden Anatomie des Brustschulter-apparates und der Schultermuskeln, 5 Teil: Vogel.

Jenaischen Zeitschrift fur Naturwissenschaft, 36:290-736, plates 18-22.

Gadow, Hans1877. Anatomie des Phoenicopterus roseus Pall, und seine

Stellung im System. Journalfur Omithologie, 25:382-396.

1889. On the Taxonomic Value of the Intestinal Con-volutions in Birds. Proceedings of the Zoological Societyof London, 1889:303-316, 1 figure, plate 32, 1 table.

1896. Muscular System. In Alfred Newton, A Dictionaryof Birds, pages 602-620. London: Adam andCharles Black.

Gardner, Leon L.1925. The Adaptive Modifications and the Taxonomic

Value of the Tongue in Birds. Proceedings of theUnited States National Museum, 67(19): 1-49, 16plates.

Garrod, A. H.1874. On Certain Muscles of the Thigh of Birds and on

Their Value in Classification, Part II. Proceedingsof the Zoological Society of London, 1874:111-123,plate 17, 2 tables.

George, J. C , and A. J. Berger1966. Avian Myology. 500 pages, illustrated. New York:

Academic Press.

Glauert, L., and C.F.G. Jenkins1931. Notes on the Banded Stilt (Cladorhynchus leucoce-

phalus), with a Description of Its Eggs. Journal of theRoyal Society of Western Australia, 17:1-6, 6 figures,2 plates.

Glenny, Fred H.1955. Modifications of Pattern in the Aortic Arch Sys-

tem of Birds and Their Phylogenetic Significance.Proceedings of the United States National Museum,104(3346):525-621, figures 108-119.

Goodman, Donald C, and Harvey I. Fisher1962. Functional Anatomy of the Feeding Apparatus in Water-

fowl (Aves: Anatidae). 193 pages, 11 figures, 40tables. Carbondale: Southern Illinois UniversityPress.

Hamilton, Robert Bruce1975. Comparative Behavior of the American Avocet

and the Black-necked Stilt (Recurvirostridae). Or-nithological Monographs (American Ornithologists'Union), 17:1-98, 18 figures, 19 tables.

Harrison, C.J.O., and C. A. Walker1976a. Cranial Material of Oligocene and Miocene Fla-

mingos: With a Description of a New Speciesfrom Africa. Bulletin of the British Museum {NaturalHistory), Geology, 27(4):305-314, 10 figures, plate6.

1976b. Birds of the British Upper Eocene. Zoological Jour-nal of the Lmnean Society, 59:323-351, 7 plates.

Hopkins, G.H.E.1942. The Mallophaga as an Aid to the Classification of

Birds. Ibis, 84:94-106.Howard, Hildegarde

1929. The Avifauna of Emeryville Shellmound, Univer-sity of California Publications in Zoology, 32(2):3O1-394, 55 figures, 4 plates.

1955. A New Wading Bird from the Eocene of Pata-gonia. American Museum Novitates, 1710:1-25, 8 fig-ures, 3 tables.

1971. Pliocene Avian Remains from Baja California. LosAngeles County Museum Contributions in Science, 217:1-17, 2 figures.

Howe, F. E., and J. A. Ross1931. Eggs of the Banded Stilt. Emu, 31(l):63-65, plates

16-17.Hudson, George E., Kenneth M. Hoff, James Vanden Berge,

and Edward C. Trivette1969. A Numerical Study of the Wing and Leg Muscles

of Lari and Alcae. Ibis, 111:459-524, 26 figures,10 tables.

Huxley, Thomas H.1867. On the Classification of Birds and on the Taxo-

nomic Value of the Modifications of Certain ofthe Cranial Bones Observable in That Class. Pro-ceedings of the Zoological Society of London, 1867:415-472, 36 figures.

NUMBER 316 71

Jehl, Joseph R, Jr.1968. Relationships in the Charadrii (Shorebirds): A

Taxonomic Study Based on Color Patterns of theDowny Young. San Diego Society of Natural HistoryMemoir, 3:1-54, colored frontispiece, 31 figures.

Jenkin, Penelope M.1957. The Filter-Feeding and Food of Flamingoes

(Phoenicopteri). Philosophical Transactions of theRoyal Society of London, series B, Biological Sciences,240(674) :401-493, 33 figures, 17 tables.

Jenkins, C.F.H.1975. Nesting of Banded Stilts at Lake Ballard. Western

Australian Naturalist, 13(4):94-95.Johnson, A. W., F. Behm, and W. R. Millie

1958. The South American Flamingos. Condor, 60(5):289-299, 3 figures.

Jones, J.1945. The Banded Stilt. Emu, 45(1): 1-36, colored fron-

tispiece, 2 figures.Kahl, M. Philip

1971. Food and Feeding Behavior of Openbill Storks.Journal fur Omithologie, 112(l):21-35, 7 figures.

1975. Ritualized Displays. In J. Kear and N. Duplaix-Hall, editors, Flamingos, pages 142-149. Berk-hamsted: T. and A. D. Poyser.

Kashin, G. N.1978. [The Replacement of a Preoccupied Generic

Name in the Class Aves.] Paleontologicheskij Zhurnal,12(1): 145-146. [In Russian; pages 131-132 in Eng-lish translation; Scripta Publishing Company,1979.]

Kear, Janet, and Nicole Duplaix-Hall, editors1975. Flamingos. 246 pages, 48 plates + 7 unnumbered

color plates, 18 tables, 9 appendices. Berkhamsted:T. and A. D. Poyser.

Kolichis, Nicholas1976. New Breeding Records of the Banded Stilt in

Western Australia. Western Australian Naturalist,13(5): 114-119, 4 figures.

Lambrecht, Kalman1931. Gallomis straeleni n. g. n. sp. ein Kreidevogel aus

Frankreich. Bulletin du Muse'e Royal d'Histoire Natu-relle de Belgique, 7(30): 1-6, 2 figures, 1 map.

1933. Handbuch der Palaeomithologie. 1024 pages, 209 fig-ures, 4 plates. Berlin: Gebruder Borntraeger.

Linnaeus, Carolus1758. Systema Naturae. 10th edition, v + 823 pages. [1956

Facsimile.] London: British Museum (NaturalHistory).

Lucas, Alfred M., and Peter R. Stettenheim1972. Avian Anatomy: Integument. Agriculture Handbook,

362 [2 volumes]; 750 pages, 422 figures, 2 append-ices.

Lydekker, Richard1891. Catalogue of the Fossil Birds in the British Museum

(Natural History). 368 pages, 75 figures. London:British Museum (Natural History).

Mainardi, Danilo1962. Immunological Data on the Phylogenetic Rela-

tionships and Taxonomic Position of Flamingos.Ibis, 104(3):426-428, 1 figure.

Makkink, G. F.1936. An Attempt at an Ethogram of the European

Avocet (Recurvirostra avosetta L.) with Ethologicaland Psychological Remarks. Ardea, 25(1-2): 1-62,16 figures.

Mayr, Ernst1957. Evolutionary Aspects of Host Specificity Among

Parasites of Vertebrates. In Premier Symposiumsur la specificite parasitaire des parasites des ver-tebres. International Union of Biological Sciences, series3,32:7-14.

McGilp, J. N., and A. M. Morgan1931. The Nesting of the Banded Stilt (Cladorhynchus

leucocephalus). South Australian Ornithologist, 11(2):37-53, 10 figures.

McGrew, Paul O., and Alan Feduccia1973. A Preliminary Report on a Nesting Colony of

Eocene Birds. In E. M. Schnell, editor, WyomingGeological Association 25th Conference Guidebook, pages163-164, 4 figures. Casper, Wyoming: Prairie Pub-lishing Company.

McNamara, James A.1976. Aspects of the Feeding Ecology of Cladorhynchus

leucocephalus (Recurvirostridae). B.S. thesis, De-partment of Zoology, University of Adelaide.

Miller, Alden H.1944. An Avifauna from the Lower Miocene of South

Dakota. University of California Publications Bulletin ofthe Department of Geological Sciences, 27(4):85-99, 8figures.

1963. The Fossil Flamingos of Australia. Condor, 65(4):289-299, 6 figures.

Miller, Waldron DeWitt1924. Further Notes on Ptilosis. Bulletin of the American

Museum of Natural History, 50:305-331.Milne-Edwards, Alphonse

1867-1871. Recherches anatomiques et paleontologiques pourservir a I'histoire des oiseaux fossiles de la France. 4volumes. Paris: Victor Masson et Fils.

Murphy, Robert Cushman1936. Oceanic Birds of South America. 2 volumes, 1245

pages, 72 plates + 16 unnumbered color plates.New York: American Museum of Natural History.

Newton, Alfred1896. A Dictionary of Birds. 1088 pages, illustrated. Lon-

don: Adam and Charles Black.Nice, Margaret Morse

1962. Development of Behavior in Precocial Birds.Transactions of the Linnaean Society of New York, 8: xii


+ 211 pages, 2 indices, 19 figures, 18 tables.Nitzsch, C. L.

1867. Nitzsch's Pterylography. English translation of Ger-man edition of 1840, P. L. Sclater, editor, 181pages, 10 plates. London: Ray Society.

Olson, Storrs L.1976. Fossil Woodcocks: An Extinct Species from Puerto

Rico and an Invalid Species from Malta. Proceed-ings of the Biological Society of Washington, 89(20):265-274, 4 figures.

1977. A Lower Eocene Frigatebird from the Green RiverFormation of Wyoming (Pelecaniformes: Fregati-dae). Smithsonian Contributions to Paleobiology, 35:1-33, 31 figures.

1978. The Nomenclatural Status of the Taxa of FossilBirds Attributed to Auguste Aymard. Proceedings ofthe Biological Society of Washington, 91(2):444-449.

1979. Multiple Origins of the Ciconiiformes. In Proceed-ings, 1978 Conference of the Colonial Waterbird Group,pages 165-170.

Olson, Storrs L., and Alan FeducciaIn press. Presbyomis and the Origin of the Anseriformes

(Aves: Charadriomorphae). Smithsonian Contribu-tions to Zoology, 323: 24 pages, 15 figures.

Palmer, Ralph S.1962. Loons through Flamingos. In Handbook of North

American Birds, 1: 567 pages, illustrated. New Ha-ven: Yale University Press.

Parkes, Kenneth C.1978. A Review of the Classification of the Ciconi-

iformes. In A. Sprunt, J. C. Ogden, and S. Winkler,editors, Wading Birds. National Audubon Society Re-search Report, 7:7-15.

Raikow, Robert J.1970. Evolution of Diving Adaptations in the Stifftail

Ducks. University of California Publications in Zoology,94:1-52, 32 figures, 16 tables.

Ray, John1678. The Ornithology of Francis Willughby. 448 pages, 78

plates. London: Royal Society.Reichenow, Anton

1877. Systematische Uebersicht der Schreitvogel (Gres-sores). Journalfur Orntthologie, 25:113-171; 225-277'.

Rooth, J.1965. The Flamingos on Bonaire (Netherlands Antilles):

Habitat, Diet and Reproduction of Phoemcopterusruber ruber. Natuurwetenschappelijke Studieknng voorSunname en de Nederlandse Antillen, Utrecht, 41:1-151,20 figures, 20 plates, 21 graphs, 18 tables.

Ross, R. C.1924. Occurrence and Behavior of Certain Shorebirds in

Southern California. Condor, 26:90 92.Rothausen, Karlheinz

1966. Fossile Flamingos im Tertiar des Mainzer Bekens.Zeitschnft Rheimshche Naturforschende GesellschaftMainz, 4:26-31, 4 figures.

Rutschke, E.1960. Untersuchungen iiber Wasserfestigkeit und Struk-

tur des Gefieders von Schwimmvogeln. ZoologischeJahrbucher (Abteilung fur Systematik, Okologie und Geo-graphie der Tiere), 87(6):441-506, 67 figures, 13tables.

Saiff, E. I.1978. The Middle Ear of the Skull of Birds: The Pele-

caniformes and Ciconiiformes. Zoological Journal ofthe Linnean Society, 63:315-370, 12 figures, 4 tables,1 appendix.

Shufeldt, Robert Wilson.1901. Osteology of the Flamingoes. Annals of the Carnegie

Museum, 1(2):295-324, plates 9-14.Sibley, Charles G.

1960. The Electrophoretic Patterns of Avian Egg-WhiteProteins as Taxonomic Characters. Ibis, 102(2):215-284, 25 figures, 2 appendices.

Sibley, Charles G., and Jon E. Ahlquist1972. A Comparative Study of the Egg White Proteins

of Non-Passerine Birds. Peabody Museum of NaturalHistory, Yale University, Bulletin, 39:1-276, 37 fig-ures.

Sibley, Charles G., Kendall W. Corbin and Joan H. Haavie1969. The Relationships of the Flamingos as Indicated

by the Egg-White Proteins and Hemoglobins. Con-dor, 71(2): 155-179, 6 figures, 3 tables.

Sibley, Charles G., and Christian Frelin1972. The Egg White Protein Evidence for Ratite Affin-

ities. Ibis, 114:377-387, 1 figure, plate 23.Stejneger, Leon hard

1885. In Aves. In J. S. Kingsley, editor, The StandardNatural History, volume 4. Boston: S. E. Cassinoand Co.

Strauch, Joseph George1978. The Phylogeny of the Charadriiformes (Aves): A

New Estimate Using the Method of CharacterCompatability Analysis. Transactions of the Zoologi-cal Society of London, 34(3):263-345, 36 figures, 2tables.

Stresemann, Erwin1959. The Status of Avian Systematics and Its Unsolved

Problems. Auk, 76(3):269-280.Tremaine, Mary M.

1975. Life at an Avocet Nest. Audubon, 77(2):68-75, 10illustrations.

Tugarinov, A. J.1940. New Data on the Tertiary Ornithofauna from the

USSR. Comptes Rendus (Doklady) de I'Academie desSciences de I'URSS, 26(3):307-3O9, 2 figures.

Tunnicliffe, G. A.1973. Preliminary Observations of the Phylogenesis of

Theogosis [Abstract]. Journal of Dental Research,53(3) :583.

Vanden Berge, James C.1970. A Comparative Study of the Appendicular Mus-

NUMBER 316 73

culature of the Order Ciconiiformes. American Mid-land Naturalist, 84(2):289-364, 10 figures.

1976. M. iliotibialis medialis and a Review of the M.iliotibiahs Complex in Flamingos. Auk, 93(3) :429-433, 1 figure.

Watson, George E.1975. Birds of the Antarctic and Sub-Antarctic. 350 pages,

illustrated. Washington: American GeophysicalUnion.

Weldon, W.F.R.1883. On Some Points in the Anatomy of Phoemcopterus

and Its Allies. Proceedings of the Zoological Society ofLondon, 1883:638-652, plates 49-50.

Wetmore, Alexander1925. Food of American Phalaropes, Avocets, and Stilts.

United States Department of Agriculture, DepartmentBulletin, 1359:1-20.

1926. Fossil Birds from the Green River Deposits ofEastern Utah. Annals of the Carnegie Museum, 16(3-4):391 -402, plates 36-37.

Zusi, Richard L.1967. The Role of the Depressor Mandibulae Muscle in

Kinesis of the Avian Skull. Proceedings of the UnitedStates National Museum, 123(3607): 1-28, 13 figures.

1974a. An Interpretation of Skull Structure in Penguins.In Bernard Stonehouse, editor, The Biology of Pen-guins, pages 59-84. London: Macmillan Press.

1974b. Charadriiformes. In Encyclopaedia Brittamca, 15thedition, 4:33-42.

Zweers, G. A.1974. Structure, Movement, and Myography of the

Feeding Apparatus of the Mallard (Anas platyrhyn-chos L.): A Study in Functional Anatomy. Nether-lands Journal of Zoology, 24(4):323-467, 90 figures, 3tables.

Zweers, G. A., A.F\C. Gerritsen, and P. J. van Kranenburg-Voogd

1977. Mechanics of Feeding of the Mallard (Anas platy-rhynchos L.; Aves, Anseriformes). In Contributions toVertebrate Evolution, 3: 109 pages, 20 figures, 6tables. Basel: S. Karger.

REQUIREMENTS FOR SMITHSONIAN SERIES PUBLICATION

Manuscripts intended for series publication receive substantive review within theiroriginating Smithsonian museums or offices and are submitted to the Smithsonian InstitutionPress with approval of the appropriate museum authority on Form SI-36. Requests forspecial treatment—use of color, foldouts, casebound covers, etc.—require, on the sameform, the added approval of designated committees or museum directors.

Review of manuscripts and art by the Press for requirements of series format and style,completeness and clarity of copy, and arrangement of all material, as outlined below, willgovern, within the judgment of the Press, acceptance or rejection of the manuscripts and art.

Copy must be typewritten, double-spaced, on one side of standard white bond paper,with i v i " margins, submitted as ribbon copy (not carbon or xerox), in loose sheets (notstapled or bound), and accompanied by original art. Minimum acceptable length is 30 pages.

Front matter (preceding the text) should include: title page with only title and authorand no other information, abstract page with author/title/series/etc., following the establish-ed format, table of contents with indents reflecting the heads and structure of the paper.

First page of text should carry the title and author at the top of the page and an unnum-bered footnote at the bottom consisting of author's name and professional mailing address.

Center heads of whatever level should be typed with initial caps of major words, withextra space above and below the head, but with no other preparation (such as all caps orunderline). Run-in paragraph heads should use period/dashes or colons as necessary.

Tabulations within text (lists of data, often in parallel columns) can be typed on the textpage where they occur, but they should not contain rules or formal, numbered table heads.

Formal tables (numbered, with table heads, boxheads, stubs, rules) should be sub-mitted as camera copy, but the author must contact the series section of the Press for edito-rial attention and preparation assistance before final typing of this matter.

Taxonomic keys in natural history papers should use the alined-couplet form in thezoology and paleobiology series and the multi-level indent form in the botany series. Ifcross-referencing is required between key and text, do not include page references within thekey, but number the keyed-out taxa with their corresponding heads in the text.

Synonymy in the zoology and paleobiology series must use the short form (taxon,author, yearpage), with a full reference at the end of the paper under "Literature Cited."For the botany series, the long form (taxon, author, abbreviated journal or book title, volume,page, year, with no reference in the "Literature Cited") is optional.

Footnotes, when few in number, whether annotative or bibliographic, should be typedat the bottom of the text page on which the reference occurs. Extensive notes must appear atthe end of the text in a notes section. If bibliographic footnotes are required, use the shortform (author/brief title/page) with the full reference in the bibliography.

Text-reference system (author/year/page within the text, with the full reference in a"Literature Cited" at the end of the text) must be used in place of bibliographic footnotes inall scientific series and is strongly recommended in the history and technology series:"(Jones, 1910:122)" or ".. . . Jones (1910:122)."

Bibliography, depending upon use, is termed "References," "Selected References," or"Literature Cited." Spell out book, journal, and article titles, using initial caps in all majorwords. For capitalization of titles in foreign languages, follow the national practice of eachlanguage. Underline (for italics) book and journal titles. Use the colon-parentheses systemfor volume/number/page citations: "10(2):5-9." For alinement and arrangement ofelements, follow the format of the series for which the manuscript is intended.

Legends for illustrations must not be attached to the art nor included within the text butmust be submitted at the end of the manuscript—with as many legends typed, double-spaced, to a page as convenient.

Illustrations must not be included within the manuscript but must be submitted sepa-rately as original art (not copies). All illustrations (photographs, line drawings, maps, etc.)can be intermixed throughout the printed text. They should be termed Figures and shouldbe numbered consecutively. If several "figures" are treated as components of a single largerfigure, they should be designated by lowercase italic letters (underlined in copy) on the illus-tration, in the legend, and in text references: "Figure 9JD." If illustrations are intended to beprinted separately on coated stock following the text, they should be termed Plates and anycomponents should be lettered as in figures: "Plate 9b_." Keys to any symbols within anillustration should appear on the art and not in the legend.

A few points of style: (1) Do not use periods after such abbreviations as "mm, ft,yds, USNM, NNE, AM, BC." (2) Use hyphens in spelled-out fractions: "two-thirds." (3)Spell out numbers "one" through "nine" in expository text, but use numerals in all othercases if possible. (4) Use the metric system of measurement, where possible, instead ofthe English system. (5) Use the decimal system, where possible, in place of fractions.(6) Use day/month/year sequence for dates: "9 April 1976." (7) For months in tabular list-ings or data sections, use three-letter abbreviations with no periods: "Jan, Mar, Jun," etc.

Arrange and paginate sequentially EVERY sheet of manuscript—including ALL frontmatter and ALL legends, etc., at the back of the text—in the following order: (1) title page,(2) abstract, (3) table of contents, (4) foreword and/or preface, (5) text, (6) appendixes,(7) notes, (8) glossary, (9) bibliography, (10) index, (11) legends.

Relationships and Evolution of Flamingos (Aves: Phoenicopteridae)

Documents

Transcript of Relationships and Evolution of Flamingos (Aves: Phoenicopteridae)