NYCTODACTYLUS.
The type species of this genus was described as follows by its author (loc. cit. supra):
“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 measurements of this specimen:
| Length of ulna | 187 | millim. |
| Length of metacarpal of wing-finger | 300 | |
| Antero-posterior diameter of outer condyle at distal end | 15 | |
| Transverse 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.”
In the December number of the same volume of the American Journal of Science, he described the genus as follows:
“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 distinguished from it by the scapular arch, in which the coracoid is not co-ossified with the scapula. The latter bone, moreover, has no articulation at its distal end, which is comparatively thin and expanded. The type of this species is Pteranodon gracilis Marsh, which may now be called Nyctosaurus gracilis. It was a Pterodactyl of medium size, measuring about eight to ten feet between the tips of the expanded wings.”
The specific description of this species rests solely upon the measurements; the other characters given are not only vague, but are also common to all the known species. The generic description, as it is seen, is based upon the structure of the coraco-scapula. It will also be observed that the characters are not drawn from the type specimen, as that did not include this part of the skeleton, according to the author’s statement. Of these two characters, the non-ossification of the coracoid and scapula is a somewhat doubtful one, as the same character may or may not occur in allied species, as, for example, in the species of Rhamphorhyncus (R. Muensteri Goldf.) described by the author himself. So incomplete and unsatisfactory are the characters thus given that Zittel, in his Handbuch, dismisses the genus with the brief remark, “noch unbeschrieben.”
Nevertheless, from the peculiar form of the scapula, and from my recollection of the specimens upon which the genus was based, I believe I have determined with certainty an excellent specimen in the Snow Museum of Kansas University as a member of it, and here give a sufficiently complete description to place the genus on a more secure foundation.
This specimen was collected by Professor E. E. Slosson, of Wyoming University, while a member of my party in western Kansas the past season. It was partly exposed upon a gently sloping surface of firm yellow chalk on the Smoky Hill river, in the vicinity of Monument Rocks. Originally, the nearly complete skeleton must have been preserved, but a number of the bones had been either wholly or partially washed away, in some cases leaving their imprint in the chalk. The bones uncovered, and now lying upon the chalk slab nearly in their natural relations, are a humerus, both radii and ulnae, a pteroid, the two carpals of one wrist, both wing metacarpals, a first and a last wing phalanx, both coraco-scapulae, the posterior part of the lower jaws, ilium, femur, sternum, numerous ribs and vertebrae. The two coraco-scapulae lie with their scapular ends nearly touching, and their coracoid ends separated by a space equivalent to the width of the sternal articulation. The two elements appear to have been imperfectly united and were probably not co-ossified. The inferior border of the coracoid, near the humeral articulation, has a greater expansion than is found in Pteranodon; its shaft is more rounded and less rugose, lacking especially the strong muscular markings upon the external surface. The articular surface does not appear to differ materially from that in Pteranodon. The scapula is of nearly the same length as the coracoid, but is much less stout. It is a thin, spatulate bone, slightly expanded at the distal extremity, where the margin is rounded, and without the characteristic oblique articular facet. It has no supra-glenoid expansion or process on the posterior proximal border, but has its margin nearly straight or gently concave from the articulation to its extremity. The space included between the bones of the two sides as they lie is a nearly regular, oval one, measuring ninety-five millimeters in its greater, forty-five in its lesser diameter.
The sternum lies at a little distance from the coraco-scapulae. It is an extremely thin bone, with a stout anterior, styliform projection, at the base of which, on either side, looking upward and outward, is the articular, trochlea-like surface for the sternal end of the coracoid. The width between these articular surfaces measures fifteen millimeters; the length of the process in front of the articulations is twenty-five millimeters. Immediately posterior to the articular surfaces, the bone expands nearly at right angles to the longitudinal axis to a width of about sixty millimeters. The thin lateral margins are nearly parallel with the longitudinal axis, and show three shallow emarginations between the four costal articular projections. The hind angles are nearly rectangular. The bone, as preserved, is only shallowly concave, and shows no true keel, though a more pronounced median convexity towards the front doubtless subserved the function of a carina in part.
The left humerus lies in position, and is especially characterized by its enormous deltoid crest (radial crest of Marsh), though otherwise slender. This crest is further removed from the head of the bone than is the case in species of Pteranodon. It is directed somewhat downward, and has its distal, gently convex, border about twenty-five millimeters in extent, while the width of the process midway between the extremity and the base measures but sixteen millimeters. The bicipital crest is also prominent. The bone is relatively shorter than in Pteranodon.
The humerus, as will be seen from the above description, and from the measurements given below, is remarkably like the same bone in Pteranodon nanus, as described by Marsh (l. c. supra), and but a little larger. In P. nanus, however, the coracoid and scapula are said to be firmly co-ossified, and the scapula has of course a different structure.
The skull has been, unfortunately, almost wholly washed away, a fragment of the cranial wall and the posterior part of the lower jaws alone remaining. It is impossible, hence, to say much concerning this part of the anatomy. The lower jaws show a different structure from that in Pteranodon. As they lie in their natural position, the width at the condyles is about twenty-four millimeters. The angular is less produced posterior to the articulation than in Pteranodon, indicating a less elongated and less powerful mandibular portion, an indication further borne out by the slenderness of the rami. The impression in the chalk shows the symphysis to begin ninety millimeters from the articulation. The width at this place could not have exceeded sixteen millimeters; and the entire length of the lower jaws could hardly have been more than one hundred and twenty-five millimeters. In the parts preserved, measuring seventy-five millimeters, there are no indications of teeth; yet it is not impossible that there may have been teeth in the anterior portion of the dentary, as in some species of Pterodactylus. I hardly think it probable, however.
There are seven cervical vertebrae preserved, apparently the full complement, as in Pteranodon and other members of the order. They differ in no especial respect from the corresponding vertebrae of Pteranodon, and, apparently, of Pterodactylus. The imperfectly anchylosed, possibly free, atlas shows three pieces, the odontoid process and the two slender lateral pieces. The lateral pieces are entirely free, with a thickened base and a slender, curved upper portion. The odontoid is gently concave in front, and seems to be imperfectly ossified with the axis; it occupies the lower part of the articulation, corresponding to the hypapophysis of the Pythonomorpha. The axis is the shortest of the remaining vertebrae, and has a well developed spine. The centrum is strongly convex behind, as are the remaining centra of the series. The following five vertebrae decrease gradually in length. The anterior ones have only a thin ridge or plate for the neural spine; the seventh, however, has a neurapophysis of some length. They are all, as is usually the case, somewhat distorted from pressure. The under side is flattened, apparently gently concave longitudinally, and with a lateral ridge terminating in an obtuse hypapophysis at each inferior hind angle.
In his discussion of the Pterosauria, Zittel says concerning the vertebrae: “zwischen oberen Bogen und Centrum ist keine Sutur zu bemerken.” Handbuch, iii, p. 776. In this he is in error, so far as the American forms are concerned. It is usually the case in the Kansas specimens of both genera that the neural arch of the post-cervical vertebrae is wholly or in part detached from the centrum, showing a sutural, and not anchylosed union in life. The centra of twelve vertebrae are preserved, in the present specimen, from the region back of the neck; in only five of them are the neural arches in any way attached. Three of these are evidently anterior thoracic, judging from their structure and the position in which they lie. The shortest of them, to which was attached a very large rib, and which was lying in front of the scapulae, may represent the first thoracic vertebra (a). Its centrum is fully as wide as long, is flat on the under surface, and has a large, stout, horizontal parapophysis near the anterior end. Just above this process for the attachment of the head of the rib, and separated by a deep notch, is a much more elongated, horizontal diapophysis for the tuberculum. The cup of the centrum is shallowly concave; the transverse, shallowly U-shaped ball is only a little convex.
Two other vertebrae (b), found close by the one just described, and possibly one or the other contiguous with it, differ remarkably in having no, or a rudimentary, parapophysial process, and in having the diapophyses much shorter. It is not impossible that a slight expansion at the lateral margins of the ball may represent small parapophyses. In Pteranodon there are at least four vertebrae with dia- and parapophyses. In the other vertebrae from this region the diapophyses are yet shorter and the neural spine stouter and broader. The other centra preserved are all shaped somewhat like the half of a cylinder, and are a little longer than broad. They have no distinct cup or ball. In two of them there is a very long, recurved parapophysial process, as though formed by an anchylosed rib, on each side; they are probably lumbar vertebrae.
Most of the ribs are very slender; a few are moderately thickened; one only is very stout; its measurements are given below.
The principal dimensions of this species can be got at with considerable certainty. Although two of the wing-phalanges and the bones of the foot are wanting, yet the relative proportions of those present agree so closely with those of the corresponding bones in Pteranodon, that there can be but little possibility of error in assuming the same proportions for the missing ones. The position of the ilium and femur, as also the ribs, show that they hold their natural relations to the pectoral arch. The tail, alone, can not be got at.
| Extreme expanse of wing-bones | 2400 | mm. | 7 ft. | 10 in. |
| Expanse of wings in life, approximated | 2000 | 6 | 6 | |
| Length of head, estimated | 150 | 6 | ||
| Length of neck | 128 | 5½ | ||
| Length of trunk | 165 | 6½ | ||
| Length of leg and foot, outstretched | 275 | 11 |
But one species has been described from the American Cretaceous smaller than the present one, Pteranodon nanus Marsh, in which the expanse of wings is given as not more than three or four feet. In this estimate the author is certainly in error. The size of the humerus, as given, is rather more than three-fourths that of the present species, and the expanse, hence, must be nearly five feet in life, or six feet as the bones lie outstretched.
As regards the specific determination of the present specimen, there must necessarily be some doubt until the species already named have been recognizably described. But three of the existing species can be taken into account, N. gracilis, P. comptus and P. nanus. That it can not be the last, has already been shown. In size, it agrees well with P. comptus, but the other characters throw no light upon the identity.
The measurements given of the type specimen of N. gracilis show the size to be materially greater,—a character, however, of subordinate value—greater slenderness, and a relatively shorter first wing-phalanx.
The relative lengths of wing-metacarpals, wing-phalanx and ulna in N. gracilis and the present specimen may be expressed as follows:
| Length of wing-metacarpal | 100 | 100 |
| Length of first wing-phalanx | 115.6 | 119.5 |
| Length of ulna | 62.3 | 60.4 |
It will be seen that not a single character has yet been given to distinguish the genus from Pterodactylus, and it is not at all impossible that it may prove to be the same; its location among the Pteranodontidae rests solely on the assumed absence of teeth, and that is a character yet wholly unknown.
The material now in the museum permits a fuller discussion of the relations and characters of this group of reptiles than has been hitherto attempted. Originally, they were described as constituting a new order, a view still held by its author and no one else. Lydekker, in his Paleontology and Catalogue gives them a subordinal value; Zittel only a family value, though expressing doubt as to their subordinal rank.
It seems very probable that the genus Nyctodactylus has no teeth in the jaws; it agrees in every other respect with the genus Pterodactylus, so far as known. If the genus has teeth it must be united with Pterodactylus. Now, in not a few species of this genus, the teeth are confined to the anterior end of the jaws, and their entire absence, unaccompanied by other structural differences, will hardly constitute an order, or even family.
But, leaving aside Nyctodactylus, it is very much of a question whether the differences between Pterodactylus and Pteranodon are sufficient to locate them in different families, let alone different suborders.
The two genera have the following in common: Tail short. Skull with more or less elongated, pointed jaws, and very small upper and lower temporal fossae. Narial opening large, confluent with the pre-orbital foramen. Cervical vertebrae elongated, with rudimentary spinous processes. Fore and hind extremities, quite alike.
Pteranodon differs from Pterodactylus, so far as that genus is known, in the united coracoscapulae and pubes, both of which characters are found in Rhamphorhynchus.
The sole family characters remaining then, for Pteranodon, are, absence of teeth, a supra-occipital crest, and the articulation of the upper end of the scapula. Now it seems evident that to place the pteranodonts in a group equivalent to all the other pterosaurs is unwarranted, and any classification that will not show the more pronounced relationships with Pterodactylus is faulty. I would, therefore, propose the following:
- Order Pterosauria.
- Family Pterodactylidae,
- subfamilies Pteranodontinae, Pterodactylinae.
- Family Rhamphorhynchidae.
- Family Ornithocheiridae.
As regards the geographical distribution of the Pteranodonts, they have hitherto been recognized only from Kansas, but I am firmly of the opinion that they occur in Europe, and, if so, it is very probable that the name Pteranodon must be eventually given up. In fact, a toothless form of Pterodactyl was described by Seeley as long ago as 1871, under the name of Ornithostoma. I cannot refer to his description at present, and can, therefore, give no opinion as to their identity. It seems certain that the peculiar form of the scapulae and their vertebral articulation[3] occur among some of the European forms, which would strengthen the belief that Pteranodon is also an European genus.
In view of the above, the practice of the American text-books in Geology in introducing generic names of characteristic fossils as names of the geological horizons whence they come, is very reprehensible, in my opinion. Even the late edition of Leconte’s Elements contains a long list of such names, the greater portion of which have been relegated to the limbo of synonymy by paleontologists. It is greatly to be desired that the name “Pteranodon Beds” shall not become established, so long as there is the least doubt of the validity of the name itself.
KANSAS MOSASAURS.
BY S. W. WILLISTON AND E. C. CASE.
PART I, CLIDASTES, WITH [PLATES II-VI].
The group of extinct Cretaceous reptiles known as the Mosasaurs or Pythonomorpha was defined by Cope, “to whom Science is so largely indebted for its present knowledge of this interesting order of reptiles” (Marsh), in 1869.[4] Although some of the characters assigned by him to the order have since been shown to be inapplicable, and the group to have less value, yet his name, Pythonomorpha, has been generally retained. Lydekker and Zittel have assigned to the group a subordinal value, as has also Marsh, though under a different name. Owen rejected it entirely, and Baur, more recently,[5] has united it with the Varanidae to form a super-family, as follows:
- Suborder Platynota.
- Super-family Varanoidea.
- Families Mosasauridae, Varanidae.
- Super-family Helodermatoidea.
- Family Helodermatidae.
The group, whatever may be its rank or position, includes, so far, the following genera: Mosasaurus Conyb., Liodon Owen, Platecarpus Cope, Clidastes Cope, Baptosaurus Marsh, Sironectes Cope, Plioplatecarpus Dollo and Hainosaurus Dollo. Pterycollasaurus Dollo, founded upon Mosasaurus maximilianus Goldf., is omitted as doubtful. All of these genera, save Plioplatecarpus and Hainosaurus, have been recorded from North America, Clidastes, Baptosaurus and Sironectes being peculiar to this country. Of these latter three genera, however, Clidastes alone is well known; but this genus is suspected by Lydekker of being the same as the imperfectly known European Geosaurus Cuvier. Thus it seems that the genera, or at least the most of them, have a wide distribution; Platecarpus, in fact, is said to occur in New Zealand.
In America, members of the group have been discovered in the Cretaceous deposits of New Jersey, Alabama, North Carolina, the upper Missouri region, Nebraska, Kansas and New Mexico. Probably nineteen-twentieths of all the known specimens, however, have been obtained in western Kansas. The material now in the University Museum, all from Kansas, comprises several hundred specimens of these animals, including, probably, the best ones known. It is upon this material that the following preliminary studies are chiefly based.
The genus Clidastes, as first described by Cope, was based upon two dorsal vertebrae of C. iguanavus, the type species, from New Jersey. Shortly afterward, however, he gave a full and careful generic description, as derived from an unusually good specimen of an allied species, C. propython, from Alabama. Only a little later, Marsh described a genus, which he called Edestosaurus, from Kansas, but without giving any real, distinctive differences from Clidastes, following the very reprehensible practice of naming supposed new forms in the hopes that future distinctive characters might be found. The genus Edestosaurus has been rejected by nearly all save the authors of the American text-books in Geology. It seems hardly necessary to point out the identity. The only distinctive character the author gave for his genus was the insertion of the pterygoid teeth, and even this character he modified later—“Palatine (sic) teeth more or less pleurodont.”[6]
This character, even were it real, is of very slight value; indeed it cannot be used to distinguish the species even.
Clidastes is, without doubt, one of the most highly specialized genera in the group, and, what is very interesting, is one of the latest. It occurs in Kansas in the uppermost part of the Niobrara beds, in the horizon so markedly characterized by the toothed birds. Both Platecarpus and Liodon occur, though in diminished numbers, almost to the very lowest portion, but Clidastes has never been found except towards the top. From measurements made the past season, the thickness of the beds in which these saurians occur cannot be less than six hundred feet.
The following species have been found in Kansas: none of them are known to occur elsewhere.