616

JC-NRLF

C a T3S 533

GIFT OF
Charles L. Camp

EARTH

MEMOIRS

or Tin:

CARNEGIE MUSEUM.

VOL. I. NO. 3.

THE OSTEOLOGY OF THE STEGANOPODES.
BY R, W. SHUFELDT, M.D.

At the present writing it is a little over nineteen years ago since I first paid any
attention to the osteology of the steganopodous birds. My initial paper upon this
subject was a brief one devoted to the osteology of Phalacrocorax bicristatus (Science,
Vol. 2, No. 41, Nov. 16, 1883, pp. 640-042, 3 figs.).1 The figures illustrating
that paper have since been used in text-books on ornithology and zoology to
Home extent. As will be observed in the body of the present memoir, five years
later I published another contribution to this subject, entitled "Observations upon
the Osteology of the Orders Tubinares and Steganopodes" (Proc. U. S. Nat. Mus.,
Vol. XL, 1888, pp. 253-315, 43 figs, in text); and also in January, 1889, "Notes on
Brewster's and the Blue-footed Gannet" (The Auk, Vol. VI., No. 1, p. 67). What
I have done in the matter of fossil birds of this suborder, will be chiefly found in
my memoir, " A Study of the Fossil Avifauna of the Equus Beds of the Oregon
Desert" (Jour. Acad, Nat. Sci. Phila,, Vol. XL, Pis. XV.-XVII, (4to), Phila, Oct., 1892,
pp. 389-425), and a number of abstracts of the same, subsequently published else-
where, as in The American Naturalist, and The Auk. In 1894, in the Proceedings of
the Zoological Society of London, I published a brief paper " On the Affinities of
the Steganopodes" (Feb. 20, pp. 160-162), and in the Ills of the same year, in a
paper entitled " On Cases of Complete Fibula; in Existing Birds," made a study of
the fibula in the Sulidx (Vol. VI., No. 23, London, July, 1894, Art. XXIX., pp. 361-
366, figs. 1, 2). The Auk also published a paper of mine In October^ Tr>f)4, making

1 See also ibid., p. 822 ; also Vol. III., No. 53, Feb. 8, 1884, p. 143 ; also a rejoinder to Dr. Theo. N. Gill, ibid., III.
No. 63, pp. 474, 475, Apr. 18, 1884.

109

89O420

11.0. MEMOIRS OF THE CARNEGIE MUSEUM

• • •
. • • •
• * *• •

certain references to these birds, it being entitled, " Notes on the Steganopodes and
^•'•Ffcfeil Birds' Eggs" (Vol. XL, No. 4, pp. 337-339) ; and there was likewise a note
published about them on November 6, 1894 (P. Z. S., p. 608).

Since then and up to the present time, (October, 1902), no paper or publica-
tion of any importance whatever of mine has been published about the Steganopodes,
and especially about their osteology. In this interim, however, I made, as far as
the material at hand would permit me, a more or less extensive study of the
osteology of the entire group. These researches from time to time were written out
and with them incorporated the facts brought out in my earlier publications. As
time passed on, too, Mr. F. A. Lucas, of the U. S. National Museum at Washington,
published a number of very important and interesting papers on the osteology of
the Steganopodes, and as two of the best of these were of no great length, they like-
wise, with their figures, are herewith incorporated.

Moreover, during the last seven or eight years, Mr. Lucas with great industry has
collected together for the osteological collections of the U. S. National Museum the
finest series of skeletons of steganopodous birds existing anywhere in any institution
in the world. With great courtesy and marked generosity Mr. Lucas has placed all
this material at my disposal, to be studied and utilized in the present connection,
and for this and for many other favors in the same direction, altogether too
numerous to mention, my most sincere thanks are due to that distinguished anat-
omist. I am greatly indebted, too, to the courtesy of the U. S. National Museum in
allowing me to take to my residence in Washington, from time to time, specimens
to be photographed by me, which latter, being reproduced, form the material repre-
senting many of the figures in my plates. My private collections have likewise
furnished specimens not as yet existing in the U. S. National Museum, and, as will
be seen, a number of these will also be found among my figures on the plates. I
am also indebted to other persons who have kindly collected material for me, but
I believe in each case such assistance is duly noticed in the body of the present
memoir. It also gives me pleasure to extend my thanks to my wife Alfhild for
having made a very fair and correct copy of the notes I have collected on the
osteology of the Steganopodes for a number of years past, and these may now be
presented in the following manner :

We have this suborder of birds very fully represented in the North American
avifauna. Of the first family in it to be considered in this memoir, — the Phaethon-
6rTropic~Birdc>we have at least two good species, viz: — Phaethon flavirostris,
the Yellow-billed Tropic Bird, and P. sethereus, the Red-billed Tropic Bird. Gannets
of the family Sulidx are still raore numerous, but they all belong, apparently, to the

SHUFELDT: THE OSTEOLOGY OF THE STEGANOPODES 111

genus Sula. There is S. cycmojw, 8. mla, S. brewsteri, S. gossi, S. piscator, and finally,
the well-known Gannet Sula bassana. The Anhingidss or Darters, are represented
by the common Anhinga, or Snake Bird, (Anhinga anhinga). Ornithologists have
placed all our Cormorants in the genus Phalacrocorax, (Phalacrocoracidse), and of
them, species and subspecies taken together, there seem to be nearly a dozen vari-
eties. Next we have three Pelicans of the family Pelecanidx; these are Pelccanus
erytkrorhynchus, the American White Pelican; P. fuscus, the Brown Pelican; and
lastly, P. californicus, the Californian Brown Pelican. Finally, there is the Man-
o'-War Bird, Fregata aquila of the family Fregatidse, — a conspicuous steganopod
of tropical and subtropical coasts generally.

Newton has said "that the tropic birds form a distinct family, Phaethontidx, of
the Steganopodes (the Dyspowmorphss of Professor Huxley), was originally maintained
by Brandt, and is now generally admitted, yet it cannot be denied that they differ
a good deal from the other members of the group ; 2 indeed, Professor Mivart, in the
Zoological Transactions (X., p. 364) will hardly allow Fregata and Phaethon to be
Steganopodons at all ; and one curious difference is shown by the eggs of the latter,
which are in appearance so wholly unlike those of the rest. The osteology of two
species has been well described and illustrated by Professor Alph. Milne-Edwards
in M. Grandidier's fine Oiseaux de Madagascar (pp. 701-704, pis. 279-281 a)."3

The same distinguished authority has remarked of the Sulidse that " structurally
the Gannet presents many points worthy of note, such as its closed nostrils, its
aborted tongue, and its toes all connected by a web — characters which it possesses
in common with most of the other members of the group of birds (Steganopodes) to
which it belongs. But more remarkable still is the system of subcutaneous air-cells,
some of large size, pervading almost the whole surface of the body, communicating
with the lungs, and capable of being inflated or emptied at the will of the birds
This peculiarity has attracted the attention of several writers — Montagu, Professor
Owen, (Proc. Zool. Soc., 1831, p. 90), and Macgillivray ; but a full and particular
account of the anatomy of the Gannet is still to be desired."

Some of our Sulidse, as S. sula, S. cyanops, and Sula piscator are known as Boo-
bies, from their apparent excessive stupidity, but which the writer pleases to call too
little dread or fear of the great destroyer — man. Sailors are very prone to the
taking of bird-life without stint and for no other purpose than amusement, when-
ever they get the opportunity upon lonely islands where sea-fowl abound. Myriad.

1 Sulidic (Gannet), Pelecanidx (Pelican), Plotidie (Snake- Bird), Phalacrocoracidie (Cormorant), and Fregalida;
( Frigate-Bird ).

•Newton, Al., F.R.S. Art. " Tropic-Bird," Encycl. Brit., 9th Ed., Vol. XXIII., p. 588, 1888.

112 MEMOIRS OF THE CARNEGIE MUSEUM

of Penguins and Gannets have thus perished. Superficially a Booby is distinguished
from a Gannet in that the former lacks the median stripe of naked skin over the re-
gion of the throat, so characteristic of the latter birds. As a rule, too, Boobies are
tropical species, and breed in trees, and are of wide distribution.

Morphologically, the genus Anhinga has received more attention than any other
of the Steganopods. " Beside the excellent description of the American bird's ali-
mentary canal furnished to Audubon by Macgillivray, other important points in its
structure have been well set forth by Garrod and Forbes in the Zoological Proceed-
ings (1876, pp. 335-345, pis. XXVI.-XXXIIL; 1878, pp. 679-681 ; and 1882, pp.
208-212), showing among other things that there is an appreciable anatomical dif-
ference between the species of the New World and of the Old ; while the osteology
of P. melanogaster has been admirably described and illustrated by Professor Milne-
Edwards in M. Grandidier's great Oiseaux de Madagascar (pp. 691-695, pis. 284-285).
In all the species the neck affords a feature which seems to be unique. The first seven
of the cervical vertebrae form a continuous curve with its concavity forward, but the
eighth articulates with the seventh nearly at a right angle, and, when the bird is at
rest, lies horizontally. The ninth is directed downwards almost as abruptly, and those
which succeed present a gentle forward convexity. The muscles moving this curi-
ous framework are as curiously specialized, and the result of the whole piece of
mechanism is to enable the bird to spear with facility its fishy prey." (Newton, Arts
" Gannet," and " Snake-Bird," loc. tit.., p. 188.

Again we find the same eminent ornithologist under the article " Frigate-Bird" 4
declaring that that interesting Steganopod "was placed by Linnaeus in the genus Pel-
ecanus, and until lately its assignment to the family Pelecanidne has hardly ever been
doubted. Professor Mivart has, however, now declared (Trans. Zool. Soc., X., p.
364) that, as regards the postcranial part of its axial skeleton, he cannot detect suf-
ficiently good characters to unite it with that family in the group named by Pro-
fessor Brandt Steganopodes. There seems to be no ground for disputing this
decision so far as separating the genus Fregata from the Pelecanidse goes, but sys-
tematists will probably pause before they proceed to abolish the Steganopodes, and
the result will most likely be that the Frigate-Birds will be considered to form a
distinct family (Fregatidse) in that group. In one very remarkable way the oste-
ology of Fregata differs from that of all other birds known. The furcula coalesces
firmly at its symphysis with the carina of the sternum, and also with the coracoids

* " ' Man-of-War Bird ' is also sometimes applied to it, and is perhaps the older name, bnt is less distinctive, some
of the larger Albatrosses being so called, and, in books at least, has generally passed out of use." [Man-o'-\var Bird is
the vernacular name given to this species in the A. O U. Check-list of North American Birds, and is the one in com-
mon use by ornithologists in the United States.]

SHUFELDT : THE OSTEOLOGY OF THE STEGANOPODES 113

of the upper extremity of each of its rami, the anterior end of each coracoid coales-
cing also with the proximal end of the scapula. Thus the only articulations in the
whole sternal apparatus are where the coracoids meet the sternum, and the conse-
quence is a bony framework which would be perfectly rigid did not the flexibility
of the rami of the furcula permit a limited amount of motion. That this mecha-
nism is closely related to the faculty which the bird possesses of soaring for a consid-
erable time in the air with scarcely a perceptible movement of the wings can hardly
be doubted, but the particular way in which it works has yet to be explained " (loc.
clt., Vol. IX., p. 786).

Among others, the birds we propose to consider osteologically in the present
memoir constitute the ORDER STEGANOPODES of the Check List of the American
Ornithologists' Union (1886, p. 106), where they are divided into the families Phae-
thontidse, Sulidse, Anhingidse, Phalacrocoracidss, Pelecanidw, and Fregatidx, — a family
arrangement adopted here. According to Reichenow the " Steganopodes " are an
Order (IV.) of the NATATORES, and contain but three families, the Graculidse>
Sulidte, and Pelecanidte? while Dr. Stejneger arrays them as follows :

Order X. Superfamily. Family.

Pelecanidre.
Sulidse.

STEGANOPODES.

(XVI.) Pelecanoidete.

PhalacrocoracidaB.

Anhingidse.
(XVII.) Fregatoideas.

(XVIII.) Phffitontidea?.

Professor Fxirbringer, in his great work, places the " Steganopodes " (a Gens) in
his Suborder Ciconiiformes, of the Order PELARGORNITHES, and divides them into the
four families Phadhontidss, Phalacrocoraddss, Pelecanidse, and Fregatidx. A still dif-
ferent arrangement is proposed by Mr. Seebohm, and the place they are supposed to
occupy in the system according to his views will later on be given by me in my
Osteology of the Tubinares.

Garrod has said " The Steganopodes, which do not form so natural a family, in
my eyes, as in those of many ; for their myological formula is not the same in all,
being

In Phsethon A. XY,

In Sula and Phalacrocorax AX,

In Fregata A,

•Die Vogel der Zoologiscben Garten, 1882.

PELECANIFORMES.

Phalacrocoraces -

114 ' MEMOIRS OF THE CARNEGIE MUSEUM

from which it may be inferred that Phaethon approaches the Ciconiidas and Fregota
the Accipitres. They all possess the ambiens, caeca, a tufted oil-gland, and the four
toes included in a web, which is but imperfectly developed in some. Sida and
Phalacrocorax, with Plotus, form one family, Phaethon another, Fregata a third, and
Pelecanus a fourth." (Coll. Sci. Mem., p. 221.)

Dr. Sharpe does not concur in this opinion, and in his " A Review of Recent
Attempts to Classify Birds" (Budapest, 1891) places this group betwixt the Anseri-
formes and the Cathartidiformes, thus :

Order (XXIII.) Suborders. Families.

Phaethontes - - - Phaethontidse.
Sulse ------ Sulidte.

f Phalacrocoracidse.

1 Plotidse.

Pelecani - - - - Pelecanidse.
Fregati Fregatidee.

In volume I of his " Hand List" (page 232), recently issued, this arrangement is
somewhat changed. There the Pelecaniformes stand between the Ichthyornithi-
formes (Order XXII.) and the Cathartidiformes (Order XXIV.). They are then
divided into the eight following families, viz.: 1. Phalacrocoracidse. 2. Odonto-
pterygidse. 3. Plotidse. 4. Sulidse. 5. Fregatidse. 6. Phaethontidse. 7. Pele-
canidse. 8. Pelagornithidse. Of these 2 and 8 are extinct groups (1899).

Nearly a quarter of a century before Sharpe's work appeared Huxley in his P.
S. Z. memoir (1867) had placed the steganopods in;his group Dysporomorphse of the
Desmognathse, and had said of them : " The rostrum is long and pointed and more
or less curved, and the external nasal apertures are very small.6 There are no basi-
pterygoid processes. The palate-bones unite for a considerable distance behind the
posterior nares, and send down a vertical crest from their junction."

" The maxillo-palatines are large and spongy. The angle of the mandible is
truncated. The sternum is broad, and its truncated posterior edge is either entire
or has a shallow excavation on each side of the middle line.

" The hallux is turned forwards or inwards, and is united by a web with the
completely webbed anterior toes. The ratio of the phalanges is as in the preceding
genera.

" The oil-gland is surmounted by a circlet of feathers.

6 They are now known to be entirely absent in some of the genera as Sula and others. — S.

SHUFELDT : THE OSTEOLOGY OF THE STEGANOPODES 115

"This group answers to the ' Steganopodes ' of Illiger; and since the appear-
ance of the admirable memoir of Brandt, ' Zur Osteologie der Vogel,' in 1840, no
doubt can have been entertained as to its extremely natural characters. The genera
composing it are sharply divided by the structure of the skull, described above, into
two groups — the one containing the Pelicans, the other the remaining genera"
(pp. 461-462).

Doctor Hans Gadow makes'an Order 9 — the Procellariiformes, and an Order 11,
— the Falconiformes, — between which he places his Order 10, — the Ardeiformes.
These last are thus characterized :

10. ARDEIFORMES.

Cosmopolitan Aquatic.
Young passing through a downy stage.
Oil-gland tufted. Aquito-cubital.
Humero-coracoid deep. No ectepicondylar process.
Desmognathous. No basipterygoid processes.

I. STEGANOPODES.

Cosmopolitan. Aquatic-nidicolous.

Piscivorous.

Rhamphotheca compound. Nares impervious.

No supraorbital glands. Angulare truncated.

Neck without apteria.

Legs short ; all the four toes webbed together. (Unique.)

Hypotarsus complex. Flexors type of II.

Orthoccelous type II. Tongue rudimentary.

1. Phaetontidse.

15 cervical vertebrae.
Procoracoid process large.
Garrod's symbol AXY + .

2. Phalacrocoratidte (including Sulinse, Plotinse, Phalacrocoracinse).

18-20 cervical vertebrae.
Garrod's symbol AX + .

3. Pelecanidse.

17 cervical vertebrae.
Procoracoid process small.
Garrod's symbol A + .

116 MEMOIRS OF THE CARNEGIE MUSEUM

4. Fregatidse.

15 cervical vertebrae.
Procoracoid process small.
Garrod's symbol A + .

II. HERODII.

Cosmopolitan. Waders. Nidicolous.

Zoophagous. Bill long, pointed, laterally compressed, with simple rham-

photheca. Nares pervious.
No supraorbital glands.
Neck long, with long apteria.

Downs of adults only upon the apteria. (Unique among Ardeiformes.)
Legs long ; four toes not webbed.
Hypotarsus complex. Flexors of type I. or VII.
Orthocoelous. Type II. Caeca rudimentary.
Tracheo-bronchial muscles attached to second bronchial rings.

1. Ardeidse.

19 or 20 cervical vertebrae.

Several pairs of powder-down patches.

1 1 primaries. — Cosmopolitan.

2. Scopidse.

16 cervical vertebrae.

No powder-down patches.
10 primaries. — Ethiopian.

III. PELARGI.

Cosmopolitan. Waders.
Neck long, without apteria. Nares pervious.
Rhamphotheca simple.
Legs long. Hypotarsus simple.
Intestinal type IV. telogyrous.
1. Ciconiida;.

Zoophagous. Nidicolous.

17 cervical vertebrae.
Hallux long, toes not webbed.
Flexors of type I.

Tongue rudimentary.

Caeca rudimentary.

Syrinx without Tracheo-bronchial muscles.

SHUFELDT : THE OSTEOLOGY OF THE STEGANOPODES

117

2. Phcenicopteridse.

Tropical. Nidifugous.

18 or 19 cervical vertebrae.

Hallux small, front toes webbed. Flexors of type IV.

Tongue large and thick.

Cseca functional.

Syrinx with tracheo-bronchial muscles.

In 1889, as has b3en stated above, the present writer contributed some brief
"Observations upon the Osteology of the Orders Tubinares and Steganopodes " to
the Proceedings of the U. S. National Museum (pp. 286-314), and so far as the
steganopodous birds were concerned there appeared in that article an illustrated
account of the skeleton of Sida bassana; some remarks, also with figures, on the
osteology of Cormorants and the Brown Pelican. That paper, with its figures,
will be incorporated into the present memoir, but at this writing I am better off for
osteological material wherewith to render a description of the skeletal characters of
this suborder. My observations at this time are based principally but by no means
altogether upon a study of the following skeletons and parts of skeletons.

LIST OF MATERIAL.

Name.

Phaethon flavirostris.
Phaethon sethereus.

Sula bassana.
Sula piscator.
Sula cyanops.
Sula gossi.

Sula brewsteri.

Anhinga anhinga.
Phalacrocorax urile.
Pelecanus fuscus.
Pelecanus fuscus.
Fregata aguila.

Material.

Remarks.

Perfect skeleton.
Four perfect skeletons.

Skeleton nearly complete.

Skeleton.
u

Three complete skeletons.

Skeleton.

11

Skull and mandible.
Skeleton.

No. 17,841. Smithsonian collections.
Author's collection. The gift of E. .1. Reed

Esq., of Guaymas, Mexico.
No. 16,643. Smithsonian collections.

" 18,739. " "

" 18,542. "

Author's collection. The gift of E. J. Reed,

Esq., of Guaymas, Mexico.
Author's collection. The gift of E. J. Reed,

Esq., Guaymas, Mexico.
No. 18,259. Smithsonian collections.

" 18,982. " "

Author's collection.
No. 18,483. Smithsonian collections.

" 18,485. " "

ON THE SKELETON IN PHAETHON.7

Of the Skull, etc. — Of the two species of Tropic Birds, which we have to consider
here, the red-billed one is the larger species, and this difference is quite apparent in
their skulls, though in other particulars they are very much alike. In P. xthereus

'Dr. Sharpe, in his recent Hand-List of Birds (1899, Vol. 1, p. 238), recognizes the following species of Phaethon as
representing the genus throughout the world, viz : P. rubricnuda, P. lepturua, P. fulnus, P. americanus, P. aethereus, and
P. indicm. In this enumeration P. americanus and P. flavirostris are one and the same species.

118

MEMOIRS OF THE CARNEGIE MUSEUM

the skull has an average length of about 10.8 centimeters, while in the other form,
P. flavirostris, it equals for the same measurement but 9.2 centimeters. Measuring
from the mid-point of the cranio-facial hinge, we find that in the yellow-billed
species the length of the supero-mandibular portion of the skull exactly equals in
length the cranial part, while in P. xthereus the facial portion slightly exceeds the
cranial. Viewed from above (Plate I., Fig. 3), it will be seen that the supero-man-
dibular moiety is distinctly marked off from the cranium by the ver}^ distinct trans-
verse cleft forming the cranio-facial hinge, which latter admits of considerable
mobility in the dried skull. The upper osseous bill is broad and massive at its
base, but almost immediately tapers as we proceed anteriorly towards its acute
apex. The culmen is slightly decurved, while below the latero-external margins
are cultrate. Just within these there runs a furrow nearly to the apex, which is
bounded mesially by a raised bony ridge. Internal to this again, the central
portion, is excavated, from the maxillo-palatines to the apex, being broad posteri-
orly and gradually tapering to the front. A delicate medio-longitudinal ridge

FIG. 1. Right lateral view of the skull of Phaethon lethereus. Natural size. Drawn by the author from a speci-
men in his own collection.

marks the anterior half of this space, which is about equally well marked in
the two species. In these birds there is always found a small foramen, perforating
the mandible upon either side, just anterior to the lateral terminations of the
cranio-facial cleft ; while beyond these, the narial apertures, one upon either aspect
of this osseous beak, are seen to be of no great size, being in form elongated,
ellipsoidal, communicating, and with smooth, rounded margins. Minute marginal
and culminar foramina occur in this part of the skull, while the intervening surface,
or what is really the sides of the upper osseous mandible, is delicately scrolled over
with fine anastomosing venations (Fig. 1).

The rhinal chamber is somewhat capacious, and the hollow part of this region,
clear to the apex, is more or less filled in with bony tissue of the cancellous variety.

SHUFELDT : THE OSTEOLOGY OF THE STEOANOPODES 119

The cranium is broad and spreading behind, and but moderately narrowed
anteriorly, and this form defines the shape of the superocranial aspect of the skull,
where the smooth, rolling frontal portion is succeeded, as we pass backwards, by
the flat and wide space that separates the sharp marginal borders of the orbits above.
Slight parial elevations characterize the parietal region, and this surface falls away
gradually, upon either side, towards the post-frontal processes. Deeply sculpt, the
crotaphyte fosste are mesially separated by a broad interval, smooth and unmarked.
A remarkable feature of the skull of one of these birds, and it is still further
enhanced when the lower mandible is articulated, is the appearance it has of having
been vertically truncated in the transverse plane or direction, or being cut squarely
across, reminding us of the skull in certain Herons and their allies.

Regarding the skull of a Tropic Bird, upon its lateral aspect, we are at first struck
with the large, squarish orbit, bounded below by the straight, stout zygomatic bar ;
posteriorly by the massive quadrate, and the spreading post-frontal process ; ante-
riorly, by the great, free lacrymal bone, with its projecting upper limb, and its ver-
tical portion, which latter is twice perforated posteriorly by conspicuous pneumatic
foramina. The interorbital septum invariably presents a large quadrilateral vacuity
at its central surface, which is usually distinct from the orbital foramina posterior
to it. A mere apology in bone represents the almost thoroughly aborted ethmoidal
wing, or pars plana. The foramen for the exit of an olfactory nerve is exceedingly
small as it occurs upon the anterior wall of the brain-case, and the groove leading
from it is open ; but the anterior foramen seen above the semi-aborted pars plana
is of considerable size, and is usually completely surrounded by bone. In front, the
margin of the rnesethmoid is deeply notched, being produced below as a conspic-
uous, blunt-pointed process, — just the reverse, for example, of what we find in Sula.

Turning to the inferior aspect of the skull we meet with many points of interest,
and a construction of parts by no means typically steganopodous (see Plate I., Fig. 2).
In front the maxiUopalatines are distinctly developed, being elegant concavo-convex
processes, with their convex sufaces parallel with each other but not in contact
mesially. Their concave external aspects are partially filled with a very open can-
cellous tissue of bone, and they have a broad base in each case, which appears in
the adult skull to coossify with the corresponding nasal, maxillary, and perhaps,
premaxillary. The postero-inferior angles of the latter are, upon either side, pro-
duced backwards as a prominent process which, when the lower mandible is artic-
ulated with the skull, outwardly overlaps its margin opposite to them.

The palatines, although in contact for their entire surfaces in the middle line, do
not coossify as is the case in the Cormorants, Gannets and other steganopods.

120 MEMOIRS OF THE CARNEGIE MUSEUM

Either one of this pair of bones is thickened, much cut away posteriorly, but
broadly wedged in among the usual facial elements in front. Above they are
moulded closely upon the rounded rostrum of the sphenoid, being drawn to a fine
point anteriorly. It is along this line beneath that the palatines fuse with the
vomer, an element which has much in its form to remind us of that bone in the
Laridse, being pointed in front, mesially carinated below, and grooved longitudinally
along its upper surface. Laterally, the surface of a palatine looks almost directly
outwards, and it is here that we invariably find one or more pneumatic foramina
leading into the interior of the bone. A palatine head is large and offers an ample
articulatory surface for the corresponding pterygoid. These palatal heads are closely
pressed together, but in front of them, below, the bones send down in common only
the merest suspicion of a midcarination, a character so common in Pelecanus, Cor-
morants, and other species of the suborder. The postero-e'xternal angle of a pala-
tine can hardly be said to exist in the skull of this bird, as the bone so abruptly
slopes away in that part of its body, as will be observed by referring to the figures
in the plates.

The pterygoids are long, stout, straight, subcylindrical, rod-like bones, cupped at
their extremities, and when articulated in situ, stand well above the basitemporal
area. Not a semblance of such a thing as a basipterygoidal process is to be
discovered upon either of them, or at the anterior borders of the basisphenoid. An-
teriorly, their palatine heads articulate in contact with each other ; their divergence
from this point is at an angle of about 20°.

The basitemporal area is triangular in form with a small transverse ridge cross-
ing its center. In front a broad, anteriorly rounded lip of bone underlaps the
double openings for the internal carotid. So prominent in Sula, the paroccipital
processes are here much reduced in size, and the hemispherical condyle is sessile and
unnotched on its superior aspect. The foramen magnum is subcircular in outline,
and is situated, as it were, in the middle of a shallow concavity.

A quadrate has an extensive anterior surface ; transversely elongated, antero-pos-
teriorly compressed, articular facet for the mandible ; a double mastoidal head ; and
a broad, squarely truncated orbital process. This bone is also pneumatic, and its
articular facet for a pterygoid is rounded and projecting. As in Sula, there is a
very deep pit, of no mean size, immediately in front of the mastoidal process of
either quadrate, and when those bones are in situ, they arch over the posterior thirds
of these pits, in such a manner as to have, in each case, the outer mastoidal head
articulate upon the outer border of this pit, and the inner head upon the inner
border. Either one of these deep excavations is situated above and to the outer

SHUFELDT: OSTEOLOGY OF THE STEGANOPODES

121

side of the corresponding osseous opening of the ear, and the foramen ovale. Pos-
teriorly the cranial surface, in so far as the occipital area is concerned, lies in the
vertical plane, and is distinctly reniform in outline, bounded all about by a raised
osseous ridge, with its convex curve above, and the foramen magnum situated me-
siad in the concavity below. The unpierced supraoccipital prominence is fairly
well-developed, and occupies the mid-vertical line of this reniform occipital area,
extending from the foramen to the limiting curve above.

The mandible of a Tropic Bird is somewhat acutely V-shaped in form with large
and deep articular cups for the quadrates. These cups have large pneumatic foram-
ina in their concavities, with a single such opening on the upper side of their short
inturned processes. Behind, they are vertically, as well as completely, truncated.

FIG. 2. Superior view of mandible of Pliacthon xthereus.
men in his private collection.

Outline sketch, natural size, by the author from speci-

Either rarnal moiety is very deep, and laterally compressed, — presenting, well
forward of the articular end, a large, irregular ramal vacuity. The dentary portions
of the mandible have hardly half the height of the ramal limbs ; they are thickened,
and each is deeply grooved for nearly the entire length of its superior margin.
Anteriorly, they run very close to each other, and the terminal symphysis is deep,
averaging in depth 1.5 cm. in P. xthereus, being nearly 2 cm. in some specimens.
In an eye of the species just mentioned I count some 15 osseous sclerotal plates ; they
are remarkable for the extensive manner in which they overlap each other, and
still more for the fact that the anterior ones have a depth of not more than one
third of the posterior ones, a uniform graduation taking place in the intermediate
plates, above and below, in this osseous circlet.

We find, too, an interesting structure in the bony parts of the hyoidmn apparatus
of Phaethon. In the adult, the ceratohyals have fused in ossification with the posterior
half of the glossohyal to form a peg-like bone, which is longitudinally grooved
above and convex below. The first basibranchial is a short, thick piece, expanded
in front to form an articular facet, while behind it supports the very minute, though

122 MEMOIRS OF THE CARNEOIK MUSEUM

free, second basibranchial. Either ceratobranchial is an extremely long, slender rod
of bone, bearing at its distal end a little bit of an epibranchial, which latter is fin-
ished off posteriorly by a thread-like extension in cartilage. Cartilage also tip the
glossohyal in sfront, and laterally, the ends of the ceratohyals may not entirely ossify.

For the trachea, its rings seem to be performed in elementary osseous tissue only,
and in mid-course of this tube, they are delicate structures, being markedly narrow
in depth, and very frail.

Of the Axial Skeleton. — (See Plate XXL, Figs. 5 and 6.) Representatives of the
genus Phaethon possess in the cervical and cervico-dorsal divisions of the vertebral
chain 15 bones. Three of these are cervico-dorsals, the anterior one of which
always supports an exceedingly rudimentary pair of free riblets.

In P. (ethereus they are seen to be in the same condition on the next following
vertebra, or the 14th vertebra of the neck, while in P. flavirostris they are of some
considerable length, though they do not develop epipleurse. Both species have a
well-developed pair of ribs on the 15th segment of the neck vertebrae, that support
epipleural spines on their posterior margins, to which part they are firmly coossified.

The atlas is comparatively small with rather slender neurapophyses. Its cup is
profoundly notched above, for the accommodation of the odontoidal process of the
axis. Below, it develops a low hypapophysial spine projecting backwards, which,
when the bones are articulated in situ, underlaps the centrum of the second vertebra.
A prominent, quadrate hypapophysis also is found to exist beneatn the centra of the
axis and third vertebra, but this feature is very nearly aborted on the fourth cervical.

On the axis vertebra we find a massive neural spine, with conspicuous lateral pro-
jections, each jutting upwards and outwards from above the postzygapophyses. All
these characters are moderately reduced in the third vertebra, while in the fourth
the neurapophysis is very much reduced in size, and the aforesaid lateral or ana-
pophysial processes are all but absent.

The lateral vertebral canals are but semi-closed in the axis ; they are thoroughly
so throughout the rest of the vertebral chain until we meet with the leading cervico-
dorsal, wherein the existence of a pair of rudimentary free pleurapophyses leaves
these canals open.

On the under side of the cervicals, the passage for the carotid canal is seen to
be wide open, and is confined to the 5th, 6th, 7th and 8th vertebra?. Throughout
the series, beginning with the third cervical, we find the parial backward projecting
spines of the parapophyses to be pointed, and exceedingly stumpy and very short.

On the 9th a median hyapophysis makes its appeai'ance, which is larger on the
10th, more quadrilateral in form and laterally compressed on the llth, small on the

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 1 23

12th, and still more so on the cervico-dorsals, where there coexists with it the barest
suspicion of lateral spines.

Viewed dorsad, the third cervical appears to have a quadrilateral form, which
passes to the more oblong shape in the fourth, and the bones of this part of the
column commence to shorten antero-posteriorly with sixth, they being wider than
they are long for the remainder of the spinal series. The last two cervico-dorsals
develop a low, quadrilateral neural spine, which resembles, but is antero-posteriorly
shorter than the same process as it occurs throughout the dorsal vertebrae.

P. fiavirostris has a delicate, osseous, interzygapophysial bar, extending from the
back of either prezygapophysis backwards and inwards to the corresponding ante-
rior bases of the postzygapophyses, in the fifth, sixth and seventh cervical vertebrae.
Usually it only ossifies in the fifth and sixth in P. xthereus, its place being taken by
ligament in the seventh and eighth. No such character as this is ever present in
any of the Sididte. Strong and broad diapophysial processes are first pronounced
in the first and second cervico-dorsals ; in the last cervico-dorsal these processes are
markedly shorter and shallower, while throughout the dorsal series of vertebra} they
are decidedly slenderer, and in their vertical diameters of very little thickness.

Passing to the dorsal series of vertebrae, we find in P.flamrostris six of those bones
freely articulated with each other between the last cervico-dorsal and the leading
one in the consolidated pelvic sacrum. This last-mentioned vertebra, in all my
specimens of P. xtherem, coossifies with the anterior sacral, though, as we shall see
presently, this difference does not affect either the number or the arrangement of
the thoracic ribs. Hypapophyses are fairly well developed upon all the dorsals of
the middle of the back, but they become small as we near the pelvis, to be very
inconspicuous or entirely absent from the last one or two bones. Pneumaticity
characterizes all the vertebrae between the atlas and the first caudal, and the centra
of the dorsals are but moderately compressed in the transverse direction. They
present the usual ornithic plan of articulation with each other, and the same may
be said of the manner in which the true ribs are connected with the segments of this
division of the spinal column. There are five pairs of these ribs that bear uncinate
processes, firmly attached to the middle of their hinder borders, and each pair of
these appendages is slender, slightly overlapping the succeeding rib in every case
save the last, which may or may not be of sufficient length to do so. As for the
ribs themselves, they, too, are long and narrow, sweeping well backwards, a fact that
entails the presence of lengthy hsemapophyses to reach the sternum. These latter
are of no great caliber ; the first pair being nearly straight, and the succeeding pairs
becoming gradually more curved as we approach the last ones, which are the most

124 MEMOIRS OK THE CARNKGIE MUSEUM

so. Finally, there are two pairs of thoracic ribs that are devoid of unciform proc-
esses, the costal ribs of the first of which reach the sternum, but those of the ulti-
jnate pair fail to do so by some considerable distance in both species. From what
has been said in a foregoing paragraph, it will be seen that this gives one true pair
of " pelvic ribs," while the penultimate pleurapophyses and their hsemapophyses in
P. flavirostris articulate with a vertebra of the dorsum which is free, but which in P.
sethereiis is coossified with the pelvic sacrum. So were we to have' studied skeletons
of the latter species only, it would have been said that it possessed two pairs of
"pelvic ribs," and had P. flavirostris alone been examined in this particular, the
descriptive osteologist would have recorded but one pair. This is but another
instance exemplifying the great necessity of having the proper material in sufficient
variety before one for comparison.

Viewing the pelvis from above, we find that the anterior margins of the ilia over-
lap the penultimate pair of thoraric ribs in the Red-billed Tropic Bird, not so, how-
ever, P. flavirostris; and this, taken in connection with the apparent appropriation
on the part of the sacrum of one of the dorsal vertebrae, already referred to above,
in the case of the first-named species, lends to the pelvis a difference of appearance
in the skeletons of the two birds, now under consideration, which is not otherwise
particularly borne out by this bone. In part, the pelves of these specimens are other-
wise very much alike. Disregarding, then, the vertebra to which reference has just
been made, we find the sacrum to be composed of ten others, which are thoroughly
fused together and with the ilia, one upon either side. Taken as a whole, the pelvis
is broad, not especially long, and very shallow and compressed in the vertical direc-
tion. The internal borders of the ilia are nearly parallel to each other, and the
broad sacrum keeps them well apart for their entire length. Such an arrangement,
as we might expect, gives open, shallow " ilio-neural grooves," one on either side
of the sacral crista, anteriorly. These are quite continuous with similar " groov-
ings" carried back posteriorly as far as the first caudal vertebra. In other words,
between the inner borders of either ilium, and the supero-median line of the sacrum,
the diapophyses of the vertebra? are found to be below the general surface ; while
between them, for the entire length of the pelvis, occur foramina of some size, and
more or less subcircular in outline. These single parial rows of interdiapophysial
foramina constitute a striking character upon the superior aspect of the pelvis of
the Tropic Birds, inasmuch as in the preacetabular portion of that bone in Sula,
Plotix, the Pelicans, and the Cormorants, they are out of sight on this aspect by the
internal margins of the ilia meeting the sacral crista. Thus it will be appreciated
that the breadth of the pelvis in Pliaetkon, to which allusion has been made, is

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 125

chiefly due to the broad sacrum, and not to the ilia, which bones, either in their
pre- or postacetabular parts, are quite narrow in their transverse direction. The
former area, in each one, being very slightly concaved, and in the latter, barely
convexed. Antitrochanterian processes are comparatively prominent, and each
overhangs a cotyloid cavity of moderate size, and of the usual ornithic character.
Laterally viewed, we note that the obturator foramen is complete, due to the fact
that either pubic style is in contact for nearly its entire length with the inferior
margin of the corresponding ischium. This obliterates anything like an obturator
space. Each pubic style is long and somewhat slender, and often passing an ischium
behind, curves downwards and very slightly inwards. Here the bone is appreciably
dilated in P. fethereus, which is not the case in the other species we are considering.
An ischium is also narrow, and projects some distance beyond the ilium of the same
side. A very shallow ilio-ischiac notch, on the posterior pelvic border, is formed
by this arrangement ; but by no means as well marked an one as we see in the
more typical steganopods, to be described later on in this memoir.

The ilio-ischiac foramen is large and of a subelliptical outline (see Fig. 6, PL XXL).
On the under side of the pelvis in Phaethon, we are to note the fact that the sacral ver-
tebrse, opposite the acetabula, do not throw out their processes to act as braces to
especially strengthen the ilio-pelvic walls at those points, a feature often seen in the
skeletons of other birds, and, as before found, both by Mivart and myself, to be the
case in Pelacanus, the true cormorants, in Sula, and to a somewhat less conspicuous
degree in Anhinga.

Eight free caudal vertebra, plus an elongated, ploughshare-shaped pygostyle, go to
make up the skeleton of the tail in a Tropic Bird. These vertebrae have spreading
diapophyses, especially the fifth one, and all have neural spines, while the last three
of them possess bifid haemal spines; and this last character is impressed upon the
lower end of the coccyx. The neural canal passes through the entire series. From
this, and including the coccygeal piece, it will be seen, from our account of the spinal
column, as it has been given above, to contain in these birds, forty vertebrae. They are
all pneumatic, except the atlas and those which go to form the skeleton of the tail.

The Sternum. — This bone in my specimen of Phaethon flavirostris has a body of a
quadrilateral form, being almost as wide as it is long. Its carina is deep anteriorly,
where it protrudes far forwards, while behind it merges with the ventral surface of
the body at some distance before it reaches the mid-xiphoidal process. In front the
carinal angle is thickened, and upon its superior aspect there exists a considerable
facet for articulation with infero-median surface of the osfurcula. This facet is not
at the apex, but is found about one third the distance back towards the sternal body,

126 MEMOIRS OF THE CARNEGIE MUSEUM

and the articulation, though very close and extensive, is free, and anchylosis has never
been seen by me to take place between the two bones. The coracoidal grooves decussate,
the median extremity of the right one passing beneath the median extremity of the left.
A very rudimentary manubrium is seen immediately below this point, and below it a
circular pneumatic foramen. Either costal process is broad and subtriangular in form,
and upon either costal border occur six facets for articulation with the hsemapoph-
yses which connect the sternum with the vertebral ribs. On the much-concaved
sternal body, just within the anterior border in the median line, dorsad, there runs
back longitudinally a characteristic bridge of bone, upon either side of which we
find a large subelliptical pneumatic foramen, and usually they are the only two
found in this situation. Other small ones occur, however, in the pitlets among the
hsemapophysial facets on the costal borders. Very broad external xiphoidal proc-
esses of the posterior part of the sternal body, extend further backwards than any
other part of the bone ; the middle xiphoidal process is broadly rounded. The in-
ternal pair of processes are small and delicate, and each one of the pair occupies a
mid-point in the great, round notch, on either side, existing between the mid- and
external xiphoidal process, thus creating a 4-notched sternum for this species, of a
peculiar pattern in this respect. Mivart has said in his "Axial Skeleton of the
Pelecanidfe" (T. Z. S., p. 365), that "in the Pelecanidx there is but a single lateral
xiphoid process on each side, while in Phaethon there are two on each side, in addi-
tion to the median xiphoid process." Farther on we shall see that this by no means
always holds good for P. tethereus. Passing next to the consideration of the sternum
in the species last mentioned, we observe that in the majority of its characters it
agrees with the bone as it has just been described for P. flavirostris. It is propor-
tionately larger however, and presents two striking differences. In the first place,
in all four of my specimens a large, blunt-pointed manubrium is present, which is
transversely pierced at its base b}^ a foramen. Again, the internal pair of xiphoidal
processes may be entirely absent, or only one may be present on one side, and when
such is the case, it appears only as a minute projection from the inner border of the
external xiphoidal process near its extremity. I have no specimens of the sternum
of P. sethereus wherein there is to be found a pair of internal xiphoidal processes,
and consequently its sternum is not four-notched as in its cogener, P. flavirostris.
More or less minute or scattered pneumatic foramina may be found on its thoracic
surface, above the two large ones which occupy similar positions to those noticed in
the Yellow-billed Tropic Bird.

The os furcula is of the broad U-shaped pattern, with its median portion below,
widened, and on the under surface of which occurs the facet for articulation with the

SHUFELDT : OSTEOLOGY OF THE STEGANOPOIJES

127

keel of the sternum as described above. The upper free extremities of the clavicles are
drawn out, rounded, blunt-pointed and the apex of each one, when articulated in situ,
reaches back, on either side, to the antero-mesial angle of the corresponding scapula.

A? scapula is long, very narrow, rounded, and somewhat compressed in the ver-
tical direction. Its distal end only is blade-like, very slightly expanded, and
obliquely truncated from a point within to the apex. It articulates in the usual or-
nithic manner with the coracoid, and neither it nor the os furcula appear to be pneu-
matic, as is the coracoid. This latter bone of the shoulder-girdle has a broad, much
expanded sternal extremity, which supports a moderately prominent lateral coracoid
process. The entire length of the bone is greater than half the length of the humerus,
and its shaft is compressed in the oblique antero-posterior direction. Its head or
summit is tuberous, and also somewhat compressed in the same plane with the
shaft. The process for articulation with the os furcula is large, much flattened from
side to side, and low down in the deep tendinal groove it creates, we observe a
minute antero-posterior perforation of the bone. As usual, the coracoid and scapula
form a well defined glenoid cavity for articulation with the head of the humerus. An
os humero-scapulare is not found.

Of the Appendicular Skeleton. The Pectoral Limb : Apart from the mere matter
of the difference in size, the bones composing the limbs in the skeletons of these two
species of Phaethon are in their corresponding characters .almost exactly alike, and
these again are notably different from those same characters as we find them exem-
plified in the limb-bones of the skeletons in other families of the Steganopodes.
These departures will be noted from time to time later on in this paper, after the
present account of the osteology of the Tropic Birds has been written.

In the brief subjoined table some of the lengths of the bones are compared, the
measurements being given in centimeters and fractions.

TABLE.

Measurements, taken from extreme distal points,

Species.

and not on curved contour lines.
P. flavirostna.

/'. mthereas.

Length of humerus. 8.4

9.3

Length of ulna. 8.9

9.7

Length of carpo-metacarpus. 4.3

4.8

Length of proximal phalanx of index digit.

2.5

2.9

Length of distal phalanx of index digit. 2.4

2.8

Total length of pinion (articulated). 8.8

10.1

Length of femur. 3.2

3.7

Length of tibio-tarsus.

4.7

5.3

Length of tarso-metatarsus.

23

2.8

Length of mid anterior toe.

.8.8

4.2

128 MEMOIRS OF THE CARNEGIE MUSEUM

From this it will be seen that both the proportionate and relative lengths of cer-
tain long bones of the limbs in these two species of birds slightly differ, but the dif-
ference is very slight.

There is no especial association of characters in either the pectoral or pelvic
limbs of Phaethon that in any way remind us of the corresponding parts of the skel-
eton in a Gull, or a Tern, much less any kind of an Auk or a Puffin. The several
bones seem to possess a distinctive character quite their own. The humerus shows
a slight compression in the transverse direction, rendering its humeral head narrow,
and its very moderately curved shaft ellipsoidal rather than circular on section.
The well-marked ulnar tuberosity projects directly anconad, as does the prominent
radial crest project palmad, — the long transverse axis of the first being parallel to
the plane of the latter. At the proximal end of the bone, on its palmar aspect, just
before we come to the humeral head, we meet with a well-defined groove running
at right angles to the long axis of the shaft ; and this groove becomes very deeply
marked at a point on the opposite side of the bone from the ulnar tuberosity.

The "pneumatic fossa" instead of being entirely open, as in a good many birds,
is covered across by a more or less perfect bony plate, which plate shows the single, cir-
cular pneumatic perforation at its center. This hole is smaller in proportion in my
specimen of P. flavirostris than it is in the other species. At the distal end of the
bone the fossa above the oblique and ulnar tubercles is fairly well scooped out ;
while on the ulnar side of the humerus at this end, both the pits for tendinal inser-
tion and the grooves for their passage are unusually well marked and deep. An
exceedingly rudimentary " epicondyloid process " can be seen, but it by no means is
the characteristic feature that we find it to be in the humeri of the Laridx, and
those suborders of birds most nearly related to them.

The radius is straight and of a more or less uniform caliber, while its companion
bone of the antibrachium, the ulna, is considerably bowed, with its shaft presenting
three faces more or less distinctly. These might be here designated as a palmar face
or surface, an interosseous face, and a subanconal face. The line or angle formed by
the intersection of the subanconal and palmar faces, has along it the row of papillae
of the secondary quill-feathers ; and another indistinct row of them passes down
the middle of the subanconal surface. There are fourteen such papillae in the first
mentioned row. The olecranon process is a very insignificant affair, while at the
distal extremity of the bone, at its anconal side, the end is drawn out into a dis-
tinct apophysis, which, in articulation, bends down towards the ulnare ossicle of the
wrist. Two segments compose the skeleton of the carpus in the adult, the usual
radiale and the just-mentioned ulnare.

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 1 29

Carpo-metacarpus presents us for examination the ornithic characters common to
many representives of this class of vertebrates. The coossified pollex metacarpal
is of average size, and the slender shaft of the medius metacarpal is, for the greater
part of its continuity, parallel with the far heavier shaft of index. This last, at the
anterior aspect of its distal end, develops a pronounced little process, which serves
to retain in their proper groove, the tendons which pass in that region during the
life of the individual. The free digit of pollex has the usual form seen in Aves
generally, and the expanded posterior portion of the proximal phalanx of the index
finger is entire, and not perforated as we find it in many of the Laridie. The distal
joint of this digit is long and slender, being once semibarbed near its lip, the pro-
jection standing out from its posterior angle. Long and spine-like, the free joint of
medius is seen to be proportionately more slender in the Red-billed than it is in the
Yellow-billed Tropic Bird. There do not appear to be any free terminal ungual
joints at the ends of any of the fingers in these birds ; and I have found no small
sesamoidal bones either about the carpus or near the elbow joint.

The humerus alone of the bones of the pectoral limb is pneumatic.

Passing to the consideration of the pelvic limb, we find that the femur too, is gener-
ally a pneumatic bone, and is peculiar inasmuch at the foramina occur irregularly at
the proximal end of the shaft either in front or behind, while in P. asthereus a single,
circular pneumatic foramen is found in the majority of skeletons in the popliteal fossa
near its center. For its proximal moiety, tibio-tarsus may likewise be hollow, and,
in life, air gains access to its interior. In one of my specimens a small perforating
foramen is seen between the low cnenial crests very near the summit of the bone.
Below the point indicated, the remainder of the skeleton of the pelvic limb seems
to be non-pneumatic.

The axis of the head of the femur makes a wide obtuse angle with the longitu-
dinal axis of the shaft, and the pit for the insertion of the ligamentum teres is but
very faintly marked upon the top of it. On the summit of the bone we find the
usual articular surface, and above this the trochanterian crest is scarcely at all ele-
vated. At the proximal end of the shaft, between the great trochanter and the caput
femoris, a very appreciable fossa exists, and immediately below this a very distinct
tubercle presents itself in a great many specimens especially of Pha'ethon asthereus. It
at once reminds one of the trochanter minor of the mammalia, and no such a char-
acter is ever to be seen upon the femora of any of the other steganopodous birds
judging from those before me, and I do not now recall ever having seen it upon the
femur of any other existing or extinct bird. This character is but faintly seen in
the U. S. National Museum specimen of the skeleton of P. ftavirostris (No. 17841).

130 MEMOIRS OF THE CARNEGIE MUSKUM

At the distal end of the femur the condyles are not very prominent, and the inter-
nal one is small. Nor is the rotular channel or the popliteal fossa deep, b\it on the
contrary they are both shallower than in many other birds known to me. We find
the shaft in tibio-tarsus to be straight and nearly cylindrical. Proximally, the low
cnemial process rises but a little above the summit of the shaft, while anteriorly,
below it, the pro- and ectocnemial ridges are but very inconspicuously developed.
It is between these latter that the small pneumatic foramen usually is seen. Distally
the usual characters of this bone are to be observed ; the osseous span to confine tendons,
at the antero-distal end of the shaft is present, and passes in a slightly oblique direc-
tion over the channel it bridges ; the condyles, with their usual reniform contour, are
well separated from each other, and the intercondyloid notch between them, is
marked in front, but very shallow behind. In the fibula we see a very slender and
rudimentary bone. Especially is this the case below the 'fibular ridge' on the side
of the shaft of the bone it articulates with, for it is then reduced to little more than
a flattened osseous hair, closely applied to tibio-tarsus, but apparently never actually
anchyloses with it.

Phaethonidse always possess a patella in front of the knee-joint, and it is of an
elongated form, being flat behind and convex in front, where it is distinctly marked
in the transverse direction by the groove for the tendon of the ambiens muscle.

Most remarkable of all the bones of the pelvic limb in Phaethon, however, is its
tarso-metatarsm. To describe this I select the left one from one of my skeletons of
P. tfthereus. The bone is oblong in outline; much flattened in the antero-posterior
direction ; and with its middle trochlear process but slightly lower on the shaft
than the internal one, which latter in turn is but a little lower than the outermost
one. On the summit, the shallow articular depressions for the condyles of the tibio-
tarsus are separated by a rounded eminence in front. At the back of the shaft, in
this region, we find the " hypotarsus" to consist simply of three vertical ridges, —
a short, small middle one, with longer and stouter outer ones. Anteriorly, the shaft
of the bone is deeply scooped out from above, downwards for nearly its entire
length. On the outer side below, this channel terminates in an antero-posterior
perforating foramen, — situated at a point a little above the notch dividing the
mid- and outer trochlear processes. On the inner side below, there is a small, longi-
tudinally-disposed groove that runs into the main excavation on this aspect of the
shaft at about its middle. This smaller groove also terminates below in an antero-
posterior perforating foramen, — situated at a point a little above the notch divid-
ing the mid- and inner trochlear processes. Thus we see in P. xthcreus, there are
two antero-posterior perforating foramina at the distal extremity of its tarso-meta-

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES

131

tarsus; — a very rare condition in existing birds, and even absent in P. flavirostris,
where only the usual outer foramen seems to be present.

At the back of the shaft there is a low median ridge running down from the
lowermost point of the hypotarsus to bifurcate below. One limb is lost on the pos-
terior surface of the mid-trochlear process, the other on the outer one, while between
them occurs the outer of the two perforating foramina just described above. Upon
either side of this ridge, the shaft upon this aspect is also grooved in the longi-
tudinal direction, but not so deeply as it is in front. Placed side by side in the
anterior groove of the bone, just below the head, we find a pair of perforating foram-
ina. They make their exit one upon either side of the lower part of the hypo-
tarsus behind ; each one lying in a postero-longitudinal groove, to which mention
has just been made in the last paragraph. The sides of the shaft of this bone of the
leg are more or less flattened, and the tubercle for the insertion of the tibialis anticus,
is very small. The first metatarsal is free, being long and flake-like, and articulates
by its entire outer margin, with the postero-internal edge of the shaft of the tarso-
metatarsus. Pes has what is usually termed the normal arrangement and number
of joints to the several toes, i. e., 2, 3, 4 and 5 joints, to the hallux, second, third and
fourth digits respectively. Exceedingly slender and long, the basal joint of hallux
is tipped off with a small, slightly curved ungual phalanx. The terminal pha-
langes of the three anterior toes are relatively much stouter, thicker, but exhibit
about the same proportionate amount of curvature. With respect to the shafts of
the intermediate joints, it is to be noted that they are long, slender, and exhibit but
very little curving; while the articular ends of these bones are but very slightly
enlarged. Indeed, the skeleton of the foot in the Phaethonidss is in reality deli-
cately constructed.

TABLE.

Toes and their joints.

P. flnvirostris.
Measurements in mm.

Remarks.

Hallux or First Toe.
Second or Inner Toe.

Third or Middle Toe.
Fourth or Outer Toe.

| Basal jo
( Ungual
I Basal
< Second
( Ungual
f Basal

] Third
[ Ungual
f Basal
Second
\ Third
Fourth
[ Ungual

nt

11

4
15
12
5
12
11
10
5.1
10
8
8
8.5
4

}

1

)

From this it will be seen that
the length of the skeleton of
the middle toe is about * the
length of the tibio-tarsus. On
the other hand it is longer than
either the femur or the tarso-
meta tarsus.

132 MEMOIRS OP THE CARNEGIE MUSEUM

A proportionate increase in their length and caliber seems to be the only im-
portant difference that distinguishes the skeletal toe-joints of P. sethereus from the
corresponding bones in P. flavirostris. To show the relative lengths of the pha-
langes in the last named species, see the above table, in which the measurements
are given in millimeters.

Beyond the patella, already described above, I find no other sesamoids associated
with the bones of the pelvic limb in the Phaethontidas.

OSTEOLOGY OF THE

(See Plates XXII. and XXIII., Figs. 7-12.)

Of the Skull. — Sula bassaua has a skull that averages in the adult about 186 mm.
in length ; whereas in Sula piscator 129 mm. is the average length of the skull in
the adult. Between these extremes we find the other species to fall, and a simliar
variation is size is, of course, applicable to the remainder of the skeleton in these
different specific forms of the Sulidae. Apart from this, the characters presented in
the skeletons of these birds in the main agree very well indeed, though differences
do exist, and these stamp the skeleton of each species with an individuality peculiarly
its own. The more important of these differences will be noted as we proceed.
(See Plates XXII., XXIII., Figs. 7-11.)

In form, the superior osseous mandible is flat upon its under side with cultrate
tomia, while superiorly it is convex from side to side, and tapers from base to apex
gradually to a point, being a little decurved near the extremity. Sometimes we
find it pierced by a foramen on this upper side, which leads to its hollow interior,
but Sula is without nostrils, though did they exist, their position would perhaps be
indicated by the posterior end of the longitudinal furrow that marks the mandible
upon its lateral aspect.9

An osseous, thoroughly adherent crust, appears to overlay the greater part of the
superior surface, the only smooth place being a small area in front of the cranio-
facial hinge, and even this is absent in Sula piscator and other species. Its entire
surface is marked all over by an exquisite anastomosing venation, the ramifications

8 Besides some four fossil forms, Dr. R. Bowdler Sharpe in his recent Hand List of Birds, recognizes the following
species of the genus Sula, in which are included all the birds of this group at present known to ornithologists : — viz :
S. basxana, S. serrator, 8. capenxin, S. cyanops, S. abnoti, S. piscatrix, S. webuleri, S. variegata. S. nebouxi, S. sula and S.
brewsleri. In the A. O. U. ''Check-List" we also finds, gossi, which may correspond with one of the species above
enumerated, as S. nebouxi, or S. websteri.

9 In my memoir entitled " Observations upon the Osteology of the Orders Tubinares and Steganopodes (Proc. U.
S. Nat. Mus. Vol., XL, 1888, pp. 253-315), I give figures of all the principal bones of the skeleton of Sula bassana, drawn
by me, natural size, direct from the specimens. These will not be reproduced here, while the osteology of other species
of Gannets will be illustrated in the plates of the present memoir, as S. gossi and S. brewsteri.

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 133

starting, in some instances, from minute foramina found upon its surface. This
venation is but feebly marked in S. piscator.

A lacrymal is a free bone, articulating with a roughened facet of some extent be-
neath the antero external angle of the frontal above, and by a smooth, gliding facet
on the upper side of the maxillary, which latter bone is thickened in a perpendic-
ular direction and otherwise enlarged in order to offer it the proper amount of sur-
face. As for the bone itself, it is of rather a columnar form, with the exception of
its extended anterior margin, which is roundly notched and shows on its inner side
the large pneumatic opening leading to its hollow interior. In S. cyanops this notch
is extended as a groove entirely across the outer face of the bone, and the pneu-
matic foramen is seen in this groove.

In Gannets there exists, projecting horizontally from the outer margin of the
frontal bone, on either side, from its " prefrontal process," a few millimeters posterior
to the fronto-lacryraal suture, a small rounded ledge of membrano-cartilage, which
reminds one of the horizontal portion of the true lacrymal bone in certain gallina-
ceous birds, as the Perdicinse, for example. This feature has been studied by me in
Sula brewsteri and S. gossi, recently killed specimens for which I am indebted to
Mr. E. J. Reed of Guaymas, Mexico, who kindly collected them for me. This
membrano-cartilaginous process probably never ossifies in the Sulidse.

In the adult bird it is impossible to distinguish the exact position, or any of the
borders, of the nasal bone.

The maxillo-jugal bar shows very plainly the suture between the jugal and
quadrato-jugal ; the latter is much smaller than the other portion, and shows a
strong peg-like process upon the inner aspect of its posterior end, which is about at
right angles to the axis of the bone. It fits in the deep conical socket on the side
of the quadrate. Beyond its enlargement for the lacrymal the maxillary is a thin,
horizontal plate of bone, anchylosed in the usual way at its anterior end. Here it
really enters into the apparatus of the cranio-facial hinge. An ossicle having the
appearance of a process pointing backward and apparently coming from the pre-
maxillary is seen over this horizontal plate of the maxillary on either side. Pro-
fessor Parker found this condition present also in another species of Sula, and this
eminent anatomist apparently describes this ossicle as a "post-maxillary" for those
birds. Either rudimentary or better developed, it is probably present in one form
or another in all of the SulidiS.

The interorbital septum, which is a thin, smooth plate, shows considerable of a
fenestra near its middle, and a few such openings of a very much smaller size pierce
its posterior wall in S. bassana.

134 MEMOIRS OF THE CARNEGIE MUSEUM

The orbital cavity itself in Gannets is very deep, the eaves of its roof almost
overhanging the jugal bar beneath. Its superior periphery is smooth and rounded.
All in front of the rhinal chamber is filled in by the spongy mass formed by the
united maxillo-palatines. The hinder portions of these bones are, however, still dis-
tinct in S. bassami, and they have all the appearance of these elements as they are
found in birds which possess them as concavo-convex plates facing outward.

The rostrum of the sphenoid is a hollow subcylindrical tube, united above with
the interorbital septum. As we proceed anteriorly it becomes more flattened from
side to side, and gradually rises upward. At a point about half way between the
palatines and cranio-facial hinge it terminates in a process directed forward ; above
this is the sharpened ethmoidal margin, nearly perpendicular to the long axis of
the skull. Osseous wings to the ethmoid never develop in ISula, not even rudi-
mentary traces of them being seen at their customary sites.

The cranio-facial hinge is exceedingly perfect in its construction, being composed
of a thin plate of bone occupying the full width of the skull ; the bones both above
and below are separated from each other by a small interval for the entire length of
the transverse line constituting the hinge. The part played in the mechanism by
the maxillaries has already been described above.

We find the sphenotic process to be more or less bifid and jutting directly out
from the side of the skull ; on the other hand, the mastoidal process is a crest of
bone curling forward. Between these two the very wide crotophyte valley
is seen.

The quadrate is a large, massive bone, with its mastoidal head composed of two
prominent ellipsoidal trochlese, separated from each other by an intervening notch.
Below these the shaft is seen to be rather compressed in an antero-posterior direc-
tion, and supports in front at its lower half an unusually formed orbital process.
This is a thin, triangular plate of bone placed in the vertical plane, and with its
apex directed forward. The pneumatic foramen of the quadrate usually occurs on
the posterior aspect of the shaft in most birds, but here it is situated to the inner
and lower side of this orbital process. In Sula cyanops I find two others on the an-
terior aspect of the bone, one near each articular process.

The pit for the quadrato-jugal is cylindrical and deep, and a perforation at its
bottom may lead into the hollow of the bone. On the posterior aspect of the quad-
rate we find an irregular facet for the mandible; it looks directly to the rear and
stands at the head of a longitudinal and deep groove which is found between two
similarly placed facets on the foot of the bone.

Each pterygoid is a trihedral and compressed bone with prominent borders.

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 1-^5

Regarding this skull from a superior view, we see in it a foramen in the superior
mandible, near the site of the narial opening of the majority of other birds. From
this aspect we also have a good view of the wonderfully perfect cranio-facial hinge
of this bird.

Posterior to this is a broad, smooth area, very slightly convex, and showing in
some species numerous venations like those on the bill of S. cyanosis. This surface
extends from the cranio-facial hinge to the anterior border of the crotaphyte fossa?,
while laterally it is bounded by the margins of the orbits.

This view also shows the extent and form of these crotaphyte fossae, and how
they are separated from each other in the median line, simply by an extension
backward of a very narrow strip of the general surface that lies beyond them. They
merge with each other in some specimens of S. piscator. They are bounded behind
by conspicuous and sharpened crests that curl slightly forward, and are best marked
laterally, becoming very low as they near the upper part of the supra-occipital
prominence.

The under view of the skull reveals a number of interesting points. We find
that the anterior portions of the palatines are parallel to each other, separated by a
median cleft of a width equal to either one of them, and which becomes pointed
behind.

Their anterior ends do not merge into the premaxillary beyond until they are
well past the points where the maxillaries are inserted. These anterior portions are
thin, horizontal plates, being directly continuous with the horizontal and fused
palatine bodies behind. This latter portion shows a small median carination just
in front of the united heads, and the postero-external angles are rather sharp, being
pointed directly backward.

Anteriorly, the pterygoidal heads meet each other and the fused palatines, the
three forming a groove on their upper sides for the rostrum. At their outer ends
each pterygoid offers a shallow cup to form the usual articulation with the quadrate
of the corresponding side.

Professor Parker found that " in Sida alba the basitemporals are as little devel-
oped as in the Dromseidw, less than in any other carinate bird. Behind each moiety
there is a large oval opening, not far in front of the occipital condyle; this exposes
the loose diplose within. The small Eustachian tubes open at a little distance from
each other, in a wide, shallow fossa, on the part where the three elements of the
parasphenoid meet." The description of these details agrees with the skull of the
specimen before me. Professor Parker, however, was fortunate in having the skele-
ton of the ear-parts in his specimen, and, of them, he says that "in Sula alba, the

130 . MEMOIRS OF THE CARNEGIE MUSEUM

columella auris is very long and bent. It has a small, cartilaginous, extra-supra-
stapedial process, and a long, attenuated stylohyal."

On either side, the entrance to the middle ear in Sula bassana, as in others of
the same genus, is shallow, and it is situated quite internal to the quadrate bone,
while immediately mesiad to it there is a pit of great depth, with its aperture look-
ing downwards, and its base in the vault of the cranium, which seems designed for
muscular lodgment; the positions of the usual foramina about it are peculiar, and
extremely interesting in these birds.

The bony wings that shield the entrance to the ears are large and tilted up
behind. Each one shows the double facet for the mastoidal head of quadrate, the
outer one having its inner margin encroached upon by the pit described above.

The postero-internal angle of either of these wings is connected with the side of
the elevated basi-temporal region by a bony bar. This condition can best be seen
from a posterior view. When speaking of the orbital cavity I neglected to mention
that the upper part of the septum is longitudinally marked, as in most birds, by an
open, single groove for the passage of the olfactory nerve to the rhinal space beyond.
The exit for it from the brain-case is very small, indeed, and occasionally, in S.
bassana, on one side, the bone spreads over it, rendering the nerve track, for a frac-
tion of the initial part of its course, tubular.

The brain-box itself is capacious and notable for its great width over its com-
pression in the vertical direction. Its anterior wall looks directly downward and
forward, making an angle of about 45 degrees with the horizontal palatine bodies.
Seen from behind, the skull shows, above, the extent to which the crotaphyte fossa?
approach each other in the median line and the crest that divides them from the
occipital area. This latter has the usual form seen among these cormorant-like
birds, constituting an arch over the foramen magnum, which occupies the center of
a concavity below it. The supra-occipital prominence is here distinguished by a
low, smooth median ridge, which traverses this dome-like elevation from the inter-
crotaphyte line to the superior periphery of the foramen magnum.

The plane of this latter aperture is about perpendicular to the plane of the basis
cranii. In outline the foramen is broadly elliptical, with the short axis transverse.
At its lower margin .we see a large ellipsoidal condyle, with its short axis at right
angles with that of the foramen. Below this again are the oval openings in the
basi-temporal, spoken of by Parker, with the prominent descending processes of this
region flanking them on either side.

In form, the inferior mandible is spear-shaped, its sides tapering gradually to a
sharpened apex. These latter, for the outer aspects of their anterior two thirds,

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 137

show the same character of venated surface as I described for the superior mandible.
Posterior to this, however, as well as the inner ramal aspect, the bone is smooth,
having the same appearance as in most birds.

The symphysis is short and develops a spine behind, which points directly back-
ward and is in every respect similar to the process in the same place, between the
sides of the lower jaw, in Herons and Albatrosses. Each ramus of this mandible
is very thick from side to side, but these parts are hollow, and the bone as a whole,
is very light, owing to the high state of pneumaticity it enjoys.

The principal foramina for the entrance of air to its* interior are four in number,
two on either limb, one being to the mesial side of the articular cup, and another
larger, longitudinally piaced, elliptical one just beyond this concavity on the inner
aspect of the ramus near its upper border. The superior side of an articular end
has a deep excavation at its center upon which the facets for the quadrate do not
encroach, so that, when the jaw is articulated, this pit comes opposite the notch
between the trochlese of the mandibular foot of the quadrate, creating an irregular
hollow space there between the bones of no inconsiderable size. When the quad-
rate thus covers it there are two entrances that are left open, one in front and one
behind, close to the pneumatic foramen.

The mandibular angles are truncate, very nearly perpendicularly so, their sur-
faces being concave and very broadly luniform in outline (Plate XXII., Fig. 7).

Commencing just in front of an articular cup, we find the superior border of
the ramus to be rather wide and rounded as far as the meeting with the dentary
This portion presents near its middle a double coronoid process, one being in front
of the other. The dentary portion of their border has an outer cultrate edge and
an inner and somewhat lower rounded one.

The outer edge goes to the anterior apex of the symphysis, the inner one to the
hinder termination of the same, while between the two a nearly horizontal surface
is contained, which gradually becomes narrower as we proceed in the forward
direction.

The lower borders of the mandible are rounded for their entire extent, being
produced beneath the- articular cups and continuous with the inner boundary of
either truncate angular extremity.

We find that the usual bones which surround the true ramal vacuity on the
side of the mandible in many birds, here interlock with each other so as to com-
pletely fill the fenestra in, but in rather an unusual way, and apparently, for a
definite purpose; for each ramus presents, both on its inner and outer side, an
oblique slit, these slits being opposite each other and with their anterior ends in

138 MEMOIRS OF THE CARNEGIE MUSEUM

the superior border. It is evident that this otherwise thick jaw is much weakened
at these points in each ramus, and this occurs just posterior to the hinder termina-
tion of the horny sheath of the lower beak. In other words, the hinder moieties of
the mandible are attached to the anterior or dentary portion by thin plates of bone,
consisting principally of the splenial elements, and are capable of being bent out-
ward, which, in the recent specimen can, owing to the way the quadrates are
attached, be effected to a considerable degree. Now in life these oblique slits have
their anterior ends come opposite the thin anterior insertions of the maxillaries,
and these latter are just beneath the very mobile cranio-facial hinge, so that the
whole apparatus is admirably arranged to permit an increase in size of the fore part
of the buccal cavity when a Gannet swallows the fish that constitutes its food, and
which its beak is so well fitted otherwise to capture. Moreover, this possible
increase in caliber takes place in that portion of the digestive tract where it is most
needed, or where the bony walls of the mouth would prevent the admission of a
very large morsel, unless some such mechanism existed, — that is, at the very
entrance of the buccal cavity, and just posterior to the more horny thecse of the
beak. In Gannets, however, this mobility is, to some extent, restricted by the
integumental sheath of the beak.

Sulidse have a wonderfully pneumatic skeleton, the entire structure enjoying
that condition, save the ribs; all the caudal vertebra (with the occasional exception
of the anterior one)1; and, finally, none of the bones of the pelvic limb are pneu-
matic below the femur. The entire skeleton of the wing is perfectly so.

In a specimen of Sula gossi before me, nine or ten large sderotal plates ossify in
either eyeball, and these overlap each other much as we find they do in other
birds. All the tracheal rings ossify quite perfectly, as do the bronchial semirings as
they approach the lungs. These rings look like little, narrow, bony, double-over-
lapped straps, as they are arranged in situ to form the somewhat antero-posteriorly
compressed windpipe. The usual bones of the superior larynx also ossify, and in the
syringeal portion of the tube we find strong osseous arcades overarching the com-
mencement of either bronchial tube, and parallel to them below, in the median
line, the dividing pessulus is also in bone.

Anatomists have long known that the hyoidean apparatus as found in the Sulidse
is invariably a very much aborted affair. The only parts of it that ossify are, first,
a little irregular piece which represents the first basibranchial ; and second, articu-
lating with this behind, are two simple curved rods of bone, which are the cerato-
branchials of the thyro-hyals. Beyond these, no part of the tongue of a Gannet
ever appears to ossify. It is hardly necessary to say that such parts of the skeleton

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 139

as the sclerotal plates, the tongue, or rings of the trachea, or much less certain
sesamoids we will hereafter he called upon to describe, are ever pneumatic.

Of the Remainder of the Skeleton of the Irunk in the Sulidse. (See Plate XXIII.,
Fig. 12.)

In the common Gannet there are twenty-one free vertebne in the spinal column
before we meet the one that first anchyloses to form, with the assistance of the four-
teen succeeding ones, a sacrum for the pelvic bones. Then follow seven more free
ones, devoted to the movable part of the tail. Finally, we have a long pygostyle
that probably contains at least six more. Owing to the lengthening behind of its
pelvic bones, the sacrum contains sixteen vertebrse in SuJa cyanops, and that species
has but six free caudal ones plus the pygostyle.

The sixteenth, seventeenth and eighteenth vertebrse support each a pair of free
ribs ; the next three belong to the dorsal series, and all have true vertebral ribs
articulating with costal ribs from the sternum. This is also the case with the first
three pairs that spring from the pelvic sacrum. The ribs on the sixteenth vertebra
are exceedingly rudimentary, and the last pair of " floating " costal ribs have no cor-
responding pair of vertebral ones above them.

In mid-series these ribs support epipleural appendages, attached in the usual way
to their posterior borders. (They may be anchylosed or free.) As I have already
stated above, they are completely non-pneumatic.

The neural canal is notable for being nearly cylindrical throughout the first
twenty-one vertebrae. It is only at the region of the enlargement for the brachial
plexus that it is rather compressed in the vertical direction.

The atlas has a minute perforation in its cup, and its neural arch is strikingly
broad and deep. The axis vertebra possesses a stumpy neural spine, and its hypa-
pophysis, directed somewhat backward, is very prominent.

The odontoid peg is comparatively small and nearly sessile with the centrum,
the latter presenting a concave face below it.

From the third to the fourteenth vertebra, inclusive, the neural spine is a very
inconspicuous character, while from this on it gradually makes its appearance, increas-
ing in size until we have the usual quadrate, longitudinal plate of the dorsal series.

The third and fourth vertebrse have each a prominent hypapophysis like the
one on the axis, but in the fifth this feature nearly entirely disappears.

The sixth vertebra is faintly marked by the carotid canal ; this gradually be-
comes more and more tubular in the seventh and eighth, while in the ninth to the
thirteenth inclusive it is a closed cylindrical canal of a caliber somewhat less than the
neural canal canal above it. It disappears entirely from the fourteenth vertebra.

140 MEMOIRS OP THE CARNEGIE MUSEUM

The lateral canals extend from the third vertebra to the fifteenth, inclusive ; they
are short in any of the segments, and their posterior apertures are far larger than
their anterior ones.

At the commencement of the cervical series the parial parapophyses are short and
not particularly well developed. They project backward from the inferior walls of
the lateral canals, but as the carotid canal begins to develop, these processes with-
draw from the former positions, move gradually lower down beneath the centrum,
and at the same time increase in length and importance, so that in those vertebrae
where the carotid canal exists, they project from its postero-inferior border directly
backward as parallel and not far separated spines.

The post-zygapophyses do not appear as divergent limbs until we find them so in
the eighth vertebra; in all the cervical segments anterior to this one the facets are
situated on the inferior aspect of the tuberous hinder end of the neural arch at its
lateral angles.

Metapophyses are seen on the ninth vertebra, but gradually disappear, to be
entirely absent in the fourteenth or fifteenth.

The transverse processes in the dorsal region are broad, flat, and horizontal, being
directed more and more to the rear as we approach the pelvis. The plates of the
neural spines above do not meet each other when the column is articulated, and
there is an entire absence of all interlacing, ossified tendons or metapophyses in this
region. In fact, all the vertebra? have a very clean-cut, non-angular appearance,
with the majority of projecting borders rounded.

The articular ends of the centra are constructed upon the " heterocselous " type;
the anterior faces in the ultimate cervicals and leading dorsals being notably wide
and shallow, and often riddled with foramina.

The pygostyle and the free caudal vertebrae will be spoken of after the pelvis has
been described ; in the meantime we will turn our attention for a few moments to
the description of the sternum and pectoral arch.

The Sternum. — This bone in the Gannets has a very peculiar form. A pectoral
aspect of the bone shows that the body has an oblong figure or outline, with the
average width nearly equal to half the length. Beyond the true sternal body the
anterior portion projects as a massive promontory, and a large part of the carina is
beyond this again.

The anterior moiety of the bone is convex upon the dorsal side, and correspond-
ingly concave on the ventral aspect. Behind, the body is so flattened out in S. bas-
sana as to be nearly horizontal. The costal borders look outward and slightly up-
ward, and each usually possesses six moderately well-developed facets for the costal

SHUFELDT: OSTEOLOGY OP THE. STEGANOPODES 141

ribs. In a specimen of 8. bassana, however (No.!l8045 U. S. Nat. Mus.), and in one
of S. piscator (No. 18739, U. S. Nat. Mus.), I find but five of these facets upon either
costal border. There are no pneumatic foramina in the elongated and shallow
intervals between them.

The principal orifices of this character consist in a diffuse group on the superior
aspect of the anterior projecting part, within the general concavity of the bone.

Either costal process gracefully rises from its base as a laminated and prominent
horn, curving in the anterior direction.

The posterior moieties of the lateral borders are somewhat rounded and extend
almost directly backward over the lateral processes behind.

These postero-external xiphoidal processes are very long and wide, being rounded
off at their extremities and directed a little outward. They are narrowest in S.
brewsteri, more flaring in S. gossi, and very wide in S. cyanops.

They are created by this hinder portion of the bone being so profoundly one-
notched that a general concave mai'gin has resulted, with simply a median papilli-
form process remaining. Even this latter is frequently altogether absent in S. gossi,
converting the posterior sternal border in this species into one long, well-marked
concavity.

The carina juts out very prominently in front of the bone ; its anterior angle is
concave from above downward, and develops a large facet for the furcula, which in
life articulates with it. Above this the border is again concave and sharp, while
still above this there is a compressed process that represents the manubrium. This
is the case in S. bassana only, for in the sternum of no other Gannet before us is
there the slightest semblance of such a process as the manubrium.

The lower border of the keel is straight and in the horizontal plane, being capped
off with a spreading rim. This border merges into the surface of the body of the
bone before it halfway reaches the xiphoidal notch.

The sides of the keel are smooth, and neither it nor the under side of the sternal
body show in this specimen any of the muscular lines usually present in most birds.

A broad median notch, concave from side to side, convex from before backward,
lies between the lofty superior portions of the coracoidal grooves. These latter meet
in front of it, while behind, its surface becomes directly continuous with the general
surface of the upper side of the body, and that where the group of pneumatic foram-
ina are found.

A coracoidal groove looks forward and outward for its upper portion, directly
upward for its lower, and extends rather less than half way between the base of the
costal process and the median line. It consists of two portions which are directly

142 MEMOIRS OF THE CARNEGIE MUSEUM

continuous with each other. The lower one is a shelf-like projection with a convex
border forward and its articular surface in the horizontal plane. Immediately
above this rises a much broader surface, though not so long, which is decidedly con-
vex from above downward. This portion of the facet for the coracoid is consider-
ably higher than the plane in which the borders of the body of the bone are found.
It faces forward and outward, and has one regular convexity as its limiting margin
above. Between the point of its outer termination and the apex of the correspond-
ing costal process, the border is one sweeping concavity.

This form of sternum is more or less peculiar to the Sulidx, and it departs in a
number of points from the form of the sternum of the Cormorants and of the Peli-
cans. Comparatively, the bone is not so long in Sula gossi as it is in S. bassana, S.
breivsieri and other species. Still the general pattern of the sternum is much the same
throughout the Sulidse. If we overlook the crossing of the coracoidal grooves in
the sternum of Phaethon xihereus, and its having a manubrium, there is a great
deal in the bone to remind us of the sternum of Snla gossi — and, in fact, there are
more steganopodous characters in the sternum of the Tropic Bird than there are of
any other avian group with which I am at present acquainted.

Of the Shoulder-girdle. — (Plate XXIIL, Fig. 12.) This part of the skeleton is,
like so much of the rest of it, thoroughly pneumatic, the foramina occurring at
their usual sites.

The clavicles form a broad U-shaped arch, being completely united below, Avhere,
at the median point beneath, they support an extensive facet for articulation with
the carinal angle of the sternum. This does away with any such thing as a hypo-
cleidium proper, notwithstanding the fact that the bone projects slightly over this facet.

The clavicular limbs are compressed from side to side, broader above than below,
with their anterior and posterior borders rounded off.

A clavicular head is also compressed in the same manner as its shaft, and
tapers off as a pointed process.

The most striking feature of this part of the bone is, however, the extraordi-
nary facet it supports to articulate with the coracoid. Either one of these is situated
at its outer side, upon a promontory of bone which is found there, the latter being
of a proper form to receive it. The facet is of an elliptical outline, placed vertically,
and facing directly backward. Something of a notch is found between it and the
clavicular head, in which occurs a number of the principal pneumatic foramina of
the furcula. On the anterior surface, just below the summit of a coracoid, we find
a distinct elliptical facet for articulation with a similar one just described for the
fourchette. Between this and the ear-shaped glenoid facet a considerable valley

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 143

is found. On the opposite side of the coracoidal head we find a group of pneumatic
foramina, and below these a peculiarly formed scapular process, a spine-like apo-
physis, which rather gracefully curls upward and then toward the shaft of the bone.
This latter portion of the bone is subcylindrical and smooth, dilating below into
a transverse fan-shaped sternal extremity.

A scapula offers but a very small portion of the articular surface for the glenoid
cavity ; not more than an eighth of it in all the specimens examined by me. The
head of the bone then reaches forward and inward, but only the outer two thirds
of this makes an indifferent articulation with the narrow and roughened border of
the scapular process of the coracoid. The shaft of the bone is quite stout behind
this and somewhat compressed in the vertical direction, while posteriorly it flat-
tens out into a broad paddle-shaped extremity that finally tapers to a point
behind.

Even more than it is in the case of the sternum, the bones of the shoulder-girdle
in all the species of the Sulidas at my hand are, apart from their specific variance in
size, almost identically alike in each and every one of their corresponding characters.
As for the bones of the shoulder-girdle in the PhaWionidse, we may say here, by way
of comparison, that it is only in the scapula where we see the characters which more
or less resemble the corresponding ones as we find them in the scapula? of the Sulidx.
Neither the os furcula nor a coracoid of a gannet or booby bear any special resem-
blance to those bones as they exist in the tropic birds.

Of the Pelvis and Caudal Vertebrse. — The first vertebra that anchyloses with the pelvic
sacrum anteriorly projects entirely beyond the iliac bones. Its centrum, in common
with the next three that follow it, is much compressed from side to side, and its neural
spine is continuous with the common neural ridge above the succeeding segments.

The first five vertebne that lie beneath the ilia throw out their apophyses in the
usual way for their support ; the last two of this series meet the iliac margins.
Here the neural canal and centra are large, in order to afford room for the increase
in size of the cord where the sacral plexus is thrown off.

The twenty-eighth and twenty-ninth vertebrae have their processes thrown directly
upward, so that they are scarcely visible upon direct ventral aspect.

In the thirtieth vertebra they are powerfully developed and extend directly
across the basin to abut by anchylosis against the pelvic walls immediately behind
the cotyloid cavity on either side. From this point the centra of the so-called uro-
sacral vertebras taper quite rapidly in size to the ultimate one, which, in S. bassana,
is enlarged and exhibits a big facet on its posterior aspect, intended for the first free
caudal. In S. piscator this enlargement is not evident.

144 MEMOIRS OF THE CARNEGIE MUSEUM

The extremities of their diapophyses anchylose in a very thorough manner with
the inner iliac margins, and a lateral view shows their sides to be riddled with pneu-
matic foramina between these processes.

Viewing this pelvis from above, we notice that the entire inner margins of the
iliac bones have merged into and completely anchylosed with the sacrum. This
converts the ilio-neural grooves into ilio-neural canals and gives the bone a very
compact appearance. The anterior margins of the ilia are, rounded, and are set off
with rather a deep and raised emargination in S. bassana, which is feebly marked in
S. piscator and other species.

The post- and pre-acetabular surfaces are about equal in the extent of their su-
perficial areas, except in Sula cyanops, where the postacetabular area is the more
extensive.

The anterior iliac surfaces are concave on either side, and each faces upward
and outward to about an equal degree. Sula piscator here offers another exception,
and in this species the pelvis is comparatively shorter, as well as broader and flatter
than it is in other species of Gannets.

Elevated above these anterior iliac concavities we find the postacetabular area to
be nearly horizontal. Large elliptical foramina are found between the apophyses
of the last three or four uro-sacrals, and these latter, likewise, develop quite a
prominent neural crest.

Upon the lateral aspect of this pelvis we find a very large cotyloid ring, the
inner margin of which is fully equal in size to the outer. A moderately sized anti-
trochanter occupies its usual site, with its articular surface directed downward,
forward and outward.

Behind this occurs an enormous elliptical ischiadic foramen, that occupies nearly
all of this post-ace tabular lateral aspect. Through the fenestra thus formed we are
enabled to get a good lateral view of the uro-sacral vertebra? and the extensive
pneumatic condition they enjoy.

The lower margin of the ilium is sharp and convex ; it forms the superior boun-
dary to a long, narrow, obturator space, which opens freely into the rather small
obturator foramen.

A pro-pubis does not develop in the Gannets, while the post-pubis is, for the most
of its extent, fragile and slender. It begins to increase in size just before arriving
at a point opposite the end of the ischium. At this point it offers a small facet on
its upper margin for the ischiadic postero-inferior angle, and the two bones are in
contact here during life. The post-pubis, retaining its increase in size, then curves
inward toward the fellow of the opposite side, to terminate in a cartilaginous tip.

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 145

The posterior border of this lateral aspect shows a well-marked ilio-ischiadic notch
at about the middle of its extent. This character is best marked in S. cyanops and
S. gossi, being much shallower in other species, especially in S. piscator. The outer
side of the bone between it and the ischiadic foramen is directed upward as well as
outward.

As has already been mentioned above, there are usually seven free vertebrae plus a
large pygostyle in the skeleton of the tail of most Gannets and Boobies, but an
exception is noted in the case of Sula cyanops, in which species I count in the
specimen before me, only six and the pygostyle. The characters of these vertebras
may be well studied in the skeleton of the tail in Sula brewsteri. What first strikes
one upon glancing at this part of the osseous system of any Gannet is its compara-
tive massiveness, the great size of the individual vertebrae, and the large pygostyle.
The neural spines however, are short and stumpy; in Sula bassana they are occasion-
ally bifid anteriorly. The neural arches beneath them close over the spinal canal for
the entire length of the series, and the latter, for a short distance, is seen to perforate
the pygostyle. The transverse processes are thick and strong, being very wide-spread-
ing, especially in the case of the ultimate and penultimate free ones. From four to
five of the last ones usually develop haemal spines. These become larger as we
approach the pygostyle, the first vertebra of which also has one, it being bifid,
and hooking forward, as do those on the caudals in front of it. The interarticular
facet on the faces of the centra are nearly flat, showing barely any concavity or
convexity.

The pygostyle appears to be composed of about six vertebrae, of which the three
anterior ones can be quite easily made out. It has a very unusual form in this bird,
being very long and subconical, with sharp superior border and rather decurved
apex. Below, it is broad and somewhat convex. Viewing it from in front we notice
that it has all the elements present, though in very rudimentary state, of one of the
caudal vertebra), including the large, prominent and anchylosed haemal spine just
mentioned.

TJie Appcndicular Skeleton.

The Pectoral Limb. — We find the humerus in Sula bassana to be somewhat longer
than the radius and ulna in this limb, but we shall see later on that this varies
greatly in the other species. I will write out here first an account of the appendic-
ular skeleton of Sula bassana, and close with the differences exemplified on the
part of the other Gannets and Boobies we have under consideration. In this
humerus the ulnar crest is prominent and projecting, though rather inclined to retreat
from the elongated and shallow pneumatic fossa that arches over it, as in many other

146 MEMOIRS OF THE CARNEGIE MUSEUM

water birds. The radial crest is reduced to a long, low, inconspicuous ridge, and,
in fact, this proximal end of the humerus, as a whole, merges into the shaft so
gradually from both sides, and its being so narrow withal, that we are rather im-
pressed with its lack of strength and an absence of a certain robustness so characteristic
of the bone in other birds of equal size and that lead a similar life. This in no way
applies, however, to the shaft itself, for this subcylindrical and hollow, bony tube, with
its double sigmoidal curve, carries with it the very elements of strength and power.

Its distal extremity lacks but little of being as wide as the widest part of the
head of the bone. It is without an ecto-condyloid process, has the trochlese very
prominent, and presents for examination a deep fossa to the anconal side of the
ulnar tubercle.

The shaft of radius for so long a one is unusually straight, and only a slight
curve is noticed in the proximal moiety of the ulna.

In its continuity the former bone is subtrihedral in its form, with its pneumatic
foramina situated beneath the transversely expanded portion of the distal end.
Muscular lines mark this radial shaft along its inferior aspect.

For its distal moiety the shaft of ulna is nearly cylindrical in form, but this is
gradually exchanged for the subtrihedral as we pass over the proximal half of the bone.

It presents for examination a double row of feebly marked papillae for the quill-
butts of the secondary feathers.

A long, shallow, though notable fossa is seen at the proximal and anconal side
of the shaft, which terminates in a single pneumatic foramen just beyond the promi-
nent cup-shaped articulation for the ulnar tubercle of the humerus.

Other pneumatic holes occur at the distal end of the ulna upon all sides, except
the outer one. The olecranon, though large and rather tuberous, would not particu-
larly attract our attention. A distinct canal upon the outer aspect of the distal
end of the shaft for the passage of the tendons characterizes this bone. The artic-
ular surface shows nothing of special interest.

As usual, the carpal segments are but two in number — a radiale and an ulnare.
They present the forms and facets common to these bones generally. Both are
pneumatic and have large apertures for the admission of air to their hollow interiors.

The carpo-metacarpus also presents a number of those foramina at either of its
extremities; the principal one, however, is found just below the trochlear surface
formed by os magnum upon the anconal side of the bone. A notable process occurs
immediately below it, as well as another group of these air-holes, in its outer as-
pect, near the short and inconspicuous first metacarpal.

The main shaft is straight and of good caliber ; on the palmar side it is longitudi-

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 147

nally grooved nearly its entire length for a tendon going to the fingers. This is best
marked upon the distal moiety of the bone. The metacarpal of the middle digit is
also straight for the major extent of its course; its extremities becoming enlarged in
order to allow it to make the usual connections with the metacarpal index. It is
rather slender and develops no special processes, as it sometimes does in other rep-
resentatives of the class.

The expanded portion of the proximal joint of the index digit is not perforated,
not even by the numerous pneumatic foramina which are irregularly scattered
over its surface. Below it is produced us a notable process, and a process that is
seen in some of the extinct birds, as, for instance, in Ickthyornis. The shaft of
this phalanx is broad and fiat anteriorly, and perfectly straight from above down-
ward.

Equal to half the length of the carpo-metacarpus, the distal phalanx of the index
digit is of a trihedral form, with an extensive excavation at the posterior aspect
of its proximal end, which is continued in a lesser degree the entire length of the
bone. It bears no claw below, but is finished off by a distinct little process.

The pollex phalanx has very much the same form as the one just described, but
it lacks the longitudinal excavation down its posterior aspect. Both of the bones are
pneumatic. Lastly, we have the smallest phalanx of all, belonging to the middle
finger. This, as usual, is behind the broad proximal joint of the index, and not
quite equal to half its hinder border in length.

Now the general characters of the bones of the pectoral limb are, in Stda cyanops,
S. piscator, S. gossi and S. brewsteri, in the main about the same as they have just
briefly been given for Sula bassana. We find differences, to be sure, but they are
very slight. The humerus in S. cyanops closely resembles that bone in S. bassana,
only it is considerably smaller in the first-named species. Occasionally about the
proximal end of the carpo-metacarpus the position and size of the pneumatic open-
ings may vary, but that is often seen to be the case in all large birds with highly
pneumatic skeletons.

In S. piscator the olecranon fossa of the humerus is comparatively larger and
with better defined borders than it has in S. bassana. This is likewise the case in
Sula gossi and S. brewsteri, in both of which species that fossa is especially well
marked. Beyond such trivial departures as these, we meet with little or nothing
worthy of formal record. Individual specimens of the same species, however, vary
a little ; take, for example, the carpo-metacarpus. It is seen to be rather smaller
than that of another specimen of the same species of S. bassana now at my hand for
comparison.

148

MEMOIRS OF THE CARNEGIE MUSEUM

Where remarkable differences do come in, is in the relative lengths of these bones
when we come to compare them in one species with the corresponding bone in an-
other. In speaking of Sula bassana, I said above that its humerus was longer than
the radius or ulna ; now this is not the case in other species of Gannets, as the sub-
joined table will plainly show.

MEASUREMENTS IN CENTIMETERS AND FRACTIONS.

Species.

Humerus.

Ulna.

Carpo-metacarpus.

Joints of Index
Digit.

Sula bassana.

23.5

20.4

9.6

9.6

Sula cyanops.

19.3

20.7

8.8

8.1

Sula piscator.

17.1

18.4

7.2

6.9

Sula gossi.

16.2

17.3

7.6

6.9

Sula brewisteri.

18.4

19.8

8.1

7.6

Of the Pelvic Limb. — In comparison with the general size of Sula bassana the
lower extremity is very short/though the bones composing its skeleton are none the
less strong in consequence. In the femur we find the axis of the head and neck
making an angle with the longitudinal axis of the shaft. The head is quite dis-
tinct, globular, and, as usual, excavated on top. Its surface is continual with the
broad articular surface which occupies the entire summit of the bone. ISTo trochan-
terian ridge rises above this latter, and, indeed, this character of the femur is but
poorly developed.

A pneumatic foramen is always seen at its most common site, on the anterior
aspect, just below the superior articular surface.

The shaft is cylindrical, roughened in some places by lines and diffuse tuberosi-
ties for muscular attachment, and is bent slightly to the front and somewhat to the
inner side. At its distal extremity the condyles are fashioned after the usual pattern
among birds, but all their characters in Sula present a sort of lack of strong devel-
opment. The fibular cleft is but faintly marked, the intercondyloid notch or fossa
is shallow, and the ridges in front much rounded and inconspicuous.

Something of the same condition is extended to the proximal end of the tibio-
tarsus of the leg, though not to such a marked degree. Here the cnemial process rises
but slightly above the articular summit of the bone, and the pro- and ecto-cnemial
ridges which descend below it soon merge into the shaft, and are, at the best not
very prominently developed.

The shaft of this bone is straight and smooth and somewhat compressed through-
out from before backward. It offers a long ridge to the fibula and is broad across
where it is found. The distal extremity of the bone evinces more character than

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 149

the proximal one. An oblique bridge, to confine the extensor tendons, is extended
across the deep groove that contains them during life. Nearly parallel with each
other, the condyles are wide apart, prominent and convex in front, to become sup-
pressed and low and thin-crested behind.

The fibula has the usual form seen in birds, but is here particularly interesting
from the fact that it does not anchylose with the shaft of the leg-bone until it arrives
at the middle of its lower third, and even from this low point the remainder of the
bone, including an oval "external malleolus," stands out quite prominently. This
rare condition of things has been pointed out also by me as occurring in Uri-
nator lumme.

Sala bassana has a long oval patella, obliquely marked across its anterior surface
by a groove for the tendon of the ambiens muscle.

The tarso-metatarsus in Sula is strikingly large in its proportions when compared
with the other bones of the limb. In length it is a little more than half as long
as the tibiotarsus, but being wider and broader it appears much more massive. Its
hypo-tarsus presents three short, longitudinal elevations of unequal sizes. These
inclose two tubular passages for tendons, and are grooved themselves besides. In
other specimens they are flat, and the two outer elevations may posteriorly meet,
thus creating a vertical perforation rather than a groove. The back of the shaft is
flat, but in front it is much scooped out above, where it shows two antero-posterior
perforations.

At the distal extremity three large trochlear projections present themselves.
They are separated from one another by wide clefts of about an equal depth. These
trochleje are placed nearly side by side, the middle one being the lowest down, the
inner next, and the outer one the most elevated. Their median grooves are best
marked behind, but in addition the internal trochlea presents a deep, vertical notch
upon its outer aspect.

The usual arterial perforation pierces the bone above the cleft found between the
outer and middle projections, a groove leading into it from above.

The accessory metatarsal is rather an elongated bone, swung in the usual way
by ligaments to the lower part of the shaft.

The basal joint of the hallux, which it supports, is comparatively more slender
for its length than the other joints of the foot.

For the three anterior toes these latter are, in number and arrangement, the same
as in the vast majority of the class. They present all the characters usually
attributed to the phalanges of the podal digits in birds, and are well proportioned,
both as regards their relative calibers and lengths.

150

MEMOIRS OF THE CARNEGIE MUSEUM

The pneumatic foramen in the femur of S. cyanops is usually very large, and
comparatively small in my specimen of S. piscator. This also applies to S. brewsteri,
but in Sula gossi it is notably very large again. The general characters of the femur
are much alike in all typical Gannets.

The ti bio-tarsus and fibula in Sula cyanops agree with those bones in S. bassana,
as they practically do in other species. There is a difference seen in the degree of
distinctness of the distal end of the fibula. The entire bone can easily be made out
in all my specimens of S. gossi, and only, or less than, the distal fourth of it fuses
with the tibio-tarsus in that species. The tibio-tarsal condyles are far apart in S.
piscator and S. gossi, and in most birds of this family this character is more evident
than it is in S. bassana. In this last-named species the tibio-tarsus and tarso-meta-
tarsus are all non-pneumatic ; I fail to find any pneumatic openings in the tibio-
tarsus of 8. cyanops, but the tarso-meta tarsus of that species enjoys that condition
to a very high degree, as numerous and large air-holes are found about the head
of the bone. Another remarkable fact is that the tarso-metatarsus in Sula cyanops
is rather larger than that bone in Sula bassana. It is very small in S. piscator, and
in all considerable variation may exist in the hypotarsus, for the places where the
tendons pass through may be either grooves or perforations. Whatever they are,
however, they never exceed two in number. Below I give a table showing the com-
parative lengths of these bones in the species under consideration.

MEASUREMENTS IN CENTIMETERS AND FRACTIONS.

Species.

Femur.

Tibia.

Tarso-inetatarsus.

Middle Anterior
Toe.

Sula bassana.

7.4

10.9

6.1

Sula cyanops.
Sula piscator.
Sula gossi.

6.4
4.9
5.4

9.9
6.9
7,9

6.5
3.8
4.9

7.9

Sula brewxteri.

5.9

9.1

5.7

8.3

From this study it is hardly necessary to add that the skeleton of the pectoral
and pelvic limbs of the SulidiK is entirely different from the corresponding parts and
bones in the Pha'ethonidte.

ANHINGID.E.

Osteology of Anhinga anhinga.

As will be seen by my list of material given above, I have, at the present writ-
ing, but a single skeleton of this family of birds, before me for examination. It is
complete, however, and there are at hand the accounts of the osteology of Plotus
by other authors. Doctor R. Bowdler Sharpe in his recent Hand List of Birds

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES

151

(1899) retains all the Anhingas in the family Plotidx, and, in addition to two fossil
forms, recognizes our existing species, viz : P. rufus, P. melanogaMer, P. novie-
hollandise, and P. anhinga. The Arihinga anhinga is the sole representative of this
family in the United States. A peculiarity of the skeleton of Anhinga is, it is
almost completely non-pneumatic. It is only into the base of the cranium and the
articular ends of the mandible that air gains access through minute foramina occu-
pying the usual sites.

Of the Skull. — This bird has a very perfect " cranio-facial hinge," and measuring
each way from the center of it we find the superior mandible to be about 1.5 cm.
longer than the cranium. In form the superior mandible is long, narrow and spear-
shaped, being drawn out to a sharp point. It is nearly straight. Beneath, it is flat
with cultrate tomia. The culmen is rounded off, and the nasals have so fused with
the surrounding bones of the face that most of the sutural lines cannot be dis-
tinguished in the adult. Where the external nostrils would naturally occur, there
are usually present only minute holes that do not appear to reach in so far as the
rhinal chamber. Viewed from above, we are to observe that the cranial and frontal

Flo. 3. Left lateral view of the skull, including mandible, of a specimen of Anhinga anhingn. Natural size, in
outline, by the author from specimen 18259 of the Coll. U. S. Nat. Museum, sr.m, the supramaxillary bone.
FIG. 4. The same skull, seen from above. Mandible removed.

region are somewhat narrow and elongated ; the former being smooth and moder-
ately rounded, the latter flat and measuring about half a centimeter between the
sharp edges of the superior peripheries of the orbits. For its hinder half, longi-
tudinally, this part of the skull presents a low, median prominence, created by the
approach upon either_side, of the extensive, though shallow, crotaphyte fossa. We

152 MEMOIRS OF THE CARNEGIE MUSEUM

may also see, upon this aspect, the conspicuous exoccipital processes. They project
almost directly backwards, and only slightly outwards. Posteriorly, the superior
occipital area is seen to be bounded above by a sharp and raised line, much curved,
which distinctly separates the occipital from the parietal region of the skull. The
foramen magnum is very large, and subcircular in outline, its mid-vertical diameter
being scarcely shorter than its transverse one. The unnotched condyle is of good
size, hemispherical in form, and sessile.

Seen upon lateral aspect, we are struck with the open, shallow auditory fossa,
which is overhung by a raised squamosal ridge, the squamosal process being small.
Anterior to this region the lateral wall of the brain case is extensive, smooth and
convex, otherwise presenting no special characters. The orbits are large, being con-
tinuously circular in front and above, the curve extending from the very short post-
frontal process to the lower end of the lacrymal bone. No osseous septum divides
these cavities, and a large, median, heart-shaped foramen occupies the central part
of the anterior wall of the brain-cavity, from which, during life, the olfactory nerves
pass out. Below this another median opening exists, of no inconsiderable size. It
is the anterior common aperture of the optic nerves, and it is flanked upon either
side by a vertical, scale-like flake of bone, extending forwards. The parial grooves
'for the first pair of nerves faintly mark the orbital roof above ; each one leads to
a corresponding foramen found over either pars piano, in front. A pars plana or
ethmoidal wing is oblong in form, and taken in connection with the lacrymal, makes
a very complete osseous partition standing between the eye-socket and rhinal cavity.
The mesethmoid is small, and is earned forward for a short distance as a thin median
plate of bone, with a free anterior, thin edge. In the forepart of the nasal cavity
there are two small foramina, in the otherwise solid wall, which lead forward into
the cancellous tissue filling in the upper osseous beak. They are placed side by side.
A lacrymal is a peculiarly twisted bone, articulating above with the frontal and
nasal. It has an external transverse antero-posterior deep groove above, below which
the bone becomes, as it were, twisted and expanded, into a broad quadrilateral plate.
This stands transversely and forms, with the corresponding pars plana, the anterior
orbital wall. The lacrymal reaches down to the zygoma and is a free bone. In
front it has, articulating with its antero-lower angle, a scale-like piece, that stands
vertically and rests for its entire length upon the upper surface of the maxillary.
This is the supra-maxillary. (Fig. 3 sr.m.) Each zygoma is twisted and much
flattened. Anteriorly, their broader surfaces lie in the horizontal plane, while the
posterior moieties are in the vertical one. The peg-like, articulatory nibs at their
hinder ends for the quadrates are very small. Looking at the skull from a direct

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 153

inferior view, many of the points already described can be distinctly seen. This is
due to the peculiar flattened conformation of the cranium. Indeed, the form of the
brain-case indicates that Anhinga has an elongated, vertically compressed brain, but
relatively not as much so as it is in a Cormorant, where it is also wider. The basi-
temporal area is small and triangular, with moderately prominent paroccipital
processes, each one descending from one of its postero-external angles. The anterior
apex of the basitemporals is run out as a scale-like point, which underlaps the
meeting of the thoroughly open Eustachian grooves. The usual foramina for the
facial, glossopharyngeal, vagus and hypoglossal nerves, and the internal carotid
arteries are plainly seen in this region. They are quite distinct from each other, and,
as a rule, nearly circular in outline. The rostrum of the sphenoid is a very delicate,
rod-like bone that is extended backwards in the median line under the cranium as
far as the eustachian entrance, as a sharpened osseous ridge. In front it fuses
with the mesethmoid. For their posterior moieties, the palatines are thoroughly
coossified in the middle line, where they send down a faint crest from the sutural
juncture. As thus constituted the common bone lies in the horizontal plane, with
postero-external angles but faintly indicated, but with a short process extending
directly backwards from either side of the united facets for the pterygoids. Anteri-
orly, the palatines are much narrower, where they are also in the horizontal plane,
and mesially separated by an elongated opening that leads above into the rhinal
chamber. Their distal ends run forward to fuse with the nasals, maxillaries, and pre-
maxillaries.

The maxillo-palatines are very wide apart, and each one is simply a subvertical,
thin, plate-like scale of bone for the most part free behind, but fusing with the usual
bones in front. A shallow, medio-longitudinal groove marks the united palatines
behind, and this is occupied by the slender, presphenoidal rod. A quadrate has a
small, spine-like orbital process; a compressed mastoidal head of no great size, which
is barely divided into two facets; a large, antero-posteriorly disposed facet for the
mandible which is separated by a pronounced cleft from a much smaller one, the
entire outer side of which is occupied by the pit for the end of the zygoma. The
bone I believe to be pneumatic.

The pterygoids are flake-like, flattened bones with very sharp edges, and with
very distinct ends bearing the articular facettes for the bones with which they artic-
ulate. Each pterygoid has a length of a little more than a centimeter. A circlet
of small "sclerotal plates" are found in either eye, and the hyoidean apparatus is but
feebly developed. I have not examined the ossicles of the ear. Oonforrning to the
shape of the cranium and superior mandible, we find the lower jaw assumes the form

154 MEMOIRS OF THE CARNEGIE MUSEUM

of a long and narrow V. The ramal sides are not deep ; there is no ramal vacuity ;
and the bone gradually tapers to a sharp point in front. From the nature of the
splenio-dentary articulation — the thin bones being feebly wedged together there —
the mandible is of a consequence very weak at that point. A considerable symphysis
exists, with the barest rudiment of a median, inter-ramal spine present. For the
most part, the upper and lower borders of the bone are rounded, though the supero-
dentary edges, from sphenial articulation to apex are inclined to be cultrate. Either
coronoid process is very much reduced, while between it and the articular cup for the
quadrate, on the mesial aspect of the ramus, is a long foramen of an ellipsoidal
outline, entering the inner structure of the jaw. Each hinder end of the mandible
is, as usual, constructed to articulate with a corresponding quadrate. On the upper
side of one there is a deep, rather narrow, obliquely-placed cup — its mesial end
being the anterior one. This is for the inner articular facet of the foot of the quad-
rate. In front of this is a small, nearly flat, articular surface, and it is for the outer
facet of the quadratal foot. Facing directly forward, and standing on the postero-
internal edge of the cup is still another small facet, intended for a corresponding
one on the same bone. These articular cups develop no inturned mesial pro-
cesses, but the usual pneumatic foramen is present. The articular processes standing
out behind, are of good size, of a quadrilateral form, and so twisted that the lateral
surface in either case looks slightly upward and, to a much greater degree, outward.

Of the Axial Skeleton. — Endowed as it is with special points of interest, much has
been written upon this part of the anatomy of Anhinga. It has been touched upon
with greater or less elaboration by Brandt (Mem. de 1'Acad. Imp. Sci. de St. Peters-
bourg, torn, v., 6ime Serie, Sect, de Sc. Nat., 1839), by Mivart, in his memoir in the
T. Z. S., already cited, by Garrod (Coll. Sci. Mem., p. 334), by Donitz (Archiv fur
Anat. u. Phys., 1873, p. 357), and by Hunter (Essays and Obser., 1861, v. 11, p.
328), and by others. Most of these writers have been attracted by the peculiarities
seen in the spinal column, which I will now proceed to examine.

In the atlas we find the articular cup of the condyle perforated near its upper
border. Its neural arch is nearly as deep as it is wide, and from it projects behind,
upon either side, a conspicuous little spine. The small hyapophysis of this vertebra
also projects posteriorly beyond the centrum. From the axis to the seventh inclu-
sive, the vertebrae are especially notable for being of no great caliber in point of size,
and at the same time remarkably elongated. A low, sharp neural ridge and hyapo-
physial spine characterize the axis. Its postzygapophysial part is welded into a
common plate of bone, with the articular facets upon its under side. Upon its
ventro-posterior aspect an open channel is formed by a curling downwards of bone

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 155

from the centrum upon either side. This traverses the entire length of the verte-
bral body in the third, fourth, fifth, sixth, and seventh vertebne, but becomes less
and less marked. On the eighth it is strong again, running between the enor-
mously long parapophysial processes. In the ninth it is the subvertebral arterial
channel, and is arched over with bone anteriorly. This remains to be the case to
the thirteenth inclusive, while in the fourteenth the channel disappears behind,
though the covered arterial passage still persists on the antero-ventral aspect, beneath
the centrum. Both are absent in the fifteenth vertebra. Parial parapophysial spines
first appear in the axis, where their distal apices fuse with the bone posteriorly.
These spines, though present, are by no means a prominent character in the third to
the seventh vertebra inclusive. Suddenly in the eighth they become remarkably
developed, and are nearly as long as the bone itself. In the ninth they at once
commence to shorten again, and this rapidly continues to be the case, until they,
after somewhat changing their character, have entirely disappeared in the seven-
teenth vertebra. A most remarkable series of "lateral canals" exist in these
vertebrae. I note them first in the axis, where they are of capillary dimensions
and run nearly the entire length of that elongated bone. Anteriorly, either one
opens at the base of the odontoid process, the posterior opening being on the side
of the vertebra near the postzygapophysial base. In the third vertebra these canals
are almost imperceptibly increased in caliber, and open upon either side, anteriorly,
in a slit which is seen to exist between the prezygapophysial facet and the very
much vertically compressed facet of the centrum. Behind, it opens a couple of
millimeters in front of the articular surface of the centrum. But little change
takes place in this particular in the fourth to the seventh vertebra inclusive, except
that in the fifth, sixth and seventh a minute foraminal opening pierces the canal on
either side about the middle of the bone on its ventral aspect. In the eighth
vertebra the caliber of the lateral canals is about doubled and their posterior open-
ings are moved far forwards so as to be found on the side of the vertebra, above the
base of the enormously elongated parapophysial process ; and between two fringe-
like, long, ossified tendons that are attached to and especially characterize this
bone of the vertebral chain. With their anterior openings remaining practically
the same as described above, and the hinder ones just posterior to the parapophyses,
no change is noted in the ninth except an increase in point of the size of the tube
on either side. This increase goes on gradually to include the eighteenth vertebra,
in which bone these lateral canals last appear. In the seventeenth they are very
short, truly lateral, and somewhat compressed from above downwards ; markedly
most compressed in the eighteenth.

156 MEMOIRS OF THE CARNEGIE MUSEUM

The neural spine, feebly developed in the third vertebra, becomes a prominent
character of the fourth, where it occupies the posterior moiety of the bone, as a con-
spicuous blade-like crest, with rounded superior margin. In the fifth, sixth, and
seventh this character disappears ; in the eighth it is feebly present ; in the ninth it
is a roughened tubercle, entirely vanishing again in the tenth to the fifteenth
inclusive. In the remainder of the series, back as far as the pelvis, it is large
and of an oblong form, with thickened superior border. Throughout the dor-
sal vertebrae, ossified tendons of great length are coossified with this spine, pro-
jecting backward in the leading dorsals, and both backward and forward in the
ultimate ones.

A low, sharp, haemal spine occurs on the ninth vertebra ; it being at the anterior
part of the bone, on the parapophysial bridge that closes in the passage for the
carotids. It is still better developed on the tenth ; where sharp lateral ridges begin
to show, one upon either side of it. All these processes are very pronounced in the
thirteenth and fourteenth, while in the fifteenth their entire character is changed.
In it the lateral ridges almost entirely disappear, and an enormous, quadrate haemal
spine is thrown down from nearly the entire length of the centrum. In the sixteenth
it is not quite so large, and its hinder angle is produced backwards. A remarkable
change is seen in the seventeenth vertebra, where the ventral aspect of the centrum
is very much broadened, quite flat, and the little hook-like, laterally-compressed
hsemal spine occupies a mid-position on the posterior border. In the eighteenth,
ninteenth and twentieth the process is practically absent, and the great breadth of
the centrum gradually narrows again, as its lateral margins are deflected. This form
sees its extreme in the twenty-first vertebra, where the centrum is markedly com-
pressed from side to side, and the aforesaid lateral margins are, ventrally, converted
into a double haemal spine.

In the twenty-second and twenty-third the spine is single, and the transverse
compression of the centra is most apparent, being present in a marked degree.

Anapophysial ridges are more or less conspicuously developed in the ninth to
the fourteenth vertebrae, inclusive, and in the dorsal series proper, are long and
broad, and, as said above, are 'provided with fringe-like metapophyses frequently of
considerable length.

In the first eight cervicals, the anterior articular facets are placed laterally upon
either side of the neural canal, and their surfaces face forwards and towards the
median plane. In the ninth vertebra these facets are, as it were, rotated backwards,
so as to be above the neural canal, and face towards the median plane, and very
slightly dorsad.

SHUFKLDT : OSTEOLOGY OF THE STEGANOPODES 157

So in the articulation of the eighth vertebra with the ninth, a decided angle is
made — nearly a right angle — with its salient point to the front. This also occurs
between the ninth and the tenth and so on down the chain, becoming, however, less
and less marked, disappearing entirely after the fifteenth, after which the vertebrae
articulate in a straight line. The extremities of the long parapophysial processes of
the eighth vertebra? articulate in the shallow grooves on the forepart of the ninth
vertebra, on its ventral aspect.

Garrod has said that "Donitz figures a pair of accessory bony bridges on the
dorsal surface of the vertebra following the most lengthy one, which must evidently,
therefore, be the ninth. He, however, speaks of it as the eighth, which seems to me
to be an error depending on the omission of the consideration of the atlas, because
in Plotus anhinga (both from Brandt's figure and my specimens) it is most certainly
the ninth, as it is in Plotus novse-hollandias, Pkalacrocomx carbo, and P. lugubris. I
have, however, not seen Plotus levaillantii."

"Donitz attributed the peculiar kink of the neck of the Darters, which it is
impossible to obliterate without lacerating the surrounding muscles, to the presence
of the bony bridges he describes; in this, however, he is mistaken, it depending on
the above-mentioned peculiarity in the eighth cervical vertebra, by which it is angu-
larly articulated with the seventh and ninth vertebrae, the upper genu being posterior,
and the lower anterior. In further verification of this, it may be stated that in
P. anhinga the bony bridges do not exist, and yet the kinking is most strongly
marked." (Coll. Scientif. Mem., pp. 336, 337.) The bridges here spoken of are also
absent in my specimen of A. anhinga.

The neural tube is not very large in the first eight cervicals, where it has more
or less a cylindrical form posteriorly, but becomes somewhat antero-posteriorly com-
pressed as we gradually pass towards the fore end of the vertebra. Increasing again
in size after the ninth, it seems to attain its greatest capacity in the sixteenth, seven-
teenth and eighteenth, to become small and cylindrical once more as it passes through
the dorsal series. In the ninth and tenth vertebra? the anterior opening of the neural
canal lies in a plane which is about perpendicular, in each case, to the plane in which
the posterior opening is found. That is, upon direct dorsal view of either of these
two vertebra?, the posterior opening of the neural canal is not in sight, while we look
almost directly into the anterior opening.

On the sixteenth vertebra we find the first pair of cervical ribs ; they are long
and slender and without uniform processes. We find also a large pair of free ribs
on the seventeenth cervical vertebra, which commonly have unciform appendages.
These are anchylosed to the bone, and are large and broad. Next to these two pairs

158 MEMOIRS OF THE CARNEGIE MUSEUM

of cervical ribs, we meet with three pairs of true dorsal ones, they all having large
costal processes, and all joining with the sternum by means of their haemapophyses.
There are two pairs of pelvic ribs, the first having stunted costal processes, though
its haemapophyses join with the sternum. In the last pair the unciform appendages
are always entirely absent, and the costal ribs do not reach the sternum. Briefly
then, there are seven pairs of ribs in this Darter, and this I believe to be generally
the arrangement in all typical Cormorants.

Several of the authorities I have referred to above have both described and fig-
ured the pelvis of an Anhinga, and the bone possesses a number of interesting char-
acters. Fifteen vertebrae of the spinal column fuse together to form its " sacrum."
The leading six of these throw out their transverse processes to coossify with the ven-
tral surfaces of the anterior portions of the ilia. The centra of the first three are
markedly compressed transversely, the first one being very deep, the next less so,
still less so the last. A large haemal spine is also found on the first, which becomes
rudimentary on the one behind, to be entirely absent in the one next in order.
Two or three vertebrae throw out their processes to abut, upon either side, against
the ilia at points just posterior to the cotyloid rings. Posterior to these, the outer
extremities of the processes of the vertebrae completely fuse with the inner borders
of the ilia ; while on either side, from the acetabulae all the way to the tail, occurs a
row of interapophysial foramina, a feature so characteristic of the pelves of the Cor-
morants. The last vertebra, though not free, is to some extent individualized, and
the extremities of its transverse processes may not so completely ossify with the
ilium on either hand, which latter bones are here drawn out into peculiarly elon-
gated posterior processes. Viewed upon its dorsal aspect we are to note that the an-
terior portions of the ilia are much horizontally expanded, while the narrowest part
of the pelvis is just in front of the acetabulae, where either iliac border shows a
marked concavity. Thorough fusion of the internal iliac borders and the " sacral
crista" takes place, and not even do the usual " neural canals " or "grooves" remain
open posteriorly. This is well shown in Plotus levaillanti (PI. XXL, Fig. 1). In
front the iliac borders develop a raised emargination, and a strong brush of coossified
tendons always project directly forward from the diapophyses and neural spine of
the leading vertebra of the " sacrum." Passing to the post-acetabular portion we find
that the pelvis is broader than it is in some of the Cormorants, and we are particu-
larly struck with the prominent crests formed by the union of the internal iliac
borders and the sacrum. The neural arch and the common neural spine of the lat-
ter is also conspicuously raised in this part of the pelvis. This begins moderately
about opposite the cotyloid rings, and increases as we advance towards the tail.

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 159

This elevation of the center and margins causes the formation of longitudinal de-
pressions between them, and down the center of either one of these we find the row
of interapophysial foramina, to which reference has already been made above.

Some Cormorants seem to have a fairly well marked propubic spine, and there is
an indication of the rudiment of such a process in the Anhinga. The internal cir-
cumference of either cotyloid ring is smaller than the external, and these cavities
are brought up very close to the sacrum. Of enormous dimensions is the ischiadic
foramen ; it has the effect of absorbing nearly all the ilium above, and to some ex-
tent behind it; while below, it makes the neck of the ischium very narrow indeed.
Inform this large foramen is subelliptical. The "obturator foramen" opens into
the obturator space, which latter is also extensive, and the pubic style which bounds
it below as far as the point where it meets the ischium, is very slender and frail.
Beyond this, the pubic bone is in close contact with the lower margin of the
ischium, where it is very considerably stouter, as it. also is after it becomes suddenly
deflected behind after passing the extreme distal angle of the ischium — precisely as
as it does in typical Cormorants.

A deep, triangular ilio-ischiadic notch, between the here very narrow ilium and
the far broader ischium, indents the posterior pelvic border.

The 'antitrochanters are prominent, and the facet on either one of them looks
forwards, downwards and slightly outwards.

Ventrally, the "pelvic basin" is seen to be fairly capacious; the "sacrum" is con-
siderably enlarged opposite the acetabulse and beyond ; the exits for the nerves of
the sacral plexus are double ; finally, in front, we see the horizontally spreading
ilia, with the compressed vertebrae, dipping down anteriorly so far ventrad.

There are six free caudal vertebrae in the skeleton of the tail of this Darter,
plus a large pygostyle. This latter bone is drawn out supero-posteriorly, and the
long superior edge is very sharp. Its antero-inferior angle is enlarged, with flattened
surface below. Beyond this, the inferior border gradually contracts and becomes
rounded. In front there is an extensive pit for articulation with the last caudal
vertebra, and above this there is a small opening where the spinal cord enters
this bone.

The centra and their articulatory facets are large in these vertebrae, but the
spinal canal is of no great caliber. In mid-series the neural arch and spine, and
the forward-projecting prezygapophyscs are conspicuous. A good-sized haemal spine
also characterizes the last four vertebra, and in each case it extends forwards to
underlap the bone next in advance. As a rule the transverse processes are rather
short and stumpy, being entirely rudimentary in the sixth caudal.

160

MEMOIRS OF THE CARNEGIE MUSEUM

FIG. 5. Ventral aspect of the sternum of Anhinga anlringa.

FlO. 6. Left lateral view of the same bone shown in Fig. 5.

FIG. 7. Anterior aspect of the osftircula of Anhinga anliinga.

Flo. 8. Mesial side of the scapular end of the same bone shown
in Fig. 7. All the cuts are natural size from the specimen (No.
18259, Colls. U. S. Nat. Mus.). Outline drawings by the author.

The Sternum and Shoulder
Girdle. — Many of the characters
of the sternum of a Darter es-
sentially agree with the corre-
sponding ones as they occur in
the sternum of such Cormorants
as represent the genus Phalacro-
corax. Its carina dies out on the
under surface of the body before
arriving at the mid-xiphoidal
process behind ; and is deep only
in front, where it protrudes for- .
wards, with a sharp anterior bor-
der, and bears a large facet, oc-
cupying the carinal angle, for
articulation with the os furcula.
Only the merest rudiment of a
manubrium is to be seen at the
usual site ; possibly the process
is entirely absent in some speci-
mens. A wide concave notch
separates the coracoidal grooves
mesiad, and the large, elongated
coracoidal facets above them.
Large, acutely triangular costal
processes are developed, each one
projecting outwards at an angle
of about 45°, and very slight
forwards. The costal borders
are short and much contracted ;
either one supports four trans-
verse hsemapophysial facets.
Posterior to the costal borders the
sternum widens, and its lateral
margins are sharp. Its hinder
border is also sharp. Two large,
broadly concave notches exist

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 161

here. They give rise to long, lateral xiphoidal processes, and a shorter mid-xiphoidal
process. On its thoracic aspect the sternal body as a whole is well concaved — uni-
formly so, as the position of the heel beneath is but barely indicated down the
median line. The form and proportions of the sternum in Anhinga are well shown
in my figures 5 and 6.

The osfurculais of the typical U-shaped pattern, with large, expanded clavicular
heads. There is but the barest rudiment of a hypocleidium below, while a facet is
present on the nether side of it, for articulation with the carinal angle of the keel of
the sternum. Either broad, laterally-compressed clavicular head, is bent rather
abruptly downwards ; a projecting, facet-bearing shoulder is developed for the coracoi-
dal articulation. Above either of these facets we note a strong process pointing back-
wards, which, when the os furcula is articulated in situ, rests upon the summit of
the corresponding coracoid. (See Figs. 7 and 8.)

Both Darters and Cormorants have unusually long coracoids. In Anhinga the
coracoid is longer than the femur. Its sternal end is considerably expanded, and
shows a wide and rather deep excavation on its posterior aspect for the coracoidal
groove and facet upon the sternum, while anteriorly the corresponding surface is
narrow and shallow. A costal process is not developed, but that border is convexly
rounded and sharp. The shaft of this bone is antero-posteriorly much compressed,
and in front a strong, muscular line marks it longitudinally, especially near the
sternal end. The scapular process is much aborted, but the head is quite massive and
tuberous. Its entire antero-mesial aspect is occupied by an oval facet for articula-
tion with the clavicular head of the furcula. The scapular facet is small, but the
glenoid surface is of fair size.

Quite uniform in width and moderately pointed distally, the thickish blade of a
scapula has its posterior extremity but very slightly bent outwards. The anterior
end of this bone is wide transversely, and the long acromial process decidedly
tilted up. The glenoidal process also stands out rather prominently, while the
articular facet for the coracoid is comparatively small. Just posterior and beyond
the glenoidal facet, the border of the scapula is broadly rounded, but is sharp all
along its mesial edge, clear to the tip of the acromial process. It is only the distal
third of the bone that is vertically compressed and at all blade-like.

On the Appendicular Skeleton.

Being thoroughly non-pneumatic, the long bones of the limbs in Anhinga are
solid and heavy. Although not so very far from being double the size, the humerus
of Fregata weighs about the same as the humerus of the species of Darter we are now

162

MEMOIRS OF THE CARNEGIE MUSEUM

describing. In PJialacrocorax urile the ulna is somewhat longer than the humerus,
while in this Darter the humerus is considerably longer than the ulna. These are
interesting facts.

In Anhinga anhinga the humerus has a length of about 13 centimeters, and it
presents the usual double sigmoid curve. This latter, however, is far better seen upon

a superior view of the bone, rather than upon its
anconal aspect. (See Fig. 9.)

The proximal end of the bone is narrow and
elongated, merging with the shaft very gradually.
The radial crest is a low, long ridge of uniform
height ; while the ulnar crest and tuberosity are
prominent and are bent anconad. This creates a
deep valley between the true articular humeral
head and the ulnar protuberance. The excavation
over which the latter arches, and where the pneu-
matic openings usually occur in other birds, is very
shallow, and indeed so very much so that really no
fossa may be said to exist there at all. A marked
diffuse excavation exists upon the palmar aspect
of this part of the humerus ; this is spread out
over a regular area just within the distal half of the
radial crest, but joins with a far narrower and some-
what deeper strip that bounds the humeral head
distad, and lies on the opposite side of the bone
to the ulnar crest. All this excavation is power-
fully marked in PJialacrocorax.

The humeral shaft proper is subcylindrical in
form, being of nearly uniform caliber for the
middle third of the bone. At the distal extremity
all the most usual ornithic characters are promi-
nent and pronounced. The olecranon fossa is
shallow, but the ulnar and radial tubercles jut out
very conspicuously, though not as markedly so in
proportion as in the Cormorants.

The ulna has a length of about 11.5 cm. and is
moderately bowed from one end to the other. The points of attachments for the
quill-butts of the secondary feathers present quite a striking feature, and each seems

Fia. 9. Anconal aspect of right hu-
merus of Anhinga anhinga. FIG. 10. An-
terior view of right femur of A. anhinga.
Fia. 11. Right coracoid of A. anhinga.
Seen from in front. All drawn in outline,
natural size, by the author, from Spec. No.
18259, Coll. of U. S. Nat. Museum.

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 1 63

to consist in a small pitlet, with its upper and lower edges very slightly raised, but
in a manner hardly to suggest the idea of a "papilla" or " tubercle" as I have so
frequently described them for many other kinds of birds.

The ulna develops a distinct lip of bone, that curves partly round the head of
the radius when they are articulated in situ. This latter bone, hardly half the cali-
ber of the ulna, shows barely any bowing or curving at all throughout its contin-
uity. Distally it is deeply grooved on top for the passage of tendons, while a deep
circular cup for articulation with the radial tubercle on the humerus surmounts its
head or proximal extremity.

The two usual ossicles, radiate and ulnare, compose the skeleton of the wrist in
the adult.

The carpo-metacarpus has a total length of 6.5 cm. and is characterized princi-
pally by a deep pit on the palmar size of the head, between the pollex metacarpal
and the process for muscular attachment. This is also a very general phalacrocora-
cine character. The shafts of index and medius metacarpals are very straight, and
the latter about as long as the former, has not more than about one fourth its bulk,
and is compressed throughout, the flat surface of the opposed side being presented
to the shaft of the former. The pollex metacarpal is short, and, on the whole the
bone is not very unlike what we find it to be in the Cormorants.

No perforation exists in the dorsal expansion of the proximal phalanx of the
index digit, and none of the terminal joints of either the thumb or the other two
fingers have claws at their extremities. The one belonging to the medius is com-
paratively long, and very sharply pointed.

I am not familiar with the existence of any special sesamoids about the articu-
lations in the skeleton of the anterior extremity, at places where they exist in some
Auks and other water birds.

Passing to the PELVIC LIMB we find a femur with a length of 5.5 cm.; a tibio-
tarsus with a length of 9 cm.; a tarso-metatarsus of 4 cm.; an accessory metatarsus
of 1.4 cm., which is nearly equalled by the patella — it having a length of 1.2 cm.
In the foot, the skeleton of hallux measures 3.1 cm., the inside toe 4.7 cm., mid-
dle toe 7.5 cm., and outside toe 8 cm. — the ungual joints being included in each
case.

Proportionately the femur is longer in Anhinga than it is in P. urile, and is not
bowed so much in the antero-posterior direction. The summit of the bone is capped
off with the articular surface, above which neither the caput femoris nor the trochan-
terian crest, rises. The external condyle is more prominent and lower on the bone
than the internal one, and is deeply cleft for articulation with the head of the fibula.

]64 MEMOIRS OF THE CARNEGIE MUSEUM

Neither the "rotular channel," nor the popliteal fossa are very deeply excavated ;
and on the head of the bone the pitlet for the ligamentum teres is also quite shallow.
The muscular lines of the shaft are distinctly denned both in front and behind.

With its shaft somewhat antero-posteriorly compressed, and very slightly bowed
to the front, the tibia-tarsus' presents us with pro- and ectocnemial processes well
developed, and a cnemial crest that rises above the summit of the bone. Of con-
siderable length is the conspicuous ridge for articulation with the fibula, and at the
antero-distal extremity of the shaft we observe the presence of the osseous bridge
under which the tendons pass in life. This end of the bone inclines as it were to
the inner side, so that if the mid-vertical axis of the large internal condyle were
extended it would be removed from and parallel to, the longitudinal axis of the
shaft, rather than being in the same line with it.

Both extremities of this bone, as is the case with the extremities of the tarso-
metatarsus, are more or less massive and enlarged as compared with the shafts of
the same. Coming to the fibula, it is very interesting from the fact that it is com-
plete, rather more so if anything than we find it to be in either Pandion or Urinator,
and is only anchylosed to the tibio-tarsus by means of its enlarged and extreme
distal end. This limited fusion of the two bones takes place at a point just above
the external condyle.

Anhinga has a large patella, but proportionately not as large as we find it in the
Cormorants. It is an oblong sesamoid, nearly equilateral, concave behind and
convex anteriorly, where it is transversely perforated by a minute foramen for the
passage of the ambiens muscle.

As will be seen from measurements above, the tarso-metatarsios is relatively a short
bone of the leg; it is further characterized by having its somewhat broad shaft com-
pressed in the antero-posterior direction and marked longitudinally upon both
aspects by strong muscular and tendinal lines and grooves. The hypotarsus is
strongly developed, especially the interno-lateral part of it, which, by a plate-like
extension, has a firm attachment to the upper third of the shaft. Its posterior angle
above is always thickened and doubly pierced for tendons. The externo-lateral part
of the hypotarsus is small and it also creates by its form a groove and a foramen for
certain tendons that pass through or over them during the life of the individual.

One or two small foramina pierce the upper part of the shaft in an antero-pos-
terior direction, the most constant one making its exit behind to the inner side of
and at the base of the hypotarsus. At the distal end the trochlese for the toes
stand well apart. The inner one of the three is the lowest, and juts out in a prom-
inent way from the bone, and has a tubercle projecting from its lateral aspect.

SHUFELDT : OSTEOLOGY OF THE STEGANOPODKS 165

The mid-trochlea is massive, not quite centrally located on the end of the shaft,
and is the next lowest in point of position. The highest of all is the external one,
and its outer part is produced the farther behind.

On the anterior part of the shaft in front, in the groove extending up from
between the mid- and external trochlere, occur two perforating foramina. The lower
one makes its exit in the intertrochlear notch ; the other on the shaft above it.
Nearly all existing birds have only a single foramen at this point. It is even single
in Phalacrocorax.

The free first metatarsal is large and strong, twisted upon itself, and with an elon-
gated transverse facet for articulation with the hallux. The skeleton of the pes is
remarkably well developed. The phalangeal joints, arranged upon the plan of 2, 3,
4, 5 for hallux to outer toe inclusive, are stout and strong, and the terminal ungual
joints are, one and all, handsomely curved and sharply pointed at their extremities.

We now pass to the consideration of the osteology of the Cormorants, which it
will be found agrees in many particulars with the species of Darter we have just
been describing.

Observations on tha Skelefology of the Phalacrocoracidx.10

A great many Cormorants of the world's avifauna, as I have before stated, belong
in this family, and as far as at present known it is the only one represented in the
United States, where, as heretofore noted, the nearly a dozen species and sub-
species it contains have all been restricted to the single genus Phalacrocorax. Cor-
morants are a good deal alike in their osteology, and in this part of their anatomy,
too, they have a good deal in common with the Darters. To give the salient features
of the skeleton in this group I will draw upon the skeleton of P. urile in particular,
as well as in general those of the collections of the U. S. National Museum, and
also republish a short description, together with the figures illustrating it, of P. per-

10 In his recent ( 1899) Hand-List of Birds, Dr. R. Bowdler Sharpe presents also a classification of the Cormorants
(pp. 232-235). They constitute the first group of his Order (XXIII.) Pelecaniformes, and all are relegated to the
family Pltalacrocorncitlic, and this latter is divided into four (4) genera, viz: ( 1 ) PhalacroLOrax ; (2) Pallasicarbo ;
(3) Nannopterum, and (4) Actiornia. The first of these contains forty-two (42) species of existing cormorants, and ten
(10) extinct forms ; the second is represented by a single extinct type, the P. perspicillatus ; the third contains only the
singular cormorant of Narborough and the Galapagos Islands ; while finally, the fourth genus is also represented by a
single extinct form, the A. anglicm of Lydekker. In the United States we have some six (6) species and five sub-
species of Cormorants, all of the family PlialncrocnracidiE. Through the constant energy and perseverance of Mr. Lucas,
a large proportion of these Cormorants are represented in the collections of the U. 8. National Museum by their skele-
tons, and through his kindness I have been enabled to study and compare all this material in the revision of the present
memoir. The collection in question is now doubtless the finest of the kind in the world, and has in it the skeletons
of more steganopodous birds than that of any other in existence.

166

MEMOIRS OF THE CARNEGIE MUSEUM

spidllatus, Pallas's Cormorant, given by Mr. F. A. Lucas in the Proceedings
of that institution for 1889 (pp. 88-94), also other material and figures by the
same author.

When we regard the skull of P. urile from above, we observe that the cranium
is very flat, and, across the parietal region, broad. Indeed, though large, the brain-

st.o

FIG. 12. Left lateral view of the skull of Phalacrocorax urile; nat. size. By the author, from a specimen in the
Smithsonian Institution. s(.o., the occipital style. (See also Plate XXIV., Fig. 15.)

cavity is vertically compressed, and this is one of the first things to attract our
attention as we examine the skull of this Cormorant. Between the sharp superior
edges of the orbits the frontal region measures transversely about a centimeter, and
this too is the average width of the not very mobile cranio-facial hinge. A broad-
ish, shallow, longitudinal groove traverses the fron to-parietal region, and the well-
marked crotaphyte fossre fail quite to meet on top of the prominent, convex occip-
ital elevation. The occipital line and crest are very sharp and ridge-like, and a
free, osseous, occipital style, over a centimeter and a half long, articulates at a median
point a short distance above the large foramen magnum, where these lines and
crests meet. This style is distally pointed, has sharp supero-median and lateral
borders, which give rise to supero-lateral surfaces intended for the attachment, on
either side, of the posterior part of the temporal muscle. (See Fig. 12.)

The upper bony beak is somewhat broad at its base ; is slightly longer than the
cranium ; narrow for its anterior half ; edges semi-cultrate ; very moderately decurved
at the end ; culmen broadly convex from side to side ; flat beneath ; external narial
apertures absent. A groove passes down upon either side, which is deepest at those
sites where nostrils occur in most birds that have them. There is a minute foramen
on either side, communicating with the rhinal chamber. The lateral processes of the
cranium are -almost entirely aborted, and the orbital cavity behind freely opens into
the cranio-zygomatic space, and the cranial wall here is bulging, smooth, and con-
vex. At the occiput the usual processes are fairly well developed, and the quadrato-
mandibular articulation, on either side, is located far posteriorly as in an alligator.

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES

167

In their essential characters, the quadrates, the pterygoids, the palatines, and
the maxillo-palatines agree with what we found above in Anhinga. Phalacrocorax
also has a small supramaxillary, or as Parker called it, a " post-maxillary," and
according to him it is large in P. carbo, and small in P. graculus.

FIG. 13. Palatal region of a young Pluilacrocurax urile, showing the maxillo-palatine of the right side.

FIG. 14. Dorsal aspect of anterior part of the cranium of a young I'halacrocoraj; urile, showing the absence of a
fronto-nasal hinge.

FIG. 15. Lateral aspect of beak of a young Phalacrovorax urile, showing the open condition of the nostril. All
three figures drawn by Mr. F. A. Lucas from specimen No. 12732 of the collections of the II. S. Nat. Museum, and
enlarged hy means of the camera lucida.

Regarding the maxillo-palatines and related bones in the skull of a Cormorant,
it seems that ornithotomists do not quite agree as to the exact sutural boundaries,
and it seems that Mr. W. P. Pycraft, of the British Museum, some time ago printed
an article on the osteology of certain steganopodes, in which he was at variance with
anatomists upon this point. Desiring the opinion of Mr. Lucas upon this point, he
has kindly furnished me with the following remarks, and illustrated them with
some drawings, which latter are here reproduced as Figs. 13, 14 and 15. In his
communication he says: "Some difference of opinion exists as to just what portions
of the maxillaries represent the maxillo-palatines, and while I do not like to differ
with my friend Mr. Pycraft for fear I be on the wrong side, yet, after carefully con-
sidering the subject, the conclusion is forced upon one that the posterior extensions
of the maxillaries are the maxillo-palatines."

168 MEMOIRS OF THE CARNEGIE MUSEUM

"By the kindness of Mr. R. C. McGregor and Mr. Curtis Clay Young the Museum
some time ago came into the possession of an extremely good series of Cormorant
skulls of all ages, and by means of these it is possible to trace the process of ossifica-
tion very well. These seem to show that the desmognathism of the Cormorant's
skull is caused by the growth of bone between the maxillaries, and is not due to any
outgrowth of those bones themselves."

" This ossification commences between the backward processes here considered as
the maxillo-palatines and progresses until they are firmly united. The union of the
anterior portion of the maxillaries occurs later and is due to the extension of ossifi-
cation into the tough lining of the roof of the mouth, and there is frequently a small
space between the maxillaries which remains open for a long time if not perma-
nently" (Fig. 13).

"The formation of the fronto-nasal hinge (Fig. 14) is a secondary character, the
bones so overlapping in young birds that there is no freedom of movement in this
region."

"The absence of external narial openings is also a secondary character, for the
young Cormorant possesses perfectly open nostrils while the cranium is almost as
schizo-rhinal as that of a gull (Fig. 15). As growth proceeds the narial openings
become more and more restricted, until about the time (the exact time is uncertain)
that the young birds take to the water, not only the external openings, but those of
the cranium have become completely filled."

Passing once more to the skull in the adult we find the interorbital septum is
unossified, as is the greater part of the anterior wall of the brain cavity. However,
we here find the "foramen rotundum" separated from the far larger vacuity above,
by a transverse bony bar. A subcircular vacuity also occurs in the mesethmoidal
plate ; and the sphenoidal rostrum is not as large as the zygoma.

Phalacrocorax urile differs from the Anhinga inasmuch as in it the pars plana does
not ossify ; and in the form of its lacry mal bone. One of these in this Cormorant is seen
to be fused with the frontal bone and nasal above; is laterally composed; not in con-
tact with the maxillary below, where it sends outwards a short process, and inwards
a longer and slenderer one, the mesial end of which fuses with the mesethmoid (see
Fig. 12).

This specimen (No. 18982, Coll. U. S. Nat. Mus.) lacks a vomer, and it was not to
be found in my skeleton of Anhinga; in the latter, if it ever existed, it may have
been lost — but I see in another place, where remarking upon the osteology of P.
bicristatus (now urile) (Science, N. Y., v. 11, No. 41, Nov. 16, 1883, p. 640), I have
said that in it "we observe a long attenuated vomer, terminating anteriorly in a free

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 169

pointed extremity." But I am inclined to think that the vomerine ossification is
not always a constant character in these birds. Ossifications of the eyes, ears and
tongue agree in the main with the corresponding parts in the Anhingidas, and have
been more or less fully described above.

V-shaped in pattern, the mandible is a strong bone in birds of this family. This
is due to the thickness of the rami, and not to their height, for they are rather low
than otherwise ; also to the unusual firmness of the dentary ossification with the
other bones posterior to it, and finally to the total absence of a ramal vacuity -
the mandible not being weakened at those points (see Plate XXIV., Fig. 17). Each
dentary upon its mesial aspect shows a deep longitudinal groove. The articular
ends are truncated posteriorly, though the postero-mesial angles are somewhat pro-
duced. Eacli of these articular cups appears to be connected with |its respective
ramus by a kind of neck, which is directed backwards and towards the median line.
When seen from above, and the jaw is articulated in situ, this gives rise to rather
an odd-appearing articulation. Either quadrate is thus entirely shut out of sight,
with the exception of its^externo-lateral process, the base of which does not come
in contact with the mandible at all. The sides of the rami are within the quadrato-
jugal bars, and rise somewhat above them. The pterygoids, owing to a special
notch in each articular cup, are thoroughly exposed,, but no more.

Of the Remainder of the Axial Skeleton. — Seventeen vertebrae are found in the
cervical region of P. urile, before we arrive at one that bears a free pair of ribs. A
small pair of rudimentary ones in this species are to be observed in the eighteenth
vertebra. Now in P. urile these last are not liberated, while in the nineteenth and
twentieth vertebrae the free ribs are long, well developed, and have anchylosed upon
them large unciform processes. In the leading pair, however, either upon one side
or the other, this last character may be absent. Such is the case on the left side
of the specimen at my hand.

P. urile has its entire vertebral chain in many respects quite different from any-
thing we described for Anhinga. We do not meet with the greatly elongated eight
leading cervicals, and the abrupt change in character as we pass to the ninth one.
Nothing of the kind occurs in the Cormorant, for in it the eight leading cervicals
are quite in harmonic proportions in all respects with those that follow them. We
nevertheless find in this Cormorant the same modifications, only in a far less marked
degree, in the eighth, ninth and tenth vertebras, which give this bird the power to
draw back its head and with great rapidity to thrust it forward again, the point of
flexure being between the vertebrae just mentioned. Anhingas catch all the fish
they eat by spearing them with their sharp beaks while in active pursuit under

170 MEMOIRS OF THE CARNEGIE MUSEUM

water. The head is drawn back and then thrust forward like a spear, with wonder-
ful rapidity and precision — the prey being transfixed on the closed, lance-like
mandibles. When prepared to deliver the blow, the angle formed by the seventh
and vertebrae points backwards, and the one between the eighth and ninth forwards,
the eighth vertebra itself standing subvertically between them. Gannets and
Herons have the same arrangement of these cervical vertebrae, but nothing like as
well developed as we find it in the Darters. Bitterns show it well,11 as do all our
American Herons, and they transfix their prey, although they do not pursue it
under water as the Darters habitually do.

P. urile has handsomely developed parapophysial spines on its cervical verte-
brae. They are rudimentary on the atlas and axis, coming to be of good size on the
third vertebra. From it on they gradually increase in length, but diminish in
caliber, until we come to the tenth cervical. In it they are as straight as sewing
needles, and as long as the centrum of the vertebra. They die out on the fifteenth
cervical. The lateral canals are normally developed and of good size in the third
vertebra at its anterior part. They very gradually increase in size down the series,
until we arrive at the first vertebra bearing free ribs. On the other hand the hypa-
pophysial carotid canal begins in the fourteenth vertebra, and terminates in the
seventh ; it never quite closes in so as to form a perfect tube, but comes very near it
in the thirteenth vertebra. A large haemal spine suddenly appears on the fifteenth
cervical, and the character persists down the series, to include the leading two or
three vertebrae of the pelvic sacrum. Haemal spines are also found on the atlas and
axis ; and a strong neural spine first appears on the sixteenth vertebra, while on
the eighteenth to the twenty-third inclusive they are very large, strong, oblong in
form, and much in contact with each other in the dorsal region of the spine ; the
various articulations among the vertebrae, are very close indeed. Here the centra
are much compressed from side to side ; the transverse processes narrow and spread-
ing, and these latter have strong metapophysial spinesinterlocking at their extremities.
But there is none of that luxuriant interlacing of ossified tendons that we see
there, and on the neural spines and anterior iliac margins, as in Anhinga. Phala-
crocorax is more or less free from that.

There are three pairs of large, strong dorsal ribs, bearing great, flat, anchylosed
unciform appendages. By a graduated series of costal ribs, these vertebral ribs join
with the sternum. The first pair of pelvic ribs likewise, by a very long pair of
hsemapophyses, connect with the sternum. Small unciform processes also appear
upon them, but not on the last pair of pelvic ribs that follow these. Nor do the

11 See the author's reference to this in the American Bittern in The Auk, Vol. X., No. 1, January, 1893, pp. 77-78.

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 171

costal ribs of this pair quite reach the sternum. On the left side in the specimen
before me there is also a "floating costal rib," with a length of nearly 3 centimeters.
It lies close along the posterior border of the distal moiety of the ultimate hsema-
pophysis.

This last-mentioned bone in a specimen of P. urile articulates with the sternum
on the left side, and no doubt from time to time such will be found to be the case
in other species of Phalacrocorax, i. e. — the ultimate pair of costal ribs or haema-
pophyses, upon one side or the other, or for the matter of that, upon both sides, may
join with the sternum the pelvic ribs to which they belong.

In Phalacrocorax we find upelvis very much like the one we described for Anhinga.
The ilia are horizontally spread out in front, where they may be so thin as to present
a number of perforating foramina. In the'postacetabular region these bones come
much closer together, both actually and relatively, than they do in the Darters. The
interdiapophysial foramina also are present, a row upon either side of the posterior
sacral crista, and they are very large just before we arrive at the first caudal vertebra.
The internal iliac borders are not elevated here as they are in Anhinga. Cormorants
of this genus also have the large obturator space ; the enormous ischiadic foramen ;
the rudimentary propubis ; and the same peculiarities of the postpubis.

Posteriorly, either ilium sends back a conspicuous process, and the two hold
firmly between them the first free caudal vertebra. In some species, as P. urile,
they may grasp two instead of one. An ilio-ischiadic notch is also present upon either
side of the pelvis.

There seem to be seventeen vertebra? included in the pelvic sacrum to the fifteen
we found in Anhinga. Six of these are beneath the fore part of the ilia ; two more
with aborted processes lie immediately between the large acetabulae ; two true dorsal
ones follow these, and they have their lateral processes thrown out as abutments to
the walls of the pelvis; finally seven more so-called uro-sacrals are to be counted
between these and the first caudal.

Viewed ventralwise, the pelvic basin is seen to be deep and capacious. Another
marked character on this aspect is the strong 'ridge or crest of bone on either side,
which extends longitudinally backwards from a point below the cotyloid ring to the
hindermost angle of the ischium. It is strongest directly below the great ischiadic
foramen. This character is but feebly marked in the pelves of the Darters.

There are six caudal vertebrae and a pygostyle. The latter is large ; pointed
posteriorly, triangular in form, broad and flattened at its lower part behind, in
front of which there is a haemal spine. Haemal spines are also found upon the last
three caudals, and neural ones on them all. These tail vertebrae are large and strong,

172

MEMOIRS OF THE CARNEGIE MUSEUM

and the transverse processes on the fourth and fifth extend out considerably beyond
the others of the series.

Sternum and Shoulder Girdle. — In the article in Science cited above it was said of
the sternum of P. urile that it had " two shallow excavations on either side of the
median line, occupying the entire xiphoidal margin or border" (p. 641). This no
doubt was a lapsus calami, as no true Cormorant has more than one such excavation
upon either side of the sternal keel. Apart from its greater size, the sternum in P.
urile agrees in almost every particular with that bone as we described it for A. an-
hinga. The lateral xiphoidal processes, however, are comparatively not as long nor
as narrow ; nor are the costal processes of the sternum of the Cormorant relatively as
lofty or as slender as they are in the Darter. Otherwise the two bones essentially

FIG. 16. Right lateral view of sternum and bones of the shoulder-girdle of a Cormorant ( Phalaerocorax urile), nat.
size, by the author, from a specimen in the collection of the U. S. Nat. Museum. (Compare this figure with Fig. 30
of Plate XXVI.)

agree. (Compare also the sternum of Phalaerocorax albiventris, PI. XXVI., Fig. 30.)
Again upon comparing the several bones of the pectoral arch in these two forms, we are
once more at a loss to find any noteworthy differential characters. The os furcula
in each are upon identically the same pattern ; the coracoids are alike with the ex-
ceptions that in the Cormorant the bone is proportionately not as long ; its sternal end

SHUFELDT : OSTEOLOGY OF THE STEGANOI'ODES 173

is more expanded, and it has a broadly developed " costal process." In the case with
the scapula, we find that the blade of it in the Cormorant is not as much curved,
nor is the distal end of the bone at all bent outwards. (See Fig. 16.)

On the Appendicular Skeleton of P. urile.

Very little of the skeleton of this Cormorant is pneumatic. For the most part
the bones are solid and heavy. Apparently air gains access only to certain parts of
the cranium and lower jaw, other bones of the osseous system being completely non-
pneumatic. This is especially true of the skeleton of the limbs, where all the bones
typify in the highest degree the un aerated variety.

In a number of respects the bones of the pectoral limb of P. urile present char-
acters which essentially agree with those upon the corresponding bones as found in
the arm of an Anhinga, which characters have already been described above. In
the humerus of Phalacrocorax, however, which is shorter than the ulna, the " pneu-
matic fossa" is profoundly concave, quite as much so as in those large species of
birds in which the bone is pneumatic, or even more so than in some which show that
condition. At its distal end the articular protuberance for the radius has an elon-
gation at its proximal extremity which is bent over towards the mid-longitudinal
axis of the shaft of the bone. The "oblique tubercle" for the ulna is more hemi-
spherical than we usually find it in birds. At the proximal end of the ulna we ob-
serve a conspicuous projecting lip of bone, which contributes an additional surface
for the radial articulation. Papillae down the shaft of this bone for the quill-butts
of the secondary feathers are in a single row and rather feebly developed. Both
radius and ulna are somewhat bowed, and when articulated in situ they are in con-
tact for their distal moieties while proximally a good-sized spindle-shaped interosse-
ous space occurs between them.

The characters of the skeleton of the hand are practically the same as we found
them to exist in the manus of Anhinga. There are no distinguishing characters of
any importance.

Gannets have the skeleton of their pectoral limbs in many respects like the
Cormorants and Darters, but in them all the bones are completely pneumatic.
When we come to consider the difference in the habits of the representatives of the
three families, this is not so much to be wondered at.

Cormorants are more or less like the Darters too, in the osteology of their pelvic limb.
There are more differences, however, to be found here than we discovered upon compar-
ison of the pectoral extremities. Phalacrocorax has the femur proportionately stouter,
shorter, and more bowed in the antero-posterior direction than it is in Anhinga.

174 MEMOIRS OF THE CARNEGIE MUSEUM

We also note that the pit for the insertion of the ligamentum teres is more elab-
orately scooped out, and the trochanterian protuberance is produced far more to
the front in the Cormorant, such hardly being the case at all in the Darter. The
fibular and intercondyloid notches are very deeply sculpt, though anteriorly the
rotular channel is unusually shallow.

The large trihedral patella, with its broad, flattish base is, as Garrod remarks,
often laterally pierced for the passage of the tendons of the ambiens muscle.12 The

Fia. 17. Knee joint of a Cormorant (Phalacrocorax urile); natural size. F, femur; P, patella (originally de-
scribed by the author in Science); Fb, fibula; T, tibio-tarsus. Drawn by the author from a specimen in the Smith-
sonian Institution.

FIG. 18. Knee-joint of Phalacrocorax urile, seen from in front. Natural size. Letters signify the same as in other
illustrations.

cnemial processes of the tibio-tarsus are fairly well developed and are confined to
the anterior aspect of the head of the bone, where we also see a strong cnemial crest
rising above its summit. For its entire length, the shaft is compressed in the
antero--posterior direction, and the fibular ridge is long — standing well away from
the side of the shaft. Phalacrocorax has the fibula about as well developed as we
found it in Anhinga, it being complete and fused with the tibio-tarsus only at its dis-
tal end. When the bone is held vertically the internal tibio-tarsal condyle is the
lower of the two on the extremity of the shaft.

Cormorants have a tarso-metatarsus differing in some marked particulars from that
borne in the Darters. In the first place it is pierced by but one arterial foramen at
its trochlear end, and its mid-trochlear process is the lowest on the shaft, rather than
the inner one as in the Anhingas. Tendinal grooves, and their dividing lines up

12 Garrod, A. H. Coll: Scientif . Mem., p. 198.

SHTJFELDT : OSTEOLOGY OF THE STEGANOPODES 175

and down the shaft, back and front, also slightly differ, as does the form of the
hypotarsus in a lesser degree.

The skeletal plan of the pea in Phalacrocorax essentially agrees with what we
found in the Anhingidse. We are to note the great length of the basal joint of
hallux, and for its entire continuity, how muchjthe shaft of the bone is bowed.

Mr. Frederic A. Lucas, Curator of the Department of Comparative Anatomy
in the U. S. National Museum, in describing some of the bones of P. perspidllaius
in the Proceedings of that institution for 1899 (Vol. XII., p. 88), remarked that he
had the following material upon which to base his studies, and which had been
collected by Dr. Leonard Stejneger on Bering Island in 1882.

" Rostral portion of cranium in " Right fused metacarpals, very im

advance of the fronto-nasal hinge, with perfect,
attached palatines. " Three pelves, lacking pubic bones.

" Lower mandible. " Left femur.

" Right ramus of lower mandible. " Two left tibiae.

"Two nearly complete sterna. " Right tibia.

" Right coracoid. " Two left tarsi.

" Right humerus. " Second cervical vertebra.

"Left humerus of another individual. " Third cervical vertebra.

" Right ulna. " Ninth (?) cervical vertebra.

" Right fused metacarpals.

" The more important of these are figured on the accompanying plates, all figures
being of natural size, and drawn by the author.

" The bones, although stained, are in a good state of preservation, being but
slightly weathered, and all are from thoroughly adult individuals.

"For a better and briefer description of these bones they have been compared
with those of an adult Phalacrocorax carbo, and the opportunity has been taken to
test, to some extent, the value of the subgenera Urile and Phalacrocorax, by com-
paring at the same time the corresponding bones of P. urile and P. dilophus.

"The former bird is, for the species, large and the latter somewhat undersized,
although adult.

" The rostrum of perspidllatus is nearly as long as in carbo, but much more
slender, and is readily distinguished from it by the deep, lateral, longitudinal
groove characteristic of the subgenus Urile.

"The under surface of the rostrum is less grooved, longitudinally, than that of
carbo and much less so than that of P. urile (see PI. XXIV., Figs. 13-21, and PI.
VI., Figs. 25-28).

176 MEMOIRS OF THE CARNEGIE MUSEUM

" The palatines are as long as those of carbo, anteriorly narrow and posteriorly
wider, conforming in pattern very nearly to those of urile, while dilophus resembles
carlo in this respect.

" The trans-palatine angle is more rounded than in carbo, much more than in
urile, resembling in this dilophus.

" The inner portion of the post-palatine is less produced ventrally than in carbo,
and the pterygoid articulation wider and flatter than in carbo, the palatine thus
lacking the keel present in carbo.

"The same differences are found between the same parts of urile and dilophus.

" The maxillo-jugal bar is as long as that of carbo but more slender.

"The lower mandible is slightly shorter and decidedly weaker than that of
carbo, and the lower mandible of urile is proportionately still weaker than that
of dilophus.

"The dentary portion of the mandible is more deeply grooved along the inner
surface than that of carbo, being comparatively the same as in urile.

"The cutting edges of the mandible are comparatively straight as in carbo and
dilophus, but urile differs from all three in having the mandible distinctly recurved.

"The sternum is transversely flatter than that of carbo, being a trifle more flat-
tened even than that of urile. The carina is also shorter than in urile, but in size
and general appearance the sterna of perspicillatus and urile resemble one another
very closely.

"From manubrium to meso-xiphoid that sternum is 13 mm. shorter than that
of carbo, being exactly as long as that of urile.

" The proportion of carina to total length is shorter than in either carbo or urile,
the sternum from anterior end of carina to mesoxiphoid measuring 2 cm. less than
that of carbo and 4 mm. less than that of urile.

" Between the costal borders the sternum is slightly wider than in either ca.rbo or
urile.

"The rudimentary manubrium, like that of urile, lies in the plane of the body of
the sternum, while in carbo and dilophus the manubrium lies in the plane of the
keel.

"If a line be drawn between the costal processes it will be found that the coracoid
articulations project less beyond this line and form a more obtuse angle with one
another than they do in carbo, and the same is true of urile as compared with
dilophus. The sternum is non-pneumatic, as in urile, but in carbo and dilophus
good-sized foramina pierce its dorsal face just back of the ridge formed by the cora-
coidal groove.

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES

177

" It is certainly interesting to find the sterna of these two representatives of the
snbgenera Phalacrocorax and Urile, respectively, agreeing with one another in these
slight structural points.

" Articulations are present for five pairs of ribs, the same number as in carbo. One
specimen of urile has four pairs of articulations, another has five on the left side and
four on the right, and dilophus has but four pairs of costal facets.

\

Bones of Pallas' Cormorant. Natural size. After Lucas.

Flo. 19. Ventral aspect of rostrum. FIG. 20. Lateral aspect of rostrum. Fici. 21. Right ramus of lower
mandible, external aspect. FIG. 22. Right ramus of lower mandible, internal aspect. FIG. 23. Right metacarpals.

178

MEMOIRS OF THE CARNEGIE MUSEUM

"The number of ribs articulating with the sternum is, however, subject to slight
variations, especially among water birds, and without an extensive series of speci-
mens it is a little difficult to be sure of the normal number.

Bones of Pallas' Cormorant. Natural size. After Lucas.

Flo. 24. Right coracoid, ventral aspect. FIG. 25. Sternum, ventral aspect. Flo. 26. Femur, anterior aspect.
FIG. 27. Tarsus, anterior aspect. FIG. 28. Tibia, anterior aspect.

SHUFELDT : OSTEOLOGY' OF THE STEOANOPODES 1 79

" The coracoid is of the same length as that of carbo, 10 mm. longer than in urile;
but, while the proximal end is but little heavier than in carbo, the shaft and espe-
cially the distal end are much more massive.

" The epicoracoid is prolonged upward into a sharper hook than in any of the
other species at hand, but this process is subject to considerable "variation with age
or in various individuals.

" One humerus is a little longer than that of carbo, the other is of exactly the same
length ; both are much stouter, especially in the proximate half.

"The humerus is practically non-pneumatic, the foramina being very minute,
while the pneumatic foramina of carbo, though not large, are readily seen.

"The humerus of urile differs from that of dilophus precisely as that of per-
spicillatus from carbo.

"The ulna is distinguishable from that of carbo only by its greater weight, and
the same may be said of the humerus of urile, as compared with that of dilophus.

" The fused metacarpals are slightly shorter and slightly stouter than in carbo,
and here again the same differences are observable between the metacarpals of urile
and dilophus.

"The 'sacrum,' as a whole, is as long as that of carbo, but its component parts
are more heavily built.

" It comprises six pre-sacrals, two true sacrals, and nine post-sacrals, and the
three ' sacra ' of perspicillatus agree with one another in these particulars.

" Phalacrocorax carbo has six pre-sacrals, two true sacrals, and nine or ten post-
sacrals. P. urile has six, two, eight, and dilophus six, two, nine.

" The hypapophyses of the anterior three vertebra have been broken off, but
although the compressed centra are larger than in carbo, the hypapophyses seem to
have been smaller.

" The six pre-sacrals present few salient characters, but the diapophyses of the
fourth vertebra lie at right angles to the vertebral column, while in the three other
species the diapophyses of this vertebra are directed forward.

" The sacral and immediate post-sacral vertebrae vary in the development of
their parapophyses in all four species under consideration.

" In all three specimens of perspicillatus the two true sacrals bear no parapophyses,
while the two succeeding vertebrae have them extended to, and anchylosed with,
the ilium.

"The diapophyses and parapophyses of these vertebra are united by a thin
plate of bone, but that this is due to age is shown by the condition obtaining in the
other species.

180 MEMOIRS OF THE CARNEGIE MUSEUM

" These latter also indicate that the canal formed by these processes, the centra
of their vertebrae and the ilium, is larger on the right side than on the left, and that
it is the first obliterated on the left side.

" In carbo neither the sacrals nor the second post-sacral bear parapophyses,
although these are present on the first post-sacral, uniting it firmly with the ilium.

" In one example of urile, slender, but well-marked parapophyses connect the
two sacrals with the ilia.

" In another and much smaller specimen the second sacral has a parapophysis
on the left side, there being no parapophyses on the first sacral.

" In both specimens of urile the first, but not the second, post-sacral bears para-
pophyses. Finally, dilophus has strong parapophyses on the second sacral and first
post-sacral, but none on the second post-sacral.

"The variation in the sacral region of these specimens is not only interesting in
itself, but interesting from the fact that it is unusual for parapophyses to be present
at all on the true sacral vertebrse of birds.

"Viewed from above the ridge formed by the confluent spinous processes of the
'sacrals' is wider than in carbo, and the interpophysial foramina are nearly closed,
while in carbo they are very open.

" Although these characters depend to some extent on age, they do not entirely,
and the same differences exists between the ' sacra ' of urile and dilophus as between
those of perspicillatus and carbo.

"The pelvis is much more rugose than in carbo, all attachments for muscles be-
ing strongly emphasized.

"The anti-trochanter is placed further back than in carbo, and is much more
rounded, thus affording more play to the femur.

" Just back of the anti-trochanter the outer edge of the ilium is raised and thick-
ened, forming a flat, subtriangular spot, but proportionately smaller than in per-
spicillatus.

" Back of this flattened portion the dorsal edge of the ilium is bent outward,
making this part of the ilium outwardly concave, where in carbo it is convex.

"The post-ilia of carbo and dilophus round gently outward and downward
throughout their entire length from their junction with the diapophyses.

"Viewed from the side the dorsal outline of the 'sacrum' is slightly decurved,
while that of carbo is very nearly straight and the same difference exists between
urile and dilophus.

"The acetabulum is slightly larger and forms more nearly a perfect circle than
in carbo.

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES

181

"The ilio-ischiatic foramen is subelliptical and wide, the longitudinal diameter
being nearly twice the vertical, while in carlo this foramen is more pointed pos-
teriorly and narrower, the length being more than three times the height.

" In consequence of the size of this foramen the distance from the dorsal edge of
the ilium to the ventral edge of the ischium is much greater than in carlo.

"The bar of the ischium bounding the obturator space is sharp-edged, rugose,
and concave exteriorly on the posterior portion, while the corresponding portion of
the ischium in carlo is comparatively smooth and slightly convex posteriorly.

Bones of Pallas' Cormorant. Natural size. After Lucas. FIG. 29. Left lateral aspect of pelvis. Flo. 30. Dorsal
aspect of pelvis. The long, sweeping post-pubic bone was broken off and lost from this bone when it was discovered.

182 MEMOIRS OF THE CARNEGIE MUSEUM

" The posterior border of the ischium is straighter than in carbo and the ilio-pubic
articulation one third shorter.

"The femur is 5 mm. longer than that of carbo, in every way much more massive,
and with all the muscular ridges more pronounced, while curiously enough it is
more pneumatic, having several foramina in the ventral aspect of the neck that are
lacking in carbo.

"There is nothing to distinguish the femur of wile from one of dilophus of the
same length, and of the two that of dilophus is slightly the heavier.

" But in the specimen of urile in which the humerus corresponds in length to that
of dilophus, the femur and tibia are both longer and heavier than in dilophus, and
the tar_sus a little lighter.

" The phalanges, again, are more massive in urile than in dilophus.

" The smallest of the three tibiae is slightly longer than that of carbo, the cnemial
crest is more expanded, and the cnemial ridges farther apart and more pronounced.

" The distal extremity of the tibia is also wider than in carbo, but at its smallest
diameter the shaft is no larger.

"The muscular ridges and grooves are more marked than in carbo, but in the ab-
sence of more material and making due allowance for individual variation, it is dif-
ficult to point out characters which definitely distinguish the tibiae of the two birds.

"The tarsus is of the same length as in carbo, but much wider, and, as through-
out, with all the ridges more pronounced.

" Little can be said concerning the three cervical vertebrae, except that, unlike the
other bones, they are less strongly built than the corresponding bones in carbo.

" From the foregoing notes it will be seen that the differences existing between
corresponding bones of perspicillatus and carbo also exist between the same bones of
urile and dilophus, and that conversely perspicillatus. and urile agree with one another
as do carbo and dilophus.

" The subgenera Phalacrocorax and Urile, therefore, seem to rest on good struc-
tural foundations, each being characterized by internal as well as external characters.

" Unfortunately no skull of perspicillatus is to be had, but the crania of carbo and
dilophus agree with one another, while differing strikingly from the cranium of
urile.

" From the harmony of the other parts its not assuming too much to suppose
that the skull of perspicillatus would resemble that of urile."

" With the exception of the sternum the greater size of the bones distinguishes
those of perspicillatus from those of urile, while well-marked differences of shape or
proportion exist between the corresponding bones of perspicillatus and carbo.

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES

183

MEASUREMENTS (IN MILLIMETERS) OF CORRESPONDING BONES OF PHALACROCORAX
PERSPICILLATUS, CARBO, URILE, AND DILOPHUS. [FROM LlJCAS.]

(All measurements are in a direct line and not along curves.)

P. perspi-

cillalus
( National

.Museum,
17041).

P. earbo P. urile i P. dilophus

(Yale College (National (National

Mnsenm, Museum, Museum,

535). 12502). 18050).

Rostrum :

Tip of mandible to extremity of maxillo-jugal bar 134 140

Maxillo-jugal bar j 68 69

Tip of mandible to posterior end of palatine 109 117

Width across nasals, at fronto-nasal hinge j 21 . 20

Lower mandible :

Length of ramus 139 144

Greatest height of ramus j 13 13

Sternum :

Anterior end of carina to mesoxiphoid 104 119

Manubrium to mesoxiphoid 83 97

Depth of carina 28 33

Width across articulations of first rib 64 66

Width across articulations of fourth rib 63 59

Coracoid :

Length 84 87

Breadth of sternal articulation 25 25

Greatest distal breadth 18 17

Humerus :

Length 170 170

Greatest proximal breadth 30 28

Greatest diameter of shaft midway between extremi-
ties 11 9

Greatest distal breadth... 21 20
Ulna :

Length 190 178

Greatest proximal breadth j 21 18

Greatest diameter of shaft midway between extremi-1

ties 8 7

Pelvis :

Greatest length of ilium I 151 152

From anterior border of ilium to external angle of I

anti-trochanter I 72 65

Greatest width of ilia in advance of acetabula | 48 44

Least width of ilia in advance of acetabula 23 23 +

Width between outer extremities of anti-trochanters. 43 46

Length of ilio-ischiatic space ' 42 41

Greatest width of ilio-ischiatic space 18 13

From dorsal edge of ilium, above the ilio-ischiatic

foramen, to ventral edge of iscbium 28 23

Length of ilio-pubic articulation 21 28

Between posterior terminations of ischia I 46 40

Femur :

Length 74 70

Greatest proximal width 21 19

Greatest diameter midway between extremities 11 10

Greatest distal width 22 18

Tibia :

Length 140 127

Width across cnemial ridges 15 13

Width at distal end of articulation with fibula 15 13

Least transverse diameter of shaft 8 8

Distal width 16 15

"Tarsus" :

Length 71 72

Proximal transverse width .- 19 16

Distal width 20 18

108
56
91
13

112

97
84
31
64
60

71
24
16

140

25

10
18

122

58
42
19
37
32
13

23
15 +
42

66

16

9

16

117
11
11

7
12

60
14
15

105
56
86
14

110
12

91
76
26
54
51

70
20
13

140

23

8
17

120

49
38
18
33
38
10

19
23
40

55
15
8
15 +

102

11

13

7

13

62
14
14

184

MEMOIRS OF THE CARNEGIE MUSEUM

" P. perspieillatus appears to have been a much heavier bird than carbo, and a bird
of weaker flight ; with more robust and muscular legs, and a more slender and more
feeble head and neck.

Skull and Jaw-bones of Pallas' Cormorant. (After Lucas.)

Flo. 31. Phalocrocorax perspicillatua, inferior aspect of cranium. The anterior and posterior portions are from dif-
ferent individuals. (Nat. size. )

Flo. 32. Phalacrocorax perspieillatus, left ramus of jaw, external aspect. ( Nat. size. )

FIG. 33. Phalacrocorax perspieillatus, mandible and left palatine, external aspect. (Nat. size.)

" In comparing the tables of measurements it must be said that they do not
adequately convey the impression produced by a comparison of the bones them-
selves. Thus, in the measurements of the lower mandible the greatest vertical width

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES . 185

is comparatively as in carbo, but from this point the ramus tapers rapidly either way
so that, as a whole, the mandible is much weaker than that of carbo.

"So too with the humerus, where the greatest proximal width is only 2 mm.
greater than in carbo, although the bone in its entirety is much more stoutly built."

In a following volume of the Proceedings of the U. S. National Museum, in
which Mr. Lucas described the above specimens of bones of P. perspicillatus, he
on subsequent pages also described the skull of this species, and presented other
data of importance having reference to the Phalacrocoraddss (Vol. XVIII. , pp. 717-
719, Pis. XXXIV.-XXXV.).

This second lot of bones were likewise obtained by Dr. Stejneger at Bering
Island in 1895, or thirteen years after the first specimens were collected by him.
A cranium and a sternum of P. perspicillatus are found in this second installment,
and they are very important discoveries.

The cranium is now in the collection of the Museum (No. 19417, U. S. N. M.).
"In its general contour," says Mr. Lucas in his paper referred to above, "it most
closely resembles that of P. penicillatus among existing cormorants, but is decidedly
larger, and is proportionately wider than in that species, while the beak is shorter.
As far as mere size is concerned, the skull of an adult male of P. carbo would be as
long as that of P. perspicillatus, but the latter is much wider and is more depressed.
The cranium is readily distinguished from that of P. urile by its greater size and less
depression, and by having a proportionately stouter beak, whose ridge lacks the
slight but characteristic emargination found near the base of the beak in P. urile.

"As a matter of fact, the differentiation of cormorants into species with grooved
beaks and those without does not exist, so far as the bony beak is concerned. Some
have deeper grooves than others, but all have more or less of a furrow along the side
of the mandible, and there is every degree of gradation, from such well-furrowed
beaks as those of P. albiventris and P. magellanicus to the shallow grooves of P. melan-
oleucus and P. carbo.

" Pallas's Cormorant shows a marked difference from all others examined in the
development of the lateral ethmoid. In other species the lachrymal sends a process
inward which fuses with a spur from the mesethmoid to form a more or less L-shaped
bar of bone, uniting the frontal and mesethmoid. A small spur, arising from the
inferior inner angle thus formed, represents the lateral ethmoid, and this is' usually
but little developed, being largest in P. penicillatus and obsolete in P. urile. In P.
perspicillatus there is a lateral ethmoid plate, complete save for an opening above,
being the retention by ossification of a cartilaginous plate found in the nestling of
P. urile before the nostrils have become closed. The maxillopalatines are also

186

MEMOIRS OF THE CARNEGIE MUSEUM

slightly better developed than in any existing cormorant, and while the difference
is small, still it does exist, and here again it is seen by comparison to be the develop-
ment of a character found in young birds.

34

36

Crauia of various Cormorants. ( After Lucas. ;

FIG. 34. Phalacocorax penidllatus.

Fid. 35. Phalacrocorax perspicillatus.

Flo. 36. Phalacrocorax carlo. (Two thirds natural size ; from photographs. )

" Difference exist between P. perspidllatus and other cormorants by the presence
of a narrow bar of bone forming two precranial cavities where but a single opening

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES

187

exists in allied species, and in the comparatively small size and regular lyrate form
of these openings. From these conditions it will be seen that there is in the cranium
an excess of ossification over that found in other cormorants. While no bar of bone
has been found in other species, there are hints of it in some, thus, P. penicillatus
and P. magellanicus, in the shape of a little bony spike running upward from the
alisphenoids, and it is not impossible that the complete bar may be found in some
very old individual. This is the more probable because in the young, of P. urile at
least, there is a bar of cartilage occupying the place of the bar of bone found in
Pallas' Cormorant.

MEASUREMENTS13 OF SPECIES OF PHALACROCORAX.14 [FROM LUCAS.]

STERNUM.

P. perspicillatus,
male
(U. S. N. M.,
No. 19417).

P. carbo,
male (Yale
College,
No. 535).

P. penicillatus
(U. S. N. M.,
No. 18535).

P. urile,
male
(U. S. N. M.,
No. 12502).

Anterior end of carina to mesoxiphoid

mm.
13215

mm.
119

mm.
101

mm.
104

Manubrium to mesoxiphoid. .

109

97

87

90

Depth of carina

4Q15

33

22

25

Width across articulations of first rib

84

66

60

70

Width across articulations of fourth rib...

72

59

58

65

SKULL.

1 P. perspicillaius,
male
(U. S. N. M.,
No. 19417).

P. earbo,
male

(U. S. N. M.,
No. 18851).

P. peniciUiitus
(U. S. N. M.,
No. 940).

P. urile,
male

(U. S. N. M.,
No. 12502).

mm,
Tip of mandible to occipital condyle 14816

mm.
135

mm.
141

mm,
116

Fronto-nasal hinge to articulation for occip-
ital style 69

61

62

55

Across anterior part of frontals 22

20

19

13

Across postorbital processes 39

31

32

25

Across squamosal processes. . 45

36

37

32

Across exoccipital processes... 40

33

33

26

" The sternum (No. 1941 7, U. S. N. M.) found with the present series of bones is
important, as its size indicates it to be that of a male, and shows the sternum previ-
ously described to have been that of a female, or possibly even that of a male of P.
urile. It is very much larger than any sternum of P. urile, and much larger even
than the large specimen of P. carbo, used for comparison.17 The present sternum is

13The measurements are in a straight line.

"Proc. U. S. Nat. Mus., XII., 1889, p. 88.

15 Estimated, owing to breakage.

1S Taken from rostrum of one bird and calvarium of another.

17 Proc. U. S. Nat. Mus., XII., 1889, pp. 88-94.

188 MEMOIRS OF THE CARNEGIE MUSEUM

thus in harmony with the other bones, and aids 'materially in emphasizing the
superior size of P. perspicillatus.

" The appended tables give the measurements of the cranium and sternum here
described, compared with the corresponding parts of other] species. The measure-
ments of the previously described sternum, ascribed to P. perspicillatus, are repeated
and an error of the first-given table corrected. The length from anterior end of
carina to end of meso xiphoid is said to be 104 mm., when it should have been
90 mm.

Unfortunately the skull of P. carbo now available is smaller than that of the
individual used as a term of comparison in the previous paper18 on Pallas'
Cormorant."

Additional information in regard to the osteology of the Phalacrocoracidss will
be found further on in the present memoir, as well as under the explanation of the
plates ai its close.

Osteology of the Pelecanidse.

(Plate XX VII., Figs. 31-41 ; Plate XXVIII., Fig. 42, and Plate XXX., Fig. 49.)

Sometime during the year 1864, the writer collected on Indian Cay of the
Bahama Banks, a fine adult male specimen of Pelecanus fuscus. From it I took the
skull, and have it before me at the present writing.

Measuring from the transverse cranio-facial groove we find the osseous superior
mandible in this specimen to be somewhat less than four times as long as the remain-
ing part of the skull. A vertical section made through the middle of the posterior
third of this mandible at right angles to its long axis gives an elliptical figure, with
the minor axis in the horizontal plane. The anterior two thirds has a sharp lateral
edge, while the extremity is armed with a powerful decurved hook. About half of
the fore part of this enormous beak is compressed from above downward, a com-
pression that is accompanied by a gradual widening of the bone to near the end,
where it slopes in toward the hook in the median line.

The maxillo-palatines constitute a great spongy mass that fills up a space anterior
to the rhinal chamber. They unite in the median line, are bounded above by the
premaxillary, below by the united palatines, while the anterior extremity of the
maxillary fuses with the mass at about its middle on either side. (Compare with
Figure 36, Plate XXVII.)

In form this maxillo-palatine mass is wedge-shaped, with the broad end anchy-
losed with the under side of the united nasal processes of the premaxillary.

18 Loc. cit.

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 189

Posteriorly its wall is composed of compact tissue, being at right angles to the
longitudinal axis of the skull. It slants from the under side of the cranio-facial
hinge to the anterior margin of a median foramen, seen just anterior to' the keel
which is formed by the union of the palatines behind.

This posterior maxillo-palatine wall has a cleft in its lower two thirds, while two
conical pits, placed side by side, lined with compact osseous tissue, occupy its upper
third. They have their bases opening in the rhinal chamber, and their apices are
pierced by the small subcircular nostrils, one in each conical passage.

The hinder half of the jugal bar is compressed from side to side, slightly dilated,
with its end crooked up, and in life simply bound to the upper and outer side of
the quadrate.

The body of a lacrymal fuses completely with the cranial elements above, its
upper surface assisting in forming the smooth superfices of the frontal region. This
is also the case in P. sharpei (PI. XXVII., Fig. 37). From this portion it sends down-
ward and slightly backward a descending process. This is composed of a cylindri-
cal pedicel for its upper third and an antero-posteriorly compressed portion for the
lower two thirds. It fails to reach the maxillary, its tip remaining free just above
that perpendicularly compressed bar which passes immediately beneath it.

The interorbital septum is entire, with the exception of a semicircular perfora-
tion, which is immediately in front of the aperture in the anterior wall of the brain-
case that gives egress to the optic nerves.

Each olfactory has a small foramen in either orbit at its usual site; the track for
the nerve being a broad, shallow groove beneath the orbital vault.

The mesethmoid is very deep ; its anterior border is sharp and thin. Commencing
in the aperture of the angle between the pterygoidal shafts, it is carried directly
upward and forward to the expanded portion beneath the roof of the cranio-facial
region, the edge meeting the median division of the maxillo-palatines. (For figures
illustrating the skull of the Brown Pelican see my memoir on the "Osteology of the
Tubinares and Steganopodes," Proceedings U. S. Nat. Museum, 1888, Vol. II.,
pp. 311-315.)

The lower fourth of this ethmoidal border is thickened and rounded for the
articulation of the palatine and pterygoidal heads.

Coming, as usual, from the anterior apex of the basi-temporal triangle, the other
portion of the rostrum is decurved and meets the point referred to above in the
angle between the pterygoids,

A quadrate is a very large bone with a broad, triangular process. Its mastoidal
head can hardly be said to be divided into two, as in most birds, for the division is

190 MEMOIRS OF THE CARNEGIE MUSEUM

so very slight and faintly marked; and a large pneumatic foramen is seen upon its
outer side — a very unusual place for this aperture.

Its mandibular foot is narrow antero-posteriorly and very wide transversely.
Two facets occupy its lower surface, separated from each other by a concave notch
which is deepest anteriorly.

The bone also presents a smooth articular surface for the quadrato-jugal at the
point above mentioned, while a large convex facet is offered to the pterygoidal cup
of the pterygoid of the corresponding side.

We find the external openings to the ear to be very small, and hid from sight
upon direct lateral view by the quadrate. A sphenotic process is well developed,
but the mastoidal one is simply a roughened line ; between the two is a wide crota-
phyte valley leading from the fossa of the same name, which is here small, incon-
spicuous, and entirely lateral.

The orbital cavity itself is thus seen to be deep and capacious, lacking bony walls
principally upon its inferior and anterior aspects. They are more complete in P.
sharpei (PI. XXVII., Fig. 37).

Upon its under side this skull presents a number of points of interest. The
anterior moiety of the superior mandible is here seen to be longitudinally grooved
by a broad and shallow furrow, which gradually becomes somewhat narrower as we
proceed backward, to finally merge into the convex, median portion of the hinder
half of this great rostrum. Along its median line it is marked by a few scattered,
slit-like foramina that lead into its shallow interior, which latter is largely filled with
an open mass of spongy, osseous tissue, continuous with the maxillo-palatines behind.

The palatine bodies, including their heads, fuse together for their entire extent
in the median plane. Resulting from this union we have a single, descending,
median carination, composed of the united inner keels of the palatine bodies and a
similar superior median one composed of the ascending processes of the same.

The latter is truncated just before reaching the maxillo-palatine bodies.

This skull lacks basi-pterygoid processes, while the pterygoids themselves are
short, thick set bones, with large anterior and posterior heads, and sharpened longi-
tudinal crests on the superior aspects of their shafts.

The basi-temporal triangle is small and its area concave. A thin, pointed lip of
bone eaves over the entrance to the Eustachian tubes, which are here apparently
thoroughly surrounded by bony walls.

We find the foramen magnum situated at the bottom of a broad, deep, and
transverse concavity. This excavation is bounded on either side by the dome-like
mastoid prominences, in front by the line of the base of the basi-temporal triangle,

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 191

and behind by a low, smooth ridge which arches between its lateral boundaries.
The foramen magnum cannot be seen on a direct vertical view of the base of the
skull (PL XXVII., Figs. 35 and 36).

The occipital condyle is rather large, ellipsoidal in form, and placed transversely,
while the outline of the foramen is also a broad ellipse, but with its long axis placed
just the other way. The plane passing through its periphery makes an angle with
the plane of the basis cranii of about sixty degrees.

Regarding this skull from a superior aspect we are to note the small, subcircular
openings to the nostrils, situated a little beyond the irregular line marking the
cranio-facial hinge (PI. XXVII., Fig. 40).

Their centers are about 2 mm. apart, and each one is situated at the posterior
end of a groove. These grooves extend the entire length of the superior mandible,
passing out on either side of the hook at its anterior extremity. At first each is
rather on the lateral aspect of the bone, but beyond the posterior half they gradu-
ally converge and get on top, to include between them the prominent convex cul-
men. Just before reaching the hook, however, the included surface becomes Mat
and depressed, when the lines terminate, as pointed out above.

The top of the skull in this Pelican is very flat for the frontal region, being
simply curved downward at the outer borders. As we proceed backward to the
parietal region, however, it gradually becomes more convex and dome-like, though
still retaining its absolutely smooth and polished character. This latter may also
be seen from a posterior aspect, and below it the high, arching, and equally smooth
occipital area. This latter extends down on either side over the enormous mas-
toidal elevations of this bird. We also notice that from this view we may see
directly into the foramen magnum ; the entire pterygoids are in sight, and the quad-
rates come down far below the basi-cranial plane.

The mandible from the skeleton of this Pelican is represented by a long, narrow
loop of bone, which is strikingly devoid of prominent characters. Its symphysis is
very short and decurved, being slightly excavated on its superior aspect behind (PL
XXVII., Fig. 41).

The upper and lower margins of either ramus are rounded for their entire length,
while the sides included between them become gradually narrower as we proceed in
the direction of the symphysis. These are smooth both internally and externally
and both concave in the vertical direction.

Rather more than the posterior moiety of each ramus is hollow for the admis-
sion of air, and each presents two foramina, which seem to be intended for that pur-
pose. One of these is on the inner and upper aspect of the ramal shaft, just beyond

192 MEMOIRS OF THE CARNEGIE MUSEUM

a concavity that occurs immediately anterior to the articular cup. The other, ellip-
tical in form, is on the inner and lower aspect, and about 2 cm. beyond it.

Each articular cup presents two concavities — a central one and another occupy-
ing the inturned process of this extremity. Both have pneumatic foramina at their
bases. The mandibular angle behind is truncate and much compressed in the per-
pendicular direction. The under surface of one of these ends is perfectly smooth
and gradually merges into the inner and outer surface of the ramal shaft. Almost
complete disappearance of the coronoids has taken place.

Almost the entire skull of this bird is highly pneumatic, and I have carefully
compared it with the specimen loaned me by the U. S. National Museum, and find
that they agree in all essential particulars. In the latter I find a mandible 38 cm.
long, Avhile the symphysis of the same only measures 4 mm. The superior carina-
tion of the united palatines is somewhat higher, especially in front than in my own
specimen, but I am inclined to believe that in mine the free margin has somewhat
broken off, and due allowance should be made for it in the figure illustrating the
skull of this bird in my earlier memoir (Fig. 40) showing the lateral view of the
Pelican's skull, when the supero-median palatal keel is drawn a little too low.

Fig. 20 of Huxley's 1867 paper in the P. Z. S. is a good representation of the pos-
terior half of the skull of a pelican (P. onocrotalus) seen upon its basal aspect. The
side view of the same is very indifferently drawn.

Some few of the small bones of the disarticulated National Museum specimen
at my hand may have been lost, but none of any consequence. Almost the entire
skeleton is pneumatic, and exceedingly light. The pygostyle is as light as a wafer,
and taken with all the caudal vertebne, only weighs a few grains. Among the non-
pneumatic bones we note the atlas, the fibulae, the first metatarsal, and all the toe-
joints. Some of the bones are actually riddled with air-holes, and such a long bone
as the humerus has many minute foramina at very various localities at both of its
extremities. The pelvis weighs no more than were it made of cork, and so it is with
all the other large bones.

Of the Remainder of the Axial Skeleton in Pelecanus.

(See Plate XXVII., Figs. 31-34, 38, 39; PI. XXVIIL, Fig. 42, and PI. XXX.,

Fig. 49.)

Professor Mivart found in the spinal column of a specimen of Pelecanus mitratus in
the British Museum but "forty vertebne in all, there being but three lumbar and five
caudal" ; and in a specimen of P. onocrotalus (No. 527A), forty one vertebrse, there

SHUFELDT : OSTEOLOGY OF THE STKGANOPODES 193

being but three sacro-caudals. He says, however, that in this species the general
number seems to be forty-two vertebne in all, not counting the pygostyle.

In a complete skeleton of Pelecanus fuscus before me (No. 18483), I find sixteen
true cervical vertebrae, none of which support true ribs. The seventeenth bears a
pair of large, well-developed ribs which do not connect with the sternum, and which
have long epipleural appendages low down on their shafts, where they are anchy-
losed. The eighteenth vertebra is the ultimate one of the dorso-cervical series of
the spinal column that is free. Its ribs are broad and flat, with their appendages'
coossified to the shafts at an angle of 45°, and inclined to be slender, narrow, and
pointed.

The nineteenth, twentieth, and twenty-first vertebra; are not only all fused
together, but are similarly joined with the bones forming the pelvic "sacrum."
This union is most perfect, and is considerably fortified by the complete coossification
of the metapophyses extending between the extremities of the transverse processes,
and then with anterior iliac borders, in which location the ossifying process has been
extended quite across, thus, on superior aspect, shutting the twenty-first vertebra
out from view.

These three vertebrte also bear broad and flat ribs, connecting with the sternum
as do all the thoracic ribs in this Pelican. This pair belonging to the nineteenth
vertebra are quite like those described above for the eighteenth ; they are broader,
however, which is also the case with the pair suspended from the twentieth vertebra.
These last have the epipleural spines reduced in size, while the last two pairs of
ribs, or those belonging to the twenty-first vertebra, and the pair of "pelvic ribs"
lack these appendages entirely.

Almost completely pneumatic in character, the leading costal ribs are rather short,
but the series gradually increases in length as we pass backwards, all of them being
more or less pneumatic, and all being remarkable for having their extremities
enlarged, for the purpose of affording a greater articulatory surface at those points.
For the most part, the pneumatic foramina in these hremapophyses are to be found
at their sternal ends.

The free pelvic vertebra? all have a swelled appearance, while their surfaces are
as a rule smooth, their salient angles much rounded off, and their processes quite
subordinated.

The cup of the atlas is deeply notched above, as is the articular surface of the
body behind, concaved upon the same aspect. This surface is of an elongo-reniform
outline, its transverse diameter being of the greatest length. A small hypopophysial
tip extends backwards from its mid-point below.

194 MEMOIRS OF THE CARNEGIE MUSEUM

Not especially stout, the lateral pedicels support a broad, oblong neural arch,
which has a straight anterior, and a concave posterior margin. Air gains access to
this vertebra through some holes situated in or near a small concavity, over which
articulates the ventral surface of the odontoid process of the axis when the bones
are in situ. A polishing of the surface alone indicates the points where the ante-
rior margin of the neural arch comes in contact with the occipital bone of the
cranium, and where the ventral aspect of its posterior margin rides on the upper
part of the anterior margin of the neural canal of the axis.

A broad-based, swollen, and tuberous neural spine is one of the chief characters
of the last-named vertebra, while its bpemal spine, situated posteriorly on the centrum,
is transversely compressed, sharp along its infero-mesial border, and withal rather
inconspicuous.

The articular surface on the body for the atlas is deeply concave, and always
presents a large mesial pneumatic foramen at its base, just beneath the odontoid
process. Posteriorly, the neural arch of this vertebra far overhangs the centrum
below, while either postzygapophysial portion of it, not strictly defined from the
bone, has upon its direct ventral aspect a large oval articular facet for articulation
with the corresponding ones at the prezygapophyses of the third vertebra.

We have no lateral canals in the axis vertebra of this Pelican, but a small pneu-
matic foramen is usually seen on either side, at the site of their occurrence in the
next bone of the spine behind it, at points where they open posteriorly. In the
third, fourth, fifth and sixth vertebra the neural spine is represented by a low,
lengthy, and thickened ridge, which fails in any case to reach the dorsal margins of
the neural canal, either in front or behind. In the seventh vertebra the neural
spine is much restricted, occupying only the center of the bone. It begins to dis-
appear in the eighth ; is almost absent in the ninth and tenth ; feebly reappears in
the eleventh and twelth ; begins to be tuberous in the thirteenth ; and from this
on it slowly assumes the form it has in the dorsal series, where it is low, thickened,
and wedge-shaped, with base behind and apex in front.

Far back on the centrum of the third vertebra we find a short, thick, and low
heemal process ; somewhat resembling the one on the axis. The character is quite
obliterated in the fourth vertebra, where we begin to see, anteriorly beneath the
extreme fore part of the centrum, the first indications of the formation of the carotid
passage. They are more distinct in the fifth vertebra ; still more so in the sixth ;
decidedly so in the seventh, where the remarkable peculiarity exists in that a carotid
channel with thin, conspicuous side-walls abruptly develops at the extreme hinder
end of the centrum. This character is just as well marked in the eighth vertebra,

SHTJFELDT : OSTEOLOGY OF THE STEGANOPODES 195

after which it gradually disappears (never to form a completed tube), to be gone
entirely in the thirteenth cervical.

In the eighth vertebra the anterior part of the carotid passage is a solid, closed
tube, its parapophysial walls being thick and strong, the passage of considerable
caliber, and the whole occupying more than the anterior third of the ventral aspect
of the bone. It also shows, ventrad, a mid-longitudinal ridge, with a longitudinal
gutter upon either side of it. This carotid canal remains a closed tube to include
the fourteenth vertebra, slowly changing in character as we near the dorsal series.
In the fifteenth to the seventeenth inclusive, its place is taken by a well-marked
haemal spine — a character entirely absent in the eighteenth vertebra, which is a true
dorsal one.

No evidence of a haemal spine whatever exists in the three fused dorsal vertebrae,
and in them the neural spine is but represented by a coossified, flattened ridge,
hardly at all higher than the fused metapophysial track at the ends of the dia-
pophyses, on either hand. Lateral canals begin in the third vertebra and persist to
the sixteenth inclusive. They are long and of small caliber in the third, fourth,
fifth and sixth, materially shortened in the seventh ; somewhat more so in the
eighth, where, for the first time, their anterior opening looks directly dorsad, which
latter continues to be the case in the ninth, tenth, eleventh, twelfth, thirteenth
and fourteenth, the canals increasing in caliber, but at the same time shortening in
length as we pass through the series just named. In the fifteenth and sixteenth
vertebrae, the lateral canals are parallel with the neural canal, about one third
smaller in size, their ventral floors being formed by the pleura pophysial elements.
For the most part the comparatively rather small neural canal is subcylindrical in
form, being most compressed vertically as we pass towards the distal cervicals, to
become more cylindrical again in the ultimate free dorsal. After passing the fourth
cervical, and from thence on to include the sixteenth, this tube is markedly of
greater caliber behind than it is in front, in any single vertebra of this part of the
spinal column.

This character can be particularly well seen in the eleventh cervical. Through-
out, the parapophysial styles are very short. On the third vertebra they are far
apart, but gradually approach each other and the median line to include the sixth
cervical. Their character changes in the seventh, where they become sharp, short,
and spiculiform. In the eighth, ninth and tenth they are represented by mere
spinelets, close together mesiad and are to be found on the postero-ventral border
of the carotid canal. From this on they gradually separate again, and entirely
disappear from the ultimate cervical vertebrae.

196 MEMOIRS OF THE CARNEGIE MUSEUM

The prezygapophysial facets are long, narrow and ellipsoidal in form on the
third cervical, being considerably nearer each other at their posterior ends than they
are in front. They face almost directly dorsad. As we pass to include the seventh
cervical these facets gradually become more and more circular in outline, and incline
to face towards the median plane. This they do quite abruptly in the eighth
cervical — distinctly so in the ninth, look slightly backwards in the tenth, but
again thereafter gradually right themselves so as to once more face the median
plane. Here it will be seen we have the same arrangement noted for the Cormor-
ant and the Anhinga though not nearly so well marked. The postzygapophyses
are short and stumpy in the fourth vertebra, but very gradually lengthen to include
the twelfth cervical, after which they once more shorten, to become short and
thick-set again in the last cervicals, and the free dorsal, in which last — as distinct
processes — they are practically aborted. Their facets for articulation with the cor-
responding ones on the prezygapophyses, are upon their ventral aspects, and of
course have an exact counter-facing to them.

These vertebras commence to lengthen after passing the third cervical, becoming
quite longish and narrow in mid-series — they then very gradually assume the
broad, massive, and short type found at the end of the cervical chain and in the
free dorsal. Though these latter are large they agree with all the others of the
column, from their extreme pneumaticity, in being very light in weight.

The three last dorsal vertebrae coossify, or rather in the adult Pelican are fused
with those of the pelvic sacrum, and the leading vertebra of the latter is quite
similar to one of the last dorsals, but in the next three the parapophysial processes
are thrown up against and completely fused with the ventral surfaces of the ilia
upon either side. These processes are spine-like and nearly aborted in the next
following one, while in the next three they are absent entirely. This is well shown
in the pelvis of Pelecanus sharpei (PI. XXX., Fig. 49).

Then comes a true sacral vertebra, which has them long and well-developed,
reaching out to points posterior to the acetabute where they fuse and broaden, and
otherwise develop bone which greatly strengthens the pelvis in all this vicinity.
The five remaining vertebrae of the "pelvic sacrum" also possess parapophysial proc-
esses, which with the diapophyses above them, most completely coossify with the
internal borders of the ilia upon either hand. Of good capacity, the " pelvic basin "
is of a short, oblong form, being one third narrower than it is long, and of nearly
uniform depth throughout. Viewed laterally, we find the acetabulum to be large
and circular; its internal periphery coequal with its external one. Above it is a
fair-sized antitrochanter, facing forward and downward, and very slightly outward.

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 197

The obturator foramen merges completely with the obturator space, which latter is
large and broadly spindle-form in outline. No prepubis is present, and the post-
pubic style is curved and very slender until it comes in contact at a point on the
postero-ventral angle of the ischium, after which it is turned slightly mesiad and
somewhat though not at all greatly enlarged.

An extensive, subelliptical ischiadic foramen monopolizes much space on this lat-
eral aspect of the pelvis. Behind it the surface of the bone is smooth, and finally
we notice a broad, shallow notch on the posterior border, between ilium and ischium.
Dorsad, the pelvis is broad, smooth and flat, with a thorough fusing of all the bones
composing it. Parial foramina occur between the diapophyses of the uro-sacral ver-
tebras only in the post-acetabular region, where they are of good size.

The ultimate vertebra may be more or less individually outlined, and simulate
in form the leading one of the caudal series. In the pre-acetabular region the ilia
are moderately concaved, lie more or less in the horizontal plane, and fail to come
quite in contact with each other over the "sacral crista," which latter is here quite
low and indistinct, due to the thorough fusing of all the bones in this region. From
the extreme outer point of one antitrochanter to the corresponding point on the
same process on the other side measures, on a right line, about 7.5 cm., while the
length of the " pelvic sacrum" is about 12 cm.

Thus, though steganopodous in character, this pelvis of Pelccanus differs mate-
rially in form and aspect from the pelvis of either a Cormorant or an Anhinga.

In the skeleton of the tail we find six free caudal vertebrae and a large pygostyle.
All are highly pneumatic. The neural canal, of some considerable size, passes
through the entire series and well into the pygostyle — an unusual occurrence in
birds. All possess tuberous neural spines, but only the last three have haemal ones,
and these increase in size from the last caudal to include the fourth. They extend
almost directly forwards. The transverse processes of the first five vertebra* are
bent ventral-wise, more particularly in the second and third, while in the last one
they are nearly aborted. The articular facets on the centra are concaved anteriorly,
and nearly flat on the posterior aspects.

The large pygostyle is drawn out into a blunt point at its postero-superior angle.
Its dorsal margin and the hinder half or more of its ventral margin are sharp — the
remainder of the latter being thickened. The antero-inferior angle projects forwards,
corresponding to the haemal spine of the first vertebra absorbed to form this com-
pounded bone.

Thus it will be seen that in this species of Pelican we have 41 vertebrae in its
spinal column, namely : sixteen cervicals that are without free ribs; one cervical with

198 MEMOIRS OF THE CARNEGIE MUSEUM

well-developed ribs ; one free dorsal ; three dorsals fused with the pelvic sacrum ;
fourteen in the pelvic sacrum ; and six caudals ; the pygostyle not being included
in this count.

The sternum and os furcula are intimately fused together at the carinal angle,
forming there an extensive union. This is also the case in Pelecanus sharpei (PI.
XX VI 1 1., Fig. 42). Apart from this fact, the sternum has much about it to remind us
of that bone in Plialacrocorax. It is profoundly concave upon its dorsal aspect, and
correspondingly convex upon the opposite surface. Either costal border, trans-
versely very broad, is occupied nearly its entire length by the five hsemapopliysial
facets, with the long, shallow concavities that are between them. These latter are
riddled with small pneumatic foramina, while only a few of these openings are found
elsewhere in the sternum, and they are very minute. Two wide, shallow indenta
tions mark the xiphoidal margin, and the lateral xiphoidal processes are short with
rounded ends. They extend backwards and outwards. The deep keel projects
somewhat forwards, but is merged upon the surface of the bone behind long before
it reaches the posterior sternal margin.

Each "costal process" is low and of a subquadrilateral outline. On the anterior
border of the bone there is a notable concavity, longitudinally disposed and of some
considerable breadth. It extends from the upper termination of the anterior and
rather sharp margin of the keel to lead into the general concavity of the dorsal as-
pect of the sternal body. It separates the long, narrow costal grooves in front, as
well as the rather deep costal facets above them.

U-shaped in pattern, the lower loop of the os furcula is not particularly stout,
while the chief feature of this bone is its enormously swelled and enlarged free
upper extremities (PI. XXVIII., Fig. 42).

These are entirely hollow, and each is provided with an elongated pneumatic
foramen upon its postero-mesial aspect. Postero-laterally, either one presents for
examination a subcircular facet for articulation with the head of the corresponding
coracoid, internal to which a stumpy, tuberous projection extends backwards. But
this fails by an extensive interval to meet the scapula, when the bones of the
shoulder-girdle are articulated as in life. The mesial and outer aspects of these
great clavicular heads are flat and smooth ; anteriorly, their thickened borders are
convex from above downwards, the posterior borders being concave. On top they
each present a flat, triangular surface with the apex to the front.

From its extreme pneumaticity I find a coracoid in Pelecanus fuscus to be a very
light bone, while it has an altitude of 9.7 cm., being large in its other proportions.
The sternal end is considerably dilated, and to the other side is a very small up-

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 199

turned epicoracoidal process. On the posterior side at the mesial angle there is a
prominent jutting flange bearing an articular facet upon its entire under surface for
articulation with the sternum. The inner two thirds of the inferior border is also
for articulation with the same bone. The head is more or less massive, being ob-
liquely compressed above. Well below this, anteriorly, is a non-elevated, circular
facet for the os furcula. Above the large glenoid concavity, between it and the
summit of the bone, is a deep valley passing downwards and forwards from the pos-
terior aspect. Pneumatic foramina occur in it, as they do also on the mesial aspect
of the shaft in an elongated group below the coracoidal head. The facet for the
articulation with the scapula is a deep circumscribed concavity, and in front of it a
conspicuous scapular process is developed, but it never reaches the os furcula in
articulation when the bones are in situ. Below the open valley for the tendons,
the shaft, is pierced by a foramen, a character often seen in birds of other groups.

A scapula is small in proportion for the size of the bird, thick, narrow and
short. It is bent neither right nor left, but moderately curved, its convexity being
along its dorsal aspect. It has a distinct ellipsoidal raised facet for articulation with
the above-described pit in the coracoid, while the surface it offers for the glenoid
cavity is small. Its most striking character is the very much elongated process at its
antero-mesial angle. This acromial process does, not meet the clavicle in articula-
tion. On the dorsal aspect of the blade, near its middle, is a distinct tubercle for
muscular attachment, that is quite noticeable.

Of the Appendicular Skeleton.
(See Plate XXVIL, Figs. 31-34, 38 and 39.)

In the pectoral limb the humerus is a very long, large bone showing the usual
double sigrnoidal curvature in its shaft, which for its middle third is subcylindrical
in form. (See PL XXVIL, Fig. 34.)

The radial crest is but moderately developed, while a peculiar tuberous enlarge-
ment is found on the ulnar border of the bone, just distad to the pneumatic cavity.
It is stopped abruptly immediately before we come to the shaft, by a distinct
though very narrow notch. The ulnar tuberosity is not very prominent, only parti-
ally overhanging the pneumatic fossa. Between it and the true humeral head is a
distinct notch or valley. At the distal end of the shaft, the oblique and radial
tubercles stand out with considerable prominence ; the olecranon fossa is pretty well
marked, otherwise we find only the usual ornithic characters here. The humerus
has an extreme length of about 27 centimeters.

200 MEMOIRS OF THE CARNEGIE MUSEUM

The ulna of the forearm is even much longer than this, being fully 32.5 centi-
meters in length, and from its great pneumaticity a very light bone. At its proxi-
mal extremity it has a very large pneumatic fossa just below the facet for articulation
with the radius. The olecranon process is barely at all developed. Along the
moderately curved shaft there are at least twenty-two distinct papillae, nearly equi-
distant from each other, for the insertion of the quill-butts of the secondary feathers
of the wing. On section, at its middle third, the shaft is triangular, and its distal
end presents us with nothing peculiar. (The ulna of P. sharpei is shown in PI.
XXVIL, Fig. 31.)

The radius is a long, slender bone and is also somewhat curved. Its distal end
is transversely expanded, below, which expansion is the pneumatic fossa,

The radiale and ulnare of the carpus both plainly show the pneumatic fossa
leading into them, as do all the bones of the manus. The carpo-metacarpus has
an extreme length of about 12.4 centimeters. Along its continuity it is slightly
carved anconad, while the very slender medius metacarpal stands well away
from the main shaft, being somewhat longer than it is at the point of anchy-
losis distally. The pollex metacarpal is short and bulky. The carpo-metacarpus of
P. sharpei exhibits the usual characters of this bone among the Pelicans (PI.
XXVIL, Fig. 39).

The proximal phalanx of the index digit, with a length of 5 centimeters has
a broad expansion posteriorly, and is peculiar in being perfectly flat and smooth
upon its anconal side, while upon the palmar aspect it is divided into two
deep concavities which are absolutely riddled with pneumatic foramina at their
bases.

The terminal joint of this finger is long and trihedral in form, with its pneu-
matic foramen at the proximal end. Though a little longer than the pollex
phalanx, the latter has very much the same shape, and is likewise pneumatic. The
free terminal joint of the medius digit has a broadish, triangular expansion behind,
and the holes for the admission of air pierce it upon both sides. This bone is 3.4
centimeters long, by 1.3 broad at its broadest part ; while the distal joint of index
has an extreme length of 4.6 centimeters.

In the pelvic limb, a femur although pneumatic to a certain degree, it is not so
markedly so, as either the tibio-tarsus or the tarso-metatarsus. It will, in the dried
skeleton, probably weigh more than the former, notwithstanding it is not as large,
nor so long. Its stout shaft is but very little bowed in the antero-posterior direction,
and the extremities are large and massive. The broad trochanter does not rise above
the articular summit of the bone proximally, and the pit for the round ligament on

SHUFELDT : OSTEOLOGY OF THE STEOANOPODES 201

the caput femoris is but little excavated. Distally, the condyles, • appear; 'to' -bp
somewhat antero-posteriorly compressed, and transversely spread apart from each
other. The external condyle is slightly the lower and larger of the two, and
it is deeply cleft behind for the head of the fibula. Anteriorly, the rotular chan-
nel is shallow, nor is the popliteal fossa on the other side of the bone as deep in
proportion as we find it in many other birds, both large and small, (PI. XXVII.,
Fig. 32).

A tibio-tarsus instead of being bowed in the antero-posterior direction as it often
is, it curves the other way so that the fibular border of the bone is con vexed, and
the opposite one, correspondingly, or even rather more, concaved. The extremities
are large, but the several cnemial processes at the proximal one are but very moder-
ately developed. Low and long, the fibular ridge extends down half the length of
the shaft, and its articular margin or border is roughened. The tibio-tarsus of P.
sharpei exhibits these characters very well, as they are seen in the Pelecanidse
generally (PI. XXVII., Fig. 33).

At the distal end of the tibio-tarsus, the large condyles of the usual reniform out-
line, protrude prominently to the front, while behind they almost immediately
merge into the general surface of the lower end of the shaft. The intercondylar
interval is deeply excavated between them in front ; at;their lateral aspects the points
for tendinal and ligamentous insertion are roughened and distinctly defined. A
small bony bridge, such a frequent ornithic character, spans obliquely the anterior,
longitudinal channel for the passage of tendons. In the deepest part of this recess,
and directly under this little osseous bridge, a large pneumatic foramen enters the
shaft of the bone. It is the principal aperture of this kind, by far, in the bone, and
exists in the same place in the shaft of the fellow of the opposite side.

A fibula has its head and articular surface thereon moderately well developed,
but the bone contracts rapidly as we proceed in the direction of its distal end, and
after passing the articular surface for the tibia, it comes to be little more than a
stout, bony thread, that by np means reaches the condyle of the tibio-tarsus, nor
does it anchylose with its distal end.

Pelicans have, comparatively speaking, a very small patella, which is roughly
wedge-shaped in form, with the flat base below, and the superior edge rounded, and
not especially sharp.

The torso-metatarsus is a stout, straight bone, with pronounced characteristics. The
intercondyloid process upon its summit is rounded and conspicuous. Immediately
below it, in front, terminates the usual longitudinal channel found on the anterior
aspect of the shaft. It is deep here and opens into the shaft by a double pneu-

202 ; : ; t ' MEMOIRS OF THE CARNEGIE MUSEUM

; ; for'anjen J- The inner one of these is large, and not only communicates with
ine interior of the shaft but passes also directly through it — appearing to the inner
side of the base of the hypotarsus as a large foramen. Two or three much smaller
foramina appear upon the opposite side of the hypotarsus, and they lead to similar
passages through the shaft.

A prominent feature of this bone is its large hypotarsal process. This is com-
posed of a thick, oblong piece or internal portion, which is capped upon its posterior
aspect by an elliptical, vertical cap, having slightly protruding margins.

Two very much smaller such plates are developed external to this larger one,
having each only about one half its altitude, and being not more than one fourth as
thick. By their posterior surfaces more or less ossifying across, two vertical tubes
are formed, an internal and external one, through which the tendons pass in life.
As thus formed, this hypotarsus stands out rather abruptly and perpendicularly
from the shaft, with hardly any inclination to merge with it at its lower part, as is
the case in a good many birds. As to the shaft of this bone, it is convex posteriorly
and strongly marked by the raised muscular lines ; it is concaved slightly in front,
the excavation being very deep proximally ; shallow in the middle third ; and deep-
ening a little again as it passes into the usual foramen between and above the external
and middle trochlear processes. These latter are large and normally disposed, the
mid one being the lower of the three, while the internal and external project about
equally backwards. The facet for the free first metatarsal is concave and of some
considerable size. The os metatarsale accessorium is a little over two centimeters long,
presenting the usual head and being slightly twisted upon itself distad. It agrees
with the toe-joints in being non-pneumatic. These latter are nearly all straight,
only some of the distal ones being slightly curved. For a bird as large as a Pelican
they indicate rather a feeble foot, and this is further sustained when we come to see
the weak terminal joints, which are ti'ansversely somewhat compressed, not very
long, but withal distinctly, in fact rather strongly curved. These phalangeal joints
are arranged upon the plan of 2, 3, 4, and 5 bones to the hallux, second, third and
fourth toes respectively.

Pelecanus fuscas has a femur 8.5 cm. long; a tibio-tarsus of 12.5 cm.; a tarso-
metatarsus of 7.8 cm. and a mid-anterior basal phalanx of pes of about 3.9 cm. in
length.

When one comes to think of it then, a Pelican has rather a peculiarly balanced
skeleton. Its lower jaw is considerably longer than its ulna, and the latter is in turn
considerably longer than the humerus. This last-named bone is only a little more
than half as long as the carpo-metacarpus of manus. The tibio-tarsus and the carpo-

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 203

metacarpus are almost exactly of a length, the difference being only a millimeter in
favor of the former.19

Further information in regard to the osteology of the Pelecanidee will be given
later on in the present memoir, as well as in the explanation of plates at its close.
In Volume I. of the Hand-List of Birds by Dr. R. Bondler Sharpe (1899, pp. 238,
239), I find that he recognizes ten species of existing Pelicans, and seven fossil
ones. They are found in nearly all parts of the world, including Australia, Tas-
mania and New Guinea. It is not likely that they differ very widely in their
osteology, and it is probable that in its main features it is very well exemplified in
Pelecanus fuscus as has just been set forth above.

Existing Pelicans have by ornithologists generally, all been restricted to the one
genus Pelecanus of the family Pelecanidas, where they undoubtedly belong.

On the Skeleton of Fregata.

(See Plate XXIX., Figs. 45-48 ; PI. XXX., Figs. 50, 51.)

Of all the steganopodous birds perhaps no one of them exceed Fregata aquila, in
point of interest, in so far as its osteology is concerned. In not a few particulars it
has a very remarkable skeleton, while in others it would appear to indicate that the
form or species is a more or less generalized one. For example, both superficially
and otherwise, the skull of Fregata resembles, in not a few respects, the skull in some
species of Albatrosses (Diomedeidse). This not only applies to the lower jaw, where
the similarity is very evident, but also to a number of characters in the cranium
and face. The long, powerfully hooked superior mandibles are a good deal alike,
as are the maxillo-palatines. Fregata has a vomer that approaches that bone in the
Albatrosses ; its palatines are not far off, and even still less so its pterygoids and
quadrates. The lacrymals are upon the same plan of structure, and the entire

19 There are some specimens of embryos and subadults both of Pelecanus and Cormorants in the U. S. National
Museum collections, and at my request Mr. Lucas has examined some of these for me, and writes the following letter
on the subject, for all of which it gives me pleasure to thank him.

DEAR DK. SHUFELDT :

I have examined Pdecanus sp. about three or four days old and Phalacrocorax carunculatus for supramaxillary and
found no trace in either. Pelecanus seems to mature more rapidly than Phalacrocorax. In tl)e young Pelican the lacry-
mal is well developed and free, and there is no trace of the partial hinge joint at base of bill. There are traces of the
three fused metatarsals and the calcaneum (?) is still free. In the young Cormorant the nostrils are still opeu, the
lachrymal free. The occipital style is represented by ligament and were we all ignorant of its existence, it night readily
be overlooked. The hyoid is large. It would seem then that the supramaxillary does not appear until late in life and
it may have no morphological meaning ; simply it ossifies at the time of closing of nostrils. The occipital style of 7*.
carunculatus is small in the adult. The specimen was about one third grown, and about ten days old probably.

Sincerely yours,

FREDERIC A. LUCAS.

204 MEMOIRS OF THE CARNEGIE MUSEUM

cranium proper in the Man-o'-war-Bird might well answer for that of an Albatross
but slightly removed from the typical stock. Fregata, however, lacks the deep supra-
orbital glandular fossae so characteristic of the Diomedeidze, and, from above down-
wards, the skull is somewhat more compressed than it is in, for example, such a
species as the Short-tailed Albatross (D. albatrus) (PI. XXIX., Fig. 46).

The superior mandible is broad at its base, and tapers gradually forwards, to be
armed at the apex with a powerful and decurved hook which is exceedingly sharp
at the point. This mandible is convex from side to side, moderately compressed,
and concave longitudinally along the mid-line of the culmen, from the cranium to
the base of the apical hook anteriorly. The narial apertures are small and situate
each at the bottom of a fossa-like depression, while from them running directly
forwards, one upon either side, is a distinct groove. It is rather deep, and very
narrow, not being carried upon the hooked part of the bill at its distal termination.

No distinct cranio-facial hinge exists, but at the middle point in that region
the proximal ends of the premaxillaries remain unobliterated throughout life (PI.
XXIX., Fig. 46).

Deeply concave towards each other, and with sharpened margins, the superior
borders of the orbits are separated from each other by an interval of something
more than 2 cm. Mesially, the intervening frontal surface shows a slight eleva-
tion, while in the parietal or post-frontal region there are two larger, convex and
elongated elevations placed side by side with a median, not at all deep, depres-
sion passing between them. Within the cranial casket these elevations harbor the
superior surfaces of the cerebral hemispheres, and the skull is thinner there than
elsewhere (PI. XXIX., Fig. 46).

The crotaphyte fossse are well-defined, even better so than in an Albatross (D.
albatrus), being very slightly depressed below the general surface, and, in the mid-
dle line posteriorly, do not meet by an interval of a centimeter (PI. XXIX., Fig. 45).

Apart from this character, and the fact that the squamosal processes are more
prominent and sharper in Fregata, the posterior aspect of its cranium agrees almost
exactly, character for character, with what is presented to us on that view in the
cranium of D. albatrus. In this I do not exclude the pterygoids and quadrates.

The large subvertical, subcylindrical foramen magnum is partially overarched
by the well-defined occipital area, and the condyle, although of the same shape, is
not quite as prominent as it is in an Albatross (PL XXIX., Fig. 48).

Seen upon side view, we are to note the deep valley of the temporal fossa, made
so by the large outstanding post-frontal process anteriorly and the sharp, ridge-like
squamosal projection referred to above (PI. XXIX., Fig. 45).

SHUFELDT : OSTEOLOGY OF THE STKGANOPODES 205

The lacrymal is a free bone articulating by its upper portion with the free border
formed by the frontal and nasal bones, while its descending limb is bulbous at its
lower extremity, and fails to reach quite the maxillary below it. The pars plana
is nearly aborted, there being merely a small bridge of bone left, that arches over
the anterior end of the nasal nerve as it enters the rhinal chamber. This nerve
passes in an open groove, but is again shielded by an osseous span, just as it enters
the orbit. The "foramen rotundum" is large and is distinct from a still more exten-
sive vacuity which is seen above it on the anterior wall of the cranial casket. No
foramina exist in the interorbital septum proper. Occasionally a small distinct
nerve foramen is seen to the outer side of the opening for the exit of the optic
nerve. The anterior border of the mesethmoid is both concaved and sharpened
(PL XXIX., Fig. 45).

The zygomatic bar is quite straight, being transversely compressed at its quadrato-
jugal end, to become twisted upon itself in the jugal part, and so vertically flattened
for the remainder of its extent or in the maxillary moiety. The extreme anterior
end fuses in between the premaxillary and nasal. At the hinder extremity we find
the usual little peg articulating in the pitlet on the outstanding process of the
quadrate (PI. XXIX., Figs. 45 and 48).

As has already been remarked above, one of these last-named bones bears a
very close resemblance to the quadrate, in an albatross. The orbital process
of the os quadratum in Fregata is antero-posteriorly compressed just as it is
in Diomedea, and its free extremity is expanded and finished off in the same
manner.

We notice a slight difference, however, in the form of the internal articular
facet of the mandibular end of the bone. It is more compressed from before, back-
wards in Fregata, and the articular surface presents but one common convexity,
whereas in Diomedea this facet is impressed by a deep, oblique valley, the axis of
which is parallel to the pterygoid of the same side.

As in the Albatross, a pterygoid bone is a straight, stout element, with some-
what enlarged anterior end, and a markedly expanded posterior one ; in the last
character being much more enlarged than it is in Diomedea. Both birds have their
pterygoids pneumatic, and when articulated in situ, they touch each other in the
middle line, beneath the sphenoidal rostrum.

In Fregata the basitemporal region is small and of a triangular outline ; the
eustachian passages are very open canals, and their anterior beginnings are separated
by quite an interval — the base of the sphenoid standing distinctly between them
(PI. XXIX., Fig. 48).

206 MEMOIRS OF THE CARNEGIE MUSEUM

Although morphologically very much alike, the palatines in Fregata fuse together
where they come in contact with each other posteriorly, which is not the case in any
Albatross ever examined by me. The ascending and descending laminae are not so
powerfully developed as they are in the Albatrosses, but in Fregata we still find the
broad, horizontally flattened anterior moieties of the palatines, with the mesial,
elongated and narrow interval separating them. At the fore part of this we like-
wise see the maxillo-palatines, similarly separated, with their external borders and
surfaces accurately fused with the contiguous bones. Between them, we discover
the sharp apex of the extremely slender, and long, and free, and curved vomer of the
Man-o'-war Bird, so very suggestive of being an extreme modification of that element
as we found it to exist in Diomedea. (See Figs. 16 and 17, in my memoir on the Oste-
ology of the Tubinares and Steganopodes, P. U. S. N. M., 1888, p. 279, for the vomer
in an Albatross.)

The postero-external angle of a palatine in Fregata is inclined to be somewhat
produced, whereas in the Diomedeidse as a rule that angle is rounded off (PI.
XXIX., Fig. 48).

Upon examining the mandible in Fregata it also presents a number of characters
it has in common with that bone in Diomedea. In form it is a long V-shaped struc-
ture, with truncated postero-articular ends. The rami are strong and thick; the
ramal vacuities are closed in; the rather .meager symphysisis decurved and sharply
pointed at the apex. Along the superior margin of either dentary portion, runs a
distinct groove ; these borders being cultrate in the Albatross, and the aforesaid
groove being conducted down upon the mesial aspect of the dentary element, where
it is far less pronounced, and its nature not so marked. The mesial aspect of either
dentary part of the jaw in Fregata is also deeply grooved for nearly its entire length ;
this last groove is but faintly developed in Diomedea, and that only at the anterior
third of the bone (PL XXIX., Fig. 47).

Fregata has a more pneumatic mandible than Diomedea, and the foramen at
either articular end is, in the former genus on top of the mesial process, while in the
latter it is situated at the base of the obliquely-disposed, deep, central concavity, or,
in other words, that concavity which is intended to accommodate the inner large
articular facette on the quadrate. It is single and circular in either case.

Fregata has a very simple skeleton as to its hyoidean apparatus, for the fore part is
considerably aborted, while behind all the elements do not ossify. The small glosso-
hyal is performed entirely in elementary cartilage, and only the diminutive cerato-
hyals, which are distinct from each other, ossify. The short, rather bulky, first
basibranchial ossifies, but there is no sign of a second one. Of the thyrohyal ele-

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 207

ments, the ceratobranchials are long, slightly curved, and completely ossified ; the
epibranchials are in cartilage only. In my specimen the sderotals of the eye, and
the bony elements of the internal ear, are missing.

Of the Remainder of the Axial Skeleton.

Upon examining the vertebral chain of Fregata aquila we find in it thirteen
vertebrae in the cervical region, that do not support free ribs ; the fourteenth
vertebra has a long, slender pair that are without unciform processes, the next five,
all freely movable upon each other, connect with the sternum by means of their
costal ribs or haemapophyses ; then, finally, there are two pairs of pelvic ribs, but
it is only the costal ribs of the leading or anterior pair, that connect with the
sternum. In the specimen before us, we may call them, from the fifteenth to the nine-
teenth vertebra inclusive, true dorsals, while counting, as best we may, the number
of vertebrae fused together in the pelvic sacrum, there appear to be fourteen in all.
Next follow six free caudals and a large pygostyle.

Since the first part of this memoir was written I have had placed before me, a
skeleton in the rough of the Great Wandering Albatross (D:exulans), and upon com-
paring the arrangement and number of its vertebrae with what has just been given
for Fregata, we find it to be dissimilar in many particulars. This leads me to think
that with respect to the spinal column, the ribs, and the vertebrae, no typical Alba-
tross will be found to agree with the Man-o'-War Bird.

Pneumaticity is a prominent character of the trunk skeleton of Fregata, and it
would seem that all the bones enjoyed that state. The atlas has a broad neural
arch, and its cup instead of being perforated, is extensively notched above. On the
axis the "odontoid process" is short and stumpy; and its haemal spine is also an in-
significant affair. On the dorsal aspect of this vertebra, the neural spine is broad
and tuberous, while the postzygapophysial processes are much swollen also. These
latter, in the third vertebra, are joined upon either side with the prezygapophyses
by means of a somewhat delicate interzygapophysial bar. A large haemal spine is
present here, and a low neural one. The same character is to be found in the fourth
vertebra, and both third and fourth possess short backward-projecting parapophysial
spines.

In the fourth or fifth cervical a distinct neural process exists ; it is absent again
to include the ninth vertebra; it is but faintly developed in the tenth, and from
thence backward it gradually increases in proportion and changes in form, until we
have the low, quadrate plate of the dorsal series. Throughout the latter, haemal
spines are absent, and from the fifth to the tenth vertebra inclusive, the carotid

•JOS MEMOIRS OF THE CARNEGIE MUSEUM

canal exists, unclosed as it is by the rather feebly produced parapophyses. From,
the third vertebra to the thirteenth inclusive, lateral vertebral canals are to be found,
and in the fourteenth vertebra the pleurapophysial elements free themselves as a
pair of ribs. It is only in the last three dorsals that metapophysial spines, of no
great length, extend backward from the postero-external angles of the transverse
processes, and in these three vertebrae, upon their dorsal aspects, a pneumatic pit is
found, on either side, posterior to the prezygayophyses. The mode of articulation
among the centra is of the common ornithic character, and one of the chief features
distinguishing these vertebrae is the lack of prominence, or even entire absence, of
projecting processes, so conspicuous in many other birds. A good example of this
is seen in the eighth, ninth, tenth and eleventh vertebrae, where the tendency of the
short parapophysial processes is to merge with the centra by means of osseous
bridges connecting the two ; and another is the marked absence of haemal spines.

As near as I can judge from my specimen, the arrangement of the ribs of Fregata
is as follows : On the fourteenth vertebra there is a rudimentary pair, while on the
fifteenth there is a long, slender pair of free ribs, that are without epipleural proc-
esses. These are well-developed on the succeeding pair belonging to the sixteenth
vertebra; and this pair of ribs are the first of the series to connect with the sternum,
by means of rather short haemapophyses, the ribs of the dorsal series, seventeenth to
nineteenth vertebrae inclusive, are long, very narrow, highly pneumatic, and are
characterized by the unusual length of their epiplural spines. They all have costal
pairs of ribs, gradually increasing in length, and connecting with the sternum.

Two pairs of pelvic ribs are seen, both lacking epipleural appendages, and the
ultimate pair anchylosed with the vertebra to which they belong, and with the ven-
tral surfaces of the ilia. The last pair of costal ribs fail to reach the sternum.

The pelvis is broad and flattened. Anteriorly oneientire vertebra projects beyond
the ilia, but it is coossified in its position with the "sacrum."

Viewed from above, we are to note that the ilia are widely separated from the
sacral crista, which latter are very large, rounded and depressed. Slight concavity
characterizes the preacetabular iliac areas, and the anterior emarginated borders of
these bones are very obliquely truncated from their antero-mesial angles backwards.
The interdiapophysial vacuities in the uro-sacral region are of considerable size, and
the most part they have a parial arrangement and are very large in front.

The postacetabular areas are about equal in extent to the anterior ones, and are
convexed as much as the latter are concaved. Bony fusion between the ilia and
the vertebrae is, in all this region, very perfect, extending clear to the sacral ex-
tremity, and no part of the latter projects posteriorly beyond the iliac bones.

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES

Upon lateral view of this pelvis we note that the ischium (upon either side) is
much drawn out behind, and there pointed. It may he said that hardly any ilio-
ischiadic notch exists, the usual site for it being very shallowly rounded off. For
their anterior moieties the pubic styles are very slender, but their elongated pos-
terior ends are more clubbed and thickened, where they come simply in contact
with either ischium above them. On the lateral aspect the antitrochanter is seen
to be not very prominent ; and the cotyloid ring and ischiadic foramen have their
usual ornithic characters. Posteriorly, the obturator foramen opens most com-
pletely into the space of the same name ; thus, in reality, merging the two vacuities
into one. In the capacious pelvic basin on
the ventral aspect we find the last six parap-
ophysial struts of the vertebras, when pres-
ent, very distinct and slender, being thrown
up as braces in the usual manner. This fea-
ture also obtains with all the vertebrae at the
fore part of the sacrum, and indeed, there are
only two in the series that entirely lack this
development of the parapophysial processes,
they being the twenty-sixth and twenty-
seventh of the vertebral chain as a whole.

This pelvis is very light and highly pneu-
matic; air-holes always occurring in a small
group immediately in front of either cotyloid
ring, as well as in many other places.

Moreover, in many respects the pelvis in
Fregala is a very differently characterized
bone from what we find in an Albatross.

In the skeleton before me, I find six large,
free caudal vertebrse, besides a big parallelo-
grammatic pygostyle. The vertebrae have
very long, depressed, narrow diapophyses,
the anterior and posterior borders of which are rounded off. Stumpy, bifid,
centrally perforated haemal spines are found only in the last two or three, while in
all these vertebrae, so light from the high state of pneumaticity they enjoy, the
neural spines are elevated and pointed. The neural canal, nowhere very large,
seems to extend even into the pygostyle. This latter bone has a rounded superior
border, while below and behind it is thickened and broadened. At its lower part

FIG. 37. Dorsal aspect of the pelvis of Fregata
aqttila, very slightly reduced. From a photograph
by the author of the specimen in the collection of
the U. S. National Museum (No. 18485).

. :••'•*«• MEMOIRS OF THE CARNEGIE MUSEUM

it is usually transversely perforated by an oval foramen of some size. Above its
anterior concave articular facet the margin is sharp, the corresponding posterior one
being rounded. In the more expanded part below this latter margin, we find the
numerous pneumatic perforations that lead into its interior parts. The large
pygostyle of an Albatross is triangular in outline, and the caudal vertebrae of one
of these birds bear but slight resemblance to those bones in Fregata.

A number of anatomists in various places have already invited attention to the
striking peculiarities of the sternum and shoulder girdle of the species we here have
under consideration (PI. XXX., Figs. 50 and 51).

In the specimen at hand, I find that the symphysis of the furcula is extensively
coossified with the fore part of the carina of the sternum, while its upper clavicular
extremities are in a like manner very completely fused with the mesial aspects of
the summits of the coracoids, upon either side. The sternal ends of these latter
bones form free articulations with the sternum, and the scapula? are also freely articu-
lated each with the corresponding coracoid. This description, it will he seen, differs
from what Newton found in a specimen of Fregata, and he says, " In one very re-
markable way the osteology of Fregata differs from that of all other birds known.
The furcula coalesces firmly at its symphysis with the carina of the sternum, and
also with the coracoids at the upper extremity of each of its rami, the anterior end
of each coracoid coalescing also with the proximal end of the scapula. Thus the
only articulation in the whole sternal apparatus are where the coracoids meet the
sternum, and the consequence is a bony framework which would be perfectly rigid
did not the flexibility of the rami of the furcula permit a limited amount of motion."2

My skeleton is from an adult bird, as no doubt the one examined by Professor
Newton was, and this makes the circumstance all the more remarkable.

In Albatrosses the symphysis of the furcula also fuses with the sternum, while
the sternum, coracoid, and scapulae of some of those birds bear other slight resem-
blances to the corresponding bones in the skeleton of the Man-o'-war Bird.

Fregata has a very short sternum, being about as long as it is broad. The body
above is profoundly concaved, where it shows numerous pneumatic foramina espe-
cially down the median portion. Either costal process is triangular, pointed and
lofty, and the costal borders short. They about equal in length, however, on either
side, the sharp-edged lateral margin.

The rather short xiphoidal processes are situated, one at either posterior-external
angle of the body of the bone, and they are rounded off behind. For the rest, the
xiphoidal border that joins them is nearly an unbroken transverse line, exhibiting

20 Alfred Newton, " Dictionary of Birds," Art. Frigate-Bird, Part I., pp. 293, 294, London, 1893.

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES

211

neither notches or indentations. Ventrally, the carina extends the entire length of
the bone, its lower margin being very convex and thickened ; the lateral edges of the
latter extending beyond the sides. The anterior carinal boundary is sharp and short
surmounted above by a rudimentary manubrium. Either costal groove is deep and
wide, and the coracoids, when articulated in situ, meet in the median line above, where
they separate the anterior sternal border and the small trihedral manubrial process.

The os furcula is of the U-shaped pattern, and is coossified, as has been said, with
the sternum and coracoid. Its rami are narrow and compressed, with its l^poclei-
dial portion below very broad, triangular in outline, slightly concaved in front,
and withal antero-posteriorly compressed. Above, the clavicular ends are trans-
versely very thin. Either one shows two places of bony fusion ; one at the tip, and
a still more extensive one in front of the head of the coracoid. Mesiad, between
these, there exists an oval foramen. The clavicular extremities are separated from
each other by a distance of 6 centimeters. By far the largest bone of the pectoral
girdle is the coracoid. Either one of them has a height of 8.5 cm. with a large
shaft and broad sternal end. A minute epicoracoidal process is present, and the
clavicular process, although pretty well developed and curving forwards, does not
reach the end of the clavicle by quite an interval. The glenoid cavity faces more
than usual to the rear, much as it does in an Albatross. Measuring on its chord, a
scapula has a length of about 8 centimeters. Its inner border is convex, and its
outer more decidedly concave. For its middle third, the bone is narrow and thick-
ened, the distal extremity being slightly dilated, compressed from above down-
wards, and brought to a rounded point posteriorly. Its coracoidal extremity is
also much compressed in the vertical direction, very wide from side to side, with its
inner angle drawn out into a well-developed process, and with a small glenoidal
facet occupying the other one.

The Pectoral Limb. — The bones making up the skeleton of the upper extremity
are all completely pneumatic, and, as compared with those of the pelvic limb, are
chiefly remarkable for their enormous comparative lengths. This is well shown in
the following table :

Length of Humerus 21.0 cm.

Femur 5. 7

Ulna 27.5

Tibio-tarsus 7.5

Radius 26.6

Fibula 6.5

Carpo-metacarpus 13.0

Tarso-metatarsus 2.4

Index digit 12.5

Mid-anterior toe 7.6

212 MEMOIRS OF THE CARNEGIE MUSEUM

The humerus presents the usual double sigmoidal curve seen in the humeri of
almost all birds. Its shaft is smooth and subcylindrical in form, being somewhat
compressed from side to side. The pneumatic fossa is inclined to be shallow and
the foramina in it very small; but there are pneumatic openings in other parts of
the bone, notably at the distal extremity, where large ones occur upon either aspect.
The radial crest or the crista superior is very prominent, being of a triangular out-
line, with a thickened border distad. The ulnar tuberosity or the crista inferior is also
conspicuously developed, with the tuberculum internum much produced. The caput
humeri is elongated and spindleform ; it is separated from a flat tuberous area on the
ulnar side of the palmar aspect by a deep transverse sulcus, the sulcus transversus.
At the distal end of the bone the muscular grooves are very deep on its anconal as-
pect, the ulnar side apparently much raised in consequence. A deep fossa exists
both palmad and anconad to the trochlefe. A low ectepicondylar eminence is pres-
ent, with deep, circumscribed muscular pits near it on the radial border. No very
special resemblance exists between this bone in Fregata and the humerus as it is
found in any of the typical Albatrosses. The same would seem to apply to the re-
mainder of the skeleton of this limb.

In the ulna we find the shaft somewhat more cylindrical than it is in the humerus,
and presenting but a very slight degree of curvature. It has two distinct rows of
papillae for the quill-butts of the secondary feathers, the more prominent row being
represented by eighteen of them. The concavities for the trochlese of the humerus are
deep with raised edges, but the olecranon process is low and tuberous. Its distal ends
present the usual ornithic characters, with muscle-grooves very distinctly marked.

The radius viewed upon its superior aspect presents for its length an elongated
sigmoidal curve. This curve for its distal two thirds nearly agrees with the curva-
ture of the shaft of the ulna, when the bones are articulated in situ ; while for the
proximal third it is reversed, and here occurs a long, narrow, spindle-shaped " interos-
seous space." The radial shaft is more or less cylindrical, smooth, and but faintly
marked by muscular lines or ridges. The distal end of the bone is considerably
expanded in the transverse direction, and here pneumatic foramina are to be seen.
Its proximal head is only moderately developed, and it is the middle third of the
continuity of the shaft that possesses the greatest caliber. This latter very gradually
diminishes as we approach the extremities.

Fregata when adult has in its wrist the two usual carpal segments, which, from
the high state of pneumaticity they enjoy, are extremely light. Although they have
the usual form seen in birds, they are here powerfully marked by the various mus-
cular-grooves and articular facets that characterize them.

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 213

The carpometacarpus also presents us with the common ornithic features ; though
from its great pneumaticity, it is one of the lightest bones of its size we have ever
met with in the class. The pollex metacarpal forms a large triangular projection
at its usual site, while the shafts of the index and medius are nearly straight and
parallel with each other ; the former being fully six times the bulk of the latter.
Distally, neither one projects beyond the other, as is sometimes the case in birds of
other groups.

The expanded portion of the proximal phalanx of the index digit is both per-
forated and riddled with pneumatic foramina of various sizes and shapes, and upon
both sides. This part sends down a conspicuous rounded process beyond the proper
shaft of the bone, while from the antero-distal aspect of the shaft itself, a similar
process projects forward. At the proximal extremity are two other apophyses
standing out from the bone, one being upon its anconal and the other upon its
palmar side, while the plane of the articular surface of the summit extends, to
some extent, over and upon each. The distal or terminal phalanx of this finger
is of great length ; trihedral in form ; and much scooped out posteriorly. Pneu-
matic foramina enter it at various points, one, somewhat elongated, even occurring
at its apex. The pollex digit is not more than two thirds the length of the one
just described, and its posterior aspect, instead of being scooped out, is very sharp
and prominent ; and this border in the still smaller phalanx of the medius digit,
has upon it a low tuberous apophysis for muscular attachment. None of these ter-
minal joints of the fingers appears to bear claws at their extremities.

The Pelvic Limb. — Feebleness of a very marked character stamps every bone of
the lower extremity in Freyata. The skeleton here is, moreover, largely non-pneu-
matic. It is, apparently, only the femur and proximal moiety of the tibio-tarsus
that enjoys that state at all. The former bone is quite as pneumatic as the humerus
of the arm, while the portion of the tibio-tarsus mentioned appears to be only
moderately so.

The femur is short and bulky, with its head very distinctly sessile, and its upper
part markedly scooped out for the insertion of the round ligament. The trochan-
terian ridge does not rise above the articular summit of the shaft, and it is broad
and flat in the antero-posterior direction externally. Below these parts extends the
short, straight, cylindrical shaft, more or less strongly marked by the usual muscular
lines. The condylar portion or distal end is weakly developed, though all the char-
acters common to the femora of most birds are present. A fibular notch is well
defined, and the external condyle is placed lower down on the shaft. Both the
popliteal fossa and rotular channel are very shallow ; the latter, it may be said,

214 MEMOIRS OF THE CARNEGIE MUSEUM

hardly exists at all. Pneumatic foramina are always found at the base of the
former, immediately above the internal condyle.

Lack of salient characters distinguishes the tibio-tarsus even more profoundly
than the femur. Its cnemial crest and processes are almost completely reduced, the
former barely rising above the nearly level articular summit of the bone. Com-
paratively short and quite straight, its shaft is of decidedly slender bulk, being
subcylindrical in form. The fibular ridge is confined completely within the limits
of its upper third, being but a little more than a centimeter in length. Its articu-
lation with that bone is very free. At the distal end we find a pretty well de-
veloped pair of condyles, but the osseous bridgelet in front is low down and weak,
indicating that the tendons too, which it is intended to hold in place, must also be
but feebly developed.

The fibula is nearly complete, its lower end, of hair-like dimensions, is fused with
the side of the shaft of the tibio-tarsus immediately above the external condyle of
the latter bone. Above its proximal articulation, the fibula is of larger size, stand-
ing in strong contrast with the very slender straight part below that point.

Fregata has a flake-like patella of some size, being quadrilateral in outline, and
marked obliquely across its anterior surface by a groove for the ambiens muscle.
As with the vast majority of birds, it is non-pneumatic in character.

The tarso-metatarsus is wonderfully short and thick-set. This bone is so much
abbreviated in the direction of its longitudinal axis that it hardly appears to
possess any shaft at all. Antero-posteriorly it is flattened, while transversely it is
relatively broad. On its anterior surface well-marked grooves plainly indicate the
three metatarsal elements of which it is composed, while posteriorly the short shaft
is quite flat and smooth. Its hypotarsus is bulky and circumscribed, — not extend-
ing down the bone, — while a single central canal pierces it for the passage of the
tendons. This is of some size. Distally, the trochlear processes are comparatively
large and spreading. They all lie in nearly the same transverse plane, the inner
one being the lowest in position ; the outer, the highest ; and the middle one
occupying an intermediate place. The foramen for the passage of the anterior
tibial artery is peculiar, inasmuch as its lower exit is in the sulcus between the
outer and middle trochlear projections instead of on the back of the shaft above
that point, where it is usually to be found. The accessory or first metatarsal is as
large in proportion as in those birds where the skeleton of this limb is harmoniously
developed in point of size with the rest of the osseous system. All the phalangeal
joints of the pes are well developed, being arranged upon the plan of 2, 3. 4, 5, for
the first to fourth toe respectively, and were not this part of the skeleton of the foot

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 215

ridiculously small for the size of the bird to which it belongs, it might, with truth,
be said to be very well developed.

The ungual joints are powerfully curved, and considerable curvature character-
izes the majority of the others, especially those of the first, second, and third toes.
Although a digression from the consideration of the osteology of Fregata, it is inter-
esting to note that the greatly developed podotheca of the ungual joint of the middle
toe is beautifully pectinated, — a fact, in the present instance at least, that distinctly
militates against the view, still entertained by a few ornithologists, that the function
of the toes of birds so armed is the possession of an instrument wherewith they may
rid themselves and their plumage of vermin.

Relationships of the Steganopodes .

If we base our judgment on the osteology of the birds that have been exam-
ined in the present memoir, it would seem that we are justified in regarding the
Suborder STEGANOPODES as being composed of three superfamilies. These may be
designated as, first, the Pelecanoidea ; second, the Phaethontoidea ; and third, and
lastly, the Fregatoidea.

Arranging these, and arraying the existing families of them, with their genera,
a taxonomic scheme on such a basis would stand thus :

SUPERFAMILIES. FAMILIES. GENERA.

Pelecanus.

Pelecanidse.

Phalacrocorax.
Pnalacrocoracidffl ^T

( JNannopterum.
Anhingidie.

Anhmga.

Pelecanoidea

Sula.

Phaethontoidea. Phaethontidte. Phaethon.

Fregatoidea. Fregatidse. Fregata.

The genera and families of fossil forms are not taken into consideration in this
scheme.

Ornithotomists are agreed that the Steganopodes, constitute a well-defined group,
but beyond this the majority are reticent as to the question of the affinities existing
among the families and genera making up this group, and the relations of it as a
whole to other avian groups in the system.

If from among the Phalacrocoracidse we select the genus Phalacrocorax, there is
no doubt, in so far as its osteology indicates, that it is closely related to the genus
Anhinga. This, as has been shown above, is evident from a direct comparison of

216 MEMOIRS OF THE CARNEGIE MUSEUM

the corresponding bones of the skeleton of any species of Cormorant with those of
the skeleton of Anhinga.

On the other hand, and by similar methods, there is no disguising the kinship
existing between Phalacrocorax and Sula, although the gap between these genera is
somewhat greater than the one standing between the Cormorants and the Anhingas.

Pelicans of the genus Pelicanus are aberrant forms which, as osteologically indi-
cated, have varying relations with all three genera thus far mentioned. They are
however, apparently more nearly related to Sulidse than they are to the Cormorants.

From the Pelecanoidea the passage to the Phaethontoidea is not far to seek, for
upon comparing the corresponding bones in the skeleton of such a Gannet as Sula
brewsteri with those of Phaethon flavirostris we are at once confronted with so many
points of similarity as to leave no doubt in our minds that it is between the
genera and families represented by such species as these that the linking of the two
groups takes place.

This is important, for in another direction we are led, on the one hand, through
Phaethon to the suborder LONGIPENNES, and on the other to the suborder TUBINARES;
Phaethon flavirostris having some osteological characters that strongly suggest larine
affinities, and still more that bring to mind the skeleton of a Puffinus.

Their distinct maxillo-palatines, their perforate nostrils, their hardly coalesced
palatines, their four-notcheds ternum, and their ilia widely separated from the
"sacral crista," taken in connection with numerous other important skeletal char-
acters, fully entitle the Tropic Birds to rank as a superfamily — the PIIAKTHON-
TOIDEA as given above.

There can be no doubt about Fregata, for the skeletal characters seen in its skull,
its sternum and shoulder-girdle, its pelvis and limbs, and in its trunk skeleton, as
we have in detail described them above, stamp it at once, not only as being a form
having many skeletal characters completely at variance with those found in average
steganopodous birds, as Cormorants and Gannets, but as a type likewise for which
a superfamily must be created, in order to show that these striking departures are
fully appreciated by the student of its osteology. As indicated in our scheme
above, this superfamily may be designated as the FREGATOIDEA.

The pelvis in Fregata is decidedly more like the pelvis in Phaethon than in other
Steganopodes. In its pelvic limb-bones, which are extraordinarily short and otherwise
weak, as compared with the very lengthy pectoral ones, and the size of the rest of the
bird, it stands quite unique in the suborder to which it belongs. More remark-
able than all, however, are the many characters in its skull, which powerfully re-
call the Albatrosses among the TUBINARES. These are so evident that one is almost

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES 217

led to believe, if it be not actually the case, that the strong-hooked beak in the
skull of Fregata is a diomedean character rather than a pelicanine one. Apart
from the free ends of the furcula coalescing with the coracoids, there are characters
in the sternum and shoulder-girdle of Fregata which also recall the forms of the
corresponding bones in the Albatrosses, but beyond this there appears to be nothing
else in the skeleton of the Man-o'-War Bird at all reminding us of those birds.

As this relationship exists between Fregata and Diomedea, remote as it may be,
it nevertheless, taken in connection with what has been pointed out above in regard
to Pha'ethon and Puffinus, ought to lead us to believe that the STEGANOPODES are more
closely affined with the TUBINARES than they are with the LONGIPENNES.

There are those who claim to see a kinship existing between the Accipitres and
the FREGATOIDEA, but there are surely no indications of it in so far as the osteology
of any of the representatives of the two suborders in question are concerned.

ADDENDA.

Since writing the above account of the osteology of the Steganopodes, I have per-
sonally reexamined some of the material in the United States National Museum,
and, thanks to the marked kindness of Mr. F. A. Lucas, the able Curator of the De-
partment of Osteology of that institution, I have been permitted fully to examine
and compare the entire collection there, constituting as it does the finest assortment
of material illustrating the osteology of the Steganopodes in existence. A number of
the specimens drawn from this material are now figured for the first time in the
plates to the present memoir.

To what I have already set forth in the body of the paper, there is nothing I find
in particular to add to the osteology of the Phaethontidse or the Sulidse. Among the
Plotidse (the Anhingidsn of the A. O. U. "Check-List ") I examined a skeleton of Plotus
levaillanti, and it presents all the usual characteristics of the skeleton in that family
of birds. A dorsal view of its pelvis is given in Plate I., Fig. 1, which will assist in
illustrating what I have said in the body of the memoir in regard to the skeleton
of P. anhinga. I have never had the opportunity to examine the skeletons of the
two other known species of this family. (Sharpe's "Hand-List of Birds," p. 236.)

The collection of Cormorant skeletons (Phalacrocoraddie) in the U. S. National
Museum is, as I have said above, the finest in the world at the present writing,
nearly every known species being represented. It is Mr. Lucas' intention, some-
time in the future, I understand, to monograph this group, and it will be a very
valuable contribution to the subject. Their skulls present some very interesting
variations, and the majority of these are well shown in the figures of the Plates

218 MEMOIRS OF THE CARNEGIE MUSEUM

illustrating this memoir (Plate IV., Figs. 13-21 and Plate VI., Figs. 25, 27 and 29).
They not only vary in size according to the species, but perhaps the most interest-
ing variation they offer is the marked vertical compression of the skull they exhibit,
associated, as it is, with a decided elongation and narrowness of that part of the
skeleton. This is well seen in such a species as Phalacrocorax pelagicus robustus, where
this character about reaches its maximum. At the other extreme, we find the skull
of some Cormorants to be moderately shortened, with a broad, dome-like cranium,
.which admits of an unusually capacious brain cavity. A species having this char-
acter markedly exemplified is Phalacrocorax albiventris, and its skull is shown in
Plate IV., Figs. 14 and 21, and in Plate VI., Fig. 29. A good medium type stand-
ing between these two extremes is seen in Phalacrocorax dilophus, while all the other
species of Cormorants tend either toward P. p. robustus, or toward P. albiventris in the
matter of this flatness and elongate-narrowness of their skulls, or in the shorter,
broader and more capacious cranium of the last-named species. The shading either
way is often so gradual that everything else being equal, it hardly seems to offer
justifiable grounds for generic divisions. Associated with these forms of the skull
in the Cormorants, we are to note the variation in the form of the pterygoid bones
at the inferior aspects of the same. Some Phalacrocoracidse possess long and com-
paratively slender pterygoids, while other species have them much shorter and
stouter. Phalacrocorax p. robustus is an excellent example of the first of these (Plate
VI., Fig. 28), while P. albiventris well exhibits the last-named condition, while in
P. dilophus again, they are moderately elongated and fairly stout, thus once more,
affording an example of the medium type in this particular. In some Cormorants
the mesethmoid bone is not perforated by a large median vacuity, as is seen to be
the case in others. It is thus perforated in such species as P. p. robustus and P. albi-
ventris, while it is perfectly solid in such forms as P. dilophus and P. melanolcucus.
There are also interesting differences to be noted in the comparative size and form
of the foramen magnum, the morphology of the cranio-facial hinge, and other
minor points, the majority, of which, if of any importance, are noticed in the body
of the present memoir.

The pelvis in the various species of the Cormorants also exhibits a difference in
form. In P. pelagicus, for example, it is comparatively short, narrow, and with the
ilio-ischiatic notch not very deep, while in Graculus Carbo, it is much elongated,
narrow, and the aforesaid notch very deep. Other species of the Phalacrocoracidse
tend one way or the other in this particular. The short-bodied Cormorants have
the sternum more or less broad and short, while in the long-bodied species this bone
is likewise more elongated and narrower.

SHUFKLDT : OSTEOLOGY OF THE STEGANOPODES 219

As I have shown in this memoir all Cormorants possess a big patella. Mr.
Lucas in studying this bone has observed that this sesamoid in some of the species
is perforated for the passage of the ambiens muscle, while in others it is not so, but
this perforation is not associated with other characters in such a manner that the
condition can be regarded as having any taxonomic value. Mr. Lucas at my
request kindly examined this point for me in quite a number of species of Cor-
morants, and found that the patella is thus perforated for the ambiens in Graculus
Carbo, Phalacrocorax dilophus, P. vigua, P. harrisi, P. Magellanicus, P. albiventris,
(where it is small) and in P. pelagicus, while the patella is imperforate in Phalacro-
corax melanoleucus, P. pimctalus, P. penicillatus and P. urile.

Osteologically, the Pelecanidse constitute quite a homogeneous group of birds,
and certainly a thoroughly circumscribed family of the Steyanopodes. I have ex-
amined the skeleton in several species, but more particularly in the case of Pelecanus
sharpei, P. fuscus, P. onocrotalus, and P. erythrorhynchus. In some species the mandi-
bles are broad and comparatively shorter than they are in others, and, when so, they
are compressed from above downwards. This is the case in P. sharpei (Plate VII.,
Figs. 36, 37), and in this species we note in its skull that upon the lateral aspect of
the superior mandible, just where it is joined to the maxillary bone, the compact
tissue is continuous downwards to the anterior end of the corresponding palatine,
completely overlying and -concealing, on side view, the spongy tissue of bone in the
rhinal chamber, (Fig. 37). Whereas in such a species as P. fuscus there is a large
triangular interval left open here, so that upon the same view we are enabled to see
almost the entire mass of the osseous spongy tissue of the aforesaid space, (see Fig.
40 of my memoir, " Observations upon the Osteology of the Order Tubinares and
Steganopodes," Proc. U. S. Nat. Mus., 1888, p. 312). In the drawing referred to
the upper and lower edges of the united palatines have been shaved off, but I was
quite a juvenile osteologist when I prepared that specimen ; indeed it was my first
osteological preparation (1864). The correct form and outline of the lower united
edge of the palatine bones is shown in the case of P. sharpei in Plate VII., Fig. 37,
of the present memoir. In Figs. 35 and 40 of that plate I would say that the
horny podotheca covering the superior mandible of Pelecanus fuscus has only in part
been removed, and this also applies to the anterior two thirds of the mandible shown
in Fig. 41. Beyond what has been mentioned above, in most other respects, the
skulls of P. fuscus and P. sharpei are very much alike.

In the mandible the ramal symphysis is, for the size of the bone, altogether the
weakest union of any in the entire Class Aves. For example, the ramus of the jaw
in P. sharpei has a length of 34.5 centimeters, while the symphysis joining the two
rami together, anteriorly, measures but 3 millimeters in any direction.

220 ' MEMOIRS OF THE CARNEGIE MUSEUM

EXPLANATION OF PLATES.

(All the specimens figured in the Plates (XX I.-XXX.) are reproductions of photographs made
by the author, the material being in either his own private collection, or the collection of the
United States National Museum.)

PLATE XXI. Fig. 1. Dorsal aspect of the pelvis of Plotus levaillanti. Adult. (Coll. U. S.
Nat. Museum No. 18,743.) Two of the pelvic ribs show upon the right side.
Somewhat reduced ; the mid-longitudinal line of this bone measures 9.5 cm.
not including the ossified anterior tendinal extensions.

Fig. 2. Basal view of the skull of Phaethon cethereus, adult, mandible removed,
slightly reduced. Specimen in author's collection. From the island of Sail
Pedro Martis, Gulf of California. Collected in April, 1889.

Fig. 3. Superior aspect of the skull of Phaethon Jlavirostris, adult. (Coll. U.
S. Nat. Musum No. 17,841.) Slightly reduced, the reduction being propor-
tionate in amount with the skull shown in Fig. 2.

Fig. 4. Skeleton of left foot of Fregata aquila, adult. (Coll. U. S. Nat.
Museum, No. 18,485.) Somewhat reduced. Sheath left on midanterior toe
shows the " pectinated claw."

Fig. 5. Ventral view of the sternum of Phaethon cethereus, adult (author's
collection). Very slightly reduced. Collected in April, 1889. San Pedro
Martis Isle, Gulf of California. The skull in Fig. 2 and this sternum are
from two different individuals.

Fig. 6. Left lateral aspect of the trunk skeleton of Phaethon Jlavirostris, adult.
(Coll. U. S. Nat. Museum, No. 17,841.) Very slightly reduced. The skull
shown in Fig. 2 belonged to this skeleton, but the reduction in the former is
somewhat greater.

PLATE XXII. Fig. 7. Right lateral view of the skull and mandible of Sula gossi, adult
female (author's collection). Very slightly enlarged. Mandible detached.
The superior mandible is tubular at the extremity and pervious at the
apex. Collected at San Pedro Martis Island, Gulf of California, 16
Oct., 1888.

Fig. 8. Basal view of the skull of Sula gossi, adult female (author's collec-
tion). Very slightly reduced. Collected at same time and place as the speci-
men shown in Fig. 7.

Fig. 9. Basal view of the skull of Sula Brewsteri. Same specimen as shown in
Fig. 7. Very slightly reduced. Left pterygoid bone dislocated inwards to
show the form of the corresponding quadrate.

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES

221

PLATE XXIII. Fig. 10. Superior aspect of the skull of Sula tjossi. Mandible removed. Same

specimen as the one shown in Plate II., Fig. 8. Very slightly reduced.
Fig. 11. Superior aspect of the skull of Sula brewsteri. Mandible removed.
Same specimen as the one shown in Plate II., Fig. 9. Very slightly reduced,
and in the same proportion as the skull shown in Fig. 10.

Fig. 12. Left lateral view of the trunk skeleton of Sula gossi. Reduced one
third. Belonged to the same individual, which furnished the skull shown in
Fig. 10.

PLATE XXIV. (All the bones figured in this plate are from the Coll. U. S. Nat. Mus. and their
numbers are shown upon the specimens. They are all adult, and natural
size.)

Fig. 13. Left lateral view of the skull of Phalacrocorax pelagicus robustus.

Mandible removed. (See Fig. 18.)

Fig. 14. Left lateral view of the skull of Phalacrocorax albiventris. Man-
dible removed. (See Fig. 21.)
Fig. 15. Left lateral view of tfie skull of Phalacrocorax urile. Mandible

removed. (Specimen bears no number.)

Fig. 16. Basal view of the skull of Phalacrocorax melanoleucus. Mandi-
ble removed (Australia). There is but one known species of cormorant in
the world smaller than this one.
Fig. 17. Left lateral view of the skull and mandible of Phalacrocorax

dilophus. Mandible detached. (See Fig. 19.)

Fig. 18. Superior aspect of the skull of Phalacrocorax p. robustus. Mandi-
ble removed. (See Fig. 18.)

Fig. 19. Superior aspect of the skull of Phalacrocorax dilophus. Mandi-
ble removed. (See Fig. 17.)
Fig. 20. Superior aspect of the skull of Phalacrocorax melanoleucus.

Mandible removed. (See Fig. 16.)

Fig. 21. Superior aspect of the skull of Phalacrocorax albiventris. Man-
dible removed. (See Fig. 14.)

This series of specimens is designed to show the marked variation in form
of the skull among the cormorants (Phalacrocoracidce), ranging all the
way from the much vertically compressed skull of P.p. robustus (Fig. 13)
to the lofty skull of P. albiventris (Fig. 14).

PLATE XXV. Fig. 22. Dorsal view of the trunk skeleton of Phalacrocorax urile. with chain
of cervical vertebrae naturally articulated but curved far backward and to the
right. Reduced one half. The skull shown in Fig. 15 belonged to this skeleton.
Fig. 23. Left lateral view of the trunk skeleton of Phalacrocorax urile. Re-
duced one half, and the same specimen as shown in Fig. 22.

222 MEMOIRS OF THE CARNEGIE MUSEUM

PLATE XXVI. Fig. 24. Anconal aspect of the right carpo-metacarpus of Freyata aquila (Spec-
imen No. 18,485, Coll. U. S. Nat. Museum). Adult. Reduced ; the actual ex-
treme length of the bone being 13.2 cm.

Fig. 25. Left lateral view of the skull of Phalacrocorax melanoleuctis. Man-
dible removed. Natural size. Same skull as shown in Plate IV., Fig. 20.

Fig. 26. Anconal aspect of the proximal phalanx of the medius digit of the
maims of the right pectoral limb of Freyata aquila. Slightly reduced. From
the same skeleton that the bone shown in Fig. 24 was obtained, with which
it was distally articulated.

Fig. 27. Basal aspect of the skull of Phalacrocorax dilophns. Natural size.
Same skull as the one shown in Fig. 17, Plate IV.

Fig. 28. Basal aspect of the skull of Phalacrocorax p. rooustus. Natural size.
Same skull as the one shown in Fig. 13, Plate IV.

Fig. 29. Basal aspect of the skull of Phalacrocorax albiventris. Natural size.
Same skull as the one shown in Fig. 21, Plate IV. Note how much shorter
the pterygoid bones are' in this species as compared with the corresponding
elements in the skulls shown in Figs. 27 and 28..

Fig. 30. Dorsal aspect of the sternum and os furcula of Phalacrocorax aloi-
ventris (No. 18,437, Coll. U. S. Nat. Museum). Eeduced about one fifth.
(See Fig. 21, Plate IV. Both from the same bird.)

PLATE XXVII. (Pelecanus sharpd and Pelecanus fuscus. Adults. All less than half natural
size. P. sharpei is No. 18,736 of the Collection in the U. S. Nat. Museum,
and P. fuscns (author's collection) was shot by me on Indian Cay, Bahama
Banks, in 1864.)

Fig. 31. Anconal aspect, right ulna, Pelecanus sharpei.

Fig. 32. Anterior aspect, left femur, .P. sliarpei. Same specimen as Fig.

31 and the others on the Plate.

Fig. 33. Anterior aspect of the left tibio-tarsus, P. sharpei.
Fig. 34. Palmar aspect of left humerus, P. sharpei.
Fig. 35. Basal view of the skull of P. fuscus ; mandible removed.
Fig. 36. Basal view of the skull of P. sharpei ; mandible removed.
Fig. 37. Left lateral view of the skull of P. sharpei ; mandible removed.
Fig. 38. Anterior aspect of the right coracoid ; P. sha.rpei.
Fig. 39. Anconal aspect of the left carpo-metacarpus, P. sharpei.
Fig. 40. Superior view of the skull of P. fuscus ; mandible removed ;

same specimen as Fig. 35, and the mandible shown in Fig. 41.
Fig. 41, Superior view of the mandible of P. fuscus. (See Fig. 40.)
The three bits of wax supporting this bone show beneath the articular
cups and the symphysis.

SHUFELDT : OSTEOLOGY OF THE STEGANOPODES

223

PLATE XXVIII. Fig. 42. Dorsal aspect of the sternum and the coossified os furcula of Pelecanus
sharped; reduced about one fifth. No. 18,736, Collection in the U. S. Nat.
Museum. See Figs. 31, 32 and others of Plate VI. Same skeleton.

Fig. 43. Palmar aspect left pectoral limb of Phalacrocorax umle. Reduced
about one fourth. From the same skeleton that furnished the skull shown in
Fig. 15, Plate IV., and the trunk skeleton shown in Figs. 22 and 23, PI. V.

Fig. 44. Inner aspect of the right pelvic limb of Phalacrocorax wile. Reduced
about one fourth from the same skeleton that furnished the pectoral limb
shown in Fig. 43.

PLATE XXIX. (Skull, including mandible of Fregata aquila. No. 18,485 of the Coll. U. S.
Nat. Museum. Adult. All slightly reduced ; Fig. 46 more so than the others.)
Total length of skull 17.3 centimeters, and ramus of mandible 16.5 cm.
Fig. 45. Left lateral view of the skull ; mandible detached.
Fig. 46 Superior view of the skull ; mandible removed.
Fig. 47. Superior view of the mandible.

Fig. 48. Basal view of the skull ; mandible removed. For other bones of
the skeleton of this individual see Plate I., Fig. 4; Plate VI., Figs. 24, 26 ;
Plate X., Figs. 50 and 51. In Fig. 48 the right quadrate bone and the
pterygoid of the same side have been removed.

PLATE XXX. Fig. 49. Ventral aspect of the pelvis of Pelecanus sharpei; considerably re-
duced. The longitudinal median line of the vertebral portion of this bone
measures about 19.5 cm. It is from the same skeleton that furnished the
bones shown in Plate VII.

Fig. 50. Ventral aspect of the sternum and shoulder-girdle of Fregata aquila;
slightly reduced. See remarks under Fig. 48, PI. IX.

Fig. 51. Left lateral aspect of the sternum of Fregata aquila. Natural size.
See remarks under Plate IX. and Fig. 48.

MEMOIRS CARNEGIE MUSEUM, VOL. I.

PLATE XXI.

SHUFELDT. OSTEOLOGY OF THE STEGANOPODES. PLATE I.

MEMOIRS CARNEGIE MUSEUM, VOL. I.

PLATE XXII.

fy.9.

SHTJFELDT. OSTEOLOGY OF THE STEGAXOFODEB. PLATE II.

MEMOIRS CARNEGIE MUSEUM, VOL. I.

PLATE XXI II.

fy.<o.

SHUFELDT. OSTEOLOGY OP THE STEGANOPODES. PLATE III.

MEMOIRS CARNEGIE MUSEUM, VOL. I.

PLATE XXIV.

SHUFELDT. OSTEOLOGY OF THE STEGANOPODES. PLATE IV.

MEMOIRS CARNEGIE MUSEUM, VOL. I.

PLATE XXV.

SHUFELDT. OSTEOLOGY OF THE STEGANOPODES. PLATE V.

MEMOIRS CARNEGIE MUSEUM, VOL. I.

PLATE XXVI.

SHUFELDT. OSTEOLOGY OF THE STEGANOPODES. PLATE VI.

MEMOIRS CARNEGIE MUSEUM, VOL. I.

PLATE XXVII.

fy 37.

SHUFELDT. OSTEOLOGY OF THE STEGANOPODES. PLATE VII.

MEMOIRS CARNEGIE MUSEUM, VOL. I.

PLATE XXVIII.

SHUFELDT. OSTEOLOGY OF THE STEGANOPODES. PLATE VIII.

MEMOIRS CARNEGIE MUSEUM, VOL. I.

PLATE XXIX

SHUFELDT. OSTEOLOGY OF THE STEGANOPODES. PLATE IX.

MEMOIRS CARNEGIE MUSEUM, VOL. I.

PLATE XXX.

A- si-

SHUFELDT OKTKOI.OOY OK THE STEGANOPODES. PLATE X.

14 DAY USE

RN TO DESK FROM WHICH BORROWED

This book is due on the last date stamped below, or
K^ j £ ,e date to which renewed.

newed books are subject to immediate recall.

LD 21~50m-4'63
(DC471slO)476

General Library

University of Californi

Berkeley

C- BERKELEY LIBRARIES