 CHAPTER XIV of DRAGONS OF THE AIR by Harry Seeley The stone's field slate in England, which corresponds in age with the lower part of the Great or Bath Oolite, yields many evidences of terrestrial life. Land plants, insects, and mammals preserved in a marine deposit. A number of isolated bones have been found of pterodactyls, some of them indicating animals of considerable size and strength. The nature of the limestone was unfavorable to the preservation of soft wing membranes or even to the bones remaining in natural association. Very little is known of the head of Ramphocephalus. One imperfect specimen shows a long temporal region which is wide and a very narrow interspace between the orbits, with a long face indicated by the extension of narrow nasal bones. The lower jaw has an edentulous beak or spear in front, which is compressed from side to side in the manner of the liacic forms, but turned upward slightly as endoringathis or campelonathis. Behind this extremity are sharp tall teeth, few in number, which somewhat diminish in size as they extend backward, and do not suddenly change to smaller series as in the liest genera. A few small vertebrae have been found, indicating the neck and back. The sacrum consists of five vertebrae. One small example has a length of only an inch. It is a little narrower behind than in front, and would be consistent with the animal having had a long tail, which I believe to have been present, although I have not seen any coddle vertebrae. The early ribs are like the early ribs of a crocodile or bird in the well-marked double articulation. The later ribs appear to have but one head. The shaped abdominal ribs are preserved. Much of the animal is unknown. The coracoid seems to have been directed forward, and as in a bird, it is two and one-half inches long. The humerus is three and one-half inches long, and the forearm measured six inches, so that it was relatively longer than in dimorphodon. The metacarpus is one and three-fourths inches long. The wing finger was exceptionally long and strong. Professor Huxley gave its length at twenty-nine inches. My own studies lead to the conclusion that the first finger bone of the wing was the shorter, and that although they did not differ greatly in length, the second was probably the longest, as in campylonathus. Professor Huxley makes the second and third phalanges seven and three-fourths inches long, and the first only about three-eighths inch shorter, while the fourth phalange is six and one-half inches. These measurements are based upon some specimens in the Oxford University Museum. There is only one first phalange which has a length of seven and three-fourths inches. The others are between five and six inches, or but little exceed four inches, so that as all the fourth phalanges which are known have a length of six and one-half inches, it is possible that the normal length of the first phalange in the larger species was five and one-half inches. The largest of the phalanges, which may be classed as second or third, is eight and one-half inches, and that, I suppose, may have been associated with the seven and three-fourth inches first phalange. But the other bones which could have had this position all measure five and one-half and seven and three-fourths inches. The three species indicated by finger bones may have had the measurements. One, seven and three-fourths, five and one-half, four and one-half. Two, eight and one-half, seven and three-fourths. Three, seven, five and one-half. Four, six and one-half, four and one-half. The femur is represented by many examples, one three and three-fourth inches long, and others less than three inches long, two and nine-tenths. In Campylonathus, which has so much in common with the jaw and the wing bones in size, the upper leg bone is two and eight-tenths inches. Therefore, if we assign the larger femur to the larger wing, the femur will be relatively longer in all species of Ramphocephalus than in Campylonathus. Only one example of a tibia is preserved. It is three and one-half inches long, or only one-tenth inch shorter than the bone in Campylonathus, which has the femur two and eight-tenths inches, so that I refer the tibia of Ramphocephalus to the species which has the intermediate length of wing. These coincidences with Campylonathus establish a close affinity and may raise the question whether the upper lyus species may not be included in the Stonesfield slate genus Ramphocephalus. The late Professor Phillips, in his Geology of Oxford, attempted a restoration of the Stonesfield ornithosar and produced a picturesque effect, page 164. But no restoration is possible without such attention to the proportions of the bones as we have indicated. Oxford Clay A few bones of flying reptiles have been found in the lower Oxford Clay near Peterborough and others in the upper Oxford Clay at St. Ives in Huntingdonshire. A single tail vertebra from the middle Oxford Clay near Oxford long since came under my own notice and shows that these animals belong to a long tail type like Campylonathus. The cervical vertebrae are remarkable for being scarcely longer than the dorsal vertebrae and the dorsal are at least half as long again as is usual having rather the proportion of bones in the back of a crocodile. Lithographic Slate Long tail pterodactyls are beautifully preserved in the lithographic limestone of the south of Bavaria at Solenhofen and the quarries in its neighbourhood often with the skeleton or a large part of it flattened out in the plain of bedding of the rock. Fine skeletons are preserved in the superb museum at Munich, at Heidelberg, Bonn, Harlem, and London and are all referred to the genus Ramphorincus or Tuscaphonathus. It is a type with powerfully developed wings and a long stiff tail very similar to that of dimorphodon so that some naturalists refer both to the same family. There is some resemblance. The type which is most like dimorphodon is the celebrated fossil at Bonn, sometimes called pterodactyls crassarostris, which in a restored form with a short tail has been reproduced in many textbooks. No tail is preserved in the slab and I ventured to give the animal a tail for the first time in a restoration. Page 163 published by the illustrated London News in 1875, which accompanied a report of a royal institution lecture. Afterwards in 1882, Professor Zittle of Munich published the same conclusion. The reason for restoring the tail was that the animal had the head constructed in the same way as pterodactyls with a long tail and showed differences from types in which the tail is short. And there is no known short tail pterodactyl with wrist and hand bones such as characterize this animal. The side of the face has a general resemblance to the pterodactyls from the lyus, for although the framework is firmer, the four apertures in the head are similarly placed. The nostril is rather small and elongated and ascends over the larger antorbital vacuity. The orbit for the eye is the largest opening in the head, so that these three apertures successively increase in size and are followed by the vertically elongated postorbital vacuity. The teeth are widely spaced apart and those in the skull extend some distance backward to the end of the maxillary bone. There are few teeth in the lower jaw and they correspond to the large anterior teeth of dimorphodon, there being no teeth behind the nasal opening. The lower jaw is straight and the extremities of the jaws met when the mouth was closed. The breast bone does not show the keel which is so remarkable in Ramphorincus, which may be attributed to its underside being exposed so as to exhibit the pneumatic foramina. The ribs have double heads, more like those of a crocodile in the region of the back than is the case with the bird-like ribs from Stonesfield. The second joint in the wing-finger may be longer than the first, a character which would tend to the association of this pterodactyl with species from the lyus, a relation to which was first drawn by Mr. E. T. Newton, who described the Whitby skull. The pterodactyls from the sullenholfen slate, which possess long tails, have a series of characters which show affinity with the other long-tailed types. The jaws are much more slender. The orbit of the eye in Ramphorincus is enormously large and placed vertically above the articulation for the lower jaw. Immediately in front of the eye are two small and elongated openings, the hinder of which, known as the antorbital vacuity, is often slightly smaller than the nostril, which is placed in the middle length of the head, or a little further back, giving a long dagger-shaped jaw which terminates in a toothless spear. The lower jaw has a corresponding sharp extremity. The teeth are directed forward in a way that is quite exceptional. Nonwithstanding the massiveness and elongation of the neck vertebrae, which are nearly twice as long as those of the back, the neck is sometimes only about half the length of the skull. All these long-tailed species from the lithographic stone agree in having the sternum broad, with a long strong keel extending far forward. The coracoid bones extend outward like those of a crocodile, so as to widen the chest cavity instead of being carried forward, as the bones are in birds. These bones in this animal were attached to the anterior extremity of the sternum, so that the keel extended in advance of the articulation, as in other pterodactyls. The breadth of the sternum shows that, as in mammals, the fore part of the body must have been fully twice the width of the region of the hip girdle, where the slenderer hind limbs were attached. The length of the fore limb was enormous, for although the head suggests an immense length relatively to the body, nearly equal to the neck and back together, the head is not more than a third of the length of the wing bones. The wing bones are remarkable for the short powerful humerus with an expanded radial crest, which is fully equal in width to half the length of the bone. Another character is the extreme shortness of the metacarpus, usually associated with immense strength of the wing metacarpal bone. The hind limbs are relatively small and relatively short. The femurs usually shorter than the humerus, and the tibia is much shorter than the ulna. The bones of the instep, instead of being held together firmly, as in the liest genera, diverge from each other, widening out, though it often happens that four of the five metatarsals differ but little in length. The fifth digit is always shorter. The hip girdle of bones differs chiefly from other types in the way in which those bones, which have sometimes been likened to the marsupial bones, are conditioned. They may be a pair of triangular bones which meet in the middle line, so that there is an outer angle like the arm of a capital Y. Sometimes these triangular bones are blended into a curved, bow-shaped arch, which in several specimens appears to extend forward from near the place of the articulation of the femur. This is seen in fossil skeletons at Heidelberg and Munich. It is possible that this position is an accident of preservation and that the pre-pubic bones are really attached to the lower border of the pubic bones. Immense as the length of the tail appears to be, exceeding the skull and remainder of the vertebral column, it falls far short of the combined length of the phalanges of the wing finger. The power of flight was manifestly greater in Rampharencus than in other members of the group, and all the modifications of the skeleton tend towards adaptation of the animals for flying. The most remarkable modification of structure at the extremity of the tail was made known by Professor Marsh in a vertical leaf-like expansion in this genus, which had not previously been observed, page 161. The vertebrae go on steadily diminishing in length in the usual way, and then the ossified structures, which bordered the tail bones and run parallel with the vertebrae in all the Rampharencus family, suddenly diverged downward and upward at right angles to the vertebrae, forming a vertical crest above and a corresponding keel below. And between these structures, which are identified with the neural spines and chevron bones of ordinary vertebrae, the membrane extends, giving the extremity of the tail a rudder-like feature, which, from knowledge of the construction of the tail of a child's kite, may well be thought to have had influence in directing and steadying the animals' movements. There are many minor features in the shoulder girdle, which show that the coracoid, for example, was becoming unlike that bone in the liais, though it still continues to have a bony union with the elongated shoulder blade of the back. The great German delineator of these animals, von Meyer, admitted six different species. Mr. Newton and Mr. Leidecker diminished the number to four. It is not easy to determine these differences, or to say how far the differences observed in the bones characterize species or genera. It is certain that there is one remarkable difference from other and older pterodactyls, in that the last or fourth bone in the wing finger is usually slightly longer than the third bone which precedes it. There is a certain variability in the specimens which makes discussion of their characters difficult and has led to some forms being regarded as varieties, while others, of which less material is available, are classed as species. I am disposed to say that some of the confusion may have resulted from specimens being wrongly named. Thus, there is a rampharencus called Kurtamanus, or the form with the short hand. It is represented by two types. One of these appears to have the humerus short, the ulna and radius long, and the finger bones long. The other has the humerus longer, the ulna much shorter, and the finger bones shorter. They are clearly different species, but the second variety agrees in almost every detail with a species named Hyrudnaceous, the swallow-like rampharencus. This identification shows not that the latter is a bad species, but that Kurtamanus is a distinct species which had sometimes been confounded with the other. While most of these specimens show a small but steady decrease in the length of the several wing finger bones, the species called Jaminji has the first three bones absolutely equal and shorter than in the species Kurtamanus, Launjamanus, or Hyrudnaceous. In the same way, on the evidence of facts, I find myself unable to join in discarding Professor Marsh's species Phyleris on account of the different proportions of its limb bones. The humerus, metacarpus, and third phalange of the wing finger in Rampharencus Phyleris are exceptionally short as compared with other species. Everyone agrees that the species called Longicatus is a distinct one, so that it is chiefly in slight differences in the proportions of constituent parts of the skeleton that the types of the Rampharencus are distinguished from each other. I cannot quite concur with either Professor Zittle, page 58-3, or Professor Marsh, figure 58-2, in the expansion which they give to the wing membrane in their restorations. For although Professor Zittle represents the tail as free from the hind legs, while Professor Marsh connects them together, they both concur in carrying the wing membrane from the tip of the wing finger down to the extremity of the ankle joint. I should have preferred to carry it no further down the body than the lower part of the back, there being no fossil evidence in favor of this extension so far as specimens have been described. Neither the membranous wings figured by Zittle nor by Marsh would warrant so much body membrane as the Rampharencus has been credited with. I have based my restoration, page 161, of the skeleton chiefly on Rampharencus Phyleris. The short tail types. The pterodactylia are less variable, and the variation among the species is chiefly confined to relative length of the head, length of the neck, and the height of the body above the ground. The tail is always so short as to be inappreciable. Many of the specimens are fragmentary, and the characters of the group are not easily determined without careful comparisons and measurements. The bones of the forelimb and wing finger are less stout than in the Rampharencus type, while the femur is generally a little longer than the humerus, and the wing finger is short in comparison with its condition in Rampharencus. The short tail pterodactyls give the impression of being active little animals, having very much the aspect of birds upon four legs or two. The neck is about as long as the lower jaw. The antorbital vacuity in the head is imperfectly separated from the much larger nasal opening. The orbit of the eye is large and far back. The teeth are entirely in front of the nasal aperture, and the postorbital vacuity is minute and inconspicuous. The sternum is much wider than long, and no specimens give evidence of a manubrium. The finger bones progressively decrease in length. The prepubic bones have a partially expanded fan-like form and never show the tri-radiate shape and are never ankylosed. About fifteen different kinds of pterodactyls have been described from the sullenhofen slate, mostly referred to the genus pterodactylis, which comprises forms with a large head and long snout. Some have been placed in a genus ornithosephalus, or tenodrakin, in which the head is relatively short. The majority of the species are relatively small. The skull in ornithosephalus brevarostris is only one inch long, and the animal could not have stood more than one and one-half inches to its back standing on all fours, and but little over two and a half inches standing as a biped on the hind limbs. A restoration of the species called pterodactylis scolopasaceps, published in 1875 in the illustrated London News in the position of a quadruped, which is an animal a little larger, with a body two and one-half inches high and six to seven inches long, with the wing finger four and one-half inches long. Larger animals occur in the same deposit, and in one named pterodactylis grandis, the leg bones are a foot long, and such an animal may have been nearly a foot in height to its back standing as a quadruped, though most of these animals have the neck flexible and capable of being raised like the neck of a goose or a deer, page 30, and bent down like a ducks when feeding. The type of the genus pterodactylis is the form originally described by Cuvier as pterodactylis longerostris, page 28. It is also known as P. antiquius, that name having been given by a German naturalist after Cuvier had invented the genus and before he had named the species. There are some remarkable features in which Cuvier's animal is distinct from others which have been referred to the same genus. Thus the head is four and one-half inches long, while the entire length of the backbone to the extremity of the tail is only six and one-half inches, and one vertebra in the neck is at least as long as six in the back, so that the animal has the greater part of its length in the head and neck, although the neck includes so few vertebrae. Nearly all the teeth, which are few in number, short and broad, not exceeding a dozen in either jaw, are limited to the front part of the beak and do not extend anywhere near the nasal vacuity. This is not the case with all. In the species named P. cocci, which I have regarded as the type of a distinct genus, there are large teeth in the front of the jaw corresponding to those of pterodactylus, and behind these a smaller series of teeth extending back under the nostril, which approaches close to the orbit of the eye without any indication of a separate antorbital vacuity. On those characters, the genus Diopacephalus was defined. It is closely allied to pterodactylus. Both agree in having the ilium prolonged forward more than twice as far as it is carried backward, the anterior process covering about half a dozen vertebrae as in pterodactylus longerostris. A great many different types have been referred to pterodactylus cocci, and it is probable that they may eventually be distinguished from each other. The species in which the upper borders of the orbits approximate could be separated from those in which the frontal interspace is wider. It is a remarkable feature in these animals that the middle bones of the foot, termed instep bones or metatarsals, are usually close together so that the toes diverge from a narrow breath as in p-longerostris, p-cocci, and other forms. But there also appear to be split-footed groups of pterodactyls like the species which have been named P. elegans and P. micronyx, in which the metatarsus widens out so that the bones of the toes do not diverge, and that condition characterizes the tenodrakin, pterodactylus brevorostris, to which genus these species may possibly be referred. Nearly all who have studied these animals regard the singularly short-nosed species, P. brevorostris, as forming a separate genus. For that genus, Sommerring's descriptive name, or Nithocephalus, which he used for pterodactyls generally, might perhaps have been retained. But the name tenodrakin, suggested by Mr. Leidecker, has been used for these types. Some of the largest specimens preserved at Stuttgart and Tubingen have been named pterodactyls suvicus and P. fragii. They do not approach the species P. grandus in size so far as can be judged from the fragmentary remains figured by von Meyer. For what appears to be the third phalange of the wing finger is seven and one half inches long, while in these species, it is less than half that length, indicating an enormous development of wing relatively to the length of the hind limb, which would probably refer the species to another genus. Pterodactyls suvicus differs from the typical pterodactyls in having a rounded flattened undersurface to the lower jaw, instead of the common condition of a sharp keel in the region of the symphysis. The beak also seems flattened and swan-like, and the teeth are limited to the front of the jaw. There appear to be some indications of small nostrils which look upward like the nostrils of Ramphorincus, but this may be a deceptive appearance, and the nostrils are large lateral vacuities which are in the position of antorbital vacuities so that there would appear to be only two vacuities in the side of the head in these animals. The distinctive character of the skeleton in this genus is found in the extraordinary length of the metacarpus and in the complete ossification of the smaller metacarpal bones throughout their length. The metacarpal bones are much longer than the bones of the forearm and about twice the length of the humerus. The first wing phalange is much longer than the others, which successively and rapidly diminish in length so that the third is half the length of the first. There are differences in the pelvis. For the anterior process of the ilium is very short in comparison with its length in the genus pterodactylus. And the long stalk of the prepubic bone with its great hammer-headed expansion transversely in front gives those bones a character unlike other genera, so that Cycno-Rampus ranks as a good genus, easily distinguished from Cuvier's type, in which the four bones of the wing are more equal in length, and the last is more than half the length of the first, while the metacarpus in that genus is only a little longer than the humerus and much shorter than the ulna. The pterodactylus suvicus has the neck vertebrae flat on the underside and relatively short as compared with the more slender and narrower vertebrae of P. phrasii. End of Chapter 14 Chapter 15 of Dragons of the Air by Harry Sealy This LibriVox recording is in the public domain. Chapter 15 Ornithosaurus from the Upper Secondary Rocks When staying at Swanage in Dorsetshire many years ago, I had the rare good fortune to obtain from the perbec beds the jaw of a pterodactyl, which had much in common in plan with the Cycno-Rampus phrasii from the lithographic slate which is preserved at Stuttgart. The tooth-bearing part of this lower jaw is eight inches long as preserved, extending back three inches beyond the symphysis portion in which the two sides are blended together. It is different from Professor Frost's specimen in having the teeth carried much further back and in the animal being nearly twice as large. This fragment of the jaw is little more than one foot long, which is probably less than half its original length. A vertebra nearly five inches long, which is more than twice the length of the longest neck bones in the Stuttgart fossil, is the only indication of the vertebral column. Professor Owen described a wing finger bone from these perbec beds, which is nearly one foot long. He terms it the second of the finger. It may be the third, and on the hypothesis that the animal had the proportions of the Solenhofen fossil just referred to, the first wing finger bone of the English perbec pterodactyl would have exceeded two feet in length and would give a length for the wing finger of about five feet three inches. For this animal, the name Dorado rancous was suggested, but at present I am unable to distinguish it satisfactorily from sycnoramphus, which it resembles in the forms both of the neck bones and of the jaw. Very small pterodactyls are also found in the English perbec strata, but the remains are few and scattered like these larger bones. Ornithodesmus latidens. The wielden strata, being shallow freshwater deposits, might have been expected to supply better knowledge of pterodactyls than has hitherto been available. Jaws of Ornithoceros sagittarostris have been found in the beds at Hastings and in other parts of Sussex. Some fragments are as large as anything known. The best preserved remains have come from the Isle of Wight and were rewards to the enthusiastic search of the Reverend W. Fox of Brixton. In the principal specimen, the teeth were short and wide, the head large and deep, with large vacuities, but the small brain case of that skull is bird-like. The neck bones are two-and-one-half inches long. In the upper part of the back, the bones are united together by ankylosis, so that they form a structure in the back like a sacrum, which does not give attachments to the scapula as in some pterodactyls from the chalk, but the bones are simply blended as in the frigate bird, allied to pelicans and cormorants. And then, after a few free vertebrae in the lower part of the back, succeeds the long sacrum, formed in the usual way of many vertebrae. I described a sacrum of this type from the wildin beds under the name Ornithodesmus, referable to another species, which in many respects was so like the sacrum of a bird that I could not at the time separate it from the bird type. This genus has a sternum with a strong deep keel and the articulation for the coracoid bones placed at the back of the keel in the usual way, but with a relation to each other seen in no genus hitherto known, for the articular surfaces are wedge-shaped instead of being ovate, and instead of being side-by-side, they obliquely overlap, practically as in wading birds like the heron. I have never seen any pterodactyl teeth so flattened and shaped like the end of a lancet, and from this character the form was known between Mr. Fox and his friends as Lattitans. The name Ornithodesmus is as descriptive of the sternum as of the vertebral column. The wing bones, as far as they are preserved, have the relatively great strength in the forelimb, which is found in many of the pterodactyls of the Cretaceous period, and are quite as large as the largest from the Cambridge green sand. In the Sussex species named P. sagittarostris, the lower jaw articulation was inches wide. A few pterodactyls' bones have been discovered in the neocomium sands of England and Germany, and other larger bones occur in the galt of Folkestone and the north of France, but never in such association as to throw light on the aspect of the skeleton. Ornithochiris. Within my own memory, pterodactyl remains were equally rare from the Cambridge green sand. The late Professor Owen, in one of his public lectures, produced the first few fragments received from Cambridge, and with a knowledge which, in its scientific method, seemed to border on the power of creation, produced again the missing parts so that the bones told their story, which the work of waves and mineral changes in the rock had partly obliterated. Subsequently, good fortune gave me the opportunity during ten years to help my university in the acquisition and arrangement of the finest collection of remains of these animals in Europe. Out of an area of a few acres during a year or two, came the thousand bones of ornithosaurs, mostly associated sets of remains, each a part of a separate skeleton, described in my published catalogs, as well as the best of those at York and in the British Museum and other collections in London. The deposit which yields them, named Cambridge green sand, may or may not represent a long period of time in its single foot of thickness, but the abundance of fossils obtained whenever the workmen were adequately remunerated for preserving them, would suggest that the pterodactyls might have lived like seabirds or in colonies like the penguins. If it were not that the number of examples of species found is always small and the many variations of structure suggested rather that the individuals represent the life of many lands. The collections of remains are mostly from villages in the immediate vicinity of Cambridge, such as Chesterton, Huntington Road, Coldham Common, Hassellingfield, Barton, Schillington, Ditten, Grandchester, Harston, Barrington, stretching south to Ashwell in Bedfordshire, on the one hand, as well as further north by Horningsea into the Fens. Each appears to be the associated bones of a single individual. The remains mostly belong to comparatively large animals. Some were small, though none have been found so diminutive as the smallest from the Solenhofen slate. The largest specimens with long jaws appear to have had the head measuring not more than 18 inches in length, which is less than half the size of the great toothless pterodactyls from Kansas. The Cambridge specimens manifestly belong to at least three genera. Something may be said of the characters of the large animals which are included in the genus Ornithocaris. These fossils have many points of structure in common with the great American toothless forms which are of similar geological age. The skull is remarkable for having the back of the head prolonged in a compressed median crest which rose above the brain case and extended upward and over the neck vertebrae, so as to indicate a muscular power not otherwise shown in the group. For about three inches behind the brain, this wedge of bone rested on the vertebrae and probably overlapped the first three neural arches in the neck. Another feature of some interest is the expansion of the bone which comes below the eye. In birds, this mallar or cheekbone is a slender rod, but in these pterodactyls it is a vertical plate which is blended with the bone named the quadrate bone which makes the articulation with the lower jaw in all oviparous animals. The beak varies greatly in length and in form though it is never quite so pointed as in the American genus for there is always a little truncation in front when teeth are seen projecting forward to a position somewhat above the palate. The snout is often massive and sometimes club shaped. Except for these variations of shape in the compressed snout which is characterized by a ridge in the middle of the palate and a corresponding groove in the lower jaw and the teeth, there is little to distinguish what is known of the skull in its largest English green-sand fossils from the skull remains which abound in the chalk of Kansas. This English genus ornithocaris represented by a great number of species had the neural arch of the neck bones expanded transversely over the body of the vertebra in a way that characterizes many birds with powerful necks and is seen in a few pterodactyls from Solenhofen. It is difficult to resist the conclusion that the neck vertebrae were not usually more than twice to three times as long as those of the back and it would appear that the caudal vertebrae in the English Cretaceous types were comparatively large and about twice as long as the dorsal vertebrae. Unless there has been a singular succession of accidents in the association of these vertebrae with the other remains ornithocaris had a tail of moderate length formed of a few vertebrae as long as those of the neck, though more slender, quite unlike the tail in either the long-tailed short-tailed groups of Solenhofen pterodactyls and longer than in the toothless pterodactyls of America. The singular articulation for the humerus at the truncated extremity of the coracoid bone is a character of this group as is the articulation of the scapulae with the neural arches of the dorsal vertebrae at right angles to them, page 115, instead of running over the ribs as in birds and as in other pterodactyls. The smaller pterodactyls have their jaws less compressed from side to side. The upper arm bone, the humerus, instead of being truncated at its lower end, as in ornithocaris, is divided into two or three rounded articular surfaces. That for the radius the bone which carries the wrist is a distinct and oblique rounded facet while the ulna has a rounded and pulley-like articulation on which the hand may rotate. These differences are probably associated with an absence of the remarkable mode of union of the scapulae with the dorsal vertebrae. But I have hesitated to give different names to these smaller genera because no example of scapula has come under my notice which is not truncated at the free end. I do not think this European type can be the nictodactyls of Professor Marsh in which sutures appear to be persistent between the bodies of the vertebrae and their arches because no examples have been found at Cambridge with the neural arches separated although the scapula is frequently separated from the coracoid in large animals. Ornithostoma The most interesting of all the English pterodactyl remains is the small fragment of jaw figured by Sir Richard Owen in 1859 which is a little more than two inches long and an inch wide distinguished by a concave pallet with smooth rounded margins to the jaws and a rounded ridge to the beak. It is the only satisfactory fragment of the animal which has been figured and indicates a genus of toothless pterodactyls for which the name Ornithostoma was first used in 1871. After some years, Professor Marsh found toothless pterodactyls in Kansas and indicated several species. There are remains to the number of 600 specimens of these American animals in the Yale Museum alone. But very little was known of them till Professor Williston of Lawrence in Kansas described the specimens from the Kansas University Museum when it became evident that the bones of the skeleton are mostly formed on the same plan as those of the Cambridge green sand genus Ornithochorus. They are not quite identical. Professor Williston adopts for them the name Ornithostoma in preference to Pteranodon which Marsh had suggested. Both animals have the dagger shaped form of jaw with corresponding height and breadth of the pallet. The same flattened sides to the snout converging upwards to a rounded ridge. The same compressed rounded margin to the jaw which represents the border in which teeth are usually implanted. And in both, the pallet has the same smooth character forming a single wide concave channel. Years previously, I had the pleasure of showing to Professor Marsh the remarkable characters of the jaw, shoulder girdle bones, and scapulae in the green sand pterodactyls while the American fossils were still undiscovered. I subsequently made the restoration of the shoulder girdle, page 115. Professor Williston states to me that the shoulder girdle bones in American examples of Ornithostoma have a close resemblance to those of Ornithocaris figured in 1891 as is evident from remains now shown in the British Museum. It appears that the Kansas bones are almost invariably crushed flat so that their articular ends are distorted. The neck vertebrae are relatively stout as in Ornithocaris. The hip girdle of the American Ornithostoma can be closely paralleled in some English specimens of Ornithocaris, though each prepubic bone is triangular in the American fossils as in P. Rampfestinas. They are united into a transverse bar as in Rampforincus, unknown in the English fossils. The femur has the same shape as in Ornithocaris and the long tibia terminates in a pulley. There is no fibula. The sternum in both has a manubrium or thick keel mass prolonged in front of its articular facets for the coracoid bones, which are well separated from each other. Four ribs articulate with its straight sides. The animal has four toes and the fifth is rudimentary. There are no claws to the first and second. In the restoration which Professor Williston has made, the wing metacarpal is long and in the shortest specimen measures one foot seven inches and in the longest one foot eight inches. This is exactly equal to the length of the first phalange of the wing finger. The second wing finger bone is three inches shorter, the third is little more than half the length of the first, while the fourth is only six and three fourths inches long, showing a rapid shortening of the bones, a condition which may have characterized all the cretaceous pterodactyls. The short humerus, about one foot long and the forearm, which is scarcely longer, are also characteristic proportions of Ornithostoma or Pteranodon, also known from the American specimens. Professor Williston gives no details of the remarkable tail beyond saying that the tail is small and short and that the vertebrae are flat at the ends without transverse processes. In the restoration the tail is shorter than in the short-tailed species from the lithographic slate and unlike the tail in Ornithocaris. This is the succession of pterodactyls in geological time. Their history is like that of the human race. In the most ancient nations, man's life comes upon us already fully organized. The pterodactyls begin, so far as isolated bones are concerned, in the radic strata, perhaps in the mushelcock or middle division of the trius. And from the beginning of the secondary time they live on with but little diversity in important and characteristic structures. And so far as habit goes, the great pterodactyls of the upper chalk of England cannot be said to be more highly organized than the earlier stiff-tailed genera of the lias or the oolites. There is nothing like evolution. No modification such as that which derives the one-toed horse or the two-toed ox from ancestors with a larger number of digits. On the other hand, there is little, if any, evidence of degeneration. The later pterodactyls do not appear to have lost much, although the tail in some of the Solonhofen genera may be degenerate when compared with the long-tail of dimorphodon, but the short-tailed types are found side-by-side with the long-tailed ramphorincus. The absence of teeth may be regarded as degeneration, for they have presumably become lost in the same way that birds now existing have lost the teeth which characterize the fossil birds, ichthyornus of the American greensand and archaeopteryx of the upper oolites of Bavaria. But just as some of the earlier pterodactyls have no teeth at the extremity of the jaw, such as Dorianathus and ramphorincus, so the loss of teeth may have extended backward till the jaws became toothless. The specimens hitherto known give no evidence of such a change being in progress. But just as the division of animals termed edentata usually wants only the teeth which characterize the front of the jaw, yet others like the great anteater of South America, named Myrmecophagia, have the jaws as free from teeth as the toothless pterodactyls or living birds, and show that in that order the teeth have no value in separating these animals into subordinate groups any more than they have among the monotremata, where one type has teeth and the other is toothless. The following table gives a summary of geological history and succession in the secondary rocks of the principal genera of flying reptiles. Geological Formation Upper Chalk Names of the genera North American Ornithostoma Pteranodon Lower Chalk Upper Greensand Galt Names of the genera British and European Ornithochiris Ornithostoma North American Nectodactyls Lower Greensand Wilden Purbeck Names of the genera British and European Ornithodesmus Dorado Rancus Portland Kimmeridge Clay and Solenhofen Slate Coraline Oolite Oxford Clay Names of the genera British and European Pterodactyls Tenodrakon Cycnoramphus Dioposephyllus Ramphorincus Scaphonathus Great Oolite and Stonesfield Slate Inferior Oolite Names of the genera British and European Ramposephyllus Upper Lyus Lower Lyus Names of the genera British and European Camponathus Dorinathus Dimorphodon Raitic Mushelcock British and European Bones End of Chapter 15 Chapter 16 of Dragons of the Air by Harry Sealy This LibriVox recording is in the public domain. Chapter 16 Classification of the Ornithosaria When an attempt is made to determine the place in nature of an extinct group of animals and the relation to each other of the different types included within its limits so as to express those facts in a classification attention is directed in the first place to characters which are constant and persist through the whole of its constituent genera. We endeavor to find the structural parts of the skeleton which are not affected by variation in the dentition or the proportions of the extremities or length of the tail which may define families or genera or species. It has already been shown that while in many ways the Ornithosarian animals are like birds they have also important resemblances to reptiles. They are often named pterosaria. The wing finger gives a distinctive character which is found in neither one class of existing animals nor the other and is common to all the pterodactyls at present known. They have been named Ornithosaria as a distinct minor division of backboneed animals which may be regarded as neither reptiles nor birds in the sense in which those terms are used to define a lizard or ostrich among animals which still exist. It is not so much that they mark a transition from reptile to bird as that they are a group which is parallel to birds and more manifestly holds an intermediate place than birds do between reptiles and mammals. In plan of structure bird and reptile have more in common than was at one time suspected. The late professor Huxley went so far as to generalize on those coincidences in parts of the skeleton and united birds and reptiles into one group which he named Sauropsida to express the coincidences of structure between the lizard and the bird tribes. The idea is of more value than the term in which it is expressed because reptiles are not as we have seen a group of animals which can be defined by any set of characters as comprehensive as those which express the distinctive features of birds. From the anatomist point of view birds are a smaller group and while some reptiles have affinity with them it is rather the extinct than the living groups which indicate that relation. Other reptiles have affinities of a more marked kind with mammals and there are points in the ornithosaurian skeleton which are distinctly mammalian. So that when the monotree mammals are united with South African reptiles known as Theriodontia which resemble them in a group term Theropsida to express their mammalian resemblances it is evident that there is no one continuous chain of life or gradation in complexity of structure of animals. We have to determine whether the ornithosauria inclined towards the Sauropsida is a bird reptile alliance or to the mammal reptile or Theropsidon alliance. There can be no doubt that the predominant tendency is to the former with a minor affinity towards the latter. The ornithosauria are one of a series of groups of animals living and extinct which have been combined in an alliance named the ornithomorpha. That group includes at least five great divisions of animals which circle about birds known as ornithosauria, Crocodilia, Cerescia, Aves, Ornithescia, and Aristosuchia. Their relations to each other are not evident in an enumeration but may be shown in some degree in a diagram. See page 190. The ornithomorpha. The ornithomorpha arranged in this way shows that the three middle groups Carnivorous Cerescia, Aristosuchia, Herbivorous Ornithescia, which are usually united as dinosauria intervene between birds and ornithosauria and that the Crocodilia and Ornithosauria are parallel groups which are connected with birds by the group of dinosaurs which resembles birds most closely. The ornithomorpha is only one of a series of large natural groups of animals into which living and extinct terrestrial vertebrata may be arranged and their living diagram may contribute to make evident the relations of ornithosauria to the other terrestrial vertebrata. See page 191. Herein it is seen that while the ornithomorpha approached towards mammalia through the ornithosauria and less distinctly through the Crocodilia they approached more directly to the Sauromorpha through the Plesiosauria and Hateria while that group also approaches more directly to the mammals through the Plesiosauria and Anomadons. The Aristosuchia is imperfectly known and therefore to some extent a provisional group. It is a small group of animals. Cordylamorpha are ichthyosaurs and the labyrinthodont group. Herpetomorpha include Lacertilia, Homyosauria, Dolichosauria, Chameleonoidia, Ophidia, Pythonomorpha. The Sauromorpha comprises the groups of extinct and living reptiles named Kelonia, Rincocephala, Sauropterygia, Anomodontia, Nothosauria, and Protosauria. These details may help to explain the place which has been given to the ornithosauria in the classification of animals. Turning to the pterodactyls themselves Von Meyer divided them naturally into short-tailed and long-tailed. The short-tailed indicated by the name pterodactylis he further divided into long-nosed and short-nosed. The short-nosed genus has since been named Teno-drakon, figure 59, page 167. The long-tailed group was divided into two types, the Rampharencus of the Solenhofen slate, figure 56, page 161, and the English form now known as dimorphodon, figure 52, page 150, which had been described from the liais. The Cretaceous pterodactyls form a distinct family. So that believing the tail to have been short in that group, figure 58, there are two long-tailed as well as two short-tailed families which were defined from their typical genera, pterodactylis, ornithochiris, Rampharencus, and dimorphodon. The differences in structure which these animals present are, first, the big-headed forms from the liais like dimorphodon agree with the Rampharencus type from Solenhofen in having a vacuity in the skull defined by bone placed between the orbit of the eye and the nostril. With those characters are correlated the comparatively short bones which correspond to the back of the hand termed metacarpals, and the tail is long and stiffened at its length with ossified tendons. These characters separate ornithosaurs with long tails from those with short tails. The short-tailed types represented by pterodactylis and ornithochiris have no distinct antorbital vacuity in the skull defined by bone. The metacarpal bones of the middle hand are exceptionally elongated and the tail, which was flexible in both, appears to have been short. These differences in the skeleton warrant a primary division of flying reptiles into two principal groups. The short-tailed group, which was recognized by De Blaineville as intermediate between birds and reptiles, may take the Pterodactylia, which he suggested as a convenient distinctive name. It may probably be inconvenient to enlarge its significance to comprise not only the true pterodactyls originally defined as pterosauria, but the newer ornithostoma and ornithochiris, which have been grouped as ornithochiroidia. The second order in which the wing membrane appears to have had a much greater extent in being carried down the hind limbs where the outermost digit and metatarsal are modified for its support has been named pterodermata to include the types which are arranged around ramphorincus and dimorphodon. Both these principal groups have subdivision by many characters in the skeleton, the most remarkable of which is afforded by the pair of bones carried in front of the pubes and termed pre-pubic bones. In the pterodactyl family, the bones in front of the pubes are always separate from each other, always directed forward and have a peculiar fan-shaped form like the bone which holds a similar position in a crocodile. In the ornithochiris family, the pre-pubic bones appear to have been originally triangular but were afterwards united so as to form a strong continuous bar which extends transversely across the abdomen in advance of the pubic bones. This at least is the distinctive character in the genus ornithostoma according to Professor Williston which in many ways closely resembles ornithochiris. The two families in the long-tailed order named pterodermata are separated from each other by a similar difference in their pre-pubic bones. In dimorphodon, those bones are separate from each other and remain distinct through life meeting in the middle line of the body in a wide plate. On the other hand, in rampharencus, the pre-pubic bones which are at first triangular and always slender become blended together into a slight transverse bar which only differs from that attributed to ornithostoma in its more slender two-shaped form. Thus, if other characters of the skeleton are ignored and a classification based upon the structure of the pelvis and pre-pubic bones there would be some ground for associating the long-tailed rampharencus from the upper ole lights which is losing the teeth in the front of its jaw with the cretaceous ornithostoma which has the teeth completely wanting. While the long-tailed dimorphodon would come into closer association with the short-tailed pterodactylus. The drumstick bone or tibia in dimorphodon with its slender fibula like that of a bird also resembles a bird in the rounded and pulley-shaped terminal end corresponding to the middle of the ankle bones in man. The same condition of a terminal pulley joint is found in the cretaceous pterodactyls. But in the true pterodactyls and in rampharencus there usually is no pulley-shaped termination to the lower end of the drumstick for the tarsal bones remain separate from each other like those of ossifications showing the same differences as separate dinosaurs into the divisions which have been referred to from their bird-like pelvis and tibiotarsis as ornitheschia in the one case and sariskia in the other from their bones being more like those of living lizards. End of chapter 16 Chapter 17 of Dragons of the Air by Harry Seeley This LibriVox recording is in the public domain. Chapter 17 Family relations of pterodactyls to animals which lived with them Enough has been said of the general structure of pterodactyls and the chief forms which they assumed while the secondary rocks were accumulating to convey a clear idea of their relations to the types of vertebrate animals which still survive on the earth. We may be unable to explain the reasons for their existence and for their departure from the plan of organization of reptiles and birds. But the evidence has not been exhausted which may elucidate their existence. Sometimes in problems of this kind which involve comparison of the details of the skeleton in different animals it is convenient to imagine the possibility of changes and transitions which are not yet supported by the discovery of fossil remains. If, for example, the pterodactyl is conceived of as divested of the wing finger which is its most distinctive character or that finger is supposed to be replaced by an ordinary digit like the three clawed digits of the hand which we have regarded as applied to the ground where it may be asked would the animal type be found which approximates most closely to a pterodactyl which had been thus modified? There are two possible replies to such a question suggested by the form of the foot. For the old bird Archaeopteryx has three such clawed digits but no wing finger. And some dinosaurs also have the hand with three digits terminating in claws which are quite comparable to the clawed digits of pterodactyls. The truth expressed in the saying that no man by taking thought can add a cubit to his stature is of universal application in the animal world in relation to the result upon the skeleton of the exercise of a function by the individual. Yet such is the relation in proportions of the different parts of the animal to the work which it performs so marked is the evidence that growth has extended in direct relation to use of organs and active life and that structures have become dwarfed from overwork or have wasted away from disuse seen throughout all vertebrate animals that we may fairly attribute to the wing finger some correlated influence upon the proportions of the animal as a consequence of the dependence of the entire economy upon each of its parts. Therefore if an allied animal did not possess a wing finger and did not fly it might not have developed the lightness of bone or the length of limb which pterodactyls possess. The mere expansion of the parachute membrane seen in so-called flying animals both mammals and reptiles which are devoid of wings is absolutely without effect in modifying the skeleton. But when in the bat a wing structure is met with which may be compared to a gigantic extension of the webfoot of the so-called flying frog the bones of the fingers and the back of the hand elongate and extend under the stimulus of the function of flight in the same way as the legs elongate in the more active hoofed animals with the function of running. Therefore it is not improbable that the limbs shared to some extent in growth under stimulus of exercise which developed the wing finger. And if an animal can be found among fossils so far allied as to indicate a possible representative of the race from which these flying dragons arose it might be expected to be at least shorter-legged and possibly more distinctly reptilian in the bones of the shoulder girdle which support the muscles used in flight. It may readily be understood that the kinds of life which were most nearly allied to pterodactyls are likely to have existed upon the earth with them and that flight was only one of the modes of progression which became developed in relation to their conditions of existence. The principal assemblage of terrestrial animals available for such comparison is the dinosauria. They may differ from pterodactyls as widely as the insectivora among mammals differ from bats but not in a more marked way. Comparisons will show that there are resemblances between the two extinct groups which appeal to both reason and imagination. Dinosaurs are conveniently divided by characters of the pelvis first into the order Sariscia which includes the carnivorous megalosaurus and the sediosaurus with the pelvis on the reptile plan and secondly the order Ornithiscia represented by iguanodon with the pelvis on the bird plan. It may be only a coincidence but nevertheless an interesting one that the characters of those two great groups of reptiles which also extend throughout the secondary rocks are to some extent paralleled in parts of the skeleton of the two divisions of pterodactyls. This may be illustrated by reference to the skull, pelvis, hind limb and the pneumatic condition of the bones. The Sariscian dinosauria have an antorbital vacuity in the side of the skull between the nasal opening and the eye as in the long-tailed Ornithosaurus named pterodermata. In some of the older genera of these carnivorous dinosaurs of the trius the lateral vacuities of the head are as large as in dimorphodon. But in some at least of the iguanodont or Ornithiscian dinosaurs there is no antorbital vacuity and the side of the face in that respect resembles the short-tailed pterodactylia. The skull of a carnivorous dinosaur possesses teeth which though easily distinguished from those of pterodactyls can be best compared with them. The most striking difference is in the fact that in the dinosaur the nostrils are nearly terminal while in the pterodactyl they are removed some distance backward. This result is brought about by growth taking place in the one case at the front margin of the maxillary bone to carry the nostril forward and in the other case at the back margin of the premaxillary bone. Thus an elongated part of the jaw is extended in front of the nostril. Hence there is a different proportion between the premaxillary and maxillary bones in the two groups of animals which corresponds to the presence of a beak and its absence in living reptiles. It is not known whether the extremity of the pterodactyl's beak is a single bone, the intermaxillary bone, such as forms the corresponding toothless part of the jaw in the South African reptile disynodon or whether it is made by the pair of bones called premaxillaries which form the extremity of the jaw in most dinosaurs. Too much importance may perhaps be attached to such differences which are partly hypothetical because the extinct ichthyosaurus which has an exceptionally long snout has the two premaxillary bones elongated so as to extend backward to the nostrils. A similar elongation of those bones is seen in porpoises which also have a long snout and the bones are carried back from the front of the head to the nostrils which are sometimes known as blowholes. But the porpoise has those premaxillary bones not so much in advance of the bones which carry teeth named maxillary as placed in the interspace between them. The nostrils, however, are not limited to the extremity of the head in all dinosaurs. If this region of the beak in dimorphodon be compared with the corresponding part of a dinosaur from the Permian rocks or trius, the relation of the nostril to the bones forming the beak may be better understood. In the sandstone of Elgin, usually named Trius, a small dinosaur is found which has been named Ornithosuchus from the resemblance of its head to that of a bird. Seen from above, the head has a remarkable resemblance to the condition in Ramphorincus in the sharp pointed beak and positions of the orbits and other openings. In side view, the orbits have the triangular form seen in dimorphodon and the preorbital vacuities are large as in that genus while the lateral nostrils which are smaller are further forward in the dinosaur. The differences from dimorphodon are in the articulation for the jaw being carried a little backward instead of being vertical as in the pterodactyl and the bone in front of the nose is smaller. Notwithstanding probable differences in the palate, the approximation which extends to the crocodile-like vacuity in the lower jaw is such that by slight modification in the skull the differences would be substantially obliterated by which the skull of such an ornithosaur is technically distinguished from such a dinosaur. The back of the skull is clearly seen in the Whitby pterodactyl and its structure is similar to the corresponding part of such dinosaurs as Ancusaurus or Atlantosaurus without the resemblance quite amounting to identity but still far closer than is the resemblance between the same region in the heads of crocodiles, lizards, serpents, colonians. Few of these fossil dinosaur skulls are available for comparison and those differ among themselves. The coincidences rather suggest a close collateral relation then prove the elaboration of one type from the other. They may have had a common ancestor. The Trius rocks near Stuttgart have yielded dinosaurs as unlike pterodactyls as could be imagined resembling heavily armored crocodiles in such types as the genus Belladon. Its jaws are compressed from side to side as in many pterodactyls and the nostrils are at least as far backward as in Ramphorincus. Belladon has pre-orbital vacuities and post-orbital vacuities but the orbit of the eye is never large as in pterodactyls. It might not be worthwhile dwelling on such points in the skull if it were not that the pelvis in Belladon is the basin formed by the blending of the expanded plates of the ischium and the pubis into a sheet of bone which more nearly resembles the same region in pterodactyls than does the ischio-pubic region in other dinosaurian animals like sediosaurus. The backbone in a few dinosaurs is suggestive of pterodactyls. In such genera as have been named Celerus and Callumospondylus in which the skeleton is only partially known the neck vertebrae become elongated so as to compare with the long necked pterodactyls. The cervical rib is often very similar to that type and blended with the vertebra as in pterodactyls in birds. The early dorsal vertebrae of pterodactyls might also be mistaken for those of dinosaurs. The tail vertebrae of a pterodactyl are usually longer than in long tail dinosauria. In the limbs and the bony girdles which support them there is more resemblance between pterodactyls and dinosaurs than might have been anticipated considering their manifest differences in habit. Thus all dinosaurs have the hip bone named Ilium prolonged in front of the articulation for the femur as well as behind it almost exactly as in pterodactyls and birds. See page 95. There is some difference in the pubis and ischium which is more conspicuous in form than in direction of the bones. There is a pterodactyl imperfectly preserved named pterodactylus dubius in which the ischium is directed backward and the pubis downward and the bones unite below the acetabular cavity for the head of the femur to work in but do not appear to be otherwise connected. In Rampharencus the connection between these two thickened bars is made by a thin plate of bone. In such a dinosaur as the American carnivorous seratosaurus the two bars of the pubis and ischium remain separate and diverging and there is no film of bone extending over the interspace between them. The development of such a bony condition would make a close approximation between the ornithosaurian pelvis and that of those dinosaurs which closely resemble pterodactyls in skull and teeth. Another pelvic character of some interest is the blending of the pubis and ischium of the right and left sides in the middle line of the body. There are some genera of dinosaurs like the English aristosuchus from the Weald and the American genera Celerus seratosaurus and others in which the pubic bones instead of uniting at their extremities are pinched together from side to side and unite down the lower part of their length terminating in an expanded end like a shoe which is seen to be a separate ossification and probably formed by a pair of ossifications joined in the median line. This small bone which is below the pubes and in these animals becomes blended with them we may regard as a pair of pre-pubic bones like those of pterodactyls and crocodiles. Except that they have lost the stock like portions which in those animals are developed to compensate for the diminished length of the pubic bones. The pre-pubic bones may also be developed in a guanodon in which a pair of bones of similar form remains throughout life in advance of the pubes as in pterodactyls. In those dinosauria with the bird-like type of pelvis the pubic bone is exceptionally developed sending one process backward and another process forward so that there is a great gap between these diverging limbs to the bone. In the region behind the sternum to which the ribs were attached and in front of the pelvis is a pair of bones in a guanodon shaped like the pre-pubic bones of dimorphodon. They have sometimes been interpreted as a hinder part of the sternum but may more probably be regarded as a pair of pre-pubic bones articulating each with the anterior process of the pubis. See figure 80. The small bones found at the extremities of the pubes and such carnivorous dinosaurs as Aristosuchus are blended by bony union with the pubes. The bones in a guanodon are placed behind the sternal region without any attachment for sternal ribs and the expanded processes converge forwards from the stock and unite exactly like the pre-pubic bones of ornithosaurs. While this character on the one hand may link pterodactyls with the dinosaurs, on the other hand it may be a link between both those groups and the crocodiles in which the front pair of bones of the pelvis has also appeared to be representative of the pre-pubic bones of flying reptiles. See figure 32, page 98. The resemblances between pterodactyls and dinosaurs in the hind limb are not of less interest though it is rather in the older pterodactyls such as dimorphodon, pterodactyls, and ramphorincus that the resemblances closest with the slender carnivorous dinosaurs. They never have the head of the thigh bone, the femur separated from its shaft by a constricted neck as in the pterodactyls from the chalk. In many ways the thigh bone of dinosaurs tends towards being avian while that of pterodactyls inclines towards being mammalian but with a tendency to be bird-like in the older types and to be mammal-like in the most recent representatives of the group in the chalk. The bones of the leg in ornithosaurs known as tibia in fibula are remarkable for the circumstance first that they resemble birds in the fibula being slender and only developed in its upper part towards the femur and secondly that in a genus like dimorphodon this drumstick bone has the two upper bones of the ankle blended with the tibia so as to form a rounded pulley joint which is indistinguishable from that of a bird. See page 102 There is a large number of dinosaurs in which this remarkable distinctive character of birds is also found. Only dinosaurs like iguanodon for instance have the slender fibula as long as the tibia and contributing to unite with the separate ankle bones of the similarly rounded pulley at the lower end. There are no birds in which the tarsal bones remain separated and distinct throughout life. But in pterodactyls from Solenhofen as in a number of dinosaurs especially the carnivorous genera the bones of the tarsus remain distinct throughout life and never acquired such forms as would have enabled the ankle bone termed astragalus to embrace the extremity of the tibia as it does in iguanodon. Thus the resemblance of the ornithosar drumstick is almost as close to dinosaurs as to birds. There is great similarity between dinosaurs and pterodactyls seen in the region of the end step known as the metatarsus. These bones are usually four in number parallel to each other and similar in form. They are commonly longer than in dinosaurs but among some of the carnivorous dinosaurs their length approximates to that seen in pterodactyls. In neither group are the bones blended together by bony union while they are always united in birds as in oxen and similar even hoofed mammals. Dinosaurs agree with pterodactyls in maintaining the metatarsal bones separate but they differ from them and agree with birds frequently in having the number of metatarsal bones reduced to three as in iguanodon though dinosaurs often have as many as five digits developed. The toe bones, the phalanges of these digits of the hind limb are usually longer in pterodactyls than in dinosaurs but they resemble carnivorous dinosaurs in the forms of the terminal bones for the claws which are similarly compressed from side to side. So diverse are the functions of the phorilim in dinosaurs and pterodactyls and so remarkably does the length of the metacarpal region of the back of the hand vary in the long-tailed and short-tailed ornithosaurs that there is necessarily a less close correspondence between that region of the skeleton between these two groups of animals for the pterodactyl phorilim is modified in relation to a function which can only be paralleled among birds and bats and yet neither of these groups of animals approximates closely in this region of the skeleton to the flying reptile. Under all the modifications of structure to differences of function some resemblance to dinosaurs may be detected which is best evident in the upper arm bone humorous is slight in the forearm bones ulna and radius and becomes lost towards the extremity of the limb. If the tendency of the thigh bone to resemble a mammalian type of femur 100 is a fundamental deep-seated character of the skeleton it might be anticipated that a trace of mammalian character would also be found in the humerus. For what the character is worth the head of the humerus does show a closer approximation to a monotree mammal than is seen in birds to some extent paralleled in those South African reptiles which approximate to mammals most closely. Not the least remarkable of the many astonishing resemblances of these light aerial creatures to the more heavy-bodied dinosaurs is the circumstance that the humerus in both groups makes a not dissimilar approach to that of certain mammals. These illustrations may be accepted as demonstrating a relationship between the ornithosaurs and dinosaurs now compared which can only be explained as results of influence of a common parentage upon the forms of the bones. But more interesting than resemblances of that kind is the similarity that may be traced in the way in which air is introduced into cavities in the bones in both groups. In some of the imperfectly known dinosaurs like Aristosuchus, Celerus, and Thecospondylus the bone texture is as thin as in pterodactyls and the vertebrae are excavated by pneumatic cavities which are amazing in size when compared with the corresponding structures in birds. For the vertebrae is often hollowed out so that nothing remains but a thin external film like paper for its thickness. In the dinosaurian genus Celerus this condition is as well marked in the tail and back as it is in the neck. The differential difference from birds appears to be that in the larger carnivorous dinosaurs the pneumatic condition of the bones is confined to the vertebral column while birds and pterodactyls have the pneumatic condition more conspicuously developed in the limb bones. The pneumatic skeleton however appears to be absent from the herbivorous types like iguanodon and the pterodactyl dinosaurs which have the birdlike form of pelvis and are most birdlike in the forms of bones of the hind limb. It is possible that some of the carnivorous dinosaurs also possessed limb bones with pneumatic cavities. Many of those bones are hollow with very thin walls. If their cavities were connected with the lungs the foramina are inconspicuous and unlike the immense holes seen in the sides of the vertebrae. According to the late Professor Marsh the limbs of Celerus and its allies which at present are imperfectly known are in some cases pneumatic. Therefore there is a closer fundamental resemblance between some carnivorous dinosaurs and pterodactyls than might have been anticipated. But the skull of Celerus is unknown and the fragments of the skeleton hitherto published are insufficient to do more than show that the two types were near in kindred though distinct in habit. Each has elaborated a skeleton which owes much to the common stock which transmitted the vital organs and the tendency of the bones to take special forms. But which also owes more than can be accurately measured to the action of muscles in shaping the bones and the influence of the mechanical conditions of daily life upon the growth of the bones in both of these orders of animals. Enough is known to prove that all dinosaurs cannot be regarded as ornithosaurs which have not acquired the power of flight. Though the evidence would lead us to believe that the primitive ornithosaur was a four-footed animal before the wing finger became developed in the fore limb as a means of extending a patageal membrane like the membrane which in the hind limb of dimorphodon has bent the outermost digit of the foot upward and outward to support the corresponding organ of flight extending down the hind legs. It may thus be seen that the characters of ornithosaurs which have already been spoken of as reptilian as distinguished from the resemblances to birds may now with more accuracy be regarded as dinosaurian. The dinosaurs, like pterodactyls must be regarded as intermediate in some respects between reptiles and birds. The resemblances enumerated would alone constitute a partial transition from the reptile to the bird although no dinosaurs have organs of flight. Many are heavily armored with plates of bone and few, if any, approximate in the technical parts of the skeleton to the bird class except in the hind limbs. Yet dinosaurs have sometimes been regarded as standing to birds in the relation of ancestors or as parallel to an ancestral stock. Before an attempt can be made to estimate the mutual relation of the flying reptiles to dinosaurs on the one hand and to birds on the other, it may be well to remember that the resemblance of such a dinosaur as a guanodon to a bird in its pelvis and hind limb is not more remarkable than that of pterodactyls to birds in the shoulder girdle and bones of the forelimb. The keeled sternum, the long, slender coracoid bones in scapulae are absolutely bird-like in most ornithosaurs and that region of the skeleton only differs from birds in the absence of a ferculum which represents the clavicles and is commonly named the marythot. The elongated bones of the forearm and the hand, terminating in three sharp claws are characters in which the fossil bird archaeopteryx resembles the pterodactyl rampharencus, a resemblance which extends to a similar elongation of the tail. It is remarkable that the resemblance should be so close since archaeopteryx affords the only bird skeleton known to be contemporary which can be compared with the sullenhofen flying reptiles. The resemblance may possibly be closer than has been imagined. The back of the head of archaeopteryx is imperfectly preserved in the region of the quadrate bone, malar arch and temporal vacuity. Until these are better known, it cannot be affirmed that the back of the head is more reptilian in pterodactyls than in the oldest birds. The side of the head in archaeopteryx is distinguished by the nostril being far forward, the vacuity in front of the orbit being as large as in the pterodactyl scaphonathus from sullenhofen and other long-tailed pterodactyls. End of Chapter 17