L I B HAR.Y

OF THE

U N IVER.SITY

OF ILLINOIS

■NATURAL- HISTORY SURVE

550.5
FI

v.£,cop.2.

Return this book on or before the
Latest Date stamped below.

Theft, mutilation, and underlining of books
are reasons for disciplinary action and may
result in dismissal from the University.
University of Illinois Library

JON 7 lb

L161 — O-1096

Field Columbian Museum

Publication 78.
Geological Series. Vol. II, No. 3.

ON THE

OSTEOLOGY OF NYCTOSAURUS

(NYCTODACTYLUS),

WITH NOTES ON

AMERICAN PTEROSAURS.

BY

Samuel W. Williston, M.D., Ph.D.

Associate Curator, Division of Paleontology

Professor of Paleontology, University of Chicago.

Oliver Cummings Farrington, Ph.D.
Curator of Geology.

Chicago, U. S. A.
June 1, 1903.

ON THE OSTEOLOGY OF NYCTOSAURUS (NYCTODAC-
TYLUS), WITH NOTES ON AMERICAN PTEROSAURS.

BY SAMUEL W. WILLISTON.

In previous papers* I have given a summary of the principal characters
of the genus Nyctosaurus,\ with more or less complete descriptions of cer-
tain parts of the skeleton. In the present paper I give a more detailed
description of the skeleton, based upon the unusually well-preserved and
nearly complete specimen collected by Mr. H. T. Martin, together "with
some remarks upon the relationships and habits of the American
pterosaurs, and the description of a new genus and species from the
Lower Cretaceous of Kansas.

The specimen herewith described is of unusual importance, as throw-
ing much light upon the structure, not only of the American pterodactyls,
but also upon certain characters of the European ones. It is, I believe,
for purposes of study the most complete specimen of this order of reptiles
now known, comprising as it does nearly every bone in the skeleton, for
the most part associated in their natural relations. The specimen now
forms a part of the collections of the Field Columbian Museum.

SKULL.

A full description, with illustrations, of the skull has been given in the
Journal of Geology, vol. x, 1902, p. 520. I give herewith illustrations
of the mandible, not there figured. A careful measurement of the
sides of this bone a little in front of the posterior end of the symphysis,
where they are the broadest, obtained by measuring each fractured portion,
gives a width for each side of twenty-four millimeters. The correspond-
ing width of the palate is twenty-eight millimeters. This would give, as
the greatest depth of the mandible, near the end of the symphysis,
about twenty millimeters.

♦Kansas University Quarterly, vol. i, p. 5, 1892.

American Journal of Anatomy, vol. 1, p. 297, 1902.

Journal of Geology, vol. x, p. 520, iqo2.

Text Book of Paleontology, Zittel (Eastman), vol. ii, p. 255, f. 361.

tThis genus has generally been known as Nyetodactylus. (See synonymy beyond.) Marsh
changed the name because of the supposed preoccupation of the name Nyctosaurus. In this he
was not justified. The name has never been used otherwise for a genus of animals. Doubtless he
thought the term conflicted with Nyctisauria, used for a group of sauria. It does not, however,
according to the accepted canons of nomenclature, and the original name should not be displaced.

Field Col. Mus., Geol. Ser., Vol. II, No. 3.

126

Field Columbian Museum — Geology, Vol. II.

Fig. i. Mandible of Nyctosaurus gracilis, x %. i, from above; 2, from below.

MEASUREMENTS OF MANDIBLE.

mm.

Length, as preserved -l6°

Lengths of rami to hind margin of floor 78

Width between condyles 39

HYOIDS.
Two slender, cylindrical, gently curved bones, about thirty millime-
ters in length, were found lying side by side beneath the anterior part of
the skull. They are sufficiently well shown in Plate I of the paper cited
in the Journal of Geology. Lying near the occiput in the same plate, is
seen a small, triangular, or V-shaped bone which I then thought might be

On the Osteology of Nyctosaurus. 127

a proatlas, said by Zittel to occur in the pterodactyls, but of which there is
no other evidence in this specimen. Since the removal of this bone from
the matrix, it seems quite surely to be, rather, the anterior, median element
of the hyoidean apparatus. One view, which I take to be the superior,
is shown in PI. XLI, Fig. 6, three times the natural size, and, in Fig. 7 of
the same plate, is shown one of the articular extremities (a) from the side,
much more enlarged. The bone is nearly flat, with the pointed extremity
curved upward, and with the two articular extremities much more
massive than the remainder of the bone; they are also directed upward.
Each has a smooth, synovial articular surface, doubtless for the articula-
tion of the cylindrical rods. On the opposite surface, near the articula-
tions, there is a slight longitudinal ridge, and near the middle of the
bone on each side, there is an elongated, oval surface, apparently for
muscular attachment. The slender rods seem to agree quite with the
hyoidean bones of other pterodactyls, but I cannot find that the triangu-
lar bone has ever been described.

VERTEBRAE.

Seven has usually been accepted as the number of vertebrae in the
neck of pterodactyls. If, however, we consider that vertebra which
bears the first rib articulating with the sternum to be the first dorsal,
then I believe that the prevailing number of cervicals in pterodactyls
is eight. Furbringer has already expressed the opinion that there are
eight vertebrae instead of seven:* "Falls die Patagiosaurier zum Theil
nur sieben Halswirbel besitzen, wie algemein behauptet wird, aber
meines Erachtens erst noch zu erweisen ist, so ware eventuell anzu-
nehmen, das dieselben durch eine geringgradige, kranial gerichtete
Wanderung der vorderen Extremitat ihren urspriinglich aus acht
Wirbeln bestehenden Hals um einen in das thorakale Gebiet uberge-
henden Wirbel verkurzten." He further expressed the opinion that
the vertebra which I had considered to be the first dorsal in Pteranodon
was really the eighth cervical. In a later paper I stated that the eighth
vertebra in the neck of Nyctosaurus was a functional cervical, though a
structural dorsal. ... I have usually assigned this vertebra to the dorsal
series because of its great structural differences from the anterior cervicals;
but these differences are scarcely greater than those of the last cervical in
the turtles. Owenf correctly located it in the neck. In his description
of this vertebra in Dimorphodon he says: " At the base of the neck or the
beginning of the back, the vertebrae suddenly decrease in length; the
hypapophysis disappears, or is represented only by a slight projection of

* Jena. Zeitschr. f. Naturwissenchaft, vol. xxxiv, p. 665, 1900.
t Paleontograph. Soc, i860, p. 60.

128 Field Columbian Museum — Geology, Vol. II.

the lower border of the anterior cup; parapophyses [i. e., exapophyses] are
less produced. The lower surface of the centrum is flattened and quad-
rate in form. The parapophysis, diapophysis and rudimental rib coalesce
around the vertebrarterial canal; an oblique ridge is continued from the
anterior zygapophysis downward and outward upcn the pleurapophvsis
and behind the zygapophysis. Above these developments the neural arch
contracts from before backward to an extent of five lines, compared with
a total vertebral breadth anteriorly of one inch, eight lines; it then rapidly
expands, rising vertically at its fore part, and developing at its back
part the posterior zygapophysis, the articular facets of which look more
directly outward than in the long cervical vertebrae; the superincumbent
tubercle is more distinct from the facet; the posterior zygapophyses are
also much more approximated than in those vertebrae."

It would seem evident from this description, which applies in the main
to the corresponding vertebra of the Kansas forms, that the eighth post-
cranial vertebra is really a cervical. Certainly we can hardly put a
vertebra with rudimentary ribs at the beginning of the dorsal series!
Owen afterwards ascribed eight vertebrae to the neck, or seven, counting
the united axis and atlas as one. Quenstedt, also, in 1855, gave the num-
ber as eight. O. Fraas (Paleontographica, 1878, p. 166) found eight
vertebrae in the neck of Pterodactylus suevicus, but mistook the third for
the axis, thus counting only seven.

In my earliest paper* on Nyctosaurus I said: "The centra of twelve
vertebrae are preserved from the region back of the neck. Three of these
are evidently anterior thoracic, judging from their structure and the posi-
tion in which they lie. The shortest of them, to which was attached a
very large rib, and which was lying in front of the scapula, may represent
the first thoracic vertebra." This specimen I have figured in PI. XLIII,
Fig. 7, of the present paper. The specimen is fortunately preserved with
little or no distortion, though a part of the spine is wanting. As seen
from the front (the view figured), the transverse, shallowly concave cup is
straight or gently concave on its upper margin, and convex below. Aris-
ing from the front end of the centrum there is, on each side, a very stout
parapophysis, with which the head of the rib referred to was in apposi-
tion. On the anterior side of this process, and continuous, or nearly so,
with the outer side of the cup, there is an oval, convex articular surface
for union with the exapophysis. Above this process, separated by a
notch, there is a long, flattened diapophysis, for articulation with the
well-developed tubercle of the rib. The anterior zygapophyses are much
more approximated than in the long cervical verterbrae. Their articular

* Kansas University Quarterly, vol. i, p. o.

On the Osteology of Nyctosaurus. 129

surface is oval, the plane of each nearly at right angles with that of its mate.
The spine is broad and transverse; from the front it presents a median
ridge. On the under side the centrum is flat, the flatness reaching nearly
to the extremity of the parapophysis ; the width is considerably greater
than the length. The posterior surface of the centrum in this specimen
is distinctly roughened for sutural union with the following centrum; the
posterior zygapophyses also unite with the succeeding ones suturally.
This vertebra is very clearly the first of the notarium, as Seeley has hap-
pily called the united dorsal vertebrae. " Two other vertebrae found close
by the one described, and possibly one or the other contiguous with it,
differ remarkably in having no or a rudamentary parapophysial process,
and in having the diapophysis much shorter." "In Pteranodon there are
at least four vertebrae with dia- and parapophyses." " In two other
centra there is a long, recurved, parapophysial process, as though formed
by an anchylosed rib, on each side; they are probably lumbar vertebrae."
It is very clear, however, that these last two vertebrae are not lumbars, but
separated elements of the notarium. Of the two other vertebrae men-
tioned, one is probably the eighth cervical, described below, while the
other may be the fourth dorsal.

In Pteranodon, I thought there might be two vertebrae intervening
between the elongated cervicals and the notarium. " The centrum is short
and broad, so different from the preceding one that it is possible there
may be an intervening one lost. The ball is more than four times as
broad as high, concave on the upper margin, convex below. The post-
exapophyses are large, and confluent with the articular faces of the ball,
but are concave. The convex pre-exapophyses, at the outer side of the
cup are at the base of the lower root (that is the parapophysis) of the
elongated transverse process."* This vertebra corresponds with the
eighth post-cranial vertebra of the present specimen, but an "imperfect
vertebra, evidently following the one described, has the centrum very simi-
lar, save that the spine appears to be less stout." I cannot at the present
time examine this specimen, but it is probable that it is the first separated
vertebra of the notarium.

From the foregoing, then, it seems assured that there is a free, short
vertebra in front of the notarium, in both Pteranodon and Nyctosaurus,
bearing a free, small rib, which does not unite with the sternum. This ver-
tebra is the eighth cervical, and is probably present in all pterodactyls. It
is strange that authors, in the description of the neck bones of the ptero-
dactyls should have so often spoken of the united atlas and axis as one
vertebra, thus reckoning seven instead of eight vertebrae in this part of the

* Williston, Kansas University Quarterly, vol. vi, p. 40.

130 Field Columbian Museum — Geology, Vol. II.

column. Von Ammon* thus describes the neck vertebrae of Rhamphor-
hynchus longicaudatus :

" Es sind sieben Halswirbel vorhanden. Atlas and Epistropheus
werden hierbei als ein Wirbel gerechnet. Beide miteinander zusam-
menhangen, ihre Gliederung und Form sind undeutlich; die Knochen
sind durch spathige Masse entstellt. Es folgen nun nach riickwarts fiinf
gleichgestaltete, deutlich von einander abgegrenzte Wirbel, dann kommt
(der VII in der Reihe) ein von diesen etwas verschieden aussehender,
weil anders gestellte Wirbel der den abschluss des Halses bildet; der
unter letzterem befindliche Wirbel (der achte, [in reality the ninth])
besitzt bereits eine grossere Rippe, so dass mit ihm die Reihe der Brust-
wirbel zu beginnen ist."

Atlas and axis, PI. XLI, Figs. 3—5. The atlas is coossified with the
axis for the most part in the adult specimen, though leaving a distinct
sutural line of division in the more immature specimen. In the younger
specimen, the coossification is less complete, and the axial intercentrum
is also apparent, though indistinguishable in the older specimen. The
atlas is composed of the usual number of parts, though the centrum, or
odontoid must be small. The intercentrum forms the lower third or
more of the cup; it is deeply and smoothly concave, with a sharp cotylar
rim, which is nearly semicircular in outline. Its upper border in front
has three parts: the middle one, concave in outline, joins the odontoid;
the lateral borders, oblique in position, join the bodies of the neura-
pophyses. The posterior surface is probably flat, or but slightly concave;
the sutural line with the axial intercentrum is nearly parallel to the rim,
and the width is slight, though a little greater on the ventral side, where
the under border slopes somewhat ventrad toward the axial intercentrum.

The bodies of the lateral pieces, or neurapophyses, of the atlas com-
plete the rim of the cup, each forming a little less than one-third of the
circumference. Their lower border unites with the atlantal intercentrum;
the inner, lower border, joins the odontoid; while the inner, upper bor-
der, thinner and less in extent, unites with its mate in the middle. The
dorsal processes or arches are slender, somewhat oblique, and but slightly
dilated at the distal extremity, which appears to be free. They apparently
curve backward, to be applied throughout the whole of their extent, to
the thin neural border of the axis. The bottom of the cup is formed by
the odontoid. Its anterior surface has three convex borders of nearly
equal extent, ending more acutely above. How deep or thin this bone is
cannot be said, and it is not impossible that the surface appearing in the
cup is really the anterior face of the body of the axis, the odontoid being
obsolete. The inner ends of the lateral pieces are thin, and it is impos-

*Correspondenzblatt des Naturw. Ver. in Regensburg, 38 Jahrg., 1884, p. 151.

On the Osteology of Nyctosaurus. 131

sible to say whether the odontoid appears on the floor of the neural
canal or not.

The axial intercentrum is a large, wedge-shaped piece intervening
between the atlantal intercentrum and the body of the axis. It is scarcely
keeled in the middle below and reaches as high on the sides as does the
atlantal intercentrum.

The body of the axis is short and much expanded posteriorly. Its
ball is small and widened transversely. Below the ball the body is
expanded into a very large, flaring rim, inclosing a deep concavity,
having a broad articular surface transversely. This surface evidently
corresponds to the united postexapophyses of the later vertebrae — the
"posterior parapophyses " of Owen and Plieninger. Nothing quite simi-
lar to this structure seems to have been observed in any of the European
pterodactyls, though an axis figured by Owen,* as also Seeley, seems to
approach this structure. Nor do I find a corresponding adaptation of
structure in any of the following vertebras.

The arch of the axis is broad and relatively high. The neural rim in
front, to which are applied the dorsal neurapophyses of the atlas, is thin,
forming a heart-shaped opening. The spine is broad, low in front, where
there is a thin margin, for a short distance, and is highest behind. The
sloping border is thickened, the posterior border thin. Below, on each
side, there projects backward a strong process, on the under side of which,
before its termination, is the small, oval zygapophysis.

In the figures given in PI. XLI, I have partially restored the axis and
atlas after two specimens, both complete, but compressed in different
ways. The length of the atlanto- axis is 14 mm.

Plieningerf* describes the atlas in Pterodactylus kochii as consisting of
a "ziemlich massiven Korper, sowie aus einem Nervenrohr umschlies-
senden Bogenpaar, welches aussen und oben jederseits einen Fortsatz
tragt, der wohl zur Anheftung von Muskeln diente." He also identifies
the bone considered by FraasJ as a " Schlundring " as the atlas. From
the description and figure it would seem that the atlas of Pterodactylus
is of a very different structure from that in Nyctosaurus and Pteranodon.
(I have seen the united axis and atlas in a species of the latter genus;
they resemble the same bones in Nyctosaurus.}

Third — seventh cervical vertebra, PI. XLIII, Figs. 7-8. The cervical
vertebra; were lying close together between the back of the skull and the
top of the thorax, but they were dislocated and disassociated, so that
nothing can be told of their sequence from their position. Three are

* Brit. Fos. Rept. iv. PI. viii, fig. 2.
tPaleontographica, igoi, p. 17.
% Paleontographica, xxv, PI. xxii g

132 Field Columbian Museum — Geology, Vol. II.

lying upon their ventral surface, that is, compressed dorso-ventrally, and
two were more or less crushed laterally or obliquely. There is less
difference in the lengths of the five following the axis than is the case in
Pteranodon ; I am therefore somewhat in doubt as to the precise position
of some of them. Those that are depressed have the under surface
irregularly plane or concave, with the lateral margins rounded and
concave, formed by the ridge which reaches from the anterior zygapo-
physis and exapophysis on each side to the corresponding posterior
exapophysis. At the front margin there is an oval fossa, or depression,
on each side, separated by a convex surface; this convexity seems often
to be produced into a distinct hypapophysis in the European pterodactyls.
In a previous paper I stated that the exapophyses are non-articular in this
genus. This statement, after a more careful examination of the speci-
mens removed from the matrix, I know to be erroneous, at least so far
as the processes of the posterior vertebrae are concerned. Owen has
identified these processes with the parapophyses, and Plieninger prefers
to adopt this name for them. As will be seen by an inspection of Fig.
7, PI. XLIII, the true parapophysis — that is the process for the articulation
of the head of the rib, for which the term was originally introduced by
Owen, and in which sense it is now used — is always situated at the anterior
end of the vertebra, close to the rim of the cup. This process bears on
its anterior face, close to the rim of the cup, a convex, articular surface
for union with the corresponding concave surface of the " posterior
parapophyses " of the preceding vertebra. One might with as
much propriety call these posterior prominent articular processes the
diapophyses, or posterior inferior zygapophyses, as " posterior para-
pophyses." There is no such thing as a posterior parapophysis, nor can
there be. Nor could these posterior "parapophyses" ever have arisen
as processes for rib articulation. They are very characteristic of the
pterodactyl vertebrae, taking the place of the lateral or double articula-
tion of the cryptodire testudinate cervical vertebrae. I have given these
articulations a distinctive name of exapophyses in order to save much
circumlocution in their description; in any event they should not be
called "parapophyses," since they have nothing to do with these processes,
either morphologically or functionally.

The cup and ball are widened transversely; in the depressed specimens
very much so, and this is not due to their crushing, since those lying upon
their sides, though mutilated, still preserve evidence of a lateral elonga-
tion of the articular surface. The convexity of the ball is marked dorso-
ventrally, more so near the neural side. The posterior border between the
exapophyses is thin and concave, nearly concealing the ball, when seen
from below. The anterior zygapophyses project much in front of the cup;

On the Osteology of Nyctosaurus. 133

they are widely separated from each other. A ridge extends downward
and posteriorly from the outer side of the zygapophysial process to the
outer side of the convex exapophysis or parapophysis. I find no foramen
inclosed between these processes, and can see no vestige of a cervical
rib, such as is described in certain European pterodactyls. The posterior
zygapophyses are concave and oblique, and above them, there is a more
or less prominent process, a sort of " metapophysis." They do not extend
as far back as the ball, leaving a wide space between and back of them in
which the floor of the neural canal is visible. The laminae are broad, thin,
and roof-like, meeting in the middle and forming a very thin, neural
spine. The length of the base of this spine varies not a little in the
different vertebrae, from twelve to sixteen millimeters. In only one
vertebra is there a long spine, the one figured in PI. XLI V, Fig. 1 7, which I
take to be the seventh cervical. In the others, the free, thin border slopes
slightly upward and forward, forming a rounded spine but a few milli-
meters in height.

Eighth cervical vertebra. Pl.XLIII, Fig. 8. The eighth vertebra, which,
as already explained, is a real cervical and not a dorsal, was lying close to
the first notarial vertebra, and near the presternal process of the sacrum.
It has been so much compressed that all of its characters cannot be made
out with clearness. A figure of it is given as it lay in the matrix. It
differs greatly from the vertebra preceding it in the shortness of its
centrum, the character of its spine, and in the presence of diapophy-
ses. The ball is transversely widened, and has, at each extremity,
a large, concave, articular exapophysis. The post-zygapophyses are
situated much posteriorly to the ball, differing therein remarkably from
those of the preceding vertebrae. The spine above the zygapophyses is
very short and thick; seen from behind, it is concave, and ends obtusely
nearly over the zygapophyses. On each lateral expansion, near the upper
extremity, there is an oval, rounded, smooth surface, forming a sort of
process corresponding to the " metapophyses " of the earlier vertebrae.
The front border of the spine is concave in outline, and is rounded.
There is a rather slender and moderately long diapophysis, springing
high up on the arch. In the specimen it has been compressed against
the vertebra, but seems to end in an articular process. On the under
side of the vertebra, near the mutilated diapophysis, there is a fragment
of a rib, more slender than those of the notarium, which probably belongs
with this vertebra. In the much confused anterior end of this vertebra,
there is a small process on each side, evidently the exapophysis, which may
also have served for the articulation of the head of the rib; if so, how-
ever, the head must have been small. In the specimens of the Pteranodon
vertebrae already described, this parapophysis seems either rudimentary or

134 Field Columbian Museum — Geology, Vol. II.

wanting. I suspect that the rib was single-headed here. Because of the
small size of this rib, the large size of the exapophyses, and the very free
union of the vertebra, as also because of the position of the notarium in
the specimen, apparently in articulation with the first tubercle of the
sternum, I feel confident that this vertebra does not articulate with
the sternum, and that it is a cervical.

Plieninger, 1. c, considers the eighth shortened vertebra in Ptero-
dactylus as the first dorsal, because it bears a diapophysis. No rib was
preserved in his specimen, and he does not state whether there is- a
parapophysial process for the rib. If a rib was present, it was doubtless
small, since the next three pairs of ribs are found in place, and are
" besonders kraftig." The next three or four pairs are also strong. The
posterior ribs are slender.

Dorsal vertebra. PI. XLI, Fig. i; PI. XLIII, Fig. 7. The three firmly
united vertebra? of the notarium, which are visible, lie with their ventral
side uppermost, directed a little obliquely toward the left side, and are
partly concealed beneath the sternum. The front end lies about ten
millimeters back of the front end of the presternal process of the sternum.
The first two centra are visible nearly wholly, the third only in part. The
centra are flat below, concave on the lateral margins. The first has a
concave cup margin, and on each side a stout parapophysis is continued
into a strong rib, without clear indications of sutural union. This ver-
tebra in the Kansas University specimen has the ribs free, and it was
itself separable from the following centrum through its suture; it is
figured in PI. XLIII, Fig. 7, and has already been described. The
parapophysis of the second notarial vertebra seems to be given off some-
what further back, and there are indications of its sutural union with the
rib about seven millimeters from the body. The rib is four millimeters in
width beyond its proximal part, and a length of about thirty millimeters
was preserved. A specimen of a notarial rib, probably the first, preserved
with the Kansas University specimen, has a length of forty-five millimeters
and a width of five. The corresponding rib in this specimen must have
been not less than fifty millimeters in length, and probably about sixty.
Whether the third vertebra bears similar anchylosed ribs in this specimen
cannot be said, as they are covered by the sternum, but since the third
vertebra in the Kansas University specimen has such, it is undoubtedly
also the case in this. The end of a flattened rib, about twenty-five
millimeters in length and three or four in width, is lying by the articular
margin of the sternum, and may belong to this vertebra.

Lying in the axis of the notarium, a convex rim of a dorsal vertebra
has partly protruded through the thin sternum. This vertebra is doubt-
less either in direct articulation with its preceding vertebra, or but very

On the Osteology of Nyctosaurus. 135

slightly removed. The length of the series is about fifty millimeters,
showing that five vertebrae are associated in it, and that the protruding
one is the fifth. Whether the fourth and fifth are both suturally united
in the notarium cannot be said, but in all probability the fifth at least is
free, since its convexity shows it to be of the nature of the following ones,
and because one of the small posterior ribs seems to have been articulated
with it. The sixth, seventh, eighth, and ninth vertebrae are nearly in
relation with each other, behind the sternum, lying upon their dorsal
side, but are pushed somewhat to the left. They are cylindrical in cross-
section, with smooth, evenly concave sides and inferior border, with a
deep cup and a prominent ball. They are a little longer than broad, of
nearly equal length, though a little more slender posteriorly. The ante-
rior zygapophyses extend in front of the cup. The elongated and slender
diapophyses arise high up on the arch from near the anterior part of the
vertebra, and are directed horizontally outward. These vertebra? are quite
similar to those of Pteranodon, and are doubtless quite like that figured
by Seeley in his Dragons of the Air, p. 86, Fig. 26. The last dorsal (or
first sacral) (PI. XLI, Fig. 1, dv), the tenth back of the neck, is flat on its
inferior surface, and its transverse process I think arises from the centrum
as in Pteranodon. It is firmly united with the sacrum by suture in both
genera, and is quite as properly counted with the sacrum as with the
dorsal or lumbar vertebrae. It is probable that its transverse process
unites with, or reaches to, the anterior projection of the ilium, as seems
also to be the case in Pteranodon.

The position in which the skeleton is lying indicates that the right
coraco-scapula and the pelvis are in the positions in which they were as
regards each other while yet held together by the ligaments. The right
arm, falling across the abdomen, has caused a slight dislocation of the
dorsal vertebrae, the notarium held by the ribs has been turned somewhat
obliquely, and the sternum has settled down a little to the left. The
ends of the scapulae, however, are nearly in the relative position to each
other which they must have held during life, and the axis of the sacrum
is in line nearly with the presternum.

Now, by bringing all the vertebrae back into a straight line, the ten
vertebrae fill the entire space between the neck and the true sacrum,
proving almost incontestably that there were neither more nor less than
ten, a number found in no other reptiles except the turtles and Pareia-
saurus; indeed, if we call the last a sacral, then there are fewer presacral
vertebrae in Nyctosaurus than in any other known reptile.

The sacrum proper, as described below, is composed of six vertebrae,
the lines of sutural union of which are clearly visible in the specimen.

Four caudal vertebrae are preserved, lying close to the proximal end

136 Field Columbian Museum — Geology, Vol. II.

of the left femur. The largest of these (PL XLIII, Fig. 11), that evidently
immediately succeeding the sacrum, is flat at each end, and has neither
distinct processes nor zygapophyses, and only a small, low spine on
the posterior part; its width posteriorly is much less than that anteriorly.
The other three vertebrae are in a series, slightly separated from each
other, and are mere nodules of bone, without processes of any kind,
about four millimeters in length by three in width; the last one ends in
a conical point. From the tapering of the first, and the association of
the last three close by the side of the femur it seems very probable that
the tail was short and slender.

measurements of vertebrae.

Length, Width, Expanse,
mm. mm. mm.

Atlanto-axis 14

Third cervical 24

Seventh cervical 21

Eighth cervical (about) 9

First dorsal 9

Second dorsal 10

Third dorsal 10

Length of first five dorsals (about) 53

Sixth dorsal 9

Seventh dorsal . 9

Eighth dorsal . 9

Ninth dorsal 9

Lumbar (tenth) 10

First true sacral 9

Second sacral 9

Third sacral 7.5

Fourth sacral 7

Fifth sacral 6

Sixth sacral 6

First caudal, or caudo-sacral 7

Distal caudals 4, 4, 4,

RIBS.

Thoracic. PL XLI, Fig. 1, r,r. There are apparently four pairs of
stout ribs arising from the first four dorsal vertebrae, the first three of
which, at least, are anchylosed to the centrum in the adult animal. They
were doubtless all attached to the four tubercles on each side of the
sternum.

Beginning with the fifth dorsal, the ribs are single-headed, and are
very slender and delicate. Four pairs are discernible in the specimen,
and a fifth seems to be indicated by a fragment. The largest of these,
that apparently belonging with the fifth dorsal, is gently curved. It meas-

II

II

II

8

3i

8

30

7-5

29

8

9

32

7-5

35

7-5

37

7-5

32

7

28

6.5

21

6.5

15

10(8)

3.3.3

On the Osteology of Nyctosaurus. 137

ures about 75 ram. in length, 2 across its somewhat expanded head, .8
across the shaft a little beyond the capitular thickening, and but .4 near its
distal extremity. The sixth vertebra has its rib on the left side in relation
with the extremity of the diapophysis; it is a little shorter, .75 in width
near its proximal extremity, and .3 distally. The seventh also has a rib in
relation with the left diaphophysis; it is yet shorter, and nearly straight;
measures .6 in width proximally, and .2 distally. The next rib is straight,
and yet more slender, and shorter. The tenth vertebra, that connected
with the sacrum, I do not think bore a free rib.

In a former paper I stated my belief that the slender, posterior, single-
headed ribs did not inclose the abdominal cavity, but were directed more
or less outwardly in the patagial membrane. In support of this belief it
is seen that all these ribs are remarkably slender, and but slightly if at all
curved. Their free and loose connection with the vertebrae, and their
delicacy prevented them from being of use as a support to the abdominal
contents; certainly not if the animal walked in a quadrupedal position.
Furthermore, if they extended in the walls of the abdomen, the abdomen
must have been exceedingly voluminous, and widely unprotected in front.

A better support for my contention is afforded by the position of the
various bones in this specimen as they are preserved. All these ribs lie
directed outwardly or obliquely backward, and none of them are mutilated
or broken in the least, save as may have happened in the preparation of
the specimen. The right arm lies obliquely across the abdomen; it was
evidently sustained by the soft parts until after the ligaments connecting
the forearm bones had partly decomposed, since these two bones and the
carpals only are slightly disassociated from the remainder of the wing
bones. The radius and ulna fell to the back part of the abdominal
cavity while yet the pelvis was held together, permitting the left innomi-
nate bone to fall outward over the end of the ulna and carpal. Further-
more, the ventral ribs have fallen upon the radius. Now, under these
conditions, had the ribs curved forward in the abdominal walls to meet
the ventral ribs, they surely would have been entangled and doubled up
beneath the forearm. But this is not at all the case, and it seems highly
improbable that the fleshy walls of the abdomen could have fallen apart,
and spread out laterally with the wing folded across them. From all of
which evidence, I believe that the ribs were directed laterally, supporting
the patagial membrane, perhaps as in Draco, and that they did not inclose
the abdominal cavity.

Further evidence of the same tenor is furnished by the type specimen
of Rhamphorhynchus phyllurus Marsh, which I have recently examined.
In this specimen, the right arm had also fallen across the abdomen, while
yet having the patagial membrane attached to it. Nevertheless, the very

138 Field Columbian Museum — Geology, Vol. II.

slender ribs are spread outwardly, and were not doubled up below the
wing bones. They cannot be ventral ribs, since these are large and broad
in this species of Rhamphorhynchus, according to Zittel.

In Pterodactylus spectabilis, as figured in the books, a number of very
delicate abdominal ribs are shown, as depicted by Von Meyer. " An die
unteren Halfte des Rumpfes kommen fiinf fadenformige Abdominalrip-
pen. Sie sind fadenformige, ohne sich gegen die Mitte her zu verstarken,
und bei ihre Lange und diinnen Beschaffenheit etwas verbogen. Die erste
Rippe, die langste, ergiebt 0.026."* Is it possible that the abdominal
and vertebral ribs have been confused in this description?

Ventral ribs or parastemum. PI. XLI, Fig. 2. The ventral ribs were,
for the most part, lying in position nearly contiguous with the xiphisternal
process. There appear to be four pairs, the first three lying together and
united. The first pair is somewhat uncertainly shown, broken on the left
side and partly concealed on the right below the border of the sternum.
The second pair is very distinct. They are somewhat V-shaped, co-ossified
and thickened in the middle, and terminating outwardly in a slender
point. The third pair was in position, co-ossified with the margins of the
second pair, but not united in the middle. Another pair is seen by the right
side of the tenth vertebra, partly beneath the ulna; their whole form can-
not be made out, but they are wider and longer, apparently, than the third
pair, to which they could have had no ossific union. These ribs are all
thin. In the restoration it seems evident that the last pair of these ribs
would, approach, if not actually meet at their ends, the anterior ends of
the prepubic processes. I am very much inclined to believe that the
prepubis is really not a part of the pelvis, but rather a part of the para-
sternal ossifications.

One of the best descriptions extant of the ribs of pterodactyls is that
of von Ammon (op. ct.):

" Die hintersten dreizehnte Rippe stellt einen zarten Knochenstreifen
vor. Das untere Ende der Rippen ist verdickt und bildet ein kleines
Gelenkkopfchen, an welchem der obertste Theil einer Gurtelrippe der
Bauchwand befestigt ist. . . . Durch ihre Verbindungen [d. h. der
Abdominalrippen] mit den oberen Rippen bestehen also vollstandige
Giirtel in der Bauchwand. Es schliessen sich immer zwei Bauchrippen zu
einem formlichen Halbkreis zusammen. Am Vereiningungspunkte ist
eine mediane Verdickung vorhanden, die eine nach unten Spitz auslau-
fendes langes Kopfchen bildet." The ventral ribs of this pterodactyl
{Rhamphorhynchus longicaudatus) are very thin and simple, in contrast
with the broad ventral ribs of the larger forms of Rhamphorhynchus,
according to Zittel.

*Paleontographica, vol. x, p. 4, 1868.

On the Osteology of Nvctosaurus. 139

Owen* says that '■ the ribs of Dimorphodon acquire a characteristic
tenuity beyond the sixth pair," and Seeley mentions the fact that the
posterior ribs of the pterodactyls are slender. In current restorations of
the pterodactyls, the posterior ribs are often shown as long and strong
bones, inclosing a large and deep abdomen, the outlet of which must
have been through the narrow pelvis. It has been suspected that there
are sternal ribs intervening between the vertebral ribs and the sternum,
but there is no evidence of such in the present specimen.

PECTORAL GIRDLE AND EXTREMITY.

So far as is definitely known, the pectoral girdle comprises the ster-
num, coracoid and scapula only. No clavicles (unless the ossification
hereinafter described is a clavicle), inter-clavicle, epicoracoids, or precora-
coids, even, are known. Furthermore, the bones present are so strangely
modified that they present very little resemblance to the same elements
in other reptiles, agreeing rather better with those of birds, an agreement,
however, that I believe to be homoplastic in nature.

Sternum. PI. XLII, Fig. 1. The sternum is a broad and thin bone,
with a stout anterior projection, and a thin, spatulate xiphisternal process
posteriorly. The presternal process is narrowed and rounded on the
under surface, and may have projected somewhat ventrad. The saddle-
shaped articular surfaces for the coracoids look dorsad and laterad. The
anterior borders at the sides of the presternal projection are thickened,
concave slightly in outline on the inner part, straight and more oblique
on the outer ends. The lateral margins are thin, and are nearly parallel
with the median axis of the bone. These borders have three emargina-
tions, separated by four rounded projections, the anterior one longer and
thicker than the others, which are thin and small. The posterior border
is convex on either side, ending in a short and deep concavity at each
side of the flat xiphisternal process. This process is spatulate, with the
distal end rounded or very slightly emarginate, and doubless gave attach-
ment to the series of abdominal ribs, which lie in the specimen nearly
in connection with it. The sternum is somewhat thickened in the middle
anteriorly, or back of the base of the anterior process. On the sides and
behind the bone is very thin, almost paper-like; this thinner part in the
specimen shows numerous small crinkly fractures, doubtless caused by the
flattening of the concave bone; similar crinkles are visible in the thin
prepubic bones. The sternum lies nearly in its normal position in the
skeleton. In falling backward it has been displaced a little obliquely to
the long axis of the body, as have the dorsal vertebrae, perhaps caused by

* Paleontographica Soc, 1869, p. 69.

140 Field Columbian Museum — Geology, Vol. II. '

the weight of the right arm lying across the body. The presternal
process extends about half an inch in front of the notarium.

measurements of sternum.

mm.

Length in middle 88

Greatest width, across attachments of first rib 80

Length of rib margins 28

Width at coracoidal articulations 40

Length of presternal process in front of coracoids — 16

Width of xiphisternal process 9

Length of xiphisternal process 11

Coraco- scapula. PI. XLIII, Fig. 6. This conjoined element is a mod-
erately stout, U-shaped bone, with the coracoid branch stouter and a little
longer than the scapular. The distal extremity of the scapula is slightly
expanded and flattened spatulate; the shaft is flattened and somewhat
constricted in width. There is a rounded protuberance on the outer or
upper margin, just beyond the glenoid rim, for the attachment of muscles.
This process appears to be wanting in the specimen previously described
by me of this genus, but it is possible its absence is due to some
postmortem compression. The glenoid articulation is deeply concave
from above downward, convex from side to side, and is bounded both
above and below by a prominent ridge, that on the inferior border being
much stronger than the upper one. The glenoid surface is placed
obliquely to the shaft of the bones, doubtless in life looking outward
and somewhat backward. The surface, from side to side, is narrower
below than above. A line indicating the sutural union of the two bones
passes directly through the middle of the articular surface transversely.
At the bottom of the U formed by the conjoined bones there is a
process, rather slender, arising from the inner surface of the scapula,
and reaching to the inner face of the coracoid, which it joins. It incloses
a small foramen between it and the coracoid, back of and below the
glenoid surface. A precisely similar process is found in the coraco-
scapula of Pteranodon, and in neither genus does the suture dividing the
coracoid from the scapula cross this process; it is, apparently without
doubt, either a separate ossification joined to the two bones, or else a
process from the scapula. I cannot at present examine the under sur-
face of this scapula, and have none of Pteranodon for comparison. In
Nyctosaurus, however, the process seems to be separated from the scapula
throughout a large part of its extent. If it really belongs with the scapula,
the inclosed foramen cannot be the usual supracoracoid foramen of the
reptilian coracoid.

A similar foramen, though of larger size, is shown by Owen in the
coraco-scapula of Ornithocheirus sedgwicki (Paleontograph., 1857), which

On the Osteology of Nyctosaurus. ' 141

he calls doubtfully a pneumatic foramen. Whether the part is an acro-
mion process of the scapula, or possibly a vestigial clavicle, must be left
for future research.

The coracoid has a flattened shaft, oval in cross-section near its middle,
and flattened at either extremity. The sternal articular surface is convex
in its greater diameter, gently concave in the opposite direction. Near
the outer extremity of the shaft, on the inferior border, there is a strong
process, gradually arising f-rom the shaft but with a deep concavity be-
tween it and the articular rim. A small sesamoid ossicle was found
lying near it. The shaft at its middle is narrowed and smooth; it seems
to lack the strong muscular rugosity on the external part which occurs in
the Pteranodon coracoid; the process on the outer inferior border is also
larger than in Pteranodon. It will be observed also in the figure that
the glenoid articular surface does not extend to the inferior margin of the
coracoid, resembling in this respect somewhat the European forms of this
bone. The scapula of course differs markedly from that of Pteranodon
in the non-articular distal end.

measurements of coraco-scapula.

mm.

Extent of coraco-sternal articulation 10

Least width of shaft, below middle of coracoid _ 7

Length of coracoid .; 55

Dorso-ventral diameter of glenoid surface -- 19

Length of scapula 49

Least width of shaft of scapula 14

Width at distal extremity 16

Humerus. PI. XLII, Fig. 7. The humeri both lie with the inner
side uppermost. The proximal articular surface is distinctly saddle-
shaped, the convexity directed from behind forward and inward. The
curve of the concavity is greater than that of the convexity in the speci-
men. The ulnar or median process is very stout, reaching as high
as the articular surface. It is broad above, with apparently two faces
for muscular attachment, separated by a narrow, free space. Its rounded
distal border sinks into the surface of the shaft at about its middle
third. The deltoid, radial, or lateral process is very large. It is directed
forward and a little downward, its lower margin uniting with the shaft
at about the middle third. Both upper and lower borders are deeply
concave, the upper one the longer. Its distal extremity is expanded,
and has an evenly convex border from above downward. Near the
upper end of this border, on the inner side, there is a small, oval,
oblique surface, looking inward, for muscular attachment; and, begin-
ning near the middle of the upper border, there is a slender, but well-

142 Field Columbian Museum — Geology, Vol. II.

marked ridge, also for muscular attachment, traversing the anterior face
obliquely, and terminating near the lower, outer angle of the process.
Fiirbringer says of these processes in the pterosaurs: "Proc. lateralis und
Proc. medialis, namentlich aber der erstere sind sehr kraftig entwickelt
und prominieren insbesondere proximal so stark, dass das zwischen
ihnen befindliche Caputhumeri sogar theilweise gegen sich zuriicktreten
kann; dagegen ist — sehr im Unterschiede zu den meisten Sauropsiden,
namentlich aber den Theromorphen und Dinosauriern — der Proc. later-
alis wenig in die Lange entwickelt, indem er gewohnlich nur das proximal
% bis y$ des Humerus einnimmt. Diese Konfigurationen lassen darauf
schliessen, dass die Mm. supracoracoideus (supracoracoscapularis, scapulo-
humeralis posterior und subcoracoscapularis) eine relativ hohe Entfaltung
besassen, der M. pectoralis dagegen keine so abnorme Starke darbot,
wie es von vornherein von einem fliegenden Thiere erwartet werden
konnte, und der M. deltoides nur mittelstark entwickelt war."*

The shaft below the lateral process is of nearly uniform width. A little
above the middle of each bone, on the inner or ventrolateral side, starting
from precisely the same place, and lying transversely, there is a slender
bone, less than a millimeter in width, and about twelve millimeters in
length, preserved. I do not know what it is, nor whether it really
belongs in this position, though the similarity of the bone on the two
sides suggests that it does. It may be an ossified tendon. . . Because
of the position in which both humeri are lying, the distal extremity
has been transversely compressed, effacing the articular surfaces to a great
extent. There seems to be a small depression on the inner epicondyle.

MEASUREMENTS OF HUMERUS. mm.

Length '. 87

Least width of shaft 5

Least width of lateral process - 17

Width of lateral process distally 25

Width of humerus through lateral process 42

Width of median process 5

Ulna. PI. XLII, Fig. 2. The bones of the left forearm lie nearly in
position, and in contact with the distal end of the humerus; those of the
right side have been slightly separated, the radius detached and removed
proximally. The left ulna is lying nearly upon its outer side, partly over-
lapping the radius; that of the right forearm has been compressed dorso-
ventrally. The bone is slightly expanded at the extremities. There are
two, nearly confluent, articular surfaces at the proximal end, the smaller,
nearly circular one, on the inner side, nearly confluent with the large, less

*Jenas. Zeitschrift, 1900, p. 363.

On the Osteology of Nyctosaurus. 143

deeply concave surface on the inner side. Just in front of this, near the
middle of the bone, transversely, there is a pneumatic foramen on the
flexor side. The inner side of the bone, at this extremity is expanded
into a thin, convex process. The articulation at.the distal extremity can-
not be made out, as it is largely concealed by the carpals. The shaft is
nearly straight, curved forward gently at the distal extremity.

measurements.

mm.

Length 144

Width of proximal extremity - — 44

Width of shaft -• •- 24

Radius. PI. XLI, Fig. 2, PI. I. The radius is a more slender and
slightly shorter bone than the ulna. On the left side it lies partly con-
cealed beneath the ulna and there would seem to have been no pronation
of the bone. The bone is more slender than the radius, somewhat concave
along its inner or upper border as it lies in relation with the ulna; nearly
straight as it lies compressed from above downward. The bone is more
expanded distally than proximally. The proximal articular surface is
rounded and concave. Lying near the roughened distal surface shown
in the figure there was a small sesamoid bone.

MEASUREMENTS OF RADIUS.

mm.

Length - 140

Width of shaft middle third . 18

Width distal extremity 24

Carpals. PI. XLII, Figs. 2, edt 3, PI. XLIV, Fig. 8. The two chief
carpals of the left wrist are in relation with each other and the wing meta-
carpal, and but slightly displaced from the forearm bones. On the right
side they are separated, and the proximal one is partly concealed beneath
the ischium, the distal one having its proximal surface obliquely exposed.
In the figure (PI. XLIV, Fig. 10), the one shown obliquely is that of the
other carpus figured. The wing metacarpal of the left side (PI. XLIV, Fig.
2), lying articulated with the distal carpal, has the under side uppermost,
from which it would appear that the view given is of the anterior side of
the carpus. The proximal carpal is much the larger, extending on the
under side either to articulate with the metacarpal, or nearly so. A sub-
angular space is left between the two bones on the radial side, for the
articulation of the lateral carpal or metacarpal of the first digit. This
latter bone, of the right side, is shown in PI. XLII, Fig. 3. It is a small,
elongate bone, with an articular emargination on one side, which I had
supposed was for the pteroid. I give, also, a good figure of this bone
of Pteranodon, copied from Cope (Cretaceous Vertebrata, Plate VII,

144 Field Columbian Museum — Geology, Vol. II.

Fig. 3). It resembles that of Nyctosaurus, though having a less elon-
gated process beyond the articular lateral emargination. The figures of
the carpal bones of Ornithocheirus given by Seeley (Ann. Mag. Nat. Hist.
Aug. 1870, and Dragons of the Air, p. 124) closely resemble these bones.
Professor Seeley says of the " lateral carpal ": " It is a flat, oblong bone,
attached to the inner side of the distal carpal, and, instead of being pro-
longed distally in the same direction as the other metacarpal bones, is
turned round and directed upward, so that its upper edge is flush with the
base of the radius, and gives attachment to the pteroid." I had supposed
that the larger end, as shown in the present figure, fitted into the interval
between the two carpals, and that the emargination was for the articulation
of the rounded head of the pteroid, though perhaps the position assigned
to it by Seeley is the correct one. The bone is thin on the sides shown in
the figure, so that the articulation of the pteroid must have been either in
the emargination or with the broader end. Marsh says* that it " stands
nearly at right angles with the wrist," and I am inclined to think he was
right. The structure of these wrist bones is almost identical with what it
is in Pteranodon.

measurements of. carpal bones.

mm.

Length of articulated carpus 15

Greatest diameter of distal carpal 24

Lesser diameter of same 14

Greatest diameter of proximal carpal 26

Length of "lateral carpal" 17

Width of same 8

Metacarpals. The first metacarpal, or pteroid (PI. XLII, Figs. 4, 5),
is an elongated, pointed, styliform bone, with an enlarged articular car-
pal extremity. In both arms this bone has been displaced, though so
closely associated with the corresponding extremity that the position is
assured. Its articular end is broad, its margin nearly at right angles with
the long axis of the bone. The articular surface is nearly at right angles
with the transverse diameter, in life nearly circular in outline, and decid-
edly convex. It is separated from the bone by a slight constriction.
The dorsal border of the bone is very gently concave throughout, and
the inferior border is correspondingly convex for the greater part of its
extent. The pteroid of the right arm is lying close to the lateral carpal,
as though its articular surface fitted into the lateral emargination of that
bone in life. Oscar Fraas was the first, so far as I can learn, to recognize
the real nature of this bone as belonging to the first digit,f a view after-
wards adopted by Marsh and Zittel. .

♦Amer. Journ. Sci., 1882, p. 255. *

fPaleontographica, 1878, p. 170.

On the Osteology of Nyctosaurus. 145

measurements of pteroid.

mm.
Length 104

Greatest width (at proximal end) 15

Diameter of articular surface 8

The next three metacarpals are, for the most part, wanting in this
specimen, a single complete one lying near the distal extremity of the
right wing metacarpal (PI. XLII, Fig. 6). It is a very small, splint-
like and slightly curved, pointed bone, measuring about twenty-four
millimeters in length by two in greatest width. Its resemblance to the
corresponding metacarpals of Pteranodon is so great that the whole
structure of the hand is doubtless the same in the two genera. I give
herewith a diagrammatic figure of these parts in. Pteranodon, based on a

^Y^

Fig. 2. Metacarpals of Pteranodon ingens.

specimen in which nearly all the bones were present and in position,
some of the terminal phalanges only being misplaced, and one or two of
the fourth finger missing. In a former paper I stated that the phalanges
of the hand were of two kinds, long and short. Possibly this is the case
in some of the smaller species, but I think not. I doubt not that the
small phalanges there described were from the foot, and had become dis-
placed and associated with the hand phalanges in the specimen described.
In the specimen of Pteranodon the small metacarpals seemed to lie,
not as they are diagrammatically figured, one above the other, but more
side by side. The second metacarpal is continued, as already stated by
Marsh, to the carpus, as a thin, thread-like bone. It was not continued
along the dorsal side of the bone, but seems to pass to the radial side,
where it belongs. From this it seems also probable that the position of
all three bones in life was along the radial dorsal margin of the fifth
metacarpal.

Both of the wing metacarpals (PI. XLIV, Fig. 2) lie upon the dorsal
surface, a position in which they are seldom found. The bone is much
broader at the base, tapering to beyond the middle, whence the shaft
has nearly parallel sides. The proximal articular surface cannot be
made out, though there appears to be a division into two facets. The
condyles, as in Pteranodon, are placed obliquely, the posterior one the
larger.

146 Field Columbian Museum — Geology, Vol. II.

MEASUKEMENTS.

mm.

Length 224

Width at proximal end 18, 23

Width of shaft at middle u, 12

First phalange of wing finger. PI. XLIV, Fig. 1. The first phalange, as
usual, lies upon the side in both wings. As in Pteranodon, it is the long-
est bone of the body. Its under margin is concave throughout, the con-
cavity greater near each end. The proximal expanded extremity has a
deep articular concavity, extending through an arc of about one-third of
a circle, with its radius directed at an angle about 45 ° with the long axis.
The upper, olecranon-like projection has a marked emargination of the
border above, between which and the proximal end the convexity is
greater, with the corresponding deeper concavity of the under border.

There are two articular facets, separated by a median ridge, which
fits into the trochlear depression of the metacarpal. The posterior sur-
face extends throughout the whole length of the arc, while that on the
anterior side reaches but little more than one half of the distance, and
its upper part is narrower than is the other surface. The pneumatic
foramen is on the anterior side, just beyond the end of the articular sur-
face. The dorsal border of the bone is gently and evenly convex from
a little beyond the proximal extremity, and the width on the distal three-
fifths is nearly uniform. The distal extremity of the bone is curved
downward; the articular surface in life was doubtless a vertical oval.

measurements.

mm.
Length 283

Greatest width at proximal end 27

Chord of proximal articulation 17

Width of shaft, proximal third 14

Width of shaft, distal third--- 14

Width distal extremity 18

Thickness of shaft at middle as compressed 1.7

The second wing phalange (PI. XLIV, Fig. 5) lies in position in the
right wing, but is somewhat displaced in the left. It is much smaller
and more slender than the first phalange. Its proximal border is nearly
straight and transverse to the axis of the shaft; it ib much broader than the
shaft. The under border of the bone is gently concave throughout,
except on the proximal part, where the concavity is deeper. The upper
border, gently concave at the proximal end, is nearly straight thence to
near the extremity. The distal extremity is quite like that of the first
phalange, though smaller.

On the Osteology of Nvctosaurus. 147

measurements.

mm.
Length — 211

Width proximal extremity - - 17, 18

Width at proximal third _- 9, 10

Width at distal third 8, 9

Width of distal end — 9, 10

The third phalange has the proximal extremity preserved, com-
pressed from above. The surface seems to have been only slightly con-
' cave. This phalange, like the preceding one, seems to have been pro-
portionally more slender than in Pteratiodon.

PELVIC GIRDLE AND EXTREMITY.

Pelvis. PI. XLI, Fig. 1. The pelvis lies with the bones nearly in
position, the sacrum with its ventral surface uppermost, the innominate
bones separated at their sutures and in juxtaposition with the sacrum, the
visceral surface also uppermost.

There was in the sacrum (PI. XLI, Fig. 1, s) in life evidently a
considerable concavity transversely, and possibly a slight concavity longi-
tudinally. It is composed of six firmly united vertebra?, the sutural lines
between the centra clearly distinguishable. The vertebrae decrease in
length from the first to the fifth; the sixth is slightly longer than its
predecessor. The centra are flattened in the preserved specimen, though
possibly in life they may have been slightly convex transversely. The
sacral ribs show no indications of sutural connections anywhere, and are
broadly united distally. Between them the oval foramina decrease rapidly
in size, and slightly approach the middle line, the last, that between the
fifth and sixth vertebrae, being almost punctiform, while the first is of
considerable size.

Trie first process is broad, with a broad, oblique surface, terminating
distally in a free, rounded margin. The pelvic brim is indicated by an
angular line passing outward and downward obliquely to the upper
border of the articular surface for the innominate bone. This surface is
oval, and encroaches obliquely on the front surface of the sacrum; beyond
it the thin margin of the sacrum does not present a sutural surface on the
ventral side. The lateral margins approach each other until at the hind
border the distance between them is less than half that at the brim of
the pelvis.

The ilium (PI. XLI, Fig. 1, il) has a long, flattened or somewhat pris-
matic anterior process, reaching as far forward as the second presacral
vertebra. The process is apparently curved somewhat backward in life,
and also outward; it has an obtuse anterior extremity. Posteriorly the

148 Field Columbian Museum — Geology, Vol. II.

ilium extends a short distance beyond the end of the sacrum as a short,
obtuse process. It sends an angular expansion upward and inward
opposite the end of the sixth sacral vertebra, which extends five or six
millimeters above the margin of the sacrum, possibly forming a sort of
roof over the last sacral and the first caudal, and approaching, or
possibly meeting, its mate in the middle. The flattened or ventral por-
tion (PI. XLI, Fig. i, ip) of the ischio-pubis is somewhat trapezoidal in
shape with the symphysial line slightly oblique to the sacral border.
Both bones, separated at the median suture, lie with the visceral surface
uppermost, and both present a considerable concavity. The anterior
border begins at the upper margin of the process, which articulates with
the sutural surface described on the lateral margins of the sacrum oppo-
site the first intercostal foramen. The margin is thin and concave as far
as the angular process for the attachment of the prepubis. This process
is everted from the brim of the pelvis, and is tipped with a small, oval,
cartilaginous or articular surface. From this surface the margin, some-
what thicker, recedes in a straight line to the beginning of the median
symphysis. The symphysial margin is thin and nearly straight. For
the first fourth of its length it seems to meet its mate, if at all, in an
harmonious suture, and the symphysis in front may have been somewhat
rounded or subangular; beyond this, however, the two bones meet in a
very oblique surface, attaining a width of four or five millimeters poste-
riorly, and the two bones must have formed a thin and sharp keel when
conjoined. The posterior border is somewhat thicker than the anterior
one. It begins in a rather acute rounded angle at the base of the
posterior process of the ilium, and is directed backward and ventrad to
the angular extremity of the symphysis. The large rounded ischiadic
(obturator) foramen is situated in front of the middle of the bone,
almost directly ventrad to the acetabulum; its borders are smooth and
thin. Between this foramen and the margin of the ilium, there is a con-
siderable convexity, representing the floor of the acetabulum. Its dorsal
margin must be nearly over the line of the sacrum. The junction of the
innominate with the sacrum is by a narrow concave sutural surface, fitting
on the margin of the sacrum; in front there is a considerable oval surface
on the anterior part of the bone, which fits into the depressed surface
already described, on the sides of the sacrum. Posteriorly, also, the pro-
jection of the ilium has a little broader attachment. Dorsad to this
junction the ilium seems to have projected beyond the margin of the
sacrum slightly in the middle, expanding in the angular process already
described.

Prepubes. PI. XLI, Fig. i,pp> The prepubes are very nearly as they
are in Pteranodon, consisting of a flat, transverse ribbon-like band meet-

On the Osteology of Nyctosaurus. 149

ing in a somewhat angularly expanded median symphysis, which is thor-
oughly co-ossified, and with an anteriorly directed, somewhat divergent,
flat, and obtusely pointed anterior process at each side, the continuation
of the posterior flattened portion which joins the pectineal process on
the everted margin of the pelvis. This process is somewhat thickened
at its posterior extremity, and is of about the same width as the other
parts of the bone. The bone in the specimen had been torn apart at the
symphysis, each half remaining in partial juxtaposition with the corre-
sponding half of the pelvis. As flattened out the conjoined bone is
much too wide transversely to articulate with the surfaces on the margin
of the pelvis; too wide even in the flattened-out position of these parts.
It seems certain that the extremities must have articulated with the
processes described, quite as I have already described the conditions in the
pelvis of Pteranodon, and it therefore follows that the bone in life must
have had a considerable convexity, in a broad, U-shaped form.

There are no indications of any sutural connection in the bones of
the pelvis, except such as I have described with the sacrum and between
the two halves. Nevertheless, the presence of the large foramen, the
general shape of the bone and the position of the acetabulam would be as
one would expect were there really the three elements present. A suture
between the ischium and pubis is scarcely to be expected, in consideration
of the obliteration of the sutures elsewhere in the skeleton. I am there-
fore inclined to look with favor upon the view, long held by Seeley, that
the pelvis is of the normal reptilian structure, and that the bones in front
really do not belong to it. It seems very probable that the anterior pro-
longations of the prepubes come in contact with the posterior ends of
the parasternum, and that the whole bone is merely a continuation of the
abdominal armature. However, I must admit that the persistence of the
sacral and symphysial sutures, permitting the pelvis to drop apart without
distortion, and the utter obliteration of any ischio-pubic suture is
inexplicable.

" Naturalists have been uncertain as to the number of bones in the
pelvis of pterodactyls, because the bones blend together early in life, as
in birds. Some follow the Amphibian nomenclature, and unite the
ischium and pubis into one bone, which is then termed ischium, when the
prepubis is termed the pubis, and regarded as removed from the
acetabulum. There is no ground for this interpretation, for the sutures
are clear between the three pelvic bones in the acetabulum in some
specimens, like Cycnorphamphus Fraasii, from Solenhofen and some
examples of Ornithocheirus from the Cambridge Greensand."*

By modeling in cardboard and clay the bones of the pelvis and

•Seeley, Dragons of the Air, p. 04.

150 Field Columbian Museum — Geology, Vol. II.

uniting them in a complete basin, with the necessary curvatures, it
transpires that the opening of the pelvis had a diameter of about seven-
eighths of an inch, while the outlet could have measured but a trifle more
than half an inch.

Femur. PI. XLIII, Figs. 4, 5. Both femora are preserved, separated
by a short distance from the pelvis. The neck is directed upward and at
a slight angle inward; it is cylindrical, the head only a little dilated and
with its convexity only a little oblique to the axis of the bone; that is, the
plane of the rim is nearly, but not quite, rectangular to the axis of the
shaft. The trochanter stands a little to the outer side of the middle axis
of the bone. From its upper angle a ridged process runs backward for a
short distance, to the inner side of which, a little below the free, upper
concave margin of bone, there is a small pneumatic opening. On the
outer side, the margin of the shaft is convex for a short distance, and is
then concave throughout to the extremity. On the inner side above, the
margin is correspondingly concave; below this to the condyle it is con-
vex. The upper part of the bone was evidently, in life, nearly cylindrical;
below it seems to have been wider and somewhat flattened. Near the
middle behind, beginning a little above the condylar surface, there is a
narrow, elongated, longitudinal ridge for the attachment of a muscle.
The distal articular surface cannot be clearly made out in the crushed
specimens. The bone had in life, apparently, a marked anterior curva-
ture, as in Pteranodon.

MEASUREMENTS OF FEMUR.

mm.

Length 81

Diameter of neck 6

Diameter of shaft, upper third 7

Diameter through condyles as crushed , 13

Femur of Pteranodon ingens Marsh, PI. XLIII, Figs. 1-3. The con-
vexity of the head is regular, covering nearly half of a circle transversely,
probably a little less in the conjugate diameter, the surface thus forming
an oval or ovate figure, the plane of whose base is nearly at right angles
to the long diameter of the bone. The convex surface is sharply limited
from the neck. The neck is cylindrical, rather stout, and is directed
nearly vertically downward. The rounded and moderately prominent
trochanter is placed over the middle of the shaft, descending into the
concavity at the side of the neck, and externally separated from the
margin of the bone. The shaft is nearly of equal width throughout;
doubtless in life it was nearly cylindrical, with a strong anterior curvature,
and a flattening in the popliteal region. The sharp margin of the inner
condvle encompasses nearly half a circle. The outer condyle, though

On the Osteology of Nyctosaurus. 151

extending further, is much less in extent. The inner distal articular sur-
face is the larger, and is separated from the inner surface by a distinct
ridge, especially posteriorly.

Tibia. PI. XLIV, Fig. 3. The tibia is a slender, straight bone, mod-
erately expanded at the upper extremity. The margin of the articular
surface above is nearly transverse to the longitudinal axis. The width
contracts chiefly at the expense of the posterior margin to the lower part
of the upper fourth of the bone, the shaft below being of nearly uniform
width. The trochlear surface at the distal extremity is pulley-shaped,
covering about half of a circle, perhaps more, and has a moderately
deep groove. The sutural union between the tibia proper and the prox-
imal tarsal bone, or bones, is entirely obliterated. The bones of the two
sides are compressed in different positions, the left one from the side, the
right one from nearly in front ; the width above is considerably greater
in the former than in the latter, indicating but comparatively little ex-
pansion of the upper part from side to side.

There is no trace of any fibula, either in the preserved remains or in
any tibial articulation.

MEASUREMENTS OF TIBIA.

mm.

Length T 133(132)

Width above '. 15 ( 11)

Width of shaft at upper end of middle third-- 7 ( 6)

Width, lower third 6( 6)

Antero-posterior diameter of trochlea 6( 6)

Foot. But a single metatarsal and one phalange are present in this
specimen. These, however, agree so well with the corresponding bones
of Pteranodon that I doubt not the foot was quite similar to that of the
former genus. This metatarsal is expanded at its proximal end and
gently convex on the trochlear end. It measures 35 mm. in length, 2 mm.
in width at the upper fourth and 4 mm. at either extremity. The bone is
compressed from side to side. The phalange, flattened on the side, has
a length of 14 mm. The distal end has the extremity convex, its chord
rectangular to the long axis of the bone ; its width is 3 mm. The
proximal end is truncated or gently concave ; its width is about 4 mm.
The shaft in the middle is scarcely narrowed from the proximal end. I
repeat here, briefly, the description of the foot of Pteranodon, from my
paper in the Kansas University Quarterly, vol. vi, page 50, with some
changes.

The first four metatarsals are very slender, straight, and contiguous
with each other, each having a flattened proximal articular surface and
a rounded trochlear distal extremity. The metatarsal of the fifth toe is

152 Field Columbian Museum — Geology, Vol. II.

rudimentary, subtriangular in shape, with the obtusely pointed distal part
curved, altogether resembling an obtuse claw. The first toe has a single
phalange, which is long, cylindrical, gently curved, and obtusely pointed.
The second toe has two phalanges, the first of which is elongate and
grooved with a distal trochlear articulation. The second phalange
resembles the first of the first toe, but is shorter. The third toe has four
phalanges, the proximal one of which is like that of the second toe, but
is elongate; the second is broader than long; the third is like the first, but
is shorter; the fourth is a short, scarcely curved and obtusely pointed claw.
The fourth toe has five phalanges, of which the first, fourth, and fifth
resemble the first, third, and fourth of the third toe, the second and third
the second of the same toe. This toe is the longest. The fifth has no
phalanges. I may add that the foot from which these measurements were
taken was found together, with the bones in place or but little displaced,
and has been so mounted in the University of Kansas Museum. I have
repeatedly examined these bones, and am confident that the number of
phalanges are as I have given them ; that is, there is an absence of the
claws on the first and second toes. The measurements of this specimen
are as follows :

mm.

Femur, length 260

Diameter of head 20

Width of condyles - 30

Tibia, length 362

Metatarsals, length ^ -- 101, 105,95, 80, 25

Phalange, first digit, length 41

Phalanges, second digit, length 25, 31

Phalanges, third digit, length 35, 4, 27, 12

Phalanges, fourth digit, length 40, 3, 3, 26, 12

Ossified tendons. Numerous thin, flat, striated bones are lying on
the matrix associated with the anterior extremity, which can only be
regarded as tendinous ossifications. One of these on- each side is lying
by the side or upon the pteroid bone, with its pointed extremity near
its proximal end. They measure one hundred to one hundred and twenty
millimeters in length, with a greatest breadth of twelve millimeters. Two
others are lying by the side of the proximal end of the right wing
metacarpal; they are somewhat shorter and have the flattened fimbri-
ated end distad. Another is lying by the side of the distal end of the
same bone. There is a shorter, more fan-shaped one by the right humerus,
and fragments of others are scattered about the arms. Altogether there
are seven of these of about one hundred millimeters in length. Their
texture is striate, as though composed of slender, ossified, tendinous fibers,

On the Osteology of Nyctosaurus. 153

altogether reminding one very much of similar tendinous ossifications
in the legs of many birds.

A number of other slender, rib-like, more or less fragmentary bones
are observed scattered about the slab near the body of the specimen,
of which I am not at all certain, though I suspect that some of them
are ventral ribs, perhaps attached during life to the extremity of the
flat parasternal structures already described. One of these bones, lying
by the lateral margin of the sacrum is about forty millimeters in length
and is gently curved. It is flattened at one end, measuring nearly three
millimeters in width, cylindrical at the other, and not more than one
millimeter wide. Another, lying back of the pelvis, of about the same
length is also slightly curved, one millimeter wide at one end and
about three-tenths of a millimeter in width at the other end. There
are also parts of several other bones which seem to be similar to these.

SYSTEMATIC POSITION.

Three specimens ascribed to Nyctosaurus are now known — the type
described briefly by Marsh, the one in the University of Kansas Museum
described by me in 1892, and the present one. Marsh must have known
more than one specimen, however, in which additional specimens he
later recognized the essential generic characters. The three specimens
differ materially in size, that described by Marsh being the largest, and
the Kansas University specimen the smallest. In the smallest speci-
men the coracoid and scapula are still distinguished by a separable
suture. Furthermore, the separated arches of the dorsal vertebrae, and
the distinction of the elements of the atlas all indicate a young animal.
On the other hand, the present specimen has all of its sutures obliter-
ated, except that between the atlantal intercentrum and the axial inter-
centrum, characters which may be ascribed to an adult condition. Among
the characters given by Marsh, is the separated coracoid and scapula,
but I suspect that this was derived from a smaller specimen than the
one which he originally named Pteranodon gracilis. I do not think
that the absolute differences in size are sufficient to separate the speci-
mens specifically, in that we know that similar differences are ascribed
to other pterodactyls by some authors. In order, however, to make the
comparative differences between these specimens clearer I have reduced
them to a common standard, using the humerus as 100 in each. In addi-
tion, I give like comparative measurements of certain species of Pterodac-
tylus, derived from figures given by Zittel:

154 Field Columbian Museum — Geology, Vol. II.

Su SD SU g§ -gu «

2/- Sj 5£ «*> £iS 1

gte ^ gS 85 £*

z z z £ £ £

Humerus ioo ioo ioo ioo ^ioo ioo

Ulna... 165 165 ... 136 148 153

Metacarpal 257 275 ... 210 106 105

First Phalange 324 327 ... 268 161 160

Second Phalange 241 .... ... 172 142 129

Femur 93 93 — 106 109 107

Tibia 151 151 ... 150 153 144

Metacarpal ' 100 100 100 — — —

Ulna 64 60 62

First Phalange 126 iiq 115 — — —

It is observed that there is a striking agreement in these specimens,
except in the metacarpals, the differences elsewhere being scarcely more
than one might expect in the measurement of bones differently affected
by the compression to which they have been subjected. Now, we may
consider this variation of specific value and give to each of these three
specimens a different name; or we may, as is far more reasonable,
consider the metacarpal as a variable bone in the individual, as it is
in the species and groups, and refer them all to one species. This
I have no hesitation in doing. There is, then, but a single species
of Nyctosaurus now known among the specimens referred to the genus;
it should be called Nyctosaurus gracilis Marsh.

In specimens of Pteranodon ingens there is not a little variation in
size, especially in the length of the wing metacarpal, which I have
observed to vary from 580 to 615 millimeters, the longest that I have
ever known from the Kansas chalk. It is very certain, I think, that
absolute identity of size cannot be relied upon to distinguish the differ-
ent species of pterodactyls, nor do I think that the relative or propor-
tional lengths of the finger bones should receive too much weight;
the different species evidently varied individually within certain limits.

In my first paper on Nyctosaurus, I discussed briefly the so-called
species of Pteranodon, P. nanus, and P. comptus, suspecting that one or
both of them belonged with Nyctosaurus.

A wider acquaintance with Kansas pterodactyls, both in the laboratory
and in the field strengthens my belief that both of these names are
synonyms of N. gracilis.

No small species from the Kansas chalk can be referred definitely to
Pteranodon, so far as my own knowledge goes, and I have seen hundreds
of specimens. P. nanus presents the strong lateral crest of the humerus
which I am satisfied is a diagnostic characteristic of Pteranodon. Al-

On the Osteology of Nyctosaurus. 155

though the author expressly stated that the "coracoid and scapula were
firmly anchylosed" a character seen in the present specimen of Nycto-
saurus, I feel pretty confident that the type of P. nanus does not have the
articular extremity of the scapula so characteristic of Nyctosaurus. The
description of P. comptus might apply, except in size, to either Nycto-
saurus or Pteranodon. I refer the species to the former genus because of
its small size, but its synonymy is not assured. That a better basis for an
opinion may be given the reader, I here reproduce all that Marsh has
written upon Nyctosaurus and the two small species which he referred to
Pteranodon.

" One of the smallest American species yet found is represented in the
Yale Museum by several bones of the wing, a number of vertebrae and
the nearly complete pelvis. The wing bones preserved are elongated
and very slender. The pelvis is unusually small, and there are five
vertebrae in the sacrum. The last of the series indicates that the tail was

short. The following are the principal dimensions of this specimen:

mm.

Length of ulna 187

Length of metacarpal of wing finger — 300

Antero-posterior diameter of outer condyle at distal end 15

Tranverse diameter of shaft, above condyles . 13

Length of first phalanx of wing finger . 347

Extent of five vertebrae of sacrum 57

"This species, which may be called Pteranodon gracilis, was about two-
thirds the size of P. velox Marsh. It probably measured about ten feet
between the tips of the expanded wings."*

" Pteranodon comptus, sp. nov. The smallest Pterodactyle known from
American strata is indicated by portions of three skeletons in the Yale
Museum. Among these remains are two distal ends of the characteristic
metacarpal of the wing finger, other portions of the wing bones, and two
sacral vertebrae. The large metacarpal is very slender and elongated,
and its outer distal condyle has its superior margin elevated above the
shaft, and terminated proximally in a point. The ulna is comparatively
large, and the proximal carpal has an oval air cavity on its radial side.
The sacral vertebrae have their centra short, and medially constricted.

" The principal measurements of the remains of thisspecies are as follows:

mm.

Greatest diameter of ulna at distal end 15

Transverse diameter of proximal carpal 17

Antero-posterior diameter of outer distal condyle of wing metacarpal 12 8

Longitudinal extent of condyle 11.6

Transverse diameter of shaft above condyle -- 11.5

Length of medial sacral vertebra 9

Transverse diameter of centrum 8.4

♦Marsh, Amer. Journ. Sci. June, 1876, p. 508.

156 Field Columbian Museum — Geology, Vol. II.

" The above specimens are all from the Upper Cretaceous of Western
Kansas."*

" Nyctosaurus, gen., nov." A second genus of American Pterodactyls
is represented in the Yale Museum by several well-preserved specimens.
This genus is nearly related to Pteranodon, but may be readily distin-
guished from it by the scapular arch, in which the coracoid is not co ossi-
fied with the scapula. The latter bone, moreover, has no articulation at
the distal end, which is comparatively thin and expanded. The type of
this genus is Pteranodon gracilis Marsh, which may now be called Nycto-
saurus gracilis. It was a Pterodactyl of medium size, measuring about
eight to ten feet between the tips of the expanded wings. Its locality is
in the Upper Cretaceous of Western Kansas. The type specimens of all
the above species are preserved in the Museum of Yale College." f

" In the same geological horizon with the gigantic forms (Pterano-
don beds) the remains of a single small Pterodactyl have been found.
Th;s animal was more diminutive than the Jurassic species \Dermodactylus
montanus], having a spread of wings not more than three or four
feet. The jaws were proportionally more slender than in the larger Cre-
taceous species, and no teeth have been found with them. The humerus
had a small head and an enormous radial crest, which curved downwards.
The scapula and coracoid were firmly ankylosed. Some of the trunk
vertebrae have very long transverse processes, or ankylosed ribs, curved
backward. Some dimensions of this specimen are as follows :

Length of humerus 62

Greatest diameter of head 12

Transverse diameter across radial crest 30

Greatest diameter of distal end 16

Vertical diameter of humeral glenoid cavity 13

Transverse diameter 6

" This species may be called Pteranodon nanus. Its known remains
were found by Mr. S. W. Williston, in the Middle Cretaceous of Western
Kansas." %

"The name Nyctosaurus, applied by the writer to this group, appears
to have been preoccupied, and hence may be replaced by Nyctodactylus.
The only species known is Nyctodactylus gracilis.'''' jj

It is seen that but five vertebrae are referred to the sacrum of the
type of Nyctosaurus gracilis ; if this be a real character one would not
question the distinctness of the species. Inasmuch as Pteranodon has
six vertebrae in the sacrum and one or more lumbars or false sacrals, as in

♦Marsh, 1. c, p. 509.

t Marsh, Amer. Journ. Sci. Dec, 1876, p. 480.

% Marsh, Amer. Journ. Sci. April, 1881, p. 343.

II Marsh, Amer. Journ. Sci. April, 1881, p. 343, footnote.

On the Osteology of Nyctosaurus. 157

the present specimen of Nyctosaurus, I believe that this number will be
found common to all the Kansas pterodactyls. I think that Marsh must
have been in error in ascribing but five to the sacrum. Should he, how-
ever, be correct I would suggest the name N. leptodactylus for the present
species!

In view of the foregoing I offer the following as the synonymy of
this genus and species :

NYCTOSAURUS.

Marsh, Amer. Journ. Sci., xii, p. 480, Dec, 1876. Nyctodactylus
Marsh, Amer. Journ. Sci., xxii, p. 343, April, 1881 ; ibid., xxvii, p. 423,
May, 1884. Williston, Kans. Univ. Quart., i, p. 2, 5, 1893; Journ. Anat-
omy, i, p. 297, 1902 ; Journ. Geology, x, p. 520, July-Aug. 1902.

NYCTOSAURUS GRACILIS.

Pteranodon gracilis Marsh, Amer. Journ. Sci., xi, p. 508, June, 1876.

Nyctosaurus gracilis Marsh, Amer. Journ. Sci. xii, p. 480, Dec. 1876.

Nyctodactylus gracilis Marsh, Amer. Journ. Sci. xxi, p. 343, April,
1881; Williston, Kans. Univ. Quart., i, p. ii, 1893; Zittel's Paleontology
(Eastman), vol. ii, p. 255, Fig. 361, 1902.

(?) Pteranodon comptus Marsh, Amer. Journ. Sci. xi, 508, June, 1876;
Williston, Kans. Univ. Quart, i, p. 11, June, 1893.

(?) Pteranodoti nanus Marsh, Amer. Journ. Sci., April, 1881, p. 343;
Williston, Kans. Univ. Quart, i, p. 11, June, 1893.

Of the remaining species of Kansas pterodactyls, six have been
described, two of which, P. umbrosus and P. harpyia Cope, are admittedly
identical with previously described forms, leaving ingens, occidentalis, velox,
and longiceps Marsh. No real attempt was made to separate P. longi-
ceps from those previously named, nor could there have been, since only
the skull was known in this species to the describer, while the skull was
not known in either of the other species. It may be identical with
either occidentalis or velox. possibly with both. The characters given by
Marsh to distinguish the species have little value, though I believe there
are three distinct ones of the genus, known from the Kansas chalk.
Two of these species are at once distinguishable by the structure of the
humerus, especially of its radial or lateral crest. In P. ingens this crest
is rounded and obtuse; in one or the other of the smaller forms it is more
elongate and of a different shape. Furthermore, among the smaller
specimens there appear to be two different types of terminal wing pha-
lange, one nearly straight and the other curved falciform, as in P. ingens.
If the smaller form with the curved phalange should prove not to be the

158 Field Columbian Museum — Geology, Vol. II.

young of P. ingens, as is not likely, then we have at least three distinctly
different species of the genus, perhaps a fourth in P. comptus.

A much more interesting question is : What relation do the genera,
Nyctosaurus, Ornithocheirus, and Pteranodon (Omithostoma) bear to each
other and to Pterodactylus ?

The classification which has hitherto seemed the best to display the
relationship of these forms is the following :

Suborder Pterodactyloidea. Tail short; wing metacarpal, at least as
long as the bones of the antebrachium; fifth digit of pes vestigial, with-
out phalanges.

Family Pterodactylidae. Scapula not articulating with supraneural
facet.

Genus Pterodactylus. Teeth present.

Genus JVyctosaurus. No teeth.

Family Ornithocheiridae. Scapula with an articular facet united with
neural facet of the notarium.

Genus Omitkocheirus. Jaws with teeth.

Genus Pteranodon {Ornithostoma). Jaws without teeth.

However, a better knowledge of the genus JVyctosaurus convinces
me that its relationship is closer with Ornithocheirus than with Pterodacty-
lus; that it has gone much further in specialization in the same direction
with Ornithocheirus and Pteranodon. Very little now remains to be
known of the osteology of both these genera, and not much of Ortii.
thocheirus . Their characters as a group may be summed up as follows.
While some parts of the skeleton of Ornithocheirus are yet unknown, I
am confident that future discovery will produce nothing very discrepant.

Ornithocheiridoz. Head elongate, slender; no antorbital vacuity;
internal nares large; eight cervical, ten dorsal {Pteranodon?}, six
sacral, and ten to twelve caudal vertebrae present, the last dorsal co-os-
sified with the sacrum; first three or four dorsal vertebrae united into a
notarium, with stout, anchylosed ribs articulating with the sternum; pos-
terior dorsal ribs very slender, single-headed, and nearly straight; three
or four pairs of flattened V-shaped ventral ribs present. Ilium much
elongated; ischio-pubis with an obturator foramen; prepubes band-like,
co-ossified, with an anterior, flattened projection on eaCh side. Coraco-
scapula with only a slight projection below the glenoid cavity, co-ossified
in the adult; scapula but little longer than the coracoid, not narrowed
distally. Three carpals present; forearm much longer than humerus; third
and fourth metacarpals very short, pointed proximally, not articulating
with carpus; second metacarpal thread-like proximally; fifth metacarpal

On the Osteology of Nyctosaurus. 159

much longer than forearm; phalanges proportionally long. No fibula
present; two tarsals only; fifth digit without phalanges.

Ornithocheirina. Upper end of scapula thickened, articulating
with a supraneural plate. Head with a prominent parieto-occipital crest.

Ornithocheirus. Jaws with teeth.

Pteranodon (Omithostomd). Jaws wholly edentulous.

NyctosaurincB. Scapula flattened spatulate above, not articulating
with supraneural plate. Humeral crest constricted; humerus propor-
tionally short, the digit long; no sagittal crest.

Nyctosaurus, Ornithodesmus (?).

In contradistinction to these characters I would define the Ptero-
dactylida as follows: I have never seen any specimen of this genus, and
must, therefore, rely entirely upon the descriptions and figures. The
characters hence are in a measure provisional.

Pterodactylida. A distinct antorbital foramen present, sometimes
partly confluent with the nares; occiput not produced, jaws always with
teeth. No consolidation of the thoracic vertebrae or anchylosis of the
sternal ribs, the posterior ribs less slender. More than ten vertebras in
the dorsal region and less than six in the sacrum. Prepubes never co-ossi-
fied in the middle, and without anterior prolongation. Scapula longer
than the coracoid, more or less narrowed distally. Humerus and legs
relatively large, the forearm and wing finger relatively small; all the meta-
carpals articulating with the carpus. A fibula present.

I am sure that all these characters are generalized ones, even as the
genus is older than those placed in the first group. It seems to me
especially that the shortening of the back, the consolidation of the early
dorsal vertebrae, and the elongation of the sacrum, with the greater elonga-
tion of the wing digit, are better evidences of relationships than is the
presence of the peculiar scapular articulation.

Certain conclusions as to the habits of the Ornithocheiridd, as I here
define the family, seem inevitable from a better knowledge of their struc-
ture. Some of these conclusions I have already presented in a former
paper, and a further study only strengthens my belief in their justness.

The neck was strong, with limited torsion, but with great power and
effectiveness in the antero-posterior flexion, thus rendering the beak very
effective in striking.* The marked posterior projection of the hind
zygaphophyses of the last cervical vertebra, quite unlike those of any
other vertebrae in the column, indicates, not that the neck turned back-

* Plieninger (op. cit.) misquotes me in saying that " Williston glaubt, dass diese Art der
Articulation die Bewegung des Halses in verticaler Kichtung von vorne nach binten eingeschrankt
hat." What I said was: " Such a mode of articulation would seem to limit the motion to one in a
vertical, antero-posterior plane, while greatly strengthening the joints."— Kans. Univ. Quart., vi, p. 39.

160 Field Columbian Museum — Geology, Vol. II.

ward at this place, but that it curved forward more after the way of birds,
a position compatible with the upright attitude. The whole structure
of the neck, the elongated vertebras, the transverse centra, and lateral
exapophysial articulations, the absence of transverse processes, all remind
one of the neck in the cryptodire testudinates, where the neck is capable
of much and effective antero-posterior flexion.

The articulation of the head of the femur, with the basal plane of its
convexity nearly at right angles to the long axis of the bone, and the
posterior position of the acetabulum, carry conviction to me that the
femora could not have been brought parallel in the same direction with-
out dislocation from the sockets, whatever attitude the animal may
have assumed. The articulation of the distal end of the tibia further-
more shows that the metatarsus could not have been extended to a right
angle; that is, Pteranodon could not have been plantigrade. The greater
extension must have been between the metatarsus and phalanges; in other
words, the creature must have been digitigrade in ambulation. The flat-
tened proximal ends of the metatarsals indicate a compact and closely
united foot, and the bones have been so found in specimens of Pteranodon.
The toes in Pteranodon, and doubtless in this genus also, were practically
clawless, and the outer toes were much the longer. The animal was
incapable of seizing or holding with the feet.

Nor could a very great flexion at the knee have been possible, as I
think the figures of the femur of Pteranodon ingots will indicate. If
the animal was quadrupedal, it must have been in a crawling attitude,
with both legs and arms widely extended.

APATOMERUS MIRUS, GEN. ET. SP. NOV.

In the University of Kansas Quarterly, vol. iii, p. 3, I described and
figured a remarkable bone from the Lower Cretaceous of Clark County,
Kansas, which I hesitatingly referred to some crocodile -like animal,
because I was at a total loss where else to place it, my knowledge of the
pterodactyl anatomy then being less than at present. The figure and
description are reproduced in vol. iv of the University Geological Sur-
vey of Kansas, p. 90, as follows: "The upper end of a femur found in
the same region appears to belong to the same kind of an animal, as
does the vertebra described above. The shape is not unlike that of a
human femur, with the trochanters evidently small and placed much
below the level of the head. The neck is stout, the head gently convex,
with an angular border. The shaft below the trochanters is somewhat
flattened from before backward, but becomes more cylindrical below. The
shaft is hollow, with firm walls not more than one-third of an inch in
thickness. The portion preserved measures 210 mm."

On the Osteology of Nyctosaurus. 161

I am now satisfied that the femur in question belongs to a heavy-boned
pterodactyl or some allied, hitherto unknown, animal. The very peculiar
head of the bone with its articular convexity directed upward, the posi-
tion of the trochanter, the hollow shaft and its curvature are characters
all found in pterodactyls and in no other animals of which I have any
knowledge. The entire length of the bone must have been about 350 or
400 mm., and while the walls are unusually thick, they are not propor-
tionally much more so than in Dermodactylus, the type of which I
distinctly remember. Dermodactylus is from the so-called fresh-water
beds of the Wyoming Jurassic, which I have long believed to be con-
temporaneous with the Comanche beds of Kansas, and which I correlate
with the Wealden of Europe. The forms cannot be the same, since this
is very much larger.

May 14, 1903.

In an article by Mr. F. A. Lucas in the Report of the Smithsonian
Institution for 1901, there is given an excellent drawing of a skull of
Pteranodon {Ornithostomd), made "from a specimen in the Yale University
Museum." From my recollection of various peculiarities shown in the
figure, as also from certain accidental peculiarities, such as the fracture
of the infra-narial bar, faithfully reproduced in Professor Marsh's drawings,
I have very little hesitation in saying that the specimen there illustrated
is the type of the genus Pteranodon, upon which was based the figure so
long current in text-books. Should I be in error in this identification, I
shall be glad to be corrected, hoping, however, in that case, that a true
figure of the type specimen may be published.

In the summer of 1891, Professor E. C. Case and myself uncovered
from the firm yellow chalk of western Kansas, the uninjured posterior
part of a skull of Pteranodon, with its sagittal-crest entire and evenly
rounded. A figure of this specimen was given in PI. I, Vol. I, of the
Kansas University Quarterly, to which I beg to refer the reader. The
specimen is now preserved on a slab in the museum of the University of
Kansas, and I doubt not that the crest as disclosed from the matrix and
as represented in the figure, is all that the animal possessed in life. Mr.
G. F. Eaton, however, in the July number of the American Journal of
Science takes me to task for criticising the original figures of Pteranodon,
which he admits were faulty, and for my "restoration" of the skull, assert-
ing that I was wrong. He now adds to the type a much longer crest than
did the author. The basis for Professor Marsh's restoration is shown in
Mr. Lucas's plate, if my identification is correct, and the drawing has
been faithfully made.

162 Field Columbian Museum — Geology, Vol. II.

I do not criticise Professor Marsh for making the restoration as he
did; he was perhaps justified, though it might have been better had he
stated how much of the restoration was conjectural, and had also pub-
lished a figure of the specimen as it actually was.

In his present paper Mr. Eaton says: "At the present time of writing
an incomplete Pterodactyl skull is being worked out at the Yale Museum,
which will, in all probability prove to be that of Nyctodactylus Marsh.
The crest, which is apparently entire, is of small size compared with that
of Pteranodon, the measurement from occipital condyle to tip of crest
being only 49 mm., while the length from occipital condyle to tip of
beak was approximately 45 cm. In general the skull compares favorably
with that shown in Williston's restoration of Nyctodactylus, given in the
American Journal of Anatomy, Vol. I, No. 3, May 26, 1902, where he
states that the outline is taken in part from a specimen of Pteranodon
Marsh, or Omithostoma Seeley, as the genus was then called. It is there-
fore fair to infer that the apparent similarity of the two genera led Wil-
liston to draw inadvertently upon the characters of Nyctodactylus in mak-
ing his restoration of Pteranodon."

Had Mr. Eaton done me the honor to have read more attentively the
article which he quotes; or had he examined the extended article on the
skull of Nyctodactylus with plates, published in the Journal of Geology for
August, 1902; or even had he examined the figure of the skull in Zittel's
text-book of Paleontology, published last autumn, of all of which he
seems strangely ignorant, he would have learned that Nyctodactylus has
no crest whatever, not even a vestige of one! It is very evident, therefore,
that Mr. Eaton has made the very error which he so freely suggests I
have made, and that his specimen is in reality a Pteranodon/ I am
pleased to learn that its "apparently entire" "crest" compares favorably
with that which I have assigned to the same genus.

Now, since his own observations show that Pteranodon may have a
short crest, very much as I have figured it, it will be of interest to learn
more fully his reasons for attaching the elongated crest he has to the skull
as shown in Mr. Lucas's figure. I furthermore note that Mr. Eaton con-
tinues, in his "restoration" of the skull of Pteranodon, the gravest of the
inaccuracies of the original figures, the absence of the supratemporal
fossa and arch. If he is correct in this, the original reference of the
genus to a distinct order of reptiles was justified. But we will await fur-
ther evidence before assuming that he is, and meanwhile, it will be unfor-
tunate if his restoration finds currency.

Inasmuch as Mr. Eaton has so clearly shown his ignorance of recent
and very accessible literature on the subject, it is too much to expect that
he would have mentioned the fact that the pelvis of Pteranodon has been

On the Osteology of Nyctosaurus 163

described and partly figured without certain errors into which he has
fallen, and that the peculiar "spiral" condyles of both Pteranodon and
Nyctodactylus have likewise been described and figured, without, however,
their having suggested to any one the probability that the animals dis-
located their jaws in deglutition.

Plate XL.
Nyctosaurus gracilis.

View of part of specimen, showing sternum, pelvis, dorsal vertebrae, ribs
and arm bones. Natural size.

FIELD COLUMBIAN MUSEUM.

GEOLOGY, PLATE XL.

Nyctosaurus gracilis Marsh.

.ux =
.auioAflo eu

.

in gri]

■ .
' i! [aim I,

Plate XLI.
Nyctosaurus gracilis.

Fig. I. Dorsal vertebrae, ribs, and pelvis; about three-fifths natural size,
r, ribs; dv, lumbar or false sacral; s, sacrum; ip, ischio-pubis; ob, obturator for-
amen; il, ilium; ftp, prepubes.

Fig. 2. Abdominal ribs. About three-fifths natural size.

Fig. 3. Atlanto-axis, from the side. Twice natural size.

Fig. 4. The same from behind, partly restored.

Fig. 5. The same, from in front: a, atlantal neurapophysis; ac, axial cen-
trum; at, atlantal intercentrum; axi, axial intercentrum; ^,ball of axis; o, odontoid;
ftz, postzygapophysis.

Fig. 6. Hyoid bone. Three times natural size.

Fig. 7. Extremity a of Fig. 6, more enlarged, from the side.

FIELD COLUMBIAN MUSEUM.

GEOLOGY, PLATE XLI.

S. W. W. del.

Nyctosaurus.

Plate XLII.
Nyctosaurus gracilis.

Fig. i. Sternum, from below. Three-fifths natural size.
Fig. 2. Left radius, ulna, and carpus, from inner side, the carpus somewhat
displaced: a, radius; b, ulna; c, proximal carpal; d, distal carpal.
Fig. 3. Right lateral carpal or first metacarpal.
Fig. 4. Right pteroid bone, from in front.
Fig. 5. Right pteroid, from behind.
Fig. 6. Small metacarpal. Twice natural size.
Fig. 7. Left humerus, from within.
All the figures, except as stated, natural size.

FIELD COLUMBIAN MUSEUM.

GEOLOGY, PLATE XLII.

S. W. W. del.

Nyctosaurus.

.

I

■

' .a§n

■■■
■

Plate XLIII.

Fig. i. Left femur of Pteranodon ingens, somewhat obliquely compressed,
from in front showing dermal markings. One-half natural size.

Fig. 2. The same, posterior surface. One-half natural size.

Fig. 3. The same, distal end. One-half natural size.

Fig. 4. Right femur of Nyctosaurus gracilis, from in front. Natural size.

Fig. 5. The same, from behind. Natural size.

Fig. 6. Left coraco-scapula of Nyctosaurus gracilis; s, scapula; c, cora-
coid; a, process of scapula. Natural size.

Fig. 7. First dorsal vertebra of Nyctosaurus gracilis, from in front. Twice
natural size.

Fig. 8. Eighth cervical vertebra of Nyctosaurus gracilis, as compressed,
obliquely from behind: s, spine; b, ball; d, diapophysis; e, exapophysis. Twice
natural size.

Fig. 9. Metatarsal of Nyctosaurus gracilis. Natural size.

Fig. 10. Right distal carpal of Nyctosaurus gracilis, obliquely from behind.
Natural size.

Fig. 11. Proximal caudal or caudo-sacral vertebra of Nyctosaurus gracilis,
from above. Natural size.

FIELD COLUMBIAN MUSEUM.

GEOLOGY, PLATE XLIII.

S. W. W. del.

Pteranodon, Nyctosaiims.

.

Plate XLIV.

Fig. i. Right first wing phalange of Nyctosaurus gracilis. One-half nat-
ural size.

Fig. 2. Right wing metacarpal of Nyctosaurus gracilis, from below. One-
half natural size.

Fig. 3. Tibia of Nyctosaurus gracilis, from the side. Natural size.

Fig. 4. Left femur of Pteranodon occidentalism.) Marsh, from behind. One-
half natural size.

Fig. 5. Right second wing phalange of Nyctosaurus gracilis, from behind.
One-half natural size.

Fig. 6. Posterior cervical vertebra of Nyctosaurus gracilis, from the side.
Twice natural size.

Fig. 7. Fifth cervical vertebra of Nyctosaurus gracilis, from above. Twice
natural size.

Fig. 8. Lateral carpal of Pteranodon ingens, after Cope. One-half nat-
ural size.

FIELD COLUMBIAN MUSEUM.

GEOLOGY, PLATE XLIV.

S. W. W. del.

Pteranodon, Nyctosaurus.

\o