Collected Papers
in Avian Paleontology
Honoring the 90th Birthday of

Alexander Wetmore

m

A X ' .-y

RAJS*

■v.

IH|

% ■ aJl
H mm

’H XT'** "V
*M v "%.'

i. *m. . , :
s ‘<<

Tv

11

w***

; z* 1 > ISf. -

. .

1 -m

$£* r I

* tfpi

STORRS L. OLSON

EDITOR

Jfi / Ut&afc

wsflijWg

W$*m

JSrrtfjSBZu ^

w /*ar.W ,/j

gapJSP?. ; # r JV

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY • NUMBER 27

t

SERIAL PUBLICATIONS OF THE SMITHSONIAN INSTITUTION

The emphasis upon publications as a means of diffusing knowledge was expressed
by the first Secretary of the Smithsonian Institution. In his formal plan for the Insti¬
tution, Joseph Henry articulated a program that included the following statement:
“It is proposed to publish a series of reports, giving an account of the new discoveries
in science, and of the changes made from year to year in all branches of knowledge.”
This keynote of basic research has been adhered to over the years in the issuance
of thousands of titles in serial publications under the Smithsonian imprint, com¬
mencing with Smithsonian Contributions to Knowledge in 1848 and continuing with
the following active series:

Smithsonian Annals of Flight
Smithsonian Contributions to Anthropology
Smithsonian Contributions to Astrophysics
Smithsonian Contributions to Botany
Smithsonian Contributions to the Earth Sciences
Smithsonian Contributions to Paleobiology
Smithsonian Contributions to Zoology
Smithsonian Studies in History and Technology

In these series, the Institution publishes original articles and monographs dealing
with the research and collections of its several museums and offices and of professional
colleagues at other institutions of learning. These papers report newly acquired facts,
synoptic interpretations of data, or original theory in specialized fields. These pub¬
lications are distributed by mailing lists to libraries, laboratories, and other interested
institutions and specialists throughout the world. Individual copies may be obtained
from the Smithsonian Institution Press as long as stocks are available.

S. Dillon Ripley

Secretary

Smithsonian Institution

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY • NUMBER 27

Collected Papers
in Avian Paleontology
Honoring the 90th Birthday of
Alexander Wetmore

Storrs L. Olson

EDITOR

ISSUED

MAY 2 i 1978

SMITHSONIAN INSTITUTION PRESS

City of Washington
1976

ABSTRACT

Olson, Storrs L., editor. Collected Papers in Avian Paleontology Honoring the 90th Birthday
of Alexander Wetmore. Smithsonian Contributions to Paleobiology, number 27, 211 pages, 91
figures, 38 tables, 1976.—Eighteen papers covering diverse aspects of avian paleontology—from the
earliest known bird to extinct species found in Indian middens—are collected here to honor the
90th birthday of Alexander Wetmore. These are preceded by an appraisal of the current state
of avian paleontology and of Alexander Wetmore’s influence on it, including a bibliography of
his publications in this field. John H. Ostrom analyzes the hypothetical steps in the origin of
flight between Archaeopteryx and modern birds. Philip D. Gingerich confirms that Ichthyornis
and Hesperornis did indeed bear teeth, that the palate in Hesperornis is paleognathous, and that
these Cretaceous toothed birds appear to occupy a position intermediate between dinosaurs and
modern birds. Larry D. Martin and James Tate, Jr. describe the skeleton of the Cretaceous diving
bird Baptornis advenus and conclude that the Baptornithidae belong in the Hesperornithiformes,
but are less specialized than Hesperornis. Pierce Brodkorb describes the first known Cretaceous
land bird as forming a new order possibly ancestral to the Coraciiformes and Piciformes. E. N.
Kurochkin summarizes the distribution and paleoecology of the Paleogene birds of Asia, with
particular emphasis on the evolution of the gruiform families Eogruidae and Ergilornithidae.
Pat Vickers Rich and David J. Bohaska describe the earliest known owl from Paleocene deposits
in Colorado. Alan Feduccia transfers the Eocene genus Neanis from the Passeriformes to the
Piciformes and he and Larry D. Martin go on to refer this and four other genera to a new
family of Piciformes, concluding that these were the dominant perching land birds of the
Eocene of North America. Storrs L. Olson describes a new species of Todidae from the Oligocene
of Wyoming and refers the genus Protornis from the Oligocene of Switzerland to the Momotidae,
concluding that the New World Coraciiformes originated in the Old World. Charles T. Collins
describes two new species of the Eo-Oligocene genus Aegialornis and presents evidence that the
Aegialornithidae should be referred to the Caprimulgiformes rather than to the Apodiformes,
although they might be ancestral to the swifts. In the following paper he shows that the earliest
known true swifts (Apodidae) are three nominal forms from the Lower Miocene of France which
prove to be but a single species of Cypseloides, a modern genus belonging to a primitive sub¬
family now restricted to the New World. Stuart L. Warter describes a new osprey from the
Miocene of California to provide the earliest certain occurrence of the family Pandionidae and
he treats functional aspects of the evolution of the wing in Pandion. Hildegarde Howard
describes a new species of flightless mancalline auk, also from the Miocene of California, which
is temporally and morphologically intermediate between Praemancalla lagunensis and the species
of Mancalla. Robert W. Storer analyzes Pleistocene fossils of pied-billed grebes, synonymizing
Podilymbus magnus Shufeldt with modern P. podiceps and describing a new species from
peninsular Florida. Kenneth E. Campbell, Jr., lists 53 species of birds, including new species of
Buteo and Oreopholus, from a Pleistocene deposit in southwestern Ecuador and compares this
with a fauna of similar age from northwestern Peru, both of which indicate more humid condi¬
tions in the past. Oscar Arredondo summarizes aspects of the morphology, evolution, and
ecology of the gigantic owls, eagles, and vultures recently discovered in Pleistocene deposits in
Cuba. Joel Cracraft analyzes variation in the moas of New Zealand, reduces the number of
species recognized to 13, and suggests that several “species pairs” represent examples of sexual
size dimorphism. G. Victor Morejohn reports remains of the extinct flightless duck Chendytes
lawi, previously known only from Pleistocene deposits, from Indian middens in northern Cali¬
fornia and concludes that the species became extinct through human agency less than 3800
years ago.

Official publication date is handstamped in a limited number of initial copies and is recorded
in the Institution’s annual report, Smithsonian Year. SI Press number 6137. Series cover desicn:
The trilobite Phacops rana Green.

Library of Congress Cataloging in Publication Data
Main entry under title:

Collected papers in avian paleontology honoring the 90th birthday of Alexander Wetmore.
(Smithsonian contributions to paleobiology ; no. 27)

Contents: Ostrom, J. H. Some hypothetical anatomical stages in the evolution of avian flight—
Gingerich, P. D. Evolutionary significance of the Mesozoic toothed birds.—Martin, L. D. and
Tate, J., Jr. The skeleton of Baptornis advenus (Aves: Hesperornithiformes). [etc.]

Supt. of Docs. No.: SI 1.30:27

1. Birds, Fossil—Addresses, essays, lectures. I. Wetmore, Alexander, 1886- II. Olson, Storrs L.

III. Series: Smithsonian Institution. Smithsonian contributions to paleobiology ; no. 27.
QE701.S56 no. 27 [QE871] 560'8s [568'.2] 75-619322

Contents

Page

Preface . . v

Appreciations, by S. Dillon Ripley and Jean Delacour . . . .... vii

Alexander Wetmore and the Study of Fossil Birds, by Storrs L. Olson xi
Publications in Avian Paleontology by Alexander Wetmore . . . . xvii

Index to Fossil Avian Taxa Described by Alexander Wetmore .... xxv

Some Hypothetical Anatomical Stages in the Evolution of Avian

Flight, by John H. Ostrom . . 1

Evolutionary Significance of the Mesozoic Toothed Birds, by Philip

D. Gingerich . . . . 23

The Skeleton of Baptornis advenus (Aves: Hesperornithiformes), by

Larry D. Martin and James Tate, Jr. . . . . . . 35

Discovery of a Cretaceous Bird, Apparently Ancestral to the Orders
Coraciiformes and Piciformes (Aves: Carinatae), by Pierce Brodkorb 67

A Survey of the Paleogene Birds of Asia, by E. N. Kurochkin . 75

The World’s Oldest Owl: A New Strigiform from the Paleocene of
Southwestern Colorado, by Pat Vickers Rich and David J. Bohaska 87

Neanis schucherti Restudied: Another Eocene Piciform Bird, by Alan

Feduccia . . 95

The Eocene Zygodactyl Birds of North America (Aves: Piciformes),

by Alan Feduccia and Larry D. Martin . . 101

Oligocene Fossils Bearing on the Origins of the Todidae and the

Momotidae (Aves: Coraciiformes), by Storrs L. Olson .... Ill

Two New Species of Aegialornis from France, with Comments on the

Ordinal Affinities of the Aegialornithidae, by Charles T. Collins 121
A Review of the Lower Miocene Swifts (Aves: Apodidae), by Charles

T. Collins. 129

A New Osprey from the Miocene of California (Falconiformes:

Pandionidae), by Stuart L. Warter. . . 133

A New Species of Flightless Auk from the Miocene of California

(Alcidae: Mancallinae), by Hildegarde Howard . 141

The Pleistocene Pied-billed Grebes (Aves: Podicipedidae), by Robert

W. Storer .... . . 147

The Late Pleistocene Avifauna of La Carolina, Southwestern

Ecuador, by Kenneth E. Campbell, Jr. . . . 155

The Great Predatory Birds of the Pleistocene of Cuba, by Oscar

Arredondo, translated and amended by Storrs L. Olson . 169

The Species of Moas (Aves: Dinornithidae), by Joel Cracraft. 189

Evidence of the Survival to Recent Times of the Extinct Flightless

Duck Chendytes lawi Miller, by G. Victor Morejohn . 207

in

Preface

Had contributions for this volume been sought from the associates and friends
of Alexander Wetmore in all fields of ornithology, their number would have
been much too great to permit the timely appearance of this festschrift, for the
endeavor was conceived barely in time for its proper execution. It was decided,
therefore, to limit the scope of this work to avian paleontology—a study which
has been particularly dear to Alex Wetmore for three score years. That this collec¬
tion could be assembled and set before the press in less than a year is a tribute
not only to the eagerness of the contributors to honor their esteemed colleague
in his 90th year, but also to the fact that there is currently an extensive and
active interest in the study of fossil birds—a fact that must be particularly grati¬
fying to Dr. Wetmore, who for so many years strived to keep such an interest
alive. The editor is particularly indebted to Dorsey Dunn and Joanne Williams,
who typed and retyped manuscripts with great patience and care, and Anne Curtis,
who assisted in preparing numerous illustrations. He also wishes to express his
appreciation for the fine cooperation of the contributors; their combined efforts
have here produced what is certain to be a landmark in paleornithology.

v

Alexander Wetmore

Appreciations

S. Dillon Ripley

SECRETARY,

SMITHSONIAN INSTITUTION

Alexander Wetmore is so familiar a figure to scientists as the dean of American
ornithology that it is difficult to realize that he has been directly associated with
the Smithsonian Institution as an administrator since 1924. His first responsi¬
bilities were in connection with the National Zoological Park, of which he
became Superintendent in 1924. Subsequently, Dr. Wetmore became Assistant
Secretary for Science of the Institution and Director of the Museum of Natural
History in 1925, and continued as Assistant Secretary until 1945, when he was
elected by the Regents to serve as the sixth Secretary, succeeding Dr. Charles G.
Abbot, who retired in that year.

Throughout this period, and after his own retirement from administrative
responsibilities in 1952, Dr. Wetmore has continued an extraordinarily active
career in ornithology. In addition to his many duties with the Smithsonian, he
also served as Home Secretary of the National Academy of Sciences from 1951 to
1955 and has been for many years a Trustee and Vice-Chairman of the Research
Committee of the National Geographic Society.

Throughout this career his publications on birds have continued in depth and
in great volume. Following his retirement he has continued his monographic
studies on the birds of Panama, which have culminated in the publication of
three volumes of “The Birds of the Republic of Panama” ( Smithsonian Miscel¬
laneous Collections , volume 150), with a fourth part in preparation. Even now,
Dr. Wetmore’s work is not completed and he continues to be a productive scien¬
tist in the laboratory of the Division of Birds.

In addition to the many research publications on fossil material specializing
in birds. Dr. Wetmore is known today as one of the most outstanding systematic
specialists. His renowned arrangement of the sequence of higher taxa of birds,
“A Classification for the Birds of the World” ( Smithsonian Miscellaneous Col¬
lections, 1S9 (11): 1-37, 1960), still stands virtually unchallenged. He is a winner
of the Brewster Medal of the American Ornithologists’ Union, and recently, in
May 1975, of the Hubbard Medal of the National Geographic Society.

The amount of materials contributed by him to the collections of the National
Museum is monumental. Indeed, present-day ornithologists would be staggered
to think of the production of research and study material deposited by Dr. Wet¬
more in the National Collection: some 26,058 skins from North America, Puerto
Rico, Hispaniola, the Hawaiian Islands, Uruguay, Paraguay, Argentina, Chile,
Venezuela, and Central America, with more than half, some 14,291, from Panama
alone. Of skeletal and anatomical specimens, Dr. Wetmore has prepared and
contributed 4363, an enormously important increment to the anatomy collections
in Washington. The majority of these are from North America and Puerto Rico,
but nearly 1000 are from Central and South America and 540 from Panama

vii

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

alone. Of eggs, Dr. Wetmore has collected 201 clutches from North, Central, and
South America. In this day and age when the collecting of birds has become
markedly diminished due to the general knowledge of specimens in existing
museums, as well as the varying directions taken in present-day studies in environ¬
ment and ecology which tend to preclude such collecting, Dr. Wetmore’s collec¬
tions seem large in retrospect; but they form part of the fundamental resource
on which present and future work will depend. The very magnitude of these
collections would tend to make further collecting in most areas where he has
worked superfluous. So today the specialist in taxonomic studies can be grateful
for the efforts of meticulous collectors such as Dr. Wetmore, whose work has laid
out in depth representative material. Thus, only highly specific additional col¬
lecting need be done in the future in areas where Wetmore’s work has given us
the foundation of our knowledge.

The number of species and subspecies described by Dr. Wetmore is equally
impressive. Over the years since 1914 he has described as new to science some 189
species and subspecies of recent birds. Many of these, in fact most, are from
Central and northern South America, but much of Dr. Wetmore’s most signifi¬
cant early field work was done in the Caribbean, particularly in Puerto Rico,
Hispaniola, and adjacent islands in the Greater Antilles. In addition, through
the initiative of the late Dr. Casey Wood, Wetmore worked on and described a
number of species from the Fiji Islands, as well as forms from other islands of
the Pacific. His monographic revisions of a number of species of northern Central
and North American birds, as well as Argentinian and southern South American
birds, have produced many novelties for science. A great deal of his work was
done in revising the avifauna of Venezuela with the late W. H. Phelps, Sr., with
whom he co-authored a number of new species and subspecies.

At least one of Dr. Wetmore’s discoveries, the population of Chilean Pintail
found in the vicinity of Bogota, Colombia, has subsequently gone extinct, due
presumably to hunting pressure. Many of the environments in which he worked
in Colombia and adjacent parts of northern South America are already so
radically changed that one wonders whether additional forms may not have gone
extinct as well. It is a sadness of our time that the development of tropical regions
of the world, with the consequent destruction of forests and unique habitats, par¬
ticularly in South and Central America, has been so rapid that many forms of
the accompanying avifauna may never be seen again in life. In a spirit of pre¬
science, Alexander Wetmore was an early supporter of the Pan-American Section
of the International Council for Bird Preservation, having joined T. Gilbert
Pearson, Robert Cushman Murphy, Marshall McLean, William Vogt, and Hoyes
Lloyd in helping to set up the original organization with Latin American
colleagues.

Many of his admirers have named numbers of new birds after our beloved
former Secretary, among them a long-billed rail of the Venezuelan coast, Rallus
wetmorei, which I have recently considered in my own ornithological work.
Including Rallus wetmorei, some 16 modern species and subspecies of birds have
been named in honor of Alexander Wetmore, as well as 4 mammals, 7 reptiles
and amphibians, 2 fishes, 9 insects, 5 molluscs, a sponge, a cactus, a glacier, and
a canopy bridge in the Bayano River forest in Panama. Truly the incessant and
intensive zeal which he has single-mindeclly given to the study of birds over the
years, often at very considerable personal expenditure in time and energy, will
mark the career of Alexander Wetmore as one of the most memorable in the
entire history of American ornithology.

NUMBER 27

IX

Jean Delacour

DIRECTOR EMERITUS,

NATURAL HISTORY MUSEUM OF LOS ANGELES COUNTY

I had been corresponding with Alexander Wetmore for several years before
I had a chance to meet him. This I did in Washington, D.C., in the spring of
1926. Referring to a visit I made to the National Zoological Park at that time,
I wrote as follows:

. . . the National Zoological Park is managed by the Assistant Secretary of the Smithsonian
Institution, Dr. Alexander Wetmore, one of the youngest and most accomplished naturalists in
the United States. Notwithstanding his heavy administrative obligations, Dr. Wetmore finds
enough time for study in descriptive ornithology and technical work, and observations of birds
in freedom and in captivity, all with remarkable results. I visited the Zoo under his kind
guidance. . . . ( L’Oiseau , 7(1926):205).

Dr. Wetmore himself published in the same issue of that periodical (pages
324-325), a report of the first breeding in captivity at the National Zoo of the
Blue Snow Goose, with several photographic plates. He was, therefore, awarded
a special medal by the Society Nationale d'Acclimatation de France. Dr. Wetmore
was Director of the National Zoo for two years, and before he exchanged that
function for the Assistant Secretaryship he was responsible for choosing as his
successor, Dr. William Mann, who was an outstanding Zoo director for many years.

The welcome given me by Dr. Wetmore in 1926 remains vivid in my memory,
and my mother and I visited Washington under his cordial and competent
guidance. Later on, we had many opportunities of getting together at meetings
and congresses, as we have had many interests in common. We met in Europe
and in America frequently, working together for bird preservation since the
inception of the International Council for Bird Preservation. We saw even more
of each other after 1940, when I came to live in the United States.

We are now among the few ornithologists of our generation still alive. We
sadly miss many of our old friends, particularly Frank Chapman, Tom Barbour,
Robert Cushman Murphy, James Chapin and T. Gilbert Pearson, to list only a
few who worked with us on different projects. It is to me a very special comfort
to know that Alex still is here, looking and acting and writing much as he always
has, and I wish him all the happiness he deserves. As past Secretary of the
Smithsonian Institution he joins the ranks of those others who have seemed
over the years almost immortal; thus his continuing research for many years
seems assured.

Alexander Wetmore and the Study of Fossil Birds

Storrs L. Olson

In most general discussions of paleontology or ornithology, the subject of fossil
birds is almost invariably treated with a predictable uniformity. Mention is
made of Archaeopteryx and the Cretaceous toothed birds, and occasionally some
of the large Tertiary predators like Diatryma and Phorusrhacos. This is accom¬
panied by a statement explaining that bird bones are fragile and seldom pre¬
served, thus accounting for what is alleged to be a meager and uninformative
fossil record for the entire class. Through frequent repetition, this myth has
gained such general acceptance that the uninformed find it difficult to conceive
of an avian paleontologist being able to find enough to keep himself occupied.

Yet for 60 years Alexander Wetmore has produced a steady stream of papers
on fossil birds. With over 150 such entries and nearly as many new fossil taxa
to his credit, he can without reservation be said to have contributed more to this
field than any other single person. One cannot help but be humbled to think
that this is but a fraction of his total scientific output.

Bringing together this collection of papers in avian paleontology to honor
Alexander Wetmore’s 90th birthday on 18 June 1976 provides not only an
opportunity to review his influence on paleornithology over the past six decades,
but also offers a chance to begin dispelling the fiction that fossil birds are rare
and provide little information on avian evolution.

Wetmore’s most intensive work on fossil birds took place in the period after
the waning of excitement over the spectacular 19th century discoveries of Mesozoic
birds, but before most of the renewed modern interest in avian paleontology had
been sparked. For many years Wetmore was virtually the only person anywhere
who was engaged in research on fossil birds, with the notable exception of the
California school of Loye and Alclen Miller and Hildegarde Howard. Thus it
was natural that bird fossils from all parts of the United States and from areas
of the world as diverse as Inner Mongolia, Java, St. Helena, Hawaii, and Ber¬
muda, passed through Wetmore’s hands continually. To this day, the cabinets
in his office hold a rich trove of undescribed treasures from a wide array of
horizons and localities.

For many years, Wetmore has assiduously maintained an extensive card catalog
of references from which he prepared three separate editions of a checklist of
fossil birds of North America. He also endeavored to keep his colleagues abreast
of current developments in avian paleontology through numerous addresses,
lectures, and entertaining synoptic papers—all the while maintaining a consis¬
tently high level of production of basic detailed descriptions and diagnoses of
new forms.

Wetmore’s first paper on fossil birds involved removing the large Miocene bird
described by R. W. Shufeldt as Palaeochenoides miocaenus from the Anseriformes
to the Pelecaniformes. Shufeldt, whom Wetmore knew well, was in no way

Xll

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

pleased by this, but Wetmore’s action was quite correct. Specimens possibly
representing two new species of Palaeochenoid.es have recently come to the
National Museum and it now appears that these may provide a breakthrough in
our understanding of these huge, enigmatic seabirds. Wetmore’s recognition of
the true affinities of Palaeochenoides marked the first step toward this under¬
standing. Shufeldt, it might be noted, was a singular eccentric who, although
making many contributions to avian paleontology, repeatedly made serious errors
in identification. The process of re-evaluating Shufeldt’s taxa, begun by Wetmore
and others, has continued up to the present, as seen, for example, in the papers
on Eocene Piciformes elsewhere in this volume.

The first new bird Wetmore described from osteological remains was a new
genus and species of large flightless rail, Nesotrochis debooyi, found in Indian
middens in the Virgin Islands. That such deposits may still be of interest to avian
paleontologists is clearly demonstrated by Morejohn in the final paper of the
present volume. In recent years two new species of Nesotrochis have been
described from Cuba and Hispaniola; despite this, the genus remains so distinc¬
tive that there is not yet a good clue as to its affinities within the Rallidae.

Wetmore continued to draw notice to the extinct Pleistocene birds of the West
Indies, analyzing fossil avifaunas from Puerto Rico, Haiti, Cuba, and the
Bahamas. Among the most notable of his discoveries was the giant barn owl,
Tyto ostologa, of Haiti, which he correctly diagnosed from a small fragment of
tarsometatarsus. He later described a similar species, T. pollens, along with two
new large eagles, from the Bahamas. As late as 1959, Brodkorb, in dedicating
to him a new fossil species of crow from New Providence Island, remarked that
Alexander Wetmore was “responsible for all previous knowledge of fossil birds
of the West Indies.’’ Since then, there have been many additional discoveries of
avian fossils in the Antilles, the most remarkable of which are certainly the
gigantic raptors of Cuba brought to light through the labors of Oscar Arredondo
(summarized in this volume). Among the material from the same deposits that
yielded Tyto ostologa, a new rail and a new falcon have recently been found.
There is every reason to believe that the fossil resources of the Greater Antilles
will continue to produce surprises, while as far as avian paleontology is con¬
cerned, the Lesser Antilles are terra incognita.

Perhaps the greatest proportion of Wetmore’s paleontological efforts concerned
the identification and description of Tertiary birds from North America, espe¬
cially those of the Eocene, Oligocene, and Miocene terrestrial deposits of the
western states and the marine Miocene of the east coast. In these areas he has
laid the groundwork for all future researches.

Some of the most exciting recent finds of fossil birds are from the extensive
lower Eocene deposits of the Green River Formation, for these often yield com¬
plete, articulated skeletons, as for example a particularly fine specimen of primi¬
tive frigatebird now under study by the writer. Feduccia and Martin in this
volume discuss the significance of the Green River Piciformes, which are now
coming to light with remarkable rapidity since Brodkorb’s recognition of the
first species in 1970. But perhaps the most astonishing of developments in Green
River paleornithology are the tremendous deposits of flamingo bones discovered
by Paul O. McGrew and now under study by him and Alan Feduccia. Here too,
Wetmore’s past contributions have played a part, for he described this flamingo
in 1926 as a new genus of recurvirostrid, Presbyornis. This case of mistaken
identity is understandable in view of Feduccia’s further investigations, which

NUMBER 27

xiii

have disclosed some extraordinary similarities between the skeletons of recurvi-
rostrids and flamingos, particularly those of the lower Eocene forms. This is
further confirmed by an undescribed flamingo of Bridgerian age in the National
Museum which is even more similar to recurvirostrids than is Presbyornis. These
discoveries now appear to be leading to a reappraisal of the affinities of both the
flamingos and the shorebirds.

Wetmore’s several contributions on Eocene owls resulted in his erecting a
new family, the Protostrigidae, the importance of which is only now becoming
apparent. The fossil record of owls is particularly good and we now know that
the order extends back at least as far as the Paleocene (see Rich and Bohaska’s
paper in this volume). Much unstudied material of fossil owls is to be found in
various museums, which, along with the revision of the many forms already
known, should provide an especially fruitful area of inquiry for avian paleon¬
tologists in the future. Of Wetmore’s Eocene birds, perhaps the most provocative
is Neocathartes grallator, a long-legged vulture that was based on a nearly
complete skeleton.

Wetmore’s contributions once provided just about all that was known of the
birds from the extensive Oligocene deposits of western North America. These are
now producing new and extremely interesting fossil birds almost annually (e.g.,
Olson’s paper in this volume). One of the predominant groups of birds in the
North American Oligocene was the gruiform family that Wetmore named the
Bathornithidae. Wetmore himself offered more than one interpretation of the
possible relationships of this group and Cracraft has recently proposed others.
It seems certain that the final word has not been said on this matter, but the
importance of the Bathornithidae is undisputed. Once again, it was Wetmore’s
pioneering work on the group that has made possible all subsequent investiga¬
tions. It now appears that the Oligocene limpkins (Aramidae) described by
Wetmore will soon be augmented by a new genus, known from much of a skeleton
collected in Wyoming by Dr. R. J. Emry of the National Museum. Oligocene
raptors described by Wetmore include two forms inseparable from the modern
genus Buteo, and an intriguing species, Palaeoplancus sternbergi, which was made
the type of a new subfamily of Accipitridae.

For Wetmore, some of the most interesting fossil deposits were those closest
to home—the Miocene marine beds of the Chesapeake Group. Most of what we
know of the birds of these deposits is to be found in Wetmore’s publications,
including the description of a diminutive gannet, Microsula avita, which is now
known to be relatively common in these beds. In the past few years many new
specimens, some of them highly significant, have come to the National Museum
from this area, although these are as yet undescribed. As abundant as this material
is, it is far overshadowed by the tremendous collections of Miocene and Pliocene
age that have recently been acquired from a phosphate mine in North Carolina
and which this writer has had the privilege of studying in collaboration with
Dr. Wetmore. This is probably the largest deposit of Tertiary birds in existence
and thousands of fossils of more than 50 species have so far been recovered.
These collections, along with those from Bone Valley, Florida, being studied by
Brodkorb, and those from the Pacific coast, which are constantly productive (see
the contributions by Howard and Wai ter in this volume), provide a solid basis
for making unprecedented gains in our knowledge of evolution in the Alcidae,
Procellariidae, Diomedeidae, Gaviidae, Sulidae, Phalacrocoracidae, and other
families of marine birds.

XXV

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Dr. Wetmore on a Smithsonian collecting trip to the Lee Creek phosphate mine,
Aurora, North Carolina, 26 April 1972.

In 1931, Wetmore published a large paper on the Pleistocene avifauna of
Florida in which it was shown that several birds, such as the California condor
and the huge vulture Teratornis, then known only from the west, particularly
from the tarpits at Rancho la Biea, were also present in Florida. This opened
up a very fertile area of investigation and in succeeding years the studies of
Brodkorb and others have continued to be a source of new information on the
rich Pleistocene avifauna of Florida (e.g., Storer’s paper herein). In his many
years of involvement in paleornithology, Wetmore has repeatedly been called
upon to identify material from Pleistocene caves and from Indian middens, a
task which as often as not holds few rewards but which nevertheless he pursued
with alacrity. From such studies he published numerous notes showing that the
distribution of many modern North American species was once much different
than at present, as indicated, for example, by Canada Jays, Magpies, and Sharp¬
tailed Grouse in Virginia, and Spruce Grouse in Virginia and Georgia. The sum
of these observations has proved to be a significant contribution to our knowledge
of the effects of Pleistocene climatic changes on avian distribution.

When the Central Asiatic Expeditions of the American Museum of Natural
History discovered fossil birds in the Eocene of Inner Mongolia, it was to Wet¬
more that the specimens were sent for study. The most abundant material was
that of the crane-like bird which Wetmore named Eogrus aeola, assigning it to

NUMBER 27

XV

a new family, Eogruidae. Recently, the significance of these birds as the probable
ancestors of the peculiar two-toecl running birds of the family Ergilornithidae
has been demonstrated (see Kurochkin’s paper herein) and provides one of the
most interesting examples of an evolutionary lineage in the avian fossil record.

Oceanic islands are of particular interest to the avian paleontologist because
of the rapid extinction of species after the introduction of exotic predators by
man. Most such introductions occurred before the era of scientific exploration
and thus many insular species can be known only from the study of fossil or
subfossil remains. Here Wetmore has likewise made numerous contributions. In
1943 he described an extinct goose from the island of Hawaii. This turned out
to be but a small indication of what was to come, for in the past few years the
Bishop Museum has forwarded to him for examination numerous fossils from
Molokai and Maui, which comprise one of the most extraordinary avifaunas
ever uncovered, some of the species being so anomalous as to be quite beyond
the wildest imaginations of the most whimsical fantasizer. From Pleistocene
deposits on Bermuda, Wetmore described a crane and a duck, leaving to Brod-
korb the naming of five new rails from these and other deposits on the island
(as yet undescribed). From St. Helena, in the South Atlantic Ocean, Wetmore
named a new rail to provide a first step in the elucidation of the extensive fossil
avifauna of that island, which this writer has recently had the opportunity to
expand.

We have touched on but a few of Alexander Wetmore’s contributions to avian
paleontology and their importance to present and future research. It should by
now be clear that, contrary to persistent belief, fossil birds are not uncommon,
and in the following pages it should be equally evident that there is much to be
gained from their study.

At last there is some light being shed on the study of Cretaceous land birds
(see Brodkorb’s paper herein), an area that had hitherto been a void. The
renowned Pleistocene tarpits at Rancho la Brea, California, long erroneously
held to be the only really productive source of avian fossils, now find a rival in
similar deposits in South America which portend a new era of discovery on that
continent (see Campbell’s paper in the present volume). Although these many
new finds are of paramount importance, the avian paleontologist has also
inherited a rich source of information in the fossils that have been made known
previously. Re-examination of the much discussed but widely misunderstood
Mesozoic birds, such as the Jurassic Archaeopteryx and the Cretaceous toothed
divers, has generated exciting new ideas and controversy, all of which can only
lead to a better understanding of avian evolution (see the papers by Ostrom,
Gingerich, and Martin and Tate in this volume). Long-neglected fossil birds,
such as those from the vast Tertiary collections of France and from the wealth
of material in the New Zealand Quaternary, are coming under scrutiny once
again, and in the light of modern concepts find a better place in the evolutionary
scheme (see papers herein by Collins and Cracraft). It would seem, therefore,
that avian paleontology is truly experiencing a renaissance.

In 1932, Joseph Grinnell (Auk, 49:9-13) in pondering the latest edition of the
American Ornithologists’ Union’s Checklist of North American Birds, to which
Wetmore contributed the portion on fossils, attempted to make some inferences
about future lists and the number of species they might contain. Concerning the
fossil list he queried, “And what about the number and relative acumen of
future students in avian paleontology: Will they be more numerous and more

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

alert than heretofore or will the attractions in this field wane in the face of the
ascending allurements for bright minds of bio-physics, bio-chemistry and cosmic
mathematics? These questions are more or less baffling of answer.” Forty-five
years later, the answers are apparent. We offer the present volume as testimony
to the fact that avian paleontology has quite enough allure of its own to attract
numerous and perspicacious practitioners, and that the materials they study allow
of significant advances not only in the knowledge of birds, but of biology and
paleontology as a whole. The discipline that Alexander Wetmore nurtured for
sixty years is expanding and vigorous and reaping the benefits of his devotion.

Publications in Avian Paleontology
by Alexander Wetmore

1917

1. The Relationships of the Fossil Bird Palaeochenoides miocaenus. Journal of Geology, 25(6):
555-557, 1 figure.

1918

2. Bones of Birds Collected by Theodoor de Booy from Kitchen Midden Deposits in the
Islands of St. Thomas and St. Croix. Proceedings of the United States National Museum,
54(2245):513-522, plate 82. (21 November) 1

1920

3. Five New Species of Birds from Cave Deposits in Porto Rico. Proceedings of the Biological
Society of Washington, 33:77-82, plates 2-3. (30 December)

1922

4. A Fossil Owl from the Bridger Eocene. Proceedings of the Academy of Natural Sciences
of Philadelphia, 73(3):455-458, 2 figures. (6 April)

5. Bird Remains from the Caves of Porto Rico. Bulletin of the American Museum of Natural
History, 46(4):297-333, 25 figures.

6. Remains of Birds from Caves in the Republic of Haiti. Smithsonian Miscellaneous Collec¬
tions, 74(4): 1-4, 2 figures. (17 October)

1923

7. An Additional Record for the Extinct Porto Rican Quail-Dove. Auk, 40(2):324.

8. Avian Fossils from the Miocene and Pliocene of Nebraska. Bulletin of the American
Museum of Natural History, 48(12):483-507, 20 figures. (3 December)

1924

9. Fossil Birds from Southeastern Arizona. Proceedings of the United States National Museum,
64(5): 1-18, 9 figures. (15 January)

1925

10. The Systematic Position of Palaeospiza bella Allen, with Observations on Other Fossil
Birds. Bulletin of the Museum of Comparative Zoology, 67(2): 183-193, 4 figures, plates 1-4.
(May)

11. Another Record for the Genus Corvus in St. Croix. Auk, 42(3):446.

1926

12. Descriptions of Additional Fossil Birds from the Miocene of Nebraska. American Museum
Novitates, 211:1-5, 6 figures. (11 March)

13. Fossil Birds from the Green River Deposits of Eastern Utah. Annals of the Carnegie Mu¬
seum, 16(3-4):391-402, plates 36-37. (10 April)

1 Exact dates of publication, when known, are included for papers in which new taxa aTe
proposed.

xvn

xviii

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

14. Description of a Fossil Hawk from the Miocene of Nebraska. Annals of the Carnegie
Museum, 16(3-4):403-408, plate 38. (10 April)

15. Observations on Fossil Birds Described from the Miocene of Maryland. Auk, 43(4):462-468.

16. The Fossil Birds of North America. Natural History, 26(5):525-526.

17. [Abstract of] A. Wetmore, Descriptions of Additional Fossil Birds from the Miocene of
Nebraska. Biological Abstracts, 1(1):201.

18. An Additional Record for the Fossil Hawk Urubitinga enecta. American Museum Novitates,

241:1-3, 3 figures.

1927

19. Present Status of the Check-list of Fossil Birds for North America. Auk, 44(2): 179-183.

20. Fossil Birds from the Oligocene of Colorado. Proceedings of the Colorado Museum of
Natural History, 7(2): 1-13, 23 figures. (15 July)

21. [On Cygnus paloregonus from Nampa, Idaho.] Page 267 in O. P. Hay, The Pleistocene of
the Western Region of North America and Its Vertebrated Animals. Carnegie Institution
of Washington Publication, 322B.

22. A Record of the Ruffed Grouse from the Pleistocene of Maryland. Auk, 44(4):561.

23. The Birds of Porto Rico and the Virgin Islands: Colymbiformes to Columbiformes. Pages
245-406 of part 3 in volume 9 of New York Academy of Sciences, Scientific Survey of
Porto Rico and the Virgin Islands. 1 map, 16 figures, plates 55-61. The Birds of Porto Rico
and the Virgin Islands: Psittaciformes to Passeriformes. Pages 407-598 of part 4 in volume
9 of New York Academy of Sciences, Scientific Survey of Porto Rico and the Virgin Islands.

3 figures, plates 62-65. [Includes discussion and figures of fossil species.]

1928

24. Bones of Birds from the Ciego Montero Deposit of Cuba. American Museum Novitates,

301:1-5, 2 figures.

25. Additional Specimens of Fossil Birds from the Upper Tertiary Deposits of Nebraska.

American Museum Novitates, 302:1-5, 2 figures.

26. The Tibio-tarsus of the Fossil Hawk Buteo typhoius. Condor, 30(2): 149-150, figures 58-61.

27. The Systematic Position of the Fossil Bird Cyphornis magnus. (Contributions to Canadian
Palaeontology, Geological Series Number 48). Canada Department of Mines, Geological
Survey Bulletin, 49:1-4, 1 figure. (15 March)

28. Prehistoric Ornithology in North America. Journal of the Washington Academy of
Sciences, 18(6): 145-158.

29. The Short-tailed Albatross in Oregon. Condor, 30(3): 191.

30. [List of Aves.] Page 3 in G. G. Simpson, Pleistocene Mammals from a Cave in Citrus
County, Florida. American Museum Novitates, 328.

1929

31. [Abstract of] J. F. van Bemmelen, Animaux disparus. Biological Abstracts, 3(1—3):390.

32. [Abstract of] W. v. Szeliga-Mierzeyewski, Der diluviale Kernbeisser ( Loxia coccothraustes
L.) aus Starunia in Polen (Anatomie und Histologie). Biological Abstracts, 3(1-3): 1012.

33. Birds of the Past in North America. Pages 377-389 in Smithsonian Report for 192S. 11
plates. Washington, Government Printing Office.

1930

34. The Fossil Birds of the A. O. U. Check-list. Condor, 32(1): 12-14, 1 table.

35. [and H. T. Martin.] A Fossil Crane from the Pliocene of Kansas. Condor, 32(1):62—63,
figures 23-25. (20 January)

36. [Abstract of] G. Archey. On a Moa Skeleton from Amodes Bay and some Moa Bones from
Karamu. Biological Abstracts, 4(l):287-288.

37. The Age of the Supposed Cretaceous Birds from New Jersey. Auk, 47(2): 186-188.

38. [Abstract of] M. D. d. Saez, Las Aves Corredoras Fdsiles del Santacru Cense [sic]. Biological
Abstracts, 4(3):992.

39. Two Fossil Birds from the Miocene of Nebraska. Condor, 32(3): 152-154, figures 51-56. (15
May)

40. Fossil Bird Remains from the Temblor Formation near Bakersfield, California. Proceedings
of the California Academy of Sciences, series 4, 19(8):85-93, 7 figures. (15 July)

41. The Supposed Plumage of the Eocene Diatryma. Auk, 47(4):579-580.

1931

42. [and B. H. Swales.] The Birds of Haiti and the Dominican Republic. United States Na¬
tional Museum Bulletin, 155:1^83, 2 figures, 26 plates. [Includes discussion of fossils.]

43. The California Condor in New Mexico. Condor, 33(2):76-77.

44. The Avifauna of the Pleistocene in Florida. Smithsonian Miscellaneous Collections, 85(2):
1-41, 16 figures, 6 plates. (13 April)

45. Two Primitive Rails from the Eocene of Colorado and Wyoming. Condor, 33(3): 107-109,
figures 21-29. (15 May)

46. [Report on Birds Found in a Limestone Urn at Chichen ItzA.] Page 189 in volume 1 of
E. H. Morris, J. Chariot, and A. A. Morris, The Temple of the Warriors at Chichen ItzA,
Yucatan. Carnegie Institution of Washington Publication, 406.

47. The Pleistocene Avifauna of Florida. Pages 479^183 in Proceedings of the VHth Inter¬
national Ornithological Congress at Amsterdam 1930.

48. The Fossil Birds of North America. Pages 401—472 in Check-list of North American Birds.
Fourth edition. Lancaster, Pennsylvania: American Ornithologists’ Union.

49. Bones of the Great Horned Owl from the Carlsbad Cavern. Condor, 33(6):248-249.

50. Record of an Unknown Woodpecker from the Lower Pliocene. Condor, 33(6):255-256.

1932

51. Additional Records of Birds from Cavern Deposits in New Mexico. Condor, 34(3): 141-142.

52. The Former Occurrence of the Mississippi Kite in Ohio. Wilson Bulletin, 44(2): 118.

1933

53. [and H. Friedmann.] The California Condor in Texas. Condor, 35(1):37—38.

54. A Fossil Gallinaceous Bird from the Lower Miocene of Nebraska. Condor, 35(2):64-65. (17
March)

55. Status of the Genus Geranoaetus. Auk, 50(2):212.

56. A Second Specimen of the Fossil Bird Bathornis veredus. Auk, 50(2):213—214.

57. Fossil Bird Remains from the Eocene of Wyoming. Condor, 35(3): 115-118, figure 22 (15
May)

58. Bird Remains from the Oliocene Deposits of Torrington, Wyoming. Bulletin of the
Museum of Comparative Zoology, 75(7):297—311, 19 figures. (October)

59. Development of Our Knowledge of Fossil Birds. Pages 231-239 in Fifty Years’ Progress of
American Ornithology 1883-1933. Lancaster, Pennsylvania: American Ornithologists' Union.

60. The Status of Minerva antiqua, Aquila ferox, and Aquila lydekkeri as Fossil Birds. Ameri¬
can Museum Novitates, 680:1-4, 1 figure. (4 December)

61. An Oligocene Eagle from Wyoming. Smithsonian Miscellaneous Collections, 87(19): 1-9, 19
figures. (26 December)

62. Pliocene Bird Remains from Idaho. Smithsonian Miscellaneous Collections, 87(20): 1-12, 8
figures. (27 December)

1934

63. [and E. C. Case.] A New Fossil Hawk from the Oligocene Beds of South Dakota. Contri¬
butions from the Museum of Paleontology, University of Michigan, 4(8): 129-132, 1 plate.
(15 January)

64. A Fossil Quail from Nebraska. Condor, 36(1):30, figure 5. (15 January)

65. [Review of] K. Lambrecht, Handbuch der Palaeornithologie. Auk, 51(2):261—263.

66. Fossil Birds from Mongolia and China. American Museum Novitates, 711:1-16, 6 figures.
(7 April)

67. The Types of the Fossil Mammals Described as Aquila antiqua and Aquila ferox. Journal
of Mammalogy, 15(3):251.

XX

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

1935

68. On the Genera Oligocorax and Miocorax. Auk, 52(l):75-76.

69. The Mexican Turkey Vulture in the United States. Condor, 37(3): 176.

70. The Common Loon in the Florida Keys. Auk, 52(3):300.

71. Pre-Columbian Bird Remains from Venezuela. Auk, 52(3):328-329.

72. A Record of the Trumpeter Swan from the Late Pleistocene of Illinois. Wilson Bulletin,
47(3):237.

73. Aves (Birds). Pages 275-277 in C. B. Schultz and E. B. Howard, The Fauna of Burnet
Cave, Guadalupe Mountains, New Mexico. Proceedings of the Academy of Natural Sciences
of Philadelphia, 87:273-298.

1936

74. The Range of the Sharp-tailed Grouse in New Mexico. Condor, 38(2):90.

75. How Old Are Our Birds? Bird-Lore, 38(5): 321-326, 7 figures.

76. Two New Species of Hawks from the Miocene of Nebraska. Proceedings of the United
States National Museum, 84(3003):73-78, figures 13-14. (3 November)

1937

77. The Eared Grebe and Other Birds from the Pliocene of Kansas. Condor, 39(1):40.

78. Ancient Records of Birds from the Island of St. Croix with Observations on Extinct and
Living Birds of Puerto Rico. Journal of Agriculture of the University of Puerto Rico,
21(1):5-16, 1 plate. (January)

79. The Systematic Position of Bubo leptosteus Marsh. Condor, 39(2): 84-85, figure 23.

80. Bird Remains from Cave Deposits on Great Exuma Island in the Bahamas. Bulletin of
the Museum of Comparative Zoology, 80(12):427-T41, 16 figures, 1 plate. (October)

81. The Tibiotarsus of the Fossil Bird Bathornis veredus. Condor, 39(6):256-257, figure 70.

82. A Record of the Fossil Grebe, Colymbus parvus, from the Pliocene of California, with
Remarks on Other American Fossils of This Family. Proceedings of the California Acad¬
emy of Sciences, series 4, 23(13): 195-201, 15 figures.

1938

83. A Miocene Booby and Other Records from the Calvert Formation of Maryland. Proceedings
of the United States National Museum, 85(3030):21-25, figures 2-3. (14 January)

84. Another Fossil Owl from the Eocene of Wyoming. Proceedings of the United States Na¬
tional Museum, 85(3031):27-29, figures 4-5. (17 January)

85. Bird Remains from the West Indies. Auk, 55(1):51—55.

86. A Fossil Duck from the Eocene of Utah. Journal of Paleontology, 12(3):280-283, 5 figures.
(4 May)

1939

87. A Pleistocene Egg from Nevada. Condor, 41(3):98-99, figure 29.

88. [On Marsh’s Discovery of Toothed Birds.] Page 48 in C. Schuchert, Biographical Memoir
of Othniel Charles Marsh. National Academy of Sciences of the United States of America
Biographical Memoirs, 20(1): 1-78.

1940

89. Fossil Bird Remains from Tertiary Deposits in the United States. Journal of Morphology,
66(l):25-37, 14 figures. (2 January)

90. A Check-list of the Fossil Birds of North America. Smithsonian Miscellaneous Collections,
99(4): 1-81.

91. Avian Remains from the Pleistocene of Central Java. Journal of Paleontology, 14(5):447-
450, 7 figures. (1 September)

1941

92. An Unknown Loon from the Miocene Fossil Beds of Maryland. Auk, 58(4):567.

NUMBER 27

XXI

1942

93. Two New Fossil Birds from the Oligocene of South Dakota. Smithsonian Miscellaneous
Collections, 101(14): 1-6, 13 figures. (11 May)

1943

94. Evidence for the Former Occurrence of the Ivory-billed Woodpecker in Ohio. Wilson
Bulletin, 55(1):55.

95. Remains of a Swan from the Miocene of Arizona. Condor, 45(3): 120.

96. Fossil Birds from the Tertiary Deposits of Florida. Proceedings of the New England Zoo¬
logical Club, 32:59-68, plates 11-12. (23 June)

97. The Little Brown Crane in Ohio. Wilson Bulletin, 55(2): 127.

98. The Occurrence of Feather Impressions in the Miocene Deposits of Maryland. Auk, 60(3):
440^41.

99. [Review of] L. Miller and I. DeMay, The Fossil Birds of California. Auk, 60(3):458-459.

100. An Extinct Goose from the Island of Hawaii. Condor, 45(4): 146-148, figure 39. (23 July)

101. A Second Specimen of the Fossil Guillemot, Miocepphus. Auk, 60(4):604.

102. Two More Fossil Hawks from the Miocene of Nebraska. Condor, 45(6):229-231, figures
62-63. (8 December)

1944

103. A New Terrestrial Vulture from the Upper Eocene Deposits of Wyoming. Annals of the
Carnegie Museum, 30:57-69, 10 figures, 5 plates. (24 May)

104. Remains of Birds from the Rexroad Fauna of the Upper Pliocene of Kansas. University
of Kansas Science Bulletin, 30(pt. 1, no. 9):89-105, 19 figures. (15 May)

1945

105. A Further Record for the Double-crested Cormorant from the Pleistocene of Florida. Auk,
62(3):459.

106. Record of the Turkey from the Pleistocene of Indiana. Wilson Bulletin, 57(3):204.

107. From My Cave Notebooks. Bulletin of the National Speleological Society, 7:1-5.

1948

108. A Pleistocene Record for Mergus merganser in Illinois. Wilson Bulletin, 60(4):240.

1949

109. Archaeopteryx. Pages 260-262 in volume 2 of Encyclopaedia Britannica. 2 figures.

110. Diatryma. Page 324 in volume 7 of Encyclopaedia Britannica.

111. Hesperornis. Pages 530-531 in volume 11 of Encyclopaedia Britannica.

112. Ichthyornis. Page 58A in volume 12 of Encyclopaedia Britannica.

113. Odontornithes. Page 707 in volume 16 of Encyclopaedia Britannica.

114. Phororhacos. Pages 778-779 in volume 17 of Encyclopaedia Britamiica. 1 figure.

115. The Pied-billed Grebe in Ancient Deposits in Mexico. Condor, 51 (3): 150.

1950

116. A Correction in the Generic Name for Eocathartes grallator. Auk, 67(2):235. (28 April)

1951

117. The Original Description of the Fossil Bird Cryptornis antiquus. Condor, 53(3): 153.

118. A Revised Classification for the Birds of the World. Smithsonian Miscellaneous Collec¬
tions, 117(4):3. (1 November)

xxii

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

1952

119. Presidential Address. Recent Additions to Our Knowledge of Prehistoric Birds 1933-1949.
Pages 51-74 in Proceedings of the Xth International Ornithological Congress Uppsala June
1950.

120. A Record for the Black-capped Petrel, Pterodroma hasitata, in Martinique. Auk, 69(4):460.

1955

121. The Genus Lophodytes in the Pleistocene of Florida. Condor, 57(3): 189.

122. A Supposed Record of a Fossil Cormorant. Condor, 57(6):371.

123. Paleontology. Pages 44-56 in A. Wolfson, editor. Recent Studies in Avian Biology. Urbana:
University of Illinois Press.

1956

124. A Check-list of the Fossil and Prehistoric Birds of North America and the West Indies.
Smithsonian Miscellaneous Collections, 131 (5): 1-105.

125. A Fossil Guan from the Oligocene of South Dakota. Condor, 58(3):234-235, 1 figure. (23
May)

126. Footprint of a Bird from the Miocene of Louisiana. Condor, 58(5):389-390, 1 figure.

127. The Muscovy Duck in the Pleistocene of Panamd. Wilson Bulletin, 68(4):327.

1957

128. A Fossil Rail from the Pliocene of Arizona. Condor, 59(4):267-268, 1 figure. (23 July)

1958

129. Miscellaneous Notes on Fossil Birds. Smithsonian Miscellaneous Collections, 135(8): 1-11, 5
plates. (26 June)

1959

130. Birds of the Pleistocene in North America. Smithsonian Miscellaneous Collections, 138(4):
1-24.

131. Notes on Certain Grouse of the Pleistocene. Wilson Bulletin, 71(2): 178-182, 1 table, 4
figures.

1960

132. A Classification for the Birds of the World. Smithsonian Miscellaneous Collections, 139(11):
4. (23 June)

133. Pleistocene Birds in Bermuda. Smithsonian Miscellaneous Collections, 140(2): 1—11, 3 plates.

0 July)

134. [and K. C. Parkes.] Archaeornithes. Pages 510-511 in volume 1 of McGraw-Hill Encyclo¬
pedia of Science and Technology.

135. Aves Fossils. Pages 694-695 in volume 1 of McGraw-Hill Encyclopedia of Science and
Technology. 1 figure.

136. [and K. C. Parkes.] Diatrymiformes. Page 104 in volume 4 of McGraw-Hill Encyclopedia
of Science and Technology. 1 figure.

137. [and K. C. Parkes.] Dinornithiformes. Page 108 in volume 4 of McGraw-Hill Encyclopedia
of Science and Technology.

138. Hesperornis. Pages 426-427 in volume 6 of McGraw-Hill Encyclopedia of Science and
Technology.

139. Ichthyornithes. Page 8 in volume 7 of McGraw-Hill Encyclopedia of Science and Tech¬
nology.

NUMBER 27

XX111

1962

140. Notes on Fossil and Subfossil Birds. Smithsonian Miscellaneous Collections, 145(2): 1-17,
2 figures. (26 June)

141. Birds. Pages 92, 95 in J. E. Guilday, The Pleistocene Local Fauna of the Natural Chim¬
neys, Augusta County, Virginia. Annals of the Carnegie Museum, 36(9):87—122.

142. Ice Age Birds in Virginia. Raven, 33(4):3.

1963

143. An Extinct Rail from the Island of St. Helena. Ibis, 103b(3):379—381, plate 9. (1 Sep¬
tember)

1964

144. [List of Aves.] Page 134 in J. E. Guilday, P. S. Martin, and A. D. McCrady, New Paris No.
4: A Pleistocene Cave Deposit in Bedford County, Pennsylvania. Bulletin of the National
Speleological Society, 26(4): 121-194.

1965

145. [Aves.] Pages 71-72 in volume 1 of L. S. B. Leakey, Olduvai Gorge 1951-61. Cambridge:
University Press.

1967

146. Pleistocene Aves from Ladds, Georgia. Bulletin of the Georgia Academy of Science, 25(3):
151-153, 1 figure.

147. Re-creating Madagascar’s Giant Extinct Bird. National Geographic, 132(4):488^493, 7
figures.

1968

148. [With C. E. Ray, D. H. Dunkle, and P. Drez.] Fossil Vertebrates from the Marine Pleisto¬
cene of Southeastern Virginia. Smithsonian Miscellaneous Collections, 153(3): 1-25, 2 figures,
2 plates.

149. Archaeopteryx. Pages 284-285 in volume 2 of Encyclopaedia Britannica, 2 figures.

150. Diatryma. Page 370 in volume 7 of Encyclopaedia Britannica.

151. Hesperornis. Pages 461^62 in volume 11 of Encyclopaedia Britannica.

152. Ichthyornis. Page 1055 in volume 11 of Encyclopaedia Britannica.

153. Phororhacos. Page 911 in volume 17 of Encyclopaedia Britannica. 1 figure.

1972

154. [Review of] G. G. Simpson, A Review of the Pre-Pliocene Penguins of New Zealand.
Quarterly Review of Biology, 47(1)78-79.

155. A Pleistocene Record for the White-winged Scoter in Maryland. Auk, 90(4):910-911.

Index to Fossil Avian Taxa Described
by Alexander Wetmore

The status of a number of these taxa has changed since their original descrip¬
tion and therefore only an alphabetical arrangement is attempted here. Species
are listed in the genera in which they were originally described. Taxa marked
with an asterisk are preoccupied and no longer available. Following each name
is the publication number (from the preceding bibliography) and page in which
the name was proposed.

Superfamilies, Families, and Subfamilies

Bathornithidae, 58:301

Eonessinae, 85:280

Palaeotringinae, 90:57

Bathornithinae, 20:13

Gaviellinae, 89:30

*Plegadornithidae, 140:3

Cladornithes, 132:4

Geranoididae, 57:115

♦Plegadornithoidea, 140:3

Cyphornithidae, 27:4

Nautilornithinae, 13:394

Presbyornithidae, 13:396

Eleutherornithidae, 118:3

Neocathartidae, 116:235

Protostrigidae, 60:4

•Eocatharddae, 103:69

Neocathartoidea, 116:235

Rhegminornithidae, 96:60

•Eocathartoidea, 103:69

Palaeoplancinae, 61:4

Telecrecinae, 65:14

Eogruidae, 65:30

Aphanocrex, 143:379

Palaeospizidae, 10:190

Genera and Subgenera

Gnotornis, 93:1

Palaeostruthus, 10:192

Aramornis, 12:1

Microsula, 83:25

Paractiornis, 39:153

BacListornis, 89:30

Miocepphus, 89:35

Phasmagyps, 20:30

Baeopteryx, 133:6

Nautilornis, 13:392

* Plegadornis, 140:1

Bathornis, 20:11

Neocathartes, 106:235

Presbychen, 40:92

Calohierax, 80:428

Nesotrochis, 2:516

Presbyornis, 13:396

*Eocathartes, 103:58

Palaealectoris, 39:152

Promilio, 129:3

Eocrex, 45:107

Palaeastur, 102:230

Protostrix, 60:3

Eogrus, 65:3

Palaeocrex, 20:9

Rhegminomis, 96:61

Eonessa, 85:280

Palaeogyps, 20:5

Telecrex, 65:13

Gaviella, 89:28

Palaeonossax, 125:234

Titanohierax, 80:430

Geochen, 100:146

Geranoides, 57:115

abavus, Presbychen, 40:92

Palaeoplancus, 61:1
Palaeorallus, 45:108

Species

avus, Nautilornis, 13:392

cooki, Cyrtonyx, 64:30

aeola, Eogrus, 65:30

brodkorbi, Promilio, 129:4

cursor, Bathornis, 58:310

ales, Geranoaetus, 14:403

bunkeri, Nettion, 104:92

debooyi, Nesotrochis, 2:516

anaticula, Eonessa, 85:280

calobates, Rhegminomis

effera, Proictinia, 8:504

antecessor, Plegadornis, 140:1

96:61

enecta, Urubitinga, 8:500

antecursor, Buteo, 58:298

cavatica, Tyto, 3:80

epileus, Prornilio, 129:4

anthonyi, Gallinago, 3:78

celeripes, Bathornis, 58:302

eversa, Dendrocygna, 9:3

aramiellus, Gnotornis, 93:1

concinna, Gavia, 89:25

fax, Palaeocrex, 20:9

aramus, Badistornis, 89:30

conterminus, Geranoaetus,

fratercula, Conuropsis, 12:3

atavus, Palaeastur, 102:230

8:487

geographies, Bathornis, 93:3

autochthones, Ara, 78:12

contortus, Geranoaetus,

gloveralleni, Titanohierax,

avita, Sula, 83:22

8:492

79:431

XXV

XXVI

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

grallator, Eocathartes,

103:58

grangeri, Buteo, 63:129
grangeri, Telecrex, 65:13
halieus, Pelecanus, 62:3
hesternus, Micropalama, 9:11
hibbardi, Colinus, 104:96
howardae, Palaeoborus, 75:73
inceptor, Puffinus, 40:86
incertus, Palaealectoris,

39:152

jepseni, Geranoides, 57:115
larva, Oreopeleia, 3:79
latebrosus, Polyborus, 3:77
latipes, Baeopteryx, 133:6
longurio, Aramornis, 12:1
mcclungi, Miocepphus, 89:35
micula, Chloroenas, 9:13
mimica, Protostrix, 84:27

minuscula, Branta, 9:6
nannodes, Grus, 35:62
ostologa, Tyto, 6:2
pachyscelus, Anas, 133:2
palaeodytes, Gavia, 96:64
patritus, Phasmagyps, 20:3
perpusillus, Paractiornis,
39:153

pervetus, Presbyornis, 13:396
phengites, Ortalis, 8:487
phillipsi, Rallus, 128:267
podarces, Aphanocrex,
143:379

pollens, Tyto, 80:436
prenticei, Rallus, 104:99
pressa, Chen, 62:9
primus, Eocrex, 45:107
proavitus, Nautilornis,

13:394

prodromus, Palaeogyps, 20:5
pumilus, Corvus, 3:81
quadratus, Calohierax,

80:429

ramenta, Falco, 76:75
rhuax, Geochen, 100:146
saurodosis, Minerva, 4:455
senectus, Palaeonossax,

125:234

sternbergi, Palaeoplancus,

61:1

tantala, Ortalis, 54:64
titan, Leptotilos, 91:447
tridens, Meleagris, 44:33
troxelli, Palaeorallus, 45:108
typhoius, Buteo, 8:489
vagabundus, Moris, 40:89
veredus, Bathornis, 20:11
vetustus, Neophrontops, 102:229

Collected Papers in Avian Paleontology

Some Hypothetical Anatomical Stages
in the Evolution of Avian Flight

John H. Ostrom

ABSTRACT

The five known skeletal specimens of Archae¬
opteryx provide the only presently available ana¬
tomical evidence pertaining to the earliest stages
in the evolution of the avian flight apparatus. This
evidence, together with the osteology of modern
birds, makes possible the reconstruction of some
hypothetical anatomical stages that must have
occurred during the course of avian evolution. It
is postulated that one of the most critical compo¬
nents of the flight apparatus is the coracoid. Evo¬
lutionary changes in coracoid morphology elevated
the actions of the principal humeral extensor (M.
coracobrachialis) and forearm flexor (M. biceps),
and as a consequence, caused deflection of the
course of the M. supracoracoideus, converting it
from a humeral depressor to a wing elevator. These
changes appear to have been related to predation
and feeding activities in the earliest birds, rather
than to early stages of flight. Subsequently, addi¬
tional changes in the forelimb components pro¬
vided for restricted elbow and wrist movements,
compact folding of the forelimb, and more stable
support of the remiges. These last changes appear
to have taken place after the acquisition of incip¬
ient flight capability.

Introduction

One of the most remarkable of all animal adap¬
tations is that of flight, which perhaps has reached

John H. Ostrom, Department of Geology and Geophysics and
the Peabody Museum of Natural History, Yale University,
New Haven, Connecticut 06520.

its zenith among vertebrates in the diverse kinds of
flight displayed by modern birds. Strangely enough,
there have been only a few investigations or specu¬
lations about the origins of avian flight, but per¬
haps that stems from the clear logic (Bock, 1965,
1969) of the currently favored'arboreal theory of
flight origins (Marsh, 1880). The purpose of this
paper, however, is not to explore that particular
question, which I have already reviewed elsewhere
(Ostrom, 1974), but rather it is to present purely
theoretical reconstructions of some of the anatom¬
ical stages that must have occurred during the
course of evolution of the avian flight apparatus,
and to discuss the implications thereof.

Reconstruction of such hypothetical evolution¬
ary stages is speculative to be sure, but it is a fruit¬
ful exercise in this instance because we know the
nature of the starting point, the almost non-bird
Archaeopteryx (Figure 1), as well as the “end
point,” the highly perfected flight apparatus of
modern birds. A few authors (Heptonstall, 1970;
Yalden, 1970) have investigated the possible flight
capabilities of Archaeopteryx, but apparently no
one has examined in any detail the anatomical
changes that clearly must have occurred in the
flight apparatus between the Archaeopteryx stage
and that of modern birds. In the absence of any
recognized intermediate stages within the avian
fossil record, consideration of these necessary ana¬
tomical changes assumes major significance, since
they may very well provide the only possible clues
about early selective factors that led to the develop-

1

Ficure 1.—The Berlin specimen of Archaeopteryx lithographica found in 1877 near Eichstatt,
Germany, in the Late Jurassic Solnhofen Limestones. Preservation of feather impressions, showing
remarkably fine structural details, established these as the remains of a true bird, despite the
fact that the skeletal anatomy is more like that of theropod dinosaurs than that of modern
birds. (The scale is 100 mm long.)

NUMBER 27

3

ment of powered avian flight. Conceivably, such
considerations might even shed light on the actual
beginnings of flight.

A premise that is critical for the remarks that
follow is that the several specimens of Archae¬
opteryx represent an extremely primitive stage in
the evolution of birds (Ostrom, 1973, 1975). (I also
believe that Archaeopteryx represents a preflight
stage [Ostrom, 1974], but not everyone concurs
with such an interpretation.) Some authors (de
Beer, 1954; Swinton, 1960, 1964) have maintained
that Archaeopteryx was not in the main lineage of
avian evolution, but so far not one single bit of
evidence has been found, either in the known speci¬
mens of Archaeopteryx or elsewhere, to support
such a contention. Indeed, as Simpson (1946) ob¬
served, Archaeopteryx is anatomically intermediate
between reptiles and modern birds, and regardless
of whether it is directly ancestral to modern cari-
nates, it is entirely reasonable to assume that the
early main-line ancestry of birds included an ana¬
tomical stage comparable, if not identical, to that of
Archaeopteryx. Thus, any consideration of the
evolution of avian flight must start with
Archaeopteryx.

Acknowledgments. —I gratefully acknowledge
the assistance and courtesies of A. J. Charig of the
British Museum (Natural History), London; H.
Jaeger of the Humboldt Museum fur Naturkunde,
East Berlin; T. Kress of the Solenhofer Actien-
Verein, Solnhofen, Bavaria; C. O. van Regteren
Altena of Teyler’s Stichting, Haarlem; and P.
Wellnhofer of Bayerische Staatssammlung fur Pala-
ontologie und historische Geologie, Munich, who
granted me the privilege of studying the speci¬
mens of Archaeopteryx in their care. I am also in¬
debted to Walter Bock, who read an early version
of the manuscript and offered valuable suggestions
and criticisms. These studies were funded by grants
from the Frank M. Chapman Memorial Fund of
the American Museum of Natural History, and the
John T. Doneghy Fund of the Yale Peabody
Museum.

Flight Apparatus of Modem Birds

By way of introduction to this section, certain
generalized comparisons among higher vertebrates
may be useful. In modern quadrupedal reptiles, the
proximal components of both the fore and hind

limbs extend laterally from the hip and shoulder
joints (sprawling posture), which are situated well
below the level of the vertebral column. In quad¬
rupedal mammals, both appendages are normally
positioned in near-parasagittal orientation (up¬
right posture) articulating with hip and shoulder
sockets that are close to the level of the vertebral
column. In birds, the hip and shoulder sockets are
both elevated and lie in or near the plane of the
vertebrae. But birds are peculiar in that the hind
limb projects downward in a nearly parasagittal
orientation, whereas the forelimb extends out lat¬
erally from the body. These contrasting limb
orientations in birds obviously are correlated with
the different limb movements in the two modes of
avian locomotion: terrestrial locomotion by means
of alternating (or synchronous) longitudinal limb
excursion in the hind quarters, and powered flight
by means of complex, but chiefly synchronous
(nonalternating) dorsoventral transverse move¬
ments of the forelimbs.

The avian skeleton includes a number of spe¬
cializations that are directly or indirectly involved
with powered flight: (1) The trunk region is quite
rigid due to fusion or restricted articular freedom
of the thoracic vertebrae, the solid bony connec¬
tion between the vertebral column and the sternum,
by full ossification of the ventral (sternal) as well
as the dorsal ribs, and the development of uncinate
processes on the dorsal ribs. (2) Fixation of the
shoulder joints by means of elongation of the
coracoids which have developed solid bony articu¬
lations with a fully ossified sternum; fusion of the
clavicles into a single median strut, the furcula,
which appears to function as a spring-like spacer
maintaining proper transverse spacing of the
shoulder joints. (3) Complete ossification and en¬
largement of the sternum and the development of
a deep and robust sternal keel. (4) Modification of
the forelimb skeleton into a rigid but collapsible
airfoil support in which the shoulder joint permits
humeral movements in nearly all directions (in¬
cluding limited long-axis humeral rotation), but
the elbow and wrist joints are restricted so as to
confine forearm flexion and extension chiefly to
the plane of the wing, wrist movements being
limited to flexion and extension in the wing plane
only; fusion of some carpals and metacarpals to
provide a solid platform for the attachment of the
primary remiges; and reduction of the manus to

4

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

digits I, II and III, with II and III united into a
relatively rigid structure. Elsewhere (Ostrom, in
press), I have given reasons for discounting the
suggestion by some authors (Holmgren, 1955) that
the remaining digits of the hand are II, III and IV.
(5) The caudal vertebrae are reduced in number
and coalesced into a short pygostyle, providing a
firmer and more readily controlable base of attach¬
ment for the tail feathers. (6) Of particular impor¬
tance is the great development of the coracoids and
furcula, which are constructed so as to prevent the
shoulder sockets from being pulled downward or
squeezed toward the midline by the powerful con¬
tractions of the flight muscles that originate on the
sternum.

In addition to these skeletal specializations, the
pectoral and forelimb musculature of carinates
have also been highly modified from the primitive
tetrapod condition, to the extent that in some in¬
stances homologies are very much in doubt. For¬
tunately, the establishment of homologies is not
critical for the theoretical reconstructions and in¬
terpretations that follow here. The flight muscula¬
ture of modern carinates has been studied and
described by many authorities, among them Strese-
mann (1933), Sy (1936), Fisher (1946), Hudson
and Lanzillotti (1955), Berger (1960), and George
and Berger (1966). From these studies, we may
classify the flight muscles in six broadly functional
categories as follows: (1) those that fix or adjust
the pectoral girdle and the shoulder socket; (2)
those that power the wing, producing the propul¬
sive down stroke; (3) those producing the recovery
stroke of the wing; (4) the flexors, for folding the
wing; (5) the extensors, for unfolding the wing,
and (6) the muscles that produce minor adjust¬
ments of the wing components, including the
remiges. Some thoracic and appendicular muscles
are involved in two or more of these actions. The
following tabulation summarizes the principal
muscles in each of these generalized categories. In
the discussion that follows, the emphasis will be on
those muscles that are concerned with the power
and recovery strokes of the wing, not because other
muscles are less important, but because these are
more conspicuously involved in the evolutionary
changes that occurred between Archaeopteryx and
later birds.

Shoulder Joint

Rhomboideus superficialis
Rhomboideus profundus
Serratus superficialis
anterior

Fixors and Adjustors

Serratus superficialis
posterior

Serratus profundus
Sternocoracoideus

Flight Muscles

Pectoralis superficialis

Wing Recovery Muscles

Supracoracoideus Deltoideus major posterior

Coracobrachialis anterior Deltoideus minor

Deltoideus major anterior

Wing Folders

Latissimus dorsi anterior
Latissimus dorsi posterior
Scapulohumeralis anterior
Scapulohumeralis posterior
Coracobrachialis posterior
Subcoracoideus
Subscapularis

Biceps brachii
Brachialis

Flexor carpi ulnaris
Flexor digitorum sublimis
Flexor digitorum profundus
Supinator

Wing Unfolders

Coracobrachialis anterior Deltoideus major anterior
Triceps brachii Deltoideus major posterior

Extensor metacarpi radialis Deltoideus minor
Extensor digitorum communis

Wing Adjustors

Serratus superficialis
metapatagialis
Pectoralis propatagialis
longus

Pectoralis propatagialis
brevis

Cucullaris propatagialis
Propatagialis longus
Propatagialis brevis
Expansor secundariorum

Pronator sublimis
Pronator profundus
Entepicondylo-ulnaris
Flexor carpi ulnaris
Ulnimetacarpalis ventralis
Extensor metacarpi radialis
Supinator

Extensor digitorum communis
Extensor carpi ulnaris

Powered avian flight is produced by synchronous
clown strokes of the wing caused by contraction of
the large ventral muscle complex, the M. pecto¬
ralis. This complex usually consists of three or four
distinct muscles, the M. pectoralis thoracica, or
pectoralis superficialis, being the largest and most
important. The other pectoralis muscles typically
are small slips that function to tense the protopata-
gium, thus belonging to the last category listed
above. The M. pectoralis superficialis originates
extensively on the posterior and lateroventral sur¬
faces of the sternum, the ventral half of the entire
length of the carina, the entire posterolateral

NUMBER 27

5

surface of the clavicle and the anterior margin of
the sterno-coracoclavicular membrane. The pector-
alis tendon inserts broadly on the ventral surface
over most of the length of the deltopectoral crest
(crista lateralis humeri) of the humerus. This mus¬
cle provides nearly all the force for flight and is the
largest of all avian muscles, averaging more than
15 percent of total body weight among all flying
birds (Hartman, 1961; Greenwalt, 1962). Two osse¬
ous features reflect the size and functional impor¬
tance of this muscle: the very large sternum and its
carina, and the long and prominent deltopectoral
crest of the humerus.

Wing elevation (recovery stroke) is accom¬
plished by the combined actions of several mus¬
cles: the M. supracoracoideus, M. coracobrachialis
anterior and Mm. deltoideus major and minor. Of
these, the supracoracoideus is by far the most im¬
portant. The coracobrachialis, by virtue of its ori¬
gin on the anterodorsal extremity of the coracoid
(the acrocoracoid) anterior and dorsal to the
glenoid fossa, provides some lifting of the humerus,
but its chief action is to extend or pull the hu¬
merus forward, thexeby unfolding the wing. Typi¬
cally, it is the smallest “elevator” muscle. The M.
deltoideus major usually consists of a pars anterior
and pars posterior. The pais anterior arises from a
small area on the dorsal side of the scapula adja¬
cent to the glenoid. The pars posterior originates
on the dorsal end of the clavicle and the antero¬
dorsal surface of the scapula. Accordingly, these
fibers tend to elevate the humerus and draw it
forward. The M. deltoideus minor also originates
on the anterodorsal apex of the scapula, above,
medial, and slightly anterior to the glenoid, hence
also acting to elevate the humerus.

The largest humeral abductor, as noted above,
is the M. supracoracoideus, also termed the pecto-
ralis secundus or pectoralis minor (Figure 2). This
muscle arises by extensive attachment on the dorsal
parts of the sternal carina, the anterolateral sur¬
faces of the sternum, the ventro-anteromedial sur¬
face of the coracoid and the lateral part of the
coracoclavicular membrane. Its fibers converge
dorsally, attaching to a narrow tendon that passes
backward through an osseous canal, the foramen
triosseum, between the dorsal extremities of the
coracoid and clavicle and the anterior extremity
of the scapula. From there, the tendon turns down¬
ward to insert on the dorsal surface of the hu-

Supracoracoldeus Tendon

Figure 2. —Anterolateral view of the pectoral girdle and
sternum of Columbia livia to show the general relationships
of the M. supracoracoideus. The upper arrow indicates the
course and action of the supracoracoideus tendon from the
insertion toward the triosseal canal. The lower arrows indi¬
cate the location and action of the M. pectoralis, which has
been removed in this drawing. (After Fig. III.l, George and
Berger, 1966.)

merus between the head and the deltopectoral
crest. The fact that the triosseal canal is situated
above the insertion point when the humerus is
depressed allows this ventrally placed muscle to
elevate rather than depress the humerus.

Figures 2 and 3 illustrate the structure of the
triosseal canal and its relationship to the supra-
coiacoideus muscle. Of particular importance is
the vei'y prominent dorsal process of the coracoid
(the acrocoracoid) that extends well above and
anterior to the glenoid. The medial side of this
process forms the lateral wall of the triosseal canal
and is the primary structural leason for the de¬
flected course of the supracoiacoideus tendon.
Medially, the dorsal extiemity of the clavicle ar¬
ticulates with the upper medial surface of the acro¬
coracoid, forming the dorsomedial roof of the
triosseal canal. A further factor of importance is
that two important muscles arise from the upper
anterior surface of the acrocoracoid, the M. cora¬
cobrachialis anterior and the M. biceps brachii. As
noted earlier, the coracobrachialis anterior is a
primary extensor of the humerus and the biceps is
equally important as the principal flexor of the

6 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Figure 3. —Four views of the left scapulo-coracoid of Cathartes aura to show the nature of the
triosseal canal, which is responsible for the reversed action of the M. supracoracoideus in modern
carinates: a, lateral view; b, anterior view; c, dorsal view; d, medial view. (Bic. br. = the site of
origin of the M. biceps brachii; Corac. = the site of origin of the M. coracobrachialis anterior.

forearm. It is safe to assume that the elevated posi¬
tions of these origins at the apex of the acrocora-
coid have functional significance.

Without concerning ourselves with homologies,
or the proper name for the avian “supracoracoi¬
deus,” the action of that muscle in modern cari¬
nates emerges as extremely important for recon¬
structing some of the details of avian evolution. By
the nature of its location and architecture, it is
clear that at some earlier stage in the evolution of

birds the antecedent of this muscle must have acted
to depress the arm. Therefore, its action has been
completely reversed, probably as a consequence of
the development of the pulley-like arrangement of
the triosseal canal and its interposition between
the points of origin and insertion. The avian wing
is elevated chiefly by this ventral muscle, rather
than by dorsal muscles as we would expect, and as
is the case in bats.

The fact that virtually all muscles in all organ-

NUMBER 27

7

isms follow the most direct route between the
points of origin and insertion argues strongly
against the possibility that the insertion of the
supracoracoideus gradually migrated to the dorsal
side of the humerus, without prior or concurrent
deflection of the fibers or tendon leading to that
insertion. Even if the insertion had shifted to a
dorsal position on the humerus, contraction of the
muscle would still depress, as well as rotate, the
humerus^unless the fibers approached the hu¬
merus from above. Consequently, the most logical
explanation of the peculiar organization and
action of the modern avian supracoracoideus
would seem to be that its path was altered during
the course of avian evolution. Modern carinates,
together with the specimens of Archaeopteryx,
establish that these postulated changes resulted
from drastic changes in the shape of the coracoid
and that these changes occurred subsequent to the
Archaeopteryx stage.

“Flight” Apparatus of Archaeopteryx

The portion of the skeleton of Archaeopteryx
that can be equated with the flight apparatus of
modern carinates displays a number of important
features:

1. There appears to be little or no loss of flexi¬
bility in the trunk region, either by vertebral fu¬
sion or by restriction of vertebral articular free¬
dom. Although fully ossified gastralia are present,
there is no evidence of ossification of either sternal
ribs or the sternum. Also, there are no uncinate
processes on the dorsal ribs.

2. The pectoral arch does not appear to have
been as rigidly fixed as in modern birds. The cora¬
coids are short, subquadrangular, not strut-like,
and had only cartilaginous or membranous contact
with the sternum. The clavicles, however, were
fused and fully ossified into a robust furcula, but
the nature of its contacts with the scapulocoracoid
are not known.

3. Contrary to de Beer’s (1954) interpretation,
no sternum is preserved in any of the presently
known specimens of Archaeopteryx (Ostrom, in
press). This indicates that the sternum was almost
certainly cartilaginous and probably lacked a keel.
It may even have been membranous. Furthermore,
the space anterior to the gastralia is quite short, a
clear indication that the sternum, whether ossified

or not, could not have been enlarged, as it is in all
modern carinates.

4. The forelimb is elongated, but it does not
possess any of the skeletal specializations of mod¬
ern carinates that are usually equated with avian
flight. The deltopectoral crest of the humerus is
comparable to that of small theropod dinosaurs
and is longer and more elevated above the shaft
than is typical of most carinates. The elbow and
wrist joints are unmodified, the carpals and meta-
carpals are not fused and digits I, II, and III are
separate and unfused. The London and Berlin
specimens clearly show that the forelimbs bore
large, remex-like feathers, but it is uncertain
whether these feathers were attached directly to
the forelimb skeleton as in modern birds and as
would seem to be required of true “flight” feathers.
Despite exceptional preservation of several of the
specimens, none shows anything that can reason¬
ably be interpreted as quill nodes on the ulna.
This is negative evidence only, but a further indi¬
cation that the “flight” feathers were not firmly
attached to the skeleton is the fact that imprints
of the “primaries” of both wings in the London
specimen are preserved with only slight disarray-
ment, yet the left hand is disarticulated and the
right hand is missing altogether.

5. The long reptilian tail of Archaeopteryx bore
feathers, but there is no indication in any of the
specimens that the caudal series was undergoing
reduction or fusion into a pygostyle. On account
of the feathers, we can conclude that the tail may
have functioned as an aerodynamic, rather than
an inertial, stabilizer, but this should not
be construed as proof of flight capability in
Archaeopteryx.

The more important of the above conditions in
Archaeopteryx are the nonavian form of the cora¬
coid, the absence of an ossified sternum, the un¬
fused carpometacarpus and the unfused digits of
the manus. As Figures 4 and 5 show, the coracoid
of Archaeopteryx is not elongated, and clearly did
not serve as a strong, anticompressive brace against
the sternum. It appears to have been fused with
the scapula and its sternal border, although not as
robust as the scapular margin, is well defined, but
thin. The glenoid segment is stout, a relatively
large supracoracoid foramen is present and a very
prominent lateral process occurs just anterior to
and below the glenoid. This last feature, some-

8

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Figure 4. —Three views of the pectoral girdle of Archaeopteryx as reconstructed from the Lon¬
don, Berlin, and Maxberg specimens: a , anterior view of the left coracoid; b, lateral view of the
left scapulo-coracoid; c, dorsal view of the left scapulo-coracoid.

Figure 5.—Left coracoid and glenoid of the London specimen of Archaeopteryx, as seen in
anterior view from the underside of the main slab. (The smallest divisions on the scale equal
0.5 mm.)

NUMBER 27

9

times referred to as the biceps tubercle (Walker,
1972), is of special significance because it appears
to be the precursor of the avian acrocoracoid. Con¬
trary to Bakker and Gabon’s (1974) interpretation,
the glenoid does not face downward, but is directed
laterally (Figures 4 and 5) more or less as in mod¬
ern carinates (Figure 3).

The Transition from Archaeopteryx to
Modern Birds

Changes in the Pectoral Girdle

Before attempting to reconstruct hypothetical
transitional stages in the evolution of the pectoral
arch between Archaeopteryx and modern birds, it
may be useful to review certain facts. First, the
coracoid of all lower tetrapods, including birds,
has certain constant relationships with other ele¬
ments of the trunk. It occupies a position between
the scapula (with which it usually forms the
shoulder socket) and the sternum, regardless of
whether the latter is ossified or cartilaginous. Thus,
at least two regions of the coracoid, the sternal
border and the scapular border, are unmistakable
reference points no matter what the shape or size
of the coracoid. Similarly, the glenoid portion is
always recognizable.

The second consideration is the role of the cora¬
coid in forelimb biomechanics of lower tetrapods.
Chief among the various muscles that attach to the
coracoid (most of which insert on the humerus) is
the biceps, the principal flexor of the antebrach-
ium. (A structural and functional analog, the M.
coracoradialis proprius, is present in amphibians.)
The biceps passes between the coracoid and the
internal proximal surfaces of the radius and ulna.
Even in mammals, where the coracoid is no longer
present as a separate bone, the major forearm
flexor (which also happens to be termed the bi¬
ceps) originates on the presumed relict of the
coracoid, the coracoid process of the scapula.

The final consideration is that the location of
the flexor origin relative to the glenoid fossa de¬
termines the approximate path of forearm flexion.
Thus, for any given position of the humerus, the
approximate orientation of the plane of forearm
flexion can be determined from those two points.
For example: the biceps brachii of lizards origi¬
nates on a small area adjacent to the sternal border

of the coracoid, anterior and ventral to the glenoid.
With the humerus positioned in a horizontal
transverse position, the biceps flexes the forearm
anteroventrally toward the midline. But with the
humerus extended forward, forearm flexion is
down and backward. In birds, the site of origin of
the biceps on the anterolateral surface of the acro¬
coracoid is situated in front of and above the
glenoid; consequently, forearm flexion is restricted
to a forward movement (Figure 6).

girdle of Corvus brachyrhynchos, showing the location and
action of the M. biceps brachii (heavy arrow), the chief flexor
of the forearm in modem birds.

If, as seems reasonable, we accept the so-called
biceps tubercle of Archaeopteryx as the homolog
of the acrocoracoid of modern birds and the prob¬
able site of origin of the chief flexor of the forearm
(whatever we call it), we can reconstruct the gen¬
eral nature of forearm flexion in Archaeopteryx.
Although the precise orientation of the scapulo-
coracoid in Archaeopteryx cannot be established
from any of the presently known specimens, there
can be little doubt that the biceps tubercle was
situated well below and anterior to the glenoid
(Figure 4b). Consequently, there must necessarily
have been a major downward component in fore¬
arm flexion, regardless of whether the humerus
was extended, retracted, or even adducted.

Transformation of the avian coracoid from the
condition in Archaeopteryx to that of modern
birds involved two major changes: the dorsoventral
elongation of the main body of the coracoid and
the raising of the site of origin of the M. biceps
brachii by anterodorsal prolongation of the acro¬
coracoid. Elongation of the coracoid increased the
distance between the glenoid and the sternum, pre-

10

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

C a t h a r t e s

Figure 7. —Hypothetical stages in the evolution of the avian coracoid from the Archaeopteryx
stage to that of a modern carinate ( Cathartes). The arrows indicate the hypothesized course of
the M. supracoracoideus fibers in each stage and their progressive deflection resulting from
evolutionary elevation and expansion of the biceps tubercle (= acrocoracoid). Upper arrows
indicate the line of action of the supracoracoideus at each stage. Dashed lines indicate the
acromion and adjacent regions of the scapula. All stages are of a left coracoid viewed from the
front.

sumably increasing the range of dorsoventral hu¬
meral excursion. This in turn may have been cor¬
related with the anteroposterior elongation of the
sternum, the development of the sternal keel, and
the enlargement of the ventral adductor muscles—
the M. pectoralis. Increased force of forelimb ad¬
duction, for whatever biological role, required
strengthening of the coracoid into a strong, anti-
compressive strut between the shoulder socket and
the enlarged muscle origins on the sternum.

Because the supracoracoideus of lower tetrapods
originates ventral and anterior to the glenoid, and
because it also has a ventral origin close to the
sternal border in modern birds, the primitive site
of origin of this muscle in Archaeopteryx probably
was in a similar position —ventral and somewhat
medial to the biceps tubercle. If so, then any up¬
ward expansion of the biceps tubercle would have
impinged against the supracoracoideus tendon,
gradually deflecting its course medially around the
base of the expanding “protoacrocoracoid.” Once
the base of this process reached the level of the
glenoid, the then-deflected supracoracoideus would
have pulled the humerus anteromedially, rather
than downward. Continued expansion and eleva¬
tion of the acrocoracoid would have resulted in

further deflection of the supracoracoideus. The
action of this muscle almost certainly was not re¬
versed abruptly, but probably changed gradually
from that of a humeral adductor, to an antero-
ventral extensor, to a forward extensor, to an
antero-dorsal extensor and finally becoming an
abductor of the humerus. Figure 7 illustrates how
this transformation may have taken place.

If the above reconstruction is even approxi¬
mately correct, it is clear that one of the major
factors in the evolution of avian flight structures
was the upward expansion of the acrocoracoid.
This conclusion is established beyond any doubt
by the presently reversed action of the supracora¬
coideus in modern birds. The critical question is:
What brought about the upward expansion of the
acrocoracoid? There appear to be several possibili¬
ties: (1) elevation of the anterior part of the glen¬
oid and rotation of the shoulder socket to face di¬
rectly laterally, thereby permitting unrestricted
transverse (up and down) movements of the fore¬
limb; (2) provision of an enlarged buttress at the
level of the glenoid for the furcula to brace against,
thereby insuring proper transverse separation of
the shoulder sockets; (3) raising of the levels of
humeral extension and forearm flexion by elevat-

NUMBER 27

11

Figure 8. —The furcula of Archaeopteryx as preserved in the London specimen. The exposed
surface is probably the anterior surface. (The smallest divisions on the scale equal 0.5 mm.)

ing the sites of origin of the coracobrachialis and
biceps. In all probability, none of these factors
acted alone, and other less obvious factors may
have been involved as well.

Whether enlargement of the pectoral adductor
muscles and the elongation of the coracoids into
robust struts occurred before, after, or concurrently
with upward expansion of the acrocoracoid cannot
be determined in the absence of intermediate
stages in the avian fossil record. Whatever the se¬
quence, the upward growth of the acrocoracoid
would have progressively deflected the action of
the supracoracoideus. It also brought about signi¬
ficant changes in other forelimb movements,
especially in elevating the range of humeral ex¬
tension and increasingly confining it to the cra-
niad sector. As a direct consequence, the level of
forearm flexion was also elevated to a nearly hori¬
zontal fore-aft plane more or less perpendicular to
the transverse, up and down, humeral movements
produced by the enlarged pectoral muscles.

Considering these three possibilities, it appears
that the glenoid in Archaeopteryx already faced
laterally and slightly forward (Figure 4 b,c) not
ventrolaterally, as Bakker and Gal ton (1974)
claim. The coracoid portion of the glenoid also

seems to have been elevated. Yet, the biceps tu¬
bercle was still small and located well below the
glenoid. Also, as was noted earlier, a robust furcula
is present in Archaeopteryx (as seen in the Lon¬
don [Figure 8] and Maxberg specimens), and al¬
though the nature of its articulations with the
other elements of the pectoral girdle is not clear,
there does not appear to have been any special
structure of the coracoid that might have served to
buttress it, since, as already noted, the biceps tu¬
bercle is not elevated. This, of course, raises the
question of the function of the furcula in Archaeo¬
pteryx. Did it serve as a transverse spacer between
the shoulder sockets? If so, it would appear to have
been related to some activity other than powered
flight—perhaps predation. Since both the M. cora¬
cobrachialis anterior and the M. biceps brachii
arise from the upper anterior surface of the acro¬
coracoid in all modern carinates, then by virtue of
their positions above and in front of the glenoid,
these muscles, respectively, pull the humerus for¬
ward and up, and flex the forearm forward and in¬
ward toward the midline. In Archaeopteryx, the
humerus apparently could not be extended for¬
ward and upward above the level of the shoulder
because no part of the coracoid was situated above

12

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

and in front of the glenoid (Figures 4 and 9).
Thus, of the three possibilities suggested above,
the evolutionary expansion of the avian acrocora-
coid would seem to have been most critically linked
with the actions of the coracobrachialis and biceps
muscles. There appears to have been some selec¬
tive advantage in raising the level or attitude of
forelimb extension and forearm flexion.

So far, I have given little attention to the scap¬
ula. This is because the scapula of Archaeopteryx
already had acquired a form remarkably similar to
that of modern birds, being very long, narrow and
strap-like. Its principal distinctions from the con¬
dition in most modern birds are its fusion to the
coracoid, the form of the acromion, and the shape
of the distal extremity, which is rectangular or
slightly flared in Archaeopteryx rather than ta¬
pered. The fact that the acromion is more promi¬
nent and robust than in most modern birds sug¬
gests that the M. deltoideus was perhaps a more

Figure 9. Comparison of lateral views of the pectoral arch
of Archaeopteryx and a modern carinate ( Cathartes ) to show
the respective positions of the biceps tubercle and the acro-
coracoid relative to the glenoid. The broken lines define the
approximate dorsoventral range of humeral extension and
forearm flexion possible in each as a result of contractions by
the muscle that originated on those two processes.

important humeral elevator at the Archaeopteryx
stage of avian evolution. This would be consistent
with the conclusion reached above that the supra-
coracoideus of Archaeopteryx could not have ele¬
vated the humerus (as was noted by Walker, 1972),
but rather must have been a lateral adductor.
If the deltoideus, however, was more important as
a humeral elevator at the Archaeopteryx stage
than it is in modern carinates, then it would ap¬
pear that the force of the recovery stroke must
have continued to decline in birds succeeding
Archaeopteryx, until complete deflection of the
supracoracoideus was accomplished. This implies
that there probably was no tendency at the Archae¬
opteryx stage, or immediately afterward, toward
powered flight.

It should also be noted here that the stout
acromion in Archaeopteryx may not have had any¬
thing to do with the deltoideus muscles, but might
have served as a buttress for the stout furcula.
This cannot be established on the basis of present
specimens, however.

The narrow form of the scapula, as compared
with the broad, triangular form in all other tetra-
pods except theropod dinosaurs, suggests that the
musculature that inserted or originated on the
scapular blade—and particularly on its dorsal sur¬
face—was greatly reduced. This certainly is true of
modern birds in which the M. rhomboideus and
M. scapulohumeralis (the largest dorsal shoulder
muscles) are of relatively small size. The fact that
this narrow scapular form occurs only in obligate
bipeds (birds, Archaeopteryx, and theropod dino¬
saurs), but not in facultative bipeds (such as non¬
human primates, kangaroos, or ornithopod dino¬
saurs), or in any quadrupedal animal is highly
suggestive. It indicates that strong stabilization of
the pectoral arch by muscles connecting the scapular
blade with the vertebral column and dorsal ribs,
and powerful abduction of the limb by large mus¬
cles extending between the humerus and the scap¬
ular blade, were unnecessary in obligate bipeds in
which the forelimb was no longer involved in
weight support.

Changes in the Forelimb

Comparison of the forelimb skeleton of Archaeo¬
pteryx with that of modern birds reveals several
major differences, the most conspicuous of which

NUMBER 27

13

Figure 10.—The right manus and carpus of the Berlin specimen of Archaeopteryx, seen in
dorsal aspect. Notice the separated fingers and the unfused metacarpus and carpus, as well as
the extent of lateral flexion. (The smallest divisions on the scale equal 0.5 mm. Roman numerals
identify the digits.)

occur in the hand and wrist. Figure 10 shows the
right hand and wrist of the Berlin specimen of
Archaeopteryx with its unfused metacarpus and
three separated fingers. The same construction is
present in the other three specimens in which the
hands are preserved. This construction is in sharp
contrast to the united metacarpus and manus of
modern birds (Figure 11). It is obvious that
phalanges have been lost or co-ossified in at least
the external finger (digit III) of modern birds,
but the most interesting changes have taken place
in the metacarpus and wrist. Figure 12 illustrates
the carpus and metacarpus as they are preserved
in the Berlin (Figure 12a) and Eichstatt (Figure
126) specimens, compared with the same elements
of a modern carinate, Cathartes aura (Figure
12 c,d). The first metacarpal is considerably shorter
than the other two (Figure 10), as it is in modern
forms, but it does not appear to be co-ossified with
metacarpal II, nor are the second and third meta-
carpals fused. The carpus consists of only three
elements, a large distal carpal with a distinctive
semicircular proximal profile, and two smaller
bones, which probably represent the radiale (sca-

pholunar) and the ulnare (cuneiform). Although
neither of the last two elements resemble modern
bird carpals, two features in Archaeopteryx do pre¬
view specialized conditions of the modern avian
carpometacarpus. These are the large lunate distal
carpal that is closely articulated with the first and
second metacarpals (Figure 12a,6), and the in¬
ternal expansion at the base of metacarpal I. There
can be little doubt that the lunate carpal of Arch¬
aeopteryx, by fusion with the two metacarpals, be¬
came the pulley-like trochlea of the carinate
carpometacarpus. The proximal internal expan¬
sion at the base of the first metacarpal in Archaeo¬
pteryx is almost certainly the precursor of the
large extensor process (processus metacarpalis I)
of the modern carpometacarpus. In Figure 13, I
have attempted to show how the modern avian
carpometacarpus probably evolved from the con¬
dition in Archaeopteryx.

Reconstructing the above intermediate stages is
far simpler than trying to account for the condi¬
tions that brought about such changes. The second
digit clearly was the dominant finger and ulti¬
mately became the main supporting structure of

14

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

m

Trochlea
C a r p a lis

Process

Figure 11.—The right manus and carpometacarpus of Cathartes aura in dorsal aspect, for
comparison with Figure 10. (Roman numerals identify the digits.)

the primary remiges. The third or outermost
finger gradually was reduced and metacarpal III
was modified into a posterior (external) brace
against metacarpal II. These changes could well
have come about in connection with flight and the
attachment of the primaries to the second meta¬
carpal, presumably bracing it against lift forces
that would tend to rotate the second digit and
metacarpal upward. Fusion of the lunate carpal to
the metacarpus, and its expansion into the pulley¬
like trochlea, increased the degree of flexion pos¬
sible at the wrist, but at the same time reduced
wrist mobility to the plane of the metacarpus and
the wing. The prominent extensor process of the
modern carpometacarpus is the point of insertion
of the largest muscle of the avian forearm, the M.
extensor metacarpus radialis, the action of which
extends or unfolds the hand (Figure 14).

In the discussion that follows, it is essential to
distinguish between two very different kinds of
flexing movements at the wrist: that in which the
extremity is flexed toward the ulnar or external
side of the forearm (termed lateral flexion here)
and that in which the hand and metacarpus are
“flexed” inward toward the radial side of the fore¬
arm. This last movement might be termed “me¬
dial flexion,” but for the sake of clarity it is desig¬
nated here as “hyperextension.” These terms differ
from the usual terminology applied by ornitholo¬
gists (which by convention is in terms of a laterally
extended wing), but hopefully they will be clear to
all readers. The term extension is used here in the
sense of straightening the wrist, and where neces¬
sary for clarity, it will be specified as extension

from the laterally flexed or the hyperextended
condition.

In the Berlin and Eichstatt specimens of Arch¬
aeopteryx, the hands are flexed laterally toward
the ulna at about 80 degrees to the radius and
ulna. Close examination of the wrist in each case
(Figure 12 a,b), and especially of the morphology
of the lunate carpal and the external aspect of the
ulnar extremity, reveals that in both specimens the
wrists are fully flexed. Notice that the internal
condyle or condylus metacarpalis does not extend
proximally along the outer surface of the ulnar
shaft as it does in modern birds. For contrast. Fig¬
ure 12c shows the much greater maximum degree
of lateral flexion (hyperflexion) possible in the
modern bird wrist. Also conspicuous in modern
birds is the elongated extensor process of the carpo¬
metacarpus, which greatly increases the leverage of
the principal extensor of the hand.

It is tempting to relate these features to some
aspect of flight; for example, the need for adjusting
or changing the surface area of the airfoil by im¬
proved efficiency and precision of extension and
flexion at the wrist. Once flight capability had
been achieved, increased leverage for the M. ex¬
tensor metacarpus radialis would reduce the
amount of energy required to counteract the force
of the airstream that tends to flex or fold the wing
extremity laterally. On the other hand, during the
power stroke, lift forces tend to open or extend the
wing extremities. Another possibility is that the
extensor process grew larger in conjunction with
the development of wrist hyperflexion, which in
turn was made possible by gradual expansion of

NUMBER 27

15

Extensor Process

Figure 12.—The wrists of Archaeopteryx ( a and b) and Cathartes aura ( c and d) as viewed
from above: a, left wrist of the Berlin specimen; b, right wrist of the recently recognized
Eichstatt specimen (b and c are preserved flexed laterally, toward the ulnar side of the forearm,
to the maximum degree possible); c, left wrist of Cathartes drawn in the same laterally hyper-
flexed position to show the greater degree of flexion possible in modern carinates; d, “exploded”
dorsal view of the right wrist of Cathartes, flexed to the same degree as b, to show the specialized
facets of the wrist elements, arrows indicating complementary articular facets. Notice in particu¬
lar the lengths of the external portions of the condylus metacarpalis of the ulna and also the
trochlea carpalis of the carpometacarpus, as compared with the corresponding regions in
Archaeopteryx. Also notice the large extensor process of the carpometacarpus compared with
the modest expansion on metacarpal I of Archaeopteryx. The phalanges have been omitted from
digit I in a and b. (Roman numerals identify the metacarpals; the horizontal lines equal 10 mm).

the trochlea carpalis of the carpometacarpus and
elongation of the condylus metacarpalis of the
ulna. A critical point here, however, is that ex¬
treme hyperflexion of the manus has no obvious
“flight” advantage, but it clearly is advantageous
for compact folding of the forelimb extremities to
protect the airfoil when not in use. Under these
circumstances, it would appear that the increased
extension leverage that was provided by a larger
extensor process on the carpometacarpus was not

related to the first explanation above, but prob¬
ably was advantageous for quick unfolding of a
hyperflexed wing. This interpretation is reinforced
when it is considered in conjunction with the
unique linkage between the modern avian elbow
and wrist that automatically synchronizes flexion
(or extension) at those two joints. As first observed
by Coues (1871) and Headley (1895), and con¬
firmed by Fisher’s (1957) experiments, the radius
of birds functions as a “connecting rod” between

16

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Archaeopteryx

C a t h a r t e s

Figure 13.—Hypothetical stages in the evolution of the avian carpometacarpus from the
Archaeopteryx stage to that of modern carinates.

the elbow and wrist. Because of the greater length
of the radial versus the ulnar condyle of the hu¬
merus, the radius slides distally along its axis when
the elbow is flexed by the M. biceps, thereby push¬
ing against the carpus and metacarpus and forcing
the wing extremity to flex. Extension of the wrist
and elbow are similarly linked. In fact, because of
the “connecting rod” action of the radius, and the
increased leverage of the M. extensor metacarpus
radialis provided by the enlarged extensor process,
it is possible for that muscle to function as the
primary unfolder of the entire wing, not just of the
hand. Although smaller than the M. triceps brachii,
the forearm extensor, the M. extensor metacarpus
radialis of most carinates has far better leverage
than the triceps (which inserts on the olecranon)
for extending the wing extremity.

It is not possible to establish which, if either, of
the above possibilities was the decisive factor in
the evolution of the modern avian wrist, but the
specimens of Archaeopteryx seem to provide a
clue. All four of the specimens in which the hand
is preserved show what appears to be a maximum
degree (about 80°) of flexion of the hands toward
the ulnar side of the forearm. In other words, the
hand could not be hyperflexed or folded back
tightly against the forearm as in modern birds.
Also, the extensor process is only very weakly de¬
veloped in these specimens. The nature of the
articular surfaces in the wrists of the Berlin and
Eichstatt specimens, however, indicates that the

hands almost certainly could have been hyperex-
tended medially, or bent toward the radial side of
the forearm, to about the same degree that they
are preserved flexed laterally toward the ulna, per¬
haps even more so. This last is important, because
medial hyperextension of the hand is not possible
in modern birds. In fact, the manus cannot even
be fully extended to align parallel with the radius
and ulna. From this, the most probable conclu¬
sion is that the extensor process is most important
for recovery (extension) of the avian manus from
a folded or laterally hyperflexed condition. If it
had developed for enhancing medial hyperexten¬
sion it is difficult to understand why this process
was retained, even enlarged, while at the same time

Ficure 14.—Dorsal view of the wing skeleton of Corvus
brachyrhytichos to show the position and action of the M.
extensor metacarpus radialis (heavy arrow), the chief extensor
of the hand.

NUMBER 27

17

the capacity for medial hyperextension of the hand
was being reduced and ultimately eliminated.

The rest of the forelimb appears to have been
altered in much less conspicuous ways during the
transition from Archaeopteryx to modern birds,
yet those changes that can be recognized may have
significant implications. In the ulna, the most ob¬
vious changes involved the external expansion and
elongation of the condylus metacarpalis (Figure
12), the articular facet of which permits hyper¬
flexion of the manus laterally. Less obvious is the
apparent lack of direct attachment of the second¬
ary remiges to the ulna, or of the primaries to the
metacarpus, as is indicated by the absence of quill
nodes. These conditions are lacking in the speci¬
mens of Archaeopteryx, but are well developed in
a variety of modern carinates.

The humerus of Archaeopteryx, although very
bird-like, is much simpler than that of modern
carinates. There is a long and well-defined delto-
pectoral crest, but as can be seen in Figure 15, this
crest projects farther from the shaft than is char¬
acteristic of most later birds. More important,
though, are the features that are missing from the
humerus of Archaeopteryx. There is no sign of
either the external or internal tuberosity, nor is

ARCHAEOPTERYX

Deltopectoral Crest

Internal Tuberosity

CATHARTES

Figure 15.—Comparison of the humeri of Archaeopteryx and
a modern carinate ( Cathartes ), as viewed in dorsal aspect.
Humeri are drawn to unit length for easy comparison, the
relative sizes of each being indicated by the horizontal scale
lines which equal 3 cm. The humerus of Archaeopteryx is
devoid of most of the tubercles and crests that are well de¬
veloped in most modern birds. Most of these features are the
sites of attachment of muscles that act to fold the wing.

there a bicipital crest. Distally, the ectepicondyle
is also absent, or at least there is no detectable
tubercle preserved in presently known specimens.
In view of the other related features of the distal
segments of the forelimb of Archaeopteryx, the
absence of these processes seems to have special
significance, because in modern carinates they play
a direct part in the compact folding of the wing.

The internal tuberosity (tuberculum mediale) is
the site of insertion of the three principal humeral
retractors (the M. subscapularis, M. subcoracoi-
deus, and M. coracobrachialis posterior). The ex¬
ternal tuberosity (tuberculum laterale) is the site
of insertion of the M. supracoracoideus which, in
addition to elevating the wing, also rotates the en¬
tire folded wing dorsally toward the midline in
modern birds. In Archaeopteryx, however, this mus¬
cle must have been a humeral depressor, as has
been emphasized above. The bicipital crest (crista
medialis) of modern birds is the area of insertion
of the M. scapulohumeralis posterior, which draws
the humerus back against the body. The implica¬
tions of these conditions are obvious: in the ab¬
sence of all specialized features of the humerus,
ulna, and carpometacarpus that in modern birds
are directly related to the folding of the wing, we
are forced to conclude that Archaeopteryx was
unable to fold the forelimb back against the body
as in modern birds. Add to this the absence in
Archaeopteryx of an ectepicondyle, which is the
site of origin of the M. extensor metacarpus radi-
alis, and also the weak development of the extensor
process of metacarpal I, which is the site of inser¬
tion of this same muscle in later birds. These con¬
ditions indicate that powerful or rapid extension
of the manus was unlikely, and probably not nec¬
essary, because the wrists of Archaeopteryx clearly
show that lateral hyperflexion of the manus was
not possible. On the other hand, a high degree of
medial hyperextension was retained, perhaps as a
critical action for prey catching or feeding activities.

Discussion

It would appear that the acquisition of obligate
bipedal posture and locomotion in some pre-
Archaeopteryx stage of avian evolution was re¬
sponsible in large part for the ultimate develop¬
ment of powered avian flight. An early consequence
was the narrowing of the scapula. Strong stabiliza-

18

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

tion of the scapula and shoulder joint, and
powerful abduction of the forelimb became less
critical than in obligate or occasional quadrupeds
(whether of sprawling or upright posture), where
antagonistic and synergistic interaction of dorsal
abductors and ventral adductors are necessary for
precise dynamic control of limb movements and
positions under weight-bearing conditions. By the
Archaeopteryx stage, forelimb abduction may have
been accomplished solely by the action of a re¬
duced remnant of the M. deltoideus that presum¬
ably originated on the prominent acromion.

With the assumption of upright, obligate bipedal
posture and the release of the forelimbs from a
weight-supporting role, new forelimb functions be¬
came possible. At the Archaeopteryx stage these
functions apparently involved laterally elevated
movements of the forelimb (as indicated by the
outward facing glenoid), anteroventral extension
of the humerus (as indicated by the anteriorly
facing surface of the coracoid below the level of
the glenoid—the only available site of origin for
humeral extensors), and powerful anteroventral
flexion of the forearm toward the midline (as in¬
dicated by the prominent biceps tubercle below
and anterior to the glenoid—the most probable site
of origin of the forearm flexor). The hands were
capable of nearly 180 degrees of lateral flexion and
medial hyperextension, as noted above. The ca¬
pacity for extreme hyperextension at the wrist
(not possible in modern birds), coupled with the
evidence for strong flexion of the forearm toward
the sagittal plane, appears to be especially signi¬
ficant. Perhaps even more significant is the evi¬
dence that the forelimb of Archaeopteryx probably
could not have been raised above the level of the
glenoid when in the anteriorly extended position,
simply because no part of the shoulder girdle was
situated above and anterior to the glenoid. The
strong anterior extensor (M. coracobrachialis an¬
terior) and forearm flexor (M. biceps brachii) of
modern birds have their present actions only be¬
cause of the elevated positions of their origins on
the acrocoracoid. The evolutionary upward expan¬
sion of the acrocoracoid would seem to have been
linked causally with the actions of those two mus¬
cles and most especially with that of the humeral
extensor. Selection apparently favored the eleva¬
tion of forearm and hand activities.

As observed above, it is tempting to equate such

changes with some aspect of flight. For example,
these changes might permit alterations in the sur¬
face area of the “wing” by means of flexion or
extension of distal components more or less in the
plane of the “wing.” Notice, however, that these
capabilities apparently were not yet available in
Archaeopteryx, where wrist and elbow movements
were not restricted. Another possibility is that ele¬
vation of forelimb extension and forearm and
hand flexion and extension ostensibly might im¬
prove the aerodynamic qualities of an incipient
“wing” by making possible a positive angle of at¬
tack (where the leading edge of the airfoil is above
the trailing edge, relative to the airflow or flight
path). A positive angle of wing attack is essential
for all forms of flight, whether powered or passive,
because without it there can be no lift. That being
true, then there is a critical flaw in attributing the
above anatomical changes between Archaeopteryx
and modern birds to aerodynamic adaptations. The
flaw is that there can be no lift, and thus no aero¬
dynamic selective advantage in raising the attitude
of a potential airfoil until after the smallest degree
of a positive angle of attack has been acquired.

An aerodynamic explanation of the anatomical
changes noted above is also weakened by the ab¬
sence of an ossified sternum in all specimens of
Archaeopteryx. The absence of a sternum strongly
suggests that the “flight” muscles of Archaeopteryx
were not of unusual size, a conclusion that is sub¬
stantiated by the short space available for the ster¬
num in front of the ossified gastralia, as well as by
the short nonstrut-like form of the coracoid. If all
these assessments are correct, then some biological
role other than flight must have been involved in
the initial and early phases of the upward expan¬
sion of the biceps tubercle into the future
acrocoracoid.

Aside from making lift possible, the only other
obvious consequence of raising the level of fore¬
limb extension and flexion is to place the hands
and their activities directly in front of and above
the animal. Two activities immediately come to
mind: climbing and prey-catching. Various authors
(Bock, 1965, 1969; de Beer, 1954; Swinton, 1960)
have interpreted Archaeopteryx as being an arbo¬
real animal. I have argued that there is no compel¬
ling evidence for this (Ostrom, 1974), and instead,
the skeletal anatomy of Archaeopteryx appears to
have been adapted for ground-dwelling activities.

NUMBER 27

19

Even if Archaeopteryx were arboreal, however, a
possibility that I do not deny, then it acquired its
climbing skills prior to elevation of the acrocora-
coid and the capacity of elevated forelimb exten¬
sion, and after the acquisition of obligate bipedal
posture.

Obviously the same is true of prey-catching and
feeding activities of Archaeopteryx. If Archaeo¬
pteryx were insectivorous, as seems almost certain,
it clearly must have been proficient at catching
insects, whether it did so with its mouth by quick
darting movements of the head on the long flexible
neck, or by grasping them in the hands or snaring
them beneath the forelimb plumage. Considering
the general absence of flight-related skeletal struc¬
tures in the forelimb and pectoral girdle, it does
not seem unreasonable to conclude that the fore¬
limbs of this obligatory bipedal predator must have
taken part in prey-catching activities.

If the forelimbs of Archaeopteryx were used to
catch prey, and if the original advantages behind
the enlargement of the contour feathers of the
forelimb was to enhance insect-catching skills
(Ostrom, 1974), there would be very real selective
advantages in any changes that increased the scope
of forelimb movements, especially if we think in
terms of leaping or flying insects.

At this point it is not possible to identify the
exact activities or selective advantages that pro¬
moted the upward expansion of the acrocoracoid,
but it seems clear that these were related to up¬
ward extension of the arms and hands. It also ap¬
pears that flight was not a factor in these first modi¬
fications. It was perhaps only coincidental that
once a certain degree of upward enlargement of
the acrocoracoid had been accomplished, the ac¬
tion of the coracobrachialis anterior would have
been supplemented by the newly deflected supra-
coracoideus acting as an anterodorsal extensor of
the humerus.

The various specialized features of the modern
avian forelimb skeleton mentioned above (reduced
fingers, fused carpometacarpus, novel tubercles and
crests on the humerus) seem best explained as
flight-related adaptations that appeared subse¬
quent to the dorsal expansion of the acrocoracoid
and the resultant ability to raise the attitude of
the extended forelimb, thereby achieving at least a
minimal positive angle of attack. Fusion of the
metacarpus would solidify the structural support

of the primary flight feathers and brace the second
metacarpal against long-axis rotation resulting
from lift forces. Phalangeal reduction may have
been correlated with changing the function of the
manus to that of an airfoil and the reduction of
the primitive prey-grasping role of the long, sepa¬
rated fingers. Fusion of the distal carpal to the
metacarpus reduced the amount of abduction-
adduction possible at the wrist, but at the same
time facilitated precise flexion-extension of the
manus in the plane of the wing, essentially per¬
pendicular to the powerful adductive actions of
the enlarged pectoral muscles. The capacity for
medial hyperextension of the manus was reduced
and ultimately lost, presumably as the primitive
avian hand became less and less involved with
prey-catching and feeding activities and was in¬
creasingly adapted for flight-related functions.
Later stages presumably involved development of
structures related to compact folding of the wing
and rapid unfolding—the various tubercles of the
humerus noted above, the capacity for lateral
hyperflexion of the manus, and the enlarged ex¬
tensor process of the carpometacarpus.

Summary

The existence of several specimens of Archaeo¬
pteryx, the oldest known fossil remains that are
universally accepted as avian, provides important
anatomical details of an extremely early stage in
bird evolution. Despite impressions of what appear
to have been modern-type “flight” feathers attached
to the forelimb (but possibly not attached to
the forelimb skeleton), the five presently known
specimens of Archaeopteryx show almost no osteo-
logical features that compare with the skeletal
adaptations of the modern avian flight apparatus.
The only exception is the furcula, preserved in
the two largest specimens.

Assuming that Archaeopteryx is in fact an an¬
cestral bird, and in the absence of any known
intermediate structural stages in the avian fossil
record between Archaeopteryx (of Late Jurassic
age) and the essentially modern birds of Late
Cretaceous and Early Tertiary ages, we can postu¬
late only the most obvious structural changes that
occurred during the evolution of the avian flight
mechanism. From the Archaeopteryx stage, the

20

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

following sequence of developments seems prob¬
able, with the first two possibly taking place con¬
currently or, less probably, in reversed order: (1)
Upward expansion of the biceps tubercle, thereby
raising the sites of origin of the M. biceps and M.
coracobrachialis and thus the level of humeral
extension and forearm flexion—possibly in connec¬
tion with prey-catching or feeding activities, or
perhaps to facilitate quadrupedal climbing. A di¬
rect consequence of the expansion of the acro-
coracoid was the conversion of the M. supracora-
coideus from a humeral adductor to a humeral
elevator. (2) Enlargement of the pectoral muscles
for more powerful arm adduction, accompanied by
enlargement and ossification of the sternum, and
elongation and strengthening of the coracoids to
immobilize the shoulder joints. (3) Attachment of
the remiges to the ulna and the second digit to
resist feather deflection during the wing down-

stroke. (4) Fusion of carpals and metacarpals into
a united carpometacarpus for firmer fixation of the
primaries, modification of the trochlea carpalis to
permit only planar flexion and extension at the
wrist, and loss of phalanges from all three fingers.
(5) Loss of medial hyperextension of the hand and
development of the capacity for compact folding
of the wing, due to elongation of the condylus
metacarpalis of the ulna and various tubercles on
the humerus. This was associated with enlargement
of the extensor process of the carpometacarpus to
provide leverage for rapid unfolding of the wing.

The occurrence of other changes in the musculo¬
skeletal system that affected the flight apparatus
cannot be determined in the above sequence, but
powered flight, as opposed to either gliding or
flapping leaps, almost certainly could not have
occurred before the first three of the above stages
had been completed.

Literature Cited

Bakker, R. T„ and P. M. Galton.

1974. Dinosaur Monophyly and a New Class of Verte¬
brates. Nature, 248:168-172.

de Beer, G.

1954. Archaeopteryx lithographica. xi + 64 pages, 9
figures, 15 plates. London: British Museum (Nat¬
ural History).

Berger, A. J.

1960. The Musculature. Chapter 8 in A. J. Marshall,
editor, Biology and Comparative Physiology of
Birds, Volume 1. 9 figures. New York: Academic
Press.

Bock, W. J.

1965. The Role of Adaptive Mechanisms in the Origin
of Higher Levels of Organization. Systematic Zo¬
ology, 14(4):272—287, 6 figures.

1969. The Origin and Radiation of Birds. Annals of the
New York Academy of Sciences, 167(1): 147-155.

Coues, E.

1871. On the Mechanism of Flexion and Extension in
Bird’s Wings. Proceedings of the American Associa¬
tion for the Advancement of Science, 20:278-284,
2 figures.

Fisher, H. I.

1946. Adaptations and Comparative Anatomy of the
Locomotor Apparatus of New World Vultures.
American Midland Naturalist, 35(3): 545—727, 28
figures, 13 plates.

1957. Bony Mechanisms of Automatic Flexion and Ex¬
tension in the Pigeon’s Wing. Science, 126:446, 1
figure.

George, J. C., and A. J. Berger

1966. Avian Myology, xii + 500 pages, 248 figures. New
York: Academic Press.

Greenewalt, C. H.

1962. Dimensional Relationships for Flying Animals.
Smithsonian Miscellaneous Collections, 144(2): 1-46,
17 figures.

Hartman, F. A.

1961. Locomotor Mechanisms of Birds. Smithsonian Mis¬
cellaneous Collections, 143(1): 1—91, 7 figures, 5
tables.

Headley, F. W.

1895. The Structure and Life of Birds, xx + 412 pages,
78 figures. New York: MacMillan.

Heptonstall, W. B.

1970. Quantitative Assessment of the Flight of Archaeop¬
teryx. Nature, 228:185-186, 2 figures.

Holmgren, N.

1955. Studies on the Phylogeny of Birds. Acta Zoologica,
36:243-328, 37 figures.

Hudson, G. E., and P. J. Lanzillotti

1955. Gross Anatomy of the Wing Muscles in the Family
Corvidae. American Midland Naturalist, 53:1-44,
35 figures.

Marsh, O. C.

1880. Odontornithes: A Monograph on the Extinct
Toothed Birds of North America, xv + 201 pages,
40 figures, 34 plates. Volume 7 of Report of the
C*eological Exploration of the Fortieth Parallel.
(Professional Papers of the Engineer Department,
United States Army, No. 18.) Washington, D.C.

NUMBER 27

21

Ostrom, J. H.

1973. The Ancestry of Birds. Nature, 242:136.

1974. Archaeopteryx and the Origin of Flight. Quarterly
Review of Biology, 49:27-47, 10 figures.

1975. The Origin of Birds. Pages 55-57 in volume 3 of
F. A. Donath, editor. Annual Review of Earth and
Planetary Sciences. 9 figures. Palo Alto: Annual
Reviews Inc.

In press. Archaeopteryx and the Origin of Birds. Linnean
Society Biological Journal.

Simpson, G. G.

1946. Fossil Penguins. Bulletin of the American Museum
of Natural History, 87:1-100, 33 figures.

Stresemann, E.

1933. Aves. Number 2 of volume 7 in W. Kiikenthal and
T. Krumbach, editors, Handbuch der Zoologie.
Berlin: W. Gruyter.

Swinton, W. E.

1960. The Origin of Birds. Chapter 1 in volume 1 in
A. J. Marshall, editor. Biology and Comparative
Physiology of Birds. New York: Academic Press.

1964. Origin of Birds. Pages 559-562, in A. L. Thomson,
editor, A New Dictionary of Birds. 1 figure. Lon¬
don: Nelson.

Sy, M.

1936. Funktionall-anatomische Untersuchungen am Vog-
elfliigel. Journal fur Ornithologie, 84:199-296, 52
figures.

Walker, A. D.

1972. New Light on the Origin of Birds and Crocodiles.
Nature, 237:257-263, 9 figures.

Yalden, D. W.

1970. The Flying Ability of Archaeopteryx. Ibis, 113:
349-356, 4 figures.

Evolutionary Significance
of the Mesozoic Toothed Birds

Philip D. Gingerich

ABSTRACT

Well-preserved fossils of the Mesozoic toothed birds
Archaeopteryx, Hesperornis, and Ichthyornis, and
of the bird-like dinosaur Compsognathus, dis¬
covered in the 19th century, indicated to early
evolutionary biologists that dinosaurs and birds
were closely related, and that birds in all proba¬
bility evolved from a dinosaur similar to Compso¬
gnathus. The modern ratites, sharing some distinc¬
tive similarities with Hesperornis, were regarded as
survivors of a primitive initial radiation of birds.
Several workers have subsequently challenged the
idea that the Cretaceous birds Ichthyornis and
Hesperornis had teeth or that they bore any simi¬
larity to the ratites. After careful study of the
actual fossil specimens of Hesperornis, it is clear
that this Cretaceous bird had toothed jaws and a
palaeognathous palate, the latter condition being
shared with ratites and certain dinosaurs. These
and other characters place Hesperornis, like Arch¬
aeopteryx, in a position morphologically, as well
as temporally, intermediate between dinosaurs and
typical birds. The few significant features uniting
the living ratites and tinamous all appear to be
primitive characteristics, suggesting that ratites and
tinamous are either survivors of an early radiation
of birds, or are possibly a more recently derived
artificial group in which primitive characters have
reappeared secondarily through neoteny.

Introduction

The discovery of fossil birds with teeth was one
of the most dramatic events in 19th century pale¬
ontology. In 1861 a partial skeleton of the

Philip D. Gingerich, Museum of Paleontology, The Univer¬
sity of Michigan, Ann Arbor, Michigan 48104.

feathered Archaeopteryx was discovered in the
Jurassic deposits of Bavaria. In the next 16 years,
skeletons of Ichthyornis and Hesperornis were dis¬
covered in the Cretaceous of North America and a
more complete skeleton of Archaeopteryx was
found in Germany. Surprisingly, the jaws of each
of these birds bore reptile-like teeth. Being dis¬
covered only a few years after publication of The
Origin of Species, toothed birds were much dis¬
cussed in connection with Darwin’s evolutionary
hypothesis.

As spectacular as the original discoveries were,
it is remarkable in retrospect how little detailed
study was made of the actual specimens until rel¬
atively recently. The history of the original dis¬
coveries of toothed birds, the initial recognition of
their evolutionary significance, and their subse¬
quent fate are reviewed here. The whole provides
an interesting historical comment on the treatment
of intermediate forms that do not conform to pre¬
conceived archetypical categorizations.

Acknowledgments.— I should like to acknowl¬
edge here the encouragement Dr. Wetmore gave to
continued study of the Yale collection of Mesozoic
birds when work was initiated on Hesperornis sev¬
eral years ago. My study of Hesperornis began,
curiously enough, as a tutorial with K. S. Thomson
on kinesis and jaw mechanics in fishes. Expanding
the range of comparisons, the Mesozoic bird ma¬
terial at Yale was examined to determine the form
of kinesis of primitive birds. When no simple an¬
swer was forthcoming, J. H. Ostrom authorized
Peter Whybrow to undertake further preparation
of the original specimens. Thus I am particularly
indebted to Professors Thomson and Ostrom
and to Mr. Whybrow for their assistance and
encouragement.

23

24

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

In addition, Drs. C. C. Black, T. H. Eaton, and
L. D. Martin permitted an extended loan of the
University of Kansas skull of Hesperornis. Drs.
C. E. Ray, Nicholas Hotton III, and Mr. Robert
Purdy allowed me to borrow the cranial material of
Hesperornis in the National Museum of Natural
History. Drs. W. J. Bock, Pierce Brodkorb, Peter
Dodson, J. T. Gregory, Hildegard Howard, P. V.
Rich, and M. V. Walker have all assisted one way
or another as well.

I also thank Mrs. Gladys Newton for typing the
manuscript, and Mr. Karoly Kutasi for printing
the illustrations. Margaret Egan read the manu¬
script and her comments have improved the paper
considerably.

Mesozoic Birds with Teeth

it is now generally admitted by biologists who have made
a study of the vertebrates, that Birds have come down to us
through the Dinosaurs, and the close affinity of the latter
with recent Struthious Birds will hardly be questioned. The
case amounts almost to a demonstration, if we compare, with
Dinosaurs, their contemporaries, the Mesozoic Birds, The
classes of Birds and Reptiles, as now living, are separated
by a gulf so profound that a few years since it was cited
by the opponents of evolution as the most important break
in the animal series, and one which that doctrine could not
bridge over. Since then, as Huxley has clearly shown, this
gap has been virtually filled by the discovery of bird-like
Reptiles and reptilian Birds. Compsognathus and Archaeop¬
teryx of the Old World, and Ichthyornis and Hesperornis of
the New, are the stepping stones by which the evolutionist of
to-day leads the doubting brother across the shallow rem¬
nant of the gulf, once thought impassable. (O. C. Marsh,
1877:352).

In 1859, perhaps the gravest deficiency of Dar¬
win’s hypothesis of evolutionary descent was the
rarity of intermediate forms in the geological
record. Intermediate forms linking species into
graded chains or linking major groups of animals
to a common ancestor were at that time poorly
known. Evidence remedying this deficiency was
supplied in a most spectacular way by the discovery
of several intermediate forms linking birds to a
reptilian origin. Interestingly, each discovery of
itself was insufficient to overcome archetypical cate¬
gorizations of birds and reptiles, and a truly evo¬
lutionary view of both classes was necessary in
order to interpret literally the clear evidence for
bird-reptile relationships offered by the skeletons
of Compsognathus, Archaeopteryx, and Icthyornis.

J. A. Wagner (1861) described a remarkably
complete skeleton of a very small new dinosaur,
Compsognathus longipes, from the Jurassic litho¬
graphic limestone of Solenhofen, Germany. In the
same year H. von Meyer (1861) first published a
notice on the skeleton of a bird from the same de¬
posit, which he named Archaeopteryx lithograph-
ica. Having a dinosaurian skeleton, Compsogna¬
thus was clearly a variant of the “Reptile type.”
On the other hand, Archaeopteryx, with its dis¬
tinct impressions of feathers, was from the begin¬
ning regarded as a variant of the “Bird type.” In¬
fluenced at least in part by Darwin’s dynamic view
of evolution, T. H. Huxley was able to overcome
his contemporaries’ fixed categorizations, even of
groups as large as reptiles and birds, and he found
in Compsognathus a bird-like dinosaur, and in
Archaeopteryx the most reptilian of birds. Thus,
Huxley (1868) confirmed the Darwinian expecta¬
tion of intermediate forms linking birds and rep¬
tiles in the fossil record. Although the actual
common ancestor of living reptiles and birds had
not been found, Huxley judged from their morph¬
ology that late Jurassic birds and reptiles were
clearly much more closely related than their living
descendants seemed to suggest. This closer simi¬
larity of the early forms was itself strong evidence
favoring Darwin’s dynamic view of evolutionary
descent, as opposed to the then-prevailing view
that living “reptiles” and “birds” were static groups
persisting through time within some predetermined
bounds.

There was, however, a limit to the intermediate
position even Huxley would accept for Archaeo¬
pteryx. Thus, of the single skeleton of Archaeo¬
pteryx then known, he wrote “unfortunately the
skull is lost” (Huxley, 1868:70), making no men¬
tion of an earlier paper by Sir John Evans (1865)
describing a premaxilla with four teeth preserved
among the other bones of the specimen. Evans’
note (1865:421) quotes a letter from von Meyer
himself concerning the apparent association of a
toothed premaxilla with Archaeopteryx:

Teeth of this sort I do not know in the lithographic stone
.... From this it would appear that the jaw really belongs to
the Archaeopteryx. An arming of the jaw with teeth would
contradict the view of the Archaeopteryx being a bird or an
embryonic form of bird. But after all, I do not believe that
God formed his creatures after the systems devised by our
philosophical wisdom. Of the classes of birds and reptiles as
we define them, the Creator knows nothing, and just as little

NUMBER 27

25

of a prototype, or of a constant embryonic condition of the
bird, which might be recognized in the Archaeopteryx. The
Archaeopteryx is of its kind just as perfect a creature as
other creatures, and if we are not able to include this fossil
animal in our system, our short-sightedness is alone to
blame.

The presence of teeth in the bird Archaeopteryx
was apparently too reptilian a characteristic for
even Huxley to accept.

O. C. Marsh was the first to discover the un¬
equivocal presence of teeth in primitive birds,
though he too was at the outset apparently unable
to accept the evidence. In September 1872, Pro¬
fessor Mudge of Kansas presented Marsh with
some fossils from the Cretaceous Niobrara Chalk,
the formation from which Marsh had earlier de¬
scribed the headless skeleton of a large, flightless,
diving bird as Hesperornis regalis. Marsh studied
Mudge’s new fossils and in October published a
note describing the postcranial skeleton as a new
form of smaller volant bird, Ichthyornis dispar
(Marsh, 1872a). A month later he published
another note (Marsh, 1872b) on the jaws of a new
small “reptile,” Colonosaurus mudgei, found in
association with the remains of Ichthyornis. In the
same month that Colonosaurus was described (No¬
vember, 1872), Marsh’s assistant T. H. Russell
discovered a nearly perfect skeleton of Hesperornis,
again in the Niobrara Chalk. This new skeleton
included a skull with associated toothed jaws (Fig¬
ure 1). Immediately after the discovery of this
skeleton of Hesperornis, Marsh published a short
paper in February 1873 stating that the toothed
jaws of “Colonosaurus” actually belonged to
Ichthyornis. Of Ichthyornis dispar, Marsh (1873:
162) wrote:

When the remains of this species were first described, the
portions of lower jaws found with them were regarded by
the writer as reptilian; the possibility of their forming part
of the same skeleton, although considered at the time, was
not deemed sufficiently strong to be placed on record. On
subsequently removing the surrounding shale, the skull and
additional portions of both jaws were brought to light, so
that there cannot now be a reasonable doubt that all are
parts of the same bird.

Although no mention was then made of the
toothed jaws of Hesperornis, that discovery prob¬
ably provided Marsh with the necessary corrobora¬
tion for him to accept the previously evident asso¬
ciation of toothed jaws with Ichthyornis. Two years
after the toothed jaws of Hesperornis were first

described by Marsh (1875), the Berlin specimen of
Archaeopteryx was found about 10 miles from the
original Solenhofen discovery, and its feathers, rep¬
tilian skeleton, and toothed jaws left no doubt
about the reptilian ancestry of birds.

Beyond their importance in dramatically filling
a gap in the fossil evidence of evolution originally
available to Darwin, the three early avian fossils
Archaeopteryx, Ichthyornis, and Hesperornis are
of interest for another reason. Huxley (1868:74)
originally interpreted the great similarity of
Compsognathus as indicating a dinosaurian (more
specifically, coelurosaurian) origin of birds:

Surely there is nothing very wild or illegitimate in the hy¬
pothesis that the phylum of the Class Aves has its foot in
the Dinosaurian reptiles—that these, passing through a series
of such modifications as are exhibited in one of their phases
by Compsognathus, have given rise to the Ratitae —while the
Carinatae are still further modifications and differentiations
of these last, . . .

Similarly, Marsh (1880:189) saw in the skull of
Hesperornis certain resemblances to the “Ratitae,”
a group he regarded as being survivors of an evo¬
lutionary stage intermediate between reptiles and
the true “ornithic type.”

Three principal ideas have come out of the early
work of Huxley and Marsh: (1) that ratites are
survivors of a primitive stock of birds, (2) that
Hesperornis was similar to ratites, and (3) that
Hesperornis and Ichthyornis actually possessed
jaws with teeth. All three of these views have been
challenged in the century since their first publi¬
cation by Huxley and Marsh. Disagreement with
these ideas has come in part from authorities urg¬
ing caution in attempting any interpretation at all,
but in most cases a strong contrary interpretation
has been offered, usually without critical examina¬
tion of even the evidence available to Huxley and
Marsh. Advocating ratites as a derived group of
birds, reconstructing Hesperornis with a “neogna-
thous” skull, and denying the presence of teeth in
Icthyornis or Hesperornis have a common effect
—to deny the primitiveness and the reptilian char¬
acters of the best known Cretaceous birds and to
maintain a wide gulf between birds and reptiles.
This common effect of so many studies by post-
Darwinian evolutionary biologists can only be
ascribed to a deep-seated typological conception of
“birds” and “reptiles”—an interesting comment on
the pervasiveness of typological thinking.

26

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

The Skull of Hesperornis

Our knowledge of the structure of the skull in
Hesperornis is based almost entirely on three speci¬
mens: (1) that found by Marsh and Russell in
1872, now in the Yale Peabody Museum (YPM
1206) (Figure 1); (2) the premaxillae and mandi¬
bles of a skull in the National Museum of Natural
History, Smithsonian Institution (USNM 4978);
and (3) a nearly complete but crushed skull in
the collections of the University of Kansas (KU
2287) (Figures 2 and 3). The first of these skulls
was described and illustrated in some detail by
Marsh (1880:5-12, plates 1,2), and a brief descrip¬
tion of the last two was published by Lucas (1903),
who illustrated the quadrate and pterygoid of die
Kansas specimen and the lacrimal of the National
Museum specimen. The Yale and National Mu¬
seum specimens are very nearly the same size and
both have been identified as Hesperornis regalis
by virtually all workers. The Kansas specimen, on
the other hand, is slightly smaller than the other
two and was placed by Lucas (1903) in a new ge¬

nus, Hargeria, having as its type the species Hes¬
perornis gracilis Marsh. After extensive compari¬
son of the three skulls, I agree with Gregory (1952)
that all three are of the same genus, Hesperornis.
It remains an open question whether more than a
single species should be recognized.

The Yale skull was only partly removed from
the enclosing rock by Marsh, and those pot tions
that were freed for study were subsequently te-
mounted on the original slab for display pm poses.
Consequentlv, the specimen was not teally avail¬
able for examination until relatively recently, when
it was removed from public exhibition.

The Yale skull is in many respects the best one
for study, because its components were scattered
before fossilization and are now disarticulated and
very little crushed (except for the braincase). The
major portion of the Yale skull is illustrated hete
as it was mounted for exhibition (Figure 1). The
braincase and some smaller fragments were com¬
pletely removed from the rock by Marsh and it is
not certain that their positions as shown in Figure
1 are those in which they were found. The pre-

Fic.ijri 1.—The Yale skull of Hesperornis regalis Marsh (YPM 1206), showing the individual
disarticulated hones well preserved. Premaxilla, nasal, maxilla, and vomers are illustrated in the
position in which they weic found—all have subsequently been removed and cleaned for study,
(d = dentary, f = frontal, I = lacrimal, m = maxilla, n = nasal, pi = palatine, pm = pre¬
maxilla, q = quadrate, t = tooth, v = vomer.) Note presence of teeth in dentary, as illustrated
by Marsh (1880, pi. 1). (Approximately one-half natural size.)

NUMBER 27

27

maxilla, maxilla, nasal, vomers, and palatine, how¬
ever, were never removed and thus retain their
original orientation as buried. It should be noted
that Marsh had the nasal and maxilla exposed
from both sides of the slab, but they were never
completely removed. All of the important pieces
of the Yale skull were carefully removed from
their matrix in 1971 by Mr. Peter Whybrow, and
they can now be studied freely and articulated.

The cranium of the University of Kansas skull
of Hesperornis is also in a slab of Niobrara chalk,
but unlike the Yale specimen, it was preserved in
articulation and both the braincase and the maxil¬
lary portion of the skull have suffered considerable
crushing. Furthermore, Lucas (1903) reported that
the specimen was preserved with the skull doubled
backwards against the pelvis, and that portions of
both the dorsal and the sternal ribs were crushed
into the palate. It is possible to identify most of
the bones preserved in this specimen, but the max¬
illae are conspicuously lacking—whether they are
crushed beyond recognition into the palate or lost
entirely cannot be determined.

In addition to the portions illustrated in Figure
2, the Kansas specimen includes most of the lower
jaws, a complete left quadrate, and a complete left
pterygoid, which have been fully prepared and
can be articulated with each other and also with
the left palatine preserved with the main part of
the cranium. The quadrate and pterygoid were
illustrated by Lucas (1903, figs. 1,2; the left ptery¬
goid is incorrectly identified as a right pterygoid),
and they are illustrated here in articulation (Fig¬
ure 3). The complicated S-shaped surface of the
left pterygoid (“Apl” in Figure 3) articulates with
the S-shaped proximal end of the palatine (“Apt”
in Figure 2).

The principal contribution of the USNM speci¬
men to our understanding of the skull morphology
of Hesperornis is furnished by the nearly complete
left lacrimal (illustrated by Lucas, 1903, fig. 3).

By studying all three specimens it is possible to
reconstruct the major features of the morphology
of the rostrum, the palate, and the mandible (Fig¬
ure 4). The reconstruction has been discussed else¬
where (Gingerich, 1973), but some additional
notes are added here. These notes and the illustra¬
tions of the Yale and Kansas specimens (Figures
1-3) are preliminary to a more definitive descrip¬
tion of this important material. They are intended

to provide additional documentation of the re¬
markable completeness of the preserved specimens
and to answer, in part, some questions raised by
several skeptical colleagues.

The length of the reconstructed skull was de¬
termined from the Yale specimen (YPM 1206).
The dorsal surface of the braincase in this speci¬
men is crushed forward, but without affecting the
length from the occipital condyle to the anterior
end of the frontals. The overlapping articulation
between the nasal and the frontal is outlined on
the surface of the frontal, and the two can be fitted
together as in life. The nasal-premaxillary articu¬
lation is preserved in both of the elements and
these too can be fitted together accurately. As nei¬
ther frontals, nasals, premaxillae, nor the base of
the braincase appear in any way distorted in
length, a total length of 26-27 cm is estimated for
this skull.

Regarding the possibilities of cranial kinesis,
little can be added to my previous discussion
(Gingerich, 1973) except perhaps to add a more
cautionary note. Rhynchokinesis in Hesperornis is
almost certainly ruled out by the complete ring of
bone formed by the premaxillae and nasals around
the external narial opening. Some slight prokinetic
movement might have been possible if the pre¬
maxillae and nasals were capable of being lifted
off the frontals, although I know of no modern bird
with such thick bone in the region of bending,
and the complex interdigitation of the nasal and
lacrimal in Hesperornis would likewise limit
prokinetic movement. The quadrates were clearly
streptostylic, which appears to have been corre¬
lated with a unique form of maxillokinesis whereby
the maxillae were able to slide anteroposteriorly
on rails formed by the nasal-premaxillary subnarial
bars (Gingerich, 1973). While I am reluctant to
postulate a form of kinetic motion so distinctive
from that of any other animal, the preserved oste¬
ology of the rostrum in Hesperornis is unique and
its adaptations were clearly different from those of
any known vertebrate. Maxillokinesis appears to
explain several unique features of the known fossil
material.

One of the most curious features of the upper
jaw of Hesperornis is the fact that the premaxilla
bore a horny sheath as in modern birds (indicated
by the’ vascular nature of the underlying bone),
while the teeth were confined to the maxillae

28

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Figure 2.—The Kansas skull of Hesperornis (KU 2287), ventral view as preserved, articulated on
a slab of Niobrara Chalk. Note particularly the little-disturbed contact between the premaxillae
and nasals, while the maxillae are completely missing. (Apt = pterygoid articulation of palatine,
Aq = quadrate articulation of squamosal, n = nasal, pi = palatine, pm = premaxilla, v =
vomer; approximately two-thirds natural size.)

NUMBER 27

29

Figure 3. —Articulated left pterygoid (pt) and quadrate (q) of Kansas specimen of Hesperornis
(KU 2287): a, medial view; b, lateral view. Note particularly the complicated articulation
between quadrate and pterygoid, the broad basisphenoid articulation of the pterygoid, and the
complicated S-shaped articulation of the pterygoid with the palatine. (Abs = basisphenoid
articulation of pterygoid, Am = mandibular articulation of quadrate, Apl = palatine articula¬
tion of pterygoid, Aqj = quadratojugal articulation of quadrate, Asq = squamosal articulation
of quadrate; twice natural size.)

proper. The lower jaw bore teeth throughout the
length of the dentary. Secondly, in both the Yale
and Kansas specimens, the maxillae have clearly
separated from the nasal-premaxillary subnarial
bars while, at least in the Kansas specimen, the
subnarial bars were little disturbed by crushing. It
should be noted also that the anterior end of each
maxilla was grooved to fit over anteroposteriorly
aligned keys or ridges of bone on the ventral sur¬
face of the premaxilla. This system of locking
would keep the anterior ends of the maxillae from
dropping away from the subnarial bars, while per¬
mitting anteroposterior motion of the maxillae
relative to the subnarial bars. Finally, it now seems
unlikely that the left and right vomers were fused
to each other at their anterior ends. Such fusion
would have prevented independent motion of the
left and right maxillary segments of the palate
relative to each other. The only possible functional
advantage of having the kind of maxillary kinesis
postulated here would be in moving each side inde¬
pendently. As evidenced by the unfused mandi¬
bular symphysis, such independent movement of
the lower jaws was clearly possible. Independent
movement of the maxillae would further expand

the range of possible movements used in ingesting
prey, which in this case was almost certainly fish.

A new specimen of Archaeopteryx, described re¬
cently by Wellnhofer (1974), fortunately has a
relatively well-preserved skull. Wellnhofer (1974:
185) interprets the skull as being definitely kinetic,
but in Archaeopteryx, as in Hesperornis, it is dif¬
ficult to see where bending that would lift a signi¬
ficant portion of the rostrum could have taken
place. Wellnhofer favors bending in the dorsal
processes of the premaxillae, but at most this would
lift only the tip of the upper jaw. Kinesis approach¬
ing that of modern birds seems not to have been
present in either Archaeopteryx or Hesperornis.

The present evidence bearing on Huxley’s and
Marsh’s conclusions regarding the evolutionary
position of the ratites, the relationship of Hesper¬
ornis to the ratites, and the presence of teeth in
Hesperornis and Ichthyornis can now be consid¬
ered. The skeleton of Archaeopteryx is more rep¬
tilian than avian, and the uncontestecl fact that its
jaws bear teeth is easy to believe. The skeletons of
Hesperornis and Ichthyornis, on the other hand,
are mdre typically avian. That a bird with an
avian postcranial skeleton should have jaws with

30

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

teeth has proved more difficult for some ornitholo¬
gists to accept. The quadrate is not preserved in
the original specimen of Ichthyornis and the
toothed jaws that Marsh found associated with
this skeleton thus cannot be articulated with the
remainder of the cranium. The articular regions of
the original jaws are also badly distorted. Gregory
(1952) made a careful study of the lower jaws of
Ichthyornis and concluded that they belonged to a
small mosasaur. Therefore Hesperornis alone was
left with the combination of toothed jaws and a
nearly typically avian skeleton. Inevitably, the as¬
sociation of teeth with the skull of Hesperornis was
also questioned. Bock (1969) claimed that the
teeth found with Hesperornis were not in place in
the jaws, but scattered and cemented with matrix
onto the skull. However, one need only examine
the Yale specimen to see that teeth are preserved
in the jaws as well as being scattered through the
matrix (Figure 1). Discovery of a new, uncrushed
posterior portion of a mandible of Ichthyornis
(Gingerich, 1972), and its comparison with the
mandibles of the original specimen and with those
of Hesperornis and modern birds, leaves little
doubt that Marsh was correct in associating toothed
jaws with Ichthyornis.

Interpretation of the structure of the palate in
Hesperornis has had an interesting history. Marsh
(1880:6) originally determined that the palate re¬
sembled most closely that of “Struthious” birds,
but he confused the vomers with the palatines of
his specimen of Hesperornis and gave no figure or
reconstruction of the palate. Thompson (1890),
followed by Lucas (1903), Shufeldt (1915), and
Heilmann (1926), challenged Marsh’s interpreta¬
tion of Hesperornis as indicating any relationship
to the ratites. In the course of the 36 years from
1890 to 1926, the palatal structure of Hesperornis
“evolved" rapidly in the literature, ultimately
"converging" toward the neognathous palatal type
of the modern loon ( Gavia ), a fish-eating, diving
bird with certain similar locomotor adaptations.

Fortunately, the Yale and Kansas specimens of
Hesperornis (Figures 1-3) preserve virtually in¬
tact at least one example of each of the palatal
bones. The quadrate and pterygoid are complete
in the Kansas specimen, portions of both vomers
are present in the Yale specimen (Figure 1), a
crushed left vomer remains in the Kansas speci¬
men (Figure 2), and virtually complete palatines

are preserved in both. About midway along their
length, a rounded surface is present on the medial
side of the vomers, which apparently articulated
with the parasphenoid rostrum. The left maxilla is
preserved in the Yale specimen (Figure 1) and it
fits together with, and is overlapped by, the left
vomer, as shown in Figure 4. There appears to be
an articular facet on a ventrolateral expansion of
the vomer for the narrow anterior end of the pala¬
tine (Figure 4). It is possible, but unlikely, that
the palatines articulated directly with posterior
projections of the maxillae (not preserved) rather
than with the vomers. As noted above, the maxillae
articulated with the subnarial bars formed by the
premaxillae and nasals. Returning to the pterygoid-
quadrate complex, it should be noted that each
pterygoid bears a large, round, flat surface that
articulates with a “basipterygoid" process of the
basisphenoid (Figure 3, “Abs”).

The entire reassembled palate is illustrated in
Figure 4c. Compared with that of living ratites,
the palate of Hesperornis is obviously different
from an emu or an ostrich in being much longer
and narrower. This lengthening has clearly been
accomplished by elongation of the premaxillae,
maxillae, vomers, and palatines relative to the more
posterior elements of the skull. Although having
adaptations quite different from those of any living
palaeognathous bird, Hesperornis shares with
palaeognathous birds all essential palatal characters
that distinguish them from neognathous birds:
(1) a relatively large vomer, (2) a firm pterygoid-
palatine connection, (3) palatines widely separated
from the sphenoid rostrum by the pterygoids, (4)
strong basipterygoid processes of the sphenoid ar¬
ticulating with the pterygoids, and (5) a complex
pterygoid-quadrate articulation including portions
of the orbital process of the quadrate (Figure 3).

The structure of the palate is still unknown in
Archaeopteryx, but the presence of a palaeogna¬
thous palate in Hesperornis would appear to be
strong evidence favoring the view that the pala¬
eognathous conformation is primitive in birds. Ad¬
ditional evidence bearing on the primitive struc¬
ture of the palate of birds is offered by this
structure in theropod dinosaurs. Ostrom (1973)
has compared the skeleton of Archaeopteryx with
that of reptiles and concluded that birds originated
from theropod dinosaurs, more specifically, from a
coelurosaurian stock of theropods. The palatal

NUMBER 27

31

Figure 4.—Reconstructed skull (a) and mandible ( b ) of Hesperornis regalis in lateral view; c,
reconstructed palate in ventral view, (a = angular, ar = articular, bs = basisphenoid, d =
dentary, f = frontal, j = jugal, 1 = lacrimal, m = maxilla, n = nasal, pi = palatine, pm
= premaxilla, pt = pterygoid, q = quadrate, sa = surangular, sp = splenial, v = vomer)
(From Gingerich, 1973.)

structure is not known in any coelurosaur, but it
is completely preserved in the large carnosaur
Tyrannosaurus (Osborn, 1912) and less well pre¬
served in the smaller Dromaeosaurus (Colbert and
Russell, 1969) and Deinonychus (Ostrom, 1969).
The structure of each of these skulls appears to
meet all of the criteria listed above for the palaeo-
gnathous palate. Osborn (1912:11) noted this
“analogy” implicitly in comparing the palate of
Tyrannosaurus with that of a cassowary. The pres¬
ence of a palaeognathous palate in Mesozoic thero-
pods, the “sister group” of birds, together with the
palaeognathous palate of the Cretaceous bird
Hesperornis, should leave little doubt that this
palatal conformation is truly primitive in birds.

I emphasize the strength of the evidence in this
case because Cracraft (1974) has proposed that the
living ratite birds are cladistically a “strictly
monophyletic” group on the basis of their “de¬

rived” palaeognathous palate, their unique rham-
phothecal structure, and their large ilioischiatic
fenestra. Cracraft asserts that the palaeognathous
palate is a derived state in birds, not a primitive
one, because “it is restricted to a small number of
species within this large class” (Cracraft, 1974:497).
This specious reasoning would lead one to assume
that teeth in Mesozoic birds are a derived condi¬
tion also, an unlikely hypothesis.

The unique rhamphothecal structure and other
resemblances of ratites and tinamous were inter¬
preted by Parkes and Clark (1966) rather less
stringently than Cracraft now proposes. They
(1966:469) noted that “resemblances are to be at¬
tributed to parallel evolution from a common
stock . . rather than to convergence from unre¬

lated stocks, and thus, employing Simpson’s con¬
cepts,' the group may be considered monophyletic.”
The resemblance in rhamphothecal structure of

32

ratites and tinamous provides no evidence that this
group is strictly monophyletic in Cracraft’s sense
rather than monophyletic in G. G. Simpson’s sense
(i.e., possibly paraphyletic, if indeed the unique
rhamphothecal structure is a derived state at all—
it may very well be primitive).

The third character Cracraft (1974:505) cites as
evidence that ratites and tinamous are “each
other's closest relatives’’ is their possession of a
large ilioischiatic fenestra. Archaeopteryx has long
been known to have a large ilioischiatic fenestra
(see for example Petronievics and Smith Wood¬
ward, 1917), and Cracraft (1974:503) himself notes
that this is the condition in Hesperornis and
Ichthyornis. In short, of the three “derived” char¬
acters cited by Cracraft (1974), the first and third
are almost certainly primitive and the second may
be primitive as well.

Evidence that ratites are strictly monophyletic
remains to be discovered and it is possible, even
probable, that the groups of living ratites and the
tinamous are paraphyletic. Huxley (1867:419) en¬
visioned the living palaeognathous ratites as
“waifs and strays” of an early radiation of birds,
the neognathous types representing a subsequent
radiation. Judging from the fossil record, succes¬
sive adaptive radiations replacing older stocks by

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

newer ones are common in vertebrate evolution,
and the class Aves is no exception. Although they
are sometimes highly modified from the ancestral
stock, we are fortunate to have in many groups of
vertebrates surviving “waifs and strays,” and still
more fortunate to have well-preserved archaic fossil
forms. In the absence of a more complete fossil
record, some question must remain as to whether
the modern ratites and tinamous are in fact sur¬
vivors of a primitive radiation of birds, or whether
their primitive characteristics are neotenic solu¬
tions to particular adaptive problems, since both
the palaeognathous palate and the open ilioischi¬
atic fenestra appear to be present in the develop¬
mental stages of modern nonratite birds (Jollie,
1958; Olson, 1973:35-36). To explain away the
primitive morphology of Hesperornis and ally it
with modern loons and grebes (Cracraft, 1974:497,
503), however, illustrates on the one hand the
arbitrary nature of the cladistic method of recon¬
structing a phylogeny, and on the other hand ex¬
emplifies another typological attempt to force an
archaic bird into a modern morphological category.
To paraphrase von Meyer (1861), if Hesperornis
does not fit our philosophical wisdom and if we are
not able to include this fossil in our system, our
shortsightedness is alone to blame.

Literature Cited

Bock, W. J.

1969. The Origin and Radiation of Birds. New York
Academy of Sciences Annals, 167:147-155.

Colbert, E. H., and D. A. Russell

1969. The Small Cretaceous Dinosaur Dromaeosaurus.
American Museum Novilates, 2380:1-49, 15 figures.

Cracraft, J.

1974. Phylogeny and Evolution of the Ratite Birds. Ibis,
116:494-521, 10 figures.

Evans, J.

1865. On Portions of a Cranium and of a Jaw, in the
Slab Containing the Fossil Remains of the Ar¬
chaeopteryx. Natural History Review, new series,
5:415—421, 1 figure.

Gingerich, P. D.

1972. A New Partial Mandible of Ichthyornis. Condor,
74:471-473, 2 figures.

1973. Skull of Hesperornis and the Early Evolution of
Birds. Nature, 243:70-73, 2 figures.

Gregory, J. T.

1952. The Jaws of the Cretaceous Toothed Birds, Ich¬
thyornis and Hesperornis. Condor, 54:73-88, 9 fig¬
ures.

Heilmann, G.

1926. The Origin of Birds. 208 pages. London: H. F. and
G. Witherby.

Huxley, T. H.

1867. On the Classification of the Birds; and on the
Taxonomic Value of the Modifications of Certain
of the Cranial Bones Observable in that Class.
Proceedings of the Zoological Society of London,
1867:415-472, 36 figures.

1868. On the Animals Which Are Most Nearly Inter¬
mediate between Birds and Reptiles. Annals and
Magazine of Natural History, 4th series, 2:66-75.

Jollie, M.

1958. Comments on the Phylogeny and Skull of the
Passeriformes. Auk, 75:26-35.

Lucas, F. A.

1903. Notes on the Osteology and Relationship of the
Fossil Birds of the Genera Hesperornis, Hargeria,
Baptornis, and Diatryma. Proceedings of the United
States National Museum, 26:545-556, 8 figures.

Marsh, O. C.

1872a. Notice of a New and Remarkable Fossil Bird.
American Journal of Science, 3rd series, 4:344.

NUMBER 27

33

1872b. Notice of a New Reptile from the Cretaceous.
American Journal of Science, 3rd series, 4:406.

1873. On a New Subclass of Fossil Birds (Odontornithes).
American Journal of Science, 3rd series, 5:161-162.

1875. On the Odontornithes, or Birds with Teeth. Amer¬
ican Journal of Science, 3rd series, 10:402-408,
plates 9-10.

1877. Introduction and Succession of Vertebrate Life in
America. American Journal of Science, 3rd series,
14:337-378.

1880. Odontornithes: A Monograph on the Extinct
Toothed Birds of North America. Volume 7 of
Report of the Geological Exploration of the 40th
Parallel, xv + 201 pages, 34 plates, 40 figures.
Washington: Government Printing Office,
von Meyer, H.

1861. Archaeopteryx lithographica (Vogel-Feder) und
Pterodactylus von Solenhofen. Neues Jahrbuch fur
Mineralogie, Geologie, und Palaeontologie, 1861:
678-679.

Olson, S. L.

1973. Evolution of the Rails of the South Atlantic Is¬
lands (Aves: Rallidae). Smithsonian Contributions
to Zoology, 152:1-53, 11 plates, 8 figures.

Osborn, H. F.

1912. Crania of Tyrannosaurus and Allosaurus. Memoirs
of the American Museum of Natural History, new
series, 1:1-30, plates 1-4, 27 figures.

Ostrom, J. H.

1969. Osteology of Deinonychus antirrhopus, an Unusual

Theropod from the Lower Cretaceous of Montana.
Yale University, Peabody Museum of Natural His¬
tory Bulletin, 30:1-165, 83 figures.

1973. The Ancestry of Birds. Nature, 242:136.

Parkes, K. C., and G. A. Clark

1966. An Additional Character Linking Ratites and Tina-
mous, and an Interpretation of their Monophyly.
Condor, 68:459-471, 7 figures.

Petronievics, B., and A. Smith Woodward

1917. On the Pectoral and Pelvic Arches of the British
Museum Specimen of Archaeopteryx. Proceedings
of the Zoological Society of London, 1917:1-6, 1
plate.

Shufeldt, R. W.

1915. On a Restoration of the Base of the Cranium of
Hesperornis regalis. Bulletins of American Paleon¬
tology, 5:73-85, plates 1-2.

Thompson, D’A. W.

1890. On the Systematic Position of Hesperornis. Uni¬
versity College, Dundee, Studies from the Museum
of Zoology, 10:1-15, 17 figures.

Wagner, J. A.

1861. Neue Beitrage zur Kenntnis der urweltlichen Fauna
des lithographischen Schiefers, II: Schildkroten
und Saurier. Abhandlungen der Bayerischen Akad-
emie der Wissenschaften, Munchen, 9:65-124,
plates 1-6.

Wellnhofer, P.

1974. Das Fiinfte Skelettexemplar von Archaeopteryx.
Palaeontographica, Abteilung A, 147:169-216, plates
20-23, 13 figures.

The Skeleton of Baptornis advenus
(Aves: Hesperornithiformes)

Larry D. Martin and James Tate, Jr.

ABSTRACT

Baptornis advenus is a foot-propelled diving bird
from the Late Cretaceous of Kansas. It was slightly
larger than the largest living loon and had an un¬
usually long neck. The feet were large, with only
slight modifications for toe-rotation. In this and
many other respects, Baptornis was a less specialized
diving bird than the contemporaneous Hesperor-
nis. However, examination of almost the entire
skeleton shows that Baptornis is more closely re¬
lated to Hesperornis than to any living diving bird
and should be included in the order Hesperor¬
nithiformes. It should not be regarded as the
earliest record of the Podicipediformes. We main¬
tain Baptornis in a family Baptornithidae separate
from Hesperornithidae. Both Hesperornis and
Baptornis are in many respects very primitive
birds, which in some characters appear to be little
modified from Archaeopteryx.

Introduction

In 1964 we discovered a previously unrecognized
partial skeleton of Baptornis advenus in the col¬
lections of the University of Nebraska State Mu¬
seum. This specimen was more complete and bet¬
ter preserved than any other known example of
the species and it encouraged us to review all
known specimens of the form. In other institutions
we found numerous examples of Baptornis. We
were able to study all of these except a tarsometa-
tarsus reported by Lambrecht (1933) to be in

Larry D. Martin, Museum of Natural History and Depart¬
ment of Systematics and Ecology, University of Kansas, Law¬
rence, Kansas 66045. James Tate, Jr., P.O. Box 2043, Den¬
ver, Colorado 80201.

Germany. We thus were able to reconstruct the
skeleton of Baptornis advenus, apart from most of
the skull and jaws.

After Hesperornis and Ichthyornis, Baptornis
is now the best known Cretaceous bird. It was first
described by Marsh (1877) as a new swimming
bird allied to Hesperornis, but readily separable
from that genus in having the third and fourth
trochleae of the tarsometatarsus of about equal
size. He did not illustrate the type, but did give
detailed measurements and commented that the
bird was about the size of a loon and may have had
similar habits. In 1880, Marsh illustrated the holo-
type tarsometataisus and mentioned a referred
femur and tibiotarsus. Lucas (1903:553-555) de¬
scribed a partial skeleton housed at the University
of Kansas, and illustrated the coracoid, humerus,
radius, ulna, and patella. Shufeldt (1915:9-11,
figs. 1-6) published photographs of the holotype
and discussed it in great detail. Finally, Lambrecht
(1933:258-260) summarized what was then known
about the anatomy of Baptornis.

After Lambrecht, xelatively little has been pub¬
lished on Baptornis, although Swinton (1965)
mentioned the genus briefly in his semipopular
book Fossil Birds, and suggested that the Creta¬
ceous diving bird Lonchodytes might be related.
Storer (1958), in his discussion of evolution in
diving birds, made some remarks concerning the
evolutionary position of Baptornis, although he
mistakenly spelled Baptornis and Baptornithidae
as “Bathornis” and “Bathornithidae". Walker
(1967) reviewed some of the material and pre¬
vious work, but erroneously assigned an alleged
large humerus to Baptornis. Other than the afore¬
mentioned works, the published references to Bap¬
tornis are restricted to checklists and catalogs

35

36

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

(Wetmore, 1956; Brodkorb, 1963b), and discussion
of these (Storer, 1965).

A close relationship between Baptornis and
Hesperornis was first suggested by Marsh (1877)
and went essentially unquestioned until Brodkorb
(1963b) placed the former in the Podicipedi-
formes. Brodkorb’s arrangement has been chal¬
lenged by Storer (1971) and by Martin and Tate
(1969) on the basis of the new material discussed
in detail below. Relationships also have been sug¬
gested between Baptornis and numerous other
Mesozoic and Tertiary foot-propelled diving birds
including Enaliornis, Lonchodytes, Hesperornis,
Neogaeornis, and Eupterornis. Those with Enali¬
ornis and Hesperornis seem best established, as
Baptornis seems to differ from these two genera
only at the familial level, whereas it seems to have
no affinities with the Podicipediformes.

Distribution. — Baptornis is known only from
Kansas (Gove, Logan, and Wallace counties) and
only from the Smoky Hill (upper) Member of the
Niobrara Chalk Formation (Coniacian). Other
North American Cretaceous birds ( Ichthyornis
and Hesperornis) are known from the present Gulf
of Mexico to above the Arctic Circle in Canada
(Russell, 1966), and range in age from the Turon-
ian, Greenhorn Formation, to the Campanian,
Pierre Shale (Walker, 1967; Martin and Tate, 1967).
Baptornis is known only from marine deposits. A
specimen from Logan County (KUVP 16112) and
one from Wallace County (YPM 5768) are from
very immature individuals, suggesting that Baptor¬
nis may have bred in the vicinity. Young specimens
of Hesperornis are uncommon, although Russell
(1967) reported several examples of subadult
Hesperornis from a bituminous marine shale along
the Anderson River in Canada (latitude 69° N)
and suggested that a nesting colony might have
existed nearby.

The Niobrara Chalk is a carbonate deposit with
no evidence of associated continental sediments.
The absence of a shoreline would imply that
Pteranodon, Ichthyornis, Hesperornis, and Bap¬
tornis were accustomed to venturing many hun¬
dreds of miles into open sea. Evidence for a nearby
shoreline in Kansas in the Coniacian is somewhat
sketchy and not completely convincing. The Nio¬
brara deposits do seem to thin, in an easterly direc¬
tion, toward areas that are known to have been
continental in the Early Cretaceous. It also has

produced a few examples of dinosaurs which are
presumed to have floated out to sea after death.
How far these dinosaurs may have floated is pres¬
ently unknown, but one wonders if it could have
been hundreds of miles in such a scavenger-rich
sea.

Material Examined. —All available specimens
were studied. These included partial skeletons at
the University of Nebraska State Museum, UNSM
20030; Field Museum of Natural History, FMNH
395; American Museum of Natural History,
AMNH 5101; University of Kansas Museum of
Natural History, KUVP 2290 and 16112; Yale Pea¬
body Museum, YPM 1465 and 5768. We were also
permitted to examine some additional material at
the Fick Fossil Museum, Oakley, Kansas. Most of
the available hesperornithid specimens were also
examined, and the extensive skeletal collection of
recent birds at the University of Kansas Museum
of Natural History was used for comparisons.

Acknowledgments.— The authors are deeply in¬
debted to the following persons for permission to
use specimens in their care: C. B. Schultz, Univer¬
sity of Nebraska State Museum; R. Zangerl, Field
Museum of Natural History; E. Simons and J.
Ostrom, Peabody Museum of Natural History,
Yale University; M. Walker, Sternberg Memorial
Museum; R. Shaeffer, American Museum of Nat¬
ural History. We have also benefitted from conver¬
sations with P. Brodkorb, R. W. Storer, and J.
Cracraft. M. A. Jenkinson read the manuscript
and offered many helpful suggestions. The skeletal
restoration was prepared by M. Tanner, the life
restoration by B. Dalzell, and the other drawings
by M. Tanner, D. K. Bennett, and D. Brennfoerder.

Class AVES

Subclass ODONTOHOLCAE Stejneger, 1885

Order HESPERORNITHIFORMES
(Fiirbringer), 1888

Amended Diagnosis.— Foot-propelled diving birds
with teeth; skull paleognathous (at least in
Hesperornis ); mandibular symphysis not fused; re¬
duced wings; coracoid with glenoid facet on tip of
scapular end; clavicles unfused; sternum flat;
patella large and perforated for tendon of ambiens
muscle; posterior extremities of ilium, ischium, and

NUMBER 27

37

pubis separate; tibiotarsus lacking supratendinal
bridge; tarsometatarsus lacking hypotarsal grooves
and proximal foramina.

Family BAPTORNITHIDAE American
Ornithologist’s Union, 1910

Included Genera.— Baptornis and Neogaeornis.

Amended Diagnosis. —Foot-propelled diving birds
with fully heterocoelous vertebrae; uncinate proc¬
esses turned dorsally (straight in Hesperornis);
coracoid more slender than in hesperornithids;
pelvis with the preacetabular portion of the ilium
relatively longer than in hesperornithids; patella
pyramidal in shape (much more laterally com¬
pressed in Hesperornis ); intracentral bones not
fused to caudal vertebrae; pygostyle long and later¬
ally compressed; outer trochlea of tarsometatarsus
not enlarged; distal foramen on tarsometatarsus an
open groove; toe-rotation not well developed.

Genus Baptornis Marsh, 1877

Type-Species. —Baptornis advenus.

Included Species. —Type-species only.

Amended Diagnosis. —Foot-propelled diving bird
about the size of a Yellow-billed Loon ( Gavia
adamsii ); neck elongate; wing greatly reduced, but
radius and ulna present; tarsometatarsus not as
compressed as in Neogaeornis.

Baptornis advenus Marsh, 1877

Lectotype.— Distal end of tarsometatarsus, YPM
1465.

Horizon. —Smoky Hill Member of Niobrara
Chalk, Late Cretaceous.

Type-Locality.— Wallace County, Kansas.

Diagnosis.— Same as for genus.

Morphology

Skull.— The skull of Baptornis is known from
only a few fragments. If teeth were present they
were restricted to the maxillae and dentaries, as in
Hesperornis. KUVP 16112 includes a fragment of
the bill just anterior to the nasal openings. There
are fairly large, distinct, triangular grooves on the
lateral sides of the fragment, extending towards
the tip. The sides of the bill are quite thick and

vaulted. There is no evidence that the bill came to
a point to form a spear as in the grebe Aechmo-
phorus. A fragment of the left side of the bill
(AMNH 5101) is probably from a position just
posterior to that of KUVP 16112. This fragment
shows a single large, elongate, longitudinal nu¬
trient foramen. In KUVP 16112 a large number of
smaller foramina extend over the dorsal surface of
the bill as in the top of the bill in loons. Appar¬
ently the tip of the bill was relatively short and
broad, being shaped like that of Hesperornis
(Marsh, 1880, pi. 1). It was probably covered by a
horny sheath.

AMNH 5101 includes a fragment of the frontal
bone which is very difficult to orient (Figure la).
The dorsal surface bears a faint scroll-like pattern
similar to that on the frontal bone of grebes. How¬
ever, the frontal-parietal suture is present and the
sagittal groove extends up to it, thus extending
further posteriorly than in the grebes and indicat¬
ing that the top of the skull was more similar to
that of loons. The cerebral hemispheres of the
brain seem to have been expanded, as in modern
birds, but better material is needed to confirm this.

The ventral half of a right quadrate (Figure
Ic-e) is preserved with AMNH 5101. The shaft
is broad but not as massive as in Gavia. The
orbital process has been broken off, but it had a tri¬
angular base and originated very low on the quad¬
rate. A prominent pit is present on the dorso-
medial margin of the base of the orbital foramen,
but the quadrate was not pneumatic. There is a
small, rectangular facet for the articulation of the
pterygoid. This facet is elongated dorsoventrally.
Lateral to the pterygoid articulation is a shallow,
rectangular depression from the middle of which a
low ridge connects to the socket for the quadrato-
jugal. The socket for the quadratojugal is large,
shallow, and somewhat triangular in shape
(rounded and very deep in loons and grebes). In
loons and grebes, the mandibular articulation is
divided into two parts by a groove, and the medial
facet runs anteroposteriorly. Baptornis resembles
Hesperornis in having a single comma-shaped
facet running from the pterygoid articulation.

The few fragments preserved are suggestive of a
skull close to that of Hesperornis. Hesperornis and
Baptornis have very similar quadrates, and both
differ from loons and grebes in most details of that
bone.

38

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Figure 1. —Skull elements of Baptornis advenus: a, fragment of the frontal (AMNH 5101), dorsal
view; b, stereophotographs of posterior fragment of the left ramus (FMNH 395), dorsal view;
c, stereophotographs of ventral half of the right quadrate (AMNH 5101), anterior view; d, same,
ventral view; e, posterior view. (All X 4.)

Mandibular Ramus.— FMNH 395 includes the
posterior portion of the left mandibular ramus
with the articular cotyla (Figure 15). It is slightly
crushed dorsoventrally. The only other Cretaceous
birds for which any part of the jaw is known are

Hesperornis and Ichthyornis (excluding Caenagna-
thus, which is not a bird).

In Baptornis, the surangular is tightly fused to
the articular, with a groove near the ventrolateral
border of the specimen along the suture between

NUMBER 27

39

these two bones. The articular cotyla is divided
into anterolateral and posteromedial sections,
which are separated by a low ridge. A groove,
which includes the articular foramen, lies on the
margin of the posterior cotyla and runs under the
internal articular process. Anterior to the posterior
articular surface is a small depression and a wide,
deep groove, which forms at the junction of the
surangular and the articular.

The articular surfaces are of about the same
shape and occupy approximately the same position
as those on the ramus of Hesperornis (Gregory,
1952, fig. 7). They differ from the articular cotylae
of the mosasaurs Clidastes and Platycarpus in hav¬
ing a concave, oblique, and elongate posterior
articular surface and in the fusion of the jaw ele¬
ments. Like Hesperornis, Baptornis lacks a depres¬
sion for the condyle of the quadrate. This depres¬
sion is present on the jaws of modern diving birds.
In the Double-crested Cormorant ( Phalacrocorax
auritus), for example, it is centrally located and
occupies nearly the entire posterior articulation. A
posterior articular foramen is similarly located in
Hesperornis and modern species of birds, but is
absent in the mosasaurs. The posterior portion of
the ramus of Baptornis is remarkably similar to
that illustrated for Ichthyornis (Gingerich, 1973;
fig. 2).

The mandibular articulation of Baptornis is
very similar to that of both Hesperornis and Ich¬
thyornis (Figure 2.) The jaws of Cretaceous
toothed birds are more similar to one another than
is generally supposed. Hesperornis shares with
Ichthyornis the following features: teeth flattened
from side to side, with expanded bases; anterior
end of dentary blunt and symphysis not fused;
teeth restricted to dentaries and maxillae; quadrate
articulation double. Gingerich (1972:472) lists
several additional features shared by Ichthyornis
and Hesperornis in the posterior portion of the
ramus. In so far as we can tell, Baptornis also
shares these characters, and although it cannot yet
be proven, we suspect that it was toothed.

Presacral Vertebrae (Table 1).—AMNH 5101
contains the most complete vertebral series of Bap¬
tornis, with at least 22 presacral vertebrae present.
UNSM 20030 includes the first eight vertebrae an¬
terior to the sacrum, and at least one of these is ab¬
sent from AMNH 5101 suggesting that the total
number of presacral vertebrae in Baptornis may

Figure 2.—Posterior portions of left rami: a, Ichthyornis
(from Gingerich, 1972), dorsal view, X 2.8; b, Baptornis
advenus (FMNH 395), dorsal view, X 4; c, same, medial
view, X 4.

have been 23, the same as in Hesperornis. Most of
the vertebrae with AMNH 5101 are fragmentary,
however, which may cause errors in interpreting
their number and position. It is possible that
Baptornis may have had 24 or 25 vertebrae, but
the number could not have been less than 23.
There are four thoracic and parts of six cervical
vertebrae with KUVP 16112. KUVP 2290 also con¬
tains four thoracic and six cervical vertebrae.
FMNH 395 includes the last unfused thoracic ver¬
tebra. Three thoracic and two cervical vertebrae
in AMNH 5101 and three cervicals with KUVP
2290 are preserved in articulation. These are the
only natural associations available. The relative
positions of the other vertebrae have been deter¬
mined by comparisons with Hesperornis.

Only the cervicals reported on by Lucas (1903:
553) have previously been described. He pointed

40

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Table 1.—Measurements (mm) of the thoracic vertebrae of Baptornis numbered from the synsacrum foreward

First

Second

Third

Fourth

Fifth

Sixth

Character

UNSM

20030

FMNH

395

UNSM

20030

KUVP

2290

UNSM

20030

KUVP

2290

UNSM

20030

KUVP

2290

UNSM

20030

KUVP

2290

UNSM

20030

Length centrum .

19.5

18.7

22

22

22

(22)

21.5

21

21.5

(20)

21

Width anterior articulation .

18.3

17.6

-

18

(14.5)

16

-

15

13

14

13.5

Height anterior articulation .

9

8.8

9

8.5

9

8.5

8.7

9

7.7

10

8

Width posterior articulation

14

12.8

15

-

13

15.5

12

14

10.5

-

11

Height posterior articulation .

Diameter of rib articulations

9.5

8.5

9.5

10.5

9.5

9

9.5

9

9

—

9

Head .

#

#

-

-

4.7

-

4.7

-

5.5

5

4.5

Tuberculum .

#

*

-

-

-

-

-

-

-

-

3.5

Height neural spine .

14

-

-

-

-

-

-

-

11

-

8.5

Length base neural spine .

12

12

-

-

12

-

13

-

13

-

13

Width across transverse process .

23

23.5

-

-

-

-

-

-

—

36

* Ribs do not occur on first vertebra; ( ) = measurements from crushed specimen.

out that these were somewhat more elongate than
the comparable vertebrae in Hesperornis.

The atlas (lacking the dorsal arch) and the
crushed anterior portion of the axis (Figure 3 a)
are present in AMNH 5101. The atlas is similar to
the atlas of Hesperornis illustrated by Marsh
(1880:196, fig. 40). In Hesperornis, loons, and
grebes, however, the ventral border of the atlas is
directed posteriorly to form a shelf. In Baptornis
the ventral border of the centrum is directed an¬
teriorly and does not form a shelf. The hypapo-
physis is short and broad so that it is not well
separated from the rest of the centrum. On either
side of the hypapophysis is a short posteriorly di¬
rected process. These are very small in Hesperor¬
nis and absent in the loons and grebes. Short,
blunt, transverse processes are also present. The
odontoid process of the axis is short and broad.
The axis has a large, deep, anteriorly directed pit
on its anterolateral side and a pair of small pits on
each side of the midline on the anteroventral sur¬
face. Its centrum is very narrow, with a thin blade¬
like hypapophysis. The anterior cervicals are all
narrow and elongate (Figure 3 c-e). They become
much shorter and more massive posteriorly. The
first eight vertebrae after the axis seem modified
for downward flexion (Figure 3) and the articular
facets of the prezygapophyses curve ventrally. The
anapophyses are short and somewhat clorsally di¬
rected. The sublateral crests form a triangle with
the anteroventral margin of the centrum on most
of these vertebrae. In Hesperornis, the sublateral
crests on the anterior cervicals are nearly parallel

and the centra are much broader ventrally. What
appears to be the tenth vertebra is very narrow
posteroventrally, with its posteroventral border di¬
rected downward. Articulating posteriorly with this
vertebra is a short, broad eleventh cervical with
the zygapophyses tilted laterally and not directed
downwards. The ventral surfaces of this and the
four following vertebrae are broad and short and
bear low parallel sublateral processes that do not
join each other ventrally as they do in similar ver¬
tebrae in giebes (Zusi and Storer, 1969, fig. 10).
Vertebrae 14-16 bear very large, deep, lateral de¬
pressions in their centra. The ventral border of the
centrum of the 16th vertebra is directed ventrally
and forms a distinct, thick, ventral process. In
Baptornis there is no evidence that large vertebrar-
terial canals were present in as many vertebrae as
they are in Hesperornis, loons, and grebes; how¬
ever, the vertebrae which should show these canals
are all badly damaged in the material of Baptornis.

The eighth vertebra anterior to the synsacrum
(Figure Aa-d) corresponds to the 16th cervical
vertebra of Hesperornis. Its dorsal spine is short,
thick, and broad at its base. The tubercle is not
bifurcated; the post- and prezygapophyses are set
at a low angle and are widely spaced; the neural
canal is circular in outline; the anterior articula¬
tion of the centrum is short dorsoventrally, but
very wide; the transverse processes are wide, swing
slightly upwards, and are directed posteriorly; the
posterior border is thickened while the anterior
portion nearest the centrum is slightly concave;
and there is no vertebrarterial canal. On this ver-

NUMBER 27

41

Figure 3. —Stereophotographs of vertebrae of Baptornis advenus: a, atlas and axis vertebrae
(AMNH 5101), X 4; b, cervical vertebra (AMNH 5101), anterior view, X 1; c, same, ventral
view, X 1; d, three partial cervical vertebrae (KUVP 2290), dorsal view, X 1; e, same, lateral
view, XI;/, thoracic vertebra (KUVP 2290), posterior view, X 1.

tebra the articulation for the tuberculum of the
rib is shallow, circular in outline, and slightly re¬
cessed from the outermost margin of the diapo¬
physes. The articulation for the head of the rib is
deep, circular, and placed anteriorly on the cen¬
trum. The centrum itself is short and thick, with a
strong indentation on either side. The hypapo-
physis is extremely short, stout, and directed pos¬
teriorly, practically merging in with the rest of the
centrum. The distal hypapophysis is terminated by
two stubby horizontal wings.

The seventh vertebra anterior to the sacrum cor¬
responds to the last cervical of Hesperornis and
resembles the previous vertebra. However, the an¬
terior articulation is not as wide and is higher
dorsoventrally; the neural canal is not as rounded;
the prezygapophyses are closer together and at a
more acute angle; the dorsal spine is wider and
thicker; there is a concavity on the anteroventral
margin of the transverse process above the articu¬
lation for the head of the rib; and the transverse
processes are shorter and do not swing back pos-

42

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Figure 4. —Stereophotographs of vertebrae of Baptornis advenus: a-d, dorsal, lateral, ventral, and
anterior views of eighth vertebra anterior to the synsacrum (UNSM 20030); e-h, dorsal, lateral,
ventral, and anterior views of first vertebra anterior to the synsacrum (USNM 20030). (All X 1.)

teriorly. The articulation for the tuberculum of
the rib is nearer the lateral edge of the diapophysis
than in the 16th vertebra. The hypapophysis is
thin and posteriorly directed.

The sixth vertebra anterior to the sacrum re¬
sembles the seventh except that the anterior articu¬
lation is not quite as wide and is higher dorso-
ventrally; the dorsal spine is of about the same
height but is thinner; the transverse processes
curve posteriorly; the pit above the articulation

for the rib is deeper, and the hypapophysis is thin
and directed anteriorly.

The fifth vertebra anterior to the sacrum has a
narrow, higher anterior articulation than the sixth;
the centrum itself is not as indented and the pre-
zygapophyses are closer together and at a more
acute angle; the dorsal spine is thin, high, and
directed anteriorly; the pit above the articulation
for the head of the rib is deep, as in the preceeding
vertebrae; the diapophysis and hypapophysis are

NUMBER 27

43

missing, but enough remains of the hypapophysis
to show that it was thin and large.

The fourth vertebra anterior to the sacrum simi¬
larly has a high, narrow anterior articulation; how¬
ever, the centrum is less deeply indented; a small
pit forms on the dorsal anterior margin of both of
the postzygapophyses; the pit above the articula¬
tion for the head of the rib has become shallow
and elongate. The dorsal spine, transverse proc¬
esses, right prezygapophysis, and most of the hypa¬
pophysis are broken off, but enough remains to
show that the hypapophysis was thin, posteriorly
directed, and probably short.

The third vertebra anterior to the sacrum has a
lower, but wider, anterior articulation; the centrum
is less indented; the pit above the articulation for
the head of the rib is absent; and the hypapophysis
is very short, thin, and triangular. The dorsal spine
and diapophysis are missing from the available
specimens.

The second vertebra anterior to the sacrum has
a wider anterior articulation than the preceding
vertebra. The top of the vertebra and the trans¬
verse processes are missing in the one available
specimen. The articulation for the head of the rib
is shallower than in the preceding vertebra, and
there is a shallow pit just above it. The hypapo¬
physis is represented by a small knob on the ven¬
tral border of the vertebra. The centrum of this
specimen had been broken through the middle
and repaired with glue. We separated the two
halves with acetone but found no medullary
cavity.

The first vertebra anterior to the synsacrum
(Figure 4 e-h) is distinctive. The anterior articula¬
tion is wider than in the preceding vertebra. The
dorsal spine is high and thin, and the posterior
margin is bordered by two small grooves lying on
top of the postzygapophyses. The diapophyses are
short, thick, and directed posteriorly, their ends
being flattened to buttress against the inner sides
of the ilia. There is a small pit lying directly an¬
terior to the diapophysis. The centrum is slightly
indented, the ventral border is flat, and there are
no rib articulations.

The vertebrae of Baptornis are nonpneumatic
and heterocoelous (amphicoelous vertebrae occur
in Archaeopteryx, Ichthyornis, and Enaliornis).
Compared to Hesperornis, the cervical vertebrae
are more elongate and not as deep, the anapo-

physes are less developed, and the sublateral proc¬
esses tend to converge more posteriorly. At least
the first five postaxial vertebrae are modified for
downward flexion, as are the presumed 14th and
15th vertebrae. The specimens of intervening vert¬
ebrae (7-13) are too fragmentary to be certain of
their adaptations but some must have been modi¬
fied for upward flexion.

The description of the thoracic vertebrae is
based almost entirely on the beautifully preserved
series with UNSM 20030. These all have good het¬
erocoelous articular surfaces, although circular
pits (Figure 3 f) occur in the centra of some speci¬
mens. They are not fused as in cormorants, or
fused and further immobilized by ossified dorsal
tendons, as in grebes. They also lack the very deep
lateral excavations found in the thoracic vertebrae
of Ichthyornis and can best be compared to the
thoracic vertebrae of Hesperornis. As Lucas (1903)
indicated, the hypapophyses are more anteriorly
situated, not as well developed as in Hesperornis,
and not bifurcated as in most modern diving birds.

Synsacrum.— The synsacrum (Figure 5 h-d) is
represented in KUVP 2290, FMNH 395, AMNH
5101, and a few fragments from KUVP 16112. It is
nonpneumatic and extremely narrow. Lucas
(1903) reported that the synsacrum of KUVP 2290
contains 10 fused vertebrae and that the first bore
a rib. Although he was correct in the number of
vertebrae (10 also being the number found in
FMNH 395) he was mistaken about the presence
of a rib with the first fused sacral. Rib facets do not
occur on any of the synsacra available nor even on
the last unfused thoracic vertebra.

The anterior sacral vertebra has a high, pos¬
teriorly sloping neural spine. The transverse proc¬
esses are tilted and flattened laterally to form a
broad contact with the ilium. The second sacral is
sutured dorsally to the first. It has a large, laterally
flattened, pointed transverse process projecting an¬
teriorly (Figure 56) that also abuts the ilium.
There are three vertebrae from the acetabulum
forward, and these may represent fused lumbars
with the rest of the sacrum being composed of
seven fused caudals (urosacrals). The neural
spines of the synsacral vertebrae are low and form
a median ridge bounded by projections on either
side. The posterior urosacrals lack the ventral keel
found in Gavia, and the posterior central articula¬
tions of the last two urosacral vertebrae are divided

44

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Figure 5.—Pelvis of Baptornis advenus: a, right innominate bone (UNSM 20030), lateral view.
Synsacra of B. advenus: b, dorsal view (KUVP 2290); c, lateral view (FMNH 395); d, ventral view
(AMNH 5101). All X 1.)

medially by a ventral groove.

Caudal Vertebrae.— UNSM 20030 has five cau¬
dal vertebrae and a pygostyle, as does FMNH 395.
AMNH 5101 also contains five caudals but in¬
cludes four anterior ones that are not represented
in either UNSM 20030 or FMNH 395. The first
caudal has already been discussed in the section on

the sacrum. The following three vertebrae have
low neural spines, flat ventral borders and widen
posteriorly, assuming a triangular shape. All of
these vertebrae are flanked by the pelvis and their
transverse processes abut the ilia. The vertebrae
posterior to the pelvis have high, straight, triangu¬
lar neural spines with rounded tips that become

NUMBER 27

45

Figure 6. —Caudal vertebrae and pygostyle of Baptornis advenus: a, dorsal view;

b, lateral view. (X 1.)

progressively lower posteriorly. The caudals also
have large, posteriorly directed and ventrally de¬
pressed transverse processes that, along with the
centra, become smaller posteriorly. The vertebral
centra are either amphicoelous or amphiplatyan
and the centra themselves are rounded ventrally
and bear ventrolateral depressions or pits. The
pygostyle is elongate and laterally flattened. It in¬
cludes five fused centra. The general configura¬
tion of the posterior two-thirds of the tail (Figure
6) is roughly similar to that in Gavia and quite
different from Hesperornis, which has dorsoven-
trally flattened caudals with wide, flat transverse
processes and shelf-like, fused intracentral bones.
Hesperornis has only two fused centra in its pygo¬
style. Intracentral bones are present in many living
diving birds, but they have not been found in
Baptornis. In Baptornis, the posteroventral margin
of the second fused centrum of the pygostyle bears
a distinct projection, which probably served as a
muscle insertion.

Ribs and Uncinate Processes. —Some fragments
of ribs are present with all of the specimens of
Baptornis. The best material (Figure 7) is pre¬

served in situ on slabs of matrix with UNSM 20030.
This material includes at least four different dorsal
ribs and four different sternal ribs. These are flat¬
tened and shaped about as in Hesperornis, sug¬
gesting a narrow rib cage. Six uncinate processes
representing five pairs are present in UNSM 20030.
The uncinate processes of Hesperornis and Bap¬
tornis do not fuse to the ribs—a condition similar
to that seen in grebes and loons, as well as certain
other modern birds. In Hesperornis, the uncinate

Figure 7. —Slab with ribs and uncinate processes of Baptornis
advenus (UNSM 20030).

46

processes are broad and straight. In Baptornis,
they are narrow, and bend upward at an acute
angle, thus resembling the uncinate processes of
grebes (Figure 8). Because there is no association
between ribs and uncinate processes on the slab,
none can be placed in order with certainty. The
ribs of Baptornis are heavier than those of grebes
and loons and do not expand as much ventrally.
The tuberculum is not separated as far from the
head as it is in Hesperornis or in the modern foot-
propelled divers.

Sternum.— The anterior and lateral margins of
the sternum (Figure 9 a) are preserved with UNSM
20030, the entire central portion and the posterior
margin having been destroyed.

In the area of the dorsal manubrial spine there
is only a smooth, thickened border from which a
shallow sulcus flares ventrally into a flattened,
rounded ventral manubrial spine. The coracoidal
sulci are deep, closely spaced, and lie at an angle
of 22° from a line perpendicular to the midline of
the sternum. The sternocoracoidal processes are
large, rounded, flare laterally, and are not strongly
curved. The costal margins are short, and bear
five costal ridges.

The sternum is most similar to that found in
Hesperornis. It appears to bear a shallow rectangu¬
lar depression on its anteroventral surface which
does not occur in Hesperornis. The sternum also

Figure 8. —Ribs and uncinate processes: a, grebe, Aechmoph-
orus occidentalis; b, Baptornis advenus; c, Hesperornis re-
galis. (Not to scale.)

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Table 2. — Measurements (mm) of the coracoid of Baptornis

Character

UNSM

20030

KUVP

2290

Approximate length .

_

53

4

3.5

7.6

8

Width scapular facet .

4.5

4.5

7

7.5

23

appears relatively smaller than in Hesperornis and
consequently the body of Baptornis may have been
somewhat narrower. The width of the anterior end
of the sternum is 53 mm and the length of the
costal margin is 27 mm.

Coracoid (Table 2).—UNSM 20030 includes the
scapular end and a fragment of the sternal end of
the left coracoid. KUVP 2290 includes most of the
right coracoid (Figure 9c). This specimen was illus¬
trated by Lucas (1903, fig. 6).

The head of the coracoid is small and turned
towards the procoracoid. The glenoid facet is
large, elliptical, and shallow. The furcular facet is
low, narrow, not undercut, and set almost directly
on the scapular end of the shaft so that it is com¬
pletely visible when the scapular end of the cora¬
coid is viewed from above (Figure 10a, 11a). The
surface of the scapular facet is rough and bears two
or more distinct pits on its anterior end. The pro¬
coracoid is short and recurved towards the shaft to
form part of the triosseal canal. Just below the
procoracoid is a foramen leading into the shaft of
the bone, which is probably the procoracoid fora¬
men. In KUVP 2290 this foramen penetrates from
the anterior to the posterior surface, as well as
branching into the shaft. Lucas (1903:553) incor¬
rectly states that the procoracoid process and fora¬
men are absent. The shaft is long and narrow. The
posterior surface is slightly concave, while the an¬
terior surface is curved and convex. This gives the
coracoid the appearance of a shallow spoon with a
square end. The sternal end is wide, thin, and
lacks definite facets.

The coracoid of Hesperornis is fundamentally

Figure 9. —Pectoral and wing elements of Baptornis advenus:
a (top to bottom), anterior, ventral, and left lateral views of
sternum (UNSM 20030) X 1; b, external and internal views
of left humerus, X 1; c, ventral, dorsal, and external views
of right coracoid (KUVP 2290), X 1; d, distal end of right
humerus, external view, X 5.

48

SMITHSONIAN CONTRIBUTIONS TO PAJLEOBIOLOGY

Figure 10. —Coracoids: a, tip of the scapular end of the left
coracoid of Baptornis advenus; b-g, ventral views of the right
coracoid: b, Baptornis advenus; c, Hesperornis regalis; d.
Archaeopteryx lithographica; e, Aechmophorus occidentalis;
f, Phalacrocorax auritus; g, Gavia immer. (Not to scale.)

similar to that of Archaeopteryx. In both genera
the coracoid is short and broad and the scapula
and humerus have articulations on the tip of the
scapular end. The coracoid of Baptornis resembles
Hesperornis in these features, as well as in having
the sternocoracoidal process above the midpoint of
the bone. The main longitudinal axis of the shaft
of the coracoid is perpendicular to the edge of the
sternal facet in the Hesperornithiformes and in¬
clined in all other foot-propelled diving birds for
which the coracoid is known (Figure 10). The
coracoid of Baptornis is more elongate than in
Hesperornis and the procoracoid process is smaller
and differently shaped. In Baptornis, the internal
edge of the coracoid is nearly straight, while in
Hesperornis there is a large, square, internal pro¬
jection (Figure 10c). It should be noted that Lucas
(1903, fig. 6) has illustrated the posterior view of

the sternal fragment of the coracoid with the an¬
terior view of the scapular portion (compare his
figure 6 with figure 10 of this paper).

A small area for the attachment of the clavicle
seems to be present although no clavicles are
known.

Scapula.— KUVP 2290 includes the articular end
of the left scapula (Figure Wb-d), which was de¬
scribed and figured by Lucas (1903:553-554). The
coracoidal articulation is long, narrow, and slightly
curved to fit the contour of the scapular facet on
the coracoid. It bears two large pits similar to those
found on that facet. The anterodorsal margin of
the proximal end does not show any articulation
for the furcula. This margin slopes at about 47
degrees to the main axis of the shaft, and is ter¬
minated dorsally by a small projection. The ven¬
tral border bears a long, shallow groove. The
measurements (in mm) of the scapula are: width
of neck, 7.5; depth of neck, 4.0; width of proximal
end, 10.5; length of glenoid facet, 9.4.

The neck of the scapula is wide and thick, which
led Lucas to suggest that it may have been ex¬
panded posteriorly as in penguins. The scapula of
Hesperornis is similarly thickened, but does not
expand posteriorly (Marsh, 1880:58). Therefore,
the posterior portion of the scapula of Baptornis
probably did not differ much from that of
Hesperornis.

Humerus. —KLTVP 2290 includes the distal end
and a portion of the shaft of the tight humerus.
This specimen was described by Lucas (1903:554),
who reported it as being a left humerus, but the
curvature of the shaft matches that of Marsh’s
illustration of the right humerus of Hesperornis
regalis (Marsh, 1880, pi. 8: fig. 1). UNSM 20030
includes a slightly abraded proximal end and the
greater portion of the shaft of the left humerus.
We have examined the large alleged humeri of
Baptornis reported by Walker (1967) and have
determined that they are fragmentary shafts of the
tibiotarsus.

The proximal end of the humerus is simplified
as compared with that of modern birds (Figure
9b). The shaft curves downward, then expands
noticeably and is twisted posteromedially 75 mm
from its proximal end. There is a nutrient fora¬
men situated in a groove 48.5 mm from the proxi¬
mal end in UNSM 20030. This foramen is like¬
wise present in KUVP 2290, which also has a

NUMBER 27

49

Figure 11. —Stereophotographs of pectoral girdle
of left coracoid (USNM 20030), internal view;
articular end of left scapula (KUVP 2290). (All

second groove and a foramen 17.5 mm distal to the
previously mentioned one.

A small, rounded prominence set off from the
internal condyle by a shallow groove may represent
the distal external condyle. The internal condyle is

elements of Baptornis advenus: u, scapular end
b-d, ventral, internal, and external views of
X 4).

not of the form usually found in birds and is only
slightly delineated from the rest of the distal end.
It bears a small foramen on its articular surface.
There is no distinct olecranal fossa or prominent
grooving. The shaft of the University of Nebraska

50

specimen was x-rayed and the broken ends of
KUVP 2290 and UNSM 20030 were examined.
Neither are pneumatic. By transposing the hu¬
merus of UNSM 20030 over that of KUVP 2290
so that the curvatures and foramina are in align¬
ment (Figure 9b), a reconstructed length of 118
mm is obtained. This is only 18 mm more than
Lucas’s (1903:553) estimate of 100 mm based on
KUVP 2290 alone. The width of the proximal end
is 8.9 mm and of the distal end 6.8 mm. The
greatest diameter of the shaft is 6.5 mm.

The humerus of Baptornis is relatively long,
curved, and rounded. It is not flattened and
straight as in the wing-propelled auks and pen¬
guins. Lucas (1903:554) suggests that the wing may
have been used in conjunction with the feet for
locomotion. This seems unlikely, although we
have shown it as having a stabilizing function
(Figure 20). The humerus is larger than that of
Hesperornis, which is a much bigger bird. In
Hesperornis, the distal end is also much more re¬
duced and the radius and ulna may not have been
present.

Radius. —A description of the left radius (Fig¬
ure 126) KUVP 2290, appears in Lucas (1903:
554). The humeral cotyla is large and oval, with
the bicipital tuberosity situated along its rim. The
shaft expands markedly at about the midpoint. In
this area there is a nutrient foramen on the palmar
side and also a faint intermuscular line. The lunar
depression is shallow. The scapholunar facet is
long and narrow and the distal ligamental process
is relatively large. The radius is 20.5 mm long, the
proximal end is 3.0 mm wide, and the distal end
is 3.5 mm wide.

Ulna. —Lucas (1903:554) briefly describes the
short, robust, left ulna, KUVP 2290 (Figure 12a).
The olecranon process is short, massive, and not
noticeably twisted as in many flying birds. The in¬
ternal and external cotylae on the proximal end
are only slightly separated by an intercotyla area
and almost form a single articular surface. The
proximal radial depression is slightly discernable.
The impression of the brachialis anticus is large
and oblong, and crosses the palmar surface of the
bone diagonally. A faint intermuscular line
stretches for 2-3 mm below the distal end of the
impression for the brachialis anticus. No nutrient
foramen can be discerned. The distal radial depres¬
sion is small and quite shallow. The internal con-

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

dyle on the distal end is a small projection. The
external condyle is not a distinct ridg© but bears a
large, flat articular surface which tilts toward the
anconal and ventral margins. Although we ex¬
amined the bone carefully, we could see no scars
for feather attachment, Lucas’ (1903:554) asser¬
tion of their presence notwithstanding. Measure¬
ments (in mm) of the ulna are: length, 21.6; width
proximal end, 4.2; width distal end, 2.9; depth
distal end, 3.5.

The radius and ulna of Baptornis are short,
stout bones resembling in general form their coun¬
terparts in the extinct diving goose Chendytes
milleri (Howard, 1955). The reduction is much
greater in Baptornis, however. In C. milleri the
ulna is 43 percent as long as the humerus (How¬
ard, 1955:142), while in Baptornis it is only about
19 percent of the restored length (118 mm) of the
humerus. The general form of the ulna is similar
to that found in theropod dinosaurs (Figure 13)
and it seems possible that the Hesperornithi-
formes may have split off from the line leading to
modern flying birds before the wing had developed
the adaptations seen in modern birds. The pres¬
ence of well-formed articulations on the distal ends
of the radius and ulna show that some sort of car-
pals were present, although these may not have
been fused into a true carpometacarpus.

Pelvis. —All the specimens of Baptornis we
studied included at least some fragments of the
pelvis. Most of the right side is preserved with
UNSM 20030, and our description of the pelvis is
based on this specimen (Figure 5a). The synsac-
rum is in place only in AMNH 5101, in which the
acetabulum is opposite the third fused sacral.

The pelvis is similar to that of Hesperornis, but
differs in having a much longer preacetabular por¬
tion of the ilium. The postacetabular part of the
ilium is quite long, as it is in all of the foot-
propellecl diving birds. The middorsal and the
posteriormost portions of the ilium are not known,
but were probably similar in shape to the same
region in Hesperornis, although the whole pelvis
is somewhat narrower proportionally. The pec-
tinal process is broad and blunt, as it is in Hesper¬
ornis, and the acetabulum is partially closed. The
acetabula of the loons and grebes have vertical
sides and are completely open. The antitrochanter
of Baptornis is large and rectangular, resembling
that of Hesperornis. As in Hesperornis, the ante-

NUMBER 27

51

Figure 12.—Stereophotographs of wing elements of Baptornis advenus: a, medial, palmar, and
anconal views of left ulna (KUVP 2290), X 4; b, medial, palmar, and anconal views of left
radius (KUVP 2290), X 4.

rior end of the ischium sweeps up to form the
posteroventral border of the antitrochanter, and
there is a prominent suture here in both Baptornis
and Hesperornis. On the pelvis of the Common
Loon, Gavia immer, just anterior to the antitro¬

chanter, there is a small scar for the gluteus medius
and minimus muscles. This scar is absent in the
pelvis of giebes, Baptornis, and Hesperornis. The
ilium and ischium of Baptornis are separate
throughout their length, as they are in Hesperor-

52

a b c d

Figure 13.—Left ulnae: a, c, medial and palmar views of the
theropod dinosaur Deinonychas antirrhopus (from Ostrom,
1969, fig. 58); b, d, same views of Baptornis advenus. (Not
to scale.)

nis. The ischium is long, thin, flattened internally,
and rounded externally, so that in cross-section it
appears bell-shaped. The pubis is long, heavy, and
flattened dorsoventrally. It bears a shallow groove
along the anterior quarter of its dorsal surface.

The measurements (in mm) of the pelvis are:
total length (estimated), 179; length of preace-
tabular ilium, 54; depth of ilium anterior to pec-
tinal process, 19; depth of ilium at pectinal process
27; greatest diameter of acetabulum, 14; height of
antitrochanter, 13; width of antitrochanter, 11;
length of free ischium, 103; length of free pubis,
109.

Loons and grebes resemble each other and differ
from Baptornis and Hesperornis in having the
preacetabular portion of the pelvis narrower,
twisted laterally, and spread apart anteriorly.
Baptornis also differs from the modern foot-
propelled diving birds in having the ventral mar¬
gin of the ilium turned medially (Figure 5). The
origins of the iliotrochantericus medius and anti-
cus muscles are ventral in Baptornis , but are dorsal
in loons and grebes. The dorsal surface of the pre¬
acetabular ilium is turned medially in loons and
grebes and laterally in Baptornis. The postace-
tabular portion of the pelvis of Baptornis is, on a

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

whole, narrower than in other diving birds except
Hesperornis.

Femur (Table 3).— Both femora are represented
in FMNH 395 and KUVP 2290. UNSM 20030 has
the right femur present (Figure 145) and KUVP
16112 includes fragments of both femora.

The femur of Baptornis is proportionately more
elongate than that of Hesperornis; the neck is
more constricted; the insertion of the round liga¬
ment is smaller; and the trochanter rises slightly
above the head, whereas both are of about the
same height in Hesperornis. When the femur is
articulated with the pelvis its position is more in¬
clined than in Hesperornis, but the difference is
probably not as great as Lucas (1903:554) sug¬
gested. The iliac facet occupies about the same
shape and area as it does in Hesperornis, and the
obturator ridge is very similar in form. Baptornis
resembles loons in having the lateral margin of the
trochanter close to and parallel with the axis of
the shaft, whereas the lateral margin of the tro¬
chanter extends a considerable distance away from
the axis of the shaft in Hesperornis. The anterior
intermuscular line sweeps down in a low arc from
the trochanteric ridge to the external condyle. The
posterior intermuscular line is not well defined,
but runs down the medial surface of the bone from
just below the head to the internal condyle. The
trochanteric ridge is proportionately larger and
heavier than it is in Hesperornis. There are no
large nutrient foramina evident on the shaft. The
popliteal area is broad and shallow. The fibular
condyle sends off a distinct wing in Baptornis and
Hesperornis and there is a thick connection be¬
tween the internal and external condyles (this
connection is thick in cormorants and grebes and
thin in loons). The fibular groove is broad and
shallow in Baptornis and Hesperornis. The inter¬
nal condyles are at about the same level in Hes-

Table 3. —Measurements (mm) of the femur of Baptornis

Character

UNSM

20030

FMNH

395

KUVP

2290

Length .

71

72

75

Diameter head .

10

10

11.5

Diameter distal articulation .

25

24

26

Diameter proximal end .

24.5

24.5

28

Antero posterior diameter midshaft

11.5

12

12.5

Transverse diameter midshaft

10

10.5

11

NUMBER 27

53

Figure 14.—Hindlimb dements of Baptornis advenus: a , distal end of left tibiotarsus (UNSM
20030); b (top to bottom and left to light), proximal, distal, anterior, lateral, posterior, and
medial views of light femur (I’NSM 20030); c (top to bottom and left to right), proximal, pos¬
terior, and anterior views of right fibula (UNSM 20030); d (top to bottom), anterior, posterior,
and distal views of right patella (UNSM 20030). (All X 1.)

54

perornis while the external condyle is much more
distal than the internal condyle in Baptomis.

A discussion of the muscle scars on the femur
would be valuable but the homologies are difficult
to ascertain. Those of the following scars seem
fairly certain. Along the posterior intermuscular
line, just below the head, is a small raised triangu¬
lar area which may correspond to the insertion of
the M. iliacus. On the lateral surface of the femur,
along the trochanteric ridge, is a complex region
of muscle attachments resembling the same area in
Hesperornis. The tubercle for M. piriformis is not
as prominent as it is in Hesperornis and is situated
about half way up the shaft.

The proportions of the femur of Baptomis sug¬
gest that the legs were not bound so closely to the
pelvis as in Hesperornis and may have had slightly
greater freedom of movement. In Hesperornis, the
femora may have been permanently held in the ex¬
tended position illustrated by Heilmann (1927,
fig. 34), while Baptomis may have been able to
rotate the legs ventrally for paddling and then ab¬
duct them for diving as do some diving ducks
(Raikow, 1970:6).

Patella (Table 4).—The patella of KUVP 2290
was described and illustrated by Lucas (1903).
There is also an excellently preserved patella with
UNSM 20030 (Figure I4d). This is a short tri¬
hedral bone resembling in some respects the
patella of a cormorant. The foramen for the ten¬
don of the ambiens muscle is large and perforate.
The articular surface on the base of the bone is
double, indicating that it probably articulated
with both the internal and external condyles of
the femur. In Hesperornis there is a single con¬
cave surface, which articulated with the external
condyle of the femur. Therefore, the patella of
Hesperornis would have been lateral to the main
axis of the tibiotarsus, while that of Baptomis
would almost have been centered on it (contrary
to Lucas, 1903:554). Cormorants have the patella

Table 4.—Measurements (mm) of the patella of Baptomis

Character

UNSM KUVP

20030 2290

Length .

Distal antero-posterior diameter .

Distal transverse diameter

19 20.5

10.5 13.5

16 16.5

6 7

Diameter ambiens foramen .

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Table 5.—Measurements (mm) of the tibiotarsus
of Baptomis

Character

UNSM FMNH KUVP
20030 395 2290

Length .

195 194

14 13 14

18 17 18.5

9 8.5

12 11.5

19 18

Elevation cnemial process .

Diameter proximal articulation .

Antero-posterior diameter shaft* .

Transverse diameter shaft* .

Diameter distal end .

* Below fibular ridge.

placed as in Baptomis. The patella in grebes is
shaped as in Hesperornis and articulates on the
external condyle of the femur. However, grebes
lack the ambiens muscle and therefore no fora¬
men is present in the patella.

Tibiotarsus (Table 5).—Both right and left
tibiotarsi occur in UNSM 20030 and in FMNH
395 and parts of both are present in KUVP 2290
and 16112.

The tibiotarsus (Figure 15) is like that of Hes¬
perornis in being elongate and nonpneumatic
(like Hesperornis, it has a large medullary cavity).
The proximal end flares out as in Hesperornis
due to the lateral expansion of the outer cnemial
crest. This region is not as expanded in loons and
grebes. As in Hesperornis, the inner cnemial crest
is low, so that the groove between the two crests is
broad and shallow. In loons and grebes, the inner
cnemial crest is high and the surface between the
two crests is narrow and deeply excavated. The
rotular process is lower than in loons or grebes and
is similar to Hesperornis. The external articular
surface is small and slopes ventrally. It is not as
rounded as in loons or grebes, nor is it set apart
anteriorly and posteriorly from the inner articular
facet by grooves (Figure 16a) as in Hesperornis.
The inner and outer articular facets are about
equal in size and are separated by a groove in the
interarticular area in grebes. The inner articular
facet is flat, oval, and directed posteromedially in
Baptomis, and just below the inner articular facet
is a deep roughened pit for the origin of M. plan-
taris which appears to be divided into dorsal and
ventral parts. The fibular crest extends about half
way down the shaft and is deeply grooved along its

Figure 15.—Left tibiotarsus of Baptomis advenus: a, pos¬
terior, b, lateral, c, anterior, and d, medial views, (X 1).

56

posterior margin, said groove crossing the outer
border of the shaft just below the fibular crest. The
foramen for the medullary artery lies in this
groove. The distal attachment for the spine of the
fibula is much smaller than in Hesperornis.

The distal end of the tibiotarsus (Figures 14a, 15,
165) is slightly inflected medially, but not as much
as in Gavia. In Hesperornis regalis and the grebes,
the distal end is almost centered on the shaft. The
internal and external condyles of the tibiotarsus
are spread farther apart than in Hesperornis and
the posterior crests are not as high. The tendinal
groove is broad and terminates in a deep lateral
pocket. The internal and external condyles are
roughly parallel to each other and the anterior
intercondylar sulcus is shallow (more so than in
Hesperornis). Both the external and internal lig-
amental prominences are very low. There is no
supratendinal bridge, although a large ligamental
attachment occurs above the medial side of the ex¬
ternal condyle. In FMNH 395 the high ascending
process of the astragulus is still clearly discernable
and the fusion of this tarsal element to the tibia
evidently took place later in otogeny than in mod¬
ern birds.

Fibula (Table 6). —The fibula of Baptornis
(Figure 14c) is most similar to that of Hesperornis.
The head is large and rectangular. The shaft has
two distinct ridges on its posterior proximal sur¬
face and the bicipital tubercle is large and elon¬
gate. The head is as in Hesperornis and is not
undercut posteriorly as much as in loons. There is
no tubercle for M. flexor perforatus digiti III as
there is in loons, but instead there is a large tri¬
angular roughened area as in Hesperornis and
grebes.

Tarsometatarsus and Toes (Tables 7,8).—As
Shufeldt (1915:9) noted, the holotype tarsometa¬
tarsus designated by Marsh consists of two portions
that are quite probably from different individuals,

Table 6. —Measurements (mm) of the fibula of Baptornis

Character

IINSM

20030

FMNH

395

KUVP

2290

Antero-posterior diameter

proximal end .

7

6

8.5

Transverse diameter proximal end

11

9

11

Greatest transverse shaft diameter

7.5

7

-

Antero-posterior diameter

at this point .

5

4

-

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY
Table 7.— Measurements (mm) of the tarsometatarsus

of Baptornis

Character

UNSM FMNH KUVP
20030 395 2290

I.pngth .

84 83 83 E

(10) 13.8 13.5

(17) 17.6 18

15 15.1 16

11.5 11.5 12

Proximal antero-posterior diameter

Proximal width .

Distal antero-posterior diameter

Tip trochlea II to distal end

E = restored length; ( ) = measurements from crushed
specimens.

as indicated by the facts that the fracture lines of
the two halves do not coincide and the proximal
portion is from a juvenile, while the distal portion
appears to be from an adult. If the two pieces had
been from one individual, Shufeldt estimated that
as much as a third of the shaft must be missing. At
our request, the curators of the Division of Verte¬
brate Paleontology of the Yale Peabody Museum
have agreed to retain YPM 1465 for the distal por¬
tion of this specimen, which we here designate as
the lectotype. The proximal portion has been re¬
numbered as YPM 5768.

KUVP 2290 represents a well-ossified individual
in which the proximal and distal ends of a left
tarsometatarsus are uncrushed, but in which the
middle of the shaft is missing. YPM 5768 is from a
young bird, and along with FMNH 395 (Figure
16/,g) and KUVP 16112 shows the lines of fusion
between the metatarsal bones. The tarsometatarsus
of UNSM 20030 is mature, but crushed.

The tarsometatarsus of Baptornis is compressed
laterally. The external cotyla is slightly larger than
the internal one (Figure \6e) and both tilt slightly
anteriorly; the intercotylar prominence is low and
blunt as compared to Hesperornis; calcaneal
ridges are absent, and there are no proximal
foramina; the tubercle for tibialis anticus is situ¬
ated high on the shaft; the anteroproximal face of
the shaft is deeply excavated; the attachment for
the external ligament is not clearly developed; the
outer extensor groove is long and shallow and leads
to the region of the distal foramen, with the ante¬
rior metatarsal groove running parallel to it; the
inner metatarsal groove crosses the medial side
about one-third the way down the shaft; the inter-
trochlear notch between the third and fourth
trochleae narrows and then widens so that the dis-

NUMBER 27

57

Figure 16.—Stereophotographs of hindlimb elements of Baptornis advenus: a, proximal end of
right tibiotarsus (UNSM 20030); b, distal end of right tibiotarsus (UNSM 20030); c, distal end
of left tarsometatarsus (KUVP 2290); d, left tarsometatarsus (UNSM 20030), posterior view;
e, proximal end of left tarsometatarsus (KUVP 2290) (not stereo); f, left tarsometatarsus (FMNH
395), anterior view; g, same, posterior view. (All X 1, except c and e, which are slightly enlarged.)

58

tal foramen communicates with the notch and the
metatarsal facet is high and posterior. In UNSM
20030 (Figure 16d) the inner trochlea is set about
7 mm proximal to the middle trochlea. The outer
trochlea is in line with the major longitudinal axis
of the bone and is approximately equal to the mid¬
dle trochlea in size. In mature specimens the me¬
dial rim of the inner trochlea develops a promi¬
nent posterior flange.

There is some ontogenetic variation in the
trochleae. Those of YPM 1465 and FMNH 395
have shallow tendinal grooves, are less expanded,
and lack the small flange on the inner trochlea
that is present in UNSM 20030 and KUVP 2290.
The lateral rim of this trochlea extends farther
distally than the medial rim.

The outer trochlea extends only slightly farther
distally than the middle trochlea and both are
about equal in size. In both, the tendinal groove is
shallow and continuous along both the plantar and
the acrotarsal faces of the middle trochlea, but it is
missing from the acrotarsal face of the outer troch¬
lea. The medial rim of the outer trochlea is slightly
larger than the lateral rim. The lateral rim extends
into a posterior flange in mature specimens. The
trochleae are compressed laterally and the extensor
grooves run up onto the anterior face as in loons
and Hesperornis. In grebes the trochleae are less
compressed and tend to be almost smooth on their
anterior face. In both grebes and Hesperornis, the
inner trochlea has rotated posteriorly, while in
loons and Baptornis it is more nearly parallel to
the main axis of the shaft.

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

The tarsometatarsus of Baptornis is like that of
Hesperornis in that it is laterally compressed, lacks
calcaneal ridges on the hypotarsus, lacks proximal
foramina, and the outer trochlea is in line with
the longitudinal axis of the bone. Baptornis differs
from Hesperornis in that the necks of the trochleae
are longer and more delicate.

In Hesperornis the outer extensor groove is very
large and deep, and the anterior metatarsal groove
is quite prominent, although it is only about one-
third the width of the outer extensor groove. There
is a distinct groove originating near the distal fora¬
men and extending from the medial rim of the
outer trochlea over the trochlea to its lateral rim.
This groove is much fainter in Baptornis. In both
Baptornis and Hesperornis the distal foramen is
well within the intertrochlear notch, between the
middle and outer trochleae. Its position in Baptor¬
nis is delineated by indentations in the necks of
the trochleae, but the distal margin is not closed
off to form a real foramen. In Hesperornis exten¬
sions of the sides of the trochleae meet to form the
distal margin of the foramen, and a suture line is
usually visible where they meet.

FMNH 395 includes metatarsal I. It is not as
short and flat as in Hesperornis. The bone di¬
verges from the shaft of the tarsometatarsus at a
fairly steep angle and terminates in a rounded
knob. The phalanx of the first digit is a thin, elon¬
gate, highly curved bone. None of the toe bones
were articulated, so their identification is somewhat
subjective. Fourteen phalanges, counting 2 un-
guals, are preserved with FMNH 395, 6 with

Table 8. —Measurements (mm) of the pedal phalanges (arabic numbers in boxhead) of Baptornis

Metatarsal Digit

Digit

Dicit

III

Digit IV

Character

I

FMNH

395

I

FMNH

395

II

FMNH

395

UNSM

20030

FMNH

395

UNSM 20030

FMNH

395

1

2

1

2

3

1

2

3

1

2

3

4

Length

Proximal vertical

14

22

31.7

37.0

28.2

20.5

37

25.5

24

37

25

23

23

diameter

Proximal transverse

3.8

4.4

11.5

8.8

4.3

(6.5)

(6.5)

7.3

8

8.5

6.5

5.2

diameter

3.2

2.8

8.3

5.8

6.5

3.5

(6)

(6)

5

6.5

4.5

4.5

7.2

Distal vertical diameter

-

3.5

2.6

8.5

7

3.2

5.5

4

4.5

5.5

4

4

2.3

Distal transverse

diameter

-

4

4.7

5

6

4.8

7.5

6

6

7

5

6

4.1

( ) = Measurements from crushed specimens.

NUMBER 27

59

Figure 17.—Left foot of Baptornis advenus: a, anterior view;
b, lateral view.

UNSM 20030, and 3 fragments with KUVP 16112.
Phalanx 1 of digit II appeared to be absent from
our sample. Phalanx 2 of digit II is flattened dorso-
ventrally, with a broad proximal end, which
rapidly narrows anteriorly. It has a simple distal
articulation for the claw. Phalanx 1 of digit III is
the largest of the foot. It is flattened laterally, with
a deep proximal articulation. The distal articula¬
tion consists of two parallel ridges, which would

abed

Figure 18. —Distal end of left phalanx 1 of digit IV showing
increasing modification for toe rotation from left to right:
a, loon, Gavia immer; b, Baptornis advenus; c, grebe, Aech-
mophorus occidentalis; d, Hesperornis regalis. (Not to scale.)

permit mainly anterior or posterior movements.
Phalanx 2 of digit III is a smaller version of the
first. Phalanx 3 of digit IV is about the same length
as that of digit III but neither articulates as in
loons or grebes. That of digit IV is narrower and
rounder than in the other toes, and the distal end
has the medial ridge of the articulation inclined,
suggesting that some degree of toe-rotation was
possible. Phalanges 2 and 3 of digit IV are much
shorter and deeper than the first phalanx, while
the fourth is elongate, laterally compressed, and
terminates in an articulation for the claw.

Although phalanx 1 of digit IV is more elongate
than in Hesperornis, nevertheless the third and
fourth toes of Baptornis are of about the same size
and length (Figure 17), whereas in Hesperornis
the fourth toe is much the longest. The phalanges
of the fourth toe also lack the deep ventral groov¬
ing found in Hesperornis. The claws of Baptornis
are somewhat flattened, gently curved and pointed.
They are not the broad nail-like structures that
grebes have.

Grebes have lobed feet while loons have webbed
feet. When swimming, both loons and grebes have
their toes spread for the power stroke. On the re¬
covery stroke, loons fold their toes posteriorly
while grebes rotate their foot 90° (Storer, 1958) so
that the edges of the lobes cut through the water.
These actions are reflected in the morphology of
the phalanges of digit IV. In loons the distal artic¬
ular ridges are parallel to each other and are about
the same size. This is the condition found in most
birds. In grebes the medial ridge is enlarged and
the lateral ridge is small and rounded (Figure 18c).
In Hesperornis the distal articulations of the
phalanges for digit IV are even more specialized,
with the medial ridge extending over the small,
rounded lateral ridge (Figure 18d) suggesting
highly developed toe-rotation and lobed feet.
Baptornis differs from the loons in having the me¬
dial ridges enlarged and inclined on the distal ar¬
ticulations of the phalanges for digit IV. However,
it is not nearly as specialized as either grebes or
hesperornithids (Figure 185).

Habits of Baptornis

From the skeletal remains now available, it ap¬
pears that adults of Baptornis were about one
meter long from the tip of the tail to the tip of the

60

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Figure 19.—Skeleton of Baptornis advenus (stippled areas restored after Hesperornis).

62

SMITHSONIAN CONTRIBUTIONS TO PATEOBIOLOGY

bill. We have restored the bird with teeth (Figures more solid than most of the other small bones in
19, 20) because of its relationship to the toothed the Niobrara Chalk. This is hardly surprising, for

bird Hesperornis. All Mesozoic birds for which additional weight is actually of an advantage for a

dentaries are known ( Archaeopteryx, Hesperornis, diving animal. The added weight, along with the

and Ichthyornis) had teeth. long necks and fusiform bodies of the Cretaceous

The skeletons of both Hesperornis and Baptornis divers, suggest that they might have swum partially

are nonpneumatic and are composed of relatively submerged, like modern anhingas. However, nei-

heavy, compact bone. Bones of Hesperornis are ther Baptornis nor Hesperornis appears to have

often recognized in the field because they appear had the neck modified for stabbing. This is further

Figure 21—Pelvis and right hindlimb of Baptornis advenus, dorsal view.

NUMBER 27

63

emphasized by their relatively broad, blunt bills.
In both Baptornis and Hesperornis the wing is
greatly reduced. This is often the case in marine
foot-propelled divers; various adaptations for
flight would make them too bouyant for efficient
diving, while the loss of flight is not too serious a
hardship for them. On the other hand, foot-
propelled divers that occupy disjunct bodies of
fresh water need to be able to fly in order to dis¬
perse. Lonchodytes is the only known nonmarine
foot-propelled diver from the Mesozoic. It is from
freshwater river deposits, and appears to have been
volant (Brodkorb, 1963a). The tiny wing of
Baptornis could scarcely have been used for loco¬
motion, but it may have had a stabilizing and steer¬
ing function, much like the pectoral fins in fishes.
The sternum in Baptornis is unusually small, even
for a flightless bird, and the body must have been
very long and narrow. The tibiotarsus was bound
to the pelvis as it is in loons, but it was probably
not as closely appressed to the postacetabular re¬
gion as it was in Hesperornis, and the bone itself
extended somewhat past the posterior margin of
the pelvis (Figures 19, 21). In Hesperornis the distal
end of the tibiotarsus is about even with the pos¬
terior margin of the pelvis. The above features,
along with characters of the acetabulum and femur,
suggest that Baptornis had some capacity to rotate
its feet under its body, probably for swimming on
the surface. When diving, the feet were certainly
held out to the side. It seems almost certain that
Baptornis could not walk upright on land and in
fact must have pushed itself along on its stomach
like a seal or a loon. The feet could have been
webbed or lobed. They do show some evidence of
inefficient toe-rotation and might well have been
lobed, although our restoration shows them
webbed. Hesperornis almost certainly had lobed
feet. The feet of Baptornis are exceptionally large
for the size of the bird. The tail is quite long, with
a long, laterally compressed pygostyle that prob¬
ably acted as a rudder. Loons appear to have simi¬
lar, but shorter tails, while in Hesperornis the tail
is long but is dorsoventrally compressed and has a
short pygostyle.

Included with UNSM 20030 are eight coprolites,
two of which show small fish jaw and other bones.
Most are round or elliptical in cross-section and
are elongate, except for the two containing the fish
material. None shows spiral grooving or surface

impressions. George Sternberg, the collector, in a
1937 communication preserved in the records of
the University of Nebraska State Museum, makes
the following reference to the association of these
coprolites with the Baptornis skeleton: “There are
7 or 8 coprolites; .... 2 show small fish bones.
These are small compared to other coprolites I
have seen and were found mingled with the
bones.” It seems likely that these coprolites are
correctly associated with the Baptornis skeleton; if
so, they are the only ones known for a Cretaceous
bird. However, several of them fit together to form
a long rounded structure that might be better in¬
terpreted as an intestinal cast (Figure 22a). A fish
jaw in one coprolite was identified by Orville W.
Bonner of the University of Kansas (pers. comm.,
1972) as Enchodus cf. parvus Stewart. Enchodus was
a very common genus of fish in the Niobrara seas. It
included some very small species, which probably
formed a significant item in the diet of
the Late Cretaceous toothed birds. Hesperornis had
scutellate-reticulate tarsi (Williston, 1898). We
have a fragment of matrix associated with UNSM
20030 showing what may possibly be the impres¬
sion of scutes (Figure 22b), suggesting that the
tarsometatarsus of Baptornis might also have been
scutellate-reticulate (grebes have reticulate tarsi
and loons have scutellate-reticulate tarsi).

Hesperornis and Baptornis form a natural
group of specialized divers, separate from, and
more specialized than any of the other foot-
propelled divers. Their advanced degree of spe¬
cialization is not surprising, as the Mesozoic birds
occupied a fairly stable environment for the entire
Cretaceous and must have been well adapted to it.

Figure 22.—Coprolites (a) and possible scute impressions from
the foot ( b ) of Baptornis advenus (UNSM 20030). (Reduced.)

64

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Baptornis was probably able to range far from
shore, as most of the finds indicate. At the present,
all of the known records of Baptornis are from the
Smoky Hill Member of the Niobrara Chalk For¬
mation, Upper Cretaceous of Central Kansas. The
immature specimens suggest that Baptornis may
have nested in this region.

Affinities of Baptornis

Baptornis and Hesperornis share a number of
primitive features with Archaeopteryx. These in¬
clude: (1) the nature of shoulder girdle, and in
particular the shape of the coracoid, with the
glenoid and scapular facets near the tip of the
scapular end; (2) the character of the distal articu¬
lation of the ulna in Baptornis (the ulna of Hes¬
perornis is unknown), which makes the presence of
a fused carptometacarpus doubtful; (3) the sepa¬
ration of the distal segments of the ilium, ischium,
and pubis; (4) the absence of a supratendinal
bridge on the tibiotarsus; (5) the prominence and
fusion late in ontogeny of the ascending process of
the astragalus on the tibiotarsus; (6) the absence
of hypotarsal canals, a closed distal foramen, or
proximal foramina in the tarsometatarsus (proxi¬
mal foramina are present in all modern birds we
have examined). To the above we can probably
add the various primitive features of the skull, in¬
cluding the shape of the quadrate and the presence
of teeth.

The absence of a supratendinal bridge, hypo-
tarsus, and proximal foramina indicates that these
are not necessary for foot-propelled diving. Fur¬
ther, the presence of these characters in modern
foot-propelled divers suggests that these birds may
have been derived from ancestors that were not
foot-propelled divers.

The supratendinal bridge is well developed in
loons and cormorants, and this may be related to
the fact that the distal end of the tibiotarsus in
these forms is inflected medially. The supratendi¬
nal bridge is not well developed in some grebes
which have (as do hesperornithiform birds) a
straight articulation between the tarsometatarsus
and tibia, causing the foot to be brought more di¬
rectly out to the side of the body.

Baptornis and Hesperornis also share a large
suite of derived characters. Many of these are di¬

rectly related to foot-propelled diving and can be
found in one state or another in all birds that
share this form of locomotion. Some of the char¬
acters that tend to unite Baptornis and Hesperor¬
nis , in addition to the shape and reduction of the
shoulder girdle and the wing, are: (1) the broad
shallow fibular groove on the femur; (2) the large
trihedral patella with a foramen for the tendon of
the ambiens muscle; (3) the broad, triangular
procnemial crest on the tibiotarsus; and (4) the
compressed tarsometatarsus with the anterolateral
edge developed as a high thin ridge. No particular
relationship is indicated between Baptornis and
any living group of foot-propelled divers. The foot
is more similar to that of loons than of grebes, but
the coracoid is totally unlike that of any living
bird. Throughout, the skeleton is most similar to
that of the Hesperornithidae.

Marsh did not attempt to classify Baptornis, but
clearly regarded it as a member of his order Odon-
tolcae, which included Hesperornis. His student
Williston (1898) included it in this order when he
listed the Cretaceous birds of Kansas. Lucas (1903)
later suggested that it was separable from Hesper¬
ornis at the familial and perhaps the ordinal level.
Lambrecht (1933) thought Baptornis was related
to the loons and grebes and placed it in the family
Enaliornithidae, in which he also included Enali-
ornis, Neogaeornis, and possibly Eupterornis of the
French Paleocene. Romer, in earlier editions of his
text Vertebrate Paleontology, placed Eupterornis
questionably in the Baptornithidae, but later
(1966) followed Brodkorb (1963b) in placing
Eupterornis in the Gaviiformes and Baptornis in
the Podicipediformes. Wetmore (1956) used the
superorder Odontognathae for the New World
toothed birds and placed the Baptornithidae in
the order Hesperornithiformes. Brodkorb (1963b)
treated the Baptornithidae as an ancient family of
grebes, which included two genera, Baptornis and
Neogaeornis. At the same time, he put Enaliornis
and Eupterornis in the Gaviiformes (loons), along
with the volant Mesozoic foot-propelled divers of
the family Lonchodytidae. Brodkorb (1971:39)
has since reaffirmed his opinion regarding Baptor¬
nis in stating that “the proper position of the
family Baptornithidae is surely in the order Podici¬
pediformes.’' This position follows closely a state¬
ment by Lucas (1903:555): “In the slender cervi-
cals, arrangement of the tibia and patella, and

NUMBER 27

65

general structures of the leg, Baptornis is more like
a grebe than is the contemporary Hesperornis, and
if, with the small amount of material available, it
is deemed essential to establish any connection be¬
tween groups of existing and fossil birds, it is sug¬
gested that the ancestors of Baptornis are much
more likely to have been also the progenitors of the
Colymbine group [= grebes] than are those of
Hesperornis.” However, examination of much
better material of both Hesperornis and Baptornis
than was available to Lucas has failed to substan¬
tiate his statement. Actually, Hesperornis is more
grebe-like than is Baptornis, which tends more to
resemble the loons. Our studies here confirm that
Baptornis and Hesperornis resemble each other
more than they do any modern birds and that both
are far removed from, and unrelated to, either
loons or grebes. This removes the Podicipediformes
from the Cretaceous record, their next earliest oc¬
currence being in the lower Miocene (Arikareean)
of Oregon (Brodkorb, 1963b:227).

Other than Baptornis the only genus we pres¬
ently include in the Baptornithidae is Neogaeornis
from the Late Cretaceous of Chile. We have not
examined specimens of Neogaeornis, but the
highly compressed tarsometatarsus with the outer
trochlea slightly more distal than the middle one
(Lambrecht, 1933, fig. 100) suggests affiliation
with the Hesperornithiformes. The fact that the

outer and middle trochleae are of about equal
size supports an allocation to the Baptornithidae
rather than the Hesperornithidae.

Enaliornis, from the Lower Cretaceous of Brit¬
ain, shares many features with the Hesperornithi¬
formes, including a broad shallow fibular groove
on the femur, a broad triangular procnemial proc¬
ess and the absence of the supratendinal bridge of
the tibiotarsus, and the outer trochlea of the tarso¬
metatarsus slightly more distal than the middle
trochlea. Enaliornis shares many features with
Baptornis, from which it differs in having amphi-
coelous dorsal vertebrae. We presently maintain
the Baptornithidae as a family separate from the
Enaliornithidae. Brodkorb (1963a:60) placed Enali¬
ornis in the Gaviiformes stating that "although
currently placed in the order Hesperornithiformes,
these birds are slightly less specialized than mod¬
ern loons of the family Gaviidae and should be
transferred to the order Gaviiformes.” However,
we think Enaliornis is more similar to Baptornis
than to any other known bird and regard it as the
primitive basal stalk of the Hesperornithiformes.
This leaves Lonchodytes as the onfy Mesozoic bird
with any possible affinities with loons.

We agree with Baird’s (1967) suggestion that the
order Hesperornithiformes probably represents an
archaic group that became extinct along with the
great sea-lizards at the close of the Cretaceous.

Literature Cited

Baird, D.

1967. Age of Fossil Birds from the Greensands of New
Jersey. Auk, 84:260-262.

Brodkorb, P.

1963a. Birds from the Upper Cretaceous of Wyoming.
Proceedings of the 13th International Ornithologi¬
cal Congress, 1:55-70, 10 figures, 1 table.

1963b. Catalogue of Fossil Birds. Bulletin of the Florida
State Museum, Biological Sciences, 7:179-293.

1971. Origin and Evolution of Birds. Pages 19-55 in
volume 1 of D. S. Farner and J. R. King, editors.
Avian Biology. 2 figures, 7 tables. New York and
London: Academic Press.

Gingerich, P.

1972. A New Partial Mandible of Ichthyornis. Condor,
74(4): 471-473, 2 figures.

1973. Skull of Hesperornis and Early Evolution of Birds.
Nature, 23:70-73, 2 figures.

Gregory, J. T.

1952. The Jaws of the Cretaceous Toothed Birds Ich¬
thyornis and Hesperornis. Condor, 54:73-89, 9 fig¬
ures.

Heilman, G.

1927. The Origin of Birds. 210 pages, 142 figures. New
York: D. Appleton.

Howard, H.

1955. New Records and a New Species of Chendytes, an
Extinct Genus of Diving Geese. Condor, 57:135-
143, 1 figure.

Lambrecht, K.

1933. Handbuch der Palaeomithologie. 1022 pages, 209
figures. Berlin: Gebruder Borntraeger.

Lucas, F. A.

1903. Notes on the Osteology and Relationships of the
Fossil Birds of the Genera Hesperornis, Hargeria,
Baptornis, and Diatryma. Proceedings of the United
States National Museum, 26:545-556, 8 figures.

66

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Marsh, O. C.

1877. Characters of the Odontornithes, with Notice of a
New Allied Genus. American Journal of Science,
14:85-87, 1 figure.

1880. Odontornithes, a Monograph of the Extinct
Toothed Birds of North America. Report of the
United States Geological Exploration of the For¬
tieth Parallel, 7:1-201, 40 figures, 34 plates.

Martin, L. D„ and J. Tate, Jr.

1967. A Hesperornis from the Pierre Shale. Proceedings
of the Nebraska Academy of Sciences (77th Annual
Meeting): 40.

1969. New Information on Baptornis advenus. Proceed¬
ings of the Nebraska Academy of Sciences (79th
Annual Meeting): 49-50.

Ostrom, J. H.

1969. Osteology of Deinonychus antirrhopus, an Unusual
Theropod from the Lower Cretaceous of Montana.
Bulletin of the Peabody Museum of Natural His¬
tory, Yale University, 30:1-165, 83 figures.

Raikow, R. J.

1970. Evolution of Diving Adaptations in the Stifftail
Ducks. University of California Publications in
Zoology, 94:1-52, 32 figures, 16 tables.

Romer, A. S.

1966. Vertebrate Paleontology. Third edition, 468 pages,
4 tables, 443 figures. Chicago: University of Chi¬
cago Press.

Russell, D. A.

1967. Cretaceous Vertebrates from the Anderson River,
N.W.T. Canadian Journal of Earth Sciences, 4:21—
38, 12 figures.

Shufeldt, R. W.

1915. Fossil Birds in the Marsh Collection of Yale Uni¬

versity. Transactions of the Connecticut Academy
of Arts and Sciences, 19:1-110, 15 plates.

Storer, R. W.

1956. The Fossil Loon, Colymboides minutus. Condor,
58:413-426, 4 figures.

1958. Evolution in the Diving Birds. Proceedings of the
12th International Ornithological Congress, 2:694-
707, 9 figures.

1965. [Review of] P. Brodkorb, Catalogue of Fossil Birds,
Part 1. Auk, 82:657-658.

1971. Classification of Birds. Pages 1-18 in volume 1 of
D. S. Farner and J. R. King, editors, Avian Biol¬
ogy. 21 figures, 2 tables. New York and London:
Academic Press.

Swinton, W. E.

1965. Fossil Birds. 65 pages, 25 figures, 12 plates. London:
British Museum.

Walker, M. V.

1967. Revival of Interest in the Toothed Birds of Kansas.
Transactions of the Kansas Academy of Sciences,
70(l):60-66.

Wetmore, A.

1956. A Check-List of the Fossil and Prehistoric Birds
of North America and the West Indies. Smith¬
sonian Miscellaneous Collections, 131 (5): 1—105.

Williston, S. W.

1898. Birds. The University Geological Survey of Kan¬
sas, 4(2):43-64, 1 figure, 4 plates.

Zusi, R. L., and R. W. Storer

1969. Osteology and Myology of the Head and Neck of
the Pied-billed Grebes ( Podilymbus). Miscellaneous
Publications of the Museum of Zoology, University
of Michigan, 139:1-49, 19 figures, 4 tables.

Discovery of a Cretaceous Bird,
Apparently Ancestral to the
Orders Coraciiformes and Piciformes
(Aves: Carinatae)

Pierce Brodkorb

ABSTRACT

Alexornis antecedens, new genus and species, is
described from the Bocana Roja Formation, Upper
Cretaceous (Campanian age), near El Rosario,
Baja California, Mexico. The humerus, ulna, scap¬
ula, coracoid, femur, and tibiotarsus are repre¬
sented. The fossil is referred to a new family, Alex-
ornithidae, and a new order, Alexornithiformes,
thought to be ancestral to the Tertiary and Recent
orders Coraciiformes and Piciformes. Since Caenag¬
nathus collinsi Sternberg and C. sternbergi Cra-
craft are reptiles, and Gobipteryx minuta Elzan-
owski appears to be reptilian also, Alexornis is the
only certain land bird known from the Cretaceous.

Introduction

The major deficiency in our knowledge of the
history of birds is their inadequate Mesozoic rec¬
ord. The only known Jurassic genus is Archaeo¬
pteryx, which was certainly a land bird, whether it
was arboreal (Brodkorb, 1971b) or a ground-
dweller (Ostrom, 1974). In the Cretaceous period
we have records of about 37 species of water birds,
distributed among 18 genera, 12 families, and 7
orders. Because many groups of land birds occur
early in the Tertiary, they must also have been
present and undergoing radiation during the
Cretaceous. However, none of the hitherto known

Pierce Brodkorb, Department of Zoology, University of Flor¬
ida, Gainesville, Florida 32611.

alleged birds from terrestrial Cretaceous deposits is
certainly referable to the class Aves.

Three species of supposed land birds have been
described from the late Cretaceous. These are:
Caenagnathus collinsi Sternberg (1940), Caenagna-
thns sternbergi Cracraft (1971), and Gobipteryx
minuta Elzanowski (1974). Each was described
from a single specimen of jaw or skull. Although
not comparable with any living or fossil bird, the
describers of Caenagnathus and Gobipteryx made
them the types of new families and orders of Aves.

Sternberg (1940) based Caenagnathus collinsi
on a mandible from the Belly River Series of Al¬
berta, and regarded it as representing a new order
of birds. Wetmore (1960) transferred it to the
theropod dinosaurs, near Ornithomimidae, an as¬
signment in which Romer (1966) concurred. Cra¬
craft (1971) returned the genus to Aves and
founded a second species, C. sternbergi, on a
scrap of a lower jaw from the same formation as
C. collinsi. White (1973) included Caenognathus
[sic] and Caenognathidae [sic] as a valid genus
and family of coelurosaurian theropods. Dale A.
Russell of the National Museum of Canada (pers.
comm.) informs me that new material of Caenagna¬
thus, plus specimens of related forms from Mon¬
golia, confirms that the Caenagnathidae are thero¬
pod dinosaurs having affinities with Oviraptor of
the Ornithomimidae, thus vindicating Wetmore.

Gobipteryx minuta was described on a small,
poorly preserved, flattened skull and mandible
from the Barun Goyot Formation in Mongolia.
Neither the specimen, which I have seen, nor the
published illustrations, are convincingly avian.

67

68

Elzanowski (1974) placed Gobipteryx in the “Su¬
perorder Palaeognathae” (i.e., Ratitae), but this
is certainly wrong. The specimen is quite small,
and the palate is unlike that of ratites or any other
bird. Better material is needed before it can be as¬
signed confidently to either Aves or Reptilia.

With the relegation of the above forms to Rep¬
tilia or taxa incertae sedis, there are no land birds
known from the 72-million-year span of the Cre¬
taceous period. It was therefore with great interest
that I agreed to study some possibly avian remains
from a terrestrial deposit of Late Cretaceous age
in Baja California, Mexico. The best preserved
specimens so far recovered are from a small land
bird—represented by elements of the pectoral
girdle, wings, and legs—the subject of the present
paper.

Stratigraphy.— The Rosario Group is composed
of three formations, each separated by an uncon¬
formity. In descending order these are the Rosario,
“El Gallo,” and “La Bocana Roja.” According to
Morris (1972) the last two formations were de¬
fined by Kilmer in his doctoral dissertation (1963).

The Rosario Formation is a marine deposit
whose invertebrate fauna was assigned an early
Maestrichtian or late Campanian age by Durham
and Allison (1960). The nannofossils indicate that
it is situated close to the Maestrichtian-Campanian
boundary (Morris, 1973).

The Gallo Formation has a thickness of more
than 150 m. A potassium/argon (K/Ar) date of
73 ± 2 million years is available for the middle
third of the formation (Morris, 1972, 1973). This
would make it of late Campanian age. It contains
the remains of hadrosaurian dinosaurs (cf. Lambe-
osaurus), smaller reptiles, amphibians, mammals,
and large tree trunks with well-developed root sys¬
tems. This assemblage is the only extensive Late
Cretaceous terrestrial vertebrate fauna from the
Pacific margin of North or Middle America.

The Bocana Roja Formation contained the re¬
mains of the bird described here. The type-
specimen of the theropod dinosaur Labocania
anomala Molnar (1974) also came from this for¬
mation, along with hadrosaur ribs and numerous
small vertebrate fossils, as yet unstudied. The age
of the formation is Campanian or earlier.

During the Cretaceous period the arrangement
of the continents was different from that of today,
Mexico being in proximity to North Africa and

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

bordering the western part of the Tethys Sea
(Deitz and Holden, 1970).

Acknowledgments. —Field work was supported
by the National Geographic Society through
grants to Dr. William J. Morris of Occidental Col¬
lege and was sponsored by the Instituto de Geo-
logia of the Universidad Nacional Autonoma de
Mexico and by the Natural History Museum of
Los Angeles County. Over several years the field
parties included, in addition to Dr. Morris, Dr.
Ismael Ferrusquia V. of the Instituto de Geologfa
and Harley Garbani, J. Loewe, and Robert M.
McKenzie of the Natural History Museum of Los
Angeles County (LACM). I am greatly indebted to
Dr. Hildegarde Howard for suggesting that the
fossils be sent to me for study. The beautiful photo¬
graphs of the tiny specimens are by Victor E.
Krantz of the Smithsonian Institution. Donald
Baird, Nicholas Hotton III, John H. Ostrom, and
Dale A. Russell made suggestions or furnished in¬
formation about Caenagnathus and Gobipteryx.

Alexornis, new genus

Type of Genus. — Alexornis antecedens, new
species.

Diagnosis. —As for the type and only known
species.

Etymology. —From Greek alexo (I defend)
and ornis (common gender, masculine selected
here) bird. On the occasion of his ninetieth birth¬
day this genus is dedicated to my friend Alexander
Wetmore, who, in addition to his many other ac¬
complishments, has done more to foster paleorni-
thology and has described more species of fossil
birds than any other author.

Alexornis antecedens, new species
Figure I

Holotype.— Distal 10 mm of right humerus,
LACM 33213 (Figure 1 a,b). From LACM locality
7256, 6 miles southwest of El Rosario, Baja Cali¬
fornia del Norte, Mexico. Bocana Roja Forma¬
tion, Upper Cretaceous, Campanian age. Collected
by H. Garbani and J. Loewe, 16 July 1971.

Paratype.— Distal 10 mm of left humerus, col¬
lected in association with the holotype and cata¬
loged with the same number.

NUMBER 27

69

Figure 1 .—Alexornis antecedens, new genus and species, Bocana Roja Formation, Campanian
age, near Rosario, Baja California, Mexico (LACM 33212): a, holotype right humerus, palmar
view; b, same, anconal view; c, left scapula, ventral view; d, same, dorsal view; e, left coracoid,
anterior view; /, same, posterior view; g, right ulna, internal view; h, same, external view; i,
right tibiotarsus, anterior view; /, same, posterior view; k, left femur, anterior view; l, same,
posterior view; m, same, lateral view. (All X 5.)

70

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Hypodigm. —The holotype, paratype, and the fol¬
lowing referred material, collected in association
with the types and cataloged under the same num¬
ber: proximal 4 mm of left scapula (Figure 1 c,d),
upper 4.4 mm of left coracoid (Figure 1 e,f), proxi¬
mal 3.5 mm of right ulna (Figure 1 g,h), distal 6
mm of left femur (Figure 1 k-m), and proximal
16.3 mm of right tibiotarsus (Figure More

than 20 other fragments were also collected with
the types and cataloged under the same number,
but they are left unidentified at this time.

Diagnosis. —Comparison of the hypodigm with
the homologous skeletal elements of the known
orders of birds shows that the resemblances of
Alexornis are closest to certain members of the
Piciformes and Coraciiformes. Within those two
orders the piciform family Bucconidae and the
coraciiform family Momotidae have the most simi¬
larity to the fossil. The fossil shares certain char¬
acters with both Bucconidae and Momotidae, some
with Bucconidae alone, and some with Momotidae
alone; but more of its characters are unique. In
size the fossil falls between the bucconid Mala-
coptila panamensis and the motmot Hylomanes
momotula (Table 1).

Etymology. —Latin antecedens, going before in
rank or time, ancestral, in reference to the sup¬
posed ancestry of this bird to the orders Piciformes
and Coraciiformes.

Description.— Humerus: (1) Olecranal fossa
deep (shallow in Bucconidae and Momotidae),
and (2) wide, extending toward entepicondylar
area (as in Momotidae; in Bucconidae less ex¬
tended toward entepicondyle). (3) Entepicondyle
produced distally, resembling Bucconidae (in Mo¬
motidae less produced; in Passeriformes much
more produced). (4) External condyle bulbous as
in both Bucconidae and Momotidae, but (5)
oriented transversely at an angle of about 60 de¬
grees to shaft (in Momotidae angle to shaft is
about 45 degrees; in Bucconidae condyle is more
upright at angle of 30 degrees to shaft). (6) Inter¬
nal condyle lies inclined toward entepicondyle,
resembling condition in Momotidae (in Bucconi¬
dae condyle is more transverse), (7) with a strong
facet for medial cotyla of ulna, resembling both
Momotidae and Bucconidae. (8) Ectepicondylar
prominence large and rounded, resembling Buc¬
conidae (less developed in Momotidae), (9) with
a transverse ridge across anconal surface proximal

Table 1.—Measurements (mm) of skeletal elements

Character

Alexornis

antecedens

Hylomanes
* momotula

Malacoptila

panamensis

Humerus

Distal width .

5.6, -

4.4

4.8

Shaft width .

2.2, 2.3

1.7

2.4

Shaft depth .

1.8, 1.8

1.6

1.8

Least depth through
brachial depression

1.1, 1.3

1.1

1.7

Scapula

Diagonal width across
acromion and
glenoid .

4.2

3.5

4.6

Shaft width .

1.7

1.6

1.5

Femur

Distal width

[3.1]

3.4

3.7

Shaft width .

2

1.7

2

External condyle depth

2.35

2.5

2.5

Fibular condyle depth

1.6

2.2

2.2

Internal condyle depth

2.6

2.0

2.1

External condyle

height .

2.5

1.5

1.6

Fibular condyle height

1.6

1.3

1.4

Internal condyle
height .

2

1.7

1.3

Shaft depth .

1.8

1.4

1.7

Tibiotarsus

Proximal width

3.7

2.6

2.9

Shaft width

1.9

1.3

1.6

Shaft depth .

1.6

1.2

1.6

Width through internal
cnemial crest .

4.1

3.1

4

Length through fibular
ridge .

10.5

6.7

8.1

* The first measurement of the humerus of Alexornis is of
the holotype, the second of the paratype.

Measurement in brackets estimated.

to its base (ridge absent in Bucconidae and
Momotidae).

Insofar as preserved, the humerus of Alexornis
has three unique characteristics (numbers 1, 5,
and 9 above), shares two with Bucconidae alone
(3 and 8), shares two with Momotidae alone (2
and 6), and is similar to both Bucconidae and
Momotidae in two others (4 and 7).

Ulna: (1) Olecranon straight, short, stout (as in
Bucconidae and Momotidae), (2) with a pit in the
tip (no pit in Bucconidae and Momotidae). (3)
External cotyla large and (4) strongly convex (of
moderate size and moderately convex in Bucconi-

NUMBER 27

71

dae; small and concave in Momotidae), (5) sepa¬
rated from olecranon by a deep groove (groove
absent in Bucconidae and Momotidae); (6) me¬
dial rim of external cotyla thick [lateral portion
of cotyla missing] (resembling Bucconidae; edge
very thin in Momotidae). (7) Internal cotyla small
(as in Bucconidae; large in Momotidae), (8) with
surface flat (moderately concave in Bucconidae;
strongly concave in Momotidae). (9) Proximal
radial depression deeply undercuts entire width of
rim of internal cotyla (in Bucconidae the depres¬
sion falls far short of rim of cotyla; in Momotidae
the depression extends to the medial edge of cotyla
but fails to undercut it).

Insofar as preserved, the ulna of Alexornis has
six unique characteristics (numbers 2, 3, 4, 5, 8,
and 9 above), shares two with Bucconidae alone
(6 and 7), shares none with Momotidae alone, and
is similar to both Bucconidae and Momotidae in
one characteristic (1).

Scapula: (1) Acromion rather short (long in
Bucconidae and Momotidae), with tip slightly
damaged, but (2) apparently blunt (as in Buc¬
conidae; tip forms a recurved hook in Momotidae
and Passeriformes). (3) Glenoid facet flat (cup¬
like in Bucconidae and Momotidae).

Thus the scapula, so far as preserved, has two
unique characteristics (numbers 1 and 3), shares
one with Bucconidae (2), and none with Momo¬
tidae.

Coracoid: (1) Brachial tuberosity with a re¬
curved hook directed toward area where pro¬
coracoid process would be if preserved (slightly
hooked in Bucconidae; hook absent in Momoti¬
dae). (2) Triosseal canal very deep (very shallow
in Bucconidae; flat in Momotidae). (3) Scapular
facet convex (as in Bucconidae and Momotidae),
but (4) very broad (very narrow in Bucconidae
and Momotidae).

The coracoid has three unique features (num¬
bers 1, 2, and 4), none is shared with Bucconidae
alone or with Momotidae alone, and one is shared
with both Bucconidae and Momotidae (3).

Femur: The specimen shows evidence of some
postmortem compression and distortion. (1) Shaft
stout (resembling Bucconidae and Eurylaimidae;
slender in Momotidae). (2) External condyle very
long, extending both proximally and (3) distally
far beyond both internal and fibular condyles (ex¬
ternal and internal condyles of about equal extent

in Bucconidae and Momotidae; in Eurylaimidae
external condyle lengthened distally only). (4) In¬
ternal condyle with only very slight indication of
a transverse shelf on posterior surface (resembling
Bucconidae and Eurylaimidae; shelf very promi¬
nent in Momotidae). (5) Fibular condyle small
(resembling Momotidae and Eurylaimidae; very
stout in Bucconidae). (6) Popliteal area deeply
excavated (resembling Momotidae; area nearly
flat in Bucconidae and Eurylaimidae). (7) Rotular
groove shallow (well developed in Bucconidae and
Momotidae).

The femur has three unique features (numbers
2, 3, and 7), three are shared with Bucconidae
alone (1, 4, and 5), one with Momotidae alone
(6), and none are held in common with both Buc¬
conidae and Momotidae.

Tibiotarsus: The cnemial crests are broken off,
but their bases are preserved. (1) Shaft stout (re¬
sembling Bucconidae; slender in Momotidae).
(2) Fibular crest wide (rudimentary in Bucconidae
and Momotidae), (3) extending proximally all
the way up shaft (as in Momotidae; in Bucconidae
falling far short of proximal 'end of shaft). (4)
Distal end of fibular crest merges gently with shaft
(as in Momotidae; in Bucconidae distal end joins
shaft at a rather abrupt angle). (5) Anterior and
posterior surfaces of fibular crest concave, with an
anterior and a posterior groove running along
junction with shaft (a slight anterior and posterior
groove in Bucconidae; in Momotidae anterior and
posterior surfaces of crest flat and ungrooved). (6)
Proximal internal articular surface swollen and
convex (in Momotidae slightly swollen; surface
more concave in Bucconidae), (7) without posterior
overhang (with slight overhang in Bucconidae; in
Momotidae a lip overhangs shaft posteriorly). (8)
Outer cnemial crest short (as in Bucconidae and
Momotidae). (9) Inner cnemial crest short, although
considerably longer than outer crest (resembling
Bucconidae and Momotidae).

The tibiotarsus has three unique characteristics
(numbers 2, 5, and 7), shares one with Bucconidae
alone (1), shares three with Momotidae alone (3,
4, and 6), and agrees with both Bucconidae and
Momotidae in two features (8 and 9).

Familial Position of Alexornis

The characteristics described above for Alexor-

72

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

nis are grouped in Table 2 to show the number of
features confined to a single taxon, those shared by
two taxa, and those common to all three taxa. By
far the strongest grouping of characteristics is of
those confined to a single taxon—49 percent in
Alexornis, 44 percent in Momotidae, and 37 per¬
cent in Bucconidae. I interpret this as indicating
that the three taxa are of equal taxonomic rank
and, therefore, propose Alexornis as the type of a
new family.

Table 2.—Summary of shared characteristics of Alexornis,
Momotidae, and Bucconidae (+ = similar to; — = differ¬
ent from)

Wing Pectoral Leg Total
girdle

_ (18) (?)

1. Alexornis 3 1

+ Momotidae
+ Bucconidae

(16) (41)

2 6

<u

be

w

C

o

o

ll

<D

P*

15

2. Alexornis

-I- Momotidae
— Bucconidae

2 0 4 6 15

3. Alexornis

+ Bucconidae
— Momotidae

4 1 4 9 22

4. Bucconidae
+ Momotidae
— Alexornis

4 3 4 11 27

5. Bucconidae

— Momotidae
— Alexornis

6. Momotidae

— Bucconidae
— Alexornis

7 2 6 15 37

9 3 6 18 44

7.

A lexornis

— Momotidae

— Bucconidae

9

5

6 20 49

Numbers in parentheses represent number of characters
considered.

ALEXORNITHIDAE, new family

Diagnosis. —Humerus with olecranal fossa deep;
entepicondylar area much produced distally; ex¬
ternal condyle oriented transversely at an angle of
about 60 degrees to shaft; ectepicondylar promi¬

nence with a transverse ridge across anconal sur¬
face proximal to its base. Ulna with a pit at the tip
of olecranon; external cotyla large, strongly con¬
vex, and separated from olecranon by a deep
groove; surface of internal cotyla flat; proximal
radial depression deeply undercutting the entire
width of lip of internal cotyla. Scapula with acro¬
mion rather short and glenoid facet flat. Coracoid
with brachial tuberosity hooked; triosseal canal
very deep; scapular facet very broad. Femur with
external condyle very long, extending both proxi-
mally and distally far beyond both internal and
fibular condyles; rotular groove obsolete. Tibio-
tarsus with its proximal articular surface not over¬
hanging shaft posteriorly; fibular crest wide, with
both its surfaces concave and separated from shaft
by an anterior and a posterior groove.

Ordinal Position of Alexornis

Although possessing a large number of unique
features, Alexornithidae shares some characters
with the order Coraciiformes as exemplified by the
Momotidae, others with the order Piciformes as
exemplified by the Bucconidae, and still others
with both of those orders. These similarities are
summarized in Table 2, and several different hypo¬
theses at the ordinal level could be formed from
these data.

The three taxa might be combined in a single
order, for which the name Piciformes would have
priority (for order-group synonymies see Brodkorb,
197la:248, 256). But of the 41 characters analyzed,
only 6 are shared by the 3 families (Table 2, line
1). Such a small proportion of common character¬
istics militates against merging the taxa in a single
order.

Alexornithidae might be referred to Coracii¬
formes, but such an arrangement is also supported
by 6 characters only (Table 2, line 2), and this
hypothesis is likewise discarded.

Alexornithidae might be placed in Piciformes,
as the order is currently understood. Nine char¬
acters support this combination (Table 2, line 3),
but line 4 of Table 2 argues against it, as Pici¬
formes and Coraciiformes share more characters
than any other combination of the taxa under
consideration.

The conclusion thus derived from Table 2 is
that the three taxa represent separate but related

NUMBER 27

73

orders. I therefore propose Alexornis as the type-
genus of a new order.

ALEXORNITHIFORMES, new order

Diagnosis. —Same as for the only known family,
Alexornithidae.

Remarks. —The age of Alexornis is about 81
million years bp, much earlier than the earliest
known occurrence of either the Coraciiformes and
Piciformes. The earliest record of the Piciformes
is early Eocene, about 51 million years bp, when

bucconid-like forms appear in Wyoming (Brod-
korb, 1970; Feduccia and Martin, herein). If Har¬
rison and Walker (1972) are correct in assigning
the British Halcyornis to the Coraciiformes, the
earliest record of that order is also early Eocene.
Undoubted members of the Coraciiformes occur
in European deposits of middle and late Eocene
age (Brodkorb, 1971a). Both morphology and the
temporal sequence thus suggest Alexornis as the
presumptive ancestor of the orders Coraciiformes
and Piciformes.

Literature Cited

Brodkorb, P.

1970. An Eocene Puffbird from Wyoming. Contributions
to Geology, University of Wyoming, 9(1): 13-15, 1
figure.

1971a. Catalogue of Fossil Birds, Part 4 (Columbiformes
through Piciformes). Bulletin of the Florida State
Museum, Biological Sciences, 15(4): 163-266, 1 fig¬
ure.

1971b. Origin and Evolution of Birds. Pages 19-55 in
volume 1 of D. S. Farner and J. R. King, editors.
Avian Biology. 2 figures, 7 tables. New York and
London: Academic Press.

Cracraft, J.

1971. Caenagnathiformes: Cretaceous Birds Convergent
in Jaw Mechanism to Dicynodont Reptiles. Journal
of Paleontology, 45(5): 805-809, 2 figures, plate 92.

Deitz, R. S., and J. C. Holden

1970. Reconstruction of Pangaea: Breakup and Disper¬
sion of Continents, Permian to Present. Journal of
Geophysical Research, 75:4939-4956.

Durham, J. W., and E. C. Allison

1960. Geologic History of Baja California and its Marine
Faunas. Systematic Zoology, 9:47-91.

Elzanowski, A.

1974. Preliminiary Note on the Palaeognathous Bird
from the Upper Cretaceous of Mongolia. Part 5 in
Results of the Polish-Mongolian Palaeontological
Expeditions. Palaeontologia Polonica, 30:103-109,
2 figures, plates 32-33.

Harrison, C. J. O., and C. A. Walker

1972. The Affinities of Halcyornis from the Lower
Eocene. Bulletin of the British Museum (Natural
History), Geology, 21(4): 151-169, 9 figures, 3 plates.

Kilmer, F. H.

1963. Cretaceous and Cenozoic Stratigraphy, El Rosario
Area, Baja California. Ph.D. dissertation. Univer¬
sity of California, Berkeley.

Molnar, R. E.

1974. A Distinctive Theropod Dinosaur from the Upper
Cretaceous of Baja California (Mexico). Journal of
Paleontology, 48(5): 1009-1017, 4 figures.

Morris, W. J.

1972. A Giant Hadrosaurian Dinosaur from Baja Cali¬
fornia. Journal of Paleontology, 46(5):777-779, 1
figure.

1973. A Review of Pacific Coast Hadrosaurs. Journal of
Paleontology, 47(3):551—561, 5 figures.

Ostrom, J. H.

1974. Archaeopteryx and the Origin of Flight. Quarterly
Review of Biology, 49:27-47, 10 figures.

Romer, A. S.

1966. Vertebrate Paleontology. Third edition, ix + 468
pages, 443 figures. Chicago: The University of Chi¬
cago Press.

Sternberg, R. M.

1940. A Toothless Bird from the Cretaceous of Alberta.
Journal of Paleontology, 14(1):81—85, 6 figures.

Wetmore, A.

1960. A Classification for the Birds of the World. Smith¬
sonian Miscellaneous Collections, 139(11): 1-37.

White, T. E.

1973. Catalogue of the Genera of Dinosaurs. Annals of
Carnegie Museum, 44(9): 117-155.

A Survey of the Paleogene Birds of Asia

E. N. Kurochkin

ABSTRACT

In the past few years, much new material of Pal¬
eogene birds has been collected in Mongolia and
Kazakhstan, the fossil collections now comprising
350 satisfactory specimens of postcranial elements.
The first Paleocene birds known from Asia were
obtained in Mongolia, where ^ remains referable to
several avian orders were recovered. Fossils from a
new middle Eocene site at Khaichin Ula 2, also
situated in the South Gobi of Mongolia, were iden¬
tified as various waterbirds and galliforms. A va¬
riety of birds is represented in early Oligocene
material from Mongolia and a wealth of new avian
material of middle Oligocene age has been ob¬
tained from the so-called Indricotherium beds of
Central Kazakhstan. Paleoecological conditions in
the Asian Paleogene are assessed on the basis of
avian fossils, and the sketchy picture of the pos¬
sible interrelationships of Paleogene birds from
Asia, Europe, and North America is outlined.

Abundant new fossil material has made possible
a deeper insight into the composition, characteris¬
tics, and relationships of the gruiform families
Eogruidae and Ergilornithidae, and proves the ex¬
istence of a phylogenetic continuity between the
Eocene genus Eogrus, the Oligocene genera Ergi-
lornis and Proergilornis, and the Pliocene genus
Urmiornis. A fragment of humerus assigned to the
Ergilornithidae shows that these birds were
flightless.

Introduction

Until recently, knowledge of the Paleogene birds
of Asia has been derived mainly from casual dis¬
coveries, with few attempts at a purposeful gather-

E. N. Kurochkin, Paleontological Institute of the USSR
Academy of Sciences, Leninsky Prospect, 33, Moscow, 117071,
U.S.S.R.

ing of avian remains and their subsequent study
being made. Publications on the subject have gen¬
erally been concerned with systematic descriptions
of individual specimens collected in association
with Paleogene mammals from isolated Asian lo¬
calities. The list of these publications is deplorably
short, consisting of 15 titles, in which 21 forms of
birds are described.

Discoveries of Paleogene birds in Asia have been
confined largely to the following three territories:
Central Kazakhstan east of the Aral Sea; East
Kazakhstan east of Lake Zaisan; and the southern
part of Mongolia together with adjoining provin¬
ces of China (Figure 1). Avian fossils from outside
this area consist of individual specimens from the
early middle Eocene of Fergana in Kirgizia ( Eoba-
learica tugarinovi Gureev, 1949), the middle
Eocene of Sumatra ( Protoplotus beauforti Lam-
brecht, 1931), and Gruidae genus indet. from the
Oligocene of Ordos, China (Teilhard de Chardin,
1926).

Three upper Eocene birds, two of which were
described as new, were reported by Wetmore
(1934) from China. Lower Oligocene ergilorni-
thids from East Gobi, Mongolia, were studied by
Kozlova (1960). Several species of birds of prey,
Anseriformes and rails were described by Kuroch¬
kin (1968a, 1968b) from the Oligocene of Kazakh¬
stan and Mongolia. Other specimens of Paleogene
birds from Asia were discussed by Tugarinov
(1940) and Bendukidze (1971). A record of Aquila
sp. from middle Oligocene deposits at Min Eske
Suek, Kazakhstan (Aubekerova, 1965) has not been
confirmed. Surveys of the Tertiary birds from the
U.S.S.R, including several Paleogene discoveries,
were published by Burchak-Abramovich (1958) and
the present writer (Kurochkin, 1974).

In the past few years, great quantities of new
avian material have been obtained from the Oligo-

75

76

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Figure 1.—Schematic map showing collection localities for Paleogene birds in Asia. Paleocene:

I, Naran Bulak and Tsagan Khushu. Middle Eocene: 2, Khaichin Ula 2; 3, Kolobolchi; 4,
Kalmakpai; 5, Andarak (in Fergana). Late Eocene: 6, Andreevka; 7, Irdin Manha and Ulan
Shireh; 8, Iren Dabasu; 9, Tung Gur (Miocene); 10, (inset) Sipang, Sumatra. Early Olicocene:

II, Zhongiz Shoki; 12, Ergelyeen Dzo (Ardyn Obo, Ergil Obo); 13, Khoer Dzan. Middle Olico¬
cene: 14, Kur Sai and Min Sai (Tchelkar Nura); 15, Donguz Tau; 16, Kyzyl Kak; 17, Kusto;

18. Kusto Kysylkain; 19, Tatal Gol (Shand Gol
21. Shunkht. Late Oligocene: 22. Agispe.

cene of Kazakhstan and in various other Paleogene
localities. This material was gathered by expedi¬
tions of the Paleontological Institute of the
U.S.S.R. Academy of Sciences, and also by a joint
Soviet-Mongolian paleontological expedition. The
tentative studies made so far have shown that a
small part of this material belongs to previously
known taxa, whereas the greater part represents
completely new forms.

In all, about 240 fossils of Paleogene birds have
been found in Mongolia. These include the first
Asian Paleocene birds known, found in deposits at
Naran Bulak and Tsagan Khushu, as well as in¬
teresting discoveries made in the Eocene deposits
at Khaichin Ula 2 and Kolobolchi, and a large col¬
lection from the Oligocene localities at Ergelyeen
Dzo, Khoer Dzan, Tatsin Gol, Tatal Gol, and
others. New material from the Oligocene sites at
Kur Sai and Min Sai, Donguz Tau and Kyzyl Kak
in Central Kazakhstan and several Oligocene lo¬
calities in the Zaisan basin, East Kazakhstan, in-

, Loh, Tatshin Gol); 20, Sen Zhak (Ordos, China);

eludes 80 fragments of avian bones, most of which
can be identified to the species level.

Acknowledgments.— I want to thank many per¬
sons for their valuable assistance in preparing this
survey and in gathering the fossil birds reported in
this work: Dr. N. Shevyreva, Paleontological In¬
stitute of the U.S.S.R. Academy of Sciences (PIN),
who provided me with avian material from Zaisan
basin; V. Y. Reshetov, also of the Paleontological
Institute, for his valuable contributions to the dis¬
covery of fossil birds in Naran Bulak, Tsagan
Khushu, and Khaichin Ula 2; Dr. D. Dashzeveg,
Geological Institute of the Mongolian Academy of
Sciences, who provided me with avian remains
from Tsagan Khushu and Khoer Dzan; Dr. Mal¬
colm C. McKenna, American Museum of Natural
History, for his help in obtaining casts of the or¬
iginal material of Eogrus and for editorial assist¬
ance; Dr. K. Yudin and Dr. B. Nekrasov, Zoologi¬
cal Institute of the U.S.S.R. Academy of Science,
for permitting me to study a skeleton of Strigops;

NUMBER 27

77

Dr. A. Sudilovskaya and S. V. Lutskaya, Ornithol¬
ogical Department of Moscow University, for their
kind permission and assistance in using the osteo-
logical collection of the Zoological Museum of
Moscow University; Prof. K. K. Flerov, Institute of
Evolutionary Morphology and Ecology of Animals
of the U.S.S.R. Academy of Sciences; Dr. B. Trofi¬
mov, Paleontological Institute of the U.S.S.R.
Academy of Sciences, for valuable comments on
the manuscript; T. D. Rakova for preparing the
illustrations accompanying this survey; and N. N.
Sukhareva for technical editing.

The Paleocene

Birds have not previously been known from the
Paleocene of Asia. As a result of field work by the
Soviet-Mongolian paleontological expedition, the
first such avian fossils were obtained from Naran
Bulak and Tsagan Khushu in South Gobi, in the
western part of the Nemeget basin. The geology of
these localities, which are situated about 6 km from
one another, has been described by Novozhilov
(1954), Gradzirisky, et al. (1968), and Shishkin
(1975). The mammalian assemblage from these
sites is generally referred to as the Naran Bulak
fauna (Belaeva, et al., 1974). Although immedi¬
ately after its discovery this fauna was believed to
be early Eocene in age (Efremov, 1954; Novizhilov,
1954), it is presently placed in the late Paleocene
(Gradzinsky, et al., 1968; Szalay and McKenna,
1971; Belaeva, et al., 1974). The Naran Bulak
mammalian fauna consists of insectivores, multi-
tuberculates, lagomorphs, condylarths, dinocera-
tids, pantodontids, and notoungulates. It has
many species in common with the Gashato fauna
collected by the Central Asiatic expedition of the
American Museum of Natural History (Belaeva,
et al., 1974).

In contrast to the early Oligocene avian mate¬
rial from Asia, which appears to consist exclusively
of genera and families more or less directly related
to Recent birds, a preliminary acquaintance with
the avian remains from Naran Bulak produces the
distinct impression that these birds are fundamen¬
tally different from Recent families in their struc¬
ture, although they may be assigned quite un¬
ambiguously to Recent orders. A pelecaniform
humerus is interesting in that it bears certain simi¬
larities to both the Sulidae and the Phalacrocoraci-

dae. A coracoid and a carpometacarpus referable
to the extinct anatid subfamily Romainvilliinae
exhibit obvious primitive traits, particularly as
compared to the relatively homogeneous skeletal
structure of Recent, Neogene, and Oligocene
anatids.

In washing thousands of mammalian bones at
Tsagan Khushu, only 12 avian bones were found.
Half of these were coracoids and scapulae belong¬
ing to small anseriforms possessing specific struc¬
tural details not found in Recent ducks. Also in¬
cluded were remains of a peculiar new type of
heron similar to the Ardeidae, and shorebirds hav¬
ing definite affinities with the Scolopacidae.

The Eocene

Gureev (1949) described Eobalearica tugarinovi
from a distal end of tibiotarsus from Andarak in
the Fergana valley, a site which was then consid¬
ered to be late Eocene. New stratigraphic correla¬
tions have shown it to belong to the early middle
Eocene (Hekker, et al., 1962) ; Recently, bird re¬
mains have been reported from the Zaisan basin,
where sediments bearing the so-called Obailin
fauna are rather common. Various authors date
this fauna as middle to late Eocene. Bendukidze
(1971) described Progrus turanicus from a distal
end of tibiotarsus from the Kalmakpai River, with¬
out giving an exact indication of locality. He
placed this bird in the family Geranoididae, but
this allocation would appear to require further
scrutiny. P. A. Aubekerova of the Institute of Zool¬
ogy of the Kazakh S.S.R. Academy of Sciences ob¬
tained unidentifiable fragmentary avian remains
from the late Eocene localities near Andreevka on
the Chinzhili River in Southeast Kazakhstan.
Another isolated discovery was that of Protoplotus
beauforti Lambrecht (1931), an upper Eocene
anhinga from Sipang, in the western part of Su¬
matra, which was described from an almost com¬
plete skeletal impression.

Wetmore (1934) examined the paleornithologi-
cal material collected by the Central Asiatic expe¬
dition at Ulan Shireh (upper Eocene of Irdin
Manha) in China. He referred a coracoid to the
Falconiformes and also described a femur as a new
genus and species of Rallidae, Telecrex grangeri.
Recently, Olson (1974) has provided convincing
evidence that Telecrex grangeri is a representative

78

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

of the family Numididae, showing some similari¬
ties to the modern genus Phasidus. This is the
earliest evidence of the presence of this group in
Asia. From the same locality, Wetmore (1934) de¬
scribed numerous specimens of a crane as a new
genus and species, Eogrus aeola, and reported
specimens thought to be of the same species from
another upper Eocene locality in Inner Mongolia
in the vicinity of Iren Dabasu. Material referable
to Eogrus aeola was obtained in 1959 by the Soviet-
Chinese paleontological expedition at the same
locality, Irdin Manha, in the Shara Murun region
of China. This includes 10 distal ends of tibiotarsi
and tarsometatarsi (PIN 2197-185-194), which
confirm the original characterization of Eogrus and
provide a solid basis for a comparative study of
new specimens of Eogruidae and Ergilornithidae
from the lower Oligocene.

Of interest are recent discoveries from Khaichin
Ula 2 on the southern slope of Bugin Tsav in
southern Gobi. In 1971-1973, 27 avian postcranial
bones were found in these sediments, which are
middle upper Eocene in age (Shuvalov, et ah,
1974). Most of the bones are well preserved and
can be studied in detail. Tentative identifications
of this material have shown that the majority of
the bones belong to shorebirds of two groups. The
lesser part of these are clearly referable to the
Scolopacidae, whereas the remainder may be at¬
tributed to small birds of the suborder Chara-
drioidea having no parallel among the recent fam¬
ilies of this group and represented by species of
different sizes. Coracoids and scapulae were espe¬
cially numerous and a radius and tarsometatarsus
were also found. From the same locality are re¬
mains belonging to the Ardeidae, Rallidae, and
Threskiornithidae, as well as to unidentified pe¬
culiar types of Galliformes, including synsacral
fragments belonging to a bird of the family Numi¬
didae. Several bones were found by us in 1972 in
another middle upper Eocene locality at Kolobol-
chi, in the Lake Valley near Somon Bogd,
Mongolia.

The Early Oligocene

Information on the early Oligocene of Asia,
especially of birds, is, in general, extremely scanty
(Flerov, et ah, 1974). Cygnavus formosus was de¬
scribed from the Aksiirskaya site in East Kazakh¬

stan (Kurochkin, 1968b). Kozlova (1960) de¬
scribed two peculiar, almost didactylous species of
Gruiformes, Proergilornis minor and Ergilornis
rapidus, from Ergelyeen Dzo in Mongolia. Wet-
more (1934) referred to a specimen from the same
locality as “Eogrus sp.” In 1970-1971, large scale
paleontological field work was conducted at Er¬
gelyeen Dzo (Ardyn Obo, Ergil Obo) and in other
lower Oligocene localities in East Gobi, Mongolia,
by Mongolian and Soviet paleontologists. Numer¬
ous avian remains were found, especially at Khoer
Dzan, situated 70 km northwest of the railway sta¬
tion at Zamin Ud, at a locality that we called Shu¬
valov’s Hills, after a student from Saratov State
University who first discovered them in 1971. In
all, about 200 avian bones were found at Khoer
Dzan, 150 of which were in good or satisfactory
condition and can be studied in detail. Several
avian bones were also found in the eastern part of
the classical locality at Ergelyeen Dzo.

The fossil material from Khoer Dzan contains
specimens referable to the Accipitridae, Anatidae
(including Cygninae), Rallidae, Ardeidae, and the
suborder Lari. Diurnal birds of prey are repre¬
sented by several rather large femora resembling
the recent Aegypiinae but differing from Gyps ,
Aegypius , and Gypaetus in a number of important
characteristics. A tarsometatarsus of a smaller spe¬
cies of Accipitridae was also found. Ducks, swans,
rails, and herons from this locality are fairly com¬
parable to recent families, although a fragment of
skull belonging to a large representative of the
suborder Lari has no exact parallels among the
Recent families of that group.

Most of the avian bones from Khoer Dzan be¬
long to the two gruiform families Eogruidae and
Ergilornithidae. Remains of both groups were also
found at Ergelyeen Dzo, and the more diagnostic
material of the former confirms Wetmore’s (1934)
assertion that Eogrus may be traced to deposits of
lower Oligocene age. These two families are dis¬
cussed in greater detail in a following section.

The Middle Oligocene

Information on birds from the middle Oligocene
of Asia is relatively rich. In Central Kazakhstan
avian remains were found in the classical Indri-
cotherium localities in the vicinity of the solon-
chak Tchelkar Tengiz. Agnopterus turgaiensis

NUMBER 27

79

was described from this region by Tugarinov
(1940), and Cygnopterus lambrechti, Somateria
sp., Limicorallus saiensis, and Megagallinula har-
undinea were later described by the present author
(Kurochkin, 1968b). Much new material was also
collected in this region in 1968. Among the fossils
from the sites at Kur Sai and Min Sai in Tchelkar
Nura, north of Tchelkar Tengiz, were remains
identified as belonging to the Pelecaniformes,
Anatidae (including Cygninae), Accipitridae, Gal-
liformes, Gruidae, Rallidae, and Otididae. Collec¬
tions made at Donguz Tau, another locality situ¬
ated on the northeastern part of Tchelkar Tengiz,
included avian remains referable to the Gavii-
formes, Podicipediformes, Ciconiiformes (includ¬
ing the Threskiornithidae), Anatidae (including
members of the Cygninae and Anserinae), Ralli¬
dae, Otididae, and Charadriiformes.

The fossil site at Kyzyl Kak, on the northern
slope of the basin of the same name in Central
Kazakhstan, south of the town of Dzhezkazgan, was
discovered and excavated in the past few years.
This locality was thought to be middle Oligocene
in age (Klebanova, 1965), but the mammalian
material obtained there in 1968 casts doubt on this
dating and most probably indicates a greater age.
Several scraps of avian bones were reported from
Kyzyl Kak, and were determined as belonging to
the Podicipediformes, Aquilavus sp., Gruidae,
and Otididae.

In East Kazakhstan, 60 km south of Lake Zaisan,
new avian material was obtained from Kusto and
Kusto Kysylkain, the sediments of which were
dated by comparison with the mammalian fauna
found at the middle Oligocene site at Kustovskaya.
Flamingos, ducks, swans, Cygnopterus sp. and
Eogrus sp. were identified from here. In the past
few years several different birds were also reported
from Bulkair, Tchaibulak, Akzhar, and Bobrovaya
Struya, located 90 km north of Lake Zaisan. All
the material was collected by Dr. N. Shevyreva of
the Paleontological Institute of the U.S.S.R. Acad¬
emy of Sciences.

Middle Oligocene sediments are very common
on the northern slope of the Lakes Valley at the
foot of the highest peaks of the Mongolian Altai,
Ikhe Bogdo and Baga Bogdo. The first of these
localities, Hsanda Gol and Loh (sediments of the
Hsanda Gol Formation), were discovered by the
Central Asiatic expedition in the 1920s. Small birds

of prey and rails were described from the avian
material collected in Tatal Gol in 1947 by the
Mongolian paleontological expedition of the
U.S.S.R. Academy of Sciences (Kurochkin, 1968a;
1969). Two of these birds of prey belong to the
new genera Gobihierax and Venerator. The third
species was placed in the recent genus Buteo, the
fossil record of which extends back to the middle
Oligocene of North America. This material was
not included in Mellett’s survey of the fauna of
the Hsanda Gol Formation (Mellett, 1968). New
material from the middle Oligocene sediments of
the Lakes Valley was obtained in 1972. The com¬
plete tarsometatarsus of a small owl, differing
markedly from all recent forms of Strigidae, is the
most interesting of these discoveries. In another re¬
gion of Mongolia, an unidentifiable bird bone was
recovered from the middle Oligocene site at
Shunkht, in Middle Gobi Aimak, east of Manlai
Somon.

The Late Oligocene

Birds in the late Oligocene deposits at Agispe,
on the northwest shore of the Aral Sea, are repre¬
sented so far only by Anas oligocaena and several
other species of Anatidae (Tugarinov, 1940, and
more recent unpublished data). A specimen from
this locality was identified by the present author as
being galliform.

Paleoecology and Paleozoogeography
of Asian Paleogene Birds

We can make some judgment of the life and en¬
vironment of fossil birds by analogy with Recent
birds belonging to allied groups, for in many cases,
such as with cormorants, ducks, or bustards, there
is no reason to suspect that there would have been
significant differences in the habits of the Paleo¬
gene forms. However, in studying Paleogene birds
with no analogs among the recent fauna, we can
only make certain assumptions from their osteology
as to their way of life.

The birds from Naran Bulak and Tsagan Khu-
shu provide but scanty evidence of the environ¬
ment of this region in the Paleocene. The
discovery here of Anseriformes, Charadrii, and
Ardeidae tells us no more than that there were
moist areas present. The distinctive morphology
of these birds as compared to Recent forms, makes

80

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

us otherwise very cautious in evaluating their eco¬
logical preferences. The same is true of the Eocene
birds from Khaichin Ula 2.

Along with Eogrus, the lower Oligocene locali¬
ties at Ergelyeen Dzo and Khoer Dzan in south¬
eastern Mongolia contain the remains of two grui-
form birds, Proergilornis minor and Ergilornis
rapidus. In the Paleogene ergilornithids, and in
their Pliocene descendent, Urmiornis, the inner
toe was vestigial or absent, and the birds had lost
the ability to fly, indicating that they were adapted
for running in open areas. In this respect their
ecological niche was probably perfectly analogous
to that of the modern African ostrich ( Struthio)
and of the Asian ostrich of the Pliocene (Kuroch¬
kin and Lungu, 1970). The presence of ergilorni¬
thids and Eogrus in the faunas of Ergelyeen Dzo
and Khoer Dzan is strong evidence that sparsely
vegetated open spaces were present in this area in
the early Oligocene.

The middle Oligocene avian material from
Tatal Gol in central Mongolia does little to il¬
luminate our knowledge of the ecological condi¬
tions that existed at that time. The three birds of
prey (Accipitridae) found there could have lived
either in forests or in open areas, since Recent
species of raptors can live in extremely varied
environments. Fossil remains of an owl are simi¬
larly uninformative, as this species might well have
lived under any of several ecological conditions.
The rail described from Tatal Gol appears to have
been more terrestrially than aquatically adapted.

We can hypothesize with much more confidence
about the habitat found in the middle Oligocene
of Central Kazakhstan, near the present solonchak
of Tchelkar Tengiz. The specimens from the sites
at Tchelkar Nura and those from Donguz Tau may
be treated together, as these sites are separated by
a mere 30 km and no great difference in their past
contemporaneous environments is likely. It is pos¬
sible, however, that certain taphonomic factors
could have affected the species composition of the
birds collected from these two localities. Water-
birds are numerous and diverse at both sites, par¬
ticularly various anatids and loons. In addition,
birds intimately associated with shoals and with
marshes overgrown with vegetation are also abun¬
dantly represented. These include various rails,
grebes, ciconiiforms, pelecaniforms, flamingos,
shorebirds, and probably cranes. Certain of these,
such as the Limicolae and ciconiiforms, would

have required shorelines as feeding sites, while
others, such as loons, swans, and other anatids,
indicate the former presence of large areas of open
water. Although aquatic species constitute a large
proportion of the birds found at Donguz Tau and
Tchelkar Nura, this does not necessarily reflect the
actual situation that existed, as waterbirds are
more likely to have been preserved than terrestrial
ones. Birds of open dry habitats are also repre¬
sented here, viz. Galliformes and Otididae. There
is much evidence to suggest that in the middle
Oligocene in the area of the northern coast of the
present Tchalkar Tengiz there were vast basins
with stretches of open water, alternating with
shoals and marshes and open drier savannas.

The birds of the Kyzyl Kak locality are not
numerous but are distinctive in their ecological
preferences. The presence of small forms of bus¬
tards and cranes suggests open country, perhaps
with low dense vegetation, while grebes indicate
the presence of some nearby water. The birds from
the middle Oligocene deposits of East Kazakhstan
at Kusto and Kyzylkain are almost all waterbirds,
such as ducks, swans, and flamingos, which cer¬
tainly indicate the existence of aquatic habitats.

The paleornithological data from the early Oli¬
gocene of East Gobi (Ergelyeen Dzo and Khoer
Dzan), and the middle Oligocene of Central
Kazakhstan (Donguz Tau, Tchelkar Nura, and
Kyzyl Kak) and Mongolia in the Lakes Valley
(Tatal Gol) confirms the environmental picture
derived from an analysis of the mammalian fauna
of the same age (Flerov, 1961). On the whole, this
suggests a more arid environment for Mongolia as
compared with Kazakhstan. This is evidenced not
only by the predominance of predatory and crane¬
like birds in Mongolia, but also by the absence of
Anseriformes, which are customarily encountered
in localities of this type in other areas. A variety of
aquatic and marsh birds prevails in the middle
Oligocene of Kazakhstan, whereas terrestrial birds
are almost lacking, the few bustard-like birds prob¬
ably being associated with watering places.

It is appropriate here to consider the possible
role of the Turgai area in the evolution of Pale-
arctic waterbirds in the latter half of the Paleo¬
gene. Notwithstanding the general paucity of
material, the discoveries made so far seem to sug¬
gest that in the late Paleogene a number of groups
of waterfowl may have originated on the coasts of
the Turgai strait and later spread from there to

NUMBER 27

81

other areas. Further studies are needed to confirm
or refute this surmise.

For a number of reasons it is difficult to deter¬
mine the relationships of the Paleogene birds of
Kazakhstan and Mongolia to those species from
other areas. In the first place, the material thus
far described and studied in detail is limited. As
can be seen from the preceding discussion, most of
it was obtained very recently and so far has been
studied but tentatively. The character of much of
the paleornithological material also impedes com¬
parisons, since fossil forms of the same systematic
group from different geographical areas may be
represented by different parts of the skeleton and
cannot be compared directly. Generally, the Paleo¬
gene birds of Western Europe and North America
are incomparably better known than those of Asia.
We can cite here only a few examples of Asian
Paleogene birds with apparent relatives from other
areas.

Eogrus, from the upper Eocene and lower
Oligocene of Asia, appears to bear some resem¬
blance to the genus Palaeogrus from the middle
Eocene to early Miocene of Europe. Cygnavus
formosus from the lower Oligocene of eastern
Kazakhstan is comparable to C. senckenbergi
Lambrecht from the lower Miocene of West Ger¬
many. The middle Oligocene forms of Cygno-
pterus from central and eastern Kazakhstan are
close to Cygnopterus affinis (Van Beneden) from
the middle Oligocene of Belgium. The genus
Agnopterus, known from the middle Oligocene of
Central Kazakhstan (Tugarinov, 1940) is repre¬
sented in the upper Eocene of England and France
by two other species. Aquilavus, represented by
new material from the middle Oligocene of
Kazakhstan, is widely represented by several spe¬
cies in Western Europe that range from upper
Eocene to lower Miocene. An undescribed middle
Oligocene loon from Central Kazakhstan may
prove to be referable to the genus Colymboides,
the two other species of which are known in the
upper Eocene of England and the lower Miocene
of France.

The Families Eogruidae and
Ergilomithidae in Asia

Of special interest is the evolutionary history of
the gruiform families Eogruidae and Ergilorni-
thidae in the Tertiary of Asia. The collections

Figure 2.—Left tarsometatarsus (PIN 3109-125) of Proergilor-
nis minor Kozlova, 1960; lower Oligocene, Ergelyeen Dzo:
anterior, medial, and distal views. (Reduced.)

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

82

from the early Oligocene site at Khoer Dzan are of
particular importance in this regard.

A complete tarsometatarsus of Proergilornis
minor Kozlova (1960) from Ergelyeen Dzo (Fig¬
ure 2) was used for comparison with 15 fragments
of tarsometatarsi of Eogruidae (PIN 3110-59,
3110-101-113, 3110-172) and 22 fragments of tarso¬
metatarsi of Ergilornithidae (PIN 3110-54, 3110-55,
3110-70-76, 3110-88-100) from Khoer Dzan. These
collections also include 9 distal ends of tibiotarsi of
Eogruidae (PIN 3110-77-85) and 9 of Ergilomi-
thidae (PIN 3110-64-67) (Figure 3). Unfortu¬
nately, no proximal ends of this bone were
recovered.

All of the eogruid hindlimb elements from
Khoer Dzan are readily attributable to the genus
Eogrus, although they exhibit several characters
distinguishing them from Eogrus aeola of the late
Eocene of Inner Mongolia and E. wetmorei from
the Miocene deposits of Tung Gur. It should be
noted that the remains of Eogrus from Khoer
Dzan are easily separable into two size groups.
Quite arbitrarily, we identified several other bones
(phalanges 1 and 2 of pedal digit III and a frag¬
ment of femur) as belonging to Eogrus.

The remains of both Proergilornis minor and
Ergilornis sp. from Khoer Dzan also fall readily
into two groups differing in size. As with Eogrus,
these differences are detectable in the tibiotatarsi,
tarsometatarsi, and phalanges. In the absence of
any structural differences between the correspond¬
ing bones of the two size groups, we have every
reason to believe that there were significant sexual
differences in size in the species of Ergilornithidae.

In examining numerous limb bones belonging
to Ergilornithidae and Eogruidae we experienced
considerable difficulty in distinguishing specimens
of one family from those of the other. It seems de¬
sirable therefore to detail here the structural dif¬
ferences in the distal end of the tibiotarsus and in
the tarsometatarsus that separate the two families
(Table 1). In making our comparisons we used
Kozlova s original specimens of Proergilornis and
Etgtlo'inis in the collection of the Paleontological
Institute of the U.S.S.R. Academy of Sciences, casts
of the type-material of Eogrus aeola Wetmore
(1934) and E. wetmorei Brodkorb (1967), ob¬
tained from the American Museum of Natural
History, and the referred material of Eogrus aeola
from Irdin Manha, also in the collection of the
Paleontological Institute.

Figure 3.—Distal portion of right tibiotarsus (PIN 3110-64)
of Ergilornithidae genus indet.; lower Oligocene, Khoer
Dzan: anterior, medial, and distal -views. (Natural size.)

NUMBER 27

83

Table 1.—Comparisons of the hindlimb of the Eocruidae and Ergilornithidae

Eogruidae

Ergilornithidae

DISTAL END OF TIBIOTARSUS

Posterior edge of internal condyle on the
same level with anterior edge

Inner tuberculum fixatum closer to outer
edge

Anterior part of external condyle flattened
laterally

Internal ligamental prominence small

Distal end narrow relative to shaft

Posterior edge of internal condyle slopes
distally

Inner tuberculum fixatum on mid-line

Anterior part of external condyle with lat¬
eral swelling

Internal ligamental prominence large

Distal end wide relative to shaft

PROXIMAL END OF

TARSOMETATARSUS

External cotyla round

Hypotarsus with broad plantar surface and
three tendinal grooves

External cotyla oval

Hypotarsus with narrow plantar surface and
no traces of tendinal grooves

DISTAL END OF TARSOMETATARSUS

Middle trochlea relatively large, with parallel
lateral surfaces

Outer trochlea closer to middle trochlea

In posterior view, groove between inner and
outer trochleae located more proximally
in relation to groove between middle and
outer trochleae

Middle trochlea relatively small with lateral
surfaces widening distally

Outer trochlea set farther laterally

Groove between inner and middle trochleae
on same level or ending farther distally in
relation to groove between middle and
outer trochleae

In the development of the plantar crest on the
shaft of the tarsometatarsus, the specimens of
Eogrus from Khoer Dzan are intermediate be¬
tween Eogrus aeola and the Ergilornithidae, thus
tending to bridge the differences between the two
groups.

Apart from tibiotarsi and tarsometatarsi, the col¬
lections from Khoer Dzan include 28 pedal
phalanges belonging to the Ergilornithidae, al¬
though it has not been possible to assign these
phalanges to a particular genus. These phalanges
are short and flat and are virtually identical to
those found in association with a tarsometatarsus
of the Pliocene genus Urmiornis. Figure 4 shows
the basal phalanges of digit III as reconstructed
from elements belonging to different individuals.
The phalanges of digit IV are also present and are
easily distinguishable by their shorter length,
while certain others we have rather arbitrarily
identified as those of digit II. It appears that digit
I had been completely lost in the Ergilornithidae.

Because the ergilornithids were large and had

massive hindlimbs with reduced toes and flattened
phalanges well adapted for running, it was natural
to suggest that they might have been flightless.
This supposition was borne out when a proximal
end of humerus (PIN 3110-60) of a bird in which
the locomotor function of the wing had obviously
been lost (Figure 5) was found at Khoer Dzan.
This specimen represents the second fragment of

Ficure 4.—Basal phalanges of pedal digit III (PIN 3110-61,
62, 63) of Ergilornithidae genus indet.; lower Oligocene,
Khoer Dzan: dorsal and lateral views. (Natural size.)

84

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Figure 5.—Proximal end of right humerus (PIN 3110-60) of
Ergilornithidae genus indet.; lower Oligocene, Khoer Dzan:
palmar, anconal, and proximal views. (Natural size.)

its kind, the first, which was found a day earlier,
being lost soon after its discovery. This humerus
has been so modified through structural degenera¬
tion that there is little basis for comparing it with
Recent Gruiformes. Nevertheless, we suggest that
this fragment is referrable to the Ergilornithidae.
It is characterized by general dorsoventral flatten¬
ing, a markedly reduced and distally sloping head,
the complete absence of a deltoid crest and liga-
mental furrow, and a weakly developed bicipital
surface.

We can compare this humeral fragment with
the humerus of Strigops habroptilus (Psittacidae),
a completely unrelated bird whose ability to fly has
been completely lost. Parallelism in the process of
reduction is quite evident in both. The humerus
of Strigops is similarly characterized by the elon¬
gated and flattened proximal end and reduced
head and Iigamental groove. In the ergilornithid
humerus from Khoer Dzan the process of reduc¬
tion was much farther advanced, however.

The collection from Khoer Dzan also contains a
phalanx 1 of digit I of the wing (PIN 3110-163)
which is characterized by its swollen structure and
its shortened and rounded distal apex. We suggest
that it also belongs to the Ergilornithidae.

The question of the nature of the relationship
between Proergilornis and Ergilornis naturally
arises. The present author is inclined to share
Kozlova’s (I960) assertion that there is consider¬
able reason to regard these two genera as inde¬
pendent. The complete tarsometatarsus from

Ergelyeen Dzo (PIN 3109-125) belongs to Pro¬
ergilornis minor (Figure 2). The tarsometatarsi
from Khoer Dzan also belong to this species, the
only exception being PIN 3110-55, which belongs
to Ergilornis rapidus. The fossils were obtained
from sediments belonging to two different levels
formed at different periods of time, though their
accumulation took place within the framework of
one and the same sedimentary cycle of the early
Oligocene. Presently these two levels are separated
by a section 10-15 m thick. In Shuvalov’s Hills,
most of the fossils were collected from the lower
level, which was composed of a white mass of
oligomict sand and day. Only the tarsometatarsi of
Proergilornis minor were represented in this layer.
The upper level, an ocherous mass of sands and
gravels of polymict composition, contained only
five avian bones: a tarsometatarsus of Ergilornis
rapidus, a tarsometatarsus of Proergilornis minor,
two phalanges referable to the Ergilornithidae and
Eogruidae, and a gruiform cervical vertebra. No
definite inferences could be made from this ma¬
terial. Nevertheless, it ought to be kept in mind
that the numerous bones from the lower level rep¬
resented a single genus only, whereas the few speci¬
mens from the upper level represent two genera.

The question of the possible ancestor of the
Ergilornithidae now arises. Our comparative stud¬
ies suggest that this ancestor may be found in the
Eocene forms of the Eogruidae. The structure of
the hindlimb elements of the early Oligocene rep¬
resentatives of the two groups exhibit a rather close
resemblance, both in overall appearance and in
details. If we take into consideration that in the
Eocene the Eogruidae exhibit a tendency toward
reduction of the inner toe (first noted by Wetmore,
1934), as well as toward development of a sagittal
crest on the tarsometatarsus and a medial fixing
nodule on the distal end of tibiotarsus, we may
suggest that the Eocene Eogruidae were the origi¬
nal members of the lineage leading to the Ergil-
ornithidae. At a later stage these two families were
developing in parallel, as is evident from their
mutual occurrence in the lower Oligocene of
Khoer Dzan, as well as from the presence of
Eogrus wetmorei in the Miocene at Tung Gur
(Brodkorb, 1967).

The appearance of the flightless didactylous
Ergilornithidae in early Oligocene time correlates
well with the general trends observed in the to-

NUMBER 27

85

pography, climate, and vertebrate fauna in the
transition from the late Eocene to the early Oligo-
cene (Flerov, et ah, 1974). This period was char¬
acterized by a complete disappearance of contacts
between the faunas of Asia and America, an inten¬
sive development of open landscapes, and the
advent of various groups of Artiodactyla and
Perissodactyla that later underwent further devel¬
opment on the vast arid open areas of Asia, Eu¬
rope, and Africa. Under such circumstances the
appearance of the flightless cursorial Ergilorni-
thidae may be considered as a natural product of
gruiform evolution.

The Neogene descendant of the Paleogene
ergilornithids is Urmiornis, which was described
for the first time from a lower Pliocene site at
Maragha, Iran (Mecquenem, 1925). Later, re¬
mains of representatives of this genus were found
in the U.S.S.R. in the southern part of the Ukraine
and Moldavia (Burchak-Abramovich, 1951). Addi¬
tional specimens have been reported from other
localities in the Ukraine and Caucasus and also
from middle Pliocene deposits (Bendukidze, 1972;
Umanskaja, 1973). The collections of the Palento-

logical Institute of the U.S.S.R. Academy of
Science also contain representative material of
Urmiornis from the middle Pliocene at Kalmakpai
in the Zaisan basin, East Kazakhstan, and several
fragments from the middle Pliocene locality at
Khirgis Nur in the Great Lakes Valley, Mongolia.
Evidently, Urmiornis was very common in both the
lower and middle Pliocene, throughout the entire
range of the Hipparion Fauna. Comparisons of the
original material of the Ergilornithidae with
Urmiornis leaves no doubt as to the affinity be¬
tween these didactylous gruiform birds, as the
structure of their limb bones is very similar. Brod-
korb (1967) was quite correct in placing Urmiornis
in the family Ergilornithidae. A direct phylo¬
genetic link between the Oligocene ergilornithids
and the Pliocene Urmiornis can no longer be
doubted.

Thus, a distinct phylogenetic line of Asian grui-
forms that survived at least up to the middle
Pliocene may be traced through the Tertiary as
follows; Eogrus (Late Eocene), Ergilornis and
Proergilornis (early Oligocene), Urmiornis (lower
and middle Pliocene).

Literature Cited

Aubekerova, P. A.

1965. Novye dannye ob iskopaemykh ptitsach Kazakh¬
stan. In Novosti ornitologii, Alma-Ata, 18-19. [In
Russian.]

Belaeva, E. I., B. A. Trofimov, and V. J. Reshetov

1974. Osnovnye etapy evolutsii mlekopitajutshikh v
pozdnem mezozoe-paleogene Tsentralnoj Asii. In
Sovmestnaja Sovetsko-Mongolskaja paleontologiches-
kaja ekspeditsija, trudy, 1:19—45 [In Russian.]

Bendukidze, O. G.

1971. Novyj predstavitel’ semeistva Geranoididae (Aves,
Gruiformes) iz eotsenovykh otlozhenij Zaisan.
Soobtzhenija Akademii nauk Gruzinskoj SSR, 63
(3):749—751, 1 figure. [In Russian.]

1972. O nakhodke Urmiornis maraghanus Mecq. na
Severnom Kavkaze. Soobtzhenija Akademii nauk
Gruzinskoj SSR, 66(3):741-743, 1 figure. [In Russian.]

Brodkorb, P.

1967. Catalogue of Fossil Birds, Part 3 (Ralliformes,
Ichthyornithiformes, Charadriiformes). Bulletin of
the Florida State Museum, Biological Sciences,
11 (3): 99-220.

Burchak-Abramovich, N. I.

1951. Urmiornis (Urmiornis maraghanus Mecq.) strauso-
podobnaja ptitsa gipparionovoj fauny Zakavkasja i

Juzhonoj Ukrainy. Izvestija Akademii nauk Azer-
bajdzhanskoj SSR, 6:83-94, 1 figure. [In Russian.]

1958. Tretichnye ptitsi SSSR. Uchenye zapiski Azerbajd-
zlianskogo Gosudarstvennogo Universiteta imeni
Kirova, Serija Biologija, 1:81-88, 3 figures. [In Rus¬
sian.]

Efremov, I. A.

1954. Paleontologischeskie issledovanija v Mongolskoj
Narodonoj Respublike (predvaritelnye resultaty
ekspeditsii 1946, 1948 i 1949 gg.). Trudy Mongolskoj
komissii, 50:3-32. [In Russian.]

Flerov, K. K.

1961, K voprosu o biologicheskoj i paleozoogeografiches-
koj kharakteristike indricoterievoi fauny. Paleon-
tologicheskij zhurnal, 1:12-22. [In Russian.]

Flerov, K. K., E. I. Belaeva, N. M. Janovskaja, A. A. Gureev,
I. M. Novodvorskaja, V. S. Kornilova, N. S. Schevyreva,
E. N. Kurochkin, V. V. Zherichin, V. M. Tchikvadze,
G. G. Martinson, N. V. Tolstikova, A. L. Chepalyga,
and L. I. Fotjanova

1974. Zoogeografija paleogena Asii. Volume 146 in Trudy
paleontologicheskogo instituta Akademii Nauk
SSSR. 302 pages, 162 figures, 4 plates. Moscow:
Izdatel’stvo Nauka. [In Russian.]

86

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Gradzirisky, R., J. Kazmierczak, and J. Lefeld

1968. Geological Data from the Polish-Mongolian Palae¬
ontological Expeditions. Palaeontologica Polonica,
19:33-82.

Gureev, A. A.

1949. Pervaja nakhodka ptitsy v nizhnem paleogene
SSSR. Doklady Akademii nauk SSSR, 64(2):249-251,
1 figure. [In Russian.]

Hekker, R. F„ A. I. Osipova, and T. N. Belskaja

1962. Ferganskij zaliv paleogenovogo morja Srednej Asii,
Kniga I. 332 pages, 89 figures, 22 plates. Moscow:
Izdatel’stovo Akademii nauk SSSR. [In Russian.]
Klebanova, I. M.

1965. Novoe mestonakhozhdenie sredneoligotsenovikh
mlekopitajutshchikh v urochishche Kyzyl-Kak.
Paleontologicheskij Zhurnal, 4:99-102. [In Russian.]
Kozlova, E. V.

1960. Novye iskopaemye ptitsy iz jugo-vostochnoj Gobi.
Trudy Problemnykh i Tematicheskikh Soveshchanii
Akademija Nauk SSSR, Zoologicheskij Institut 9:
323-329. [In Russian.]

Kurochkin, E. N.

1968a. Iskopaemye ostatki ptits iz Mongolii. Ornitologija,
9:323-330, 4 figures. [In Russian.]

1968b. Novye oligotsenovye ptitsi Kazakhstana. Paleon-
tologicheskij Zhurnal, 1:92-101, 5 figures. [In Rus¬
sian].

1969. Novoe nazvanie dlja roda Tutor Kurochkin, 1968.
Paleontologicheskij Zhurnal, 2:122. [In Russian.]

1974. Obzor paleogenovykh ptits Azii. Ornitologija, 11:
317-328, 1 figure. [In Russian],

Kurochkin, E. N., and A. N. Lungu

1970. Novyj straus iz srednego sarmata Moldavii. Paleon¬
tologicheskij Zhurnal, 1:118-126, 2 figures, 1 plate.
[In Russian.]

Lambrecht, K.

1931. Protoplotus beauforti n.g. n.sp., ein Schlangen-
halsvogel aus Tertiar von W.-Sumatra. Dienst van
den Mijnbouw in Nederlandsch-lndie, Wetenschap-
pelijke Mededeelingen, 17:15-24, 3 plates.
Mecquenem, R.

1925. Contribution a l’^tude des fossiles de Maragha
[continuation from 1924]. Annales de Paleontolo-
gie, 15(1): 26—36.

Mellett, J. S.

1968. The Oligocene Hsanda Gol Formation, Mongolia:
A Revised Faunal List. American Museum Novi-
tates, 2318:1-16.

Novozhilov, N. I.

1954. Mestonakhozhdenie mlekopitajuschchikh nizhnego
eotsena i verkhnego Paleotsena Mongolii. Trudy
Mongol’skoj komissii, 59:33^16. [In Russian.]

Olson, S. L.

1974. Telecrex Restudied: A Small Eocene Guineafowl.
Wilson Bulletin, 86(3):246-250, 3 figures.

Shishkin, M. A.

1975. Stratigrafija i tafonomija verkhnepaleotsenovogo
mestonakhozhdenija pozvonochnikh Naran-Bulak
(Juzhnaja Gobi, MNR). Sovmestnaja Sovetsko-
Mongolskaja paleontologicheskaja ekspeditsija,
trudy, 2:225-244. [In Russian.]

Shuvalov, V. F., V. J. Reshetov, and R. Barsbold

1974. O stratotipicheskom razreze nizhnego paleogena na
jugo-zapade MNR. Sovmestnaja Sovetsko-Mongol-
skaja paleontologicheskaja ekspeditsija, 1:60-71. [In
Russian.]

Szalay, F. S., and M. C. McKenna

1971. Beginning of the Age of Mammals in Asia: The
Late Paleocene Gashato Fauna, Mongolia. Bulletin
of the American Museum of Natural History,
144(4):269—318.

Teilhard de Chardin, P.

1926. Description des Mammiferes Tertiares de Chine et
de Mongolie. Annales de Paleontologie, 15:1-52.

Tugarinov, A. J.

1940. Novye dannye dlja tretichnoj ornitofauny SSSR.
Doklady Akademii nauk SSSR, 23(3):314—316, 2
figures. [In Russian].

Umanskaja, A. S.

1973. Novaja nakhodka ostatkov urmiornisa iz neogeno-
vikh otlozhenij Ukrainy. Vestnik Zoologii, 4:30-33,
1 figure. [In Russian.]

Wetmore, A.

1934. Fossil Birds from Mongolia and China. American
Museum Novitates, 711:1-16, 6 figures.

The World’s Oldest Owl: A New Strigiform
from the Paleocene of Southwestern Colorado

Pat Vickers Rich and David J. Bohaska

ABSTRACT

Among the fossils recovered from a small, mid-
Paleocene fissure filling in southwestern Colorado
is the oldest known owl, Ogygoptynx wetmorei,
new genus and species. This form, represented by
a single tarsometatarsus, does not clearly belong
in any of the known families of Strigiformes and
may represent a new higher category of owls that
provides a link between the Strigidae and the
Tytonidae.

Introduction

In 1916, Walter Granger, following up the ex¬
ploratory work of J. W. Gidley, examined sediments
around Tiffany in southwestern Colorado and lo¬
cated several fossil-bearing areas that produced a
moderately diverse mammalian fauna (Simpson,
1935). This assemblage became important in the
definition of a North American land mammal age,
the Tiffanian (Wood, et al., 1941). Among the
bones found by Granger was a small avian tarsome¬
tatarsus. The following paper describes this fossil
and considers its relationships.

Acknowledgments. —Thanks are due Mr. Earl
Manning, who first recognized the avian affinities
of Ogygoptynx, to Dr. Malcolm C. McKenna,
American Museum of Natural History (AMNH),
for allowing us to borrow the specimen for study,
and Ms. Charlotte Holton for arranging the loan.
Comparative material of the living genus Phodilus
was kindly furnished by Dr. Ned K. Johnson, Mu-

Pat Vickers Rich and David J. Bohaska, Geosciences Depart¬
ment, Texas Tech University, Lubbock, Texas 79409.

seum of Vertebrate Zoology, University of Cali¬
fornia, Berkeley. The photographs and chart were
prepared by Lloyd Logan and Robert Suddartha.
Sincere thanks are also due Dr. Don E. Russell,
Institut de Pal^ontologie, Museum National d’His-
toire Naturelle, Paris (PM) for lending the Eu¬
ropean Paleogene owls briefly surveyed in this
paper and presently under more extensive study.
The manuscript was read by Hildegarde Howard
and typed by Carol Baldwin.

Order STRIGIFORMES
Family incertae sedis
Ogygoptynx, new genus

Type-Species. —Ogygoptynx wetmorei, new spe¬
cies.

Distribution. —Known only from the Tiffanian
(mid-Paleocene) of Colorado.

Diagnosis. —See Table 1.

Characters differing from the Strigidae but simi¬
lar to the Tytonidae (including the Phodilinae):
tarsometatarsus lacking a supratendinal bridge on
the proximal end (three strigids also lack this
bridge; Ford, 1967); posterior metatarsal groove
not separated from the proximal articular surface
by a marked bony ledge; anterior metatarsal groove
deep across the entire width of the shaft at the
proximal end.

Characters differing from the Phodilinae and
Strigidae but resembling the Tytoninae: tarso¬
metatarsus elongate and slender; posterior meta¬
tarsal groove only slightly excavated, lacking a
lateral wall near the proximal end.

Characters resembling the Phodilinae and inter-

87

88

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Table 1.—Comparison of the tarsometatarsus in Ogygoptynx with several families and subfamilies of owls

Character

Protostrigidae

Strigidae

Tytoninae

Phodilinae,
Phodilus badius

Ogygoptynx

wetmorei

Proximal End

Intercotylar

-

Set back from

Occupies most

Occupies most

Occupies anterior

prominence

anterior margin
of proximal end,
located midway
between anterior
and posterior
margins

anterior part
of proximal end;
protrudes
anterior to
cotyla

anterior part of
proximal end;
protrudes
anterior to
cotyla

margin of
proximal end;
not protruding
anterior to
cotyla

Anteroposterior

-

Narrow

Deep

Deep

Narrow

width across
intercotylar area

Groove lateral
to intercotylar
prominence

Distinct

Indistinct

Indicated slightly

Indicated

moderately,

intermediate

between

Strigidae and
Phodilinae

Supratendinal

bridge

Ossified in all but

3 species (see
Ford, 1967)

Not ossified

Not ossified

Not ossified

Width of

-

Wide

Narrow

Narrow

Wide?

calcaneal canal

Orientation of
external cotyla

—

Directed laterally

Directed

posteriorly

Directed

posteriorly

Directed laterally

Degree of

-

Highly excavated

Slightly excavated

Highly excavated

Slightly excavated

excavation of
posterior
metatarsal groove

Degree of separation

-

Bony ledge sepa-

No bony ledge

No bony ledge

No bony ledge

of posterior
metatarsal groove
and proximal

rating groove
from intercotylar
area

present

present

present

articular surface

Depth of anterior

-

Deep only at inner

Deep across full

Deep across full

Deep across full

metatarsal groove

part of supra-

width of

width of

width of

near proximal end
of tarsometatarsus

tendinal bridge

proximal end,
undercuts
cotylar area

proximal end,
undercuts
cotylar area

proximal end,
undercuts
cotylar area

Lateral wall of

-

Present

Absent

Present

Absent

posterior metatarsal
groove

Distal End

Relative distal

Middle trochlea

Middle trochlea

Inner trochlea

Middle trochlea

Inner trochlea

extension of

extends farther

extends farther

extends slightly

extends slightly

extends much

trochleae

distal than inner

distal than inner

more distal than

middle

farther distal

than inner

farther distal

than middle

Groove on middle

Anterior part not

Anterior part

Anterior part not

Anterior part with

Anterior part not

trochlea

noticeably
grooved, blends
smoothly into
shaft

noticeably

grooved

noticeably
grooved, blends
smoothly into
shaft

shallow groove

noticeably
grooved, blends
smoothly into
shaft

NUMBER 27

89

Table 1. —Continued

Character

Protostrigidae

Strigidae

Tytoninae

Phodilinaf.,
Phodilus badius

Ogygoptynx

wetmorei

Shape of outer
trochlea in
lateral view

Distal margin
highly rounded

Distal margin
flattened or
slightly grooved

Distal margin
strongly grooved

Distal margin
strongly grooved

Distal margin
moderately
rounded

Curvature across
trochleae in
distal view

Highly curved;
channel created
by such

curvature deep;
outer trochlea
only slightly
curved (medially
concave)

Moderately
curved; channel
created by such
curvature

shallow; outer
trochlea moder¬
ately curved

Highly curved;
channel created
by such

curvature deep;
outer trochlea
highly curved

Moderately
curved; channel
created by such
curvature deep;
outer trochlea
highly curved

Highly curved;
channel created
by such

curvature deep;
outer trochlea
only slightly
curved

General

Proportions of
tarsometarsus

—

Variable, generally
short, stout

Elongate, slender

Short, stout

Elongate, slender

mediate between the Strigidae and Tytoninae:
shallow groove slightly lateral to the intercotylar
prominence.

Characters differing from the Tytonidae but re¬
sembling the Strigidae: intercotylar area shallow
anteroposteriorly, due in part to the intercotylar
prominence not protruding anterior to cotylar mar¬
gins; distal margin of outer trochlea in lateral view
flattened and only slightly rounded, not grooved;
outer trochlea in lateral view only slightly curved.

Characters differing from the Protostrigidae:
distal margin of outer trochlea in lateral view
flattened and only slightly, rather than greatly,
rounded.

Characters within the Strigiformes unique to
Ogygoptynx: proximal end in lateral view shaped
like a parallelogram with unequal angles, not rec¬
tangular; outer trochlea in distal view not smoothly
rounded but slightly grooved laterally, not tapering
to a point but broadened posteriorly; anterior por¬
tion of shaft just proximal to middle trochlea ele¬
vated farther anteriorly than the remainder of the
shaft, but distal portion of shaft medial to this re¬
gion markedly planar and not rounded; inner
trochlea decidedly more elongate than middle
trochlea.

Etymology. —From Greek, Ogyges, mythical king
of Thebes, suggesting ancient or primeval, and
ptynx, an owl.

Ogygoptynx wetmorei, new species
Figures 1, 2

Holotype. —Nearly complete right tarsometa-
tarsus, AMNH 2653, missing only a small part of
the calcaneal ridge and possibly some of the
midshaft.

Type-Locality and Stratigraphic Position.—
Mason Pocket near old Mason schoolhouse, north¬
ern drainage slope of the San Juan River, 6.5 to 8
km north of Tiffany, Section 20, T33N, R6W, La
Plata County, southwestern Colorado. Found in
“a small pocket of gray shale imbedded in a stratum
of mottled purplish and brownish clay,” which has
been thought by some workers to be a fissure filling
of some type. “Tiffany Beds,” Tiffanian, mid-
Paleocene in age (Simpson, 1935).

Diagnosis. —As for the genus.

Measurements. —Maximum width of proximal
end, 9.0 mm; maximum depth of external cotyla,
4.9; maximum depth of internal cotyla, 5.1; maxi¬
mum length from intercotylar prominence to distal
end of attachment of tibialis anticus, 13.8; maxi¬
mum width of distal end, 9.4; maximum width of
inner trochlea, 4.2; maximum width of middle
trochlea, 3.7; depth of internal border of inner
trochlea, 3.2; depth of external border of inner
trochlea, 3.6; depth of internal border of middle
trochlea, 3.5; depth of external border of middle

NUMBER 27

91

Figure 2 .—Ogygoptynx wetmorei, new genus and species,
stereo pairs of holotype right tarsometatarsus (AMNH 2653):
a, proximal view; b, distal view. (X 3.)

trochlea, 3.9; depth of external border of outer
trochlea, 4.0; overall length from intercotylar
prominence to distal end of inner trochlea, at
least 48.2.

Etymology. —Named in honor of Dr. Alexander
Wetmore, who has done much to further the
knowledge of fossil owls as well as other birds.

Comparison with Modern Strigiformes

Ogygoptynx, known only from a tarsometa¬
tarsus, clearly belongs in the order Strigiformes as
indicated by the broadly and deeply excavated ante¬
rior metatarsal groove; the single, slender cal¬
caneal ridge; the shallow, narrow middle trochlea
relative to the inner and outer trochleae; and the
inner trochlea extending distally nearly as far as,
or farther than, the middle trochlea.

Within the order, Ogygoptynx appears to be
intermediate between the Tytonidae (including
the Phodilinae; Ford, 1967) and the Strigidae,
since it possesses a mosaic of character states of
each of these families (Table 1).

One modern owl, Phodilus (subfamily Phodi¬
linae), presently allocated to the Tytonidae (Ford,
1967), is to some degree a strigid-tytonid mosaic,

Figure 1 . — Ogygoptynx wetmorei, new genus and species,
holotype right tarsometatarsus (AMNH 2653): a, anterior
view (stereo pair); b, posterior view (stereo pair); c, medial
view; d, lateral view. (X 2.)

but unlike Ogygoptynx, it is clearly most closely
related to the Tytonidae. Strigid features seen in
Phodilus include the deep proximal excavation of
the posterior metatarsal groove, development of a
lateral wall on the metatarsal groove near the
proximal end, the moderately arched distal end,
and the overall short, stout configuration of the
tarsometatarsus. In most of its characters, how¬
ever, Phodilus resembles the Tytonidae (Table 1),
and thus, at least as evidenced by the tarsometa¬
tarsus, the allocation of Phodilus to the Tytonidae
appears to be correct.

Because the tarsometatarsus of Ogygoptynx has
such a mingling of strigid and tytonid character
states, it cannot clearly be referred to any of the
higher categories of modern Strigiformes.

Comparison with Fossil Strigiformes

The mid-Paleocene Ogygoptynx is the earliest
known owl (Figure 3) and it is certain that by the
early Eocene a number of other strigiforms had
made their appearance in North America. Proto-
strix and Eostrix, in the family Protostrigidae, are
known from a number of localities in the Ameri¬
can West (Wetmore, 1938; Brodkorb, 1971; Martin
and Black, 1972; Rich, unpublished data). The
distal end of the tarsometatarsus is known for both
genera and thus they may be compared directly
with Ogygoptynx. Unfortunately, the proximal end
of the tarsometatarsus is not known in any of the
protostrigids. The smallest genus, Eostrix, closely
approximates Ogygoptynx in size, but along with
Protostrix it differs in the relative distal extension
of the inner and middle trochleae, in the shape of
the outer trochlea, and in lacking those character
states unique to Ogygoptyrix (see “Diagnosis”) It
is difficult to assess the value of many of these
unique character states clue to the partial crushing
of the type of O. wetmorei, as well as to our in¬
complete knowledge of the tarsometatarsus in the
Protostrigidae.

A number of other fossil owls, previously classi¬
fied in the Strigidae (Brodkorb, 1971), show a mix¬
ture of strigid and tytonid character states suggest¬
ing that they may bear some relationship to
Phodilus and Ogygoptynx. Brodkorb (1970)
erected the genus Paratyto for Bubo arvernensis
Milne-Edwards, from the lower Miocene of France,
placing it in the “Phodilidae” on the basis of its

92

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

m.y. Stage/Age Epoch

NORTH AMERICA

EUROPE

Epoch Stage/ Age

m.y.

30

35 H

40 M

45 M

50 H

55 M

60 K

65 W

WHITNEYAN/

ORELLAN

CHADRONIAN

DUCHESNEAN

UINTAN

BRIDGERIAN

WASATCHIAN

CLARKFORKIAN

TIFFANIAN

TORRJONIAN

PUERCAN/
DRAGON IAN

u

O

O

u

O

u

O

<

CL

Necrobyas
Strigogyps
“ Otus ” henrici
“ Otus ”, species
“Bubo” incertis

Protostrix

lEostrix

I Ogygoptynx, new genus

Evernden,
e» al., 1964
McKenna,1973

V

O

0

V

0

u

O

UJ

—i

<

a.

STAMPIAN

BARTONIAN

LUTETIAN

YPRESIAN

THANETIAN

k 30

k 35

k 40

N 45

DANO-MONTI AN

M 50

k 55

k 60

k 65

Berggren, 1972
McKenna, 1973

Figure 3. — Temporal distribution of early Tertiary owls.

lacking an ossified supratendinal bridge on the
tarsometatarsus (as in the Tytoninae and Phodi-
linae) and on the stoutness of the tarsometatarsus
(as in the Strigidae and Phodilinae). He did not
mention, however, that three species of strigids also
lack an ossified supratendinal bridge, nor did he
discuss other characters useful in separating the
Strigidae from the Tytonidae (including the Phodi¬
linae). Thus, reexamination of Paratyto would be
useful for determining whether it is a strigid or a
phodiline.

At least four other European Tertiary owls (Fig¬
ure 3), Necrobyas harpax, N. rossignoli, “Otus”
henrici, and “Otus” sp. (PM 3120) (Milne-
Edwards, 1892), are based entirely or in part on
tarsometatarsi that are stout and also lack an
ossified supratendinal bridge. The tarsometatarsus

of Ogygoptynx is narrower than in these owls, with
N. harpax exhibiting proportions most similar to
those of the new Paleocene form. Another tytonid
character seen in the four European owls and
Ogygoptynx is the lack of a ledge proximal to the
posterior metatarsal groove on the tarsometatarsus.
The calcaneal groove is narrow in the four Eu¬
ropean owls, as in the tytonids, but is probably
wide in Ogygoptynx, as in the strigids. In all of the
above four species, the external cotyla points pos¬
teriorly, as in the Tytonidae, except in two speci¬
mens assigned to “Otus” henrici (PM 3117 and
PM 3118) in which the cotyla points laterally as in
the Strigidae and Ogygoptynx. The position of the
intercotylar prominence in all four species is inter¬
mediate between that in the tytonids and strigids
and very similar to that in Ogygoptynx. A fifth

NUMBER 27

93

owl, Necrobyas edwardsi Gaillard (1908, 1939),
from the Tertiary of France, also based on a tarso-
metatarsus, was unavailable for study.

Other possibly intermediate genera of owls in¬
clude Prosybris Brodkorb (1970) and Lechusa
Miller (1956). Prosybris, from the lower Miocene
of France, is placed with the Tytonidae but has a
somewhat stouter tarsometatarsus than Tyto; like
Ogygoptynx, its external trochlea is elongate.
Lechusa stirtoni, based on a coracoid from the
Pliocene of San Diego, California, was classified by
Miller (1956) as a tytonid, although both he and
Wetmore (in Miller, 1956:620) noted several
strigid characters in this form. Miller stated, how¬
ever, that Lechusa was definitely not a phodiline.
Direct comparison between Lechusa and Ogygo¬

ptynx is impossible because of the lack of corre¬
sponding elements.

An extensive review of the above-mentioned fos¬
sil forms and some other European fossil owls is
needed before the interrelationships of these, as
well as modern forms, can be interpreted.

Ogygoptynx, the oldest known owl, does not
clearly belong in any of the presently established
families within the order Strigiformes (Proto-
strigidae, Tytonidae, Strigidae) and may well be
distinct enough to rate separate familial rank.
Until a thorough revision of the early Tertiary
European owls is completed, however, we will re¬
frain from making any such designation and only
point out the mosaic nature of this early owl.

Literature Cited

Berggren, W. A.

1972. A Cenozoic Time Scale—Some Implications for
Regional Geology and Paleobiogeography. Lelhaia,
5:195-215.

Brodkorb, P.

1970. Two Fossil Owls from the Aquitanian of France.
Quarterly Journal of the Florida Academy of
Science, 32(2): 159-160.

1971. Catalogue of Fossil Birds, Part 4 (Columbiformes
through Piciformes). Bulletin of the Florida State
Museum, Biological Sciences, 15(4): 163-266.

Evernden, J. F„ D. E. Savage, G. H. Curtis, and G. T. James

1964. Potassium-Argon Dates and the Cenozoic Mam¬
malian Chronology of North America. American
Journal of Science, 262:145-198.

Ford, N. L.

1967. A Systematic Study of the Owls Based on Compara¬
tive Osteology. Ph.D. dissertation, The University
of Michigan, Ann Arbor. [University Microfilms
68-7595, 128 pages.]

Gaillard, C.

1908. Les oiseaux des phosphorites du Quercy. Annales
de I’Universite de Lyon, new series, 1(23): 1-178, 37
figures, 8 plates.

1939. Contribution a l’^tude des oiseaux fossiles. Archives
du Musee d’Histoire Naturelle de Lyon, 15(Mem-
oire 2):1-100, 34 figures.

Martin, L. D., and C. C. Black

1972. A New Owl from the Eocene of Wyoming. Auk ,
89(4): 887-888, 1 figure.

McKenna, M. C.

1973. K/Ar Recalibration of Eocene North American
Land-Mammal “Ages” and European Ages. Geo¬
logical Society of America Abstracts, 5(7):733.

Miller, L.

1956. A Collection of Bird Remains from the Pliocene
of San Diego, California. Proceedings of the Cali¬
fornia Academy of Sciences, 28(16):615-621, 1 figure.

Milne-Edwards, A.

1892. Sur les oiseaux fossiles des depots Eocenes du phos¬
phate de chaux du sud de la France. Pages 60-80
in volume 2 of 2 me Congres Ornithologique Inter¬
national, Budapest, 1891, Partie Scientifique.

Simpson, G. G.

1935. The Tiffany Fauna, Upper Paleocene, I: Multi-
tuberculata, Marsupialia, Insectivora, and ?Chirop-
tera. American Museum Novitates, 795:1-19, 6
figures.

Wetmore, A.

1938. Another Fossil Owl from the Eocene of Wyoming.
Proceedings of The United States National Mu¬
seum, 85(3031):27-29, 2 figures.

Wood, H. E„ 2nd, R. W. Chaney, J. Clark, E. H. Colbert,
G. L. Jepsen, J. B. Reeside, Jr., and C. Stock

1941. Nomenclature and Correlation of the North Ameri¬
can Continental Tertiary. Bulletin of the Geologi¬
cal Society of America, 52:1-48, 1 chart.

Neanis schucherti Restudied:
Another Eocene Piciform Bird

Alan Feduccia

ABSTRACT

Neanis schucherti, an avian fossil from the lower
Eocene Green River Formation of Wyoming, was
described in 1913 under the genus Hebe by Shu-
feldt as the earliest representative of the order
Passeriformes. The name Neanis has since been
substituted for Hebe for reasons of priority. The
specimen is here restudied and is found to be of
piciform, not passeriform, affinity. It is assigned to
the extinct Eocene family Primobucconidae. Of
the two other previously described lower Eocene
Piciformes, Primobucco kistneri Feduccia 1973 is
reassigned to Neanis, while P. mcgrewi Brodkorb
1970 is retained in Primobucco, which genus is
recognized by its larger size.

Introduction

The oldest North American avian fossil referred
to the order Passeriformes and the only one as¬
signed to the Neotropical family Rhinocryptidae
(tapaculos) is Neanis schucherti (Shufeldt, 1913)
from the lower Eocene Green River Formation of
Wyoming. This fossil was first described under the
generic name Hebe. Brodkorb (1965) substituted
Neanis for Shufeldt’s genus, which was preoccupied
by Hebe Risso 1826 (Crustacea). The type of
Neanis schucherti consists of a small slab and
counterslab (YPM 1233) containing impressions
of bone and feathers and some poorly preserved
pieces of bone (Figures 1, 2). These were collected
in 1874 by F. A. C. Richardson of the Powell Ex-

Alan Feduccia, Department of Zoology, University of North
Carolina, Chapel Hill, North Carolina 27514.

pedition. The original label reads: “Found five
miles west of Green River City, Wyoming. In the
fish cut of the R.R. . . . Associated with insects de¬
scribed by Scudder.’’ On the back of the label is the
following inscription: “Compare Pteroptochidae;
see Ibis, 1874, p. 191 (July), for sternum with 2
emarginations in sternum.” The citation is to
Sclater’s (1874) paper on the tapaculos and it was
no doubt the above inscription that led Shufeldt to
place the fossil in the family Pteroptochidae
(= Rhinocryptidae). He based this conclusion al¬
most entirely on the presence of a four-notched
sternum (i.e., with “2 emarginations” on each
side), and on an alleged “large manubrium which
is bifurcated anteriorly” (Shufeldt, 1913:647). I
have recently had the opportunity to examine the
type of Neanis schucherti, and I present here my
conclusion that the affinities of Neanis are with the
Piciformes rather than the Passeriformes.

Acknowledgments. —The loan of the type of
Neanis schucherti was arranged through the kind¬
ness of Drs. Charles G. Sibley and John H. Ostrom
of the Yale University Peabody Museum of Natu¬
ral History (YPM). Dr. Paul O. McGrew of the
University of Wyoming Department of Geology
kindly lent the types of Primobucco tncgreun and
Primobucco kistneri. Skeletons of modern species
were made available through the courtesy of Dr.
Pierce Brodkorb (collection of Pierce Brodkorb),
Dr. R. L. Storer (University of Michigan Museum
of Zoology), and Dr. R. L. Zusi (National Museum
of Natural History, Smithsonian Institution). I
wish to thank Drs. Pierce Brodkorb, Larry D. Mar¬
tin, and Helmut C. Mueller for their critical re¬
view of the manuscript. This work was supported
by a grant from the University of North Carolina
Research Council.

95

96

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Description and Comparisons

Two facts render the assignment of Neanis to the
Rhinocryptidae questionable or suspect. First, the
characters used in the identification, namely the
presence of a four-notched sternum and a bifurcate
manubrium ( = forked spina externa of the ster¬
num), are unique neither to the Rhinocryptidae
nor the Passeriformes. Second, in recent years sev¬
eral primitive piciform birds have been described
from the lower Eocene Green River Formation of
Wyoming, the same horizon and locality from
which Neanis was recovered. Piciform birds typi¬
cally have a four-notched sternum and a forked
spina externa. The Green River species Primo-
bucco mcgrewi Brodkorb (1970) and Primobucco
kistneri Feduccia (1973)* provide the earliest
records of the order Piciformes. Among the struc¬
turally primitive piciform families, the Bucconidae
has been used in the past to accommodate these
lower Eocene zygodactyl birds, but they are now
considered to merit their own family, the Primo-
bucconidae, which also includes three middle
Eocene genera (Feduccia and Martin, p. 101,
herein).

The presence of a four-notched sternum is an
unreliable taxonomic character even at the ordinal
level. Both two- and four-notched sterna occur in
many orders not related to the passerines and their
allies (e.g., Ciconiiformes and Charadriiformes),
and four-notched sterna are commonly found in
certain coraciiform birds (e.g., rollers and king¬
fishers). All members of the Piciformes possess a
four-notched sternum (Feduccia, 1972), and
within the Passeriformes the four-notched sternum
is found within the Formicariidae and Rhino¬
cryptidae (Heimerdinger and Ames, 1967). I am
able to confirm that Neanis schucherti does indeed
possess a four-notched sternum although to deter¬
mine this requires “very close and careful examina¬
tion with a high-power lens’’ (Shufeldt, 1913:646).
Shufeldt (1913:646) goes on to point out that “the
sternum of [Neanis'] differed in this particular
from all typical existing passerines as they now
occur in North America, at least north of Costa
Rica.”

In birds, the spina externa of the sternum may

* Misspelled " kisterni” in two places in the original ref¬
ence.

be forked (as in most passerines) or a simple rod
(as in most nonpasserines). Olson (1971) has ex¬
amined the spina externa of the sternum and
found it to be highly variable. It is typically forked
(bifurcate) in the Passeriformes (exclusively so
in the suborder Passeres) but within the “subos-
cines” an unforked spina externa occurs within the
Eurylaimidae, Cotingidae, Philepittidae (Ames,
1971), and Dendrocolaptidae (. Xiphocolaptes pro-
meropirhynchus, (personal observation). However,
Smithornis of the Eurylaimidae has a forked spina
externa and within the Cotingidae the character is
intragenerically variable in Procnias (Olson, 1971).
Olson (1971:509) also points out that, “the forked
versus simple spina externa is variable in other
orders as well. In the Piciformes either conforma¬
tion may be found in the Picidae, Capitonidae,
Bucconidae and Galbulidae. The character is also
variable in the Coraciiformes and Trogoniformes.”
Thus, perhaps the only taxonomic use of the na¬
ture of the spina externa would be to exclude forms
with the simple, unforked spina externa from the
suborder Passeres. In any case, I am unable to con¬
firm the presence of a forked spina externa in
Neanis schucherti. I have found the structure Shu¬
feldt evidently intended (Figure 1), but the area is
so crushed that some imagination is necessary to
envision it as a bifurcate spina externa.

The two slabs containing the type of Neanis
schucherti are illustrated in Figures 1 and 2. Fig¬
ure 1 depicts the slab containing the actual bones;
Figure 2 shows the counterslab with bone and
feather impressions. The bones present on the slab
are as follows: sternum (left side: ventral aspect);
both coracoids (left, 10.8 mm); furcula (12.1 mm
from furcular process to scapular tuberosity); left
scapula; left humerus (proximal portion as pre¬
served, 13.3 mm; width of proximal end, 5.4 mm;
least width of shaft, 1.8 mm); right ulna (19.1 mm);
right radius (approximately 18.4 mm); bones of
manus (very faint). Unfortunately, no useful ratios
could be obtained. Furthermore, few useful osteo-
logical characters are exhibited in the bones ex¬
posed, except for the form of the proximal end of
the humerus, which is preserved in palmar view. It
is in the humerus that one finds characters that ally
Neanis with the structurally primitive families
of the Piciformes, such as the modern family
Bucconidae. Neanis conforms with the characters

NUMBER 27

97

Figure 1.—Holotype of Neanis schucherti (YPM 1233). The slab containing the actual bones,
top to bottom: right radius, right ulna, right coracoid, furcula, left coracoid, and left humerus
(actual length of proximal portion, 13.3 mm), left scapula. The sternum is to the left, and the
approximate boundaries of the posterior notches of the left side of the sternum are outlined in
ink. The position of the supposed spina externa is indicated by an arrow.

that Brodkorb (1970:13) used to place Primobucco
mcgrewi in the Piciformes and Bucconidae: “ (1)
proximal end inflected, so that entire caput humeri
is medial to inner border of shaft (head of hu¬
merus more nearly in line with shaft in other fami¬
lies of Piciformes); (2) shaft more curved than in

other piciform families; (3) deltoid crest long, bent
near its mid-length at an angle of about 150 degrees
(deltoid crest nearly parallel with shaft in other
piciform families).” The other characters outlined
by Brodkorb are not clearly visible in Neanis; how¬
ever, it is clear from the palmar view of the hu-

98

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Figure 2. —The counterslab of the holotype of Neanis schucherti, containing the impressions of
the bones of the slab in Figure I and some feather impressions.

merus (Figure 1) that Neanis is not passerine, but
piciform, and of living families is most similar to
the Bucconidae, as are the two other species of
Piciformes from the lower Eocene Green River
Formation.

The length of the humerus of Primobucco
mcgrewi is 26.7 mm, and that of Primobucco
kistneri was accurately estimated at 18-19 mm, as
the entire outline of the bone was extant. In out¬
line, the humerus of Neanis schucherti is some¬
what similar to that of Primobucco kistneri, but
the bones of the latter are so crushed that the com¬
parison is unsatisfactory; however, the two forms
were of the same general size, which, as I stated of
Primobucco kistneri (Feduccia, 1973:503), “would

probably best approximate . . . some of the modern
African barbets (Capitonidae) of the genus Pogo-
niulus (including Viridobucco), which are ap¬
proximately 4-5 inches in total length.” In the ab¬
sence of the evidence to the contrary, it seems best
for the present to regard Neanis schucherti and
Primobucco kistneri as distinct species; however,
because of their general similarity in size, I recom¬
mend that P. kistneri be included in the genus
Neanis, which has priority over Primobucco. Be¬
cause Primobucco mcgrewi is considerably larger
than either Neanis schucherti or Neanis kistneri,
I strongly recommend the retention of the genus
Primobucco to represent the large lower Eocene
piciform birds from the Green River Formation.

NUMBER 27

99

Literature Cited

Ames, P. L.

1971. The Morphology of the Syrinx in Passerine Birds.
Bulletin of the Peabody Museum of Natural His¬
tory. 37:1-194.

Brodkorb, P.

1965. New Taxa of Fossil Birds. Quarterly Journal of
the Florida Academy of Sciences, 28(2): 197-198.

1970. An Eocene Puffbird from Wyoming. University of
Wyoming Contributions to Geology, 9(1): 13-15.

Feduccia, A.

1972. Variation in the Posterior Border of the Sternum
in Some Tree-trunk Foraging Birds. Wilson Bulle¬
tin, 84(3):315-328.

1973. A New Eocene Zygodactyl Bird. Journal of Paleon¬
tology, 47(3): 501-503, 1 figure, 1 plate.

Heimerdinger, M. A., and P. L. Ames

1967. Variation in the Sternal Notches of Suboscine Pas¬
serine Birds. Postilla, 105:1-44.

Olson, S. L.

1971. Taxonomic Comments on the Eurylaimidae. Ibis,
113(4):507-516.

Sclater, P. H.

1874. On the Neotropical Species of the Family Pterop-
tochidae. Ibis, series 3, 4:189-206.

Shufeldt, R. W.

1913. Fossil Feathers and Some Heretofore Undescribed
Fossil Birds. Journal of Geology, 21(7):628-652.

The Eocene Zygodactyl Birds of North America

(Aves: Piciformes)

Alan Feduccia and Larry D. Martin

ABSTRACT

Recent discoveries of zygodactyl birds in the
Eocene of Wyoming, along with reinterpretation
of previously described taxa, show that these forms
belong to an extinct family, affiliated with the
Bucconidae, for which we here propose the name
Primobucconidae. The genera Primobucco, Ne-
anis, Uintornis, Botauroides, and a new genus,
Eobucco, are assigned to this family, and three new
species, Primobucco olsoni, Uintornis marionae,
and Eobucco brodkorbi are described. Primobuc-
conids appear to have been the dominant small
perching birds of the Eocene of North America.

Introduction

Recent discoveries of piciform zygodactyl birds
from the lower Eocene Green River Formation of
Wyoming brought to light an entirely new element
in the avifauna of the North American Tertiary.
Brodkorb (1970a) described the first of these forms
as a new genus and species of the Bucconidae,
which is structurally the most primitive family of
the Piciformes. This species, Primobucco mcgreivi,
provided the earliest record of the order Piciformes
and the only fossil record of the Bucconidae. In ad¬
dition to describing Primobucco mcgrewi, Brod¬
korb (1970a) suggested that Uintornis lucaris
Marsh (1872), from a much higher level in the
Eocene than P mcgrewi, might also be referable to

Alan Feduccia, Department of Zoology, University of North
Carolina, Chapel Hill, North Carolina, 27514. Larry D. Mar¬
tin, Museum of Natural History and Department of Sys-
tematics and Ecology, University of Kansas, Lawrence,
Kansas, 66045.

the Bucconidae. Later, Feduccia (1973) described
a new zygodactyl bird, Primobucco kistneri, from
the same formation as P. mcgrewi. While P.
mcgrewi was based on a right wing, the type of P.
kistneri included much of the skeleton on a slab.
Although the bones were poorly preserved, the
zygodactyl condition of the toes could be clearly
discerned for the first time in any known fossil.
Few useful osteological characters were present in
this fossil, but by using ratios of the hindlimb ele¬
ments it was at least possible to show that P.
kistneri was a “perching” piciform bird, closely re¬
sembling the Bucconidae and Capitonidae in
proportions. Being from the same approximate
horizon and locality as P. mcgrewi, it seemed rea¬
sonable to assume that the two were related, al¬
though P. mcgrewi was larger than the modern
bucconids Notharchus tectus or Malacoptila pana-
mensis (Brodkorb, 1970a: 14), whereas P. kistneri
was much smaller, being approximately the size of
some of the modern African capitonids of the ge¬
nus Pogoniulus, which are about 100-130 mm long.

Feduccia (pp. 95-99, herein) examined the type
of Neanis schucherti (Shufeldt, 1913), also from
the lower Eocene of Wyoming, which was origi¬
nally described as belonging to the Rhinocryptidae,
thus supposedly representing the earliest record of
the order Passeriformes. He found, however, that
this species is not a passerine, but a piciform, prob¬
ably congeneric with P. kistneri. Because both these
species are much smaller than P. mcgrewi, P.
kistneri was removed to the genus Neanis, and the
genus Primobucco was reserved for larger lower
Eocene forms the size of P. mcgrewi.

The problematical genus Uintornis was origi¬
nally affiliated with the woodpeckers (Picidae) by
Marsh (1872). Shufeldt (1915:51) stated emphat-

101

102

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

ically that this assignment was erroneous, but left
the question of the true affinities of the genus un¬
settled. Cracraft examined the type and agreed
with Shufeldt that it was not from a woodpecker,
suggesting instead that it belonged with the Cuculi-
formes (Cracraft and Morony, 1969:6). On Cra-
craft’s suggestion, Brodkorb (1970b; 1971) placed
Uintornis in the Cuculidae. Our present studies of
the type of Uintornis lucaris show that it is not a
cuculid and that Brodkorb’s (1970a) original as¬
signment of it to the Bucconidae was more nearly
correct. We found three other tarsometatarsi, also
from middle Eocene deposits in North America, to
be affiliated with Uintornis at the family level. One
of these was originally described by Shufeldt (1915)
as a new genus and species of heron, Botauroides
pawns; the second represents an undescribed spe¬
cies of Uintornis; while the last represents a new
genus.

Two species of piciform birds from the Miocene
of Europe have been placed in an extinct family
Zygodactylidae (Brodkorb, 1971), based on the ge¬
nus Zygodactylus (Ballman, 1969a; 1969b). These
forms are distinctly different from the above spe¬
cies and further study of them will be necessary in
order to clarify their affinities with other groups of
modern and Tertiary zygodactyl birds. Mean¬
while, the morphology of the Eocene forms pre¬
cludes their assignment to any known family of
the Piciformes and the erection of a new family is
therefore made necessary.

Acknowledgments.— We are greatly indebted to
S. W. Shannon of the Geological Survey of Ala¬
bama for bringing to our attention the type of
Primobucco olsoni and placing it at our disposal
for study. P. O. McGrew of the University of Wyo¬
ming, Department of Geology, kindly lent the types
of Primobucco mcgrewi and Neanis kistneri. Skele¬
tons of modern species were made available
through the courtesy of Pierce Brodkorb (Univer¬
sity of Florida), R. W. Storer (University of Michi¬
gan Museum of Zoology), and R. L. Zusi (National
Museum of Natural History, Smithsonian Institu¬
tion). C. B. Schultz made the University of
Nebraska specimen available. D. Adams and D.
Bennett rendered the illustrations. Abbreviations
are as follows: Geological Survey of Alabama Type
Collection (GSATC), University of Kansas Mu¬
seum of Natural History (KUVP), University of
Nebraska State Museum (UNSM), University of

Wyoming Geological Museum (UWGM), and Yale
Peabody Museum (YPM).

Order PICIFORMES
Suborder GALBULAE

PRIMOBUCCONIDAE, new family

Included Genera.— Primobucco, Neanis, Uin¬
tornis, Botauroides, Eobucco.

Diagnosis.— Small perching birds with the fol¬
lowing combination of characters: (1) humerus
(Figure la) with shaft curved, the head inflected
medially, and the deltoid crest low, slightly rounded
and proximally located; (2) radius and ulna (Fig¬
ure la) slender and elongate; (3) metacarpals II

Figure 1.— a, Restoration of the right wing of Primobucco
olsoni, palmar view. Palmar views of right humeri: b, Cheli-
doptera tenebrosa (Bucconidae); c, Megalaima lincata (Capi-
tonidae); d, Tauraco sp. (Musophagidae); e, Tapera naevia
(Cuculidae).

NUMBER 27

103

and III (Figure la) nearly parallel to each other,
with only a narrow intermetacarpal space; (4)
phalanx 1 digit II of manus (Figure la) broader
proximally than in most other Piciformes; (5)
tarsometatarsus (Figure 2e, f) relatively short,
broad, and flat; (6) hypotarsus with a square lat¬
eral block of bone and a low ridge leading distally
from it (Figure 6), probably with only a single
tendinal canal (not clearly determinable in any
known specimen); (7) papilla for tibialis anticus
on the extreme internal margin of the tarsometa¬

tarsus (Figure 2e, /); (8) tarsometatarsus with a
single proximal foramen (Figure 6); (9) distal end
of tarsometatarsus flared, with large intertrochlear
spaces and trochleae lying in the same anterior-
posterior plane (Figures 2.e,f,k; 5; 6); (10) middle
trochlea the most distad, with the inner and outer
trochleae about subequal (Figures 2 e,f; 5; 6); (11)
facet for metatarsal I entirely medial (Figures 5,
6); (12) inner trochlea grooved distally and pos¬
teriorly, middle trochlea deeply grooved (Figures
5, 6); (13) outer trochlea inflected inwards, with-

Figure 2. —Proximal views of left tarsometatarsi: a, Notharcus macrorhynchos (Bucconidae); b,
Megalaima lineata (Capitonidae); c, Tauraco sp. (Musophagidae); d, Piaya cayana (Cuculidae).
Restoration of the distal end of the left tibiotarsus and tarsometatarsus: e, Primobucco olsoni.
Anterior (f-j) and distal ( k-o ) views of left tarsometatarsi: f, k, Eobucco brodkorbi (Primobuc-
conidae); g, l, Notharcus macrorhynchos (Bucconidae); h, in, Megalaima lineata (Capitonidae);
i, n, Tauraco sp. (Musophagidae); j, o, Piaya cayana (Cuculidae).

104

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

out a well-developed sehnenhalter, but with a dis¬
tinct groove separating the posterior portion of
the trochlea from the remainder.

Remarks. —-Characters 1, 6 , and 13 of the diag¬
nosis eliminate all families of the Piciformes ex¬
cept the Bucconidae. In the other piciformes, the
bill is long (except in indicatorids and capitonids),
the shaft of the humerus tends to be straight with
the deltoid crest parallel to it (Figure 1c), there
are two or more hypotarsal canals arranged one be¬
hind the other (Figure 2b), and the sehnenhalter
is well developed (Figure 2m). The fossils tend to
be bucconid-like in characters 1, 2, 3, 5, 6 , 7, 8 , and
13. They differ from the Bucconidae, however, in
having a shorter mandibular symphysis, and in
characters 4, and 9 through 12. The bucconids and
the other piciforms differ from the Primobucconi-
dae in that phalanx 1 digit II of the manus is
narrower, especially at its proximal end; the inner
and middle trochleae of the tarsometatarsus are
placed closer together and the outer trochlea is
elevated; the anterior face of the middle trochlea
is not deeply grooved; and the facet for the first
metatarsal is posteriorly situated.

Because Uintornis was suggested as being culculi-
form (Cracraft and Morony, 1969), we have com¬
pared all members of the Primobucconidae with
nine genera of Cuculidae and three genera of
Musophagidae. The Primbucconidae differ from
both of these families in all but character 10 in the
above diagnosis. The musophagids and cuculids
differ from the Primobucconidae in that the ulna is
short, strongly curved, and has large feather papil¬
lae; metacarpal III is strongly curved so that there
is a large intermetacarpal space; the supratendinal
bridge is much above the proximal margin of the
condyles of the tibiotarsus; the tarsometatarsus is
elongate (Figure 2 i,j); the inner and middle
trochleae are close to each other, and the outer
trochlea is elevated and inflected inwards, with no
separation of the posterior portion from the rest of
the trochlea (Figure 2 n,o); the facet for metatarsal
I is located posteriorly and the inner trochlea is not
prominently grooved. The anterior face of the tar¬
sometatarsus is often deeply excavated in cuculi-
forms and there are usually two proximal foramina.
The cuculids always have two large enclosed hypo¬
tarsal canals (Figure 2d), a feature that is certainly
absent in the only primobucconid in which it can
be checked. Also, musophagids have a very triangu¬

lar and distally situated deltoid crest (Figure Id).

Clearly the Primobucconidae belong in the
Piciformes and are most closely related to the Buc¬
conidae. We can find no substantive evidence for a
relationship between Uintornis (or any of the other
forms of Primobucconidae) and any group of the
Cuculiformes.

The family Primobucconidae includes five gen¬
era. The species of two of these genera, Primobucco
and Neanis, are preserved as crushed specimens on
slabs and represent medium-sized and small forms,
respectively, of lower Eocene age. Three genera,
Uintornis, Botauroides, and Eobucco, are medium
to large in size and are middle Eocene in age; all
are represented by fairly well-preserved tarsometa-
tarsi only. Obviously, distinguishing the lower
from the middle Eocene forms is difficult because
the material is not strictly comparable. Neverthe¬
less, size differences and such characters of the
tarsometatarsus as can be discerned in the lower
Eocene genera will permit them to be distinguished
from most, if not all, of the middle Eocene forms.
For this reason, and because of the time element
involved, we believe it is best to recognize five
genera in the Primobucconidae.

Primobucco Brodkorb, 1970a

Type-Species. — Primobucco mcerewi Brodkorb,
1970a.

Included Species.— P. mcgrewi, P. olsoni.

Amended Diagnosis.— Medium-sized primobuc-
conids, larger than Neanis and probably smaller
than Uintornis, Botauroides, or Eobucco.

Primobucco mcgrewi Brodkorb, 1970a

Holotype.— Right wing, UWGM 3255.

Type-Locality and Horizon. —From fish quar¬
ries in SE 14 of Sec 18, T21N, R117W, near Fos¬
sil, Lincoln County, Wyoming; lower beds of
Green River Formation, lower Eocene (Late
Wasatchian) (Brodkorb, 1970a).

Primobucco olsoni, new species

Figures 1 a , 3, 4

Holotype.— Two slabs containing a nearly com¬
plete skeletal impression and counterimpression,

NUMBER 27

105

Geological Survey of Alabama Type Collection,
GSATC 217 (Figures 3, 4).

Type-Locality and Horizon. —The “first bluff”
north of US Highway 30 north, across from Nug¬
get, Lincoln County, Wyoming; Green River For¬
mation, lower Eocene. Collected by Mr. George
Moravec.

Diagnosis. —Smaller than Primobucco mcgrewi
but larger than Neanis. Wing more slender and
humerus longer than in Primobucco mcgrewi.
Middle trochlea of tarsometatarsus extending far¬
ther distally than the other trochleae.

Description.— Skeleton preserved on two slabs
with numerous feather impressions; mandible

broad with a short symphysis; humerus with a
low, gently curved deltoid crest almost parallel to
shaft (Figure lo); head of humerus inflected me¬
dially; radius and ulna slender, straight, and elon¬
gate, the ulna with no evidence of feather papil¬
lae; carpometacarpus long and slender with a
large, straight process for metacarpal I and a nar¬
row intermetcarpal space; phalanx 1 digit II of
the manus broad (narrower proximally in most
bucconids); tibiotarsus short and robust with the
supratendinal bridge straight, lying just above the
condyles; tarsometatarsus broad, short, and flat
with a high intercotylar prominence, anterior face
of shaft with a low medial ridge, a single medial

Figure 3.—Holotype slab of Primobucco olsoni, new species (GSATC 217), viewed ventrally.
(a.t. = anterior toes, co = coracoid, cm = carpometacarpus, d — digit II, h = humerus, m =
mandibular ramus, p.t. = posterior toes, r.u. = radius and ulna, tm = tarsometatarsus, tt =
tibiotarsus)

106

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Figure 4. —Holotype counterslab of Primobucco olsoni, new species, viewed dorsally.

(Abbreviations as in Figure 3.)

proximal foramen, the papilla for tibialis anticus
along the medial edge, the facet for metatarsal I
situated medially, the middle trochlea situated dis-
tally, and the trochleae deeply grooved. Compared
to Primobucco mcgrewi the humerus of P. olsoni
is longer (28.8 vs. 26.7 mm), the shaft more slender
(least width c. 2.2 vs. 3.0 mm), the proximal width
greater (c. 7.5 vs. 6.7 mm), the ulna shorter (c.
32.0 vs. 34.2 mm), and phalanx 1 digit II of
manus shorter (c 6.7 vs. 7.0 mm). The approxi¬
mate length of the coracoid is c. 15.0 mm and that
of the mandibular ramus c. 34.5 mm. The esti¬
mated toe lengths measured (in mm) through the
arc are: digit I, 11.0; digit II, 16.4; digit III, 18.8;
and digit IV, 13.3. Equivalent estimates of the toe
arcs for Neanis kistneri are 6.8, 7.4, 8.6, and 8.2,

respectively. The chord of the left wing of P.
olsoni is estimated at 92 mm; of the species of Buc-
conidae listed in Ridgway (1914), the Barred
Puffbird, Nystalus radiatus, is the nearest in size,
with the wing chord of males averaging about 92
mm.

Etymology.— The specific name is in honor of
Storrs L. Olson for his contributions to avian
paleontology.

Remarks.— No accurate measurements of the
tibiotarsus were possible but by comparing the
two sides we were able to estimate the total length
of the tibiotarsus very roughly as 28.5 (± 3) mm.
The tarsometatarsus measures c. 15.5 mm, giving a
ratio of tarsometatarsus to tibiotarsus of approxi¬
mately 0.54. This rules out an affinity with the

NUMBER 27

107

Picidae (average ratio of 8 species, 0.67) or the
Cuculidae (average of 7 species, 0.69). The same
ratio for Neanis kistneri is 0.56; the average for 6
species of Bucconidae, 0.58; 3 species of Galbulidae,
0.51; five species of Capitoniclae, 0.58; and 2 spe¬
cies of Indicatoridae, 0.59 (Feduccia, 1973). These
ratios indicate only that the fossil does not belong
to either the Cuculidae or the Picidae, but is a
“perching” piciform bird. In woodpeckers, the dif¬
ferent ratio results from the tibiotarsus being pro¬
portionately reduced. Comparative measurements
of bucconids are given in Brodkorb (1970a).

The holotype of Primobttcco olsoni is especially
important because it permits us to associate the
skeleton of Primobucco with those primobucconids
known only from the tarsometatarsus. Features in
which Primobucco resembles Eobucco and Uintor-
nis are the short, broad tarsometatarsus with a
high intercotylar prominence, the single lateral
proximal foramen, the small and very medially
situated tubercle for the tibialis anticus, and the
medially situated facet for metatarsal I.

Neanis Brodkorb, 1965

Synonym. — Hebe Shufeldt, 1913 (preoccupied).

Type-Species. — Hebe schucherti Shufeldt, 1913.

Included Species.— N. schucherti, N. kistneri.

Amended Diagnosis. —Wing relatively shorter
than in Primobucco; tarsometatarsus with large
hypotarsus.

Remarks. —Better preserved material is needed
for a full diagnosis of the very small zygodactyl
birds referred to Neanis.

Neanis schucherti (Shufeldt, 1913)

Holotype. —YPM 1233, partial skeleton on a
slab and impression on counterslab.

Type-Locality and Horizon. —Fish cut of the
railroad, 8 km west of Green River City, Wyoming;
Green River Formation, lower Eocene.

ing; Tipton Tongue Member of the Green River
Formation, lower Eocene (Wasatchian).

Amended Diagnosis. —Smaller than N. schu¬
cherti or any of the other primobucconids.

Uintornis Marsh, 1872

Type-Species. — Uintornis lucaris Marsh, 1872.

Included Species.— U. lucaris, U. marionae.

Amended Diagnosis. —Zygodactyl birds with the
outer trochlea of the tarsometatarsus not as
strongly rotated as in other primobucconids; mid¬
dle trochlea situated distally; distal foramen
located above the outer trochlea.

Uintornis lucaris Marsh, 1872

Holotype.— YPM 617, distal end of right tarso¬
metatarsus (Figure 5 f,g).

lcm

Neanis kistneri (Feduccia, 1973)

Holotype. —UWGM 3196, partial skeleton on a
slab.

Type-Locality and Horizon. —N i/ 2 , NW 14 ,
Sec 6 , T23N, R104W, Sweetwater County, Wyom-

Figure 5.—Tarsometatarsi of Uintornis and Botauroides: a-d,
holotype partial left tarsometatarsus of Uintornis marionae,
new species (KUVP 26906), anterior, posterior, medial, and
distal views; e, holotype, partial left tarsometatarsus of
Botauroides parvus (YPM 1030), anterior view; f-g, holotype
partial right tarsometatarsus of Uintornis lucaris (YPM 617),
anterior and posterior views.

108

Type-Locality and Horizon. —Henry’s Fork,
Sweetwater County, Wyoming; Bridger Formation,
middle Eocene.

Amended Diagnosis. —Outer trochlea of tarsome-
tatarsus inflected inwards but with inner and outer
trochlear ridges not widely separated; middle
trochlea located much farther distally than the
others and distal foramen located more proximad.

Description. —Shaft of tarsometatarsus broad
and flat; inner extensor grooves deep; distal
foramen more proximal than the outer trochlea;
distal end expanded with wide intertrochlear
notches; trochleae deeply grooved; inner trochlea
small, smooth anteriorly and grooved posteriorly,
with a small medially directed posterior flange;
middle trochlea deeply grooved, with high, thick,
trochlear ridges, the inner trochlear ridge distal to
the outer and with a short, thick neck; outer
trochlea large, at the same level as the inner troch¬
lea and inflected medially; outer trochlea grooved,
with the outer ridge produced posteriorly as a
flange or incipient sehnenhalter; facet for metatarsal
I medially situated. Measurements (in mm) of
type: width of distal end, 4.77; width of shaft at
distal foramen, 3.3; depth and width of inner
trochlea, 1.36 and 1.27; depth and width of middle
trochlea, 2.00 and 1.90; depth of outer trochlea
2.90.

Uintornis marionae, new species

Figure 5 a-d

Holotype. —KUVP 26906, distal end of right
tarsometatarsus lacking outer trochlea (Figure
5 a-d).

Type-Locality and Horizon.— Sage Creek, Sweet¬
water County, Wyoming; Bridger Formation, mid¬
dle Eocene.

Diagnosis.— Resembles Uintornis and differs from
Botauroides in having a more slender shaft, a small
inner trochlea, and a distinct groove proximal to
the outer trochlea. Differs from Uintornis lucaris
in being about 25 percent larger, and in having
the facet for metatarsal I less deeply impressed and
the outer ridge of the outer trochlea more medially
inclined and elongated.

Description.— Tarsometatarsus broad and flat;
distal foramen lying proximal to outer trochlea in
anterior view and not visible in posterior view;
facet for metatarsal I situated laterally; inner

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

trochlea grooved posteriorly and smooth anteri¬
orly; intertrochlear spaces wide; middle trochlea
deeply grooved and situated distally; trochleae not
arched. Measurements (in mm) of type: width of
shaft at distal foramen, 4.18; depth and width of
inner trochlea, 1.83 and 1.34; depth and width of
middle trochlea, 2.25 and 2.02.

Etymology.— Named for Marion A. Jenkinson,
who has often assisted us in our work on fossil
birds.

Remarks. —The size difference and various quali¬
tative features of the specimen separate it from
Uintornis lucaris. After Eobucco, described later in
this paper, it is the largest known member of the
Primobucconidae. Uintornis seems to be the least
specialized genus of the family in terms of toe
rotation.

Botauroides Shufeldt, 1915

Type-Species. — Botauroides parvus Shufeldt,

1915.

Included Species.— B. parvus.

Amended Diagnosis. — Botauroides differs from
Uintornis in having a proportionately wider shaft,
the notch for the facet of metatarsal 1 shallower
and slightly more posterior, the inner trochlea at
about the same level as the middle trochlea, and
outer trochlea not as elevated. The inner ridge of
the outer trochlea is slightly more rotated and does
not project as far anteriorly as it does in Uintornis.
The outer trochlea is proportionately smaller and
the medial ridge of the middle trochlea swings
farther medially. Uintornis has a shallow groove
just proximal to the outer trochlea, and the inner
trochlea is relatively smaller than in Botauroides.

Botauroides parvus Shufeldt, 1915

Holotype. —YPM 1030, distal end of left tarso¬
metatarsus (Figure 5e).

Type-Locality and Horizon.— Spanish John
Meadow, Wyoming; Bridger Formation, middle
Eocene.

Diagnosis.— As for the genus.

Description. —Shaft very broad and flat, not ex¬
panded distally; distal foramen situated far proxi-
mally; inner trochlea larger and grooved posteri¬
orly; inner and middle trochleae at same level;

NUMBER 27

109

outer trochlea relatively small, rotated medially
and not very elevated; trochleae not arched.

Remarks. —Cracraft (pers. comm.) directed our
attention to the similarities between Botauroides
and Uintornis. Shufeldt (1915), with his uncanny
ability to err, had referred it to the Ardeidae,
where it appears in Brodkorb’s (1963) catalog.

Eobucco, new genus

Type-Species.— Eobucco brodkorbi, new species.

Diagnosis. —Largest known primobucconid; re¬
sembles Uintornis and differs from Botauroides in
having the facet for metatarsal I deeply impressed
and the middle trochlea located farther distally.
Differs from Uintornis in having the ridges of the
middle trochlea not extending as far proximally
on the anterior side; in having the outer ridge of
the outer trochlea inclined more medially and
elongated until it extends past the outer ridge of
the middle trochlea (it does not reach this trochlea
in Uintornis); in having the outer ridge of the
inner trochlea extending posteriorly as a distinct
flange; and in the far distal position of the distal
foramen.

Eobucco brodkorbi, new species
Figure 6

Holotype.— UNSM 20046 (Figure 6), left tarso-
metatarsus.

Type-Locality and Horizon. —56 km north of
Green River, Sweetwater County, Wyoming;
Bridger Formation, middle Eocene.

Diagnosis. —As for the genus.

Description.— Shaft of tarsometatarsus short,
broad, and flattened; intercotylar prominence
high and large; hypotarsus damaged but with a
large lateral square of bone, which may have in¬
cluded a closed canal (the two closed canals, such
as found in cuculids, could not have been present);
a shallow groove rather than a ridge leads distally
from the hypotarsus; anterior face of shaft grooved,
with a high lateral ridge present; medial edge of
shaft thin; a single large proximal foramen present
near the midline; tubercle for tibialis anticus very
small and distally situated along the medial edge
of the shaft; facet for metatarsal I high and deeply
impressed into the medial side of the shaft; distal

foramen small, level with the outer trochlea;
trochleae widely spread, not arched; inner trochlea
relatively large, anterior face smooth, posterior
face grooved; proximal part of the outer rim of the
outer trochlea produced into a large posterior
flange; middle trochlea large with high trochlear
ridges diverging posteriorly; middle trochlea more
distal than other trochleae; large outer trochlea at
the same level as the inner trochlea; outer trochlea
rotated medially with its elongate outer ridge ex¬
tending medially past the outer rim of the middle
trochlea; outer trochlea grooved. Measurements
(in mm) of type: total length, 26.95; width of
proximal end, 6.75; width of distal end, 6.00; depth
and width of the inner trochlea, 1.36 and 1.27;
depth and width of middle trochlea, 2.00 and 1.90;
depth and width of outer trochlea, 2.90 and 1.80.

Etymology.— Named for Pierce Brodkorb, who

lem

Figure 6. —Holotype left tarsometatarsus of Eobucco brod¬
korbi, new genus and species (UNSM 20046): a, anterior view;
b, posterior view; c, lateral view; d, distal view; e, proximal
view.

110

was the first to recognize the true affinity of the
primobucconids.

Remarks.— This is the largest and most advanced
of the known primobucconids. It shares with Uin-
tornis the distal position of the middle trochlea but
has the outer trochlea much larger and more me¬
dially rotated. The single proximal foramen is re¬
markable for its large size and central position,
thereby resembling the proximal foramen in
Galbula. The medial side of the shaft is deeply in¬
dented under the internal cotyla where it is re¬
duced to a thin blade. The tubercle for the tibialis
anticus is extremely reduced. Eobucco possesses a
combination of primitive and specialized char¬
acters that probably preclude its having given rise
to any modern group of piciforms.

Conclusion

The allocation of Neanis schucherti, Uintornis
lucaris, and Botauroides parvus to the Piciformes,
and the description of the new forms Primobucco
olsoni, Uintornis marionae, and Eobucco brod-
korbi, brings the total number of species of North
American Eocene zygodactyl birds to eight: two
very small forms, Neanis schucherti and Neanis
kistneri; two medium-size forms, Primobucco
mcgrewi and Primobucco olsoni; two large forms,

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Eobucco brodkorbi and Uintornis marionae, and
two others, Uintornis lucaris and Botauroides par¬
vus, slightly smaller than the last two.

The order Piciformes probably arose in the New
World and its forms occupied the ‘‘perching” ar¬
boreal adaptive zone in the early Tertiary of
North America when tropical and subtropical cli¬
mates predominated. Later, the order spread to the
Old World where the Miocene Zygodactylidae and
the modern families Picidae, Capitonidae, and
Indicatoridae are represented, the last named
being the only modern piciform family not found
in the New World. Probably through climatic
change and competition with more advanced land
birds, e.g., passerines, the piciforms retreated
mainly to tropical zones of the New and Old Worlds.
The most primitive living piciform families, the
Bucconidae and Galbulidae, are presently confined
to the New World tropics. The Bucconidae
are structurally closest to the Eocene family
Primobucconidae.

The preponderance of evidence now indicates
that the typical "perching” birds of the early Terti¬
ary of North America were primitive piciforms,
rather than passerine birds. Thus, it was probably
not until the mid-Tertiary that the passerines took
over in North America as the predominant “perch¬
ing” group.

Literature Cited

Ballmann, P.

1969a. Les oiseaux mioc£nes de La Grive-Saint-Alban
(Is£re). Geobios 2:157-204.

1969b. Die Vogel aus der altburdigalen Spaltenfiillung
von Wintershof (West) bei Eichstatt in Bayern.
Zitteliana 1:5-60.

Brodkorb, P.

1963. Catalogue of Fossil Birds, Part 1 (Archaeopterygi-
formes through Ardeiformes). Bulletin of the Flor¬
ida State Museum, Biological Sciences, 7(4): 180-293.

1965. New Taxa of Fossil Birds. Quarterly Journal of the
Florida Academy of Science, 28:197-198.

1970a. An Eocene PufFbird from Wyoming. University of
Wyoming Contributions to Geology, 9(1): 13-15, 1
figure.

1970b. The Paleospecies of Woodpeckers. Quarterly Jour¬
nal of the Florida Academy of Sciences, 33(2): 132-
136, 1 figure.

1971. Catalogue of Fossil Birds, Part 4 (Columbiformes
through Piciformes). Bulletin of the Florida State
Museum, Biological Sciences, 15(4): 163-226.

Cracraft, J., and J. J. Morony, Jr.

1969. A New Pliocene Woodpecker, with Comments on
the Fossil Picidae. American Museum Novitates,
2400:1-8, 1 figure.

Feduccia, A.

1973. A New Eocene Zygodactyl Bird. Journal of Paleon¬
tology, 47(3):501-503, 1 figure, 1 plate.

Marsh, O. C.

1872. Notice of Some New Tertiary and Post-Tertiary
Birds. American Journal of Science, 4(3):256-262.

Ridgway, R.

1914. The Birds of North and Middle America. Bulletin
of the United States National Museum, 50 (6):
1-882.

Shufeldt, R. W.

1913. Fossil Feathers and Some Heretofore Undescribed
Fossil Birds. Journal of Geology, 21:628-652, 12
figures.

1915. Fossil Birds in the Marsh Collection of Yale Uni¬
versity. Transactions of the Connecticut Academy
of Arts and Sciences, 19:1-110, 15 plates.

Oligocene Fossils Bearing on the Origins
of the Todidae and the Momotidae
(Aves: Goraciiformes)

Storrs L. Olson

ABSTRACT

A new genus and species of tody, Palaeotodus emryi,
is described from the “middle” Oligocene (Orellan
land mammal stage) of Wyoming, providing the
first record of the modern family Todidae out¬
side the West Indies. The fossil bird Protornis
glarniensis from the lower Oligocene of Switzer¬
land is removed from the Alcedinidae to the
Momotidae to provide the first occurrence of
the latter family outside the New World. The
Todidae and Momotidae appear to be more closely
related to each other than either is to any other
family of Coraciiformes. The Momotidae were evi¬
dently derived from the Old World. The Todidae
appear to have been derived from a momotid-like
ancestor in the Oligocene or earlier. The present
distribution of these two families in the New
World tropics is relictual. The Coraciiformes ap¬
pear to have been one of the prevalent groups of
small land birds in the Oligocene.

Introduction

The five modern species of todies (Todidae),
endemic to the Greater Antilles, are among the
most intriguing birds of the West Indies. The Mo¬
motidae of Central and South America and the
Todidae are the only families of Coraciiformes
confined to the New World. Apart from late Pleis¬
tocene remains of modern species, there has
hitherto been no fossil record of either family.

Storrs L. Olson, Department of Vertebrate Zoology, National
Museum of Natural History, Smithsonian Institution, Wash¬
ington, D.C. 20560.

Now, a recently discovered fossil and a new inter¬
pretation of one of the first fossil birds to be de¬
scribed, provide us with increased information
bearing on the evolution and geographic origins of
both the Todidae and the Momotidae.

Acknowledgments.— I am indebted to Robert J.
Emry for calling the Wyoming specimen to my at¬
tention, permitting me to work on it, and pro¬
viding much information and assistance. The
manuscript has had the benefit of his comments
and those of John Farrand, Jr., Alan Feduccia, and
Pierce Brodkorb. Robert W. Storer (University of
Michigan Museum of Zoology) kindly lent me a
skeleton of Hylomanes for study, and casts of Swiss
fossils were generously lent by Malcolm C.
McKenna (American Museum of Natural His¬
tory). I am grateful to L. B. Isham for his skillful
illustrations accompanying this paper and to Anne
Curtis for rendering Figure 3.

An Oligocene Tody from Wyoming

In June of 1972, Dr. Robert J. Emry of the
Smithsonian Institution collected several blocks of
matrix containing great concentrations of bones of
small vertebrates from a deposit of Orellan age
(“middle” Oligocene) in east-central Wyoming.
Present in these samples are the abundant re¬
mains of at least two species of squirrels, various
smaller rodents, small marsupials, and insectivores.
That this great concentration of bone may be at¬
tributed to the work of owls is virtually certain
since the blocks also contain the beautifully pre¬
served skeletons of at least four small owls, possibly
of two species.

Ill

112

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

On the surface of one of these blocks, Dr. Emry
noticed the bill of a small bird and later found a
crushed avian cranium that fit perfectly with the
bill. After the two portions had been reunited,
an incomplete avian humerus was found attached
to the lower surface of the skull. There is no rea¬
son to believe that the humerus and skull are not
from the same individual. Were they not, the owl
that cast the pellet containing these remains would
have had to eat two different birds at the same
time, since it is rare for elements from a single prey
item to be found in two successive pellets (Rac-
zyriski and Ruprecht, 1974). This would seem un¬
likely, particularly in view of the absence of birds
other than owls in the remainder of the samples.

The distinctively shaped, flattened bill of the
fossil was immediately reminiscent of the Todidae,
but since bills of similar shape have evolved inde¬
pendently in several groups of birds, many of them
passerines, identification was made cautiously.
After careful comparisons, I concluded that this
specimen is indeed referable to the family Todidae.

Palaeotodus, new genus

Type. — Palaeotodus emryi, new species.

Diagnosis. —Similar to modern Todus but with
different proportions, the wing apparently being
better developed. Bill proportionately shorter and
broader, not as pointed as in Todus. Mandibular
rami not as flattened, the anterior portions
grooved, so as to form a distinct dorsal shelf. Three
ridges on the ventral side of the interorbital bridge
separate, rather than coalesced as in Todus.

Etymology.— Greek, palaeos, ancient, plus To¬
dus, the genus of modern todies. See Newton
(1896:970, footnote) for the etymology of Todus.

Palaeotodus emryi, new species
Figure 1

Holotype. —Incomplete and partially crushed
skull with most of the anterior portions of the ros¬
trum intact, including the mandibular symphysis
and parts of both rami; crushed posterior portion
of cranium with ventral surface of interorbital
bridge well preserved; much of the rest of the skull
crushed, jumbled, and displaced ventroanteriorly.
Vertebrate Paleontological Collections of the Na¬

tional Museum of Natural History, Smithsonian
Institution, USNM 205608. Collected in NW 1/4,
SE 1/4, Sec. 27, T32N, R71W, about 5.6

km SSE of Douglas, Converse County, Wyoming
(42°42'55' / N; 105°2T15"W) on 12 June 1972 by
Robert J. Emry (Field No. WYO. 72-246) and
Leroy Glenn.

Horizon. —Brule Formation, Orellan land-
mammal stage, “middle” Oligocene.

Paratype. —Somewhat distorted right humerus
with the shaft crushed and lacking the distal end;
same number and data as the holotype.

Measurements. —Overall length of skull as pre¬
served 34.5 mm, length of bill from anterior of nos¬
tril 10.0, length of mandibular symphysis 7.8, width
of mandible at beginning of symphysis 5.6, width
of mandibular ramus 1.7, proximal width of hu¬
merus 6.7.

Etymology. —After Dr. Robert J. Emry, the col¬
lector, in recognition of his significant contribu¬
tions to our knowledge of the Oligocene fauna of
North America.

Description.— Bill flat, broad, nearly spatulate,
with a broader more rounded tip than in Todus.
Internarial bar long, slender, terete, and somewhat
heavier than in Todus, continued out the rostrum
as a slightly elevated ridge. Mandible flattened, the
symphysis shorter and broader than in Todus;
mandibular rami deeper than in Todus, grooved
anteriorly to form a dorsal shelf. Ventral surface
of interorbital bridge with three ridges, the middle
one terminating in a pointed process (tip broken
off in the type), the outer ones flaring laterally to
form the edges of the huge anterior cranial fenes¬
tra. The condition in Todus is essentially similar,
but the three ridges are not as distinct anteriorly
and coalesce to form a narrower, deeper interor¬
bital bridge. The middle process in Todus is a

Figure 1.— Skull (holotype) and humerus (paratype) of
Palaeotodus emryi, new genus and species (USNM 205608),
compared with the same elements of Todus subulatus
(USNM 292589): a, dorsal view of skull of P. emryi; b,
dorsal view of skull of T. subulatus; c, lateral view of skull
of P. emryi; d, ventral view of mandible of P. emryi; e, f,
cutaway views of ventral side of interorbital bridge and dor¬
sal part of anterior cranial fenestra of T. subulatus; g, h,
same views of P. emryi; i, anconal view of humerus of P.
emryi; j, proximal end of humerus of P. emryi, viewed with
distal portion tilted further upward; k, proximal view of
same; i, anconal view of humerus of T. subulatus. (All figures
approximately X 3.)

v Hb

_ J

J

j « | J j

f 1

114

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

point of attachment for a narrow ligament that
joins with similar ligaments from the parasphenoid
rostrum and the lateral edges of the cranial fenes¬
tra to form a tenuous cross holding the anterior
part of the brain in place. The humerus of Palaeo-
todus is similar to that of Todus, but is much
larger; the internal tuberosity is heavier and much
less lateromedially elongate; and the ligamental
furrow and the scar on the external tuberosity are
both considerably deeper.

Remarks.— Few nonpasserine landbirds have the
bill anywhere near as flattened as in Todus and
Paleotodus. Those with the most flattened bills are
Electron, Hylomanes, and Aspatha (Momotidae),
Jacamarops (Galbulidae), and Myioceyx (Alce-
dinidae). In all of these forms the bill is deeper and
the internarial bar broader and shorter than in
either of the two genera of Todidae. The bony
structure of the bill in passerines, as for example in
Todirostrum (Tyrannidae), is quite different from
that of Todus and Palaeotodus, being deeper and
more arched, with a more troughlike mandible and
a shorter, wider, and less terete internarial bar.

Palaeotodus agrees with Todus and differs con¬
spicuously from the Momotidae, Alcedinidae, Gal¬
bulidae, and indeed from all of the rest of the
Coraciiformes and Piciformes, in having the ante¬
rior wall of the cranium and the interorbital sep¬
tum unossified. In the Momotidae, Alcedinidae,
and Galbulidae the interorbital septum is partially
or very heavily ossified. The anterior wall of the
cranium is ossified in all of these families, whereas
in Todus and Palaeotodus there is a great, open
fenestra.

The humerus of Palaeotodus is broken and dis¬
torted, with the head crushed down distally relative
to the internal and external tuberosities. Although
in Todus there is slightly more of a depression be¬
neath the head than in other Coraciiformes, the
humerus can in no sense be regarded as having a
double fossa, as stated by Bock (1962), and has a
single pneumatic opening beneath the internal
tuberosity. Palaeotodus is similar to Todus in this
respect. The humerus of Palaeotodus differs from
that of the Passeriformes in the less bulbous head,
the much lesser development of the bicipital crest,
the higher and more pronounced external tuber¬
osity, and the narrower, more ventrally projecting
internal tuberosity. In the Piciformes the deltoid
crest is much more expanded, the head more

bulbous, the capital groove much deeper, and the
internal tuberosity less perpendicular to the shaft
than in Palaeotodus, although these differences are
considerably less pronounced in the Galbulidae and
Bucconidae. Within the Coraciiformes the humeri
of the Upupidae, Phoeniculidae, and Meropidae
have larger, more triangular deltoid crests than in
Palaeotodus, while in the Coraciidae the bicipital
crest is more extensive. The humerus in the
Alcedinidae has the head more bulbous and situ¬
ated much higher above the external tuberosity,
the shaft straighter, the internal tuberosity much
heavier, and the bicipital surface much less pro¬
duced than in Palaeotodus. The humerus of
Palaeotodus is most similar to that of the Todidae
and the Momotidae. In the conformation of the
internal tuberosity it is more similar to the Momo¬
tidae, whereas in the greater excavation of the ex¬
ternal tuberosity and ligamental furrow it more
closely resembles the Todidae.

The manner in which the skull was crushed in
the type of Palaeotodus makes it appear smaller,
while the distortion of the humerus is such as to
make it appear wider and thus larger. Nevertheless,
it is quite evident that the proportions of Palaeoto¬
dus are different from those of Todus, the wing
being much larger in relation to the head.

This difference in proportions may be due at
least in part to the small size of Todus being sec¬
ondarily derived, since the species of this genus are
the smallest members of the order Coraciiformes.
In the evolution of vertebrates, body size usually
changes at a more rapid rate than head size, so that
small forms derived from larger ones tend to have
proportionately larger heads, and vice versa. Many
authors have remarked on the large-headed appear¬
ance of Todus in life. In the Oligocene, the Todi¬
dae were possibly more diverse than at present and
probably included larger, more actively flying
forms with better developed wings than the strictly
sedentary modern todies.

An Oligocene Motmot from Switzerland

In 1839, von Meyer called attention to the re¬
mains of what he thought to be a passerine bird
from slate deposits (Glarner Fischschiefer) in
Switzerland, then considered to be of Cretaceous
age. In a subsequent publication he named this
specimen Protornis glarniensis (von Meyer, 1844).

NUMBER 27

115

Later (von Meyer, 1856), he emended the name to
P. glaronensis and this spelling was in general use
until Brodkorb (1971), whom I have followed, re¬
vived the original orthography. Lambrecht (1933)
maintained Protornis as a genus incertae sedis in
the Passeriformes. At that time the deposits from
which the type of P. glarniensis was derived were
regarded as upper Eocene in age. Subsequent
studies have shown them to be of lower Oligocene
age (Peyer, 1957).

The type of P. glarniensis consists of a slab con¬
taining the bones of all four limbs, the pectoral
girdle, a complete mandible, the quadrates, and a
few vertebrae and ribs. These were insufficiently
exposed when von Meyer studied them, but Peyer
(1957) undertook further preparation of the type,
illustrating his results with numerous photographs
and x-radiographs. The fossil is slightly distorted
from stresses imposed on the rock after deposition;
Stiissi (1958) and Baumann (1958) have offered
mathematical and optical corrections, respectively,
for this distortion. Another less complete specimen
was referred to Protornis, possibly glarniensis, by
Peyer (1957). I have had access to casts of both
these specimens, as well as to Peyer’s excellent illus¬
trations. The casts were made by a copper electro¬
plate process and appear to be very accurate rep¬
resentations of the original specimens.

After his study of the type of P. glarniensis,
Peyer (1957) concluded that it belonged with the
Alcedinidae (kingfishers) and more particularly
that it was nearest to Dacelo. I agree with the as¬
signment of this form to the Coraciiformes, but
numerous characters of its skeleton show conclu¬
sively that Protornis cannot be a kingfisher.

Protornis glarniensis is a small bird, slightly
smaller than the modern motmot Hylomanes
momotula. As detailed by Peyer, many aspects of
its structure demonstrate that it does not belong
with the Passeriformes and the clearly anisodactyl
feet eliminate the Piciformes from consideration.
The mandibular symphysis is broad, flattened and
somewhat spatulate, differing from most non¬
passerine landbirds except the Momotidae and
Todidae. The overall conformation of the man¬
dible is in fact, markedly similar to that of the
motmot genera Electron and Hylomanes and is
quite distinct from that of the kingfishers, includ¬
ing the flat-billed genus Myioceyx (Figure 2). The
symphysis is broader than in the Todidae and

somewhat shorter than in the modern genera of
Momotidae, being most similar in this respect to
Hylomanes, which genus is generally conceded to
be the most primitive of living motmots.

In the mandibular articulation of Protornis the
internal process is a long, thin splint set off from a
well-developed retroarticular process by a distinct
notch, with the actual articulating surface for the
quadrate much reduced. This is exactly the condi¬
tion seen in the Momotidae and Todidae. In the
Alcedinidae the articular cup for the quadrate is
large and deep, the retroarticular process virtually
absent, and the internal process wide, heavy, and
triangular, quite unlike Protornis or the motmots
and todies. Bee-eaters, Meropidae, have a fairly
long, slender internal process, but it is not set off
from the retroarticular by a notch, and the articu¬
lar cup is deep, as in kingfishers. Furthermore, the
bill shape of Protornis is not at all like that of the
Meropidae.

The shape of the hypotarsus in Protornis is ex¬
actly as in motmots and differs from that of the
kingfishers, in which it projects above the cotylae
in a distinct point. The tarsometatarsus of Pro¬
tornis is only slightly shorter than the middle toe
with claw, as in the motmots. In todies the tarsus
is longer than the middle toe with claw, whereas in
the kingfishers and bee-eaters the tarsus is squat
and much shorter than the middle toe. The pro¬
coracoid process appears to be nearly absent in
Protornis, as in motmots and todies, whereas it is
better developed in kingfishers.

From the evidence detailed above it is clear that
Protornis does not belong with the Alcedinidae,
where Peyer (1957) placed it. The proportions of
the bill and of the hindlimb and toes preclude its
assignment to the Todidae. In all of its important
featuxes it agrees with the Momotidae. It differs
from the modern forms of the family mainly in the
shorter mandibular symphysis and the higher, more
expanded sternocoiacoidal process of the coracoid.
Protornis glarniensis should, theiefore, be assigned
to the family Momotidae.

A second fossil from the Glarner Fischschiefer,
consisting of a slab with both hindlimbs, the right
wing, some ribs, and portions of the pelvis super¬
imposed on the sternum, was assigned to the genus
Protornis by Peyer (1957), who suggested that it
might.be referable to the species P glarniensis.
This is plainly impossible, for the second specimen

116

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Figure 2. —Dorsal views of mandibles: a, Electron platyrhynchum; b, Protornis glarniensis
(adapted from Peyer, 1957); c, Hylomanes momotula; d, Todus subulatus; e, Myioceyx lecontei;

j, Dacelo novaegumeae. (Not to scale.)

is much too large to belong to the same species as
P. glarniensis; the carpometacarpus, for example,
is twice as long. Furthermore, in the right foot of
this specimen it can be clearly seen that both digits
I and II are reversed (Figure 3)—a condition

found only in the Trogonidae among modern birds.
The rest of the skeleton of this specimen is gen¬
erally similar to that of modern trogons, although
differing in some details. This specimen deserves a
great deal more attention since it provides the

NUMBER 27

117

Figure 3. —Diagram of right foot of the so-called second
specimen of Protornis showing the heterodactyl condition
typical of the Trogonidae. (Adapted from Peyer, 1957, and a
cast of the specimen; the distal portion of the fourth toe is
present as an impression in the matrix.)

earliest evidence of the occurrence of the hetero¬
dactyl foot. It obviously cannot be assigned to Pro¬
tornis or the Momotidae and for the present should
be regarded as belonging to the Trogonidae. Four
fossil species of trogons are known—all from France
(Brodkorb, 1971). Three of these, in the genus
Archaeotrogon, are from the Phosphorites du
Quercy, which range in age from upper Eocene to
lower Oligocene, and are thus possibly contempor¬
aneous with the Swiss specimen. The fourth spe¬
cies, Paratrogon gallicus, is from lower Miocene
(Aquitanian) deposits at Langy.

A second species of Protornis, P. blumeri, was
described from the Glarner Fischschiefer in 1865
by Heer (1876). The type appears to have been
poorly preserved and has not been restudied, its
whereabouts being unknown. Brodkorb (1971)
placed this species, along with P. glarniensis, in the
Alcedinidae, but considered its position uncertain.
From the original illustration one cannot even as¬
certain that the specimen was avian. In view of
this, and since more than one family of birds occurs
in the Glarner deposits, P. blumeri should be rele¬
gated to the category of Aves incertae sedis.

Discussion

The ten families of the order Coraciiformes fall
into several diverse groups whose relationships
within and without the order are as yet uncertain.
Sibley and Ahlquist (1972:230) maintained that,
■‘no compelling evidence exists to ally any group
of the Coraciiformes more closely to a non-
coraciiform than to other members of the Coracii¬
formes.” On the basis of biochemical analysis of
egg-white proteins, Sibley and Ahlquist concluded,
as have other taxonomists in the past, that within
the Coraciiformes, the Alcedinidae, Todidae, Mo¬
motidae, and Meropidae appear to form a natural
but distantly interrelated group. Feduccia (1975)
discovered that these families possess a highly de¬
rived stapes, which is shared only with the Trogon¬
idae, and concluded that all five families are closely
related. Contrary to most earlier opinions, Sibley
and Ahlquist (1972:230) suggested that the Todi¬
dae are more closely allied to the Alcedinidae than
to the Momotidae. The osteology of these families
does not support this contention, and along with
their distributional history strongly indicates that a
fairly close affinity exists between todies and mot-
mots and that these families differ considerably
from the kingfishers.

Seven of the ten families of Coraciiformes are
confined to the Old World. Of approximately 89
Recent species of Alcedinidae, only six, in two
genera, are found in the New World. Of these, two
are in the genus Ceryle, which also contains two
Old World species, while the genus Chloroceryle,
which is only weakly differentiated from Ceryle,
contains four species endemic to the New World.
Clearly the kingfishers are an Old World family
that has only recently invaded the Americas. Thus,
the Todidae and Momotidae are the only truly
New World families of modern Coraciiformes.

Recent motmots are neotropical in distribution,
ranging from southern Mexico south through Bra¬
zil. In a classical exercise in zoogeography. Chap¬
man (1923) analyzed the distribution of the genera,
species, and subspecies of motmots, concluding that
they had originated in Central America, the few
South American forms having been derived from
the north. Lonnberg (1927), noting that Central
and North America probably presented a more or
less continuous tropical environment in the Ter¬
tiary, felt that the motmots could as easily be con-

118

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

sidered North American in origin. A North or
Central American origin of the modern members
of the family, as opposed to a South American one,
has properly gained general acceptance. Neverthe¬
less, this leaves unanswered the question of the
origins of ancestral motmots.

The modern todies, confined to the West Indies,
are also thought of as being North American in
origin. This is due in part to their presumed re¬
lationship with the motmots and in part to the
North American derivation of most of the rest of
the Antillean avifauna (Bond, 1966). It might then
be asked whether the todies evolved their distinc¬
tive characteristics from some momotid-like ances¬
tor after arriving in the Greater Antilles, or had
already assumed these characteristics before colon¬
izing the islands.

Because of the small size and feeble flight of
modern todies, Bond (1966) expressed reservations
about their ability to cross even narrow water gaps
and suggested that they might have been rafted to
the West Indies from Central America. However,
because the five species of modern todies are very
similar in plumage and morphology, one must as¬
sume that members of the genus Todus have been
able to cross the water barriers between the Greater
Antilles within relatively recent geologic time.
Moreover, since there are two species of Todus on
Hispaniola, one must either assume sympatric
speciation or a double invasion of the island.
Bond’s own remarks (1974) on the greater similar¬
ity of the voice of Hispaniolan T. angustirostris to
that of Cuban T multicolor than to that of the
other Hispaniolan species, T. subulatus, suggest a
double invasion. Thus, if todies were able to cross
the water barriers between the islands of the
Greater Antilles they might as easily have crossed
from the mainland. Furthermore, the evidence pro¬
vided by Palaeotodus shows that in the Oligocene,
todies were larger and had proportions suggestive
of greater powers of flight. It therefore seems pos¬
sible that todies might have colonized the West
Indies over water as easily as, say, kingfishers, and
it is not necessary to invoke rafting to explain their
present distribution.

Without doubt, the order Coraciiformes, as tra¬
ditionally conceived, arose in the Old World. The
existence of Protornis in the lower Oligocene of
Switzerland now provides evidence that the family

Momotidae, presently confined to the New World,
actually had its origins in the Old World. The
place and time of origin of the Todidae are un¬
certain. The Orellan land-mammal stage repre¬
sents a geologically very short span of time follow¬
ing the much longer Chadronian stage and pre-
ceeding the Whitneyan stage, the latter also
representing a very short span of time. The de¬
posits from which Palaeotodus was recovered are
about 30 million years old and have traditionally
be^n regarded as middle Oligocene, although there
is as yet no paleontological correlation between the
North American terrestrial deposits of so-called
Oligocene age and those of Europe. By the reduced
ossification of the skull, Palaeotodus certainly
seems to be referable to the Todidae rather than
the Momotidae, but perhaps with material from
earlier in the Oligocene it would not be possible
to distinguish the two families, the family Todidae
having assumed its characteristics since that time.

Probably through a combination of climatic
change and competition with more advanced land
birds, the motmot-tody group was entirely sup¬
planted in the Old World. The deterioration of
tropical conditions in North America in the late
Tertiary left motmots only in Central America,
from whence they have spread into South America
since the closing of the Panamanian seaway in the
late Pliocene. Similar factors affected the North
American todies and only the isolated West Indian
relicts of the genus Todus have survived up to the
present.

Feduccia and Martin (p. 110, herein) have shown
that the predominant order of small land birds of
the Eocene in North America was the Piciformes.
It is now becoming evident that the Oligocene was
similarly important in the evolution of the Coracii¬
formes. Although the evidence is far from con¬
clusive, if the Coraciiformes (including the Tro-
gonidae) were not the predominant perching land
birds of the Oligocene, they were certainly much
more prevalent than today. Recently I have ex¬
amined a number of fragments of small land birds
of Chadronian and Orellan age from the western
United States. All of these appear to be referable
either to the Coraciiformes or Piciformes and defi¬
nitely are not passerine. Thus, it would appear that
the Passeriformes may not have gained a strong
foothold in North America until the Miocene.

NUMBER 27

119

Literature Cited

Baumann, E.

1958. Affine Entzerrug mit einfachen optischen Mitteln.
Schweizerische Palaeontologische Abhandlungen,
73:17-21, 4 figures. [Separate.]

Bock, W. J.

1962. The Pneumatic Fossa of the Humerus in the
Passeres. Auk, 79(3):425-443, 2 figures.

Bond, J.

1966. Affinities of the Antillean Avifauna. Caribbean
Journal of Science, 6:173-176.

1974. Nineteenth Supplement to the Check-list of Birds
of the West Indies (1956). 12 pages. Philadelphia:
Academy of Natural Sciences of Philadelphia.

Brodkorb, P.

1971. Catalogue of Fossil Birds, Part 4 (Columbiformes
through Piciformes). Bulletin of the Florida State
Museum, Biological Sciences, 15(4): 163-266.

Chapman, F. M.

1923. The Distribution of the Motmots of the Genus
Momota. Bulletin of the American Museum of
Natural History, 48:27-59, 4 figures.

Feduccia, A.

1975. Morphology of the Bony Stapes (Columella) in the
Passeriformes and Related Groups: Evolutionary
Implications. The University of Kansas Museum of
Natural History Miscellaneous Publication, 3:1-34,
7 figures, 16 plates.

Heer, O.

1876. The Primaeval World of Switzerland. Volume 1.
London: Longmans, Green and Co. [English trans¬
lation of Die Urwelt der Schweiz, Volume 1. Zurich,
1856 (not seen).]

Lambrecht, K.

1933. Handbuch der Palaeornithologie. xix + 1024 pages.
Berlin: Gebriider Borntrager.

Lonnberg, E.

1927. Some Speculations on the Origin of the North
American Ornithic Fauna. Kungliga Svenska Ven-
tenskapakademiens Handlingar, series 3, 4 (6): 1-24.

von Meyer, H.

1839. Ein Vogel im Kreideschiefer des Kantons Glaris.
Neues Jahrbuch fur Mineralogie, Geognosie, Geol¬
ogic und Petrefaktenkunde, 1:683-685.

1844. [Letter.] Neues Jahrbuch fur Mineralogie, Geog¬
nosie, Geologie und Petrefaktenkunde, 6:329-340.

1856. Schildkrote und Vogel aus dem Fischschiefer von
Glarus. Palaeontographica, 4(3):83-95, 2 plates.

Newton, A.

1896. A Dictionary of Birds. 1088 pages. London: Adam
and Charles Black.

Peyer, B.

1957. Protornis glaronensis H. v. Meyer Neubeschreibung
des Typusexemplares und eines weiteren Fundes.
Schweizerischen Paldontologischen Abhandlungen,
73:1-47, 26 figures, 11 plates.

Raczyriski, J., and A. L. Ruprecht

1974. The Effect of Digestion on the Osteological Com¬
position of Owl Pellets. Acta Ornithologica, 14(2):
1-36.

Sibley, C. G., and J. E. Ahlquist

1972. A Comparative Study of the Egg White Proteins
of Non-Passerine Birds. Peabody Museum of Nat¬
ural History Yale University Bulletin, 39:1-276, 37
figures.

Stiissi, F.

1958. Die Entzerrung von Fossilien am Beispiel des
Protornis glaronensis. Schweizerischen Paldontolo¬
gischen Abhandluyigen, 73:1-16, 13 figures. [Sepa¬
rate.]

Two New Species of Aegialornis from France, with Comments
on the Ordinal Affinities of the Aegialornithidae

Charles T. Collins

ABSTRACT

Collections from the upper Eocene-lower Oligo-
cene phosphorite deposits of Quercy, France, in¬
clude numerous fossil elements attributed to two
species of Aegialornis. An examination of the hu¬
meri in this series disclosed the presence of two
unrecgonized species, which are newly described
here as Aegialornis wetmorei and A. broweri. Pre¬
liminary study of the other skeletal elements pre¬
viously assigned to Aegialornis indicates that at
least some of them are probably referable to the
Charadriiformes and the Coraciiformes. The hu¬
meri of Aegialornis show closer similarity to Chor-
deiles than to any members of the Hemiprocnidae
or Apodidae, and, therefore, the Aegialornithidae
is removed from the Apodiformes and placed in
the Caprimulgiformes near the Caprimulgidae.

Introduction

In the Museum National D’Histoire Naturelle,
Paris, and the British Museum (Natural History),
are extensive collections of bird fossils from the
upper Eocene to lower Oligocene phosphorite de¬
posits of Quercy, France. These include numerous
distinctive humeri and some additional material
referred to two species in the genus Aegialornis:
A. gallicus Lydekker and A. leenhardti Gaillard.
An additional form, Primapus lacki, was later de¬
scribed from the lower Eocene London Clay of
Britain (Harrison and Walker, 1975). Further
study of the Quercy material indicates the presence
of two additional species of Aegialornis. This ge-

Charles T. Collins, Department of Biology, California State
University, Long Beach, California 90840.

nus has been placed in a distinct family, the
Aegialornithidae, the taxonomic history of which
has been summarized by Harrison (1975). The
family was first proposed by Lydekker (1891) who
treated it as incertae sedis near the Laridae. The
subsequent view of Milne-Edwards (1892) and
Gaillard (1908) that Aegialornis is more properly
included in the Apodiformes has been widely,
though seemingly uncritically, accepted. Brodkorb
(1971), on the basis of the evidence now presented
here, included the Aegialornithidae in the Capri¬
mulgiformes, a placement recently disputed by
Harrison (1975). It is the purpose of this paper to
review the species of Aegialornis and to comment
on the possible affinities of the Aegialornithidae.

A wide array of fossil and recent material was ex¬
amined in this study. Included were the types of
Aegialornis gallicus, A. leenhardti, and Tachyornis
hirundo, and much of the additional material re¬
ferred to these species. Recent skeletons examined
included many genera of Caprimulgiformes, partic¬
ularly Chordeiles, Caprimulgus, and Phalaenopti-
lus, and from one to several species of swifts and
crested swifts in the genera Hemiprocne, Cypse-
loides, Streptoprocne, Apus, Aeronautes, Reinarda,
Hirundapus, and Chaetura.

Acknowledgments.— The study of these fossils,
part of a wider study of the biology and evolution
of swifts, has been generously supported by research
grants and a postdoctoral fellowship from the
Frank M. Chapman Memorial Fund of the Ameri¬
can Museum of Natural History. Fossil material
was obtained on loan from the Institut de Paleon-
tologie, Museum National D’Histoire Naturelle,
Paris (PM); the Department of Paleontology,
British Museum (Natural History) (BMNH); and
the Musee D’Histoire Naturelle de Montauban.

121

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

122

Recent comparative material of Hemiprocne was
obtained from the National Museum of Natural
History, Smithsonian Institution. I am indebted to
the curators of these institutions for their kind¬
nesses. C. J. O. Harrison and C. A. Walker gener¬
ously provided advance copies of their papers and
photographs of Primapus. John Auth provided
essential photographic assistance.

AEGIALORNITHIDAE Lydekker, 1891

Aegialornis Lydekker, 1891

Synonyms.— Tachyornis Milne-Edwards, 1892;
Belornis Milne-Edwards, 1893.

Aegialornis gallicus Lydekker, 1891

Figures 1, 4 b

Type-Locality.— Departement de Lot, Bach

near Lalbenque, France.

Synonym. — Tachyornis hirundo Milne-Edwards,
1892. France, Phosphate de Chaux (= Phosphor¬
ites du Quercy, fide Gaillard, 1908).

This species is represented by at least 20 hu¬
meri: the type-series of 13 and 2 additional speci¬
mens in the British Museum, and 5 specimens, in¬
cluding the type of Tachyornis hirundo, in the
Paris Museum. Referred material includes 2 cora¬
coids, 3 ulnae, 11 carpometacarpi, and 3 proximal
phalanges of digit II (BMNH); and 26 tarsometa-
tarsi (PM). As noted by Lydekker (1891) and
Harrison and Walker (1975), the humerus is short
and stout with a long, prominent, angular deltoid
crest; deep ligamental furrow; large, flattened
ectepicondylar process; laterally compressed head;
broad bicipital surface and bicipital crest; and deep
brachial depression. The humeri of A. gallicus are

a

b

c

Figure 1. —Left humerus of Aegialornis gallicus: a, anconal view; b, palmar view;

c, external view. (X 4.)

NUMBER 27

123

Table 1.—Ranges and means (in parentheses) of measurements (mm) of humeri in Aegialornis

Character

A. leenhardti

n = 2

A. gallicus
n = 20

A. wetmorei

n = 3

A. broweri

n = 1

Overall length .

29.2-29.8

(29.5)

24.3-27.4 (25.9)

21.8-22.4 (22.0)

19.1

Shaft width .

3.7-3.7

(3.7)

2.5-3.3 (3.1)

2.6-2.8 (2.7)

2.3

Shaft depth .

Width of

2.7-2.8

(2.8)

2.2-2.5 (2.4)

2.0-2.1 (2.1)

1.7

distal end .

Depth of

7.0-7.3

(7.2)

5.6-6.3 (5.9)

4.9-5.0 (5.0)

4.2

distal end .

Height of
ectepicondylar

3.9-4.1

(4.0)

3.6-4.0 (3.7)

3.2-3.4 (3.3)

2.7

process .

Height of

ectepicondylar
process as % of

5.9-6.1

(6.0)

5.5-6.3 (6.0)

6.1-6.2 (6.2)

4.8

total length

20.33

23.04

28.03

25.1

smaller than those of A. leenhardti (Table 1), but
larger than those of other species of Aegialornis or
Primapus. The referred elements, not all of which
appear to be properly assigned to Aegialornis , are
discussed below.

Aegialornis leenhardti Gaillard, 1908

Figures 2, 4 a

Synonym.— Originally proposed as Aegialornis
leehnardti Gaillard, 1908; spelling emended to
leenhardti by Brodkorb, 1971:233.

Type-Locality. —Phosphorites du Quercy,
France.

The holotype right humerus (Musee D’Histoire
Naturelle de Montauban) and a previously un¬
recognized left humerus (PM) agree in being
larger than A. gallicus and in having a larger and
more distally located ectepicondylar process (Table
1). Additional material of this species (not ex¬
amined in this study) is present in other museum
collections (Gaillard, 1908; P. Ballmann, pers.
comm.). A left tarsometatarsus from Caylux (Mu¬
seum de Lyon) figured by Gaillard (1908) appears
similar to the numerous tarsometatarsi he referred
to A. gallicus and which were examined in this
study. These specimens are, in my opinion, from
an undescribed species possibly belonging in the
Coraciiformes.

Figure 2- —Holotype right humerus of Aegialornis leenhardti,
palmar view. (X 4.)

124

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Aegialornis wetmorei, new species

Figures 3 a , 4 c

Holotype.— Complete right humerus (PM 15478)
from the upper Eocene-lower Oligocene Phos¬
phorites du Quercy, France.

Measurements of Holotype.— Overall length
from head to internal condyle 21.8 mm, width and
thickness of shaft at midpoint 2.8 X 2.1 mm, great¬
est width of distal end 5.0 mm, thickness through
internal condyle 3.4 mm, height of proximal edge
of ectepicondylar process above distal edge of
ectepicondyle 6.3 mm.

Paratypes.— Two nearly complete left humeri
(PM 15479 and 15480) slightly abraded on deltoid
crest, bicipital crest, and internal tuberosity; from
the same deposits as the type.

Etymology.— This species is named after Dr.
Alexander Wetmore on the occasion of his nine¬
tieth birthday, in recognition of his many contri¬
butions to the field of paleornithology.

Diagnosis and Description.— These humeri are
appreciably smaller and more slender than the

a b

Figure 3.—Holotype right humeri of Aegialornis , anconal
views: a , A . wetmorei , new species; b , A . broweri , new species.
(X 4.)

smallest humerus of A. gallicus (the type of Tachy-
ornis hirundo) or that of the still larger species
A. leenhardti. The ectepicondylar process is less ro¬
bust than in A. leenhardti or A. gallicus and is lo¬
cated farther proximally, being well above the
proximal edge of the brachial depression, whereas
the ectepicondylar process is located at the level of
the proximal end of the brachial depression in A.
gallicus and A. broweri and is somewhat more distal
in A. leenhardti (Table 1).

Aegialornis broweri, new species

Figures 3 b , 4 d

Holotype. —Nearly complete right humerus (PM
15481) from the upper Eocene-lower Oligocene
Phosphorites du Quercy, France.

Measurements of Holotype.— Overall length
from head to internal condyle 19.1 mm, width and
thickness of shaft at midpoint 2.3 X 1.7 mm, great¬
est width of distal end 4.3 mm, thickness of distal
end through external condyle 2.6 mm, thickness
through internal condyle 2.7 mm, height of proxi¬
mal edge of ectepicondylar process above distal
edge of ectepicondyle 4.8 mm.

Etymology.— This species is named after Dr.
Lincoln P. Brower in recognition of his contribu¬
tions to other fields of biology and also for instill¬
ing in me a way of thinking I have tried to follow
throughout my career.

Diagnosis and Description.— The single known
humerus of A. broweri differs from A. leenhardti,
A. gallicus, and A. wetmorei in being smaller, with
a proportionately more slender shaft. The ectepi¬
condylar process is more proximally located than
in either A. leenhardti or A. gallicus, but is not as
far proximal as in A. wetmorei. The brachial de¬
pression is less excavated and the muscle attach¬
ments of the proximal end are less well defined
than in the other species of the genus. The type
shows no signs of immaturity and must therefore
pertain to an additional small species of Aegialor¬
nis in this fauna. The lower Eocene species Prima-
pus lacki is still smaller, the humerus being little
more than two-thirds the length of that of A.
broweri.

Discussion

It is perhaps surprising that there should be four

NUMBER 27

125

such closely related species of Aegialornis (Figure
4) in the same fauna. The differences in the posi¬
tion of the ectepicondylar process in these forms,
however, make it unlikely that the apparent spe¬
cies limits are simply breaks in a continuum of one
or two highly variable or sexually dimorphic spe¬
cies. Primapus lacki from the lower Eocene of Brit¬
ain differs from the four species of Aegialornis in
being much smaller and in having a slightly bi-
lobed appearance to the bicipital crest and the
entepicondyle projecting distally beyond the in¬
ternal condyle (Harrison and Walker, 1975). The
putative swift, Cypselavus gallicus Gaillard, from
the upper Eocene-lower Oligocene Phosphorites
du Quercy, was not examined in this study, but as
noted elsewhere (p. 131, herein), it appears from
the published illustrations that it may be a small
member of the Aegialornithidae, about the same
size as Primapus lacki. The earliest known modern
swift (Apodidae) is Cypseloides ignotus (Milne-
Edwards) from the lower Miocene (Aquitanian)
of France.

The affinity of the Aegialornithidae to the Apod¬
idae and Hemiprocnidae of the suborder Apodi,
has been accepted, largely uncritically, since the
early suggestions of Milne-Edwards (1892) and
Gaillard (1908). This action has recently been en¬
dorsed by Harrison (1975) on the basis of a review
of the humeri and other referred elements of
Aegialornis gallicus and Primapus lacki. From my
study of the referred material of A. gallicus I am
convinced that the coracoids, the proximal pha¬
langes of digit 2, and the tarsometatarsi belong to
species in the orders Charadriiformes and Coracii-
formes, and thus cannot be used to elucidate the
ordinal affinities of Aegialornis. The similarity of
some of these elements to those of the Laridae was
noted by Lydekker (1891) in the original descrip¬
tion of Aegialornis. Until it is possible to restudy
all of the referred material, it seems wisest to con¬
fine discussion of the possible affinities of Aegialor¬
nis to characters of the humerus, the type-element
in all the species of the Aegialornithidae.

The original allocation of Aegialornis to a fam-

a bed

Figure 4. —Size comparison of palmar views of right humeri of the four species of
Aegialornis: a, A. leenhardti, holotype; b, A. gallicus;,c, A. wetmorei, new species, holo-
type; d, A. broweri, new species, holotype. (X 3.)

126

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

ily within the Apodi seems to have been based
principally on the superficial resemblance of the
short and stout humeri to those of the Hemiproc-
nidae and Apodidae. Harrison (1975) also cites the
prominent ectepicondylar process, shorter and
more abruptly projecting deltoid crest, less prox-
imally deflected internal tuberosity, and the pres¬
ence of the distinct flange on the bicipital crest as
characteristics shared with the Apodi. Although
there are some definite similarities between the hu¬
meri of Aegialornis and those of the Apodi, par¬
ticularly He?niprocne, I feel there is a greater re¬
semblance between Aegialornis and some members
of the Caprimulgidae, particularly Chordeiles and
related genera (the Chordeilidae of Oberholser,
1914).

In Aegialornis the head is deflected anconally
and the distal end is directed palmarly. This con¬
dition, somewhat accentuated by the longer shaft,
is also found in Chordeiles. The humerus is notably
straight in Hemiprocne and the primative swifts of
the subfamily Cypseloidinae; only a slight anconal
deflection of the head is present in the Apodinae
and Chaeturinae. The deep ligamental furrow in
Aegialornis and Chordeiles extends well out onto
the internal tuberosity, where it curves to ap¬
proach the distal margin. In the Apodi the liga¬
mental furrow is shorter and straighter, ending
near the proximal base of the internal tuberosity.
In Chordeiles the internal tuberosity is deflected
slightly more proximally at the tip than in Aegia¬
lornis, and the pnenumatic fossa is more exposed. In
the Apodi the internal tuberosity projects laterally
or is deflected distally and bears little resemblance to
that of Aegialornis. A distinct similarity exists be¬
tween Aegialornis and Chordeiles in the thickened
median crest and broadly excavated capital groove
proximal to it. A slight projecting flange on the bi¬
cipital crest of Aegialornis can also be noted in
some specimens of Chordeiles. Although this flange
is usually much more highly developed in Aegialor¬
nis, considerable variation is shown in the speci¬
mens examined in this study, with some individuals
showing only slight development of this feature.
The shape of the deltoid crest is very similar in
Chordeiles and Aegialornis and lacks the more
abrupt taper and concave proximal edge of the
Apodi. The deltoid crest is appreciably different in
other genera of the Caprimulgidae (e.g., Capri-
mulgus and Phalaenoptilns), which have a more

flattened lateral edge and a longer, more gradual
slope to the distal edge. Thus, only some of the
genera of Caprimulgidae have the “longer and
more smoothly rounded” profile of the deltoid crest
incorrectly attributed to the entire family by Har¬
rison (1975).

The distal end of the humerus shows many simi¬
larities between Aegialornis, Hemiprocne, and to a
lesser extent Streptoprocne, in the flared external
tricipital groove and medially expanded entepi-
condyle. Aegialornis and Chordeiles are alike in
having a broader and more protruding attachment
for the anterior articular ligament and a deeper
intercondylar groove. A large peglike ectepicon¬
dylar process, the single most distinctive character¬
istic of the humeri of swifts and Aegialornis, is
accompanied by a distinct, raised lateral muscle
scar at its base in Aegialornis which is lacking in
the Apodi. A small ectepicondylar process and as¬
sociated muscle scar is also present in Chordeiles
and, as noted by Harrison (1975), in Podager. In
the Apodi the ectepicondylar process is never as
thickened as in Aegialornis and is always substan¬
tially more proximally located. In some of the
Apodi there is also a secondary process located dis¬
tal to the ectepicondylar process. This is particu¬
larly well developed in the Hemiprocnidae and
Cypseloidinae but completely absent in Aegialor¬
nis. Within the Caprimulgidae there is consider¬
able difference in the appearance of the distal
portion of the humerus, as well as in the develop¬
ment of the ectepicondylar process. A strong re¬
semblance to Aegialornis can be found in Chor¬
deiles and related genera, but not in Caprimulgus,
Phalaenoptilns, and Eurostopodus.

Although the Aegialornithidae show some simi¬
larities to the Hemiprocnidae, I feel that the ma¬
jority of the characters of the humeri indicate a
closer relationship with the Chordeiles group of
the Caprimulgidae. I therefore place the Aegial¬
ornithidae as a family within the Caprimulgi-
fornres, possibly allied to the Caprimulgidae.

With the tentative removal of Cypselavus galli-
cus from the Apodidae to the Aegialornithidae,
the earliest fossil swifts appear in the lower and
middle Miocene deposits of France (p. 131, herein).
There is thus no longer any evidence to support
the earlier notion that the Apodidae and Aegial¬
ornithidae were contemporaneous during the late
Eocene or early Oligocene. Therefore, the possi-

NUMBER 27

127

bility exists that the Aegialornithidae are repre¬
sentatives of a caprimulgiform lineage that later
gave rise to the swifts and crested swifts. Although
a close relationship between the Caprimulgiformes
and the Apodi is not supported by presently avail¬
able biochemical information (Sibley and Ahl-

quist, 1972), neither does this information pro¬
vide any conclusive evidence of the affinities of
swifts to other groups. A caprimulgiform-
apodiform relationship should be reviewed further
when additional fossil elements are found that can
definitely be assigned to the Aegialornithidae.

Literature Cited

Brodkorb, P.

1971. Catalogue of Fossil Birds, Part 4 (Columbiformes
through Piciformes). Bulletin of the Florida State
Museum, Biological Sciences, 15(4): 163-266.

Gaillard, C.

1908. Les oiseaux des phosphorites du Quercy. Annales
de I’Universite de Lyon, new series, 1(23): 1-178, 37
figures, 8 plates.

Harrison, C. J. O.

1975. Ordinal Affinities of the Aegialornithidae. Ibis,
117(2): 164-170, 5 figures.

Harrison, C. J. O., and C. A. Walker

1975. A New Swift from the Lower Eocene of Britain.
Ibis, 117(2): 162-164, plates 14-15.

Lydekker, R.

1891. Catalogue of the Fossil Birds in the British Mu¬
seum (Natural History). 368 pages. London: Taylor
and Francis.

Milne-Edwards, A.

1892. Sur les oiseaux des depots Eocenes des phosphates
de chaux du sud de la France. Pages 60-80 in
volume 2 of C. R. 2 me International Ornithological
Congress, Budapest 1891.

1893. [Letter.] Bulletin of the British Ornithologists’
Club, 1:53-54.

Oberholser, H. C.

1914. A Monograph of the Genus Chordeiles Swainson,
Type of a New Family of Goatsuckers. Bulletin of
the United States National Museum, 86(1): 1-123,
6 plates.

Sibley, C. G., and J. E. Ahlquist

1972. A Comparative Study of the Egg White Proteins
of Non-passerine Birds. Peajpody Museum of Nat¬
ural History, Yale University, Bulletin, 39:1-276,
37 figures.

A Review of the Lower Miocene Swifts
(Aves: Apodidae)

Charles T. Collins

ABSTRACT

Three nominal species of swifts have been de¬
scribed from lower Miocene (Aquitanian) depos¬
its of France. Re-examination of these forms,
Cypselus [ = Apus] ignotus Milne-Edwards, Collo-
calia incerta Milne-Edwards, and Cypselavus inter-
medius Gaillard, indicates that they are attrib¬
utable to a single species, ignotus, referable to the
modern genus Cypseloides. This provides the first
occurrence of the Cypseloidinae in the fossil record
and indicates a possible origin in the Old World
for this primitive group of swifts, presently re¬
stricted to the New World.

Introduction

Up to now, five species of fossil swifts have been
described, all coming from Tertiary deposits in
France. The present paper is aimed at reviewing
the three nominal species from deposits of early
Miocene age, with comparisons being made with a
much wider array of skeletal material of modern
swifts than were available to the original describers
of the fossil forms. Recent swifts examined in this
study included Cypseloides rutilus, C. cherriei,
C. niger, and Streptoprocne zonaris in the Cypse¬
loidinae; from one to several species in the genera
Apus, Aeronautes, Cypsiurus, Tachornis, Reinarda,
and Panyptila in the Apodinae; and Chaetura, Col-
localia, and Hirundapus in the Chaeturinae.

Acknowledgments. —I thank the Frank M. Chap¬
man Memorial Fund of the American Museum of

Charles T. Collins, Department of Biology, California State
University, Long Beach, California 90840.

Natural History for support and I am also indebted
to Dr. J. P. Lehman of the Institut de Pal^ontol-
ogie, Museum National D’Histoire Naturelle, Paris,
for lending the types of Cypselus ignotus and
Collocalia incerta.

Family APODIDAE
Subfamily CYPSELOIDINAE
Cypseloides ignotus (Milne-Edwards, 1871)

Cypselus ignotus Milne-Edwards, 1871:394, pi. 177: figs. 9-13.
Apus ignotus .—Paris, 1912:286.

Collocalia incerta Milne-Edwards, 1871:394, pi. 177: figs. 1-8.
Cypselavus intermedius Gaillard, 1939:42, fig. 20.

From the Aquitanian deposits at St.-Gerand-le-
Puy, Departement de Allier, France, Milne-
Edwards (1871) described a new species of swift,
Cypselus [=Apus~\ ignotus. This was based on a
complete right carpometacarpus and a left ulna
with the proximal end badly chipped. When I ex¬
amined these specimens, a second left ulna, excel¬
lently preserved, had somehow been associated
with the two syntypes. This is identical to the first
ulna and I therefore refer it to the species ignotus
also. From the same deposits, Milne-Edwards
(1871) named a second species of swift, Collocalia
incerta, based on a single well-preserved left tibio-
tarsus. This was characterized as being much too
small to have come from the same species as the
wing elements assigned to Apus ignotus. Consider¬
ably later, Gaillard (1939) reported a left humerus
of a swift from Aquitanian deposits at Chavroches,
also in'the Departement de Allier, which he de¬
scribed as a new species, ihtermedius, in the

129

130

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Eocene-Oligocene genus Cypselavus (Gaillard,
1908).

I have studied the original material of Apus
ignotus and Collocalia incerta as well as Gaillard’s
(1989) description and illustrations of Cypselavus
intermedins. While all these fossils clearly belong
to the Apodidae, it is equally apparent that none is
referable either to Apus or to Collocalia.

In size and conformation, the two ulnae of
ignotus (Figure lfl) are very similar to those of
some of the smaller species of the genus Cypse-
loides. They lack the well-developed olecranal
process found in the subfamilies Apodinae and
Chaeturinae. This condition is typical of the
Cypseloidinae. The fossil ulnae are slightly longer
and stockier than the ulnae of Cypseloides rutilus
or C. cherriei, and the prominence for the anterior
articular ligament is more shelf-like. Also, the ex-

Ficurf. 1.—Bones of Cypseloides ignotus (Milne-Edwards): a,
referred left ulna; b, syntype right carpometacarpus; c, left
tibiotarsus (holotypc of Collocalia incerta), anterior view; d,
same, posterior view. (Approximately X 3.5, c and d at
slightly different magnifications.)

ternal cotyla appears slightly more undercut at its
palmar edge and the distal ligamental attachment
of the carpal tuberosity is less laterally flared.

The carpometacarpus of swifts shows less marked
distinctions between the subfamilies than does the
ulna. That of ignotus (Figure 1 b) is somewhat
longer and stockier than in Cypseloides rutilus or
C. cherriei, but it does have the more widely flared
pollical facet of metacarpal I and the more pro¬
nounced tendinal groove on metacarpal II typical
of the Cypseloidinae. The fossil also has a broader
and more flared tuberosity of metacarpal II, pro¬
viding a wider articular facet for digit II, which is
typical of the two smaller species of Cypseloides
studied, but not of swifts of other subfamilies.

As noted by Milne-Edwards (1871) the tibio¬
tarsus of Collocalia incerta (Figure 1 c,d) is indeed
more delicate than would be expected for any
member of the Apodinae or Chaeturinae of the
size of Apus ignotus. However, the tibiotarsus in
the Cypseloidinae is proportionately more slender
than in the other subfamilies of swifts, particularly
the Apodinae. There is, in fact, a very close agree¬
ment in overall size and morphology between the
type of Collocalia incerta and Recent specimens of
Cypseloides rutilus. The posterior intercondylar
groove of incerta is not deeply excavated as it is in
members of the Apodinae and Chaeturinae, in¬
cluding Apus and Collocalia. The proximal por¬
tion of the shaft is straight, as in Cypseloides, and
not distinctly bent laterally as typical of many
other swifts. The fossil element is slightly smaller
and stockier than in C. rutilus (C. cherriei has a
much longer tibiotarsus than C. rutilus in spite of
its having wing elements similar in size to both
C. rutilus and A. ignotus), and the internal liga¬
mental prominence is less developed but more ex¬
cavated under the lip of the rotular crest.

The wing elements of ignotus are clearly those
of a small swift belonging to the genus Cypseloides.
The tibiotarsus of incerta similarly shows affinities
to Cypseloides particularly to C. rutilus. Con¬
trary to Milne-Edwards (1871), it is entirely
probable that these fossils, which are from the
same locality and horizon, come from the same
species. This species should now be known as
Cypseloides ignotus (Milne-Edwards) with incerta
becoming a junior synonym, ignotus being chosen
on the basis of line priority.

In the referred ulna of Cypseloides ignotus, the

NUMBER 27

131

maximum length is 17.9 mm, distal width 3.2 mm,
proximal width 3.7 mm, and shaft width 1.6 mm.

No accurate length could be determined for the
chipped ulna in which the distal width is 3.2 mm,
proximal width 3.7 mm, and shaft width 1.8 mm.

The single carpometacarpus measures 16.4 mm in
total length, proximal height 5.2 mm, proximal
width 2.35 mm, and distal width 3.95 mm. The
tibiotarsus has a total length of 21.1 mm, width
across condyles 2.2 mm, width across proximal ar¬
ticular surfaces 2.5 mm, and shaft dimensions of
0.9 X 1.0 mm at the narrowest point and 1.0 X 1.5
mm at the middle of the fibular crest.

The type-humerus of Cypselavas intermedius
Gaillard (1939) was not examined in this study,
but from the original figures it appears to have the
distinctively longer and narrow proportions char¬
acteristic of the species of Cypseloides. As was
noted by Lowe (1939:324), the ectepicondylar
process of intermedius is much more distally po¬
sitioned than in any of the modern forms of the
Apodinae or Chaeturinae but is only slightly more
distal than in Cypseloides, a genus that was not
compared by earlier workers. The measurements
of the type of C. intermedius as given by Gaillard
(1939:43) are: total length 11 mm, proximal width
4.5 mm, distal width 3 mm. Thus, this specimen
agrees closely in size with specimens of modern
Cypseloides rutilus and C. cherriei, and it would
therefore also be of the same approximate size as
C. ignotus. The type of Cypselavus intermedius
comes from the same horizon and from a locality
close to that of Cypseloides ignotus. Since it also
appears to belong to the genus Cypseloides and is
of the same size as C. ignotus, I feel that Cypsela¬
vus intermedius should also be synonymized with
Cypseloides ignotus. As a result, the genus Cypsela¬
vus Gaillard is reduced to a single species, C. galli-
cus, from the upper Eocene or lower Oligocene
(Phosphorites du Quercy) of France; the genus
Collocalia is deleted from the fossil record; and the

Literature

Brodkorb, P.

1971. Catalogue of Fossil Birds, Part 4 (Columbiformes
through Piciformes). Bulletin of the Florida State
Museum, Biological Sciences, 15(4): 163-266.

Brooke, R. K.

1970. Taxonomic and Evolutionary Notes on the Sub-

earliest fossil possibly attributable to Apus now be¬
comes Apus gaillardi (Ennouchi) from the upper
middle Miocene (Tortonian) of France (Brod¬
korb, 1971).

Although the specimens of Cypselavus gallicus
and Apus gaillardi were not examined in this
study, the published illustrations are sufficient to
determine that neither species shows any similari¬
ties to Cypseloides ignotus or the modern Cypselo-
idinae. In fact, Cypselavus gallicus shows a dis¬
tinctly closer resemblance to the Aegialornithidae,
the humerus agreeing in size with the newly
described small aegialornithid Primapus lacki,
from the lower Eocene of Britain (Harrison and
Walker, 1975). In the published illustrations (Gail¬
lard, 1908), the humerus of Cypselavus gallicus
appears to lack the prominant ectepicondylar
process seen in the Aegialornithidae, but this
could well be the result of damage. The illustra¬
tions of the humerus of Apus gaillardi (Ennouchi,
1930) show it to have the general proportions of
the modern Apodidae and Chaeturinae. This spe¬
cies, and an additional swift from the upper Mio¬
cene of Italy, are currently under review elsewhere
(P. Ballmann, pers. comm.).

The Cypseloidinae (see Brooke, 1970:14-15 for
use of this term) appears to be the most primitive
subfamily of the Apodidae. It is therefore not un¬
expected that an extinct species of Cypseloides be
among the earliest known swifts. Like the vultures
of the family Cathartidae [= Vulturidae], the
modern species of Cypseloidinae are confined to
the New World; but also like the Cathartidae
(Cracraft and Rich, 1972), they can now be shown
to have had a past distribution and possible origin
in the Old World. Further elucidation of the origin
and evolution of the Apodidae will have to await a
review of additional modern forms and the remain¬
ing fossil swifts, as well as the swift-like members of
the Aegialornithidae (see Harrison and Walker,
1975; Harrison, 1975, Collins, pp. 121-127, herein).

Cited

families, Tribes, Genera and Subgenera of the
Swifts (Aves: Apodidae). Durban Museum Novitates,
9(2): 13-24.

Cracraft, J., and P. V. Rich

1972. The Systematics and Evolution of the Cathartidae
in the Old World Tertiary. Condor, 74(3):272-283,
10 figures.

132

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Ennouchi, E.

1930. Contribution a I’etude de la faune du Tortonien
de La Grive-St-Alban (Isere). 135 pages, 6 plates.
Paris: Les Presses Modemes.

Gaillard, C.

1908. Les oiseaux des phosphorites du Quercy. Annales
du Universite de Lyon, new series, 1(23): 1-178, 37
figures, 8 plates.

1939. Contributions a l’etude des oiseaux fossiles. Ar¬
chives du Museum d’Histoire Naturelle de Lyon,
15(m£moire 2): 1-100, 34 figures.

Harrison, C. J. O.

1975. Ordinal Affinites of the Aegialornithidae. Ibis,
117(2): 164-170, 5 figures.

Harrison, C. J. O., and C. A. Walker

1975. A New Swift from the Lower Eocene of Britain.
Ibis, 117(2): 162-164, 2 plates.

Lowe, P. R.

1938. On the Systematic Position of the Swifts (Suborder
Cypseli) and Hummingbirds (Suborder Trochili),
with Special Reference to their Relation to the
Order Passeriformes. Transactions of the Zoological
Society of London, 25(4):307-348, 4 plates.

Milne-Edwards, A.

1871. Recherches anatomiques et paleontologiques pour
servir a I’histoire des oiseaux fossiles de la France.
4 volumes (1869-1871). Paris: Victor Masson et Fils.

Paris, P.

1912. Oiseaux fossiles de France. Revue Frangaise
d’Ornithologie, 4(37):283-298.

A New Osprey from the Miocene of California
(Falconiformes: Pandionidae)

Stuart L. Warier

ABSTRACT

Two nearly complete humeri and two partial ulnae
from Barstovian age Miocene deposits at Shark-
tooth Hill, near Bakersfield, Kern County, Cali¬
fornia, are described as the first known extinct spe¬
cies of the modern genus Pandion. Possible func¬
tional implications of the morphological differ¬
ences observed between the fossil species and
modern P. haliaetus are discussed and the fossil
record of the Pandionidae is reviewed.

Introduction

In 1973 an avian fossil owned by a private col¬
lector was brought for identification to the Natural
History Museum of Los Angeles County (LACM)
by Mr. Raj Guruswami-Naidu. The specimen, from
the Miocene Sharktooth Hill beds, was identified by
Dr. Hildegarde Howard and me as a right humerus
closely resembling that of a modern osprey, Pan¬
dion haliaetus. The specimen was cast and re¬
turned, subsequently to be obtained anew by the
collector and original owner, Mr. William Hawes,
who donated it to the LACM, along with portions
of a left humerus and parts of right and left ulnae
that were found associated with it. Through the
courtesy of Dr. Howard, Dr. Lawrence Barnes and
Dr. David Whistler, all of the Department of Ver¬
tebrate Paleontology, LACM, the specimens were
made available to me for study.

Upon detailed comparison, the bones, which

Stuart L. Warter, Department of Biology, California State
University, Long Beach, California 90840, and Research As¬
sociate, Natural History Museum of Los Angeles County, Los
Angeles, California 90007.

bear a remarkable resemblance to those of modern
Pandion haliaetus, were found to differ from that
species in a number of subtle, but apparently sig¬
nificant features. These were considered suffi¬
ciently important to warrant recognition of a new
species, thus extending the history of the genus
Pandion back as far as the Miocene.

The terminology used follows that of Howard
(1929) and Fisher (1946). Twelve specimens of
modern P. haliaetus were examined, four at the
LACM and eight at the University of California,
Los Angeles. Appreciation is expressed to the cu¬
rators of these collections for their cooperation.
Detailed comparisons are based on skeleton LACM
Bi 268, which is typical of larger specimens of P.
haliaetus.

Pandion homalopteron, new species
Figures 1-3

Holotype.— Nearly complete left and right hu¬
meri and proximal portions of left and right ul¬
nae, all associated; LACM 42815; collected by Mr.
William Hawes of Bakersfield, California. Right
humerus entire, but shattered and filled in two
places with plaster; surfaces of head and internal
tuberosity sufficiently intact to permit reasonably
accurate total measurement (151 mm), but other
contours of both ends badly eroded; deltoid crest
missing. Left humerus consisting of three pieces,
plus fragments; shaft and distal end joined by ac¬
tual contact along external surface, proximal por¬
tion joined by comparison with companion right
humerus and with left humerus of recent Pandion.
Head of left humerus entire, but other contours of
proximal end severely eroded; all articular sur¬
faces of distal end intact; only a small portion of

133

Figure J.—Holotypc humeri of Pandion hornaloptcron, new species (LACM 42815): a, palmar
view and, b, anconal view of left and right elements. Natural size. (Courtesy of Natural History
Museum of Los Angeles County)

base of deltoid nest remaining. Left ulna badly
shattered, but nearly complete (1 HO nun), lacking
approximately 20 mm or less of the distal end;
proximal articular surfaces largely intatt, tip of
olecranon and tip of external cotyla broken. Right
ulna less complete (120 mm) with proximal articu¬

lar surface largely intact, but olecranon, tip of in¬
ternal cotyla, and edge of external cotyla missing;
distal IK mm of shaft lacking the anconal surface,
last 2>2 mm displaced to the palmar side and joined
only by matrix. Colors variable: right humerus
light tan; right ulnar fragment brown; left hu-

NUMBER 27

135

Figure 2.—Holotype humeri of Pandion homalopteron, new species (LACM 42815): a, external
view and b, internal view of left and right elements. Natural size. (Courtesy of Natural History
Museum of Los Angeles County)

merits with proximal segment light tan, shaft and
distal segment brown; left ulna brown proximally,
fading to light tan distally.

Locality and Age.— From the Sharktooth Hill
bone bed, middle Miocene (Barstovian age, Savage
and Barnes, 1972:133). Round Mountain Silt,

Sharktooth Hill, near Bakersfield, Kern County,
California; LACM locality 3205.

Measurements of Holotype. —See Table 1. The
brachialis scar of the right humerus is 13 ± 2.5
mm long by 6 ± 1 mm wide (margin indistinct)
and the length of the brachialis scar of the right

136

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Table 1.—Measurements (mm) of the holotype humeri (R
= right; L = left) of Pandion homalopteron, new species,
compared with modern P. haliaetus

Character

Width of head from
the external to the
internal tuberosity
Length from head to
distal margin of

Ratio of above
measurement to
total length (= %

Width of shaft at
distal end of
deltoid crest
Length of

ectepicondyle from
proximal margin
of extensor
metacarpi radialis
to distal margin of
flexor metacarpi

radialis scar .

Length of
entepicondyle
from proximal
margin of
pronator brevis
scar to distal
margin of flexor
carpi ulnaris scar
Width of distal end

P. homalopteron

P.

haliaetus

mean

range

n

151 (R)

145.9

135-154

8

27.5 (R)

27.1

25-28.5

9

63 (R&L)

63.3

58-69

8

41

42.9

40-44

8

12 (R & L)

11.5

11-12

8

14.5 (R&L) 15.8 14.5-16.5 8

12.5 (L)

24.5 (L)

14.6 13-16 8

23.8 21.5-24.5 10

ulna is 30 mm. These measurements in P. haliae¬
tus are variable and may differ between the right
and left sides of the same individual: the brachialis
scar of the humerus ranges from 15 to 17.5 mm in
length and 7 to 8.5 mm in width (n = 10); that of
the ulna ranges from 28 to 36 mm in length
(n = 10).

Diagnosis. — Pandion with humerus and ulna re¬
sembling those of large individuals of modern P.
haliaetus, but showing evidence of weaker muscu¬
lature and other osteological features that prob¬
ably permitted less extension at the elbow and less
rotation at the shoulder.

Etymology. —Greek homalos, even, level; and
pteron, wing; referring to the more level configu¬
ration of the wing that would result from a reduced

ability to raise the wrist during soaring, thereby
reducing or eliminating the “kinked-wing” appear¬
ance often presented in flight by members of the
modern species.

Description. —Humerus with head more tri¬
angular, less rounded than in the modern form;
capital groove and ligamental furrow shallower,
less deeply excavated; anconal surface of internal
tuberosity in internal view less tapering, more
nearly perpendicular to main axis of shaft; capital
groove and median crest not extending below
pneumatic foramen as they do in P. haliaetus.

Distal end of humerus with internal condyle
higher, more rounded than in P. haliaetus; ole-
cranal fossa in palmar-distal view shallower and
wider; border of fossa in anconal-distal view less
rounded, more triangular; brachial depression
noticeably smaller and less excavated, its external
margin situated more externally; external condyle
in palmar view rotated, its long axis at a greater
angle from the axis of the shaft; viewed from the
external side the external condyle rounder, less
squared, and less deep than in the modern form.
Ectepicondylar and entepicondylar prominences
shorter, the scars for M. extensor metacarpi radi¬
alis and M. pronator brevis closer to the distal end
of the bone; facet of anterior articular ligament
wider and shorter, its surface flat to concave, this
facet in P. haliaetus being longer, narrower, and
convex.

Proximal end of ulna with surface of internal
cotyla shallower, its lip (palmar surface) more ex¬
tensively flared; palmar lip of radial depression
less enlarged; surface of external cotyla less angled
from the axis of the shaft, more steeply inclined
from the surface of the internal cotyla; scar for the
insertion of M. brachialis ( = M. brachialis anti-
cus) shorter than in most modern specimens of
equivalent size; prominence for anterior articular
ligament with shorter, wider facet; proximal half
of radial surface of ulna convex in cross-section,
whereas flattened or concave in the modern form;
olecranon apparently less robust.

Discussion

The modern Osprey, Pandion haliaetus, is a
highly specialized fish hunter. It is capable of hov¬
ering over one spot with rapidly beating wings held
high over the back. Also, it is capable of increasing

NUMBER 27

137

Figure 3. —Holotype ulnae and humerus of Pandion homalopteron , new species (LACM 42815):
a, dorsal view and b, palmar view of left and right ulnae; c, distal end of left humerus in
distopalmar view. Natural size. (Courtesy of Natural History Museum of Los Angeles County)

138

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

the angle of attack of the wings in level soaring
flight by raising the wrists through rotation of the
long outstretched forelimbs. In this position the
wings present the “kinked” or “crooked” appear¬
ance for which it is so well known.

The distinctive shape of the deltoid crest of P.
haliaetus is undoubtedly related to the aforemen¬
tioned abilities. It is a large, triangular plate, de-
flexed from the external surface in a palmar
direction and beginning abruptly at a point more
distad than in other falconiforms. While the del¬
toid crest is missing from both humeri of P ho-
malopteron, enough remains of the base of the
crest of the left humerus to tell that the entire
process was similarly positioned, enlarged and de-
flexed (Figure 1).

A large deltoid crest usually is indicative of
strong pectoral musculature and well-developed
powers of flight. Perhaps paradoxically, such a crest
occurs in Pandion along with a rounded sternal
carina, a feature that in soaring birds like cathar-
tid vultures often is associated with a small deltoid
crest and a relatively weak flight mechanism. In
Coragyps (Fisher, 1946:603), the palmar surface of
the deltoid crest provides the area of insertion of
the superficial layer of M. pectoralis (= pectoralis
superficialis), and on the anconal surface of the
crest, M. deltoideus minor and M. deltoideus major
originate on the crest and on extensive areas of the
shaft of the humerus proximal, distal, and posterior
to the deltoid crest. Judging from muscle scars on
the humerus of P. haliaetus, however, the enlarged
deltoid crest serves as the area of insertion for
much, if not all, of the large M. deltoideus major,
whereas M. deltoideus minor is small and inserts
anconally along the shaft anterior to the crest.

In Coragyps, the anterior portion of M. deltoi¬
deus major “is more important in elevating the
leading edge of the wing since the anterior exten¬
sion of the deltoid crest provides a longer lever
arm” (Fisher, 1946:590). An important function of
the expanded, deflexed, deltoid crest in Pandion,
then, is to provide a lever arm for increasing the
upward rotational ability of the humerus through
the action of M. deltoideus major. A similar in¬
crease in downward rotational ability would prob¬
ably be conferred to the humerus by M. pectoralis.

The humerus of Pandion also has an enlarged
internal tuberosity, the function of which is to in¬
crease the lever arm for several small muscles in¬

serting upon it which, in vultures (Fisher, 1946:
603), serve to depress the trailing edge of the wing,
thereby raising the leading edge. The described
differences in the humeral head, capital groove,
ligamental furrow, and internal tuberosity of P.
homalopteron may indicate a lesser degree of mus¬
cular development and rotational ability in the
shoulder than in P. haliaetus.

The differences in the morphology of the elbow
joint provide additional evidence of some degree of
functional dissimilarity between the two species of
Pandion. The configuration of the joint surfaces
would seem to indicate a lesser degree of extension
at the elbow in P homalopteron. This appears to
be borne out by mechanical manipulation of the
bones. The forearm of P. haliaetus exhibits a much
greater degree of extension at the elbow than does
that of Buteo. The robust olecranon of P. haliaetus
fits closely into its corresponding depression on the
humerus, possibly serving as a bony stop against
further extension. The degree of this extension in
P. homalopteron is also greater than in Buteo, but
less than in P. haliaetus. In spite of the olecranon
being incomplete, the ulna of P homalopteron
could not be extended to the same degree as that
of P. haliaetus without partially disarticulating the
joint. The observed differences in the size and con¬
figuration of the attachment for the anterior ar¬
ticular ligament also might be related to the de¬
creased ability to extend the elbow.

In the fossil form, the convex radial surface of
the ulna, the smaller M. brachialis scars, and pos¬
sibly the more distal origins of Mm. pronator brevis
and extensor metacarpi radialis, may indicate
weaker intrinsic musculature. All three of the above
muscles are involved in flexion of the forearm, while
Mm. brachialis and pronator brevis also are in¬
volved in supination and pronation, respectively,
of the manus (Fisher, 1946:591-594).

Intrinsic rotational movements of the hand and
forearm in birds are limited (Bellairs and Jenkin,
1960:258), and the degree to which they occur has
not been determined (George and Berger, 1966:
14). However, P haliaetus may have greater abili¬
ties to raise the wrist through intrinsic rotation, as
well as by rotation at the shoulder of an entire,
more extended wing, than did P. homalopteron.

There is no reason to assume that P. homalop-
teron was any less variable in its dimensions than
is modern P. haliaetus. To the extent that the

NUMBER 27

139

single available specimen can be considered typical
of the Miocene population, the species P. homalop-
teron appears to have been larger in absolute
gross skeletal dimensions than an average-sized
modern Osprey, but was smaller than average in
other measurable features. Several of these features
may indicate a lighter wing musculature relative to
bone size than is found in P- haliaetus. This, along
with the seemingly lesser powers of extension and
rotation of the wing, presents a picture of a bird
similar to the modern Osprey in size, but one with
a more level wing and less refined powers of soaring
and hovering. Such a bird could be ancestral to P
haliaetus.

The Fossil Record of the Pandionidae

Pandion homalopteron provides the only Ter¬
tiary record of the family Pandionidae founded on
adequate and diagnostic material. Brunet (1970)
has placed the species Palaeocircus cuvieri Milne-
Edwards, based primarily on an incomplete carpo-
metacarpus from the upper Eocene of France, in
the Pandionidae, stating that the type, while spe¬
cifically distinct, is scarcely separable from Pan¬
dion at the generic level. Storrs Olson (pers.
comm.) believes that Brunet’s illustrations of the
specimen indicate to the contrary, however, since
both the proximal and distal symphyses between
metacarpals II and III are longer than in Pandion.
The assignment of Palaeocircus to the Pandionidae

should be regarded with caution, particularly since
the family is at present monotypic and the addi¬
tion of another genus would require redefinition
of the family.

A record of Pandion from the middle Pliocene
Bone Valley Formation in Central Florida (Brod-
korb, 1972) is based on a single claw (Storrs Olson,
pers. comm.). Another claw, kindly lent to me by
the National Museum of Natural History, Smith¬
sonian Institution, (USNM 192193), comes from
the Lee Creek phosphate mine near Aurora, Beau¬
fort County, North Carolina. Middle Miocene and
Pliocene fossiliferous deposits are exposed there,
and in this case it is not certain from which level
the specimen was derived (Storrs Olson, pers.
comm.). This claw is from digit III of the right
foot, but has the tip broken so that an accurate
measurement of the chord is not possible. It is ref¬
erable to the genus Pandion and is of a size ap¬
propriate for either P. homalopteron or P. haliae¬
tus, but since this element is not diagnostic and its
age is uncertain, no specific identification can be
made.

The only other fossil records for the Pandionidae
are Pleistocene remains of the modern species
Pandion haliaetus. To the various localities listed
in Brodkorb (1964:260) may be added a pre¬
viously unreported left tarsometatarsus (LACM
27082) from Pleistocene deposits at Kelly Springs,
Kelly Park, Orange County, Florida (LACM lo¬
cality 7119).

Literature Cited

Bellairs, A. D’A., and C. R. Jenkin

1960. The Skeleton of Birds. Pages 241-300 in Volume
1 of A. J. Marshall, editor, The Biology and Com¬
parative Physiology of Birds. New York: Academic
Press.

Brodkorb, P.

1964. Catalogue of Fossil Birds, Part 2 (Anseriformes
through Galliformes). Bulletin of the Florida State
Museum, Biological Sciences, 8(3): 195-335.

1972. New Discoveries of Pliocene Birds in Florida [Ab¬
stract], Page 64 in Proceedings of the XVth Inter¬
national Ornithological Congress. Leiden: E. J.
Brill.

Brunet, J.

1970. Oiseaux de l’Eocfene Superieur du Bassin de Paris.
Annales de Paleontologie (Vertebres ), 56(1): 1-57, 4
plates.

Fisher, H. I.

1946. Adaptations and Comparative Anatomy of the

Locomotor Apparatus of New World Vultures.
American Midland Naturalist, 35(3): 545-727, 28
figures, 13 plates.

George, J. C., and A. J. Berger

1966. Avian Myology, xii + 500 pages. New York: Aca¬
demic Press.

Howard, H.

1929. Avifauna of the Emeryville Shellmound. University
of California Publications in Zoology, 32(2):301—
394, 54 figures.

Savage, D. E., and L. G. Barnes

1972. Miocene Vertebrate Geochronology of the West
Coast of North America. Pages 124-145 in E. H.
Stinemeyer, editor, Proceedings of the Pacific Coast
Miocene Biostratigraphic Symposium. 364 pages.
Bakersfield, Cal.: Society of Economic Paleontolo¬
gists and Mineralogists.

A New Species of Flightless Auk
from the Miocene of California
(Alcidae: Mancallinae)

Hildegarde Howard

ABSTRACT

Praemancalla wetmorei is described from the late
Miocene of Orange County, California, with hu¬
merus and ulna as holotype and paratype, and
radius, carpometacarpus, and coracoid referred.
The species, although less specialized as a flightless
diver than the geologically younger genus Man-
calla, appears to be more advanced then Praeman¬
calla lagunensis, which is believed to be derived
from slightly older deposits.

Introduction

Since Lucas (1901) described the first humerus
of Mancalla, knowledge of the flightless mancalline
alcids (Mancallinae) has increased to include
nearly all skeletal elements and to involve five spe¬
cies and two genera.

The type-genus, Mancalla, is known from four
species— M. californiensis Lucas (1901), M. di-
egense (Miller, 1937), M. milleri Howard (1970)
and M. cedrosensis Howard (1971)—and is recorded
from Humboldt County in northern California to
Cedros Island, Mexico. The Humboldt County
site, with a single humerus assigned to M. diegense
(Howard, 1970), is believed by Kohl (1974:217) to
be Pleistocene in age. The other records are middle
to late Pliocene. Praemancalla is known from the
single species, P. lagunensis Howard (1966), de-

Hildegarde Howard, Chief Curator Emeritus, Natural His¬
tory Museum of Los Angeles County, 900 Exposition Blvd.,
Los Angeles, California 90007.

scribed from a late Miocene deposit in Leisure
World, Laguna Hills, Orange County, California.

Recently, three other Miocene sites in Orange
County have yielded mancalline bones. These sites
are in Laguna Niguel, approximately 5 km south
of the Laguna Hills locality. The specimens from
these sites are in the collections of the Natural His¬
tory Museum of Los Angeles County. The catalog
and locality numbers are listed under Los Angeles
County Museum (LACM).

The associated avifauna from these sites includes
the same families recorded at the type-locality of
Praemancalla lagunensis (LACM Loc. 1945), but
the species represented are not identical. None of
the species described as new from locality 1945 has
appeared in the Laguna Niguel localities. On the
basis of associated marine mammals, it is suggested
that these sites may represent a later subdivision of
the late Miocene than locality 1945 (Barnes, et al.,
in prep.).

The mancalline skeletal elements from Laguna
Niguel include humerus, ulna, radius, carpometa¬
carpus, and coracoid, all of which have been pre¬
viously described for Mancalla. Only for the
carpometacarpus, coracoid, and distal end of the
humerus is there comparable material of Praeman¬
calla. The newly found specimens suggest a gen¬
erally larger form than any previously described
mancalline species. Qualitative characters show dis¬
tinction from comparable elements of all species of
Mancalla. Distinctions are also apparent with re¬
spect to Praemancalla lagunensis, but the degree of
adaptation towards restriction of the wings for
swimming is closer to Praemancalla than to Man¬
calla. Possibly a third genus is indicated. At the

141

142

present state of knowledge, however, it seems wiser
to assign the specimens to the genus Praemancalla
under a new species name.

Acknowledgments. —I am grateful to the Earth
Sciences Division of the Natural History Museum
of Los Angeles County for placing the museum’s
collections at my disposal and for the many cour¬
tesies of the staff members. I particularly appre¬
ciate the assistance of Dr. Lawrence Barnes in dis¬
cussion of matters pertaining to the field work in
Orange County. The photographs were taken by
Lawrence Reynolds, museum photographer. Spe¬
cial thanks go to my husband, Henry Anson Wylde,
for the art work in connection with the prepara¬
tion of the plate.

Praemancalla Howard

In describing Praemancalla lagunensis, the spe¬
cific diagnosis of the holotype humerus and para-
type carpometacarpus served also as the generic
diagnosis. In the specimens from Laguna Niguel
now at hand, the following characters are in agree¬
ment with that diagnosis:

Humerus with groove separating base of ectepi-
condyle from external condyle, brachial impres¬
sion faint and running diagonally from ectepi-
condylar prominence to a point slightly proximal
to attachment of articular ligament, with no pap¬
illa present above condyles. The tricipital grooves
and ridges are broken in the humerus from Laguna
Niguel, so the characters of this area set forth in
the original diagnosis cannot be assessed.

Carpometacarpus with distinct, blunt pisiform
process, trochlear area having narrow, deep groove
between internal and external crests posteriorly,
metacarpal II relatively broad with more rounded
anterior contour and more angular internal con¬
tour than in Mancalla, and process of metacarpal
I relatively shorter.

The following additional characters observed in
the specimens now at hand are considered to be of
generic value when compared with Mancalla: hu¬
merus with head only slightly extended over capi¬
tal groove, deltoid crest weakly developed, area of
anterior articular ligament slightly swollen; ulna
with prominent olecranon process; radius lacking
prominent crest on convex contour; coracoid with
scapular facet facing dorsally, coracohumeral at¬
tachment flat and angular in outline.

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Praemancalla wetmorei, new species
Figure 1 a, b, e-g, i-k

Holotype.— Humerus, LACM 42653, complete
except for tricipital area of distal end (Figure

i hi)-

Type-Locality. —LACM Loc. 6906, site of exca¬
vation for North American Rockwell building
(now U.S. General Services Administration build¬
ing) on El Lazo Road, Laguna Niguel, Orange
County, California; 914 m north of junction of
Aliso Creek and Sulfur Creek, in yellow sands and
laminated gray shale. Latitude 33°33’43” N, longi¬
tude 117°42’44” W. In the NE 1/4 NE 1/4 SE
1 /4 of unsurveyed Sec 16, T7S, R8W, San Juan
Capistrano quadrangle, USGS 1948, 1:24000.

Formation and Age. —Monterey Formation, late
Miocene.

Paratype.— Proximal end of ulna LACM 32429
from type-locality (Figure \e).

Diagnosis. —Humerus broad proximally; medial
profile of capital groove a wide open curve; depth
through deltoid crest only 5 percent greater than
depth of shaft above distal end; ectepicondylar
prominence notably protuberant at its proximal
tip and slightly lateral in position with respect to
palmar surface of shaft; groove between external
condyle and base of ectepicondyle more constricted
and less distal in extent than in P. lagunensis;
shaft breadth above ectepicondylar prominence
53 percent of shaft depth at same point; shaft
depth 113 percent of breadth of distal end.

Ulna laterally compressed, with short brachial
impression partially palmad in position and bor¬
dered palmad by heavy ridge; olecranon blunt but
protruding proximally beyond cotylae and dis¬
tinctly set off from cotylae by lateral depression
both externally and internally.

Measurements.— Humerus: length to external
condyle 82.7 mm, greatest proximal breadth from
pectoral to bicipital crests 22.2 mm, breadth across
head 19.6 mm, breadth through distal condyles
8.5 mm, breadth and depth of shaft above ectepi¬
condylar prominence 5.1 mm and 9.6 mm, respec¬
tively, height of ectepicondylar prominence above
distal end 16.9 mm, greatest depth through deltoid
crest 10.1 mm, breadth of shaft at same point 5.5
mm. Ulna: proximal breadth across cotylae 7.5
mm, proximal depth through olecranon 11.3 mm,

NUMBER 27

143

Figure 1.—Skeletal elements of Praemancalla and Mancalla: a, b, coracoid (LACM 37637) of
P. wetmorei, new species, medial and dorsal views; c, coracoid (LACM 15289) of P. lagunensis,
dorsal view; d, coracoid (LACM 2310) of M. diegerise, dorsal view; e, paratype ulna (LACM
32429) of P. wetmorei, internal view; f, referred carpometacarpus (LACM 52216) of P. wetmorei,
internal view; g, radius (LACM 53907) of P. wetmorei, palmar view, h, humerus (LACM 15367)
of M. cedrosensis, palmar view; i, j, holotype humerus (LACM 42653) of P. wetmorei, palmar
and anconal views; k, referred humerus (LACM 32432) of P. wetmorei, anconal view; I, humerus
(LACM 2331) of M. diegerise, anconal view. (Approximately natural size.)

144

breadth and depth at middle of shaft 4.2 mm and
6.2 mm respectively.

Referred Material. —From Laguna Niguel,
Orange County, California, late Miocene, Mon¬
terey Formation. Proximal end of humerus LACM
32432 (Figure Ik) from type-locality (LACM Loc.
6906). Complete radius LACM 53907 (Figure lg)
and scapular end of coracoid LACM 37637 (Figure
la,b) from LACM locality 6902 at northwest end
of El Lazo Road, 365 m northwest of type-locality
and 244 m east of Aliso Creek, in basal 0.5-1.5 m of
coarse yellow sand directly overlying laminated
gray shale. Proximal % of carpometacarpus LACM
52216 (Figure 1/) from LACM locality 3185, in
tributary gully west of Aliso Creek in coarse yellow
sand.

Etymology.— I take pleasure in naming this new
species in honor of Dr. Alexander Wetmore, who
has done so much to further the science of pale-
ornithology and who has generously given advice
and counsel to me throughout my years of study in
this field.

Description. — Compared with Mancalla, the hu¬
merus of P. wetmorei is relatively, as well as ac¬
tually, broader both proximally and distally (ratio
of greatest proximal breadth to length 26.8 percent
in P. luetmorei, 23-25 percent in Mancalla; ratio of
distal breadth to length 10.2 percent in P wet¬
morei, 8.2-9.6 percent in Mancalla) and exceeds in
length all but one specimen of Mancalla (the maxi¬
mum of M. diegense). It is, however, 8 percent
longer than the average for M. diegense and 12
percent longer than the average for M. cedrosensis
(Table 1), and 32 percent longer than the much
smaller M. milleri.

The lesser protrusion of the head over the capi¬
tal groove is reflected in the wide, open curve be¬
tween the head and internal tuberosity as seen in
palmar and anconal views; this condition contrasts
with the narrow, U-shaped curve found in Mancalla
(Figure 1 h-l). Further distinction from Mancalla is
seen in the deltoid crest which, in P. ivetmorei, de¬
scribes a low, even arc and is not expanded towards
its distal termination. Distally, the greater breadth
of the humerus is observed not only in the width
through the condyles but also in a slight expansion
in the region of the attachment of the anterior lig¬
ament. In this character, as well as in the lateral
slant of the brachial impression and absence of a
prominent papilla above the condyles, P. wetmorei

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

resembles Praemancalla lagunensis. It is distin¬
guished from that species in the greater projection
of the ectepicondylar prominence from the shaft,
narrower groove between the base of the ectepi-
condyle and external condyle, and relatively nar¬
rower and deeper shaft (relative breadth to depth
of shaft 53 percent in P. wetmorei, 66 percent in P.
laguensis). In depth of shaft relative to breadth of
distal end, P wetmorei is intermediate between
Praemancalla lagunensis and the several species of
Mancalla (99 percent in P. lagunensis, 113 percent
in P wetmorei, 126-140 percent in Mancalla).

The prominence of the olecranon immediately
distinguishes the ulna of P. wetmorei from all spe¬
cies of Mancalla, but the palmad position of the
brachial impression and the shortened lip of the
external cotyla assign the element to the subfamily
Mancallinae rather than the typical alcids.

The radius (LACM 53907), although short and
laterally compressed as in Mancalla, is less blade¬
like and lacks the prominent crest on its convex
contour. The ulnar depression is broader and
deeper than in Mancalla. Neither the ulna nor the
radius is known for Praemancalla lagunensis.
Those assigned to P. wetmorei both show less modi¬
fication towards a flipper-like wing than in Man¬
calla, and in this regard are in keeping with the
character of the other elements known for
Praemancalla.

The radius is 12 percent longer than the maxi¬
mum known for any species of Mancalla (Table
1). Using the radius as a guide, and comparing the
relative size of ulna to radius in the type of Man¬
calla cedrosensis (associated skeletal elements of
one individual), it is suggested that the ulna of P.
wetmorei attained a length of 36.5 mm.

Carpometacarpus LACM 52216 differs from that
of Mancalla and resembles Praemancalla in the
presence of a distinct, blunt pisiform process,
rounded anterior contour of shaft of metacarpal
II and deep narrow groove between the internal
and external crests of the trochlea posteriorly. It is
distinguished from P. lagunensis by the narrower
shaft and the relatively longer process of meta¬
carpal I, with more than half its length distal to the
level of the metacarpal symphysis; also, the troch¬
lea extends higher above that process and the lat¬
eral surface of the internal crest of the trochlea is
more broadly and less deeply depressed.

In coracoid LACM 37637 the furcular facet is

NUMBER 27

145

Table 1.—Skeletal measurements (mm) of Praemancalla wetmorei compared with P. lagunensis,
Mancalla diegense, M. cedrosensis, M. californiensis

Character

P. wetmorei

P. lagun-

M.

diegense

M.

cedrosensis

M. call•

ensis

min.

mean

max.

min.

mean

max.

forniensis*

Humerus

Length .

Greatest proximal

82.7

-

71

76.5

85.2

69.5

73.5

80

-

breadth .

22.2

-

17.3

18.7

20.3

17

17.9

20.1

19.4

Distal breadth

8.5

7.8

6.4

6.7

8

6.9

7

7.2

-

Ulna

Proximal depth ....

11.3

-

9f

9.3

8.8

9.7

10.2

9.9

Proximal breadth

7.5

-

5.9

6.4

6.6

6.2

6.6

7.2

6.5

Radius

Greatest length
Greatest shaft

35.8

-

29.6

30.9

31.8

27.3

29.3

31.1

29.7

depth

5.8

6.3

6.35

6.4

5.4

5.6

6.1

6.8

Shaft breadth .

Carpometacarpus

3

2.3

2.5

2.7

2.5

2.6

2.8

2.3

Length process

Metacarpal I .

15.7

14

15.2

15.3

15.5

15.4

15.6

16

17.1

Proximal depth

12.1

11.9

9.7

10.3

11

10.2

10.5

11.1

11.2

Shaft breadth .

Coracoid

Length from below
scapular facet

4.3

4.5

3.1

3.4

3.7

3.4

3.5

3.8

3.4

to head .

Breadth below

20.9

18.8

15.4

17.3

19.5

15.2f

16.7

18.8

furcular facet

Breadth furcular

5.8

6

5.4

5.8

6.3

5.2

5.5

6

5.9

facet .

10.3

-

7.3

7.8

8.6

7.1+

7.3

8.7

* Only one specimen of each element of M. californiensis, except carpometacarpus (average of
four).

f Only two specimens measurable for this dimension.

broad and deep; it extends ventrally well beyond
the triosseal canal, is strongly thrust mediad above
the canal, and is markedly undercut. Below the
facet the bone narrows and the ventral border of
the triosseal canal is sharply angular. The species
of Mancalla vary in development of the furcular
facet and the bordering of the triosseal canal. The
greatest overhang of the facet and the least angu¬
lar border of the triosseal canal are found in M.
cedrosensis; the least overhang and most angular
border of the canal occur in M. californiensis. In
no specimen of Mancalla is the furcular facet as
ventrally extended as in P. wetmorei. This facet is
broken ventrally in the single known coracoid of
P. lagunensis, but the portion that remains is deep

and has a strong overhang; below the facet, how¬
ever, the area is broader and more rounded than
in P. wetmorei. In direct dorsal view (with dorsal
surface of shaft held horizontally) the scapular
facet in P. wetmorei is more dorsally and less lat¬
erally directed than in Mancalla, and the triosseal
canal faces more mediad. Resemblance is closer to
Praemancalla lagunensis, although the canal is
even more medially directed in the latter species.
The attachment of the coracohumeral muscle in
P. wetmorei is broad, flat, and angular in outline at
its anterior end, as in P. lagunensis, but is rela¬
tively longer, and narrows near the glenoid facet.
In Mancalla the attachment is narrow and
rounded.

146

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Conclusions

Four wing elements and a coracoid from three
correlative localities of late Miocene age in Laguna
Niguel, Orange County, California, represent a
new species, Praemancalla wetmorei, in the alcid
subfamily Mancallinae. The degree of specializa¬
tion towards a flipper-like wing is less than in the
Pliocene genus Mancalla, and reflects a stage of
development closer to the Miocene genus Prae¬
mancalla. Differences noted with respect to Prae¬

mancalla lagunensis, however, suggest a slight
advance towards the more specialized wing of
Mancalla. The humerus is more compressed, the
triosseal canal of the coracoid more dorsally ro¬
tated and narrower, and the process of metacarpal
I of the carpometacarpus longer.

This suggested evolutionary trend is in keeping
with the evidence presented by the associated
faunas of the localities involved, which indicates a
slightly greater age for the type-locality of P.
lagunensis than for the Laguna Niguel sites.

Literature Cited

Barnes, L. G., D. P. Domming, H. Howard, R. W. Huddles¬
ton, and C. A. Repenning

In prep. Correlation and Characterization of Late Mio¬
cene (Clarendonian Correlative) Marine Vertebrate
Assemblages in California.

Howard, Hildegarde

1966. A Possible Ancestor of the Lucas Auk (Family
Mancallidae) from the Tertiary of Orange County,
California. Los Angeles County Museum Contri¬
butions in Science, 101:1-9, 1 figure, 2 tables.

1970. A Review of the Extinct Avian Genus Mancalla.
Los Angeles County Museum Contributions in
Science, 203:1-12, 1 figure, 4 tables.

1971. Pliocene Avian Remains from Baja California. Los

Angeles County Museum Contributions in Science,
217:1-17, 2 figures, 1 table.

Kohl, Roy F.

1974. A New Late Pleistocene Fauna from Humboldt
County, California. Veliger, 17(2):211—219, 2 maps,
1 table.

Lucas, Frederic A.

1901. A Flightless Auk, Mancalla californiensis, from the
Miocene of California. Proceedings of the United
States National Museum, 24(1245): 113-134, 3 figures.

Miller, Loye

1937. An Extinct Puffin from the Pliocene of San Diego,
California. Transactions of the San Diego Society
of Natural History, 8(29): 375-378, 2 figures.

The Pleistocene Pied-billed Grebes
(Aves: Podicipedidae)

Robert W. St over

ABSTRACT

Pleistocene specimens of pied-billed grebes ( Podi-
lymbus) were compared with a series of skeletons
of the modern North American form, Podilymbus
podiceps podiceps. Most of the fossils agreed closely
with this form and are allocated to it. The co-types
of Podilymbus magnus Shufeldt also fall within the
range of variation of this form, hence P. magnus
becomes a synonym of P. podiceps. A new species,
Podilymbus wetmorei, characterized by a wide tar-
sometatarsus and a heavy femur, is described from
the Pleistocene of Florida.

Introduction

The Pied-billed Grebe ( Podilymbus podiceps) is
widely distributed in the New World from Canada
to southern South America. The only other living
species of the genus, the Atitlan or Giant Pied¬
billed Grebe (P gigas), is confined to Lake
Atitlan, Guatemala. The genus is represented in
upper Pliocene deposits of Idaho by a large species,
P. majusculus (Murray, 1967), and in numerous
Pleistocene deposits. Most of the Pleistocene
specimens have been assigned to the living species,
P. podiceps, but a few have been referred to an
allegedly larger extinct species, P. magnus. The
latter was first described by Shufeldt (1913:136-
137) on the basis of two tarsometatarsi and a cora¬
coid from Fossil Lake, Oregon. Later, Wetmore
(1937:198-199) synonymized P. magnus with P.
podiceps, pointing out that there is considerable

Robert W. Storer, Museum of Zoology, The University of
Michigan, Ann Arbor, Michigan, 48104.

sexual dimorphism in the genus and that Shufeldt
had only one skeleton (a female) of the living spe¬
cies with which to compare his fossil material. Wet-
more found that the larger of the tarsometatarsi
described by Shufeldt was only slightly larger than
those of two males of the living North American
subspecies ( P. p. podiceps) and was matched by an
example of the slightly larger South American race
(P. p. antarcticus). More recently, Brodkorb (1959:
273-274) revived the name P. magnus for twelve
bones from Arredondo, Florida, using eight skele¬
tons of the living North American form for com¬
parison. He (1963a: 113) also referred material
from the Santa Fe River, Florida, to P. magnus.
McCoy (1963:337) in his report on the fossil avi¬
fauna of the Itchtucknee River, a tributary of the
Santa Fe, referred two tarsometatarsi to P. magnus
and 47 other bones (including two other tarsometa¬
tarsi) to P. podiceps. Subsequently, Brodkorb
(1963b:230) wrote that “specimens from Fossil
Lake and some of the Floridian localities average
large and are perhaps recognizable as a temporal
subspecies, Podilymbus podiceps magnus Shufeldt.”

The availability of a series of 39 skeletons of the
modern North American form ( Podilymbus p.
podiceps) from Michigan and Wisconsin has per¬
mitted a better estimate of variation within a living
population of this species than was heretofore pos¬
sible, as well as providing a comparison of skeletal
elements of this population with a large number of
fossil elements from late Pleistocene deposits. The
following fossil material has been examined:Cali¬
fornia: McKittrick, 1 tarsometatarsus; Rancho La
Brea, 1 femur; Florida: Reddick, 3 coracoids, 1
humerus, 1 tibiotarsus, 2 tarsometatarsi; Haile, 1
coracoid, 1 ulna, 1 tibiotarsus; Arredondo, 2 cora¬
coids, 3 humeri, 1 ulna, 3 carpometacarpi, 1 femur,

147

148

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Table 1 .- —Measurements (mm) of modern and late Pleistocene Pied-billed Grebe bones

Character

Modern

Fossil

n

max.

min.

mean ± (7 m

a

n

max.

min.

mean±(7m

<7

Coracoid

MALES

Length

23

33.9

30.1

31.98±0.25

1.18

6

33.4

30.9

31.97

-

Width sternal end

32

12.0

10.4

11.28±0.01

0.47

5

12.2

10.7

11.26

-

Least width shaft .

23

3.0

2.4

2.65±0.03

0.14

25

3.0

2.4

2.72±0.03

0.15

Width head .

23

3.5

2.7

3.10±0.05

0.22

9

3.3

2.7

3.04

-

Depth head .

23

5.5

4.6

5.17±0.05

0.24

10

5.5

5.0

5.18±0.05

0.16

Humerus

Length .

23

81.2

74.3

77.63±0.48

2.26

6

77.7

74.4

75.96

-

Width head .

23

14.8

13.0

13.97±0.10

0.46

10

14.3

13.6

13.93±0.07

0.24

Least width shaft .

23

3.3

2.9

3.09+0.03

0.12

27

3.3

3.0

3.10±0.01

0.10

Width distal end .

23

7.5

6.6

7.20±0.05

0.25

26

7.6

6.7

7.12±0.04

0.25

Ulna

Length .

20

76.4

69.1

71.87±0.04

2.03

-

-

-

-

-

Least width

20

2.6

2.2

40

i+

o

o

40

0.10

5

2.6

2.3

2.52

-

Width proximal end

20

6.0

5.3

5.69±0.04

0.18

1

-

-

5.8

-

Depth proximal end .

20

5.8

4.6

4.98+0.06

0.26

1

-

-

5.1

-

Width distal end .

20

4.6

3.9

4.31 ±0.05

0.20

5

4.5

4.0

4.24

-

Depth distal end

20

4.4

3.9

4.20±0.04

0.17

5

4.6

4.3

4.40

-

Carpometacarpus

Length

23

37.6

33.2

34.98±0.28

1.32

4

35.5

34.5

35.05

-

Width head .

23

3.4

2.9

3.21 ±0.03

0.14

6

3.35

3.0

3.13

-

Length metacarpal I .

23

5.5

4.6

4.98±0.05

0.26

6

5

4.9

4.92

-

Depth head .

23

7.0

6.1

6.57 ±0.05

0.23

6

6.9

6.3

6.67

-

Depth distal end

23

4.2

3.4

3.95±0.04

0.19

4

4.1

3.8

4.0

-

Femur

Length

23

42.7

38.9

40.42±0.28

1.29

8

42.4

39.0

41.05

-

Width head .

23

10.0

8.8

9.47±0.08

0.36

9

10.0

9.5

9.74

-

Least width shaft

23

4.0

3.2

3.67±0.04

0.20

15

4.0

3.5

3.76±0.03

0.11

Width distal end

23

10.7

9.2

9.94±0.07

0.32

9

10.7

9.2

9.92

-

Tibiotarsus

Length to articulation

22

75.7

69.8

72.68±0.37

1.70

2

72.0

71.1

71.55

-

Total length .

22

88.7

80.9

84.48±0.43

1.99

1

-

-

84.1

-

Length crest*

22

12.8

11.1

11.80±0.13

0.59

2

12.1

11.8

11.95

-

Width head

23

7.9

6.8

7.30±0.07

0.31

5

7.9

6.9

7.48

-

Width distal end .

23

8

6.9

7.56±0.05

0.23

7

7.8

7.4

7.61

-

Least width shaft

22

4.3

3.6

3.91 ±0.04

0.18

4

4.25

3.7

3.93

-

Least depth shaft .

22

2.7

2.3

2.53±0.03

0.12

5

2.6

2.45

2.53

-

Tarsometatarsus

Length

23

44.2

38.8

41.59±0.30

1.46

11

43.8

40.0

41.91 ±0.32

1.08

Least width shaft

23

3.6

2.95

3.30±0.03

0.17

19

3.6

3.2

3.48±0.03

0.13

Coracoid

FEMALES

Length

15

30.0

27.4

28.47±0.23

0.85

3

28.7

26.8

28.0

—

Width sternal end

14

11.1

9.6

I0.14±0.12

0.44

5

10.5

9.6

10.02

-

Least width shaft

16

2.6

2.2

2.36±0.03

0.11

11

2.5

2.3

2.40±0.02

0.08

Width head

16

3.2

2.6

2.87±0.04

0.15

4

2.9

2.5

2.75

-

Depth head .

16

4.9

4.4

4.68±0.03

0.13

2

4.7

4.6

4.65

-

NUMBER 27

149

Table 1.—Continued

Modern

Fossil

Character

n

max.

min.

mean±CT m

a

n

max.

min.

mean± ffm

a

Humerus

Length .

15

74.3

65.6

69.17±0.63

2.34

8

70.9

66.5

69.62

-

Width head .

15

13.0

11.8

12.28±0.09

0.35

7

12.6

11.9

12.44

-

Least width shaft .

15

3.0

2.6

2.75±0.03

0.12

21

3.0

2.4

2.76±0.03

0.15

Width distal end .

15

6.9

6.3

6.48±0.05

0.18

19

7.2

6.1

6.50±0.07

0.30

Ulna

Length .

16

67.2

61.6

64.61 ±0.04

1.72

1

-

-

62.7

-

Least width .

16

2.3

1.9

2.20±0.03

0.10

4

2.4

2.1

2.26

-

Width proximal end

16

5.4

4.8

5.08±0.05

0.18

1

-

-

5.3

-

Depth proximal end .

16

4.8

4.2

4.43±0.05

0.18

1

-

-

4.5

-

Width distal end .

16

4.1

3.7

3.90±0.04

0.15

1

-

-

3.6

-

Depth distal end .

16

4.1

3.5

3.80±0.04

0.17

3

4.0

3.6

3.82

-

Carpometacarpus

Length .

16

32.8

29.2

31.13±0.27

1.05

1

-

-

33.2

-

Width head .

16

3.0

2.7

2.87±0.03

0.10

1

-

-

2.9

-

Length metacarpal I .

16

4.9

4.0

4.47±0.06

0.24

1

-

-

4.5

-

Depth head .

16

6.1

5.7

5.99±0.04

0.15

1

-

-

6.2

-

Depth distal end

16

3.7

3.3

3.57±0.03

0.13

1

-

-

3.7

-

Femur

Length .

15

39.2

35.2

36.56±0.30

1.08

5

37.3

35.9'

36.8

-

Width head .

15

9.0

8.1

8.54±0.07

0.27

5

8.9

8.3

8.64

-

Least width shaft .

15

3.4

3.0

3.25 ±0.04

0.15

6

3.5

3.1

3.33

-

Width distal end .

15

9.1

8.3

8.68±0.06

0.23

5

9.0

8.5

8.82

-

Tibiotarsus

Length to articulation .

14

68.2

62.1

64.81±0.49

1.85

4

63.9

63.0

63.50

-

Total length .

14

79

72.2

75.18±0.52

1.96

2

73.4

73.2

73.3

-

Length crest* .

15

10.8

9.7

10.25±0.09

0.33

3

10.4

9.6

100.33

-

Width head .

15

6.8

6.2

6.56±0.05

0.18

3

6.9

6.55

6.78

-

Width distal end .

15

7.3

6.3

6.83±0.07

0.28

4

7.1

6.7

6.90

-

Least width shaft .

14

3.8

3.3

3.49±0.04

0.16

5

3.6

3.3

3.44

-

Least depth shaft .

14

2.3

2.1

2.19±0.02

0.08

5

2.25

2.1

2.19

-

T ARSOMETATARSUS

Length .

15

39.0

35.8

37.43±0.27

1.05

5

37.5

36.0

36.98

-

Least width shaft .

15

3.2

2.7

2.92 ±0.03

0.13

9

3.05

2.85

2.98

-

* Calculated by subtracting the length measured from the articulation from the total length.

3 tarsometatarsi; Catalina Lake, 1 coracoid;
Hornsby Spring, 1 femur, 1 tibiotarsus; Itchtucknee
River, 28 coracoids, 58 humeri, 7 ulnae, 7 carpo-
metacarpi, 18 femora, 23 tibiotarsi, 20 tarsometa¬
tarsi; Jenny Spring, 1 humerus; Lake Monroe, 2
humeri, 1 femur, 1 tibiotarsus; Rock Spring, 2
coracoids, 5 humeri, 1 ulna, 6 tibiotarsi; St. John’s
Lock, 3 humeri, 1 ulna, 1 tibiotarsus, 1 tarsometa-
tarsus; Vero Beach, 1 humerus. Nevada: Smith

Creek Cave, 1 coracoid. Oregon; Fossil Lake, 4
tarsometatarsi (including the types of P. magnus).
Total: 220 bones.

Using dial calipers reading to one tenth of a
millimeter, 33 measurements were taken from the
fossil and modern bones (Table 1). Each fossil
specimen was “sexed” by comparing it with the
largest and smallest element for each sex in the
modern sample. There is very little or no overlap

150

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

between the sexes in the total lengths of the ele¬
ments measured, males being larger. In instances
in which a fossil was within a zone of overlap in
length, it was almost always outside a zone of over¬
lap in another measurement. Therefore, I believe
that few, if any, of the fossils were assigned to the
wrong sex and only a few fragmentary specimens
could not be “sexed” in the above manner. Vari¬
ability within the sexes of the modern material and
the fossils “sexed” in this way proved similar, a
further indication that most, if not all of the fossils
were assigned to the correct sex.

The means and, where practical, the standard
errors of the means and standard deviations of the
measurements for these bones were calculated and
compared with those of the modern material
(Table 1).

Acknowledgments. —I am indebted to Pierce
Brodkorb of the University of Florida (PB) and to
the curators of the American Museum of Natural
History (AMNH), the Natural History Museum of
Los Angeles County (LACM), the Museum of
Comparative Zoology (MCZ), the University of
Florida (UF), and the University of Michigan Mu¬
seum of Paleontology for permission to use ma¬
terial under their care. The figure was prepared by
Mark Orsen. Part of this work was subsidized by
the National Science Foundation through Grant
GB-8269.

Discussion

The late Pleistocene material is discussed below
by skeletal element. Unless otherwise indicated, all
specimens in the following sections are from the
Itchtucknee River.

Coracoid.— In both sexes, the means of all meas¬
urements of the modern and Pleistocene coracoids
are very close. One specimen (PB1209 “ 9 ”) falls
outside the expected range (i.e., mean ± 2 a) of
the modern sample in the width of the head (2.5
mm as opposed to 2.6 mm in the smallest modern
specimen).

Humerus. —With five exceptions, the fossil hu¬
meri agree with the modern series in size. A distal
portion (UF 15309) measures 3.0 mm in shaft width
and 6.5 mm in distal width, the former measure¬
ment being near the mean for males and the latter
near the mean for females in the modern series.
(This bone was not included in the calculations of

the fossil sample because it could not be assigned
with confidence to either sex.) Humerus PB 8005
has a shaft width of 3.4 mm, which is approxi¬
mately 2.5 standard deviations above the mean for
males, whereas its other measurements are well
within the range of modern males. (It was not in¬
cluded in the calculations of the fossil sample be¬
cause its identification was not certain; it might
belong to the new species described herein.) UF
15297 has a shaft width of 2.4 mm, three standard
deviations below the mean for modern females, and
a distal width of 6.3 mm, which is within the range
of modern females. UF 15280 and UF 15307 have
distal widths of 7.2 and 7.0 mm, respectively, which
are somewhat outside the expected range of the
modern sample.

Ulna. —One partial ulna (PB 7687 from Rock
Spring) has a distal width of 4.6 mm or 2.35 stand¬
ard deviations above the mean for modern males.
In other measurements, it is within the expected
range of modern specimens, as are the other fossil
ulnae examined.

Carpometacarpus.— The fossil carpometacarpi
are all within the expected range of variation of
the modern sample.

Femur. —With two exceptions (UF 15214 and UF
15220), all the fossil femora fall within the ex¬
pected range of modern specimens. The two excep¬
tions are referred to the new species and are not
included in the calculations of the fossil sample.

Tibiotarsus. —Four fossils are outside the range
of the modern sample. UF 15251 “ $ ” is more than
2 standard deviations above the means for males in
distal width ( 8.2 mm) and least width of shaft
(4.4 mm). Two modern specimens are 8.0 mm in
distal width and one is 4.3 mm in shaft width.) UF
15254 is 68.7 mm in length, measured from the
articulation, which is between the ranges of the
two sexes but slightly nearer the mean of females,
3.7 mm in width of shaft, within the range of either
sex, and 2.4 mm in depth of shaft, within the range
of modern males. MCZ 2606 “ $ ” and PB 1851 “ 9 ”
have cnemial crests well below the range expected
for their assigned sexes. In other measurements,
they are within the range of modern specimens or
differ by only a tenth of a millimeter. These four
fossils were not included in the calculations of
means for the fossil samples.

Tarsometatarsus.— All but three fossils of this
element fall within the expected range of the mod-

NUMBER 27

151

ern sample. Two of these (UF 15223 and PB 1762
from Reddick) are very broad and are referred to
the new species. The third (PB 1854) measures
44.8 mm in length and 3.5 in least width of shaft.
The former measurement is 2.2 and the latter 1.2
standard deviations above the mean for the mod¬
ern series. These three fossils are not included in
the calculations of the fossil sample.

Thus, most of the Pleistocene fossils agree well
with the sample of modern specimens. These fossils
include the type tarsometatarsi of P. mcignus,
which measure 43.6 and 42.9 mm in length and 3.5
and 3.2 mm in least width. I have not examined
the coracoid Shufeldt (1913, pi. 38:fig. 449) tenta¬
tively assigned to P. magnus, but his photograph
of it agrees better with the modern form than two
other coracoids (his figures 461, 462) he did assign
to P. podiceps. Wetmore (1937:199) referred this
specimen to the modern form. The 12 fossils from
Arredondo referred to P. magnus by Brodkorb
(1959:273-274) fall within the expected range of
variation of the modern form. I have not seen the
fossil (or fossils) from the Santa Fe River referred
without comment to P. magnus by Brodkorb
(1963a: 115). The tarsometatarsi from the Itch-
tucknee River referred to P. magnus by McCoy
(1963:337) also fall within the expected range of
the modern form, although one (PB 1854) is 0.6
mm longer than the longest modern specimen meas¬
ured. Both measure 3.5 mm in least width and are
equaled or surpassed by both fossil and modern
specimens in this dimension. Thus, I can find no
evidence that the late Pleistocene birds were sig¬
nificantly larger than the modern ones. Because
the types of P. magnus are indistinguishable from
the modern form, Podilymbus magnus Shufeldt
must be considered a synonym of Podilymbus podi¬
ceps (Linnaeus).

Of the few fossils falling outside the range of the
modern series, several are heavier than the corre¬
sponding elements of P p. podiceps and are estab¬
lished here as a new species.

Podilymbus wetmorei, new species

Figure 1 b,d

Holotype.— Nearly complete left tarsometatar-
sus in the collection of Pierce Brodkorb (PB 1762)
from the Dixie Lime Products quarry, locality IA,

Table 2. —Measurements (mm) of Podilymbus wetmorei

Character

“Male”

“Female”

Tarsometatarsus

UF 15223

PB 1762 (type)

Length .

42.0

36.8

Least width shaft .

3.75

3.5

Femur

UF 15214

UF 15220

Length .

42.4

36.8

Width head .

10.2

8.4

Least width shaft .

42.0

3.7

Width distal end .

10.8

9.0

1.6 km south of Reddick, Marion County, Florida.
Collected by Pierce Brodkorb, 2 March 1957.

Age.— Late Pleistocene (Rancholabrean) fide
Webb (1974:13).

Range.— Known so far only from two localities in
peninsular Florida.

Diagnosis.— Similar in length and general con¬
figuration to the tarsometatarsus of Podilymbus
podiceps , the shaft much heavier (Figure 1; Table
2). Differs from Podilymbus majusculus Murray
(1967), from the upper Pliocene of Idaho, in being
shorter, comparatively wider in the shaft, and in
having the ridge along the external side of the ante¬
rior surface of the bone much reduced for its distal
third. Shorter and relatively heavier than the tarso¬
metatarsus of P. gigas.

Referred Material.— One tarsometatarsus (UF
15223) and two femora (UF 15214 and UF 15220)
all from the Itchtucknee River, Columbia County,
Florida. Like the type, the second tarsometatarsus
has a notably heavy shaft; the femora have thick
shafts and wider heads and distal ends than P.
podiceps (Table 2). Two exceptional humeri men¬
tioned above (UF 15309 and PB 8005 from the
Itchtucknee River) have wide shafts and may like¬
wise represent P. wetmorei. Two tibiotarsi (MCZ
2606 and PB 1851 from the Itchtucknee River)
have short cnemial crests. The former is wide at
both articulations and has a thick (deep) shaft,
suggesting that it may also belong to the new form.
However, the second bone is small in all these di¬
mensions and probably is an aberrant or worn ex¬
ample of P. podiceps. The placement of the other
fossils that fall outside the range of the modern
form is uncertain, and will probably remain so
until more material becomes available.

Remarks.— Having both a thick tarsometatarsus

152

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Figure 1. — Left tarsometatarsi and left femora of species of Podilymbus: a, c, P. podiceps
(UMMZ 151328); b, P. wetmorei, new species (holotype, PB 1762); d, P. wetmorei (UF 15220).
(Scale = 10 mm.)

and a heavy femur is paradoxical in a foot-
propellecl diving bird, because the complex of
structural modifications which appear adaptive for
rapid locomotion under water includes a short,
heavy femur, a long cnemial crest, and a narrow
tarsometatarsus. In the case of P. wetmorei, a thick¬
ened but not shortened femur might have provided
a greater area for the attachment of muscles used
in swimming (notably the gastrocnemius), which
could compensate for the area lost through the
shorter cnemial crest. Modern pied-billed grebes
inhabit bodies of water with considerable emergent
vegetation and have wider tarsometatarsi than
glebes inhabiting more open water. Presumably this
is an adaptation for diving almost directly down¬
ward, breast first, rather than springing forward

like most other grebes. The even wider tarsometa¬
tarsi of P. wetmorei probably represent a further
extension of this adaptive trend already evident in
modern species of Podilymbus.

The presence of two species of Podilymbus in
the same Rancholabrean deposits in Florida sug¬
gests that one (P. wetmorei) may have been a resi¬
dent form that differentiated from the widespread
P. podiceps, which latter may only have wintered
there. The situation between the fossil form known
as Gallinula brodkorbi, also from the Itchtucknee,
and G. chloropus (Olson, 1974:174) suggests an
interesting parallel and might be accounted for
similarly.

Etymology. —I take great pleasure in naming this
species in honor of Alexander Wetmore.

NUMBER 27

153

Literature Cited

Brodkorb, P.

1959. The Pleistocene Avifauna of Arredondo, Florida.
Bulletin of the Florida State Museum, Biological
Sciences, 4(9):269-291.

1963a. A Giant Flightless Bird from the Pleistocene of
Florida. Auk, 80(2): 111-115.

1963b. Catalogue of Fossil Birds, Part 1 (Archaeopterygi-
formes through Ardeiformes). Bulletin of the Flor¬
ida State Museum, Biological Sciences, 7(4): 179-293.
McCoy, J. J.

1963. The Fossil Avifauna of Itchtucknee River, Florida.
Auk, 80(3):335—351.

Murray, B. G., Jr.

1967. Grebes from the Late Pliocene of North America.
Condor, 69(3):277-288.

Olson, S. L.

1974. The Pleistocene Rails of North America. Condor,
76(2): 169-175.

Shufeldt, R. W.

1913. Review of the Fossil Fauna of the Desert Region of
Oregon, with a Description of Additional Material
Collected There. Bulletin of the American Museum
of Natural History, 32(6): 123-178, 35 plates.

Webb, S. D.

1974. Pleistocene Mammals of Florida, x + 270 pages.
Gainesville: University Presses of Florida.

Wetmore, A.

1937. A Record of the Fossil Grebe, Colymbus parvus,
from the Pliocene of California, with Remarks on
Other American Fossils of This Family. Proceed¬
ings of the California Academy of Sciences, series
4, 23(13): 195-201, 15 figures.

The Late Pleistocene Avifauna
of La Carolina, Southwestern Ecuador

Kenneth E. Campbell, Jr.

ABSTRACT

A collection of fossils from the late Pleistocene site
known as La Carolina, located on the arid Santa
Elena Peninsula of southwestern Ecuador, con¬
tains 53 species of birds, representing 16 families
and 42 genera, including 7 extinct species pre¬
viously recorded only from the Talara Tar Seeps
of northwestern Peru. New species of Buteo and
Oreopholus are described. The genus Protocon-
urns Spillman is synonymized with Aratinga.

Seventy-two percent of the species recorded from
La Carolina were also recorded from the Talara
Tar Seeps. The resemblance between the two avi¬
faunas suggests a similarity in age, habitat, and
climatological conditions at the two sites at the
time of deposition. Evidence suggests that during
glacial periods the currently arid Santa Elena Pe¬
ninsula was part of a broad, forested coastal sa¬
vanna extending from central Ecuador south to
northern Peru.

Introduction

While on a visit to Quito, Ecuador, in early 1970,
in connection with my research on the late Pleisto¬
cene avifauna of the Talara Tar Seeps of north¬
western Peru (Campbell, in press), I had the pleas¬
ure of meeting and talking with Professor Gustav
Orc^s V. of the Escuela Polit^cnica Nacional. Dur¬
ing the course of our conversations I learned that
he had in storage at the Escuela a collection of
fossil birds from a site known as La Carolina, lo¬
cated on the Santa Elena Peninsula. The collection

Kenneth E. Campbell, Jr., Department of Zoology, University
of Florida, Gainesville, Florida 32611.

was made by Dr. Robert Hoffstetter in the course of
his work on the Pleistocene mammals of Ecuador
(Hoffstetter, 1952). I was most graciously granted
permission to take the fossil birds back to the Uni¬
versity of Florida where I could work on them in
conjunction with the Talara avifauna.

The La Carolina site is located on the north side
of the Santa Elena Peninsula near the town of La
Libertad (Figure 1). Hoffstetter (1952) and Ed¬
mund (1965) have described the location and geo¬
logical history of this and other nearby sites. Spill¬
man (1942) also contributed to our knowledge of
the paleontology and geological history of the
Santa Elena Peninsula, including descriptions of
what he regarded as two extinct genera and species
of birds.

I have not visited the site personally and must
therefore refer the reader to the description pub¬
lished by Hoffstetter (1952). The fossil beds consist
of fine-grained aeolian and fluviatile sands that
were deposited in a broad, shallow estuarine en¬
vironment. Certain zones of the deposit are im¬
pregnated with pitch and some evidence suggests
that this occurred subsequent to the deposition of
the sands (Edmund, 1965). Preservation of the fos¬
sil material is generally excellent, indicating rapid
burial with little disturbance. In addition to the
specimens recorded below, there are a large num¬
ber of small, delicate specimens of various species
of Passeriformes that are not sufficiently prepared
to include in this report.

Species distributions were taken from Chapman
(1926), Marchant (1958), and Meyer de Schauen-
see (1966). For osteological characters of the
various species the reader is referred to a pending
paper 6n the paleoavifauna of the Talara Tar Seeps
(Campbell, in press). The La Carolina specimens

155

156

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

are currently uncataloged and will be returned to
Professor Orc£s in the near future.

Acknowledgments. —I wish to thank Professor
Gustav Orces V. for his kind permission to study
the avian material from La Carolina. Special
thanks also go to Pierce Brodkorb for his assistance
in the preparation of this manuscript and for ac¬
cess to his skeletal collection. I am very grateful to
the Frank M. Chapman Fund of the American
Museum of Natural History, the Society of Sigma
Xi, and the Organization for Tropical Studies, Inc.
(Pilot Research Proposal No. 69-23) for providing
the crucial financial assistance that made possible
my field studies in Ecuador and Peru. The National
Museum of Natural History and the American
Museum of Natural History kindly loaned speci¬
mens that were critical for identification of some of
the fossil material reported on here. A. Gordon Ed¬
mund of the Royal Ontario Museum provided nu¬
merous useful comments.

ARDEIDAE

Nycticorax nycticorax (Linnaeus)

Material.— One complete right coracoid, hu¬
meral end of 1 left coracoid, 1 complete left carpo-
metacarpus, proximal ends of 1 right and 1 left
femur.

Remarks.— N. nycticorax is resident in Ecuador.

ANATIDAE

Dendrocygna autumnalis (Linnaeus)

Material.— One left scapula, 1 complete right
coracoid, proximal end of 1 right ulna, distal end
of 1 left tibiotarsus.

Remarks.— D. autumnalis is resident in Ecuador.

Anatidae genus and species

Material. —-Two complete right and 2 complete
left coracoids, humeral end of 1 right coracoid, 1
complete left ulna, 1 complete left radius, 1 com¬
plete right carpometacarpus, proximal ends of 1
right and 1 left carpometacarpus.

Remarks. —This new genus and species of Ana¬
tidae is being described as a form of Tadorninae

Figure 1.—Location of the La Carolina fossil site, south¬
western Ecuador, relative to the Talara Tar Seeps of north¬
western Peru. The dotted line marks the approximate
eastward limit of the arid zone of the Santa Elena Peninsula.

from the Talara Tar Seeps of northwestern Peru
(Campbell, in press). The presence of this and the
other entinct species listed below indicate that
paleoclimatic conditions at La Carolina were simi¬
lar to those at the Talara Tar Seeps, much as
similar conditions exist at the two sites today.

Anas species 1

Material.— One complete right coracoid, prox¬
imal end of 1 right radius, distal end of 1 right
tibiotarsus, 1 complete left tarsometatarsus.

Remarks. —The proximal end of the radius is not
included in the description of Anas species 1, from
the Talara Tar Seeps (Campbell, in press). The
present specimen is referred to this species on the
basis of size (proximal width, 4.4 mm), and be¬
cause it differs from the radius being referred to
Anas species 2 (Campbell, in press) by having:
(1) ligamental prominence high, (2) ulnar depres¬
sion deep, (3) scapholunar facet broad and flat.

Anas species 3

Material. —One left scapula, 1 complete left hu¬
merus (holotype), 2 complete left carpometacarpi.

NUMBER 27

157

Remarks. —The complete humerus from La
Carolina is being used as the holotype in the de¬
scription of Anas species 3 (Campbell, in press),
because it is in better condition than any of the
material from the Talara Tar Seeps. Elements of
this species not found in the Talara Tar Seeps in¬
clude the scapula and carpometacarpus. The small
size of these specimens indicates that they may be
safely referred to Anas species 3. In addition to its
small size, the scapula is characterized by having:
(1) acromion of moderate length and width, with
anteroventral corner a prominent projection; (3)
attachment of Lig. furculoscapulare dorsale oval
in shape, located on midline of shaft. Measure¬
ments of the scapula are: proximal height, 7.7
mm; proximal width, 2.4 mm.

The carpometacarpus is characterized by hav¬
ing: (1) process of metacarpal I moderately high
and long; (2) proximal edge of metacarpal I slop¬
ing slightly proximad; (3) proximal metacarpal
symphysis of moderate length, distal symphysis
long; (4) internal face of external rim of carpal
trochlea angular proximally; (6) area covered by
external cuneiform ligament moderately convex;
(7) posterior carpal fossa of moderate width for its
length; (8) surface of carpal trochlea extending
deep into posterior carpal fossa; (9) internal rim
of carpal trochlea with posterior edge bowing ex¬
ternally. Measurements (in mm) of the two carpo-
metacarpi are as follows: length, 35.0 and 35.2;
height through metacarpal I, 8.1 and broken;
proximal width, 3.9 and 3.9; least width of shaft,
2.5 and 2.7; length of distal fornix, 5.0 and 4.9.

The numbered characters listed above corre¬
spond to those being used in the descriptions of
Anas species 1, and Anas species 2 (Campbell, in
press). The absence of Anas species 2, from La
Carolina may be a result of the small sample size,
although it was more common at the Talara Tar
Seeps than Anas species 3.

Anas bahamensis Linnaeus

Material. —Portions of 12 crania, 9 right and 4
left scapulae, 23 complete right and 27 complete
left coracoids, humeral ends of 5 right and 7 left
coracoids, 4 complete right and 2 complete left
humeri, proximal ends of 1 right and 2 left humeri,
distal ends of 1 right and 4 left humeri, 16 com¬
plete right and 18 complete left ulnae, proximal

ends of 3 right and 3 left ulnae, distal ends of 2
right and 3 left ulnae, 11 complete right and 3 com¬
plete left radii, proximal ends of 7 right and 4 left
radii, distal ends of 3 right and 5 left radii, 28 com¬
plete right and 34 complete left carpometacarpi,
proximal ends of 3 left carpometacarpi, 5 complete
right and 8 complete left femora, proximal ends of
2 left femora, distal ends of 3 left femora, 3 com¬
plete right and 1 complete left tibiotarsus, 1 al¬
most complete right and 1 almost complete left
tibiotarsus, proximal ends of 1 right and 1 left
tibiotarsus, distal ends of 4 right and 3 left tibio-
tarsi, 19 complete right and 24 complete left tarso-
metatarsi, proximal ends of 2 right and 4 left
tarsometa tarsi.

Remarks. —In terms of numbers of specimens
and of individuals, A. bahamensis is the second
most abundant species in the La Carolina deposits.
This species is currently found in large numbers in
southwestern Ecuador when standing bodies of
fresh water occur (Marchant, 1958).

I am elsewhere (Campbell, in press) expressing
my belief that it was material of this species that
Spillman (1942) named Archeoquerquedula lam-
brechti. The absence of any large, extinct duck
from the La Carolina deposits upholds the view
that Archeoquerquedula lambrechti should be
synonymized with Anas bahamensis.

VULTURIDAE

Vulturidae genus and species indeterminate

Material.— Proximal end of 1 right tarsometa-
tarsus.

Remarks.— This specimen is too broken to be
identified further. It represents a condor the size of
a large individual of Vultur gryphus Linnaeus.
The intercotylar prominence is very broad and
low, unlike that found in either Vultur Linnaeus,
Gymnogyps Lesson, or the new genus of condor
being described from the Talara Tar Seeps (Camp¬
bell, in press). I am convinced the specimen rep¬
resents a new genus and species, but consider the
specimen too fragmentary to describe.

Coragyps cf. atratus (Bechstein)

Material. —Distal ends of 1 right and 1 left
carpometacarpus.

158

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Remarks. —These specimens resemble those from
the Talara Tar Seeps in differing slightly from the
subspecies of C. atratus currently resident in south¬
western Ecuador and northwestern Peru.

Cathartes aura (Linnaeus)

Material.— One right scapula, proximal end of
1 left radius, 1 complete left tarsometatarsus, distal
ends of 2 left tarsometatarsi.

Remarks.—C. aura is resident in southwestern
Ecuador.

ACCIPITRIDAE

Accipitridae genus and species

Material.— Proximal end of 1 left femur.

Remarks. —This extinct new genus and species
of large eagle is being described from the Talara
Tar Seeps (Campbell, in press). Hoffstetter (1952:
40) mentioned an enormous eagle from La Caro¬
lina. The specimens he was referring to probably
belong to this species.

Geranoaetus melanoleucus (Vieillot)

Material.— Proximal end of 1 right carpometa-
carpus, distal end of 1 left tibiotarsus.

Remarks.— G. melanoleucus is resident in west¬
ern Ecuador today.

Buteo polyosoma (Quoy and Gaimard)

Material.— Proximal ends of 1 right and 1 left
humerus, proximal end of 1 right radius, 1 com¬
plete left carpometacarpus, distal end of 1 right
tibiotarsus, 2 complete left tarsometatarsi.

Remarks.— B. polyosoma is a common resident
in southwestern Ecuador.

Buteo hoffstetteri, new species

Figure 2

Holotype.— Right tarsometatarsus lacking me¬
dial portion of proximal end. Uncataloged.

Paratype. —Almost complete left tarsometatarsus.

Diagnosis.— Tarsometatarsus agrees with that of

Buteo Lac^pede and differs from that of all other
South American genera of Accipitridae by having
those characters of the genus Buteo as listed by
Campbell (in press).

Tarsometatarsus characterized by having: (1) in¬
ternal cotyla very concave (moderately concave in
B. lineatus, very concave in B. polyosoma ); (2) an¬
terior metatarsal groove very deep immediately
distal to intercotylar prominence and distal to tu¬
bercle for tibialis anticus (moderately deep in both
areas in B. lineatus and B. polyosoma)-, (3) shaft
narrow abruptly distal to tubercle for tibialis anti¬
cus (narrows gradually in B. lineatus, abruptly in
B. polyosoma ); (4) shaft with posterior metatarsal
groove moderately deep (very deep in B. lineatus,
moderately deep in B. polyosoma); (5) inter-

trochlear notches wide (narrow in B. lineatus and
B. polyosoma ); (6) internal trochlea with very

prominent angular proximolateral projection (not
as prominent in B. lineatus, very prominent in
B. polyosoma ); (7) external trochlea narrow,
short, and projecting posterodistad at approxi¬
mately 45° to axis of shaft (wide, long, and at
approximately 85° to axis of shaft in B. lineatus
and B. polyosoma ); (8) shaft very slender (similar
in B. lineatus, wider in B. polyosoma).

Measurements (in mm) of the holotype and
paratype (in parentheses) are as follows: length,
74.6 (76.3); proximal width, broken (11.1 ± 1.0);
distal width, 10.7 (12.2 ± 0.5); least width of shaft,
4.0 (4.2).

Referred Material.— Distal end of 1 left and 1
right tibiotarsus.

Characters.— Tibiotarsus characterized by hav¬
ing: (1) shaft only slightly concave at postero-
proximal end of internal condyle (moderately to
deeply concave in B. lineatus and B. polyosoma);
(2) internal condyle short (long in B. lineatus, of
moderate length in B. polyosoma); (3) external
condyle merging gradually with shaft antero-
proximally, not forming a lip (merges abruptly,
forming a lip in B. lineatus and B. polyosoma).
The only measurement that can be taken is the
distal width (10.6 mm) of one specimen.

Etymology.— This species is named for Dr.
Robert Hoffstetter of the Museum National d’His-
toire Naturelle, Paris, France, in recognition of his
paleontological work in Ecuador and his role in
the preservation of the La Carolina fossils.

Remarks. —Although Buteo lineatus does not oc-

NUMBER 27

159

« »

Figure 2.—Holotype right tarsometatarsus of Buteo hoffstet-
teri, new species (uncataloged), in anterior, external, and
posterior view, X 1.

cur in South America, it was used in the diagnosis
because its tarsometatarsus resembles that of B.
hofjstetteri in being long and slender rather than
short and stout. While the tarsometatarsus of B.
polyosoma is the same length as that of B. linecitus,
it is much more robust. In all other species of
Buteo with tarsometatarsi of similar length, the
bone is even more robust than in B. polyosoma and
these species must therefore also differ from B.
hoffstetteri.

Circus cinereus Vieillot

Material. —Proximal end of 1 left carpometa-
carpus, 1 complete right femur, distal end of 1 left
tibiotarsus.

Remarks. — C. cinereus is resident in western
Ecuador, but has not been recorded from the Santa
Elena Peninsula.

FALCONIDAE

Falco peregrinus Tunstall

Material.— Two left scapulae, sternal end of 1
right coracoid, proximal end of 2 left humeri, dis¬
tal end of 1 right humerus, 1 complete left ulna,

proximal end of 1 right ulna, distal ends of 2 left
ulnae, proximal end of 1 right radius, 2 complete
right carpometacarpi, proximal ends of 1 right and

1 left carpometacarpus, distal end of 1 right carpo-
metacarpus, 1 complete left femur, distal end of 1
right tibiotarsus, 1 complete right tarsometatarsus,
proximal ends of 2 left tarsometatarsi.

Remarks. — F. peregrinus occurs in Ecuador both
as a northern and a southern migrant.

Falco femoralis Temminck

Material.— One complete left coracoid, proxi¬
mal end of 1 left tibiotarsus.

Remarks. — F. femoralis is resident in western
Ecuador.

Polyborus plancus (Miller)

Material. —Five right and 4 left scapulae, 2 com¬
plete right and 2 complete left coracoids, humeral
ends of 4 right and 1 left coracoid, proximal ends
of 3 right and 1 left humerus, distal ends of 1 right
and 1 left humerus, 1 complete right ulna, proxi¬
mal ends of 2 left ulnae, distal ends of 3 right and

2 left ulnae, 1 complete right and 3 complete left
carpometacarpi, proximal ends of 2 left carpometa¬
carpi, proximal ends of 2 left femora, distal end of
1 right femur, distal ends of 2 right and 1 left tibio¬
tarsus, proximal end of 1 left tarsometatarsus, dis¬
tal ends of 3 right and 4 left tarsometatarsi.

Remarks. —These specimens differ from the sub¬
species presently found in southwestern Ecuador
and northwestern Peru in the same way as do the
specimens from the Talara Tar Seeps. The com¬
plete right carpometacarpus is considerably lighter
in build than the rest of the carpometacarpi from
this site as well as those from the Talara Tar Seeps
and is perhaps sufficiently different to warrant con¬
sideration as a distinct species. However, the great
variation found in Polyborus makes it unwise to
describe a new species without a larger sample. P.
plancus is common in southwestern Ecuador.

Milvago species

Material. —Two left scapulae, 1 complete left
carpometacarpus.

Remarks. —This species is being described from

160

the Talar a Tar Seeps (Campbell, in press). No
species of Milvago presently occurs west of the
Andes Mountains in Peru or Ecuador.

CHARADRIIDAE

Pluvialis dominica (Statius Muller)

Material. —One complete right and 3 complete
left coracoids, 8 complete right and 4 complete left
humeri, distal ends of 1 right and 2 left humeri, 5
complete right and 5 complete left carpometacarpi,
2 complete left femora, proximal ends of 2 right
tibiotarsi, distal ends of 3 right and 1 left tibio-
tarsus, 1 complete right and 5 complete left tarso-
metatarsi, proximal ends of 1 right and 1 left
tarsometatarsus, distal ends of 2 right tarsometa-
tarsi.

Remarks. —As North American migrants, P.
dominica and the following 3 species of plovers
occur as seasonal visitors or residents along the
coast of Ecuador.

Pluvialis squatarola (Linnaeus)

Material.— Proximal end of 1 right humerus.

Charadrius vociferus Linnaeus

Material. —Proximal end of 1 left tibiotarsus.

Charadrius semipalmatus Bonaparte

Material.— One complete left humerus, distal
end of 1 left tibiotarsus.

Oreopholus orcesi, new species

Figure 3

Holotype.— Complete left femur. Uncataloged.

Diagnosis. —Femur agrees with that of Oreo¬
pholus and differs from that of all other genera of
South American plovers (with possible exception
of Zonibyx Reichenbach, Pluvianellus Gray, and
Phegornis Gray, which were not available for com¬
parison) by having: (1) head disc-shaped, project¬
ing mediad, perpendicular to shaft, without pro¬
truding dorsad above level of flat iliac facet; (2)

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Figure 3.—Holotype left femur of Oreopholus orcesi, new
species (uncataloged), in anterior, external, and posterior
views, X 2.

attachment of M. flexor ischiofemoralis elevated
along anterior edge; (3) shaft with proximal two-
thirds straight, and marked anteroposterior flexure
in distal third, in lateral view; (4) shaft with distal
end turned sharply mediad posterior to antero-
proximal termination of internal condyle, in ante¬
rior view; (5) internal condyle with posteroproxi-
mal corner projecting proximad more than external
condyle; (6) shaft very wide immediately proximal
to condyles.

Femur differs from that of Oreopholus ruficollis
(Wagler) by having: (1) head much smaller; (2)
attachment of M. iliacus marked by sharp, high
ridge along corner of shaft (ridge absent and at¬
tachment much larger in area in O. ruficollis)-,
(3) shaft with greater anteroposterior flexure; (4)
shaft with marked concavity at point of flexure me¬
dial to internal condyle, resulting in internal
condyle terminating proximally in high, narrow
ridge (not as excavated in O. ruficollis, but with
internal condyle more rounded); (5) internal
condyle deeper anteroposteriad, with posterior
articular surface flat (convexity present on surface
in O. ruficollis); (6) external condyle projecting
less proximad, and at a greater angle to axis of
shaft, in posterior view; (7) attachment of M.
gastrocnemius, pars externa, smaller in area, more
elevated, and positioned more posteriad; (8) rotu-
lar groove narrower and deeper.

Measurements (in mm) of the holotype, with

NUMBER 27

161

those of one specimen of O. ruficollis (in paren¬
theses), are as follows: length from external con¬
dyle to iliac facet, 29.7 (30.0); width of proximal
end, 5.9 (6.2); width of distal end, 6.0 (6.1); least
width of shaft, 2.5 (2.4).

Etymology. —This species is named for Pro¬
fessor Gustav Orces V., of the Escuela Politecnica
Nacional, Quito, Ecuador, for his pioneering work
in the natural history of Ecuador.

Remarks. —Although only one specimen of O.
ruficollis was available for comparison, the major¬
ity of the characters cited above are considered
unlikely to vary with a larger sample. The species
of Zonibyx, Pluvianelhis, and Phegornis are all
much smaller than O. orcesi. In the event that the
species described here does not belong in Oreo-
pholus, but in one of the three genera listed above,
it would still represent a new species, if based solely
on size.

SCOLOPACIDAE
Tringa solitaria Wilson

Material.— Proximal end of 1 left tibiotarsus,
proximal end of 1 left tarsometatarsus, distal end
of 1 left tarsometatarsus.

Remarks. —As North American migrants, T.
solitaria and the following 12 species of the family
Scolopacidae occur as seasonal visitors or residents
along the coast of Ecuador.

Totanus flavipes (Gmelin)

Material.— One complete right coracoid, 1 com¬
plete right and 2 complete left humeri, 1 complete
right and 1 complete left carpometacarpus, shaft
and proximal end of 1 right femur, distal ends of
3 right and 1 left tarsometatarsus.

Remarks.—I consider the osteological differences
between Tringa and Totanus as sufficient to sepa¬
rate them at the generic level.

Totanus melanoleucus (Gmelin)

Material.— Humeral end of 1 right coracoid, 1
complete left humerus, proximal end of 1 right
humerus, distal end of 1 left humerus, 1 complete
left carpometacarpus.

Catoptrophorus semipalmatus (Gmelin)

Material.— Distal ends of 2 right and 1 left hu¬
merus, 3 complete right and 3 complete left carpo-
metacarpi, proximal end of 1 left carpometacar¬
pus, shaft and proximal end of 1 right femur, distal
end of 1 right tibiotarsus, proximal end of 1 right
tarsometatarsus.

Calidris canutus (Linnaeus)

Material. —One complete left coracoid, 6 com¬
plete right and 6 complete left humeri, 5 almost
complete right and 5 almost complete left humeri,
proximal ends of 2 right and 3 left humeri, distal
ends of 4 right and 5 left humeri, 5 complete right
and 3 complete left carpometacarpi, proximal ends
of 2 right carpometacarpi, distal ends of 7 right
and 3 left tibiotarsi.

Erolia melanotos (Vieillot)

Material.— Three complete right and 8 com¬
plete left humeri, 1 almost complete right humerus,
proximal ends of 1 right and 1 left humerus, distal
end of 1 left humerus, 1 complete right and 2 com¬
plete left carpometacarpi, 1 complete right femur,
distal end of 1 left femur, distal end of 1 right
tibiotarsus, 1 complete left tarsometatarsus.

Remarks. —I consider the osteological differences
between Calidris, Erolia, and Ereunetes as suffi¬
cient to separate them as three distinct genera.

Ereunetes pusillus (Linnaeus)

Material. —One complete right and 1 complete
left humerus, distal end of 1 right humerus.

Ereunetes mauri (Cabanis)

Material.— Two complete right and 4 complete
left humeri, distal ends of 1 right and 3 left humeri.

Micropalama himantopus (Bonaparte)

Material. —One complete right femur.

Bartramia longicauda (Bechstein)

Material.— One complete right humerus, distal

162

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

end of 1 right humerus, 1 complete right and 1
complete left carpometacarpus, distal ends of 2 left
tibiotarsi, proximal end of 1 left tarsometatarsus,
distal end of 1 left tarsometatarsus.

Numenius phaeopus (Linnaeus)

Material.— Two complete left coracoids, 1 com¬
plete right carpometacarpus, distal end of 1 left
tibiotarsus.

Limosa fedoa (Linnaeus)

Material. —One complete right and 1 fragmen¬
tary left carpometacarpus.

Arenaria interpres (Linnaeus)

Material.— Two complete left and 1 complete
right coracoid, humeral end of 1 right coracoid,
proximal end of 1 left carpometacarpus, 1 com¬
plete left femur, 1 complete left tarsometatarsus.

Scolopacidae genus and species

Material. —One complete left coracoid, 1 com¬
plete right femur.

Remarks. —A new genus and species of Scolo¬
pacidae is being described from the Talara Tar
Seeps (Campbell, in press) on the basis of a cora¬
coid and a tarsometatarsus. The coracoid from La
Carolina agrees in all characters with the holotype
coracoid, differing only in being more robust. Meas¬
urements (in mm) of the La Carolina coracoid
are as follows: length, 17.3; head to scapular facet,
6.7; depth of proximal end, 4.9; least width of
shaft, 2.4; length of sternal facet, 5.8.

The above femur agrees with the other elements
assigned to the new genus in superficially resem¬
bling Limnodromus more than any other genus of
scolopacid. The femur differs from that of L. gris¬
eus by having: (1) neck longer and much more
constricted; (2) head directed more anteriorly;

(3) shaft flexed anteriad at level of attachment of
M. iliacus, in lateral view (straight in L. griseus)-,

(4) shaft curved gradually posteriad in distal half,
in lateral view (curvature more localized in L.
griseus)-, (5) attachment of M. gastrocnemius, pars
externa, more elevated, facing more posteriad, but

lying more on corner of shaft, farther from base of
external condyle; (6) external condyle more dis¬
tinctly set off at base posteriorly. The internal
condyle is too broken to yield any characters.

Measurements (in mm) of the femur are as fol¬
lows: length from iliac facet to external condyle,
28.9; width of proximal end, 5.3; width of distal
end, 4.8; least width of shaft, 2.2; depth of proxi¬
mal end, 4.0.

KECURVIROSTRIDAE

Himantopus mexicanus (Statius Muller)

Material.— One complete right coracoid.

Remarks. — H. mexicanus is generally distributed
along the coast of Ecuador.

PHALAROPODIDAE

Phalaropus fulicarius (Linnaeus)

Material.— One complete right humerus, shaft
and distal ends of 2 right humeri, 2 complete left
carpometacarpi.

Remarks. —As North American migrants, P. fuli¬
carius and the following two species of phalaropes
occur in western Ecuador as seasonal visitors or
residents.

Lobipes lobatus (Linnaeus)

Material.— Two complete right and 6 complete
left humeri, 1 almost complete left humerus, 1
complete right and 4 complete left coracoids,
proximal ends of 1 right and 1 left humerus, distal
ends of 1 right and 2 left humeri, 2 complete right
and 2 complete left carpometacarpi, proximal end
of 1 right carpometacarpus, 1 complete right and
3 complete left tarsometatarsi.

Steganopus tricolor Vieillot

Material.— Five complete right and 4 complete
left coracoids, 19 complete right and 20 complete
left humeri, 3 almost complete right and 2 almost
complete left humeri, proximal ends of 1 right and
2 left humeri, distal ends of 3 right and 7 left hu¬
meri, 5 complete right and 9 complete left carpo-

NUMBER 27

163

metacarpi, 2 complete right and 2 complete left
femora, distal ends of 6 right and 5 left tibiotarsi,
2 complete right tarsometatarsi, 1 almost complete
right and 1 almost complete left tarsometatarsus,
proximal ends of 1 right and 1 left tarsometatarsus,
distal end of 1 right tarsometatarsus.

metatarsus of the two species of Steganopus, see
Table 1.

BURHINIDAE

Burhinus superciliaris (Tschudi)

Steganopus species

Material.— One complete right and 1 complete
left humerus, distal end of 1 left humerus, 1 com¬
plete right tarsometatarsus.

Remarks. —This species is being described from
the Talara Tar Seeps on the basis of a complete
femur (Campbell, in press). As the above elements
differ significantly from S. tricolor they are here
referred to the new species.

The humerus of the new Steganopus differs
from that of S. tricolor by having: (1) median
crest larger, less excavated; (2) internal tuberosity
larger, more rounded, directed less dorsad, and
projecting less anconally; (3) head not undercut
by capital shaft groove; (4) attachment of M.
proscapulohumeralis positioned closer to base of
median crest; (5) attachment of M. triceps, exter¬
nal head, larger dorsoventrad and bordered ven-
trally by distinct lip of bicipital crest, resulting in
a larger bicipital crest; (6) deltoid crest longer,
thicker, and higher; (7) impression of M. brachi-
alis anticus shallower, not as deeply inset at the
edges; (8) attachment of anterior articular liga¬
ment shorter and higher; (9) attachment of M.
pronator brevis positioned more distad; (10) in¬
ternal and external condyles longer; (11) shaft
much larger.

The tarsometatarsus of the new species differs
from that of S. tricolor by having: (1) internal
trochlea larger, but projecting less posteriad; (2)
internal trochlea less excavated adjacent to middle
trochlea; (3) middle trochlea more rounded pos¬
teriad in medial view; (4) internal trochlea posi¬
tioned close to, and directed more toward,
midline of shaft. The proximal ends of the tarso¬
metatarsi of Steganopus apparently do not ossify
completely except in older individuals. Although
the present specimen showed only a slight indica¬
tion of pitting, no reliable characters could be ob¬
tained from the proximal end.

For measurements of the humerus and tarso-

Material.— Proximal end of 1 left humerus, dis¬
tal ends of 1 right and 3 left humeri, 1 complete
right and 3 complete left coracoids, proximal end
of 1 left coracoid, distal end of 1 right scapula, dis¬
tal end of 1 right ulna, 1 complete right and 2

Table 1.—Measurements (mm) of the humerus and tarso¬
metatarsus of Steganopus (OR = Observed range, M =
Mean, N = Number of specimens)

Character and
measurement

Steganopus

species

Steganopus

tricolor

Recent

Steganopus
tricolor
La Carolina

Humerus

Length
OR

M .

N .

Width of proximal end

OR .

M .

N .

Width of distal end

31.6-32.5

32

2

7.1-7.4
7.3
2

32.1-38

34.6

8

6.9-8.2
7.4
8

31.6-37.1

34.3

35

6 . 8-8

7.3

35

OR

5-5.1

4.5-5.6

4.7-5.6

M .

5.1

4.9

5.1

N .

2

8

35

Least width of shaft

OR .

2.4

2-2.3

2-2.6

M .

2.4

2.2

2.3

N .

2

8

35

Tarsometatarsus

Length

OR

30.6

31-34.7

32.8-34.5

M .

-

32.2

33.6

N .

1

8

2

Width of proximal end

OR .

3.5

3.2—3.7

3.5-3.8

M .

-

3.4

3.6

N .

1

8

4

Width of distal end

OR .

3.2

3-3.3

3.3-3.5

M .

-

3.2

3.4

N .

1

8

2

Least width of shaft

OR.

1.2

1.2-1.3

1.3

M .

-

1.2

1.3

N .

1

8

4

164

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

complete left carpometacarpi, 1 complete left
femur, distal end of 1 left tibiotarsus, distal end of
1 right tarsometarsus.

Remarks. — B. superciliaris is common in the
semi-arid region of Ecuador.

THINOCORIDAE

Thinocorus rumicivorus Eschscholtz

Material.— One complete right and 1 complete
left humerus, 1 complete right and 1 complete left
carpometacarpus.

Remarks. — T. rumicivorus is resident in the arid
regions of Ecuador.

LARIDAE

Lams pipixcan Wagler

Material.— One complete right carpometacar¬
pus, proximal end of 1 right carpometacarpus.

Remarks. —A North American migrant, L. pi¬
pixcan is a seasonal resident of Ecuador.

Larosterna inca (Lesson)

Material.— Humeral end of 1 left coracoid.

Remarks.— A small, shallow attachment of Lig.
humerocoracoideum anterius superius, located on
the anteroventral corner of the head of the cora¬
coid, is characteristic of Larosterna. This is the first
fossil record of the genus. L. inca is found along
the coast of southern Ecuador.

COLUMBIDAE

Zenaida auriculata (Des Murs)

Material.— Two right and 2 left scapulae, 1 com¬
plete right and 8 complete left coracoids, humeral
ends of 1 right and 1 left coracoid, 1 complete right
and 6 complete left humeri, proximal ends of 2
right and 3 left humeri, distal ends of 1 right and
1 left humerus, 4 complete right and 5 complete
left ulnae, 2 complete right and 2 complete left
radii, 5 complete right and 6 complete left carpo¬
metacarpi, 1 complete right and 1 complete left
femur, 1 complete right and 1 complete left tibio¬

tarsus, distal end of 1 right tibiotarsus, 4 com¬
plete right and 1 complete left tarsometatarsus,
shaft and proximal end of 1 left tarsometatarsus.

Remarks.— Z. auriculata is resident in western
Ecuador.

Zenaida asiatica (Linnaeus)

Material.— Proximal end of 1 left carpometa¬
carpus.

Remarks.— Z. asiatica is resident in southwest¬
ern Ecuador.

Columbina cruziana (Knip and Prevost)

Material.— One complete right humerus.

Remarks.— C. cruziana is common in south¬
western Ecuador.

PSITTACIDAE

Aratinga roosevelti (Spillman)

Material.— Portions of 11 crania, 12 premaxil-
laries, 14 mandibles, 24 right and 17 left scapulae,
35 complete right and 46 complete left coracoids,
humeral ends of 2 left coracoids, 25 complete right
and 24 complete left humeri, proximal end of 1
right humerus, distal ends of 3 left humeri, 55 com¬
plete right and 56 complete left ulnae, proximal
ends of 1 right and 1 left ulna, distal end of 1 left
ulna, 30 complete right and 28 complete left radii,
proximal ends of 5 right and 1 left radius, distal
ends of 2 right and 2 left radii, 85 complete right
and 71 complete left carpometacarpi, proximal
ends of 3 right and 2 left carpometacarpi, shafts of
2 left carpometacarpi, 29 complete right and 22
complete left femora, distal ends of 1 right and 1
left femur, 36 complete right and 24 complete left
tibiotarsi, proximal ends of 1 right and 1 left tibio¬
tarsus, distal ends of 4 right and 1 left tibiotarsus,
21 complete right and 19 complete left tarsometa-
tarsi, proximal ends of 2 left tarsometatarsi, distal
ends of 3 left tarsometatarsi.

Remarks. —Spillman (1942) reported a large col¬
lection of parrot fossils from a site on the Santa
Elena Peninsula near La Carolina. He placed this
material in a new genus and species, Protoconurus
roosevelti. From his descriptions and illustrations

NUMBER 27

165

it is reasonably certain that the present material
belongs to the same species. I could not detect
sufficient differences between this fossil material
and Recent specimens of the genus Aratinga to
warrant the placement of the former in a different
genus. Therefore, I suggest that Protoconurus be
synonymized with Aratinga.

It is not possible at this time to say whether the
fossil material represents an extinct species, as
Spillman believed, or an extant species. This can¬
not be determined until a satisfactory series of
comparative material is available. The fossil ma¬
terial is slightly larger than the one available speci¬
men of A. erythrogenys from southwestern Ecuador
and differs from that species in many ways.

One coracoid of those listed above is signifi¬
cantly smaller than the others and may represent
a different species.

Aratinga species

Material.— One complete left coracoid, distal
end of 1 right humerus.

Remarks. —These two specimens appear to rep¬
resent a species of Aratinga, but may possibly rep¬
resent another genus, as not all genera of Ecua¬
dorian parrots were available for comparison. They
are considerably larger than the corresponding
elements of A. erythrogenys.

TYTONIDAE
Tyto alba (Scopoli)

Material. —One complete right coracoid.

Remarks. — T. alba is resident in western
Ecuador.

STRIGIDAE

Speotyto cunicularia Gloger

Material.— One complete right and 1 complete
left carpometacarpus, distal ends of 2 right tibio-
tarsi, distal end of 1 left tarsometarsus.

Remarks.— S. cunicularia is resident in western
Ecuador.

Discussion

Ecological Considerations. —La Carolina is lo¬
cated in the most arid region of Ecuador. Vegeta¬
tion is very sparse, except in the larger dry river
valleys. Ecological and climatological conditions at
the site are very similar to those found at the
Talara Tar Seeps, except that La Carolina is much
closer to the coastline and receives slightly more
rainfall. The major difference between the arid
region of Ecuador and that of coastal Peru is in
the greater area of the latter. Whereas the Peru¬
vian coastal desert is thousands of kilometers long,
the Ecuadorian desert is essentially limited to the
Santa Elena Peninsula (Figure 1). The transition
from barren desert to heavy forest is very abrupt, in
some places occurring over only a few kilometers.

The principal references concerning the modern
avifauna of southwestern Ecuador are the works of
Chapman (1926) and Marchant (1958, 1959,
1960). A point stressed by both authors is the ra¬
pidity with which the arid and semi-arid regions
are colonized by birds from the adjacent forests
following the rare periods of rainfall. Marchant
also documents how large numbers of birds are at¬
tracted to standing bodies of fresh water on the
peninsula, such as an artificial lake.

The importance of the La Carolina avifauna in
interpreting the paleoecology of southwestern
Ecuador is difficult to assess. This is because the
desert region is so small, the transition to forest so
sharp, and the speed of colonization under the
proper conditions so rapid, that only a few years of
heavy rainfall are needed to change the desert into
a lush savanna with seasonally flowing rivers. Such
rivers would in turn form wide deltas in the low-
lying, flat coastal zone. This type of environment
undoubtedly attracts birds from the surrounding
areas and provides excellent stop-over points for
the North American migrants that make up over
43 percent of the paleoavifauna.

A climatological phenomenon known as El
Nino (Chapman, 1926; Murphy, 1926) periodically
brings heavy rains to southwestern Ecuador and
northern coastal Peru. This supposedly occurs
every seven years, but in reality the phenomenon
may take place in any year. When rains do come to
southwestern Ecuador they occur during the
northern hemisphere winter. This increases the
probability of migrant species encountering favor-

166

able conditions during passage to and from their
wintering grounds.

If the avifauna of the Talara Tar Seeps were not
known, one would probably interpret the avifauna
of La Carolina, and possibly the mammalian fauna
as well, as having been deposited during El Nino
years. However, the Talara avifauna strongly sug¬
gests that the entire climatological regime of
northern coastal Peru, and thus by inference that
of southwestern Ecuador, was quite different from
that found today. Instead of desert, the area was
probably a lush savanna or, in the case of the Santa
Elena Peninsula, more probably a tropical dry
deciduous forest. Archeological work on the Santa
Elena Peninsula also indicates a moister climatic
regime in the past (Meggers, Evans, and Estrada,
1965; Richardson, 1973). The causes of the present
dry climate, as opposed to the wetter climate
present during glacial periods, are the subject of
another paper currently in preparation.

Comparison with the Avifauna of the Talara
Tar Seeps. —Of the 53 species from La Carolina,
38 (72%) were also found at the Talara Tar Seeps
(Table 2). This includes 7 extinct species common
to both sites that were first recorded from the
Talara Tar Seeps. The high degree of similarity
between the paleoavifaunas of the two localities,
as well as that between the modern avifaunas, con¬
firms that there must have been a great deal of
interchange between coastal Ecuador and coastal
Peru. A similar situation is observed between the
mammalian faunas of the two sites (Lemon and
Churcher, 1961).

The straight line distance between La Carolina
and the Talara Tar Seeps is approximately 273
kilometers. At the present time the Gulf of
Guayaquil separates the arid region of Ecuador
from that of northern Peru. During glacial periods,
however, the lowered sea level exposed as dry land
the area currently covered by the Gulf of Guaya¬
quil, allowing easy movement between the two
areas. The formation of the Gulf of Guayaquil at
the end of the Pleistocene resulted in the isolation
of the avifauna of the arid Santa Elena Peninsula
from that of coastal Peru. Consequently, a great
deal of subspeciation has occurred between the two
areas (Chapman, 1926).

The differences observed in the composition of
the two paleoavifaunas can be attributed to two
major causes—the means of entrapment and the

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

location relative to the coastline. The active tar
seeps at Talara attracted large numbers of scav¬
enger species that came to feed upon other animals,
including large mammals, trapped in the tar. The
drawing power of La Carolina in this regard must
have been weak. As the Talara site is located ap¬
proximately 32 kilometers inland, it has a higher
percentage of terrestrial species and fewer strictly
marine species that do not venture inland. The
opposite is true of La Carolina which was on, or
very near, the coastline.

The similarities between the avian and mam¬
malian faunas of La Carolina and the Talara Tar
Seeps, particularly in the extinct species, suggest
that material was being deposited at the two sites
contemporaneously. The Talara site has been
dated at c. 13,900 b.p. (Churcher, 1966). If the
age of La Carolina differs significantly from that of
the Talara Tar Seeps, I believe it will prove to be
younger rather than older, for climatological rea¬
sons to be discussed in a later publication.

Entrapment of the Birds at La Carolina.—
How birds became trapped, or their remains con¬
centrated, at La Carolina is not clear. The un¬
disturbed matrix available with the collection was
extremely rich in avian specimens, and contained
only a few fragmentary mammalian bones. At
other nearby sites, however, Edmund (pers. comm.)
found only mammalian and no avian material.

None of the fossil material I studied shows any
large concentration of pitch, although there are
some indications of its presence in the matrix sur¬
rounding some specimens. The absence of heavy
pitch concentrations on the bones indicates that
the birds were not trapped in active tar seeps, as
at the Talara Tar Seeps and Rancho La Brea. It
is quite possible, however, that oil seeps may have
coated the water with a surface layer of oil and that
upon landing in this the feathers of the birds would
become oil-soaked, preventing further flight and
resulting in their drowning. As the bodies decom¬
posed, the bones would be freed from the oil and
drop to settle in the soft mud. Such a film of oil
would also help to explain the lack of mammalian
specimens in the concentration of avian remains.
Mammals are better able to sense and avoid oil-
covered water than are birds alighting from flight.
For larger mammals, passage through oil covered
water would be no more than an unpleasant
inconvenience.

NUMBER 27

167

Table 2. —Fifty-three nonpasserine species from La Carolina, Ecuador

Species

*Nycticorax nycticorax .

*Dendrocygna autumnalis .

Anatidae genus and species

*Anas species 1 .

* Arias species 3 .

*Anas bahamensis .

Vulturidae gen. et sp. indet. ...

*Coragyps cf. atratus .

*Cathartes aura .

•Accipitridae genus and species

*Geranoaetus melanoleucus .

*Buteo polyosoma .

Buteo hoffstetteri, new species

Circus cinereus .

*Falco peregrinus .

*Falco femoralis .

*Polyborus plancus .

*Milvago species .

*Pluvialis dominica .

*Pluvialis squatarola .

*Charadrius vociferus .

*Charadrius semipalmatus .

Oreopholus orcesi, new species

*Tringa solitaria .

*Totanus flavipes .

*Totanus melanoleucus .

*Catoptrophorus semipalmatus .
Calidris canutus .

No. of
Specimens

Minimum

No. of

Individuals

Species

No. of
Specimens

Minimum
No. of
Individuals

5

1

*Erolia melanotos .

22

9

4

1

Ereunetes pusillus .

3

2

10

3

*Ereunetes mauri .

10

7

4

1

*Micropalama himantopus .

1

1

4

2

Bartramia longicauda .

8

2

326

37

Numenius phaeopus .

4

2

1

1

Limosa fedoa .

2

1

2

1

* Aren aria interpres .

7

2

5

3

♦Scolopacidae genus and species ....

2

1

1

1

Himantopus mexicanus .

1

1

2

1

Phalaropus fulicarius .

5

3

7

2

*Lobipes lobatus .

28

8

4

1

*Steganopus species .

4

2

3

1

Steganopus tricolor .

102

29

22

3

*Burhinus superciliaris .

18

4

2

1

*Thinocorus rumicivorus .

4

1

52

6

*Larus pipixcan .

2

1

3

2

Larosterna inca .

1

1

47

9

*Zenaida asiatica .

1

1

1

1

*Zenaida auriculata .

64

9

1

1

*Columbina cruziana .

1

1

2

1

A rat in ga roosevelti .

724

88

1

1

Aratinga species .

2

1

3

1

*Tyto alba .

1

1

11

3

*Speotyto cunicularia .

5

2

5

2

5

2

57

15

Total .

1611

284

♦Also recorded from the Talara Tar Seeps, Peru.

Perhaps the most remarkable aspect of the avi¬
fauna from La Carolina is the tremendous number
of specimens of parrots, making up approximately
44 percent of the entire collection. Parrots must
have been attracted to the site in large flocks, per¬
haps for water, and subsequently trapped by the
oil. From such large flocks of parrots one could

reason either that water was scarce and La Carolina
was the only watering site available, the parrots
thus having to travel long distances to reach it, or
water may have been abundant and La Carolina
close to the roosting or feeding areas of the parrots.
If the latter were true, it would indicate the pres¬
ence of forest on the presently barren peninsula.

Literature Cited

Campbell, K. E., Jr.

In press. The Pleistocene Avifauna of the Talara Tar
Seeps, Northwestern Peru. Royal Ontario Museum,
Toronto, Life Sciences Division, Contribution.

Chapman, F. M.

1926. The Distribution of Bird-life in Ecuador: A Con¬
tribution to a Study of the Origin of Andean
Bird-Life. Bulletin of the American Museum of
Natural History , 55:1-784.

Churcher, C. S.

1966. The Insect Fauna from the Talara Tar-seeps, Peru.
Canadian Journal of Zoology, 44:985-993.

Edmund, A. G.

1965. A Late Pleistocene Fauna from the Santa Elena
Peninsula, Ecuador. Royal Ontario Museum, To¬
ronto, Life Sciences Division, Contribution, 63:1-21.
Hoffstetter, R. F.

1952. Les mammifoes Pleistocenes de la Republique de
’ l’Equateur. Memoires de la Societe Geologique de
France, new series, 66:1-391.

168

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Lemon, R. R. H., and C. S. Churcher

1961. Pleistocene Geology and Paleontology of the Talara
Region, Northwestern Peru. American Journal of
Science, 259:410-429.

Marchant, S.

1958. The Birds of the Santa Elena Peninsula, S. W.
Ecuador. Ibis, 100:349-387.

1959. The Breeding Season in S. W. Ecuador. Ibis, 101:
137-152.

1960. The Breeding of Some S. W. Ecuadorian Birds.
Ibis, 102:349-382, 584-599.

Meggers, B. J., C. Evans, and E. Estrada

1965. Early Formative Period of Coastal Ecuador: The
Valdivia and Machalilla Phases. Smithsonian Con¬
tributions to Anthropology, 1:1-234.

Meyer de Schauensee, R.

1966. The Species of Birds of South America with Their

Distribution. 577 pages. Wynnewood, Pennsylvania:
Livingston Publishing Company.

Murphy, R. C.

1926. Oceanic and Climatic Phenomena Along the West
Coast of South America During 1925. The Geo¬
graphical Review, 16:26-54.

Richardson, J. B.

1973. The Preceramic Sequence and the Pleistocene and
Post-Pleistocene Climate of Northwest Peru. Pages
199-211 in D. W. Lathrap and J. Douglas, editors,
Variation in Anthropology: Essays in Honor of
John C. McGregor. Urbana Illinois: Illinois Archae¬
ological Survey.

Spillman, F.

1942. Contribucidn al conocimiento de fdsiles nuevos
de la avifauna Ecuatoriana en el Pleistoceno de
Santa Elena. Proceedings of the Eighth American
Scientific Congress, Washington, 4:375-389.

The Great Predatory Birds of the Pleistocene of Cuba

Oscar Arredondo

translated and amended by Storrs L. Olson

ABSTRACT

Recent paleontological investigations in Cuba
have shown that the island was formerly inhabited
by large populations of rodents and edentates.
Based on discoveries of giant fossil raptorial birds
(Strigiformes and Accipitriformes) in Cuba and
other Antillean islands, the theory is put forth
here that these predators were the principal agents
in controlling the abundance of these native mam¬
mals. Details are given of the discovery in Cuba of
the gigantic owl Ornimegalonyx, two species of
giant barn owls ( Tyto ), a giant species of eagle
(.Aquila borrasi), and a vulture ( Antillovultur)
similar in size to the Andean Condor ( Vultur
gryphus). The relationships and possible origins of
gigantism in these birds are discussed. Descriptions,
geographic distribution, tables of measurements,
and a list of the extinct faunas found in each of the
type-localities are offered for each species.

Se expone la tesis, basada en los ultimos descu-
brimientos realizados en Cuba y otras Antillas sobre
aves fosiles gigantes de los orclenes Strigiformes y
Accipitriformes, de que dichos taxones consti-
tuyeron el elemento faum'stico primoridal que con-
trolara con su accion predatora a roedores y des-
clentados que en estas islas existian, manteniendo
asi el equilibrio biologico necesario. Se ofrecen
datos concretos probatorios de la extremada abun-
dancia en las Antillas de poblaciones de roedores y
desdentados. Se dan detalles del descubrimiento en
Cuba del gigantesco bubo Ornimegalonyx, de dos
especies gigantes de lechuzas del genero Tyto, de
una especie de aguila gigante y de un buitre simi¬
lar en talla a Vultur gryphus. Se plantea una dis-
cusion sobre el motivo que pudo originar el
gigantismo en estas aves. Se ofrecen descripciones
espedficas de todas estas grandes aves predatoras,
incluyendose, distribucion geografica, tablas de
medidas y una relacion de la fauna extinguida
hallada asociada en las localidades tipicas.

Introduction

Recent studies of fossil material from numerous
caves in Cuba have convincingly demonstrated the
former extraordinary abundance of endemic mam¬
mals there. These include rodents of the genera
Heteropsomys, Capromys (including Geocapro-
mys), and Macrocapromys; ground sloths (Eden¬
tata) of the genera Cubanocnus, Miocnus,
Mesocnus, and Megalocnus ; and insectivores of the
genera Solenodon and Nesophontes. These genera
were represented throughout the island by species
very numerous in terms of individuals. From such
caves as Paredones and El Tunel in the province of
Habana it is no exaggeration to say that tens of

Oscar Arredondo, Grupo Espeleologico, Martel de Cuba,
Avenida 43, No. 5847, Apartado 4, Mariano 14, Habana,
Cuba.

thousands of mandibles of Capromys pleistocenicus
can be extracted from a single small chamber, al¬
though most of these specimens are poorly pre¬
served. An examination of only a part of the fossil
material taken from a small cave in Cayo Salinas,
a few miles east-southeast of Caibarien, Las Villas,
yielded the remains of over 200 individuals of the
edentate genus Mesocnus (Acevedo, Arredondo,
and Gonzalez, 1975). Further confirming the for¬
mer abundance of native mammals is the wide¬
spread occurrence in Cuba of remains of blood¬
eating vampire bats of the genus Desmodus
(Koopman, 1958; Arredondo, 1958b; Woloszyn
and Mayo, 1974). These bats would necessarily
have required numerous, large, warm-blooded
mammals to sustain them. Similar abundant faunas
of large’ rodents and ground sloths are also known
from cave deposits in Hispaniola and Puerto Rico.

169

170

Obviously, some natural regulator must have
acted to maintain a biological equilibrium between
these prolific herbivorous mammals and the vegeta¬
tion on which they fed. In the Antilles, however,
there are virtually no native carnivorous mammals.
Exceptions are Cubacyon transversidens (Arre¬
dondo and Varona, 1974), a canid known from a
single fragment of maxilla found in association with
extinct vertebrates in a cave in the province of
Habana, and the extant populations of Procyon
known from the Bahamas, Barbados, and Guade¬
loupe, some or all of which may have been intro¬
duced to these islands by man (Varona, 1974;
Arredondo and Varona, 1974). A mandible found
in an Indian midden in Camagiiey, Cuba, and at¬
tributed to Procyon lotor by the naturalist Andres
Poey (Harrington, 1935), may similarly have
pertained to an introduced animal; the record is
dubious in any case, since the specimen can no
longer be found. Regardless, small procyonids
could not have been an influential factor in the
control of the endemic Antillean rodents and
edentates.

The numerous and dense deposits of smaller
vertebrates (Table 1) frequently found in Cuban
caves abound with remains of the rodents 1 Capro-
mys pleistocenicus, Capromys nanus, Heteropso-
mys torrei, and H. offella; the insectivores Neso-
phontes micrus and N. major; several genera of
bats; passeriform birds, particularly of the genera
Mimocichla, Quiscalus and Dives; as well as non¬
passerines such as Crotophaga and Glaucidium.
These remains are certainly attributable to the
actions of medium-sized owls still existing in the
island such as Tyto alba and Asio stygius, and per¬
haps to others now extinct, such as Pulsatrix
arredondoi (Brodkorb, 1969).

The most outstanding accumulations of bones,
however, are those of the larger mammals (Table
1) such as Capromys pilorides, C. columbianus,
Macrocapromys acevedo, and occasional examples
of Cubanocnus gliriformis, as well as large birds of
various orders. Such remains suggest that many
larger predators were once active in Cuba.

Much of the accumulation of vertebrate remains

1 The mammalian classification used in this paper follows
Varona (1974) and in some instances is at variance with
Professor Arredondo’s preferences.—Ed.

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

in cave deposits in Haiti was attributed to the
depredations of the extinct giant barn owl Tyto
ostologa, first described by Wetmore (1922).
Another giant barn owl, Tyto pollens Wetmore
(1937), first reported from Great Exuma and later
from New Providence Island (Brodkorb, 1959),
must have been an effective predator of Capromys
in the Bahamas, as no doubt were the extinct
diurnal raptors (Accipitridae) of the genera Calo-
hierax and Titanohierax (Wetmore, 1937). The
presence of such large raptors elsewhere in the
Antilles suggested that the Cuban cave deposits
might be attributable to similar avian predators.
This was vividly confirmed by the discovery in
Cuba of remains of five species of predatory birds
of truly tremendous proportions. These birds
occupied the niches which in continental areas are
usually filled by various groups of carnivorous
mammals. The absence of mammalian competitors,
combined with a superabundance of large prey,
are no doubt the principal factors contributing to
the gigantism observed in these birds. The larger
forms may have functioned particularly to keep
the populations of edentates in check.

Brief History of Certain Discoveries

On 2 January 1954, the remains of a gigantic
owl, the largest known, were discovered for the first
time in the depths of a large cavern known as Pfo
Domingo Cave, located in the Sierra de Sumidero,
opposite Pica-Pica Valley in Pinar del Rfo. These
bones (Figure 5) were found in place, fastened to
the calcareous surface of the floor by travertine,
just as were those of the various edentates en¬
countered in the immediate vicinity. This owl was
described under the name Ornimegalonyx oteroi
(Arredondo, 1958a) and was at first erroneously
regarded as a member of the Phorusrhacidae, an
extinct family of flightless South American carniv¬
orous birds. Dr. Byran Patterson, who studied some
of these remains, afterwards informed me (pers.
comm.) of his belief that they actually pertained to
a great owl, two times larger in linear dimensions
than Tyto ostologa of Haiti. Brodkorb (1961),
recognizing the validity of the nomenclature pro¬
posed in 1958, established that the species belonged
in the family Strigidae, where he maintained it in
later publications (Brodkorb, 1969, 1971). Addi-

NUMBER 27

171

Table 1.—Associated fauna found in the type-localities of the large extinct species of Cuban birds of prey

Species

Ornimegalonyx

oteroi

Cavema de Pio
Domingo,

Pinar del Rio

Tyto

riveroi

Cueva de
Bellamar,
Matanzas

Tyto noeli and
Aquila borrasi

Cueva del Tiinel,
Habana

Antillovultur

varonai

Cueva de

Paredones, Habana

Amphibia

Bufo sp .

-

-

X

X

Reptiua

Iguanidae gen. and sp. indet.

-

—

X

-

Epicrates cf. angulifer .

-

-

X

X

Geochelone cubensis .

-

X

X

X

Crocodylus sp.

-

-

X

X

Aves

Cathartes aura .

_

_

X

—

Antillovultur varonai .

-

—

-

X

Vulturidae gen. and sp. indet.

-

-

X

-

Aquila borrasi .

?

-

X

X

Grus cubensis .

?

—

—

-

Ornimegalonyx oteroi .

X

-

X

X

Ornimegalonyx sp.

-

-

-

X

Pulsatrix arredondoi .

-

-

-

X

Gymnoglaux sp.

-

-

-

X

Tyto alba .

-

-

-

X

Tyto noeli .

-

-

X

X

Tyto riveroi .

-

X

-

-

Mammalia

Solenodon cf. cubanus .

X

—

X

X

Nesophontes micrus .

X

-

X

X

Nesophontes major .

-

-

-

X

Phyllonycteris poeyi .

-

-

X

-

Natalus lepidus .

-

-

X

-

Cubacyon tranversidens .

-

-

X

-

Acratocnus sp.

-

-

X

-

Miocnus antillensis .

X

-

X

X

Cubanocnus gliriformis .

X

X

X

X

Mesocnus torrei .

X

X

X

X

Mesocnus browni .

—

-

X

X

Neomesocnus brevirostris .

—

-

-

X

Megalocnus rodens .

X

X

X

X

Megalocnus sp.

-

-

X

-

Capromys pilorides .

X

-

X

X

Capromys prehensilis .

X

-

-

-

Capromys nanus .

-

-

-

X

Capromys columbianus .

X

X

X

X

Capromys pleistocenicus .

-

X

X

X

Capromys sp. A.

-

-

X

-

Capromys sp. B .

X

-

X

-

Capromys sp. C .

-

-

X

X

Capromys sp. D .

-

-

-

X

Macrocapromys acevedo .

-

-

-

X

Heteropsomys torrei .

X

-

X

X

Heteropsomys offella .

X

-

X

X

X = present; ? = uncertain; - =

not present.

172

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

tional fossils of this giant owl were later found in
various caves in the provinces of Habana, Matan-
zas, Las Villas, and Camaguey, as well as in the Isle
of Pines. It is possible that two unrecognized spe¬
cies of Ornimegalonyx may exist among these re¬
mains; one of these seems to be somewhat smaller
than O. oteroi, while the other is larger. Consider¬
ing the enormous size of this owl, particularly of
its claws, it is quite conceivable that it could have
made victims of juvenile edentates, notwithstand¬
ing the fact that Ornimegalonyx appears to have
been incapable of flight.

In July 1954, explorations in the Cueva de
Paredones in San Antonio de los Banos, revealed
for the first time the fossil bones of an eagle larger
than any of the living species of the family Accipi-
tridae. Other bones of it were found a few years
later in the Cueva del Tunel, La Salud, Habana.
This eagle was named Aquila borrasi (Arredondo,
1970). According to its tarsometatarsus, it was very
similar in size and morphology to the recently de¬
scribed species Garganoaetus freudenthali from the
upper Miocene of Italy (Ballmann, 1973).

A fossil vulture from Cueva de Paredones was
recently described as a new genus and species,
Antillovultur varonai Arredondo (1971), and was
the size of an Andean Condor. Various bones found
in a cave in Habana and now under study, possibly
indicate another large species of vulture.

Two species of giant barn owls of the genus Tyto
were discovered a little later (Arredondo, 1972a,
1972b). One of these, Tyto noeli, founded on
abundant bones from two caves in Habana, was
similar in size to Tyto ostologa of Haiti. The other
species, Tyto riveroi, based on the distal portion of
a tarsometatarsus from Cueva de Bellamar, Matan-
zas, was truly gigantic, being larger than any of the
fossil or living species of the genus. It is the one
strigiform that most closely approaches the size of
Ornimegalonyx oteroi, and like that species could
also have captured small edentates. The eminent
paleornithologist Alexander Wetmore (1959) was
the first to report fossil remains of large barn owls
from Cuba, but these were not named. 2

" These specimens, a. humerus and a femur, are still at the
National Museum of Natural History. They are very much
larger than T. ostologa or T. noeli and most probably per¬
tain to the distinctive species T. riveroi. At a later date I
hope to be able to describe these fossils further, along with
abundant unpublished material of T. ostologa. —Ed.

The discovery of three species of tytonids in the
upper Miocene of Italy (Ballmann, 1973), one of
which, Tyto robusta, is equal in size to T. noeli
and T. ostologa, and the other, Tyto gigantea,
being only, slightly smaller than T. riveroi, partly
contradicts the hypothesis that gigantism in Antil¬
lean barn owls is attributable to insular evolution
from smaller species that responded to the great
abundance of food and the lack of competition
from carnivorous mammals. The genus Tyto
evidently had already evolved giant species in Eu¬
rope, millions of years before the beginning of the
Pleistocene. The following conclusions could there¬
fore be drawn: (1) either the giant Antillean barn
owls evolved in parallel with those of Europe, ar¬
riving through convergence at species of approxi¬
mately the same size, or (2) the Antillean forms
are descended from Tertiary European forms that
established themselves in North America and
colonized the Antilles before or during the Pleisto¬
cene. Against this last suggestion is the absence on
the American continent of giant species of Tyto.

Ornimegalonyx is truly exceptional for its ex¬
tremely large size. It appears to have evolved its
gigantism in Cuba from some remote smaller ances¬
tor. An affinity of Ornimegalonyx with any of the
living genera of large continental owls is not
clearly evident and its relationships may lie closer
to some extinct form rather than with any presently
living.

The study of the origins, evolution, and paleo-
ecology of the giant raptorial birds of the Antilles
is of great interest and significance to our under¬
standing of the environment and evolution of
many of the terrestrial vertebrates of those islands.
It is hoped that this summary of what is known of
the Cuban birds will aid in that understanding.

Abbreviations used are as follows: Academia de
Ciencias de Cuba (ACC), Departamento de Pale-
ontologia de la Universidad de la Habana
(DPUH), Museo del Grupo de Exploraciones
Cientificas “Pedro Borras Astorga’’ (GEC), Mu¬
seum of Comparative Zoology (MCZ), Museo
Felipe Poey de la Academia de Ciencias de Cuba
(MFP), Museo Montand de la Universidad de la
Habana (MMEJH), personal collection of Oscar
Arredondo (OA), Sociedad Espeleologica de Cuba
(SEC).

NUMBER 27

173

Order ACCIPITRIFORMES
Family VULTURDDAE
Genus Antillovultur Arredondo, 1971

Antillovultur varonai Arredondo, 1971

Holotype. —Proximal portion of left tarsometa-
tarsus, DPUH 1254.

Type-Locality.— Cueva de Paredones, San An¬
tonio de los Banos, Habana, Cuba.

Age. —Late Pleistocene.

Other Material.— GEC (unnumbered), distal
portion of left humerus; OA 847, external trochlea
of left tarsometatarsus; OA 848, body of seventh
cervical vertebra. All specimens from the type-
locality.

Description. —The type (Figure 1 a,b) is a proxi¬
mal portion of a tarsometatarsus, 42.5 mm in
length, lacking the proximal articulating surface,
hypotarsus, and slightly more than half the distal
portion of the bone. The estimated total length is
141 mm, or about equal to that of Vultur gryphus
and longer and slightly more robust than in
Gymnogyps californianus (Figure 2, Table 2).
From these two species and Cathartes aura it differs
in having the groove in the anterior face of the
bone narrower and deeper and the internal tuber¬
cle for M. tibialis anticus more expanded. In ante¬
rior view, the surface of the shaft between the
internal border and the groove is notably thick
and rounded, whereas in Vultur, Cathartes, and
Teratornis it is narrower and sharp-edged. The
surface of the shaft delimited by the external bor¬
der and the groove is more slender than the
internal ridge, contrary to the condition in the
other genera mentioned. The shaft in medial view
is proportionately more slender than in Vultur or
Cathartes. These characters are considered to be of
generic value.

The distance between the proximal border of
the larger tubercle for M. tibialis anticus and the
proximal extremity of the anterior groove is
greater than in Vultur or Teratornis and equal to
that in Gymnogyps. Antillovultur has an addi¬
tional two proximal foraminae situated above the
usual two. Vultur similarly possesses another aper¬
ture above the lateral proximal foramen but lacks
the medial proximal foramen of Antillovultur.

Figure 1.—Specimens of Antillovultur varonai, Cueva de
Paredones: a, holotype fragmentary proximal end of left
tarsometatarsus (DPUH 1254), anterior view; b, same, pos¬
terior view; c, left humerus lacking proximal end (GEC
unnumbered), palmar view. (Natural size.)

174

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Table 2. —Measurements (mm) of the tarsometatarsus of Antillovultur varonai compared with other large New World vultures

Character

Antillovultur

varonai

Type, DPUH 1254

Vultur

gryphus

Gymnogyps

californianus

Teratornis

merriami

Maximum length .

141

141

124.2

132

Greatest proximal width .

32.2

30.2

27

28

Least width of shaft (at break) . .

17

17

16.8

15

Length of outer trochlea from the angle of the
middle trochlea .

9

9.5

6

9

Greatest width of outer trochlea .

8

8

8

8

Greatest width across trochlea .

35*

35.4

30

29.5

* Estimated.

Cathartes aura has only two proximal foramina.
Teratornis differs in having three great united
foramina arranged so as to form a kind of circle.
In Antillovultur the medial tubercle for M. tibi¬
alis anticus is better developed than in either
Vultur, Gymnogyps, or Cathartes, and is situated
over a slight protruberance on the internal border
of the anterior groove. In posterior view the sur¬
face of the bone from the lateral foramen to the
external border is flat and does not slope downward
as in Vultur. On the opposite side, the surface from
the medial foramen to the external border is very
depressed compared to that in Vultur. An external
trochlea from a left tarsometatarsus of this species
is similar in size and shape to that of Vultur.

abed

Figure 2. —Comparison of tarsometatarsi of New World vul¬
tures: a, Gymnogyps californianus; b, Vultur gryphus; c,
Antillovultur varonai; d, Teratornis merriami (based on Wet-
more, 1931). (Half natural size.)

A seventh cervical vertebra attributed to Antil¬
lovultur consists of the complete body lacking all of
the processes. It is similar to vertebrae of Vultur
in size and morphology.

A distal portion of a left humerus (Figure 1c)
is from a specimen very similar in size to Vultur.
This fragment has a length of 184 mm from the
distal end to a point a little beyond the protruber¬
ance at the distal extremity of the deltoid crest.
Taking this protruberance as a point of reference,
the complete length of the bone can be estimated
as 265 mm, which is about 20 mm less than in the
specimens of Vultur compared. The least width of
the shaft is 20 mm (21.7 in Vultur ); the maximum
distal width is 45.8 mm (54.3 in Vultur). The
bone, although almost as large as that of Vultur,
is more slender. The distal protruberance of the
deltoid crest is at the same level as in Vultur.
Antillovultur differs from Vultur in having the
ectepicondylar prominence less pronounced, re¬
calling that of Cathartes. The internal and exter¬
nal condyles are slightly smaller than in Vultur but
with the distal borders more prominent. The in¬
ternal condyle has a smooth and extensive depres¬
sion on the entepicondylar side, which is present
in Cathartes but absent in Vultur. The attachment
of the anterior articular ligament is as large as in
Vultur, but situated closer to the internal condyle,
the space between them being less than in Vultur
and more similar to Cathartes. The foramen located
in this space is isolated from the brachial depres¬
sion by a ridge that is absent in Vultur but some¬
what evident in Cathartes. The brachial depression
is deeper and more pronounced than in Vultur,
particularly in the proximal region, but is notably
less expanded than in either Vultur or Cathartes.
The olecranal fossa is similar to that of Vultur and

NUMBER 27

175

less dilated than in Cathartes. The tricipital grooves
are similar to those of Vultur but the external one
is somewhat deeper and appears like that of
Cathartes.

Family ACCIPITRIDAE
Genus Aquila Brisson

Aquila borrasi Arredondo, 1970

Holotype. —Left tarsometatarsus lacking troch-
leae, DPUH 1250.

Type-Locality.— Cueva del Tunel, La Salud,
Habana, Cuba.

Other Localities.— Cueva de Paredones, San
Antonio de los Banos, Habana. Cueva de Pio
Domingo, Sumidero, Pinar del Rio.

Age.— Late Pleistocene.

Other Material.— Right femur lacking con¬
dyles, SEC P-26; ungual phalanges, SEC P-31,
P-32, P-35, P-1147, and ACC 1000a; subterminal
phalanx, ACC 1000b; distal end of tarsometatarsus,
SEC P-40.

Description.— Tarsometatarsus (Figure 3) gen¬
erally similar to that of Aquila chrysaetos but
notably longer, since even without the distal end it
measures 97.7 mm. The estimated total length of
the bone is 130 mm, or 34 mm longer than that of
Aquila chrysaetos and larger than that of any liv¬
ing species of eagle (Table 3). The proximal artic¬
ular region is similar to that in Aquila but with
the proximal foramina located only 5 mm from the
internal cotyla, whereas in A. chrysaetos this dis¬
tance is nearly twice as great. That which remains
of the base of the hypotarsus in the type indicates
that this process was probably similar to that of A.
chrysaetos.

The tarsometatarsi of the fossil species Buteo
typhoius and B. contortus, from the upper Miocene
of Nebraska, and of B. conterminus, from the up¬
per Pliocene of Nebraska (Wetmore, 1923), are
larger than those of Recent species of Aquila and
Haliaeetus, but are more slender proximally, the
articular region in proximal view being of a dif¬
ferent shape than in Aquila and also differing in
the form and position of the middle trochlea. The
tarsometatarsus of B. contortus measures 113 mm
in length, which is 17 mm less than in Aquila bor¬
rasi. Although the Cuban species has certain simi¬

larities to B. contortus in the general structure of
the tarsometatarsus, it is larger and more robust
than that species and appears more like A. chrysa¬
etos. According to the published figures, the re¬
cently described species Garganoaetus freudenthali
Ballmann (1973), from the upper Miocene of
Italy, is similar morphologically to B. borrasi, al¬
though its tarsometatarsus is slightly more robust.
Aquila borrasi was a gigantic form within its genus,
the only other known fossil forms of which are
Aquila delphinensis and A. pennatoides, described
by Gaillard (1939) from tarsometatarsi from the
upper Miocene of France.

The femur of Aquila borrasi is larger and more
robust than that of any living eagle. Although the
one known specimen is incomplete, its maximum
length is estimated at about 155 mm, as opposed to
125 mm in Aquila chrysaetos, 114 mm in Haliae¬
etus leucocephalus, and 96 mm in Spizaetus
ornatus. This is even larger than in the two im¬
mense living eagles Harpia harpyja (131 mm)

Figure 3. —Holotype left tarsometatarsus of Aquila borrasi
(DPUH 1250), Cueva del Tunel. (Anterior view at natural
size.)

176

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Table 3. —Measurements (mm) of the tarsometatarsus and claws of Aquila borrasi
compared with other species of living eagles

Character

Aquila borrasi

Pithecophaga

jefferyi*

Harpia

harpyja*

Aquila

chrysaetos

Haliaeetus

leucocephalus

Spizaetus

omatus

Tarsometatarsus

Type, DPUH 1250

Length from the upper border of

the internal cotyla to the begin-

ning of the scar for the hallux

97.7

80

70

64.5

55

68.2

Proximal width .

22.4

24

36.6

21.5

20.5

18.5

Least width of shaft .

13.2

13.2

19.3

10

10

9.6

Ungual Phalanx of Digit I

GEC

SEC P-31 (unnumbered)

Dorsal arc .

52.6 51.2

47

64

38.6

34.5

40.2

Ventral arc .

46.7 47.2

39

51.5

33.5

31.2

33.5

* Measurements from specimens in the USNM collections. It can be seen that the tarsometatarsus of Aquila borrasi, while
decidedly longer, is proportionately much more gracile than in either Pithecophaga or Harpia —Ed.

and Pithecophaga jefferyi (130 mm). 3 Other meas¬
urements (in mm) of this specimen are as fol¬
lows: total length as preserved, 140, proximal
width 45.4, vertical diameter of head 16.3, width of
neck 17.2, maximum width of pneumatic foramen
8.0, length of pneumatic foramen 14.0, least width
of shaft 19.8.

The femur differs from that of allied genera by
the lesser projection of the trochanter above the
head and by the greater width between the anterior
border of the head and the apex of the trochanter
(twice that of Aquila chrysaetos or Haliaeetus leu-
cocephalus). The pneumatic opening is roughly
oval in shape but wider distally. It is located at the
base of the trochanter as in Aquila but differs in
being not perfectly oval and in being oriented semi-
obi iquely towards the external border of the tro¬
chanter. The foramina of the upper region of the
trochanter are larger and deeper than in the other
species compared. The rugose intermuscular line
on the anterior face of the bone angles below and
near the pneumatic opening along the external
border of the shaft almost to its midpoint. In con¬
trast, this line in A. chrysaetos originates farther
above the upper border of the pneumatic opening
and descends straight to the midpoint of the shaft.
The head is massive and the neck is thick and
oriented slightly upwards.

3 Measurements supplied from specimens in the collections
of the National Museum of Natural History, Smithsonian
Institution.—Ed.

The ungual phalanges are very well developed
(Figure 4), being almost two times larger than
those of A. chrysaetos. They resemble those of
Harpia harpyja in having the same degree of cur¬
vature. The ungual phalanx of digit IV is larger,
while that of digit I is smaller than in Harpia
(Table 3). The shape of the articular region and
the ventral process of the first ungual phalanx of
A. borrasi more closely resembles that of Aquila
than Haliaeetus, Spizaetus, or Buteo. The ungual
phalanges of digits II and IV are likewise similar
to those of Aquila and differ from those of the
other genera examined. The ungual phalanx of
digit IV measures 33 mm through the ventral arc
and 35 mm through the dorsal arc.

Order STRIGIFORMES
Family STRIGIDAE
Genus Ornimegalonyx Arredondo

Ornimegalonyx oteroi Arredondo, 1958a

Synonym.— Ornimegalonyx arredondoi Arre¬

dondo, 1958.

Lectotype. —Left tarsometatarsus SEC P-383E.
Lectotype designated by Brodkorb (1961); depos¬
ited in the Museum of Comparative Zoology,
Harvard University.

Type-Locality. —Caverna de Pio Domingo, Si-

NUMBER 27

177

Figure 4. —Phalanges of Aquila borrasi from Cueva de Pare-
dones: a, ungual phalanx of digit I (SEC P-31); b, ungual
phalanx of digit II (SEC P-32); c, subterminal phalanx (SEC
P-35). (Natural size.)

erra de Sumidero, Ensenada de Pica-Pica, Pinar del
Rio, Cuba.

Other Localities. —Sierra de Anafe, Guanajay,
Habana; Cueva de Paredones, San Antonio de los
Banos, Habana; Cueva del Tunel, La Salud, Ha¬
bana; Cueva de Isla, Punta del Este, Isla de Pinos;
Cueva de Quinto, Boca de Camarioca, Mantanzas;
Canteras de los Hornos de Cal, Sancti Spiritus, Las
Villas; Sierra de Cubitas, Camaguey.

Age.— Late Pleistocene.

Other Material.— Caverna de Pio Domingo,
SEC P-383E, various pieces from the same indivi¬
dual as the lectotype as follows: left tarsometa-
tarsus and fragments of the right, fragments of
both tibiotarsi, left femur and fragments of the
right, articular portions of the mandible, fragments
of the sternum, scapula, carpometacarpus, and
ungual and subterminal phalanges. From Pare¬
dones, El Tunel, Quinto, and Isla caves there are
unnumbered pieces of tarsometatarsi, tibiotarsi,
femora, phalanges, synsacra and other bones, all in
private collections. From the collection of the
author there are three subterminal phalanges from
El Tunel and Paredones (OA 671, 676, 800).

Description.— Eight tarsometatarsi of Ornimega-

lonyx have come to light so far: the lectotype from
Pio Domingo; one from Cueva de Quinto; two
proximal halves from Paredones; a distal half from
Cueva del Tunel and a proximal half from the
same locality. Ultimately, two complete specimens
from the same individual were found in a cave in
the Sierra de Cubitas, Camaguey, along with other
elements.

The dimensions of these bones indicate the enor¬
mous size that this bird had in relation to all other
known Strigiformes, living or fossil (Figures 5-9,
Tables 4-6). The tarsometatarsus is almost double
the length of that of Bubo bubo, or more than
double if one considers the specimen from Cueva
de Quinto (GEC unnumbered) (Figure 7 a) or that
from Paredones. These bones are more than three
times the size of the corresponding element of
Nyctea scandiacci, more than four times that of
Asio otus, and eight and a half times the size of
Glaucidium siju.

In spite of its gigantic size, the tarsometatarsus of
Ornimegalonyx is proportionately less robust than
that of Bubo, Nyctea, or Pulsatrix, the difference
being due to the relative lengthening of the shaft
in Ornimegalonyx. If the tarsometatarsus of Orni¬
megalonyx were reduced to the length of that of
Asio otus, the two elements would be seen to be
very similar in proportion, whereas if one magni¬
fied the tarsometatarsi of Bubo, Nyctea, or Pulsa¬
trix to the size of that of Ornimegalonyx, they
would appear much stronger, wider, and more
robust.

In Ornimegalonyx the internal trochlea is pro¬
portionately shorter and wider than in Bubo,
Nyctea, or Asio, being more similar to Pulsatrix.
The middle trochlea is narrow and placed very
close to the outer trochlea. The distal foramen is
somewhat lower than in Bubo, Nyctea, or Asio,
and the ossified bridge on the anteroproximal re¬
gion of the bone recalls that of Bubo and differs
from Nyctea in that it is stronger and more circu¬
lar. The internal cotyla is similar to that in Bubo
and Asio, but somewhat lower than in Nyctea. The
wide, deep groove on the posterior face of the bone
is more pronounced than in Bubo, Nyctea, or
Asio.

The tibiotarsi of the type individual are frac¬
tured into proximal and distal portions and shafts
(Figure 5, Table 5). In a complete state they would
have measured some 250 mm in length. The tibio-

178

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Table 4. —Measurements (mm) of the tarsometatarsus of Ornimegalonyx oteroi compared with other owls

Ornimegalonyx oteroi

Bubo

Nyctea

Character

SEC P-383.E
Lectotype

SEC P-39

GEC (un¬
numbered)

MFP (un¬
numbered)

bubo

scandiaca

Total length .

147*

177

177

178

83.7

55.7

Proximal width .

c. 32

36

35

-

22.6

19.9

Distal width across trochleae .

c. 34

-

33

32

24.4

20.8

Least widtli of shaft .

c. 15

c. 16

15

18

12.2

11.5

Length of middle trochlea from angle with inner
trochlea .

_

_

10

10

6.2

5.5

Least width of middle trochlea .

-

-

13

13

8.9

8

Length of outer trochlea from angle with middle
trochlea .

c. 6

6

6

3

2.5

Width of outer trochlea .

c. 9

-

8

9

5.4

6.8

Length of inner trochlea from angle with middle
trochlea .

. ,

9

9

6

5.5

Width of inner trochlea .

-

-

14

c. 13

6.7

6

* Estimated.

tarsus from Cueva de Quinto has a length of 272
mm, which is almost twice that of Bubo bubo or
Nyctea scandiaca. Its proximal width of 40 mm is
likewise twice that of those species. The tibiotarsus
of the lectotype individual appears to have been
straight, while that from Cueva de Quinto is
slightly curved and subtly twisted. Compared to
the modern genera examined, the tibiotarsus of
Ornimegalonyx (Figure 6a) has the fibular articu¬
lation more pronounced, the rotular crest, and the
cnemial crest more elevated, and the fossa proximal
to the condyles on the anterior face deeper.

Eight femora of Ornimegalonyx have been
found so far. The left femur of the type individual
is in the MCZ. Of the right, only a part of the
proximal end remains in Cuba. The largest femur,
represented by the proximal end only (Figure Id),
was found to the east of Sancti Spiritus and is de¬
posited in the Museo Montane of the University
of Havana (MMUH 5072). Two incomplete fem¬
ora from Paredones (formerly SEC P-37 and SEC
P-38, but now in the MCZ) are smaller than those
of the type and for now are perhaps best referred
to as Ornimegalonyx sp.

The femur of Ornimegalonyx differs principally
from that of other owls in being much larger and
notably more robust (Table 6). That of the type
individual is almost one and a half times larger
than the femur of Bubo bubo, while the large
femur (MMUH 3072) from Sancti Spiritus is al¬
most double the size of that of Nyctea scandiaca.
It is curious that when reduced, the tarsometatarsi

of Ornimegalonyx are less robust than in Bubo or
Nyctea, while in the femora the opposite occurs.

Fragments of the sternum, as well as parts of the
scapulae, ribs, vertebrae and carpometacarpi, were
associated with the type individual in Pio Do¬
mingo cave. The most important sternal fragments
are an anterior portion with the articulations for
the coracoids, and another fragment of the left
side containing four costal facets (the fifth having
been fractured off). The costal facets vary slightly
in size, the largest being 6X6 mm. Through these
fragments it has been possible to reconstruct the
sternum (Figure 8) as being wide, almost flat (both
dorsally and ventrally), with a vestigial keel, which
indicates that the bird was hardly able to fly. Its
estimated length is 120 mm (vs. 47 mm in Tyto
alba), the estimated width 75 mm (30 mm in T.
alba)-, and the height at the keel some 30 mm
(25 mm in T. alba). The similarity of this last
measurement in two species which otherwise differ
so greatly in size is a further indication of the great
extent of the atrophy of the keel in Ornimegalonyx.

The carpometacarpus of Ornimegalonyx is
small in proportion to the enormous size of the
body. Its total length is estimated at about 90 mm
whereas in Bubo bubo, a smaller volant species, it
is 85 mm.

Ficure 5. —Tibiotarsus, tarsometatarsus, femur, and phalanges
(SEC P-383E) from the same individual as the lectotype of
Ornimegalonyx oteroi, Caverna de Pio Domingo. The speci¬
mens are covered with calcareous concretions. (Natural size.)

NUMBER 27

179

180

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

As with the other bones, the phalanges of Orni-
megalonyx stand apart from those of other owls by
their great size. The length of the ungual pha¬
langes of digits II and III, measured through the
dorsal arc, range from 37 to 39 mm, while that of

a

Figure 6 . —Hindlimb elements of Ornimegalonyx oteroi from
Cueva de Paredones: a, right tibiotarsus lacking distal end
(SEC P-28); b, posterior view of left tarsometatarsus lacking
distal end (SEC P-39). (Natural size.)

digit I measures 40 mm. The proximal height of
these phalanges ranges from 15-17 mm.

It was not until the beginning of 1969 that frag¬
ments of mandibles were found among the remains
of the type individual of Ornimegalonyx oteroi.
These consist of the two articular portions of both
rami. These fragments permit for the first time a
very approximate estimate of the size of the man¬
dible and ultimately of the whole skull (Figure
9a). These mandibles are very similar in overall
morphology to those of the diminutive genus Glau-
cidiam. They differ from Bubo bubo in that the
internal angular process is greatly lengthened. The
posterior angular process, in comparison to that of
Bubo, is notably more robust, and the portions
that remain of the surangular and angular are
greatly thickened at the point of the fracture.
Judging by comparison with recent owls, the length
of the entire mandible of Ornimegalonyx oteroi
would have been some 115 mm and the distance be¬
tween the external borders of the articulations was
approximately 100 mm, or about two times larger
than in Bubo bubo.

A portion of a cranium of Ornimegalonyx was

f

Figure 7.—Some variation in the hindlimb of Ornimegalonyx:
a, Ornimegalonyx cf. oteroi, left tarsometatarsus (GEC un¬
numbered), Cueva del Qtiinto; b, c, Ornimegalonyx oteroi,
left tarsometatarsus and right femur of the type individual
(SEC. P-38E), Pio Domingo cave; d, Ornimegalonyx sp.,
proximal end of left femur (MMUH 3072), Camera de los
Hornos de Cal, Sancti Spiritus.

NUMBER 27

181

Table 5.—Measurements

(mm) of the

tibiotarsus

of Ornimegalonyx oteroi

compared with

other owls

Character

Ornimegalonyx oteroi

Bubo

bubo

Nyctea

scandiaca

SEC P-383.E

SEC P-28

SEC P-29

MFP (un¬
numbered

GEC (un¬
numbered

Total length .

-

-

-

272

155

129

Proximal width .

c. 36

39

-

-

37

19

17.6

Least width of shaft .

c. 13

14

15*

13*

c. 15

9.5

8

Distal width across trochleae

30

-

31

c. 30

c. 30

20.8

19.2

* Estimated.

Table 6. —Measurements (mm) of the femur of Ornimegalonyx oteroi compared with other owls

Ornimegalonyx oteroi

Ornimegalonyx

Bubo

bubo

Nyctea

scandiaca

Character

SEC P-383E

MFP

(unumbered

sp.

MMUH 3072

Maximum length .

Proximal width between antero-external
border of head and postero-external

154

160

168*

108

96.4

border of trochanter

42

45*

46

23.7

20.5

Antero-posterior diameter of head

c. 11

12*

c. 12

7.5

8

Vertical diameter of head

15

—

c. 15

9

8.3

Least width of shaft .

16

16*

20*

9.6

9.8

Distal width through condyles .

35*

38*

-

22.5

21.5

* Estimated.

Figure 8.—Ventral and lateral views of the sternum of Ornimegalonyx oteroi, as reconstructed
from two fragments from Pio Domingo and Paredones caves. (Natural size.)

••

NUMBER 27

183

found in Cueva de Paredones in 1959 (Figure
9b,c). For the time being it is best referred to only
as Ornimegalonyx sp., for it is apparent that it
does not correspond to the species O. oteroi.
Rather, it appears to belong to a smaller form, as
also evidenced by the femora found in the same
locality. The specimen consists of the posterior
portion of a cranium from the postorbital arc to
the occiput and including the basisphenoid, the
foramen magnum, and the occipital condyle.
Viewed from the front, the great thickness of the
walls of the cranium are seen in the region of the
break. Compared to Bubo bubo it is larger, and in
ventral view it has the postorbital process better
developed. The opisthotic process is rather promi¬
nent and its extremity bends notably, hanging in
the form of an ear. The basipterygoid processes are
well developed. The foramen magnum is some¬
what higher than wide, as opposed to Glaucidium
and Tyto in which it is wider than high. As in
Nyctea, the occipital condyle is very well developed
in relation to the foramen magnum, whereas it is
proportionately much smaller in Glaucidium and
Tyto.

From the actual and estimated measurements of
the various bones of Ornimegalonyx it can be es¬
tablished that this great owl stood some 1100 mm
high in life. Although the general aspect of its skel¬
eton is similar to that of living owls, it is distin¬
guished from them by the long and robust hind-
limbs, provided with long, heavy toes armed with
the most powerful claws possessed by any strigiform
bird. Although the sternum is larger than in any
living owl, it is actually small in proportion to the
rest of the bones of the skeleton. Its semiflat struc¬
ture and reduced keel show that Ornimegalonyx
was little or not at all capable of flight. In accord¬
ance with this, the bones of the wing are poorly
developed, particularly the carpometacarpus.

Some of the differences in morphology and size
that are observed between individuals of Orni¬
megalonyx are probably attributable to sexual di¬
morphism, since in other owls the females are

Figure 9. —Skulls of Ornimegalonyx: a , hypothetical recon¬
struction of the skull of Ornimegalonyx oteroi to show the
size as extrapolated from the two mandibular articulations
associated with the type; b, c, anterior and ventral views of
the cranium (GEC unnumbered) of Ornimegalonyx sp. from
Cueva de Paredones (the left quadrate is incorrectly articu¬
lated). (All figures natural size.)

larger than the males. Nevertheless, other bones
that are either larger or smaller than those of the
type individual and have distinct differences from
it, probably indicate additional species—the cran¬
ium and femora from Paredones and the femur
from Sancti Spiritus being examples.

Ornimegalonyx had to have been the scourge
and terror of most of the larger mammals of the
Pleistocene of Cuba and the claws and mandibles
of this bird would have constituted a terrible com¬
bination of superior destructive power.

Family TYTONIDAE
Genus Tyto Billberg

Tyto noeli Arredondo, 1972a

Holotype.— Right tarsometatarsus, DPUH 1251.

Type-Locality. —Cueva del Tunel, La Salud,
Habana, Cuba.

Other Localities. —Cueva de Paredones, San
Antonio de los Banos, Habana; Cueva del Indio,
Reparto El Globo, Calabazar, Habana; quarries
near Sancti Spiritus, Las Villas.

Age. —Late Pleistocene.

Other Material.— Cueva del Tunel: OA 818,
right femur; OA 812, distal portion of left tibio-
tarsus; OA 804, distal portion of right humerus;
OA 806, proximal fragment of right humerus; OA
822, shaft of right tibiotarsus; OA 815, distal por¬
tion of right ulna. Cueva de Paredones: OA 828,
proximal portion of right tarsometatarsus; OA 827,
proximal portion of right tibiotarsus; OA 839,
right coracoid. Cueva del Indio: OA 1027, right
femur.

Description.— Similar to the living species Tyto
alba in its general skeletal configuration, but much
larger (Figures 10, 11, Tables 7 and 8), equaling in
size the extinct species Tyto ostologa of Haiti and
T. pollens of the Bahamas. The tarsometatarsus
was between 90 and 100 mm long and was similar
to that of T. pollens, but more slender, even in
specimens that are longer than in T. pollens. The
tibiotarsus is likewise similar to that of T pollens
but is less robust. This slenderness of the hindlimb
is the most notable difference between the two
species.

In the femur, humerus, ulna, coracoid, and
claws, the only pronounced difference from Tyto

184

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Figure 10.—Paratypes of Tyto noeli from Cueva del Tunel and Cueva de Paradones compared
with a Recent specimen of Tyto alba furcata (on the left in each pair): a, fragments of right
humerus (OA 804 and 826); b, fragments of right ulna (OA 806 and 815); c, fragments of
tibiotarsus (OA 827, 812, and 822); d, right femur (OA 818); e, right coracoid (OA 839). (Natural
size.)

alba is in size. Likewise, a fragment of the anterior
portion of a sternum of T. noeli from quarries near
Sancti Spiritus has the same conformation as that
of T. alba but is larger. One might expect to find
greater distinctions in the skull, but so far only
fragments of the skull of T. noeli have been found.

Brodkorb (1959:357) suggested that T. pollens

may possibly be differentiated from T ostologa
only at the subspecific level. The same could be
suggested for T. noeli. From the upper Miocene of
Italy, a new species of giant barn owl, Tyto robusta
Ballmann (1973), has been described that is near
the size of T. noeli. The Cuban species is somewhat
larger and heavier, however.

NUMBER 27

185

Table 7. —Measurements of limb bones of Tyto noeli compared with Tyto alba

Character

Tyto noeli

Tyto alba furcata

Femur

OA 818

OA 834

OA 1027

Total length .

74.3

73*

-

61

Proximal width .

14.4

13.1

14

11.4

Antero-posterior diameter of head

5

5

5.5

3.2

Vertical diameter of head

5.8

5

5.7

4.6

Breadth through trochanter .

9.2

8.2

9.2

6.8

Least width of shaft .

6.4

6

6.5

5

Distal width .

14.8

14

-

12.1

Tibiotarsus

OA 827

OA 831

Total length .

147*

-

108

Proximal width .

16

-

10.9

Least width of shaft .

6.4

-

5.5

Distal width .

-

15

11.3

Humerus

OA 804

OA 826

Total length .

137*

-

100.1

Proximal width .

-

23

16.9

Width of shaft .

8.5

-

6.2

Distal width .

20.5

-

15.7

* Estimated.

Table 8. —Measurements (mm) of the tarsometatarsi of the three Cuban species of Tyto

Character

Tyto riveroi

DPUH 1252
Type

Tyto noeli

DPUH 1251 OA 828
Type

Tyto alba furcata

Maximum length .

125*

91.7

100*

78.4

Proximal width .

22*

14.6

16

11.1

Distal width .

22

17.4

-

13.7

Least width of shaft .

9

6.7

7.3

5.2

Length of middle trochlea from
angle of internal trochlea .

6.8

5

_

2.5

Width of middle trochlea .

8.5

6.4

-

4.8

Length of outer trochlea from
angle of middle trochlea

11.5

c. 4

_

3

Width of outer trochlea .

12.3

4.8

-

3.9

Length of inner trochlea from
angle of middle trochlea

6.5

4.4

_

2.8

Width of inner trochlea .

9.8

7.5

—

5

* Estimated.

Tyto riveroi Arredondo, 1972b

Holotype. —Distal portion of a left tarsoraeta-
tarsus, DPUH 1252.

Type-Locality. —Cueva de Bellamar, Mantanzas,
Cuba.

Age.— Late Pleistocene.

Description. —Tarsometatarsus with the general

aspect of that of Tyto alba and still more similar to
that of T. noeli. Except for size, significant morph¬
ological differences from the above species are al¬
most absent; however, the measurements of the
type of T. riveroi notably exceed the limits of
either (Figure 11, Table 8). The estimated total
length of this bone would be approximately 125

186

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

mm. The following slight morphological distinc¬
tions are also noted: greater separation of the in¬
ternal and external trochleae from the middle
trochlea, the intertrochlear spaces being narrower
in T alba and T noeli; shaft proportionately
wider and thicker. Compared with T. pollens, the
same slight differences are apparent.

Tyto gigantea, recently described from the upper
Miocene of Gargano, Italy (Ballmann, 1973), was
an enormous barn owl, equal in size to T. riveroi.
According to the published figures, the Italian
species has the distal foramen somewhat more ele¬
vated and the middle trochlea lower and more
elongate than in the Antillean species.

Figure 11.—Comparison of the tarsometatarsi of the three Cuban species of Tyto: a, Recent
Tyto alba furcata; b, Tyto noeli, holotype (DPUH 1251), Cueva del Tunel; c-e, Tyto riveroi,
holotype (DPUH 1252), Cueva de Bellamar, anterior, posterior, and lateral views. (Natural size.)

Literature Cited

Acevedo, M., O. Arredondo, and N. Gonzdlez

1975. La Cueva clel Tunel. Volume 1, 74 pages, 17 fig¬
ures, 2 maps, 6 tables. Havana: Editorial Pueblo y
Educacidn. Instituto del Libro.

Arredondo, O.

1958a. Aves gigantcs de nucstro pasado prehistdrico. El
Cartero Cubano, 17(7): 10-12, 6 figures.

1958b. El Vampiro Cubano. Scout (Havana), pages 6-7, 10,
6 figures.

1970. Nueva especie de ave pleistocdnica del orden Ac-
cipitriformes (Accipitridae) y nucvo genero para
las Antillas. Ciencias, series 4, Ciencias Bioldgicas,
8:1-19, 10 figures.

1971. Nuevo genero y especie de ave fdsil (Accipitriformes:
Vulturidae) del Pleistoceno de Cuba. Memoria de
la Sociedad de Ciencias Naturales La Salle, 31(90):
309-323, 5 figures.

1972a. Nueva especie de ave fdsil (Strigiformes: Tytonidae)
del Pleistoceno superior de Cuba. Boletin de la
Sociedad Venezolana de Ciencias Naturales, 29(122—
123):415-431, 5 figures.

1972b. Especie nueva de lechuza (Strigiformes: Tytonidae)
del Pleistoceno cubano. Boletin de la Sociedad
Venezolana de Ciencias Naturales, 30(124-125):
129-140, 4 figures.

NUMBER 27

187

Arredondo, O., and L. S. Varona

1974. Nuevos g^nero y especie de mamffero (Carnivora:
Canidae) del Cuaternario de Cuba. Poeyana, 131:
1-12, 3 figures.

Ballmann, P.

1973. Fossile Vogel aus dem Neogen der Halbinsel Gar-
gano (Italien). Scripta Geologica, 17:1-75, 16 figures,
7 plates.

Brodkorb, P.

1959. Pleistocene Birds from New Providence Island,
Bahamas. Bulletin of the Florida State Museum,
Biological Sciences, 4(11):349—371.

1961. Recently Described Birds and Mammals from
Cuban Caves. Journal of Paleontology, 35(3):633—
635.

1969. An Extinct Pleistocene Owl from Cuba. Quarterly
Journal of the Florida Academy of Sciences, 31(2):
112-114, 1 figure.

1971. Catalogue of Fossil Birds, Part 4 (Columbiformes
through Piciformes). Bulletin of the Florida State
Museum, Biological Sciences, 15(4): 163-266.

Gaillard, C.

1939. Contribution a 1’^tude des oiseaux fossiles. Archives
du Museum d’Histoire Naturelle de Lyon, 15(mem-
oire 2): 1-100, 34 figures.

Harrington, M. R.

1935. Cuba antes de Colon. 290 pages. 111 figures, 108
plates. Havana: Coleccidn de Libros Cubanos,

XXXII, Cultural.

Koopman, K. F.

1958. A Fossil Vampire Bat from Cuba. Breviora, 90:
1-4, 1 plate.

Varona, Luis S.

1974. Catalogo de los mamiferos vivientes y extinguidos
de las Antillas. 139 pages. Havana: Academia de
Ciencias de Cuba.

Wetmore, A.

1922. Remains of Birds from Caves in the Republic of
Haiti. Smithsonian Miscellaneous Collections, 74
(41): 1-4, 2 figures.

1923. Avian Fossils from the Miocene and Pliocene of
Nebraska. Bulletin of the American Museum of
Natural History, 48:483-507, 20 figures.

1931. The Avifauna of the Pleistocene in Florida. Smith¬
sonian Miscellaneous Collections, 85(2): 1-41, 16

figures, 6 plates.

1937. Bird Remains from Cave Deposits on Great Exuma
Island in the Bahamas. Bulletin of the Museum of
Comparative Zoology, 80(12):427-441, 16 figures, 1
plate.

1959. Birds of the Pleistocene in North America. Smith¬
sonian Miscellaneous Collections, 138(4): 1-24.

Wotoszyn, B. W., and N. A. Mayo

1974. Postglacial Remains of a Vampire Bat (Chiroptera:
Desmodus) from Cuba. Acta Zoologica Cracoviensia,
19(13):253-265, 5 figures, 1 plate.

The Species of Moas (Aves: Dinornithidae)

Joel Cracraft

ABSTRACT

In order to provide a basis for future studies on the
functional morphology and evolutionary history of
the New Zealand moas, the species-level system-
atics of the family is reviewed. Based on a study of
museum collections and an analysis of intraspecific
variability, only 13 species are considered to be
valid: Anomalopteryx didiformis, A. oweni,

Megalapteryx didinus, M. benhami, Pachyornis
elephantopus, P. mappini, Euryapteryx curtus,
E. geranoides, Emeus crassus, Dinornis struthoides,

D. torosus, D. novaezealandiae, and D. giganteus.
The taxa accepted as valid in this paper probably
fit the biological species concept more closely than
do those of previous classifications, in which as
many as 29 species have been recognized. The
“species-pairs,” Pachyornis mappini-P septentrio-
nalis, Euryapteryx curtus-E. exilis, E. geranoides-

E. gravis, and Emeus crassus-E. huttonii probably
represent examples of sexual size dimorphism.

Introduction

Birds provide a number of classic examples of
insular adaptive radiations, the best known being
the Hawaiian honeycreepers (Drepanididae) and
the Galapagos finches (Geospizinae). A unique
radiation of this sort is also found in the New Zea¬
land moas (Dinornithidae)—unique because a
spectacular radiation took place not only in feed¬
ing mechanisms, but also in body size and propor¬
tions, in contrast to the well-known passerine
radiations. This occurred within a group that was
completely flightless, and indeed it is probable that
being flightless enhanced selective pressures toward
divergence in cranial morphology and body size.

Joel Cracraft, Department of Anatomy, University of Illinois
at the Medical Center, Chicago, Illinois 60680.

Moas are primitive ratite birds whose closest
living relatives are probably the kiwis (Cracraft,
1974). I have postulated elsewhere that the com¬
mon ancestor of the kiwis and moas probably had
a distribution in the Cretaceous that included west¬
ern Antarctica and parts of South America and
that moas and kiwis were isolated on New Zealand
following northward continental drift of that land-
mass beginning in the Late Cretaceous (Cracraft,
1973a; 1974). If this is so, then the radiation of
moas probably began in the Tertiary, but it is my
opinion that the taxonomic and morphological
diversity observed in the Holocene assemblages of
moas is the result of a relatively recent, probably
Pleistocene, episode of speciation (Cracraft, in
prep.). Thus, species formation within moas prob¬
ably involved repeated cycles of isolation and
sympatry among populations of the two major is¬
lands and among populations that very likely were
isolated in forest refugia at glacial maxima. It is
within such a model of their evolution that I have
considered the species-level systematics of moas.

Despite the fact that hundreds of papers have
been written about moas (see summaries in Lam-
brecht, 1933; Oliver, 1949), there is comparatively
little known about their functional morphology or
evolutionary history. Upon initiating such studies,
it quickly became clear that the complexities and
confusion of presently accepted species-level syste¬
matics would hinder any advances in morphologi¬
cal or evolutionary investigations. It thus became
necessary to review the status of the various species
of moas, and this paper presents the results of that
study.

The species-level systematics of moas has suf¬
fered from a century of typological thinking and a
lack of application of modern concepts of popula¬
tion biology. No less than 60 specific names have
been applied to a group that almost certainly con-

189

190

tains fewer than 20 biological species (see Brod-
korb, 1963, for citations to the taxonomic litera¬
ture). Each variant, each newly discovered bone in
some cases, was frequently given a new specific
name. Unfortunately, the taxonomic philosophy of
certain influential recent workers, rather than
clarifying the systematics of these birds, merely con¬
founded matters further. Oliver (1949:132-134),
for example, believed that “it is the work of the
systematist . to define the units that make up
[a series of continuous sizes between extreme
forms],” and that “in dealing with fossil species we
should not hesitate to give specific names to forms
that differ only slightly from one another . . .”
(1949:164). It was this philosophical approach
that enabled Oliver as late as 1949 to describe one
new genus and six new species of moas, all of
which were based on a small number of isolated
bones. As will be seen, none of these taxa appears
to be valid.

It is my purpose here to propose species-limits
within moas which it is hoped will reflect the bio¬
logical structure of those species. I have attempted
to obtain some estimate of intraspecific variability
and to apply this to recognizing species-limits. Spe¬
cies based on isolated bones or on inadequate ma¬
terial are critically evaluated for validity; likewise,
geographic representatives on separate islands,
previously given separate species names, are here
considered conspecific unless there is good evidence
to the contrary. Species-limits of moas undoubtedly
will remain a matter of personal opinion for some
time to come, and I make no pretense at having
arrived at a definitive picture. There is a need for
additional studies, and the systematics of these
birds would benefit especially from a comprehen¬
sive numerical-multivariate approach. I do believe,
however, that the species-limits proposed in this
paper considerably clarify moa taxonomy and
facilitate more interesting studies of their
evolution.

Materials and Methods.— I have studied mate¬
rial of moas in the following museums (abbrevia¬
tions used in the text follow in parentheses):
British Museum (Natural History) (BM); Ameri¬
can Museum of Natural History; Field Museum of
Natural History; Canterbury Museum, Christ¬
church; Otago Museum, Dunedin; and National
Museum of New Zealand (formerly Dominion
Museum) (DM), Wellington. During this study the

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

only major collection I was unable to examine was
that at the Auckland Museum (AM), Auckland.
Fortunately, Archey’s (1941) valuable monograph
on the moas is based almost entirely on the Auck¬
land collection and I was, therefore, able to incor¬
porate much information on that material into
this study.

Of the species recognized by Oliver (1949) I
have examined material of all except Pachyornis
murihiku, Anomalopteryx antiquus, Megalapteryx
hectori, and M. benhami. All of these were de¬
scribed from isolated bones and it is probable that
none represents a valid species, with the possible
exception of M. benhami. Thus, I was able to study
the majority of taxa in need of critical evaluation.

In addition to using standard univariate statisti¬
cal procedures, I have employed several multi¬
variate morphometric techniques in order to
characterize patterns of intra- and interspecific
variability in more detail. The theory and meth¬
odology of multivariate approaches and their ap¬
plication to biological problems are discussed by
Blackith and Reyment (1971) and Oxnard (1973).
Basically, these techniques describe patterns of
variation or degrees of similarity (or difference)
for many variables taken simultaneously over
many taxa. I have used two techniques: (1) prin¬
cipal components analysis (BMDOIM; Dixon,
1970) in order to examine the structure of varia¬
tion within groups, primarily to investigate prob¬
lems of sexual dimorphism in size and shape
within a species; (2) discriminant function-
canonical analysis (BMD07M; Dixon, 1970), in
order to examine the patterns of variation among
groups that are defined prior to the analysis. I em¬
ployed this approach to examine the nature of the
separations among the presumed species of a genus
to evaluate species distinctness and the presence of
sexual dimorphism. Part of the output of
BMD07M is a posterior probability classification
which allows one to discover whether individuals
assigned to one group prior to the analysis are in
fact closer to the means of another group.

In all examples employing multivariate tech¬
niques, I undertook the analysis of each hindlimb
element based either on my own data or that in
Archey (1941) and Oliver (1949) using the follow¬
ing four variables: bone length, breadth of proxi¬
mal end, breadth of shaft at midpoint, and breadth
of distal end.

NUMBER 27

191

Acknowledgments. —It is with pleasure that I
dedicate this paper to Dr. Alexander Wetmore in
honor of his ninetieth birthday, his many contribu¬
tions to avian paleontology and ornithology, and
especially for the affection he has shown us all.

I want to thank the following institutions and
individuals for making their collections available
to me and for making my visits enjoyable and pro¬
ductive: American Museum of Natural History
(Wesley Lanyon, Malcolm C. McKenna); British
Museum (Natural History) (Alan J. Charig,
Cyril A. Walker); Canterbury Museum (Roger S.
Duff, Ronald J. Scarlett, Michael Trotter); Field
Museum of Natural History (John Bolt, Melvin A.
Traylor); National Museum of New Zealand
(John C. Yaldwyn, R. K. Dell); and Otago Mu¬
seum (Roy Forster, J. Darby). I am grateful to
Hildegard Howard and Pat V. Rich, John C.
Yaldwyn, and Sir Robert Falla for providing help¬
ful comments on an early draft of this paper. My
trip to New Zealand and subsequent data analysis
were made possible by financial support from the
Australian Academy of Science (through H. J.
Frith), the Chapman Fund of the American Mu¬
seum of Natural History, and the National Science
Foundation (grant GB-41089).

Intraspecific Variability

The relative variability of one species of kiwi
and four species of moas, all of which are believed
to represent “good” biological species, are ex¬
amined here. By gaining some understanding of
the degree of variability within species recognized
to be valid by nearly all previous workers, a basis
of comparison can thereby be provided for assessing
species-limits among the more controversial taxa
considered in the following section. For reasons to
be discussed below, these assessments of variability
within “good” species cannot be used as absolute
limits or criteria of species variability, but they can
serve as guidelines.

Causes of Variability in Moas

It has been known for 100 years or more that
moas are highly variable. Moreover, some workers
have seemed to appreciate the fact—although they
seldom stated so explicitly—that this variation is
complex in nature and cannot be attributed to any

single factor. The main obstacle to understanding
this variation is that of dealing with fossil (perhaps
more correctly, subfossil) populations and their
well-known problems of sampling in space and
time. Some workers have sought to solve these
problems either by ignoring them or by naming
new species for each variant and thereby elimina¬
ting the necessity of delimiting or explaining intra-
specific variability (note comments by Oliver cited
above).

At this time it is not possible to make quantita¬
tive estimates of the separate factors contributing
to patterns of variability within species of moas.
Sample sizes for some species over their entire
range, or for local populations of most species, are
usually too small; stratigraphic control is lacking
for all but a few of the moa specimens collected so
far. Consequently, one is forced to estimate intui¬
tively the relative importance of temporal, geo¬
graphic, and individual components of variation.
Within most fossil populations, particularly of
moas, the following five factors seem important.

Intrapopulational Variation.— This is the vari¬
ation observed in individuals of the same local
interbreeding populations. We might expect the
amount of such variation to be relatively low com¬
pared to those samples in which the influences of
geographic or temporal variation are also present.
Most of the samples discussed in this section do not
represent single local populations; such samples
probably do exist for certain species, but restric¬
tions of time while in New Zealand did not permit
me to study this aspect of variation in detail. Cer¬
tain samples of Apteryx australis and Euryapteryx
curtus (including E. exilis ) might provide exam¬
ples of this type of variation, and these are dis¬
cussed below.

Sexual Dimorphism.— Since sexual size dimor¬
phism is common in other ratites (with either
males or females being larger), one would expect
moas to show sexual size dimorphism also. Unfor¬
tunately, sample sizes are usually inadequate to
lend support to this, although evidence is pre¬
sented below of four probable examples of sexual
size dimorphism in moas. The question of how
much size difference to expect between sexes must
be considered when dealing with fossil populations:
if there were too great a difference between two
particular samples, then we might be inclined to
recognize two species rather than two sexes. But

192

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

what is “too great” a difference? Almost no quanti¬
tative information has been published about di¬
morphism in other ratites. In one study of the
Emu (. Dromiceius novaehollandiae), data gathered
by Long (1965) show that coefficients of variation
(CV) of combined male-female samples range from
3.5 for bill length to 19.14 for body weight. Co¬
efficients of variation for lengths of a few skeletal
elements range from 4.77 for the tarsometatarsus
to 5.28 for the tibiotarsus. These CVs for the skele¬
tal elements are generally less than those calcu¬
lated for the expanded species of moas discussed in
the next section.

Geographic Variation. —The major factor in
this type of variation in moas is probably inter¬
island differentiation. It has previously been noted
that the bones of North Island forms tend to be
shorter and less stout than those of comparable
taxa from the South Island (Archey, 1941:62, 71;
Scarlett, 1972:20; Oliver, 1949:164). In addition,
some intra-island differentiation may have oc¬
curred, but samples are too limited to confirm this.
I believe geographic differentiation contributes
greatly to the large variability in the samples of
certain species discussed below.

Temporal Variation.— It is difficult to assess the
importance of temporal variation in affecting
variability within moas. The chronology of natural
moa deposits is not well known. One can be reason¬
ably certain that these assemblages are no older
than 7000-8000 years, and most are undoubtedly
much younger, apparently less than 4000 years old
(Fleming, 1962). Hence, it may be that temporal
variation contributes relatively little to the vari¬
ability of the available samples of moas.

Variation and Recency of Sympatry.— If spe-
ciation in moas has resulted from isolation in forest
refugia during glacial maxima, then it is reason¬
able to assume that variability would have in¬
creased as a result of this isolation, thus providing
an example of intra-island geographic variation.
Many samples of moas probably are composed of
different populations that had come in contact
following the last glaciation. Thus, we may be
sampling birds that had recently diverged morpho¬
logically, and some of the variation observed may
be the result of recent character displacement in
size following this contact.

Analysis of Species

Basic statistical data for the femora, tibiotarsi,
and tarsometatarsi of Apteryx australis and four
species of moas are given in Table 1. Of particular
importance for the discussions that follow are the
coefficients of variation (CV), which are measures
of relative variability independent of size.

Apteryx australis: This sample (housed in the
National Museum of New Zealand) of 24—32 indi¬
viduals referable to the modern Brown Kiwi, comes
from the Castle Rocks cave deposit on the South
Island. It is evident from Table 1 that A. australis,
with CVs ranging from 4.68 to 7.6, exhibits less
variability than any of the moas. I attribute this to
the relative lack of geographic and temporal in¬
fluences since the sample comes from a single lo¬
cality and was probably deposited over a relatively
short span of time. Of the species studied, this
sample of A. australis possibly comes closest to rep¬
resenting only intrapopulational variation. The
degree of variability in this sample is similar to
that shown by fossil populations of some species of
gruiforms (Cracraft, 1973b). In A. australis, meas¬
urements of length are less variable than those of
breadth.

Megalapteryx didinus: This sample is taken
from many localities in the South Island (data
from Archey, 1941; Oliver, 1949). In a later section
I synonymize M. hectori with M. didinus, but speci¬
mens assigned to the former are not included in
this sample. This species may be one of the more
variable of moas in that the lowest CV is 5.79
while the highest is 12.04. Most of this variability
is probably attributable to geographic variation.
As with A. australis, measurements of length are
less variable than those of breadth.

Anomalopteryx didiformis: The analysis of this
species is based on a large sample collected from
sites on both islands. Included in this series by
Archey (1941, table A) are a few specimens that
might be assigned to Oliver’s (1949) species A.
parvus. Few workers accept A. parvus as a distinct
species (see below) and my analysis is based on all
the specimens listed by Archey. This species also
shows a fairly high degree of variability, with CVs
ranging from 6.79 to 9.3. Again, lengths are less
variable than other measurements. The variability
of this sample is probably influenced substantially
by geographic and intrapopulational variation.

NUMBER 27

193

Table 1.—Statistics for Apteryx australis and four species of moas (measurements in mm)

Character

Apteryx

australis

Megalapteryx

didinus

Anomalopteryx

didiformis

Dinornis

torosus

Pachyornis

elephantopus

Femur

Length

n .

24

16

38

16

21

x .

89.45

245.69

238.45

295.38

308.86

SD .

4.43

15.29

16.2

17.73

15.54

cv .

4.96

6.22

6.79

6

5.03

Proximal breadth

n .

24

13

37

12

20

x .

20.1

71.62

82.27

108.83

129.55

SD .

1.53

6.19

7.65

10.21

12.11

CV .

7.6

8.64

9.3

9.38

9.35

Distal breadth

n .

24

15

38

12

21

x .

21.45

75.73

85.53

117.58

145.57

SD .

1.13

7.72

7.89

11.75

8.84

CV .

5.25

10.2

9.26

9.99

6.07

Tibiotarsus

Length

n .

32

15

39

14

27

x .

129.93

385.33

377.21

588.36

542.93

SD .

6.08

22.3

27.33

36.77

37.97

CV .

4.68

5.79

7.25

6.25

6.99

Proximal breadth

n .

32

12

33

10

19

x .

17.31

85.58

101.27

139.7

167.42

SD .

1.06

7.75

8.9

8.34

14.32

CV .

6.11

9.06

8.79

5.97

8.56

Distal breadth

n .

32

15

38

13

27

x .

15.71

51.93

56.71

82.07

97.33

SD .

0.88

6.25

4.89

6.86

9

CV .

5.59

12.04

8.62

8.36

9.25

T ARSOMETATARSUS

Length

n .

30

18

43

16

29

x .

64.85

179.39

179.02

303.13

230.55

SD .

3.3

10.9

13.96

22.75

15.75

CV .

5.09

6.07

7.8

7.51

6.83

Proximal breadth

n .

29

16

41

13

27

x .

17.17

53.25

59.15

88.54

105.85

SD .

0.98

5.04

5.34

5.08

10.25

CV .

5.73

9.46

9.03

5.73

9.68

Distal breadth

n .

30

17

42

14

29

x .

21.05

72.41

77.43

112.71

134.52

SD .

1.11

7.4

6.26

7.62

12.06

CV .

5.29

10.22

8.08

6.76

8.97

194

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Dinornis torosus: The data for this species are
based on a small sample collected from many sites
in the South Island (Archey, 1941; Oliver, 1949).
Variability is fairly high, with CVs ranging from
5.73 to 9.99. This is the only species in which
lengths tend to be more variable than breadths.
Variability in this sample is probably affected by
geographic and temporal components.

Pachyornis elephantopus: This sample is from
various sites on the South Island (Archey, 1941;
Oliver, 1949; and measurements by the author).
Coefficients of variation are comparable to those of
A. didiformis and D. torosus, ranging from 5.03 to
9.68. Once again, length measurements are less
variable than those of breadth. Intra-island com¬
ponents probably contribute substantially to vari¬
ability in this species.

Summary

Variability within individual species of moas ap¬
pears somewhat higher than is found in most other
birds so far studied. For example, in six fossil spe¬
cies of gruiforms, CVs for hindlimb elements
averaged 6-7 (Cracraft, 1973b: 97-107). Simpson
(1946) presented data for the King Penguin
(.Aptenodytes patagonicus) and Goodge (1951)
analyzed variability in the Common Murre (Uria
aalge ); both workers found low CVs, averaging
2-4, in these extant species.

Data for A. didiformis, D. torosus, and P. ele¬
phantopus would seem to suggest that in moas CVS
generally fall in the range of 6-10, but if the sam¬
ple of M. didinus were used as a standard, then
CVs as high as 10-12 might be expected within a
single species. Indeed, estimates based on the above
samples may be conservative in that they possibly
include only individuals of a single sex, specimens
of the other sex previously having been considered
as forming a distinct species.

During growth, the length of long bones ap¬
parently reaches its maximum value before maxi¬
mum body weight is obtained (Cock, 1963). Bone
breadths of adults, on the other hand, probably
reflect final body weight since they have a mechani¬
cal relationship to the amount of weight that can
be supported. It can thus be expected that breadth
measurements will tend to be more variable than
those for length, since the former depend upon
variation in body weight at the termination of

growth. Estimates of variation in length might be
preferred over those for breadth as a more precise,
and more conservative, measure of intraspecific
variability. For this reason the comparisons in the
following section will be based on measurements
of length.

Systematics

In this section I have attempted to formulate
species limits based on information from univariate
and multivariate analyses of variation, personal
examination and comparison of many complete
and partial skeletons, and an evaluation of pre¬
viously published opinions on moa systematics. De¬
tailed discussions of morphology (particularly at
the generic level and higher), natural history, and
taxonomic synonymies can be found in Archey
(1941), Oliver (1949), Brodkorb (1963), or papers
cited therein, and are not included here unless they
bear directly on the subject of species limits.

Family DINORNITHIDAE
Subfamily ANOMALOPTERYGINAE

Anomalopteryx Reichenbach, 1852

Summary. —Two species of Anomalopteryx are
tentatively admitted here—a larger form, A. didi¬
formis, found on both North and South islands,
and a smaller, less common form, A. oiueni, found
only on the North Island.

Anomalopteryx didiformis (Owen, 1844)

Synonyms. — Anomalopteryx parvus (Owen,
1883), Anomalopteryx antiquus Hutton, 1892.

Of the two species of the genus recognized here,
this was the more common and occurred on both
the North and South islands. Archey’s (1941:14-29)
discussion, although not employing statistical
methods, amply demonstrated the great variation
present within this species.

Anomalopteryx parvus, based on fairly extensive
material from both islands, is included in this spe¬
cies, following Archey (1941), Brodkorb (1963),
and Scarlett (1972). This is contrary to Oliver
(1949:138), who maintained A. parvus as distinct.
Oliver (1949:144-145, figs. 115-116) pictured bones

NUMBER 27

195

of the two species for comparative purposes, and
indeed these appear quite different in size. Oliver’s
method of analysis and argumentation, however,
was to compare the extremes in size or shape, thus
magnifying the differences. There is a continuity
in size and morphology in specimens assigned to
these two species (Archey, 1941:18); thus, unlike
examples to be described below in other genera, it
is not possible to recognize two size groups within
A. didiformis that could represent sexual size
dimorphism.

According to Scarlett (1972:22), A. antiquus is
now considered to come from lower Pleistocene
deposits rather than being Miocene or Pliocene in
age as previously thought. Scarlett further indicated
diat A. antiquus is ‘‘doubtfully distinct from didi¬
formis” and that there are no significant morpho¬
logical differences between them. Likewise, Archey
(1941:29) noted close similarities between the two
species. I did not locate the type during my stay at
the Canterbury Museum. At present I believe that
compelling evidence is lacking to maintain A.
antiquus as a separate species.

Anomalopteryx oweni (Haast, 1885)

There has been some controversy about the
generic assignment of this species. As did all other
early workers, Haast (1885, 1886) placed this spe¬
cies in Dinornis. Shortly thereafter, Lydekker
(1891:280) put oweni in Anomalopteryx on the
basis of its skull morphology. Archey (1941:44)
transferred the species to Pachyornis without com¬
ment; Brodkorb (1963:211) followed Archey.
Oliver (1949:134-135) returned oweni to Anomal¬
opteryx, claiming that the type cranium illustrated
by Haast (1886) shows the diagnostic features of
Anomalopteryx, as does the associated premaxilla,
and that most of the skeletal material assigned to
oweni by Archey belongs to Pachyornis septentri-
onalis ( = P. mappini of this paper).

Although I was unable to examine the type-
material of A. oweni in the Auckland Museum,
several comments on the species can still be made.
The cranium and premaxilla illustrated by Haast
do appear to be more similar to Anomalopteryx
than to Pachyornis, although the rounded anterior
border of the temporal fossa in dorsal view and the
markedly sloping nasal region in lateral view do
not resemble species of either genus. Some of the

hindlimb elements listed by Archey as belonging to
oweni are within the size range of Pachyornis
mappini (including P. septentrionalis ), whereas
others appear to be too small to be referred to that
species. Thus, until the systematics of oweni can
be clarified by restudy of the type and comparison
with other material, I tentatively include it as a
valid species in the genus Ano?nalopteryx.

Megalapteryx Haast, 1886

Summary.— Two species of Megalapteryx are ac¬
cepted here—a small one, M. didinus, and a larger
one, M. benhami. Both are known from the South
Island, the alleged presence of M. didinus in the
North Island being doubtful.

Megalapteryx didinus (Owen, 1883)

Synonym.— Megalapteryx hectori Haast, 1886.

Megalapteryx diclinus is known from a moderate
number of bones from the South Island, few of
which were found in association. The presence of
this species on the North Island is suspect (Oliver,
1949:151-152).

Archey (1941) synonymized M. hectori with M.
didinus without comment, but Oliver (1949:149)
maintained the species. Brodkorb (1963) and
Scarlett (1972) followed Archey. I also believe
there is insufficient evidence to justify recognition
of M. hectori. Only a few limb bones are presum¬
ably applicable to this species (I did not examine
the type, which is in the Nelson Museum), and
their size, although somewhat smaller than typical
bones of M. didinus, probably falls within the lim¬
its of variability for that species. For example, CVs
for the lengths of the hindlimb elements for the
combined sample of M. didinus and M. hectori are
comparable to CVs of the other moas listed in
Table 1: e.g., femur, 8.45; tibiotarsus, 6.76; and
tarsometatarsus, 7.40. I therefore follow the au¬
thors cited above in synonymizing hectori with
didinus.

Megalapteryx benhami Archey, 1941

This species was described on the basis of a femur
and a tibiotarsus, not positively associated, from
the Mt. Arthur region, South Island. Oliver (1949)
also lists a femur from Wairanga, South Island.

196

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

The bones were placed in Megalapteryx on the
basis of several morphological characters associated
with the rotular groove and muscle scars (Archey,
1941:35), and it would be important for future
workers to confirm the validity of these characters
in distinguishing genera. As many authors have
noted, there is great variability in the form of the
femur and it is frequently difficult to identify this
element to genus unless found in association with
more diagnostic bones. As it now stands, the spe¬
cies M. benhami can be tentatively accepted, as its
larger size (mean femur length of 296.5 mm;
length of tibiotarsus, 454 mm) is very probably out¬
side the range of variation for M. didinus. Future
workers need to restudy the material of this species
in order to verify its validity and generic assign¬
ment.

Pachyornis Lydekker, 1891

Summary.— Two species of Pachyornis are rec¬
ognized here. The larger, P. elephantopus, is
known only from the South Island. A smaller spe¬
cies, P mappini, showing what appears to be
pronounced sexual dimorphism in size, was re¬
stricted to the North Island.

Pachyornis elephantopus (Owen, 1856)

Synonyms. — Pachyornis murihiku Oliver, 1949;
Pachyornis australis Oliver, 1949.

Pachyornis elephantopus was restricted to the
South Island and was the larger of the two species
of the genus recognized here. It was also one of
the more common species of moas and is repre¬
sented by a number of complete skeletons from the
Pyramid Valley Swamp.

Oliver (1949:67) described P. murihiku for a
single skeleton said to be from a “full-grown but
not quite mature” individual from Southland,
South Island. The type was supposedly in the
Southland Museum, Invercargill, but according to
Scarlett (1972:21) it cannot now be found. Both
Brodkorb (1963) and Scarlett (1972) accepted P.
murihiku as a distinct species. The measurements
of P. murihiku indicate that it was only slightly
smaller than P elephantopus (Oliver, 1949:59,
86-87); furthermore, if the measurements of the
limb bones of P. murihiku are included in the

sample of P. elephantopus in Table 1, the CVs of
bone length are not appreciably increased; viz.
femur, 6.52; tibiotarsus, 8.04; and tarsometatarsus,
7.67. Oliver (1949:68-70) listed the following char¬
acters as distinguishing P. murihiku from P. ele¬
phantopus: (1) culmen rises at a higher angle,
(2) the “front of the body [of the sternum] is not
bent upwards so far as to bring it to a right angle
with the rest of the body,” and (3) the ischia and
pubes are widely diverging. It is difficult to evalu¬
ate these features in terms of species differences,
especially since the type of P. murihiku is from an
immature individual. In the absence of additional
material it may be questioned whether the above
differences should be accepted as being consistent
between the two species. There is a great deal of
variation in skeletons of P. elephantopus, and I be¬
lieve that the above features of P. murihiku must
be verified by additional specimens before this spe¬
cies is accepted as valid.

Oliver (1949:70) described another species of
Pachyornis, P. australis, from a single well-
preserved cranium from Takaka River, South
Island. Both Brodkorb (1963) and Scarlett (1972)
accepted P. australis as a distinct species, although
Scarlett indicated that it may be a variant of P.
elephantopus. I have compared the type (DM 26)
with a large series of skulls of P. elephantopus and
can find no significant differences that can be re¬
garded as being of specific value. Even though the
skull of P. australis is at the lower end of the size
range for P. elephantopus, several skulls assigned
to the latter species by Oliver himself (1949:84)
are of comparable size (e.g., DM 95, DM 198,
DM 333). Crania of P. elephantopus exhibit con¬
siderable variability in shape and in the develop¬
ment of processes and muscle scars. For example,
within a series of skulls of P. elephantopus in the
National Museum of New Zealand it was possible
to find the following characters attributed to “P.
australis” by Oliver (1949:70-72): (1) skull wider
in proportion to length, (2) evenly rounded
cranium, (3) wide space between lambdoidal and
temporal ridges, and (4) narrow temporal fossa.
Furthermore, there are similarities in rostral
shape and considerable variation in the develop¬
ment of the transverse process of the basisphenoid
rostrum. Therefore, it does not seem prudent to
accept P. australis as a valid species.

NUMBER 27

197

Pachyornis mappini Archey, 1941

Synonym.— Pachyornis septentrionalis Oliver,
1949.

Archey (1941:41) proposed this species for a
small North Island form of Pachyornis. The type
(AM 124) is an almost complete skeleton. Most of
the remaining material assigned to P. mappini
consists of isolated elements and many of those in
the National Museum of New Zealand have been
acquired recently and are as yet undescribed.

Oliver (1949:61) described a new species, P
septentrionalis, for a partial skeleton (DM 129),
also from the North Island. He included in this
species those bones from the lower end of the series
that Archey (1941) placed in P mappini. Oliver
(1949:61) stated that bones of P. septentrionalis,
in addition to being smaller, are also more slender
than those of P. mappini. Brodkorb (1963) ac¬
cepted both species, while Scarlett (1972) suggested
that the two might be conspecific.

An analysis of the skeletal measurements of these
nominal species (Archey, 1941:139; Oliver, 1949:
86; measurements by the author) indicates that
two separable populations do appear to exist, the
major difference between them being in size
(Table 2). Using principal components analysis of
each element of the hindlimb (Figure 1), a mod¬
erately well-defined separation between the two
forms can be demonstrated along the first principal
component, which in this case is a size axis. The
second component is a shape axis, mainly defining

Figure 1. —Principal components analyses of logarithmically
transformed measurements of the femur (a), tibiotarsus (6),
and tarsometatarsus (c) of Pachyornis septentrionalis (dark
circles) and P. jnappini (open squares). The first principal
component of each analysis is graphed along the abscissa and
is a size axis; the second principal component is graphed
along the ordinate and is a shape axis, primarily indicating
relative robustness. (Note that the two taxa are separated by
size but not by shape. See text for details.)

relative robustness, and it is evident that the two
populations do not differ significantly in this re¬
spect. Thus, Oliver’s (1949:61) claim that bones of
P. septentrionalis are more slender than those of P.

Table 2. —Statistics for Pachyornis mappini and P. septentrionalis (measurements in mm)

Character

P. mappini

P. septentrionalis

P. mappini +

P. septentrionalis

Length of femur

n .

12

14

26

x .

206.25

174.07

188.92

SD .

12.88

11.63

20.27

CV .

6.24

6.68

10.73

Length of tibiotarsus

n .

12

12

24

x .

365.5

291.67

328.58

SD .

30.44

14.79

44.38

CV .

8.33

5.07

13.51

Length of tarsometatarsus

n .

7

10

17

x .

156.29

133

142.59

SD .

12.57

3.71

14.37

CV .

8.04

2.79

10.08

198

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Table 3. —Posterior probability classification of stepwise discriminate function analysis for bones
assigned to Pachyornis mappini and P. septentrionalis

Character

P.

mappini

P.

septentrionalis

n

Percent

misclassified

Femur

P. mappini .

11

1

12

8.3

P. septentrionalis .

1

13

14

7.1

Tibiotarsus

P. mappini .

12

0

12

0

P. septentrionalis .

0

12

12

0

T arsometatarsus

P. mappini .

7

0

7

0

P. septentrionalis .

0

10

10

0

mappini is not substantiated. In order to examine
further the distinction between the two forms, I
analyzed the data using a step-wise discriminate
function-canonical analysis. The distinctness of the
two groups was further verified and few of the ele¬
ments were misclassified (Table 3).

What is the meaning of these differences? If, in
fact, the two taxa do not represent distinct species,
then very likely we are dealing with sexual size dif¬
ferences. Table 2 presents some basic statistical
data for lengths of the hindlimb bones. The com¬
bined sample of measurements for the lengths of
the femur and tarsometatarsus do not show CVs
much higher than those for P mappini alone, or
for those of other moas (Table 1). It can be noted
also that the CV of 2.79 for the tarsometatarsus
length of P. septentrionalis (Table 2) is suspi¬
ciously low compared to CVs of other moas. It is
my belief that these two skeletal populations prob¬
ably represent different sexes and that Oliver’s
(1949) description of a new species was unwar¬
ranted. I therefore synonymize septentrionalis with
mappini until firm evidence can be offered that
they are distinct.

Euryapteryx Haast, 1874

Synonym.— Zelornis Oliver, 1949.

Oliver (1949:117-128) created the genus Zelor¬
nis for the species Euryapteryx exilis Hutton (the
genotype) and Emeus haasti Rothschild. Archey
(1941) considered the former to be a valid species
in the genus Euryapteryx, while the latter he
treated as a synonym of Euryapteryx gravis (p. 54).
The diagnostic feature separating Zelornis from
Euryapteryx was said to be the high arched culmen.

Oliver (1949:110) admitted that there were no
differences in the shape of the postcranial ele¬
ments. The type-specimen of Z. exilis is a skeleton
from Wangaehu in the Wanganui Museum. The
skull was figured by Oliver (1949, figs. 92-94)
where it is readily apparent that the premaxilla is
considerably broken and abraded. The premaxilla
of Z. exilis does not appear to differ in shape from
those referred to Z. haasti. I have examined nearly
all of the cranial material assigned to Zelornis
haasti by Oliver (1949:127) and can find no impor¬
tant differences in size or shape that will distin¬
guish it at the generic level from Euryapteryx. To
my knowledge Scarlett (1972) is the only recent
author to synonymize Zelornis with Euryapteryx,
and I concur with his decision.

Summary. —Two sexually dimorphic species are
recognized here—a moderately large form, E. gera-
noides, present on both North and South islands,
and a small species, E. curtus, confined to North
Island.

Euryapteryx curtus (Owen, 1846)

Synonyms.— Euryapteryx exilis Hutton, 1897;
Euryapteryx tane Oliver, 1949.

Euryapteryx curtus was a small species of moa,
apparently confined to the North Island. The only
morphological difference between E. curtus and E.
exilis is in size (Archey, 1941:60), E. exilis being
slightly larger (Table 4). In order to assess the
morphological similarities in size and shape in
Euryapteryx, I analyzed the measurements of the
femur, tibiotarsus, and tarsometatarsus given by
Archey (1941) and Oliver (1949), using principal
components and canonical analyses. Figure 2 plots

NUMBER 27

199

Table 4.—Statistics for species of Euryapteryx (measurements in mm)

E. tane

Character
Length of Femur

n

x

SD

CV

3

190.67

9.29

4.87

Length of tibiotarsus
n .

x

SD

CV

3

328

9

2.74

Length of tarsometatarsus

n

x

SD

CV

3

149

7.55

5.07

E. curtus

15

167.6

12.82

7.65

20

268.8

16.5

6.14

19

124.74

8.29

6.65

group centroids for the first two canonical axes.
The multivariate analysis substantiates the separa¬
tion of E. exilis and E. curtus, but the centroids
themselves as projected onto the first axis are only
about two and a half SD units from each other
(except for the tibiotarsus where they are almost
four units apart). If one examines the statistics of
the combined sample (Table 4), the population
E. exilis + E. curtus has CVs of about 10-12 for
bone lengths. This variability is comparable to
that of Pachyornis mappini, and I believe it is
likely that the difference between E. exilis and E.
curtus is one of sexual size dimorphism. Most of
the specimens in Archey’s series of these two forms
(1941, tables H, I) come from Doubtless Bay,
North Island.

Oliver (1949:105) described a separate species,
E. tane, for a small number of specimens that were
larger than those of E. curtus. He (1949:123) noted
that some of the leg bones of E. tane possibly be¬
longed to what he called Zelornis exilis. It is readily
apparent that the samples of E. tane and E. exilis
are virtually identical (Table 4; Figure 2). There is
little question, therefore, that E. tane should be
synonymized with E. curtus as defined here.

Euryapteryx geranoides (Owen, 1848)

Synonyms.— Euryapteryx gravis (Owen, 1870),
Zelornis haasti (Rothschild, 1907).

The type-material of E. geranoides consists of a

E. exilis E. exilis + E. geranoides E. gravis E. gravis -f
E. curtus E. geranoides

18

33

8

13

21

198.17

184.27

236.63

280.15

263.57

10.89

18.24

12.53

10.89

24.4

5.5

9.9

5.3

3.89

9.26

17

37

5

12

17

331.65

297.68

387.6

475.17

449.41

12.67

34.97

17.99

21.57

45.74

3.82

11.75

4.64

4.54

10.18

15

34

9

13

22

148.4

135.12

174.78

209.62

195.36

4.39

13.64

12.34

16.16

22.68

2.96

10.09

7.06

7.71

11.61

cranium, premaxilla, and mandible collected at Te
Rangatapu, North Island, and housed in the Brit¬
ish Museum (Natural History). According to Oliver
(1949:106) the mandible belongs to Anomalop-
teryx didiformis, and the cranium and premax¬
illa, which may or may not be associated, belong to
Euryapteryx. To my knowledge a lectotype has not
yet been designated, but the cranium (BM 21687)
would be a logical choice. The type cranium and
premaxilla are smaller than most of the material
attributed to E. gravis but larger than in E. curtus.
Therefore, E. geranoides has been accepted as a
distinct species of intermediate size by Archey
(1941), Oliver, (1949), and Brodkorb (1963). It
should be pointed out that there is apparently no
postcranial material directly associated with cra¬
nial material attributable to E. geranoides (sensu
stricto).

Material assigned to the large form known as E.
gravis is abundant in South Island localities (the
type-skeleton in the British Museum is from Kaka-
nui) but very rare on the North Island (Archey,
1941:54-56; Oliver, 1949:108-112). Statistics pre¬
sented in Table 4 and the results of the canonical
analyses shown in Figure 2 confirm the intermedi¬
ate position of E. geranoides between “E. exilis”
( =E. curtus) and E. gravis. The question is
whether E. geranoides is a distinct species, and if
not, to which species—the larger E. gravis or the
smaller E. curtus—this skeletal population belongs.
Recently, Scarlett (1972:21) suggested that E. gera-

200

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Figure 2. —Group centroids of five nominal taxa of Euryapteryx plotted against the first
(abscissa) and second (ordinate) canonical axes for logarithmically transformed measure¬
ments of the femur (a), tibiotarsus (fi), and tarsometatarsus (c). Scales are in standard
deviation units. Note the closeness of group centroids of E. tane and E. exilis, the closeness
of E. curtus to E. tane-E exilis, and the intermediate position of E. geranoides between
E. gravis and the smaller forms. (See text for details.)

noides might be united with E. curtus once the
gaps were eliminated.

I would like to suggest here that E. geranoides is
conspecific with E. gravis and that these forms rep¬
resent another case of sexual size dimorphism. The
evidence is two-fold. First, measurements of the
combined sample exhibit CVs very similar to those
seen in P. mappini-P. septentrionalis and E.
curtus-E. exilis, the two other presumed examples
of sexual size dimorphism (Table 4). Secondly,
bones attributed to E. geranoides and E. gravis
occur on both North and South islands. If E. gera¬
noides were conspecific with E. curtus (or with E.
exilis, if this form were distinct from E. curtus ),
then the absence of E. curtus from the South Island
is unexplained. If E. geranoides and E. exilis rep¬
resented different sexes, then both should be pres¬
ent on the South Island. Present evidence, there¬
fore, is more consistent with the hypothesis that E.
curtus-E. exilis constitute one sexually dimorphic
species and E. geranoides-E. gravis another. In the
case of the latter, the older name, geranoides, has
priority.

As noted above, the cranial material of Zelornis
haasti is very similar to that of E. geranoides. The
femur and tibiotarsus of the one skeleton of Z.
haasti are somewhat larger than typical ”E. gravis”
(Oliver, 1949:128; see also Table 4), but the as¬
sociated tarsometatarsus is easily within the size
range of that form. Hence, it is likely that the few
bones assigned to Z. haasti are large, perhaps aber¬
rant bones of E. geranoides, and I follow Archey
(1941) in synonymizing haasti.

Emeus Reichenbach, 1852

Emeus crassus (Owen, 1846)

Synonym.— E. huttonii (Owen, 1879).

There has been little difference of opinion about
species-limits within Emeus. Most recent authors
(Archey, 1941; Oliver, 1949; Brodkorb, 1963) have
accepted two species, the large crassus and the
smaller huttonii. Only Scarlett (1972:22) has com¬
bined the two species, stating that several speci¬
mens from Pyramid Valley are intermediate in

NUMBER 27

201

Table 5. —Statistics for species of Emeus
(measurements in mm)

Character

E. crassus

E. huttonii

E. crassus +

E. huttonii

Length of femur
n .

21

10

31

272.81

238.4

261.71

SD .

12.23

11.92

20.24

CV .

4.48

5

7.73

Length of tibiotarsus

n .

21

11

32

464.24

386.64

437.56

SD .

23.08

19.01

43.16

CV .

4.97

4.92

9.86

Length of tarsometatarsus

n .

22

10

32

213.55

184.4

204.44

SD .

11.58

13.46

18.22

CV .

5.42

7.3

8.91

size. Both forms occurred in the South Island and
are known primarily from the Canterbury and
Otago districts. Although according to Archey
(1941:51) several bones referable to Emeus have
allegedly been found at Martinborough and Te
Aute on the North Island, Yaldwyn (1956) does
not list Emeus from the Martinborough Caves, nor
did Oliver (1949) make note of Emeus on the
North Island. On the South Island E. crassus has
been reported as moderately common, whereas E.
huttonii was apparently less so.

Statistics presented in Table 5 show that the
bones assigned to E. crassus and E. huttonii by
Archey (1941) and Oliver (1949) comprise two
distinct populations. Multivariate analysis of the
hindlimb measurements verifies the univariate re¬

sults, and in the posterior probability classification
of the discriminate function analysis, few bones are
misclassified (Table 6).

The two forms of Emeus could represent two
species, or alternatively, they might represent
sexual dimorphism within a single species. It is my
opinion that the latter hypothesis is more probable.
First, the combined sample of the two populations
does not exhibit variability that could be consid¬
ered outside the limits for a single species of moa,
all CVs being fairly low (Table 5: femur length,
7.73; tibiotarsus length, 9.86; tarsometatarsus
length, 8.91). Secondly, the two forms appear to
have been broadly sympatric, both occurring to¬
gether in the larger fossil deposits. To my knowl¬
edge there is no good evidence that one had a
distribution exclusive of the other. If the two forms
represent dimorphic sexes, then the larger form,
crassus , would appear to have been the female, as
an egg was found preserved in association with a
skeleton of this form at Pyramid Valley (Falla,
1941). I therefore follow Scarlett (1972) in tenta¬
tively synonymizing E. huttonii with E. crassus.

Subfamily DINORNITHINAE
Dinornis Owen, 1843

Summary. —Four species of Dinornis are recog¬
nized here. The smallest, D. struthoides, was con¬
fined to the North Island. A second species, D.
torosus, was only slightly larger than D. struthoides
and was restricted to the South Island. A larger
form, D. novaezealandiae, and the largest species
of moa, D. giganteus, were both found on the
North and South islands.

Table 6.—Posterior probability classification of stepwise discriminate function analysis for bones

assigned to Emeus crassus and E. huttonii

E. crassus

E. huttonii

n

Percent

misclassified

Femur

E. crassus .

18

1

19

5.3

E. huttonii .

0

10

10

0

Tibiotarsus

E. crassus .

20

0

20

0

E huttonii .

0

10

10

0

Tarsometatarsus

F.. rrassus .

21

1

22

11

4.5

9.1

E. huttonii .

1

10

202

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Dinornis struthoides Owen, 1844

Synonym. — Dinornis gazella Oliver, 1949.

Because of a decision by the International Com¬
mission of Zoological Nomenclature (Hemming,
1954), the name D. struthoides refers to all those
specimens included under the name D. novaezea-
landiae in Archey (1941), Oliver (1949), and
Brodkorb (1963). This species was the smallest of
the genus and was found on the North Island.
Oliver (1949:170) recorded two bones from the
South Island which he assigned to this species,
noting that some of these bones were “indistin¬
guishable in size and proportions’’ from those of
the North Island. Neither Archey (1941), Brod¬
korb (1963), nor Scarlett (1972) listed this species
from the South Island and it may be that Oliver’s
specimens are referable to small individuals of D.
torosns. Thus, the presence of D. struthoides on
the South Island needs to be verified (unfortu¬
nately, I did not examine the relevant material
while in New Zealand). In any case, if D. struth¬
oides were present on the South Island, it was evi¬
dently uncommon.

Oliver (1949:166) described a new species from
the North Island, D. gazella, based on a pelvis and
some referred bones that are smaller than those
typical of D. struthoides (Table 7). The size dif¬
ferences are slight, however, and the variation
exhibited by the combined sample of struthoides
and gazella is easily within the range of a single

species (Table 8). I compared a series of bones
(DM 108: 5 tarsometatarsi, 3 tibiotarsi from Te
Aute) referred by Oliver to D. gazella with bones
of D. struthoides and found that those of the for¬
mer had thinner shafts although the two samples
were nearly equal in length. The differences ap¬
pear to be entirely related to age, as the bones
referred to D. gazella are those of immature indi¬
viduals. Therefore, I suggest that D. gazella be
merged with D. struthoides.

Dinornis torosus Hutton, 1891

This species is only slightly larger than D. struth¬
oides and has been considered the South Island
counterpart of that species (e.g., Archey, 1941:62;
Oliver, 1930). Indeed, I would be inclined to merge
D. torosus with D. struthoides if it were not for
some significant differences that apparently exist
in cranial structure.

As I will detail in subsequent papers on evolu¬
tionary relationships and cranial morphology, D.
torosus appears to be more advanced in cranial
structure than D. struthoides but in certain other
respects is more primitive than the two larger spe¬
cies D. novaezealandiae and D. giganteus. The
skull of D. torosus differs from that of D. struth¬
oides in having the basisphenoid rostrum moder¬
ately inflated, and the mandible stouter and less
deflected ventrally. The most important and con¬
sistent difference seems to be in the nature of the

Table 7-

—Statistics

for species

of Dinornis

(measurements

in mm)

Character

D.

struthoides

D.

gazella

D.

torosus

D. novae¬
zealandiae

D.

robustus

D.

Hercules

D.

maximus

D.

giganteus

Length of femur

n .

12

1

15

15

12

1

23

8

x .

265

231

295.33

340.53

354.5

353

406.13

390.88

SD .

17.41

-

18.35

13.14

15.47

-

20.73

22.2

cv .

6.57

-

6.21

3.86

4.36

-

5.1

5.68

Length of tibiotarsus

n .

9

3

14

21

15

5

32

10

x .

520.67

469.67

588.36

701.67

718.6

779.8

866.63

875.2

SD .

29.08

-

36.77

40.87

22.8

37.99

46.54

71.57

CV .

5.59

-

6.25

5.82

3.17

4.87

5.37

8.18

Length of

tarsometatarsus

n .

15

5

16

13

14

3

29

10

x .

282.2

259

303.13

368.08

382.64

420.33

463.41

486.1

SD .

19.27

12.43

22.75

30.54

17.72

-

38.78

30.08

CV .

6.83

4.8

7.51

8.3

4.63

-

8.37

6.19

NUMBER 27

203

Table 8.—Statistics for combined species of Dinornis (measurements in mm)

Character

D. struthoides +

D. gaze lla

D. novaezealandiae +

D. robustus +

D. Hercules

D. giganteus +

D. maximus

Length of femur .

n .

13

28

31

x .

262.38

346.96

402.19

SD .

19.15

15.38

21.82

CV .

7.3

4.43

5.42

Length of tibiotarsus

n .

12

41

43

x .

507.92

717.39

866.91

SD .

34.43

42.17

53.32

CV .

6.78

5.88

6.15

Length of tarsometatarsus

n .

20

30

39

X .

276.4

380.1

469.23

SD .

20.3

27.79

37.73

CV .

7.35

7.31

8.04

basisphenoid rostrum. In D. giganteus, and to a
lesser extent in D. novaezealandiae, there is a
marked inflation of the basisphenoid rostrum as
compared to the condition seen in D. torosus.
Based on the comparative series I was able to ex¬
amine, these differences are consistent between
species.

Dinornis novaezealandiae Owen, 1843

Synonyms. —Dinornis ingens Owen, 1844; D.
robustus Owen, 1846; D. Hercules Oliver, 1949.

As a result of the above-mentioned decision of
the International Commission of Zoological No¬
menclature (Hemming, 1954), D. ingens now
becomes a synonym of D. novaezealandiae. Conse¬
quently, the latter name belongs to those forms of
Dinornis that were larger than D. struthoides and
D. torosus but smaller than D. giganteus. Bones
referred to D. novaezealandiae (=D. ingens of
Archey, Oliver, and Brodkorb) are known from
both the North and South islands, although they
were apparently less common in the latter. I in¬
clude D. robustus in this species because it appears
to be little more than the South Island representa¬
tive of D. novaezealandiae. Bones referred to the
two species overlap considerably in length (Table
7), but those included under the name D. robustus
are somewhat stouter than those assigned to D.
novaezealandiae (Archey, 1941:71, Oliver, 1949:
171).

Oliver (1949:174) described an additional spe¬
cies, D. Hercules, from a few limb bones from the
North Island. The type tibiotarsus (DM 217) is
about the same length as some tibiotarsi of D.
novaezealandiae but has the shaft more curved; in
general, bones assigned to D. Hercules by Oliver are
slightly larger than those of D. novaezealandiae
(Table 7). Scarlett (1972:21) suggested that D.
Hercules may be a “bow-legged variant” of D. gi¬
ganteus, but I believe most of the specimens as¬
signed to D. Hercules are closer to D. novaezea¬
landiae in size and I here include it with that
species. I was able to compare the type of D. Her¬
cules with other species of Dinornis and in my
opinion the differences in stoutness and the curva¬
ture of the shaft of the type are attributable to
individual variation.

The combined sample of bones of D. novaezea¬
landiae, D. robustus, and D. Hercules exhibits very
little variation, all CVs being less than 7.50 (Table
8). This variability is well within that for a single
species of moa.

Dinornis giganteus Owen, 1844

Synonym.— Dinornis maximus Owen, 1867.

Dinornis giganteus is the North Island repre¬
sentative of the largest species of moa and D.
maximus is its South Island form. Bones of the
latter are somewhat stouter, but measurements of
the two overlap greatly (Table 7). I can see little

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

204

value in considering these minor variations to be
indicative of species differences. The combined
samples show a relatively small amount of varia¬
bility, certainly within the limits of a single species
(Table 8). Judging from the available collections,
the North Island form was less common than the
South Island form.

Conclusions

In contrast to the 20 species recognized by
Archey (1941) and 29 species recognized by Oliver
(1949), I here accept only 13 species as being valid.
The present arrangement is actually fairly similar
in parts to Archey’s, but combines several North
and South islands counterparts, while several
“species-pairs” are regarded as examples of sexual
size dimorphism. Further study may show that
Anomalopteryx oweni and Megalapteryx benhami,
which are based on somewhat dubious material,
perhaps do not deserve recognition. Certainly the
acceptance of the large number of species advo¬
cated by Oliver is untenable.

The systematic results of this paper can be sum¬
marized by the following classification:

Family Dinornithidae

Subfamily Anomalopteryginae
Genus Anomalopteryx Reichenbach, 1852
A. didiformis (Owen, 1844)

A. oweni (Haast, 1885)

Genus Megalapteryx Haast, 1886
M. didinus (Owen, 1883)

M. benhami Archey, 1941

Genus Pachyornis Lydekker, 1891
P. elephantopus (Owen, 1856)

P. mappini Archey, 1941
Genus Euryapteryx Haast, 1874
E. curtus (Owen, 1846)

E. geranoides (Owen, 1848)

Genus Emeus Reichenbach, 1852
E. crassus (Owen, 1846)

Subfamily Dinornithinae
Genus Dinornis Owen, 1843
D. struthoides Owen, 1844
D. torosus Hutton, 1891
D. novaeiealandiae Owen, 1843
D. giganteus Owen, 1844

The 13 species recognized here appear to have
been distributed as follows:

North Island
Anomalopteryx didiformis
A. oweni

Pachyornis mappini
Euryapteryx geranoides
E. curtus

Emeus crassus (uncertain)
Dinornis struthoides

D. novaeiealandiae
D. giganteus

South Island
Anomalopteryx didiformis

Megalapteryx didinus
M. benhami

Pachyornis elephantopus
Euryapteryx geranoides

Emeus crassus
Dinornis struthoides
(uncertain)

D. novaeiealandiae
D. giganteus
D. torosus

It is of interest to note that the smaller and topo¬
graphically less diverse North Island had fewer
species than the South Island. Also, many of the
North Island forms appear to have been smaller
than their South Island counterparts.

Literature Cited

Archey, G.

1941. The Moa: A Study of the Dinornithiformes. Bulle¬
tin of the Auckland Institute and Museum, 1:1-145.
Blackith, R. E., and R. A. Reyment

1971. Multivariate Morphometries, ix + 412 pages. New
York: Academic Press.

Brodkorb, P.

1963. Catalogue of Fossil Birds, Part 1 (Archaeopterygi-
formes through Ardeiformes). Bulletin of the
Florida State Museum, Biological Sciences, 7(4):
177-293.

Cock, A. G.

1963. Genetical Studies on Growth and Form in the
Fowl, 1: Phenotypic Variation in the Relative
Growth Pattern of Shank Length and Body Weight.
Genetical Research, 4:167-192.

Cracraft, J.

1973a. Continental Drift, Paleoclimatology, and the Evo¬
lution and Biogeography of Birds. Journal of Zool¬
ogy, 169:455-545.

1973b. Systematics and Evolution of the Gruiformes (Class
Aves), 3: Phylogeny of the Suborder Grues. Bulle¬
tin of the American Museum of Natural History,
151:1-127.

1974. Phylogeny and Evolution of the Ratite Birds. Ibis,
116:494-521.

Dixon, W. J.

1970. BMD Biomedical Computer Programs. Los An¬
geles: University of California Press.

Falla, R. A.

1941. The Avian Remains. Pages 339-353 in R. S. Allan,
E. Percival, R. S. Duff, and R. H. Falla, Prelimi-

NUMBER 27

205

nary Report of Excavations at Pyramid Valley
Swamp, Waikari, North Canterbury. Records of
the Canterbury Museum, 6.

Fleming, C. A.

1962. The Extinction of Moas and Other Animals Dur¬
ing the Holocene Period. Notornis, 10:113-117.

Goodge, W.

1951. Variation in Skeletal Measurements of the Common
Murre. Condor, 53:99-100.

Haast, J. von

1885. Preliminary Notice of Paper on Dinornis oweni
Haast. Proceedings of the Zoological Society of
London, 1885:482.

1886. On Dinornis oweni, a New Species of the Dinor-
nithidae, with Some Remarks on D. curtus. Trans¬
actions of the Zoological Society of London, 12:
171-182.

Hemming, F., editor

1954. Opinion 229: Acceptance of the Lectotype Selected
by Lydekker (1891) for the Nominal Species Dinor¬
nis novaezealandiae Owen, 1843 (Class Aves.) Inter¬
national Commission of Zoological Nomenclature,
Opinions and Declarations, 4(19):221:230.

Lambrecht, K.

1933. Handbuch der Palaeornithologie. xix + 1024 pages.
Berlin: Gebriider Borntrager.

Long, J. L.

1965. Weights, Measurements and Food of the Emu in
the Northern Wheatbelt of Western Australia.
Emu, 64:214-219.

Lydekker, R.

1891. Catalogue of the Fossil Birds in the British Mu¬
seum (Natural History), xxvii + 368 pages. Lon¬
don: British Museum (Natural History).

Oliver, W. R. B.

1930. New Zealand Birds, viii 4- 541 pages. Wellington:
Fine Arts (N.Z.) Limited.

1949. The Moas of New Zealand and Australia. Do¬
minion Museum Bulletin, 15:1-206.

Oxnard, C. E.

1973. Form and Pattern in Human Evolution, ix + 218
pages. Chicago: University of Chicago Press.

Scarlett, R. J.

1972. Bones for the New Zealand Archaeologist. Canter¬
bury Museum Bulletin, 4:1-69.

Simpson, G. G.

1946. Fossil Penguins. Bulletin of the American Museum
of Natural History, 87:1-99.

Yaldwyn, J. C.

1956. A Preliminary Account of the Sub-fossil Avifauna
of the Martinborough Caves. Records of the Do¬
minion Museum, 3(1): 1-7.

Evidence of the Survival to Recent Times
of the Extinct Flightless Duck Chendytes lawi Miller

G. Victor Morejohn

ABSTRACT

Skeletal remains of the extinct late Pleistocene
flightless diving duck, Chendytes lawi, were dis¬
covered at two northern California Indian midden
sites. Carbon 14 dates of midden shell and aspartic
acid racemization of the Chendytes bones showed
that this species lived into the Holocene and be¬
came extinct sometime after 3780 years bp. The
remains from these northern California middens
and a newly discovered Pleistocene tibiotarsus from
the Port Orford Formation of Oregon extend the
known range of Chendytes laid from the Channel
Islands of southern California northward some 720
kilometers.

Introduction

Study of avian skeletal remains from an Indian
midden on the south shore of Laguna Creek, Santa
Cruz County, California, has revealed many ele¬
ments which unquestionably belong to the extinct,
flightless diving duck, Chendytes lawi Miller. Two
bones of this species were also found in another
midden on Ano Nuevo Point, about 29 km north
of the city of Santa Cruz. Previously, all certain
records of Chendytes were from Pleistocene depos¬
its. The present account documents its persistence
into the Holocene and its contemporaneity with
aboriginal man.

Acknowledgments.— I wish to thank Kenneth
R. Lajoie and Edward Helly of the U.S. Geological
Survey for providing Carbon 14 dates and assistance
in field work. Jeffery L. Bada of Scripps Institution

G. Victor Morejohn, Department of Biological Sciences, San
Jose State University, San Jose, California 95192.

of Oceanography kindly determined the age of
bone fragments through aspartic acid racemization.
Burton L. Gordon of San Francisco State Univer¬
sity provided many bone fragments found at the
Laguna Creek midden, some of which belonged to
Chendytes. Hildegarde Howard assisted in the
original identifications of bones and supplied use¬
ful comments on the study. Eric Anderson, Janice
Cross, Howard Hutchison, Diana Matthiesen, and
Bruce Welton were helpful in the field or assisted
me in the museum collections.

Description and Age of the Sites

The Laguna Creek midden (State of California
Archeological site designation—Santa Cruz 7, here¬
after abbreviated as CA-SCr-7) is situated atop a
large sand dune (Sandhill Bluff) on a marine ter¬
race some 25 meters above sea level at latitude
36°58'30" N and longitude 122°09'10" W, approxi¬
mately 12 km north of Santa Cruz, Santa Cruz
County, California, and about 230 meters west of
U.S. Highway No. 1. The dune occupies an area of
about 915 m 2 , with the midden capping the highest
part and occupying an area of about 120 m 2 . The
windswept seaward side of the midden is greatly
eroded, in places to a depth such that the under¬
lying dune substrate is visable. Most of the bones
were collected from this exposed face. Bones of
Chendytes lawi were found from 15 cm below the
surface to the basal layer of the midden, and several
others were found scattered on the surface of the
dune below the exposed face.

Most of the midden consists of broken mollusk
and barnacle shells tightly packed to a depth vary¬
ing between 1.5 and 1.75 m. Early reports (Stearns,
1873-1874:157) indicate that this midden was

207

208

originally some 6.5 m high, but recent local use
for shell to feed poultry has greatly reduced the
size of the midden and removed the younger upper
layers. Samples of mollusk and barnacle shells from
the upper and lower midden levels were used for
Carbon 14 dating. The uppermost layer was aged
at 3780 ± 95 years bp and the basal layer was aged
at 5390 ± 100 years bp. Two bones of Chendytes
from the basal layer were determined by aspartic
acid racemization (Bada, et al., 1974) to be approxi¬
mately 6000 years old, which corresponds closely
with the Carbon 14 dates of other material from
the same layer. Fragments of a human femur and
an incomplete skull lacking the rostrum and lower
jaw were also found in the basal layer, but the
aspartic acid age of these fragments was approxi¬
mately 4000 years bp, obviously indicating secon¬
dary introduction through burial. The above
datings establish CA-SCr-7 as the oldest known site
of Indian occupation on the central California
coast (M. J. Moratto, in litt., 19 December 1974).
The Indian midden on Ano Nuevo Point, San
Mateo County, California, has not been studied.
Archeologists have estimated its age at about 2500
years bp, but it is probably older. Persistent sea¬
sonal winds expose part of the midden or cover it
with sand. Only a few vertebrate fragments have
so far been collected.

The following bones of Chendytes were recov¬
ered from the Laguna Creek and Ano Nuevo mid¬
dens: humeri, 1 left, 2 right proximal ends; femora,
8 left, 7 right nearly complete, 4 left, 5 right in¬
complete; tibiotarsi, 1 left, 5 right proximal ends,
3 left, 3 right distal ends, 10 partial shafts; tarso-
metatarsus, 1 left incomplete; synsacrum, 1 incom¬
plete; cervical vertebra, 1 incomplete.

Discussion

Because comparisons were initially made with
reference specimens of living taxa, the elements of
Chendytes collected at CA-SCr-7 at first proved im¬
possible to identify. Later, while studying anseri-
form fossils with Dr. Hildegarde Howard at the
Natural History Museum of Los Angeles County,
I examined specimens of Chendytes that I then
recognized as being similar to the unidentified
bones from CA-SCr-7. In size (Table 1) and
morphology, the midden bones were found to con¬
form with those previously described for Chendytes

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

Figure 1.—Comparisons of femora and tibiotarsi of Chendytes
lawi (on the left in each pair) and Melanitta fusca: a, cross-
section of femora at midshaft showing cancellous nature of
bone and heavy compact peripheral layer present in the
flightless C. lawi; b, anterior view of proximal end of tibio-
tarsus contrasting the development of the outer cnemial
crest; c, posterior views of proximal ends of tibiotarsi with
cross-sections proximal to and through the fibular crest.

laiui by Miller (1925; 1930), Miller, et al. (1961),
and Howard (1947, 1949, 1955). I later noted sev¬
eral characteristics of the tibiotarsi and femora that
had not previously been considered by these
authors.

Miller (1925) first recognized the affinities of
Chendytes with the diving ducks, particularly the
scoters. Compared to Melanitta fusca, the largest
of the scoters, I found great reduction in the mar¬
row cavities of the tibiotarsi and femora, with a
concomitant increase in the thickness of the peri¬
pheral compact bone in C. lawi (Figure la). In
fact, certain cross-sections of the femora of Chen¬
dytes bear a similarity to those of mammalian

NUMBER 27

209

femora. The density of the bone in Chendytes is
probably an adaptation for diving. Cross-sections
of the tibiotarsus proximal to the fibular crest show
M. fusca to have a pronounced posterior ridge,
whereas in Chendytes this ridge is low and rounded
(Figure lc). Cross sections through the fibular
crest show M. fusca and Chendytes to have a flat¬
tened anterior and a ridged posterior surface, but
in Chendytes the inner cnemial crest extends dis-
tally to about the middle of the fibular crest and
angles anteriorly (Figure 16). The outer cnemial
crest in anterior view is also markedly different in
the two species (Figure Id).

Howard (1955) reported a total of 156 fossil
skeletal remains of Chendytes (C. lawi, 89; C.
milleri, 67) from eleven localities in southern Cali¬
fornia. Since then, another 130 specimens have
been recovered from Anacapa Island (Howard,
1964). While most of the bones were clearly from
Pleistocene localities, fifteen were reportedly col¬
lected from two Indian middens (one at Malaga
Cove in the Palos Verdes area of Los Angeles
County, the other on San Nicolas Island). How¬
ever, these bones are well mineralized, unlike those

Table 1. —Comparison of measurements (mm) of bones of
Chendytes lawi from northern and southern California
localities

Character

n

Northern

California

mean range

Southern

California*

range

Humerus

Breadth of proximal
end from external
to internal
tuberosity .

3

13.3

13-13.5

14.3

Breadth of shaft
below external
tuberosity .

3

4.3

4-4.5

4.8

Femur

Length .

5

73.2

69.6-77.5

65.7-76.4

Breadth of proximal
end .

9

16.1

15.2-17.8

14-17

Breadth of distal end ..

6

19

17.7-20.4

16.3-18.2

Tibiotarsus

Breadth of proximal

end .

3

15

14.4-15.7

14.8-15.5

Breadth of distal end .

3

13.7

13.3-14.5

13.3-14.5

* From Howard, 1955.

typically found in middens, and they were con¬
sidered by Howard (1955) to have been secondarily
associated with the Indian midden material. In
contrast, bones of C. lawi from the Laguna Creek
and Ano Nuevo Point Indian middens in northern
California were not mineralized.

Although bones of a variety of birds have been
recovered from other Indian middens along the
California coast (Howard, 1929; Howard and
Dodson, 1933; and D. M. Howard and Cook, 1971)
and demonstrate that a number of avian species
were eaten by Indians (Table 2), no other ele¬
ments of Chendytes have thus far been found.
Howard (1929), for lack of comparative anseri-
form skeletal material, chose not to attempt identi¬
fication of waterfowl in her study of the avian
remains from the Emeryville shellmound. I re¬
cently reexamined the anseriform material from
this site in an attempt to find elements of Chen¬
dytes, but met with no success.

The Emeryville site was estimated to be no
younger than 2310 ± years (Hubbs, et al., 1962).
Chendytes may have been extinct by this time, or
its absence from the Emeryville site may have been
due to the lack of appropriate habitat. The site is
located on an alluvial plane (Howard, 1929) with-

Table 2-—Relative abundance of avian species found associ¬
ated with remains of Chendytes lawi at the Laguna Creek
Indian midden

Species

Number of specimens

Gavia stellata

1

Puffinus griseus

6

Fulmarus glacialis

2

Unidentified albatross

23

Diomedea albatrus

10

Unidentified cormorant

2

Phalacrocorax penicillatus

3

Phalacrocorax pelagicus

1

Unidentified anseriform

12

Unidentified duck

6

Melanitta fusca

2

Chendytes lawi

46

Unidentified falconiform

1

Buteo jamaicensis

1

Unidentified galliform

I

Unidentified charadriiform

1

Uria aalge

29

Cepphus columba

6

Ptychoramphus aleutica

1

Unidentified gull

26

Unidentified bird

99

210

SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY

out nearby reefs or rocky shorelines such as were
probably required by Chendytes (Miller, et al.,
1961). The elements of Chendytes lawi from CA-
SCr-7, which may be at least as young as 3780 years,
provide the latest known occurrence of the genus.
At some time after deposition of these bones, per¬
haps 2500 to 3000 years ago, Chendytes became
extinct.

The known fossil localities of C. lawi are con¬
centrated in southern California and the range of
the species was formerly considered by Miller, et al.
(1961:10) to be “south of Point Conception, from
Ventura County to Orange County.’' Although the
material from CA-SCr-7 extends this range some
320 km farther northward, a hitherto unreported,
nearly complete, fossilized tibiotarsus of C. lawi
(University of California Museum of Paleontology
No. 112026) collected by David Taylor from the
lower Pleistocene Port Orford Formation in Curry
County, Oregon, near Cape Blanco (UCMP verte¬
brate locality V-74042) extends the range of this
species some 720 km north of the southern Cali¬
fornia sites (Figure 2).

Evolution of flightlessness in smaller birds is
usually associated with insular distribution (Olson,
1973:31-36). Being flightless, the breeding sites of
Chendytes almost certainly had to be restricted to
offshore islands, and it may be assumed that the
Channel Islands were of particular importance in
this regard. There is no geological evidence of
former islands in the immediate vicinity of the
Laguna Creek area. Therefore, C. lawi was prob¬
ably present in this area as a nonbreeding migrant.

The frequency of occurrence of bones of Chen¬
dytes lawi at CA-SCr-7 may indicate selective pre¬
dation by Indians there. The bird was flightless,
was no doubt relatively easy to capture, and had
large goose-sized legs which would have made it a
preferred food item. The preponderance of leg
bones at CA-SCr-7 supports this contention. Prob-

Figure 2. —The known distribution of Chendytes along the
Pacific coast of California and southern Oregon.

ably for the preceding reasons, the remains of C.
lawi were more abundant in this midden than any
other species of bird (Table 2). The birds probably
were caught in the surf zone, rather than on land,
inasmuch as the California Indians were known to
have captured waterfowl with nets (Heizer, 1974).
The extinction of C. lawi prior to the deposition of
the other coastal middens thus far studied in Cali¬
fornia may account for its absence in them. Over¬
harvest by aboriginal man was probably the prin¬
cipal factor contributing to the extinction of this
species.

Literature Cited

Bada, J. L., R. A. Schroeder, and G. F. Carter
1974. New Evidence for the Antiquity of Man in North
America Deduced from Aspartic Acid Racemization.
Science, 184(4138):791-793, 2 tables.

Heizer, R. F., editor

1974. The Costonoan Indians. Local History Studies, 18:
1-115, 23 figures. Cupertino, California: California
History Center, De Anza College.

Howard, D. M„ and S. F. Cook
1971. The Archeology of the Hudson Mound. Monterey
County Archeological Society Quarterly, 1:1-11, 7
figures.

Howard, H.

1929. The Avifauna of Emeryville Shellmound. Univer¬
sity of California Publications in Zoology, 32(2):
301-394, 4 plates, 55 figures.

211

NUMBER 27

1947. Wing Elements Assigned to Chendytes. Condor,
49(3):76-77, 1 figure, 1 table.

1949. Avian Fossils from the Marine Pleistocene of South¬
ern California. Condor, 51(1):20—28.

1955. New Records and a New Species of Chendytes, an
Extinct Genus of Diving Geese. Condor, 57(5): 135—
143, 3 figures, 2 tables.

1964. Further Discoveries Concerning the Flightless “Div¬
ing Geese” of the Genus Chendytes. Condor, 66(9):
372-376.

Howard, H., and L. M. Dodson
1933. Birds Remains from an Indian Shellmound near
Point Mugu, California. Condor, 35(11):235.

Hubbs, C. L., G. S. Bren, and H. E. Suess
1962. La Jolla Natural Radiocarbon Measurements II.
Radiocarbon, 4:208.

Miller, L.

1925. Chendytes, a Diving Goose from the California
Pleistocene. Condor, 27(7): 145-149, 1 figure.

1930. Further Bird Remains from the Upper San Pedro
Pleistocene. Condor, 32(3): 116-118, 1 figure.

Miller, L., E. D. Mitchell, and J. H. Lipps

1961. New Light on the Flightless Goose Chendytes lawi.
Los Angeles County Museum Contributions in
Science, 43:3-11, 2 plates.

Olson, S. L.

1973. Evolution of the Rails of the South Atlantic Is¬
lands (Aves: Rallidae). Smithsonian Contributions
to Zoology, 152:1-53, 8 figures, 11 plates, 4 tables,
2 appendices.

Stearns, R. E. C.

1873-1874. Remarks on the Upper Tuolumne Canon.
Proceedings of the California Academy of Sciences,
first series, 5:155-158.

■fr U.S. GOVERNMENT PRINTING OFFICE: 1976-211—599/7

Publication in Smithsonian Contributions to Paleobiology

Manuscripts for serial publications are accepted by the Smithsonian Institution Press, sub¬
ject to substantive review, only through departments of the various Smithsonian museums.
Non-Smithsonian authors should address inquiries to the appropriate department. If submission
is invited, the following format requirements of the Press will govern the preparation of copy.

Copy must be typewritten, double-spaced, on one side of standard white bond paper, with
V/i" top and left margin, submitted in ribbon copy with a carbon or duplicate, and accom¬
panied by the original artwork. Duplicate copies of all material, including illustrations, should
be retained by the author. There may be several paragraphs to a page, but each page should
begin with a new paragraph. Number consecutively all pages, including title page, abstract,
text, literature cited, legends, and tables. The minimum length is 30 pages, including typescript
and illustrations.

The title should be complete and clear for easy indexing by abstracting services. Taxonomic
titles will carry a final line indicating the higher categories to which the taxon is referable:
“(Ammonoidea: Goniatitidae).” Include an abstract as an introductory part of the text.
Identify the author on the first page of text with an unnumbered footnote that includes his
professional mailing address. A table of contents is optional. An index, if required, may be
supplied by the author when he returns page proof.

Two headings are used: (1) text heads (boldface in print) for major sections and
chapters and (2) paragraph sideheads (caps and small caps in print) for subdivisions. Further
headings may be worked out with the editor.

In taxonomic keys, number only the first item of each couplet; if there is only one couplet,
omit the number. For easy reference, number also the taxa and their corresponding headings
throughout the text; do not incorporate page references in the key.

In synonymy , use the short form (taxon, author, date:page) with a full reference at the
end of the paper under “Literature Cited.” Begin each taxon at the left margin with sub¬
sequent lines indented about three spaces. Within an entry, use a period-dash (.—) to separate
each reference. Enclose with square brackets any annotation in, or at the end of, the entry.
For references within the text, use the author-date system: “(Jones, 1910)” and “Jones
(1910).” If the reference is expanded, abbreviate the data: “Jones (1910:122, pi. 20: fig. 1).”

Simple tabulations in the text (e.g., columns of data) may carry headings or not, but they
should not contain rules. Formal tables must be submitted as pages separate from the text, and
each table, no matter how large, should be pasted up as a single sheet of copy.

Use the metric system instead of, or in addition to, the English system.

Illustrations (line drawings, maps, photographs, shaded drawings) can be intermixed
throughout the printed text. They will be termed Figures and should be numbered con¬
secutively; however, if a group of figures is treated as a single figure, the components should
be indicated by lowercase italic letters on the illustration, in the legend, and in text references:
"Figure 9b.” If illustrations (usually tone photographs) are printed separately from the text as
full pages on a different stock of paper, they will be termed Plates, and individual components
should be lettered (Plate 9b) but may be numbered (Plate 9: figure 2). Never combine the
numbering system of text illustrations with that of plate illustrations. Submit all legends on
pages separate from the text and not attached to the artwork. An instruction booklet for the
preparation of illustrations is available from the Press on request.

In the bibliography (usually called “Literature Cited”), spell out book, journal, and
article titles, using initial caps with all words except minor terms such as “and, of, the.” For
capitalization of titles in foreign languages, follow the national practice of each language.
Underscore (for italics) book and journal titles. Use the colon-parentheses system for volume,
number, and page citations: “10(2) :5-9.” Spell out such words as “figures,” “plates,” “pages.”

For free copies of his own paper, a Smithsonian author should indicate his requirements
on “Form 36” (submitted to the Press with the manuscript). A non-Smithsonian author will
receive 50 free copies; order forms for quantities above this amount with instructions for
payment will be supplied when page proof is forwarded.