CHAPTER XI. Shoulder-Girdle and Fore Limb

SHOULDER-GIRDLE AND FORE LIMB

STERNUM

The sternum is always a distinguishing part of the bony structure of the breast. In Crocodiles it is a cartilage to which the sternal ribs unite; and upon its front portion a flat knife-like bone called the interclavicle is placed. In lizards like the Chameleon, it is a lozenge-shaped structure of thin bony texture, also bearing a long interclavicle, which supports the clavicular bones, named collar bones in man, which extend outward to the shoulder blades. Among mammals the sternum is usually narrow and flat, and often consists of many successive pieces in the middle line, on the under side of the body. Among Bats the anterior part is somewhat widened from side to side, to give attachment to the collar bones, but the sternum still remains a narrow bone, much narrower than in Dolphins, and not differing in character from many other Mammals, notwithstanding the Bat's power of flight. The bone develops a median keel for the attachment of the muscles of the breast, but something similar is seen in burrowing Insectivorous mammals like the Moles. So that, as Von Meyer remarked, the presence of a keel on the sternum is not in itself sufficient evidence to prove flight.

Among birds the sternum is greatly developed. Broad and short in the Ostrich tribe, it is devoid of a keel; and therefore the keel, if present in a bird, is suggestive of flight. The keel is differently developed according to the mode of attachment of the several pectoral muscles which cover a bird's breast. In several water birds the keel is strongly developed in front, and dies away towards the hinder part of the sternum, as in the Cormorant and its allies. The sternum in German Pterodactyles is most nearly comparable to these birds.

In the Solenhofen Slate the sternum is fairly well preserved in many Ornithosaurs. It is relatively shorter than in birds, and is broader than long; but not very like the sternum of reptile or mammal in form. The keel is limited to the anterior part of the shield of the sternum, as in Merganser and the Cormorant, and is prolonged forward for some distance in advance of it. Von Meyer noticed the resemblance of this anterior process to the interclavicle of the Crocodile in position; but it is more like the keel of a bird's sternum, and is not a separate bone as in Reptiles. In Pterodactyles from the Cretaceous rocks, the side bones, called coracoids, are articulated to saddle-shaped surfaces at the hinder part of the base of this keel, which are parallel in Ornithocheirus, as in most birds, but overlap in Ornithodesmus, as in Herons and wading birds.

Showing the strong keel and the facets for the coracoid bones on its hinder border above the lateral constrictions]

The keel was pneumatic, and when broken is seen to be hollow, and appears to have been exceptionally high in Rhamphorhynchus, a genus in which the wing bones are greatly elongated. Von Meyer found in Rhamphorhynchus on each side of the sternum a separate lateral plate with six pairs of sternal ribs, which unite the sternum with the dorsal ribs, as in the young of some birds. The hinder surface of the sternum is imperfectly preserved in the toothless Pterodactyles of Kansas. Professor Williston states that the bone is extremely thin and pentagonal in outline, projecting in front of the coracoids, in a stout, blunt, keel-like process, similar to that seen in the Pterodactyles of the Cambridge Greensand. American specimens have not the same notch behind the articulation for the coracoid to separate it from the transverse lateral expansion of the sternal shield. The lateral margin in the Cambridge Greensand specimens figured by Professor Owen and myself is broken; but Professor Williston had the good fortune to find on the margin of the sternum the articular surfaces which gave attachment to the sternal ribs. The margin of the sternal bone thickens at these facets, four of which are preserved. The sternum in Ornithostoma was about four and a half inches long by less than five and a half inches wide. The median keel extends forward for rather less than two inches, while in the smaller Cambridge species of Ornithocheirus it extends forward for less than an inch and a half.

A sternum of this kind is unlike that of any other animal, but has most in common with a bird; and may be regarded as indicating considerable power of flight. The bone cannot be entirely attributed to the effect of flight, since there is no such expanded sternal shield in Bats. The small number of sternal ribs is even more characteristic of birds than mammals or reptiles.

THE SHOULDER-GIRDLE

The bones which support the fore limb are one of the distinctive regions of the skeleton defining the animal's place in nature. Among most of the lower vertebrata, such as Amphibians and Reptiles, the girdle is a double arch--the arch of the collar bone or clavicles in front, and the arch of the shoulder-blade or scapula behind. The clavicular arch, when it exists, is formed of three or five parts--a medium bar named the interclavicle, external to which is a pair of bones called clavicles, reaching to the front of the scapulae when they are present; and occasionally there is a second pair of bones called supraclavicles, extending from the clavicles up the front margins of the scapulae. Thus the clavicular arch is placed in front of the scapular arch. The supraclavicles are absent from all living Reptiles, and the clavicles are absent from Crocodiles. The interclavicle is absent from all mammals except Echidna and Ornithorhynchus. Clavicles also may be absent in some orders of mammals. Hence the clavicular arch may be lost, though the collar bones are retained in man.

The scapular arch also is more complicated and more important in the lower than in the higher vertebrata. It may include three bones on each side named coracoid, precoracoid, and scapula. But in most vertebrates the coracoid and precoracoid appear never to have been segmented so as to be separated from each other; and it is only among extinct types of reptiles, which appear to approximate to the Monotreme mammals, that separate precoracoid bones are found; though among most mammals, probably, there are stages of early development in which precoracoids are represented by small cartilages, though few mammals except Edentata like the Sloths and Ant-eaters, retain even the coracoids as distinct bones. Therefore, excepting the Edentata and the Monotremes, the distinctive feature of the mammalian shoulder-girdle appears to be that the limbs are supported by the shoulder-blades, termed the scapulae.

Among reptiles there are several distinct types of shoulder-girdle. Chelonians possess a pair of bones termed coracoids which have no connexion with a sternum; and their scapulae are formed of two widely divergent bars, divided by a deeper notch than is found in any fossil reptiles. Among Lizards both scapula and coracoid are widely expanded, and the coracoid is always attached to the sternum. Chameleons have the blade of the scapula long and slender, but the coracoid is always as broad as it is long. Crocodiles have the bone more elongated, so that it has somewhat the aspect of a very strong first sternal rib when seen on the ventral face of the animal. The bone is perforated by a foramen, which would probably lie in the line of separation from the precoracoid if any such separation had ever taken place. The scapula, or shoulder-blade, of Crocodiles is a similar flat bone, very much shorter than the scapula of a Chameleon, and more like that of the New Zealand Hatteria. Thus there is very little in common between the several reptilian types of shoulder-girdle.

In birds the apparatus for the support of the wings has a far-off resemblance to the Crocodilian type. The coracoid bones, instead of being directed laterally outward and upward from the sternum, as among Crocodiles, are directed forward, so as to prolong the line of the breast bone, named the sternum. The bird's coracoid is sometimes flattened towards the breast bone among Swans and other birds; yet as a rule the coracoid is a slender bar, which combines with the still more slender and delicate blade of the scapula, which rests on the ribs, to make the articulation for the upper arm bone. Among reptiles the scapula and coracoid are more or less in the same straight line, as in the Ostrich, but in birds of flight they meet at an angle which is less than a right angle, and where they come in contact the external surface is thickened and excavated to make the articulation for the head of the humerus. There is nothing like this shoulder-girdle outside the class of birds, until it is compared with the corresponding structure in these extinct animals called Pterodactyles. The resemblance between the two is surprising. It is not merely the identity of form in the coracoid bone and the scapula, but the similar angle at which they meet and the similar position of the articulation for the humerus. Everything in the Pterodactyle's shoulder-girdle is bird-like, except the absence of the representative of the clavicles, that forked #V#-shaped bone of the bird which in scientific language is known as the furculum, and is popularly termed the "merry-thought." This kind of shoulder-girdle is found in the genera from the Lias and the Oolitic rocks, both of this country and Germany.

In the Cretaceous rocks the scapula presents, in most cases, a different appearance. The coracoid is an elongated, somewhat triangular bone, compressed on the outer margin as in birds, but differing alike from birds and other Pterodactyles in not being prolonged forward beyond the articulation for the humerus. In these Cretaceous genera, toothed and toothless alike, the articulation for the upper arm bone truncates the extremity of the coracoid, so that the bone is less like that of a bird in this feature. Perhaps it shows a modification towards the crocodilian direction. The scapula, which unites with the coracoid at about a right angle, is similarly truncated by the articular surface for the humerus; but the bone is somewhat expanded immediately beyond the articulation, and compressed; and instead of being directed backward, it is directed inward over the ribs to articulate with the neural arches of the early dorsal vertebrae in the genera found in strata associated with the Chalk. As the bone approaches this articulation, it thickens and widens a little, becoming suddenly truncated by an ovate facet, which exactly corresponds to the transversely ovate impression, concave from front to back, which is seen in the neural arches of the dorsal vertebrae on which it fits. This condition is not present in all Cretaceous Pterodactyles. It does not occur in the Kansas fossil, named by Professor Marsh, Nyctodactylus. And it appears to be absent from the Pterodactyles of the English Weald, named Ornithodesmus.

An ossification which gives attachment to the scapulae seen in the early dorsal vertebra of Ornithocheirus

(From the Cambridge Greensand)]

Showing how the scapulae articulate with a vertebra and the articulation of the coracoids with the sternum. The humeral articulation with the coracoid is unlike the condition shown in other Ornithosaurs]

There is no approach to this transverse position of the scapulae among birds. And while the form of the bones in the older genera of Ornithosaurs is singularly bird-like, the angular arrangement in this Cretaceous genus is obtained by closely approximating the articulations on the sternum, so that the coracoids extend outward as in reptiles, instead of forward as in birds; and the extremities of the scapulae similarly approximate towards each other. This rather recalls the relative positions of scapula and coracoid among crocodiles. If crocodile and bird had been primitive types of animals instead of surviving types, it might almost seem as though there had been a cunning and harmonious blending of one with the other in evolving this form of shoulder-girdle.

THE FORE LIMB

The bones of the fore limb, generally, correspond in length with the similar parts of the hind limb. The upper arm bone corresponds with the upper leg bone, and the fore-arm bone is as long as the fore-leg bone; then differences begin. The bones which correspond to the back of the hand in man, termed the metacarpus, are variable in length in Pterodactyles--sometimes very long and sometimes short. The wing metacarpal bone is always stout, and the others are slender. The extremity of the metacarpus was applied to the ground. Three small digits of the hand are developed from the three small metacarpal bones, and terminate in large claws.

The great wing finger was bent backward, and only touched the ground where it fitted upon the wing metacarpal bone. It appears sometimes to have been as long as the entire vertebral column.

Owing to the circumstance that the joint in the arm in Pterodactyles was not at the wrist as among birds, but between the metacarpus and the phalanges, it follows that the fore limb was longer than the hind limb when the metacarpus was long; but the difference would not interfere with the movements of the animal, either upon four feet or on two feet, for in bats and birds the disproportion in length is greater.

HUMERUS OR UPPER ARM BONE

The first bone in the fore-arm, the humerus, is remarkable chiefly for the compressed crescent form of its upper articular end, which is never rounded like the head of the upper arm bone in man, and secondly for the great development of the external process of bone near that end, termed the radial crest. Sir Richard Owen compared the bone to the humerus of both birds and crocodiles, but in its upper articular end the crocodile bone may be said to be more like a bird than it is like the Pterodactyle. In flying reptiles the articular surface next to the shoulder-girdle is somewhat saddle-shaped, being concave from side to side above and convex vertically, while most animals with which it can be compared have the articular head of the bone convex in both directions. A remarkable exception to this general rule is found in some fossil animals from South Africa, which, from resemblance to mammals in their teeth, have been termed Theriodonts. They sometimes have the head of the bone concave from side to side and convex in the vertical direction. To this condition Ornithorhynchus makes a slight approximation. The singular expansion of the structure called the radial crest finds no close parallel in reptiles, though Crocodiles have a moderate crest on the humerus in the same position; and in Theriodonts the radial crest extends much further down the shaft of the humerus. No bird has a radial crest of a similar kind, though it is prolonged some way down the shaft in Archaeopteryx. In Pterodactyles it sometimes terminates outward in a smooth, rounded surface, which might have been articular if any structure could have articulated with it. There is also a moderate expansion of the bone on the ulnar side in some Pterodactyles, so that the proximal end often incloses nearly three-fourths of an ovate outline. The termination of the radial crest is at the opposite end of this oval to the wider articular part of the head of the bone, in some specimens from the Cambridge Greensand. The radial crest is more extended in Rhamphorhynchus. All specimens of the humerus show a twist in the length of the bone, so that the end towards the fore-arm, which is wider than the shaft, makes a right angle with the radial crest on the proximal end, which is not seen in birds. The shaft of the humerus is always stouter than that of the femur, though different genera differ in this respect.

The humerus in genera from rocks associated with the Chalk presents two modifications, chiefly seen in the characters of the distal end of the bone. One of these is a stout bone with a curiously truncated end where it joins the two bones of the fore-arm; and the other is more or less remarkable for the rounded form of the distal condyles. Both types show distinct articular surfaces. The inner one is somewhat oblique and concave, the outer one rounded; the two being separated by a concave channel, so that the ulna makes an oblique articulation with the bone as in birds, and the radius articulates by a more or less truncated or concave surface.

ULNA AND RADIUS

The bones of the fore-arm are similar to each other in size, and if there be any difference between them the ulna is slightly the larger. There is some evidence that in Rhamphorhynchus the upper end of the ulna was placed behind the radius, probably in consequence of the mode of attachment of those bones to the humerus. The ulna abutted towards the inner and lower border, while the radius was towards the upper border, consequent upon the twist in the humerus. This condition corresponds substantially with the arrangement in birds, but differs from birds in the relatively more important part taken by the radius in making the articulation. The bones are compared in Dimorphodon with the Golden Eagle drawn of the same size (Fig. 42). In birds the ulna supports the great feathers of the wing, and this may account for the size of the bone. The ulna is best seen at its proximal end in the specimens from the Cambridge Greensand, where there is a terminal olecranon ossification forming an oblique articulation, which frequently comes away and is lost. It is sometimes well preserved, and indicated by a suture. The examples of ulna from the Lias show a slight expansion of the bone at both ends, and at the distal end toward the wrist the articulation is well defined, where the bone joins the carpus. The larger specimens of the bone are broken. The distal articular surface is only connected with the proximal end of the bone in small specimens: it always shows on the one margin a concavity, followed by a prominent boss, and an oblique articulation beyond the boss. On the side towards the radius, on the lower end of the shaft there is an angular ridge, which marks the line along which the ulna overlaps the radius. The lower end of the radius has a simple, slightly convex articulation, somewhat bean-shaped. No rotation of these bones on each other was possible as in man. There is a third bone in the fore-arm. This bone, named the pteroid, is commonly seen in skeletons from Solenhofen. It was regarded by Von Meyer as having supported the wing membrane in flight. Some writers have interpreted it as an essential part of the Pterodactyle skeleton, and Von Meyer thought that it might possibly indicate a fifth digit in the hand. The only existing structure at all like it is seen in the South African insectivorous mammal named Chrysochloris capensis, the golden mole, which also has three bones in the fore-arm, the third bone extending half-way up towards the humerus. In that animal the third bone appears to be behind the others and adjacent to the ulna. In the German fossils the pteroid articulated with a separate carpal or metacarpal bone, placed on the side of the arm adjacent to the radius, and the radius is always more inward than the ulna. If the view suggested by Von Meyer is adopted, this bone would be a first digit extending outward and backward towards the humerus. That view was adopted by Professor Marsh. It involves the interpretation of what has been termed the lateral carpal as the first metacarpal bone, which would be as short as that of a bird, but turned in the opposite direction backward. The first digit would then only carry one phalange, and would not terminate in a claw, but lie in the line of the tendon which supports the anterior wing membrane of a bird.

The third bone in the fore-arm of Chrysochloris does not appear to correspond to a digit. The bone is on the opposite side of the arm to the similar bone of a Pterodactyle, and therefore cannot be the same structure in the Golden Mole. The interpretation which makes the pteroid bone the first digit has the merit of accounting for the fifth digit of the hand. All the structures of the hand are consistent with this view. The circumstance that the bone is rarely found in contact with the radius, but diverging from it, shows that it plays the same part in stretching the membrane in advance of the arm, that the fifth digit holds in supporting the larger wing membrane behind the arm.

According to Professor Williston, the American toothless Pterodactyle Ornithostoma has but a single phalange on the corresponding first toe of the hind foot, and that bone he describes as long, cylindrical, gently curved, and bluntly pointed. There is some support for this interpretation; but I have not seen any English or German Pterodactyles with only one phalange in the first toe.

The wing in Pterodactyles would thus be stretched between two fingers which are bent backward, the three intermediate digits terminating in claws.

THE CARPUS

The wrist bones in the reptilia usually consist of two rows. In Crocodiles, in the upper row there is a large inner and a small outer bone, behind which is a lunate bone, the remainder of the carpus being cartilaginous. Only one carpal is converted into bone in the lower row. It is placed immediately under the smaller upper carpal. In Chelonians, the turtle and tortoise group, the characters of the carpus vary with the family. In the upper row there are usually two short carpals, which may be blended, under the ulna; while the two under the radius are commonly united. The lower row is made up of several small bones. Lizards, too, usually have three bones in the proximal row and five smaller bones in the distal row.

The correspondence of the distal carpals with the several metacarpal bones of the middle hand is a well-known feature of the structure of the wrist.

Von Meyer remarks that the carpus is made up of two rows of small bones in the Solenhofen Pterodactyles; while in birds there is one row consisting of two bones. The structure of the carpus is not distinct in all German specimens; but in the short-tailed Solenhofen genera the bones in the two rows retain their individuality.

In all the Cretaceous genera the carpal bones of each row are blended into a single bone, so that two bones are superimposed, which may be termed the proximal and distal carpals. One specimen shows by an indication of sutures the original division of the distal carpal into three bones; and the separated constituent bones are very rarely met with. Two bones of the three confluent elements contribute to the support of the wing metacarpal, and the third gives an articular attachment to the bone which extends laterally at the inner side of the carpus, which I now think may be the first metacarpal bone turned backward towards the humerus. The three component bones meet in the circular pneumatic foramen in the middle of the under side of the distal carpal. There is no indication of division of the proximal carpal in these genera into constituent bones.

This condition is somewhat different from birds. In 1873 Dr. Rosenberg, of Dorpat, showed that there is in the bird a proximal carpal formed of two elements, and a distal carpal also formed of two elements. Therefore the two constituents of the distal carpal in the bird which blends in the mature animal with the metacarpus, forming the rounded pulley joint, may correspond with two of the three bones in the Cretaceous Pterodactyle Ornithocheirus.

The width of a proximal carpal rarely exceeds two inches, and that of a distal carpal is about an inch and three-quarters. Two such bones when in contact would not measure more than one inch in depth. The lower surface shows that the wing had some rotary movement upon the carpus outward and backward.

METACARPUS

The metacarpus consists of bones which correspond to the back of the hand. The first digit of the hand in clawed animals has the metacarpal bone short, or shorter than the others. Among mammals metacarpal bones are sometimes greatly elongated; and a similar condition is found in Pterodactyles, in which the metacarpal bone may be much longer than the phalange which is attached to it. Two metacarpal bones appear to be singularly stouter than the others. The first bone of the first digit, if rightly determined, is much shorter than the others, and is, in fact, no longer than the carpus (Fig. 43). It is a flat oblong bone, attached to the inner side of the lower carpal, and instead of being prolonged distally in the same direction as the other metacarpal bones, is turned round and directed upward, so that its upper edge is flush with the base of the radius, and gives attachment to a bone which resembles a terminal phalange of the wing finger. According to this interpretation it is the first and only phalange in the first digit. The bone is often about half as long as the fore-arm, terminates upward in a point, is sometimes curved, and frequently diverges outward from the bones of the fore-arm, as preserved in the associated skeleton, being stretched towards the radial crest of the humerus. This mode of attachment of the supposed first metacarpal, which is true for all Cretaceous pterodactyles, has not been shown to be the same for all those from the Solenhofen Slate. There is no greater anomaly in this metacarpal and phalange on the inner side being bent backward, than there is in the wing finger being bent backward on the outer side. The three slender intervening digits extend forward between them, as though they were applied to the ground for walking.

The bone which is usually known as the wing metacarpal is frequently stouter at the proximal end towards the carpus than towards the phalange. At the carpal end it is oblong and truncated, with a short middle process, which may have extended into the pit in the base of the carpal bone; while the distal terminal end is rounded exactly like a pulley. There is great difference in the length of the metacarpus. In the American genus Ornithostoma it is much longer than the fore-arm. In Rhamphorhynchus it is remarkably short, though perhaps scarcely so short as in Dimorphodon or in Scaphognathus. The largest Cretaceous examples are about two inches wide where they join the carpus. The bone is sometimes a little curved.

Between the first and fifth or wing metacarpal are the three slender metacarpal bones which give attachment to the clawed digits. They bear much the same relation to the wing metacarpal that the large metatarsal of a Kangaroo has to the slender bones of the instep which are parallel to it.

The facet for the wing metacarpal on the carpus is clearly recognised, but as a rule there is no surface with which the small metacarpals can be separately articulated. One or two exceptional specimens from the Cambridge Greensand appear to have not only surfaces for the wing metacarpal, but two much smaller articular surfaces, giving attachment to smaller metacarpals; while in one case there appears to be only one of these additional impressions. It is certain that all the animals from the Lias and Oolites have three clawed digits, but at present I have seen no evidence that there were three in the Cretaceous genera, though Professor Williston's statements and restoration appear to show that the toothless Pterodactyles have three. Another difference from the Oolitic types, according to Professor Williston, is in the length of the slender metacarpals of the clawed phalanges being about one-third that of the wing metacarpal, but this is probably due to imperfect ossification at the proximal end; for at the distal end the bones all terminated on the same level, showing that the four outer digits were applied to the ground to support the weight of the body. The corresponding bone in the Horse and Oxen is carried erect, so as to be in a vertical line with the bones of the fore-arm; and the same position prevails usually, though not invariably, with the corresponding bone in the hind limb, while in many clawed mammals the metacarpus and metatarsus are both applied upon the ground. In Pterodactyles the metatarsal bones are preserved in the rock in the same straight line with the smaller bones of the foot, or make an angle with the shin bone, leading to the conviction that the bones of the foot were applied to the ground as in Man, and sometimes as in the Dog, and were thus modified for leaping. Just as the human metacarpus is extended in the same line with the bones of the fore-arm, and the movement of jointing occurs where the fingers join the metacarpus, so Pterodactyles also had these bones differently modified in the fore and hind limbs for the functions of life. The result is to lengthen the fore limb as compared with the hind limb by introducing into it an elevation above the ground which corresponds to the length of the metacarpus, always supposing that the animal commonly assumed the position of a quadruped when upon the earth's surface.

This position of the metacarpus is a remarkable difference from Birds, because when the bird's wing is at rest it is folded into three portions. The upper arm bone extends backward, the bones of the fore-arm are bent upon it so as to extend forward, and then at the wrist the third portion, which includes the metacarpus and finger bones, is bent backward. So that the metacarpus in the Pterodactyle differs from birds in being in the same line as the bones of the fore-arm, whereas in birds it is in the same line with the digit bones of the hand. It is worthy of remark that in Bats, which are so suggestive of Pterodactyles in some features of the hand, the metacarpals and phalanges are in the same straight line; so that in this respect the bat is more like the bird. But Pterodactyles in the relation of these bones to flight are quite unlike any other animal, and have nothing in common with the existing animals named Reptiles.

THE HAND

From what has just been said it follows that the construction of the hand is unique. It may be contrasted with the foot of a bird. The bone which is called, in the language of anatomists, the tarso-metatarsus, and is usually free from feathers and covered with skin, is commonly carried erect in birds, so that the whole body is supported upon it; and from it the toes diverge outward. It is formed in birds of three separate bones blended together. In the fore limb of the Pterodactyle the metacarpus has the same relation to the bones of the fore-arm that the metatarsus has to the corresponding bones of the leg in a bird. But the three metacarpal bones in the Pterodactyle remain distinct from each other, perhaps because the main work of that region of the skeleton has devolved upon the digit called the wing finger, which is not recognised in the bird. In the Pterodactyles from the Solenhofen Slate there is a progressive number of phalanges in the three small digits of the hand, which were applied to the ground. This number in the great majority of species follows the formula of two bones in the first, three bones in second, and four in the third; so that in the innermost of the clawed digits only one bone intervenes between the metacarpal and the claw. The fingers slightly increase in length with increase in number of bones which form them.

The terminal claw bones are unlike the claws of Birds or Reptiles. They are compressed from side to side, and extremely deep and strong, with evidence of powerful attachment for ligaments, so that they rather resemble in their form and large size the claws of some of the carnivorous fossil reptiles, often grouped as Dinosauria, such as have been termed Aristosuchus and Megalosaurus. In the hand of the Ostrich the first and second digits terminate in claws, while the third is without a claw. But these claws of the Ostrich and other birds are slender, curved, and rather feeble organs. In the Archaeopteryx, a fossil bird which agrees with the Pterodactyles in retaining the separate condition of the metacarpal bones and in having the same number of phalanges in two of the fingers of the fore limb, the terminal claws are rather more compressed from side to side, and stronger than in the Ostrich, but not nearly so strong as in the Pterodactyle. The Archaeopteryx differs from the Pterodactyle in having no trace of a wing finger. The first metacarpal bone is short, as in all birds; and the first phalange scarcely lengthens that segment of the first digit of the Bird's hand to the same length as the other metacarpal bones. It therefore was not bent backward like the first digit in Pterodactyles. The wing finger, from which the genius of Cuvier selected the scientific name--Pterodactyle--for these fossils, yields their most distinctive character. It is a feature which could only be partly paralleled in the Bat, by making changes of structure which would remove every support to the wing but the outermost digit of that animal's hand. In the Bat's hand the membrane for flight is extended chiefly by four diverging metacarpal bones. There are only two or three phalanges in each digit in its four wing fingers. In Pterodactyles the metacarpal bones are, as we have seen, arranged in close contact, and take no part in stretching the wing.

THE WING FINGER

In Birds there is nothing whatever to represent the wing finger of the Pterodactyle, for it is an organ external to the finger bones of the bird, and contains four phalanges. The first phalange is quite different from the others. Its length is astonishing when compared with the small phalanges of the clawed fingers. The articular surface, which joins on to the wing metacarpal bone, is a concave articulation, which fits the pulley in which that bone ends. The pulley articulation admits of an extension movement in one direction only. Many specimens show the wing finger to be folded up so as to extend backward. The whole finger is preserved in other specimens straightened out so as to be in line with the metacarpus. This condition is well seen in Professor Marsh's specimen of Rhamphorhynchus, which has the wing membrane preserved, in which all bones of the fore-arm metacarpus and wing finger are extended in a continuous curve. The outer surface of the end of the first bone of the wing finger overlaps the wing metacarpal, so that a maximum of strength and resistance is provided in the bony structures by which the wing is supported. There is, therefore, in flight only one angular bend in the limb, and that is between the upper arm bone and the fore-arm.

An immense pneumatic foramen is situate in a groove on the under side of the upper end of the first phalange in Ornithocheirus, but is absent in specimens from the Kimeridge clay. This bone is long and stout. It terminates at the lower end in an obliquely truncated articular surface. Specimens occur in the Cambridge Greensand which are 2 inches broad at the upper end and nearly 1-1/2 inch wide at the lower end. An imperfect bone from the Chalk is 14-1/2 inches long. The bones are all flattened. Specimens from the Chalk of Kansas at Munich are 28 inches long. The second phalange is concave at the upper articular end and convex in the longer direction at the lower end. The articular points of union between the several phalanges form prominences on the under side of the finger in consequence of the adjacent bones being a little widened at their junction. It should be mentioned that there is a proximal epiphysis or separate bone to the first phalange, adjacent to the pulley joint of the metacarpal bone, which is like the separate olecranon process of the ulna of the fore-arm. It sometimes comes away in specimens from the Chalk and Cambridge Greensand, leaving a large circular pit with a depressed narrow border. On the outer side of this process is a rounded boss, which may possibly have supported the bone, if it were applied to the ground with the wing folded up, like the wing of a Bat directed upward and backward at the animal's side.

The four bones of the wing finger usually decrease progressively in length, so that in Rhamphorhynchus, in which the length of the animal's head only slightly exceeds 3-1/2 inches, the first phalange is nearly as long, the second phalange is about 3-1/4 inches, the third 2-3/4 inches, and the fourth a little over 2 inches. Thus the entire length of the four phalanges slightly exceeds 11 inches, or rather more than three times the length of the head. But the fore-arm and metacarpus in this type only measure 3 inches. Therefore the entire spread of wings could not have been more than 2 feet 9 inches.

The largest Ornithosaur in which accurate measurements have been made is the toothless Pterodactyle Ornithostoma, also named Pteranodon, from North America. In that type the head appears to have been about 3 or 4 feet long, and the wing finger exceeded 5 feet; while the length of the fore-arm and metacarpus exceeded 3 feet. The width of the body would not have been more than 1 foot. The length of the short humerus, which was about 11 inches, did not add greatly to the stretch of the wing; so that the spread of the wings as stretched in flight may be given as probably not exceeding 17 or 18 feet. A fine example of the wing bones of this animal quite as large has been obtained by the (British Museum Natural History). Many years ago, on very fragmentary materials, I estimated the wings in the English Cretaceous Ornithocheirus as probably having a stretch of 20 feet in the largest specimens, basing the calculation partly upon the extent of the longest wings in existing birds relatively to their bones, and partly upon the size of the largest associated bones which were then known.