A Contribution to the Morphology and Development of the Pectoral Skeleton of Teleosteans

The method pursued was that of reconstruction in wax from serial sections as described in my previous papers.

Swirski, Wiedersheim, and Ducret are the only workers who have dealt with the same subject during recent years. Of these the first alone attempted to reconstruct the pectoral skeleton at different stages; the others confined themselves to descriptions and to figures of sections.

Before proceeding to the main body of the paper I should like to record my indebtedness to the Research Committee of the Royal Society for a grant towards the expenses incurred, to Prof. T. W. Bridge and Prof. W. C.M’Tntosh for much useful material, and finally to Mr. G-. A. Boulenger for his kindly criticism.

Younger stages than any at my disposal were described by Ducret. From his observations it is seen that the cartilaginous skeleton first begins to appear as procartilage in embryos 9·6 mm. long. When they are 13·4 mm. long the skeleton has become cartilaginous, and consists of a lamina lying in the fin, and of a basal plate in the body wall.

Stage I (figs. 1 and 2).—At this stage the fin is a halfmoon-shaped fold which, on account of the relative smallness of the embryo as compared with the yolk-sac, points dorsally. Consequently the morphologically ventral side of the fin faces laterally. It is largely supported internally by a thin concavo-convex plate of hyaline cartilage (f.p.) with its concave face towards the inner side. Its shape calls to mind that of the larval Polypterus (Budgett), and it differs chiefly in the breadth of its base of attachment. The fin-plate is continuous with the girdle (sc. and co.), which lies in the body wall, and is also hyaline. These two regions are far from being sharply marked off from one another either by histological structure or difference of orientation. The scapulo-coracoid region is but a continuation of the tin-plate, and with that forms the same concavo-convexity. Such figures as that of Ducret’s (pl. II, fig. 12) give the impression that the fin-plate is attached at right angles to the plane of the girdle or “plaque vasale,” and therefore that the portion of the girdle dorsal to the line of attachment is the scapula, and that which is ventral is the coracoid. Such an impression is erroneous and is due to the study of sections without the aid of reconstructions.

The boundary between these two regions can therefore be ascertained only by following the line of attachment of the fin-fold to the body. This is represented by a dotted line (g. b.) in the figures, and may be spoken of as the glenoid line. It is along this that the glenoid articulation is formed at a later stage, though there is no sign of its formation at present. At this early stage the glenoid line is practically parallel to the long axis of the animal’s body.

The girdle is triradiate with a large fenestra (fn. 4) at the junction of the radii. Through this there passes a branch of the fourth spinal nerve to supply the ventral muscles of the fin, and also a small artery and vein.

The dorsal radius is the scapulo-coracoid region (sc. co.). It is continuous with the fin-plate along its upper, and is folded inwards on its lower, border (fig. 2). The inner edge of the fold bears a slight outgrowth—a kind of scapular, process(sc.) near its anterior end. Posteriorly it is much thickened, aud bears a much stronger outgrowth (m.), which is the first trace of the mesocoracoid bar. At a point (fn. 1—3) lying a short distance internal to this folded edge, and between the two outgrowths, a stout nerve forks, sending one branch to the true dorsal, and the other to the ventral, surfaces of the fins. This nerve is formed by the union of the first three spinal nerves.

The anterior radius (pr. p.) is long and rod-like, and looks almost like an antero-ventral continuation of the mesocoracoid thickening. At a later stage it trends medianly and meets its fellow of the opposite side in the middle line. At this stage, however, owing to the size of the yolk-sac, it bends somewhat laterally. Gegenbaur, and Swirski following him, calls this the procoracoid. The latter describes it as arising independently from the rest of the pectoral skeleton in the pike. But though Wiedersheim has investigated the same fish he finds no evidence for this independence. Neither from Ducret’s observations nor my own on the salmon is there anything in support of this. It will be sufficient, therefore, to regard it as a process of the coracoid, and to call it the præcoracoid process.

The posterior radius (po.p.) stands nearly at right angles to the general plane of the girdle and the fin-plate. It runs for a short distance from the post-axial border of the latter into the body wall. It corresponds to the “coracoid” of

Swirski, the “Sockelstück,” or “processus posticus ”of Wiedersheim, and the “processus ensiformis ”of Ducret. For the sake of uniformity it may be called the post-coracoid process.

Stage II (figs. 3 and 3a).—The skeleton, as a whole, is no longer concave towards the inner side, but both fin-plate and girdle are flat. The boundary between them is marked by an angle (fig. 3a, g. b.) due to the fact that the fin is no longer vertical, but stands out laterally.

The glenoid line is still practically parallel to the long axis of the body.

The fin-plate has undergone a considerable change in shape (cf. figs. 1 and 3), consisting mainly in a rounding off of the posterior angle due to a relative shortening of the posterior border (P.) and a growing out of the distal border. Near the front border (A) a large foramen has appeared, which marks off the region of the future first radial (r.).

The girdle is still triradiate, but now all its radii lie in the same plane, which is nearly vertical (3a, co., pr. p.), tending to slope inwards towards the mid-ventral line. The foramen for the branch of the fourth spinal nerve (fn. 4), which loomed so large at the previous stage, is now extremely small, and has taken up its final position near the base of the post-coracoid process, and may, therefore, be conveniently referred to as the post-coracoid foramen.

The cartilage of the scapular region has grown forwards so that the ventral branch from the common trunk of the first three spinal nerves has become half surrounded. The deep notch thus formed (fn. 1—3) eventually becomes a foramen—the scapular foramen.

The coracoid region has grown more rapidly than the scapular, as may be seen by comparing the region anterior to the post-coracoid foramen in figs. 1 and 3. This growth has brought the anterior border (f. b.) close against a blood-vessel which supplies the ventral muscles of the fin.

The mesocoracoid swelling, which at the previous stage was situated near the anterior margin of the coracoid, is no larger, and is situated just ventral to the middle of the glenoid line (fig. 3 a. m.) and halfway between the scapula and postcoracoid process.

The præcoracoid process has undergone no change beyond a relative increase in size and a slight median rotation.

The postcoracoid process also is slightly larger, and is now in a straight line with the præcoracoid process.

Stage III (figs. 4 and 5).—Hitherto the fin-plate has tended to be the dominant feature of the pectoral skeleton, but now it has yielded to the girdle. Both regions are still continuous, and the line between them horizontal. They are both orientated in the same way as in Stage II, though the hinder half of the fin-plate has begun to turn in such a way as to face somewhat posteriorly.

In the fin-plate the first radial element (r) has become a separate piece of cartilage. Another fenestra has appeared in the centre of the plate, and depressions on either side of it indicate the position of future fenestræ. Thus the four radials are blocked out.

Along the distal border of the second and third radials three nodules of cartilage have been differentiated. These are the distal radials. The first two belong to the first radial, the third distal to the third radial. There is a close similarity in everything but size between these distal elements and the first radial.

The scapulo-coracoid has now grown round the nerve and blood-vessel which lay in front of it, and thus the scapular (fn. 1—3) and præcoracoid (f. b.) foramina have been formed. In consequence of this growth the large space which formerly existed between the coracoid and cleithrum (cl.) has almost disappeared.

The mesocoracoid process (fig. 5, m.), though still small, is larger than before. Owing to the growth of the coracoid it is no longer situated close under the glenoid line. The apparent reduction of the præcoracoid process is due to the fact that its proximal portion has been taken up into the extending coracoid. The distal portion is remarkably slender, but does not yet meet its fellow of the opposite side.-For the greater part of its length it is, in contact with the cleithrum.

The most striking feature at this stage is the great length of thq postcoracoid process, which rivals even the præ-coracoid process. It is interesting to note that this process forms a posterior continuation of the cartilage bordering the glenoid line.

Stage IV (figs. 6 and 7).—The pectoral skeleton.now bears a close resemblance to that of the adult, and differs from the previous stages chiefly in the orientation of parts. Whilst the scapular region still retains an upright position, the coracoid region, and with it the fin, has rotated inward? into an almost horizontal position (fig. 6). Consequently the glenoid line, which has hitherto been parallel to the long axis of the body, now slopes obliquely backwards and inwards towards this axis. This change was foreshadowed in Stage III, when the hinder part of the fin-plate faced slightly posteriorly.,

The fin-plate has now broken up completely into the four radials indicated in Stage III. These are now rod-like, and increase in length from the pre-axial to the post-axial element. At its distal end each bears a pair of distal radials —mere nodules of cartilage. These increase in size in a reverse direction, viz. from post-to pre-axial border, a fact which suggests that the first formed radial element also belongs to the distal series. In support of this it may be pointed out that like the other distal radials it appears before the radials themselves; that it is not in the same line with the latter, but with the former; and finally that, like all the distal elements, it is enclosed by the base of the dermal fin-ray.

The scapular region has now, for the first time, risen above the pre-axial corner of the fin skeleton, which is thus removed some distance from the cleithrum (cf., figs. 4 and 7).

The coracoid region has extended still further forward, so that the præcoracoid foramen is a considerable distance from its anterior border.

The mesocoracoid (fig. 6, m.) is now fully developed, having extended up to and united with the top of the scapula. It is a significant fact that the mesocoracoid remains dormant so long and then develops so rapidly. The rapidity of its development may be judged from the fact that the young in my Stages I—III vary in length from 15·5 mm. to 22 mm.j whilst those of Stage IV are only 25 mm. long. This is no doubt associated with the equally rapid rotation of the coracoid into the horizontal position; and just as the scapula rises to form a buttress for the outer end of the glenoid border, so the mesocoracoid does likewise for the inner end. In this connection it is interesting, as further evidence, to compare the position of the base of the mesocoracoid as seen in figs. 5, 6, and 8, and to note its increasing proximity to this border. Ducret observed the late appearance of the mesocoracoid (p. 21).

The growth of the præcoracoid process has been almost stationary, and so in consequence of the forward growth of the scapulo-coracoid region it has been pulled away from that intimate relation to the cleithrum which it has had up to this stage. Nevertheless it now meets its fellow in the middle line.

The postcoracoid process is still present, but is greatly reduced.

The Adult (fig. 8).—Already in the salmon 52 mm. long, the girdle is as fully ossified as in the adult, and unfortunately I have no stages between Stage IV and this. Concerning the adult little need be said.here, for it has been well described by Parker, though he overlooked the præcoracoid foramen.

The Development of the Pectoral Skeleton in Gasterosteus aculeatus.

Stage I (fig. 9).—Comparing this with the corresponding stages of the salmon one is struck at one time by the great resemblances, at another by the equally great differences. In both the girdle and fin together form a continuous whole; concavo-convex in shape with the concavity facing inwards. In both the fin predominates over the girdle, though here this predominance is carried much further. This no doubt is associated with the fact that the stickleback begins to lead an active free-swimming life at a much earlier stage than does the salmon.

On the other hand in the stickleback there is an entire absence of both præ- and meso-coracoid processes. The glenoid line tends to be more vertical. The postcoracoid process is enormously developed, being about the same size as the whole of the remainder of the girdle. The postcoracoid foramen is unrepresented, and the fourth nerve passes down to a muscle which runs along by the side of the process. The level at which it reaches this muscle is indicated by an asterisk (n. 4).

The common trunk for the first three spinal nerves forks at the position indicated (fn. 1—3).

Stage II (fig. 10).—Owing to the growth of the girdle the disproportion between it and the fin-plate has diminished considerably. There is still no trace of a division between them, and together they are concavo-convex as before.

The fin-plate shows no sign of differentiation into radials.

The scapulo-coracoid region has. grown much further anteriorly and dorsally, thus bringing about a rotation of the glenoid line into an almost vertical position. The præcoracoid process has no actual existence, but is represented by the antero-ventral angle of the coracoid.

The postcoracoid process is relatively larger than before and has become bent downwards.

Stage III (fig. 11).—The fin and girdle still form a continuous piece of cartilage without any sign of either foramina or radials. It is now much flatter.

The glenoid line makes about the same angle with the vertical as it did in Stage II, but its dorsal end comes much nearer to the upper corner of the fin-plate.

The scapulo-coracoid is very little altered, and the whole of its anterior margin is applied to the slender spicular cleithrum. Its antero-ventral corner is much more prominent (pr.p.), and may with reserve be spoken of as a process.

The postcoracoid process has attained astonishing proportions. It has rotated upwards. Its exact position is between the anterior portion of the trunk muscles and the skin (figs. 13—14). It is evidently a structure of considerable importance, but in what way is difficult to tell.

Stage IV (fig. 12).—The skeleton now bears a close resemblance to that of the adult, and, as in that, the scapulo-coracoid forms by far the greater portion. It is flat and plate-like, and stands in a vertical position comparable with that of the scapula in Stage IV of the salmon. The glenoid articulation has been formed by the breaking down of the cartilage.

The tendency for the glenoid line to rotate from a nearly horizontal into a vertical position which we have followed through the earlier stages has now reached its consummation.

The fin-plate has been perforated in three places preparatory to the formation of the four radials.

The cartilage of the scapular region has extended dorsally and anteriorly with the result that its upper end is on a level with the top corner of the fin skeleton. Its dorso-anterior border has been folded inwards and continued into a process (sc/). A similar infolding was noted by Swirski in the pike. He cautiously suggested that it might be a relic of the mesocoracoid (p. 31). I have also seen it in the adult pike, and it seems to me more likely that it serves to supply a larger surface of contact with, and secures a firmer hold to, the cleithrum.

The forward growth of the scapula has brought about the complete enclosure of the nerve, which up till now has run round its anterior border.

The præcoracoid process is now more worthy of the name, though still very small as compared with that of the salmon. It has no relation to its fellow. The whole front border of the girdle from the tip of the scapula to the end of the process is in the closest contact with the cleithrum, thus furnishing a striking contrast to the state of affairs in the salmon.

Perhaps the most interesting feature at this stage is the ossification of the still very long postcoracoid process. This extends later to the coracoid region, thus giving rise to the bone of that name.

Close to, but by no means continuous with, the postcoracoid process, is the “interclavicle ”(i. cl.), which is a true dermal bone as opposed to the coracoid which, as development shows, is a cartilage bone.

The Adult.—The pectoral skeleton of the adult has been accurately described by Parker. It differs from that of Stage IV in the much greater extension of the scapulo-cora-coid in an antero-posterior direction. Also the scapula has: risen high above the fin skeleton, and the foramen has become extraordinarily large.

The “interclavicle” is now completely co-ossified with the çoracoid.

This process was first seen and described by Swirski in the development of the pike. In that fish it is the most striking feature of the earlier stages, and remains prominent until a fairly late stage. By comparison with Silurns and through that with sturgeon he came to regard it as homologous with Gegenbaur’s “third process,” and Parker’s “coracoid.” Wiedersheim found it also in developmental stages of grayling, catfish, and pike, and called it “processus posticus.” He absolutely disagreed with Swirski in his interpretation, because the process was directed caudally, and ran parallel to the long axis of the body. He also examined larvæ of the sturgeon, and not finding it in these he regarded it as a new structure peculiar to the Teleosts, developed for purposes of fixing and support (p. 170). Ducret found it also in the salmon, and called it “processus ensiformis.”

It is present during development in the cod and herring, but I have failed to findit in either Siphonostomaor Anguilla. Both the latter fish, however, are of a very special type, and are characterised by the great reduction of cartilage in all parts of the skeleton. In a larva of Amia 19 mm. long there is no sign of it.

To Wiedersheim’s objection to Swirski’s interpretation it may be added that whereas in those forms which, according to Gegenbaur, have a well-developed “under process ”it is closely associated with the cleithrum, the postcoracoid process is never, at any stage, related to that bone.

Owing to the researches of Humphrey, Thacher, and Regan, no doubt remains concerning the similar nature of the median and paired fins. This is perhaps best shown by the pelvic fin of Psephurus (Regan). In this fish there is very little to choose in the anatomical details between the pelvic and the anal fins. In both each ray of the fin consists of three segments, which—using Bridge’s terminology—may be called proximal (axonost), mesial (baseost), and distal (marginal) radials. In both there is a tendency towards the fusion of the anterior proximal radials. This is carried further in the pelvic than in the anal fin. The plate formed by the fusion of these elements is the very primitive pelvic girdle. The essential community of structure between the pectoral fins and the median is not so obvious; but, bearing in mind the researches of palæontologists, e. g. on Cladoselache and’ Cladodus, we have no difficulty in seeing it between the pectoral and pelvic fins of Psephurus. The distal and mesial elements are still present in the fin, and, as pointed out by Regan, the metapterygium and girdle represent the proximal elements, though in a much more specialised condition than in the pelvic fin. The metapterygium, therefore, has the same origin as the girdle, and is fundamentally the posterior continuation of it. The postcoracoid process has precisely the same relationships (e. g. fig. 9), and may be regarded as the homologue of the metapterygium.

Further facts may be given in support of this conclusion.

In the adult sturgeon the metapterygium lies almost in the region where the base of the fin passes into the body-wall, and has not yet rotated much outwards to form part of the free fin. It thus occupies the same position as the process.

It is significant that in the Ganoids which possess a metapterygium, development, so far as known, has revealed no postcoracoid process. In the Teleosts, which have no metapterygium, the process is very generally present, and present in widely diverse types.

The fact that the process is continuous with the girdle is not a serious objection to the view, for in early development the radials are continuous with one another and with the girdle.

This view does not agree with Gegenbaur’s conclusion (p. 161) that the metapterygium of Ganoids has become one of the radials of the Teleosts. Based, as it is, upon the study only of adult living forms, his conclusion is so far sound. But all diversities of structure are not wrapped up in Lepidosteus and Amia. Other types have existed, and many facts of Teleostean morphology will probably find their explanation only among them.

At its first appearance the pectoral skeleton of the sturgeon consists of five short rays and a plate, the “primäres basale.” From Mollier’s figures it is evident that the line of attachment of the rays to the plate is parallel to the long axis of the body. The skeleton at this stage, then, bears a close resemblance to the pelvic skeleton, and presents such a condition as the fin-fold theory of the origin of paired fins requires. Mollier tells us that the second to the sixth spinal nerves converge to the “primäres basale,” and bifurcate medially to it, sending one branch to the dorsal and the other to the ventral side. In the adult (Pl. 23, fig. 15) the second and third nerves, united with a branch from the hypoglossal or first spinal, form a common trunk (n. 1—3), which bifurcates in the muscle canal (m. c.), sending one branch (n. 1—3d) to the dorsal muscles, and. the other (n. 1—3D) through the scapular foramen to the ventral muscles.

The fourth nerve (n. 4)likewise bifurcates, sending one branch to a plexus which supplies the dorsal muscles, and the other through a passage (fn. 4) to the ventral muscles. The entrance to this passage was seen by Parker and called the “coracoid foramen” but was overlooked by Gegenbaur. It evidently corresponds to the postcoracoid foramen. The fifth spinal nerve (n. 5) repeats the conditions of the fourth with this difference, that its ventral branch has no foramen, but passes to the ventral fin muscles by way of the inner side of the metapterygium. Thus the anterior part of the “primäres basale” is represented by that portion of the girdle which lies between the glenoid facettes and the scapular and postcoracoid foramina, whilst the posterior part becomes the metapterygium. This is proved to be the case by Mollier’s figures.

Whilst the metapterygium has retained the primitive position parallel to the axis of the body, the girdle has rotated so that the glenoid border, whilst still horizontal, is nearly at right angles to this axis. The radial elements upon it no longer project at right angles, as they must have done in early fishes, but are almost parallel to the metapterygium. Herein, then, lies a cause for the reduction in number and importance of the radials borne by the metapterygium. Thus the latter, instead of rotating out to become the main bearer of the fin skeleton, as in other fishes than the Teleostomes, remains in the body-wall, and becomes a minor factor.

All these changes are probably associated with the increasing importance of the dermal fin rays, and with the acquirement of the power to lay that portion of the fin flat against the side of the body, or to extend it so that it may act as a brake instead of a keel.¹

It is much to be regretted that Mollier did’not make careful reconstructions of the developing pectoral skeleton in the sturgeon. The study of sections alone cannot give a clear idea of the changes in place relations.

In both salmon and stickleback at the earliest stage the glenoid approximates to the most primitive position, viz. parallel to the long axis of the body. In the former, as development advances, the fin-plate and coracoid region rotate, so that the glenoid border slopes obliquely inwards. In the latter no such movement takes place, and the glenoid line rotates into a vertical position.

Unfortunately, Swirski, in making reconstructions for the pike, did not insert the fin plate; nevertheless, examination of his figure shows that in it the movements of the glenoid line are the same as those in the stickleback.

We thus have two absolutely divergent tendencies exhibited by the developing pectoral skeleton of the salmon and sturgeon on the one hand, and by the stickleback and pike on the other. These find their concisest expression in the statement that, whilst both start with the glenoid line in practically the same position, in the one it rotates horizontally round the pre-axial end, in the other it rotates vertically round the post-axial end. The following diagram will serve to make the movements clear.

Attention has already been drawn (p. 369) to the close relationship which exists between the presence of a mesocoracoid and the horizontal rotation of the glenoid border. If the facts of development are any guide at all, it is very difficult to see how the type of girdle without a mesocoracoid has been derived from that with a mesocoracoid, and vice versa. In the development of the former there is no sign of a vanishing mesocoracoid or of an inwardly moving glenoid border.

In the most recent classification of Teleosts (Boulenger) all the non-ostariophysous fishes with a mesocoracoid are placed in the sub-order Malacopterygii. The lowliest forms without it are placed in the Haplomi. In the definition of this suborder this sentence occurs:—“ The absence of the mesocoracoid arch distinguishes the Haplomi from the Malaco-pterygii, with which they are united by various authors.” This statement alone suffices to show that, apart from the absence of this arch, the Haplomi, at least as represented by the pike and Galaxias, are as primitive as the general run of Malacopterygii. If a list be made of the features which indicate lowly affinity among the latter, it will be seen that most, if not all, are found among the Haplomi. Again, in my paper on Oromeria (p.270) these words about Galaxias occur:—“ In some respects, e. g. forward extension of the cranial cavity, and the condition of the articular head of the hyomandibular, it is as lowly as, or even more lowly than, the salmon.”

Whatever may be said about the classification of other Teleosts, it is generally acknowledged that the Ostariophysi form a natural group. Though every conceivable modification of form and external character occurs amongst them, yet the mesocoracoid and the features associated with its presence remain constant. Again, among Malacopterygii there are forms which are extraordinarily specialised, but they nevertheless retain the mesocoracoid. These facts speak strongly for the stability of this structure.

The absence of a mesocoracoid in Polypterus was pointed out by Gegenbaur, and indicates the possibility that forms may have existed even among Holostei exhibiting the same peculiarity. Unfortunately fossils have as yet yielded no information on the primary pectoral girdle of these fishes.

If there be any truth in the view just propounded, then we are able to recognise among Teleosts three distinct offshoots from the Holostean stock, viz. one including the Ostariophysi, another the Malacopterygii, and a third the Haplomi and all those teleosts with the same type of pectoral girdle.

This bone was first described in the stickleback by Parker, and because of its resemblance to the true clavicle (interclavicle) of the sturgeon he called it “interclavicle.” Recently Starks has disputed this decision, and after a careful study of the pectoral skeleton of Hemibranchii has come to the conclusion that this bone has no separate existence, but is only an extension of the coracoid (hypocoracoid). A glance at Pl. 23, fig. 12 shows him to be wrong in this conclusion for there it will be seen that the “interclavicle ”is quite separate from the coracoid.

This reopens the question of its homologies once more.

This bone is related to the post-coracoid process, which, as we have seen, very probably represents the metapterygium. Now the “interclavicle ”of the sturgeon is related only to the lower tip of the coracoid (Pl. 23, fig. 15, co.), and does not come near the metapterygium. The two bones, therefore, are not homologous, and the “interclavicle” of Gaste rosteus and its al lies shouldbe called by a different name, viz. “infradeithrum.”

Mr. Regan has called my attention to the presence of a similarly placed bone in Eurypholis (Woodward, p. 207). There can be little doubt that it is an infracleithrum also.

1. The earlier stages in the development of the pectoral skeleton of the salmon are fundamentally the same as those of the stickleback (p. 370).

2. The mesocoracoid appears late in the development of the salmon, and is associated with the rotation of the glenoid border into a transversely horizontal position (p. 368).

3. There is no sign of a mesocoracoid during development in the stickleback, and this is associated with the rotation of the glenoid border into a vertical position (pp. 370 and 371).

4. Those Teleosts without a mesocoracoid probably constitute a group separate and not originating from those Teleosts with a mesocoracoid (pp. 377 and 378).

5. The postcoracoid process, which is so prominent a feature during development, represents the metapterygium (pp. 373 and 374).

6. The so-called “interclavicle” of the stickleback is not a part of the coracoid, but arises as a separate dermal ossification. It is not homologous with the bone of the same name in the sturgeon, and should therefore receive a different name, e.g. infracleithrum (p. 379).

Illustrating Dr. H. H. Swinnerton’s paper, “A Contribution to the Morphology and Development of the Pectoral Skeleton of Teleosteans.”

A. Preaxial border of the fin. Cleithrum, co. Coracoid or coracoid region. /. b. Foramen for blood-vessel, or præcoracoid foramen, f. n. Foramen for spinal nerve, fn. 1-3. Scapular foramen, fn. 4. Postcoracoid foramen. f.p. Fin plate, g.b. Glenoid line or border, i. cl. Infracleithrum. m. Mesocoracoid arch or bone. m. c. Canal for dorsal fin muscles, n. 1. First spinal or hypoglossal nerve, n. 1-5. Spinal nerves 1-5. P. Post-axial border of fin. po. p. Postcoracoid process, pr. p. Præcoracoid process. r. First formed radial, r.â. Distal radial, r.m. Mesial radial, r.p. Proximal radial, sc. Scapula or scapular region. sc’. Scapular process, sp.c. Spinal cord. st. Stomach, t. tn. Trunk muscle.

Figs. 1—7, 9—11 were drawn from wax models reconstructed from serial sections.

FIGS. 1—8.—Pectoral skeleton of Salmo salar.

FIG. 1.—Stage 1. 15·5 mm. long, × 68. External view.

FIG. 2.—Ditto. Internal view.

FIG. 3.—Stage II. 19·0 mm. long. × 37. External view.

FIG. 3A.—Stage II. Diagrammatic vertical section.

FIG. 4.—Stage III. 22·0 mm. long. × 32. External view.

FIG. 5.—Ditto. Internal view.

FIG. 6.—Stage IV. 25·0 mm. long. × 30. Internal view.

FIG. 7.—Ditto. Ventro-lateral view.

FIG. 8.—Adult. Internal view.

FIGS. 9—12.—Pectoral skeleton of Gasterosteus aculeatus.

FIG. 9.—Stage I. 3·6 mm. long. × 157. External view.

FIG. 10.—Stage II. 4·0 mm. long. × 157. External view.

FIG. 11.—Stage III. 5·4 mm. long, × 118. External view.

FIG. 12.—Stage IV. 11·0 mm. long. × 52. External view.

FIG. 13.—Gasterosteus aculeatus, Stage IV. Transverse section through the post-pectoral region. × 30.

FIG. 14.—Ditto. Portion of the same section, as in Fig. 13. × 150.

FIG. 15.—Acipenser sturio. Pectoral girdle and associated nerves posterior view, × ⁠.