The Fossil Waterfowl (Aves: Anseriformes) from the Eocene of England

16 Pages·2001·0.63 MB·English

by  Gareth J. Dyke

PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY
CENTRAL PARK WEST AT 79TH STREET, NEW YORK, NY 10024
Number 3354, 15 pp., 7 figures December 31, 2001
The Fossil Waterfowl (Aves: Anseriformes) from
the Eocene of England
GARETH J. DYKE1
ABSTRACT
The taxonomy and phylogenetic positions of the fossil waterfowl (Aves: Anseriformes)
described from the Eocene of England are reviewed. Although a total of six species within
five genera have been described, only the phylogenetic positions of the two taxa Anatalavis
oxfordi Olson and Headonornis (Lydekker) can be hypothesized with confidence within the
order Anseriformes. Anatalavis oxfordi is considered to be the sistertaxon to a cladecompris-
ing another fossil waterfowl, Presbyornis Wetmore, and the extant true ducks, Anatidae,
whereas Headonornis may be congeneric with Presbyornis. Fossil material referred to the
genera Palaeopapia Harrison and Walker and Paracygnopterus Harrison and Walker is con-
sidered Aves incertae sedis.
INTRODUCTION ard, 1955; Olson and Parris, 1987; Olson,
1999).
The avian orderAnseriformesincludesthe
From rocks of Eocene age (ca. 55 million
extant waterfowl, the screamers, ducks, and
years ago), the fossil remains of anseriform
geese. Considered by most systematiststobe
birds have been recovered from a number of
one of the most basal clades within modern
birds (Neornithes sensu Cracraft, 1988; see deposits in both Europe and North America.
also reviews in Sibley and Ahlquist, 1990; From the Lower Eocene Green River For-
Ericson, 1997; Livezey, 1997; Groth and mationofWyomingandUtahcomesperhaps
Barrowclough, 1999), the fossil record of the most famous of the fossil anseriforms,
this group has been considered to extend thetaxonPresbyornispervetusWetmore.For
back into the Mesozoic period, prior to the a number of years, thistaxonwasconsidered
Cretaceous-Tertiary (KT) boundary (How- to be ‘‘intermediate’’ in morphology,provid-
1DivisionofVertebrateZoology (Ornithology),AmericanMuseumofNaturalHistory.
Copyright(cid:113)AmericanMuseumofNaturalHistory2001 ISSN0003-0082
2 AMERICAN MUSEUM NOVITATES NO. 3354
ing evidence for an evolutionary link be- MATERIALS AND METHODS
tween a number of traditional avian orders
Thematerialdiscussedinthispaperisheld
by the time of the early Eocene (Olson and
in the Palaeontology DepartmentCollections
Feduccia, 1980; Feduccia, 1995, 1996).
of The Natural History Museum, London
However, the recentinclusionofPresbyornis
(formerly the British Museum of Natural
withincladisticanalysesofRecentwaterfowl
History; BMNH PAL), the Sedgewick Mu-
has demonstrated that this taxon can be
seum, Cambridge (SMC), and in the United
placed well within the order Anseriformes, States National Museum, Washington D.C.
closely related to the extant true ducks (An- (USNM). Recent comparative osteological
atidae; Ericson, 1997; Livezey, 1997). specimens were examined in the collections
FromtheLowerEoceneLondonClayFor- of The Natural History Museum, Tring. The
mation and contemporary deposits in Eng- anatomicalnomenclatureemployedisasout-
land, the remains of fossil birds have been lined by Baumel and Witmer (1993), with
known for over 100 years (Koenig, 1825; some modifications to English following the
Owen, 1841; Andrews, 1899). However, standardworkofHoward(1929).References
much of this material remainedunstudiedaf- to Recent avian taxa and higher-level groups
ter this time until the various works of Har- follow Sibley and Monroe (1990). Rules
rison and Walker in the 1970s and 80s (e.g., governing the synonymy and designation of
Harrison, 1982a, 1982b, 1983; Harrison and taxonomic names are used as outlined by the
Walker, 1976, 1979a, 1979b, 1979c). Recent International Commission on ZoologicalNo-
considerations of the fossil bird assemblage menclature (ICZN, 1999).
from the London Clay Formation include
Mayr and Daniels (1998), Dyke and Cooper DISCUSSION AND SYSTEMATICS OF
FOSSIL TAXA
(2000), and Dyke and Waterhouse (2001).
Notably, Harrison and Walker (1976, 1979a) AVES
namedfourspeciesofwaterfowlwithinthree
ANSERIFORMES
genera from the Tertiary of England (Hea-
donornis hantoniensis, Paracygnopterus Anatalavis Shufeldt, 1915
scotti, Palaeopapia eous and Palaeopapia
ANATALAVISOXFORDI Olson, 1999
hampsteadiensis) and remarked on the pos-
sible referral of the taxon Proherodiusoweni The holotype specimen (BMNH PAL
Lydekker to the order. Recently, Olson 5922;fig.1A–C)ofAnatalavisoxfordiOlson
(1999) described a new species within the is one of the best preserved fossil birds
genus Anatalavis Shufeldt (Anatalavisoxfor- known from the London Clay Formation.
Unfortunately, however, it is also one of the
di) on the basis of material collected from a
most fragile fossil bird specimens known.
basal member of the London Clay Forma-
Although both the partially complete skull,
tion.
sternum, and pelvis (along with a number of
In this paper, the fossil records of water-
forelimb elements) have been prepared,
fowl from the Eocene of England are re-
studying this material is complicated by the
viewed in light of recent phylogenetic hy-
delicate preservation of the material. This
pothesesregardingtherelationshipsofextant
specimen was collected from bed A of the
anseriforms. In cases where material is com-
London Clay at Walton-on-the-Naze, Essex,
plete enough (i.e., preserves characters that
England(King,1981)in1991(Olson,1999).
have been employed in phylogenetic consid-
FOSSIL MATERIAL: BMNH PAL 5922 (fig.
erations of Recent taxa), fossil birds can be 1A–C), an incomplete skeleton consisting of
used in the production of minimum age es- skull (fig. 1A), incomplete rostral portion of
timates for the divergences of internal nodes mandible, pterygoid, atlas, axis, thoracicand
within cladograms, and are therefore invalu- caudal vertebrae, complete furcula (fig. 1B),
able for unravelling the pattern of the evo- left and right coracoids (fig. 1C), left and
lutionary radiation of the modern birds as a right scapulae, incomplete sternum (lacking
group (Dyke, 1998, 2001). posterior portions), incomplete anterior por-
2001 DYKE: FOSSIL WATERFOWL 3
Fig. 1. Parts of the holotype specimen of Anatalavis oxfordi Olson (BMNH PAL5922). A, skullin
lateral views; B, furcula; C, coracoid in dorsal and medial views. Scale bars are 10 mm.HYPindicates
hypocleidium (apophysis furculae); FP, foramen pneumaticum.
tion of pelvis, complete right and incomplete thatthefossilrecordofthisextantfamilycan
left humeri, distal ends of left radius and be traced back to the Lower Eocene of the
ulna, left ulnare, complete leftcarpometacar- northern hemisphere, and (2) the origin of
pus, complete phalanges, and minor digitsof this group of waterfowl can be inferred to
leftforearm.Forcompletedescriptionandil- have occurred inthenorthernhemisphere(of
lustration of this material, see Olson (1999). course, the distribution of the family is cur-
TAXONOMIC HISTORY: Olson (1999) con- rently restricted to the southern continents).
sidered the holotype specimen of Anatalavis Olson (1999) further proposed that the early
oxfordi and referred it to within the extant fossilrecordoftheorderAnseriformesinthe
family Anseranatidae, as the sole member of northern hemisphere may have been domi-
a new subfamily, the Anatalavinae (fig. 2). nated by anseranatid-like taxa during this
The only other member of this family of an- time, to be replaced later by the members of
seriforms is the extant Australian Magpie the Anatidae, or true ducks.
Goose,Anseranas semipalmata.Onthebasis Referral of the holotype specimen of An-
of this referral, a number of significant con- atalavis oxfordi to the Anseranatidae was
clusions were proposed by Olson (1999): (1)
made by Olson (1999: 232) on the basis of
the following osteological characters: the V-
shaped conformation of the furcula, the size
of the symphysis (fig. 1B), and the presence
of a distinct foramen pneumaticum on the
dorsalsurfaceofthedistalcoracoid(fig.1C).
All of these features were considered by Ol-
son (1999) to be uniquely derived within the
Anseranatidae. In particular, with regard to
the presence of a foramen pneumaticum on
the dorsal coracoid, he noted that (p. 232)
‘‘although a similar condition exists in mod-
ern Anhimidae [screamers], which have one
of the most pneumatized skeletons in any
Fig. 2. Phylogenetic position of Anatalavis group of birds, this foramen is absent in Eo-
oxfordi proposed by Olson (1999). cene Anhimidae, which are evidently com-
4 AMERICAN MUSEUM NOVITATES NO. 3354
pletely nonpneumatic’’. Although it is true
that the this character is present in the extant
screamers (Livezey, 1997), the extent of var-
iation of this feature in the known fossil
formsisunclearsincelittlematerialhasbeen
described to date. Indeed, examination of
material described and illustrated by Alvar-
enga (1999) as a fossil screamer from the
Oligocene of southeastern Brazil shows that
this foramen is clearly present (albeitbroken
distally: Alvarenga, 1999: figs. 3, 4). Al-
thoughOlson(1999)citedasyetunpublished
material of fossils referred informally by
some to screamers or similar birds from the
Eocene of England and Wyoming (p. 232),
clearly the distribution (and primitive ab-
sence) of this character requires further in-
vestigation.Intheabsenceofacladisticanal-
ysis including all available osteological evi-
dence, as well as the characters cited by Ol-
son (1999), the referral of this specimen to
the extant family Anserantidae cannot be
considered unequivocal. Fig. 3. Single tree (138 steps) resulting from
PHYLOGENETIC ANALYSIS: The phylogenet- reanalysis of Livezey (1997) and including the
ic relationships of the exant genera of water- London Clay fossil Anatalavis(consistencyindex
CI (cid:53) 0.761; retention index RI (cid:53) 0.815).
fowl have been considered most recently by
Ericson (1997) and Livezey (1997, 1998).
Both of these workers included the fossil results of the analysis). Consideration of
Presbyornis pervetus Wetmore (see above) these features in the holotype skull will rep-
within osteological character analyses and resent an area for further work.
agreed that it should be placed as the sister Parsimony analysis of the osteologicalda-
taxontotheRecentAnatidae(trueducks:Er- taset presented by Livezey (1997) by use of
icson, 1997; Livezey, 1997). the software package PAUP (version 4.0b;
By use of the osteological matrix present- Swofford, 1999: all characters unordered us-
edbyLivezey(1998),theholotypespecimen ing Tinamiformes as the outgroup), and in-
of Anatalavis oxfordi (BMNH PAL 5922) cluding the holotype of Anatalavis oxfordi,
was coded for inclusion in aparsimonyanal- resulted in the production of a single most
ysis. Of the 96 osteological characters pre- parsimonious tree (MPT) of length 138 steps
sented by Livezey (1998), only 22 (23 %) (fig. 3). The inclusion of Anatalavis within
are informativefor Anatalavisbecauseofthe theanalysishasnoeffectontherelationships
incomplete and fragile preservation of the of the remaining extant taxaandPresbyornis
holotype. These character codings are given (see Livezey, 1997): the London Clay fossil
in appendix 1. Some small changes were is hypothesized to occur as the sister taxon
made to the character-state definitions pro- to the extant Anatidae and Presbyornis (fig.
videdbyLivezey(1997)fortheanalysispre- 3), in a more derived position with respect
sented here: characters 3 (presence / absence to the Anseranatidae (Anseranas). The posi-
of ventral prominence of processus paraoc- tion for this taxon hypothesized by Olson
cipitalis relative to plane of os paraspenoi- (1999) is not supported by parsimony anal-
dale) and 20 (presence/absence of processus ysis.
retroarticularis) were amended as outlined to REVISED TAXONOMY OF ANATALAVIS OXFOR-
simple binary states because of uncertainties DI: Although consideration of the phyloge-
in the coding of these characters in the skull netic position of Anatalavis oxfordi Olson is
ofAnatalavis(thesechangesdonotaffectthe complicated by incomplete preservation of
2001 DYKE: FOSSIL WATERFOWL 5
theholotype,thepositionofthistaxonwithin and the Anatidae. Thesecond ofthetwofea-
the single MPT (fig. 3) is unequivocal. On tures discussed by Olson (1999), the pres-
the basis of the osteological characters (and ence of a foramen pneumaticum on the dor-
analysis) presented by Livezey (1997), sal surface of the coracoid (character 70 of
monophyly of the waterfowl (order Anseri- Livezey, 1997), is hypothesized to diagnose
formes) is supported by 11 unambiguous the clade Anseriformes (cid:49) Galliformes andis
characters (see Livezey, 1997, for details), lost in Presbyornis and the Anatidae. Hence,
oneofwhich,thepresenceofalongandven- bothoftheosteologicalfeaturespresentedby
trallyterminatinglaminabasiparasphenoidal- Olson(1999)insupportofthereferralofAn-
is (relative to the plane of the parasphenoi- atalavis to the Anseranatidae are primitive
dale; character 3, Livezey, 1997), is pre- among Anseriformes.
served in Anatalavis (appendix 1). Olson
(1999: 232) noted that the skull of the ho- Headonornis Harrison and Walker, 1976
lotype has all of the features typical of the
Headonornis hantoniensis (Lydekker,
order Anseriformes, to which it clearly be-
1891); Harrison and Walker, 1976
longs. These include (according to Olson,
1999) a ‘‘duck-billed’’ upper jaw shape; the Agnopterus? hantoniensis Lydekker (1891)
configuration of the quadrato-mandibular ar-
Agnopterus? hantoniensis Lydekker; listed
ticulation; an enlarged,deep,curved,retroar-
by Lambrecht (1933)
ticular process; and an enlarged rounded or
ovoid ‘‘basipterygoid process’’ on the par- Agnopterus? hantoniensis Lydekker; listed
asphenoid rostrum, with a correspondingly by Brodkorb (1964)
enlarged facet on the pterygoid.
Headonornis hantoniensis Harrison and
The position of Anatalavis within the
Walker (1976)
cladeAnseres(truewaterfowl;seefig.3)can
be inferred by the presence of a prominent The genus Agnopterus was erected by
crista fossa parabasalis of the exoccipitale Milne-Edwards (1869–1871) for the recep-
(derived state for character 2 of Livezey, tion of thedistal portionofarighttibiotarsus
1997) and a prominent processus corono- from the Upper Eocene of Montmartre,
ideus of the mandible (derived stateforchar- France. Milne-Edwards (1869–1871) named
acter 18 of Livezey, 1997). A sister group the species Agnopterus laurillardi on the ba-
relationship between Anatalavis, Presbyor- sis of this material, and he noted a resem-
nis, and the clade Anatidae (true ducks) can blance with the extant Phoenicopterus (fla-
be hypothesized because of the presence of mingo). Later, Lydekker (1891) named the
aspinaexterna(asacompressedphlange)on species Agnopterus? hantoniensis on the ba-
the rostral end of the sternum (derived state sis of a partially complete right coracoid
for character 60 of Livezey, 1997), and an (BMNH PAL 30325; fig. 4A), and referreda
elongate acromion on the cranial end of the cast of a proximal left femur (BMNH PAL
scapula (reversal within Anseriformes for 144) to this taxon. Harrison and Walker
character 68 of Livezey, 1997). (1976) removed this material from the genus
Both of the features discussed by Olson Agnopterus, making the coracoid the holo-
(1999) in support of the referralofAnatalav- type of the new genus and species Headon-
is to the Anseranatidae were included within ornis hantoniensis within the Anseriformes.
the osteological character matrix of Livezey They also referred an incomplete right hu-
(1997) and have been demonstrated to have merus (BMNH A 3686; fig. 4B). The speci-
a wider distribution within Anseriformes. men (BMNH PAL 144) is a cast of an orig-
The presence of a hypocleideum on the fur- inal held in the Institute of Geological Sci-
cula (character 67 of Livezey, 1997), al- ences, London (IGS GSM 113109; formerly
though seen in Anatalavis and Anseranas, is the Geological Survey Museum, London;
hypothesized to have been present primitive- Harrison and Walker, 1976).
ly within the group: this character, tested by MATERIAL: BMNH PAL 30325, a partially
the congruence of others, has been second- complete right coracoid lacking extremities
arily lost in Anhima, Chauna, Presbyornis, of processus lateralis and angulus medialis
6 AMERICAN MUSEUM NOVITATES NO. 3354
Fig. 4. Some of the fossil material referred to Headonornis hantoniensis Harrison and Walker. A,
holotype right coracoid in dorsal view (BMNH PAL 30325); B, referred distal end of left humerus in
cranial view (BMNH PAL 5105); C, referred right humerus in caudal view (BMNH PAL 3686). Scale
bars are 5 mm.
(holotype; Harrison and Walker, 1976; fig. end(referredspecimen:HarrisonandWalker,
4A), purchased from the Hastings Collection 1976; fig. 4B), collected from Upper Eocene
in 1855 but collected sometime prior to this deposits near Milford, Hampshire, and pre-
from the Upper Eocene of Hordle (Hord- sented by J. Athersuch in 1970. BMNHPAL
well), Hampshire. BMNH PAL 3686, an in- 4989,theproximalendofarightscapula(re-
complete right humerus lacking the distal ferred specimen: Harrison and Walker,
2001 DYKE: FOSSIL WATERFOWL 7
1979a), collected from the Lower Oligocene noted that, in the morphology of the distal
Hampstead Beds on the Isle-of-Wight by S. humerus, this species does not differ fromP.
L. Wood in 1925 (presented in 1933). SMC pervetus other than in size and the distinc-
C20412, the sternal end of a coracoid (re- tivenessofthefossaolecrani.Thelatterchar-
ferred specimen: Harrison and Walker, acter may also be related to the large size of
1979a; collection data unknown). As noted the specimen. In particular, on the distal end
by Harrison and Walker (1979a), this speci- of the humerus of both Presbyornis pervetus
men was described by Seeley (1866) under and P. isoni, the fossa m. brachialis is deep
thegenericnamePtenornis(nospeciesname and marked, and the condylus dorsalis is
given), that was later regarded by Lydekker prominent and strongly developed proximal-
(1891) as a nomen nudum. BMNH PAL ly (Olson, 1994; Ericson, 1999a).
5105, the distal end of a left humerus (re- Harrison and Walker (1979a) noted that
ferred specimen: Harrison and Walker, Headonornis from the London Clay is ‘‘a
1979a;fig.4C),collectedfromtheLowerOl- member of the Presbyornithidae’’. However,
igocene Bembridge Marls at Burnt Wood, on the basis of the material avialabletothem
Isle of Wight, by R. Ford in 1978. at the time, their conclusion cannot be con-
TAXONOMICREMARKS:Asdiscussedabove, firmed. As discussed above, since the holo-
the original holotypespecimenofH.hanton- type on Headonornis hantoniensis is a cor-
iensis is a partially complete right coracoid acoid, it is not certain whether the two re-
(Harrison and Walker, 1976). ferred humeri (Harrison and Walker, 1976,
As noted by Harrison and Walker (1976), 1979a) can be associated to the taxon on any
thisspecimenisverysimilartothecoracoids basis other than their large size; the coracoid
of Recent anseriforms, such as Cygnus and andhumeriareclearlyfrombirdsofasimilar
Anseranas. The cotyla scapularis is deepand size. Subsequent investigation, however,
rounded (as appears to be primitiveformod- shows that the second referred humerus,
ern birds in general; Mayr, 1999), the im- BMNH PAL 5105 (Harrison and Walker,
pressio m. sternocoracoidei is poorly devel- 1979a), is identical both in size and pre-
oped and shallow, and the processus procor- served morphology to the holotype of Pres-
acoideus is blunt, having some ventral cur- byornis isoni (Olson, 1994).
vature at its tip. The morphology of this At the time of his description of P. isoni,
specimen agrees with that of Presbyornis Olson (1994) omitted discussion of the Lon-
pervetus(ascodedbyLivezey,1997,andfig- don Clay material from his paper. However,
ured by Ericson, 1999a) in the presence of a themorphologyofanotherspecimen,BMNH
small foramen pneumaticum on the proces- PAL 6240, does serve to confirm the suspi-
sus procoracoideus and in the absence of a cions of Harrison and Walker (1979a) that at
large and excavated foramen on the sternal least some of the London Clay material
face (dorsal surface). More generally, should be referred to the Presbyornithidae.
BMNH PAL 30325 has a smallandsemicir- BMNH PAL 6240 was collected from the
cular processus acrocoracoideus and a non Totland Bay Member of the Middle-Upper
excavated articularis clavicularis. With re- Eocene Headon Hill Formation at Hordle,
spect to these specific features, and overall, Hampshire,in1995byA.Jord.Thismaterial
this element is very similar (other than in consists of a partially complete, but frag-
size) with the known coracoids of P. perve- mented, right humerus with both the proxi-
tus (Ericson, 1999b). Hence, the holotype of mal and distal ends, as well as much of the
H. hantoniensis can be referred to the family shaft.Theproximalendpreservesaboutone-
Presbyornithidae Wetmore. third of the crus dorsale fossa (broken be-
A similar-sized member of this family is neath the fossa pneumotricipitalis), and the
known from the Tertiary of North America. distal end lacks the medial area lateral to the
Olson (1994) described a ‘‘giant’’ species of condylus dorsalis (fig. 5). Nevertheless, the
the genus Presbyornis, P. isoni, on the basis morphology of this specimen confirms the
of a left humerus lacking the distal end association between the two specimens orig-
(USNM 294116: holotype) and the left pha- inally referred to H. hantoniensis (BMNH
lanx of the alar digit (USNM 294117) from PAL 3686 and PAL 5105; Harrison and
the Palaeocene of Maryland. Olson (1994) Walker, 1976, 1979a), as does the presbyor-
8 AMERICAN MUSEUM NOVITATES NO. 3354
Fig.5. AdditionalfossilmaterialreferredtoHeadonornishantoniensis(BMNHPAL6240).A,right
proximalhumerusincaudalview;B,cranialview;C,referreddistalendoflefthumerusincranialview
(BMNH PAL 5105). Characters based on the hypothesis of Ericson (1999) are labeled. Scale bars are
5 mm. Abbreviations as follows: a, incisura capitis humeri; b, impression of M. scapulohumeralis
cranialis; c, tuberculum supracondylare ventrale; d, epicondylus dorsalis.
nithid nature of this material on the basis of byornithid characters (fig. 5), including the
the osteological features discussed by Eric- presence of an excavated area beneath the
son (1999a). On the basis of the hypothesis caput humeri and a prominent scar for the
of Ericson (1997, 1999a), BMNH PAL 6240 attachment of M. flexor carpi ulnaris.
exhibits the presence of four derived pres- In addition, in size and morphology
2001 DYKE: FOSSIL WATERFOWL 9
Fig.6. TheoriginalholotypesternumofProherodiusoweniLydekker(BMNHPAL43164)inright
lateral view. Scale bar is 10 mm.
BMNH PAL 6240 is identical to both the Proherodius Lydekker, 1891
originally referred specimens of H. hanton-
Proherodius oweni Lydekker, 1891
iensis and the holotype of P. isoni (figs. 4,
5). This match is further confirmed by com- Proherodius oweni Lydekker (1891)
parison with the illustrations of a complete
(but crushed) humerus referred to P. isoni Proherodius oweni Harrison and Walker
from the late Paleocene of North Dakota by (1977)
Benson (1999).
Proherodius oweni Harrison and Walker
Having confirmed that these fossil humeri
(1978)
from the Eocene of England are similar to
those described for Presbyornis isoni (see Owen (1846) illustrated and described an
Olson, 1994; Benson, 1999), the synonymy incomplete cranial portion of sternum
of this material remains. Since the holotype (BMNH PAL 43164; fig. 6) as a small wad-
of the species hantoniensis within the genus ing bird, and he referred to the similarity of
Headonornis is a coracoid that can only be the overlapping sulcus carinae seen on the
placed tentatively within the Presbyornithi- specimen to the condition seen in living her-
dae (see above) and cannot be associated ons (family Ardeidae). Owen (1846) did not
with certainty to the originally referred hu- namethisspecimen,leavingthistoLydekker
meri (other than by size), the original genus
(1891) who formally referred it to thefamily
and species is retained here on the basis of
astheholotypeofthetaxonProherodiusow-
the holotype alone. The three fossil humeri
eni. This assignment met with the agreement
(BMNH PAL 3686, 5105, and 6240) are
of Harrison and Walker (1977), who de-
therefore referred to the taxon Presbyornis
scribed and re-figured the specimen afterad-
isoni Olson. It is likely,however,thatthege-
ditional preparation.
neric name Headonornis Harrison and Walk-
er is a junior synonym of Presbyornis Wet- TAXONOMIC REMARKS: Harrison and Walk-
er (1978, 1979c) remarked on the possible
more, and that the species name isoni Olson
referraloftheholotypeofProherodiusoweni
is preoccupied by the earlier name hanton-
to the family Presbyornithidae, and they
iensis Harrison and Walker; however, the re-
(1979c) noted that ‘‘the shape and angle of
covery of additional fossil material will be
required to confirm this. articulation of the surfaces of the coracoid
The remaining fossil specimen referred to sulci, in differing from those of herons,
H. hantoniensis by Harrison and Walker matchedthoseofthelargeHeadonornishan-
(1979a: BMNH PAL 4989, the proximal end toniensis of the Upper Eocene’’. It is unclear
of a right scapula) is regarded as Aves in- what to make of this remark: Harrison and
certae sedis. This specimen cannot be re- Walker (1976) made no mention of the re-
ferred to the holotype of H. hantoniensis ferral of a sternum to H. hantoniensis, and
with certainty. no museum catalog number was given in the
10 AMERICAN MUSEUM NOVITATES NO. 3354
later paper either (Harrison and Walker, portion of sternum) within his fossil gavi-
1979c). iform (loon) taxon Colymboides anglicus.
The possibleaffinityofthisspecimenwith The genus Colymboides had been created on
thefamilyPresbyornithidaewasnotedbyEr- the basis of a humerus from the Lower Mio-
icson (1999b), but in this paper the taxon cene of France by Milne-Edwards (1867–
name Proherodius oweni Lydekker was in- 1871). The sternum (BMNH PAL 4355)
correctly attributed to Harrison and Walker specimen was originally cataloged as part of
(1978). Of the features of the sternum listed BMNH PAL 30330, along with a partially
by Ericson (1999b: 4), only the character complete left coracoid, the holotype of Co-
‘‘right sulcus articulariscoracoideuscrossing lymboides anglicus Lydekker. The sternum
belowleftpastmidpointoftherostrumsterni was removed by Harrison and Walker
witharidgepresentbetweenthetwo’’(How- (1976), made the holotype of Howardi eous,
ard, 1955; Ericson, 1997) is preserved in and placed within the order Anseriformes.
Proherodius (Harrison and Walker, 1978, Both the specimens BMNH PAL 30330 and
1979c). This feature was listed by Ericson 4355 are from the same locality, the Upper
(1997: 471) as present in the membersofthe Eocene of Hordle (Hordwell), Hampshire,
Presbyornithidae, as well as in Phoenicop- and were purchased from the Hastings Col-
terus and Threskiornis, but absentin remain- lection in 1885.
ing anseriforms. However, although this When it subsequently proved that the ge-
specimen is of a similar size with respect to neric name Howardia was preoccupied by a
the known specimens of Presbyornis perve- bug (Hemiptera), Harrison and Walker
tus (Howard, 1955; Ericson, 1999b), the (1979b) proposed the replacement name Pa-
presence of crossed coracoidal sulci is wide- laeopapia for this taxon. They later tenta-
spread among basal Neornithes, seen also in tively referred an additional specimen,anin-
Lithornithidae, for example (Houde, 1988). complete left coracoid (SMC C20949; cast
At present, more anatomical details are re- BMNH PAL 4405; fig. 7A), to the genus
quired to confirm the placement of Proher- (Harrison and Walker, 1979a) collected from
odius within Anseriformes and Presbyorni- the Lower Oligocene Hamstead Beds, Isle-
thidae (Dyke, 2000). This situation would be of-Wight (date unknown).
complicated further by the resulting synon- TAXONOMIC REMARKS: Harrison and Walk-
ymy of the two existing generic names— er (1976) noted that BMNH PAL 4355
Proherodius Lydekker (Lydekker, 1891) ‘‘shows no similarity to gaviid sterna’’ but
would clearly have priority over Presbyornis they did not dicuss this further. When Ly-
Wetmore, as stated by Harrison and Walker dekker (1891) named the taxon C. anglicus,
(1978). At present, this material is consid- he noted (p. 193) that the holotype coracoid
eredtobeAvesincertaesedis,becauseofthe (BMNH PAL 30330) fitted perfectly intothe
lack of diagnostic characters seen in the ex- preserved sulcus carinae of the incomplete
isting material. sternum. Since this is the case, I agree with
theobservationsofLydekker(1891)thatthis
Palaeopapia Harrison and Walker, 1979b specimen must be retained as the single re-
ferredspecimenofC.anglicus,placedwithin
Palaeopapia eous (Harrison and Walker,
the Gaviidae until further materialpertaining
1976); Harrison and Walker, 1979b
to this taxon becomes available. The incom-
Colymboides anglicus Lydekker (1891) plete left coracoid tentatively referred to Pa-
laeopapia eous by Harrison and Walker
Colymboides anglicus Lydekker; listed by
(1979a) is not diagnostic at the ordinal level
Lambrecht (1933)
and is regarded here as Aves incertae sedis.
Howardi eous Harrison and Walker (1976) Although this specimenisverysimilartothe
coracoidsoftheRecentteal(e.g.,Anas),hav-
Palaeopapia eous Harrison and Walker
ing both a cup like cotyla scapularis and an
(1979)
oblique facies articularis sternalis (fig. 7A),
Lydekker(1891:192)placedthespecimen neither of these characters has been hypoth-
BMNH PAL 4355 (an incomplete anterior esised as a synapomorphy of Anseriformes