Popular Science Monthly/Volume 60/April 1902/The Evolution of Fishes



WHEN a fish dies he leaves no friends. His body is at once attacked by hundreds of creatures ranging from the one-celled protozoa and bacteria to individuals of his own species. His flesh is devoured, his bones are scattered, the gelatinous substance in them decays, and the phosphate of lime is in time dissolved in the water. For this reason few fishes of the millions who die each year leave any trace for future preservation. At the most, a few teeth, a fin spine or a bone buried in the clay might remain intact or in such condition as to be recognized.

But now and then it happens that a dead fish may fall in more fortunate conditions. On a sea bottom of fine clay the bones, or even the whole body, might be buried in such a way as to be sealed up and protected from total decomposition. The flesh would in any case disappear and leave no mark or at the most a mere cast of its surface. But the hard parts might persist, and now and then they do persist, the lime unchanged or else silicified or subjected to some other form of chemical substitution. Only the scales, the teeth, the bones, the spines and the fin rays can be preserved in the rocks of sea or lake bottom. In a few localities, as Green River in Wyoming, Monte Bolca in Tuscany and Mount Lebanon in Syria, with certain quarries in Scotland and lithographic stones in Germany, many skeletons of fishes have been found, pressed flat in layers of very fine rock, their structures traced as delicately as if actually drawn on the smooth stone. Fragments preserved in ruder fashion abound in the clays and even the sandstones of the earliest geologic ages. In most cases, however, fossil fishes are known from detached and scattered fragments, many of them, especially of the sharks, by the teeth alone. Fishes have occurred in all ages from the Silurian to the present time and no doubt the very first lived long before the Silurian.

No one can say what the earliest fishes were like, nor do we know what was their real relation to the worm-like forms among which men have sought their presumable ancestors, nor to the Tunicates and other chordate forms, not fish-like, but still degenerate relatives of the primeval fish.

From analogy we may suppose that the first fishes which ever were bore some resemblance to the lancelet, for that is a fish-like creature with every structure reduced to the lowest terms. But as the lancelet has no hard parts, no bones, nor teeth, nor scales, nor fins, no traces of its kind are found among the fossils. If the primitive lish was like it in important respects, all record of this has necessarily vanished from the earth.

The next group of living fishes, the Cyclostomes, including the hagfishes and lampreys,—fishes with small skull and brain but without limbs or jaws,—stands at a great distance above the lancelet in complexity of structure, and equally far from the true fishes in its primitive simplicity. In fact the lamprey is farther from the true fish in structure than a perch is from an eagle. Yet for all that it may be an offshoot from the primitive line of fish-descent. There is not much in the structure of the lamprey which may be preserved in the rocks. But the cartilaginous skull, the backbone, fins and teeth may sometimes leave their traces in soft clay or lithographic stone. These parts are actually represented in a fossil form bearing some likeness to a lamprey found in the Devonian of Scotland known as Palæospondylus. This early form was a highly specialized one with well-defined vertebrae, a fact which shows that the group had reached a relatively high development even in these early times. The few existing Cyclostomes may even be looked upon as degraded types as compared with their Devonian ancestors.

Besides the lampreys the Devonian seas swarmed with mysterious creatures covered with a coat of mail, fish-like in some regards, but limbless, without true jaws and having next to nothing in common with the fishes of to-day. These are called Ostracophori, and most recent authorities, as Woodward and Dean, are inclined to regard them as highly modified or specialized lampreys, a side offshoot which has left no descendants among recent forms. Cope has compared these Ostracoderms to the larva of Tunicates, and Patten has suggested their affinity to spiders, or rather to the horseshoe crab (Limulus), which is at least as much spider as crab. The weight of evidence would rather place them nearest the lampreys, but certainly not in the same natural class.

Among these forms in coat of mail are some in which the jointed and movable angles of the head suggest the pectoral spines of some catfishes. But in spite of its resemblance to a fin, the spine of the pectoral fin in Pterichthys is an outgrowth of the ossified skin and has no more homology with the spines of fishes than the mailed plates have with the bones of a fish's cranium. In none of these fishes has any trace of an internal skeleton been found. It must have retained its primitive gelatinous character. There are, however, some traces of eyes, and the mucous channels of the lateral line indicate that these creatures possessed some other special senses.

Whatever the Ostracophores may be, they are not true fishes and they should not be included within the much-abused term Ganoidei, a word once used in the widest fashion for all sorts of mailed fishes, but little by little restricted to the hard-scaled relatives and ancestors of the gar-pike of to-day.

Dimly seen in the vast darkness of Paleozoic time are the huge figures known as Arthrognaths. These are mailed and helmeted fishes, limbless so far as we know, but with sharp, notched, turtle-like jaws quite different from those of the fish or those of any animal alive to-day. These creatures appear in Silurian rocks and are especially abundant in the fossil beds of Ohio, where Newberry, Claypole, Eastman and Dean have patiently studied the broken fragments of their armor. Most of them have a great casque on the head with a shield at the neck and a movable joint connecting the two. Among them was almost every variation in size and form.

These creatures have been often called ganoids, but with the true ganoids like the gar-pike they have seemingly nothing in common. They are very different from the Ostracophores. To regard them as derived from ancestral Dipnoi is to give a possible guess as to their origin and a very unsatisfactory guess at that. But they have all passed away in competition with the scaly fishes and sharks of later evolution, and it seems certain that they, like the mailed Ostracophores, have left no descendants.

Next after the lampreys, but a long way after in structure as in time, come the sharks. With the sharks appear for the first time true limbs and the lower jaw. The upper jaw is still formed from the palate, and the shoulder girdle is attached behind the skull. "Little is known," says Professor Dean, "of the primitive stem of the sharks and even the lines of descent of the different members of the group can only be generally suggested. The development of recent forms has yielded few results of undoubted value to the phylogenist. It would appear is if paleontology alone could solve the puzzles of their descent."

Of the very earliest sharks in the Upper Silurian age the remains are too scanty to prove much save that there were sharks in abundance and variety. Spines, teeth, fragments of shagreen, show that in some regards these forms were highly specialized. In the Carboniferous age, according to Dean, 'there occurred the culminating point of their differentiation when specialized sharks existed, whose varied structures are paralleled only by those of the existing bony fishes,—sharks fitted to the most special environment, some minute and delicate, others enormous, heavy and sluggish, with stout head and fin spines and elaborate types of dentition.'

The sharks are, however, doubtless evolved from the still more primitive shark without limbs and with the teeth slowly formed from modification of the ordinary shagreen prickles. In determining the earliest among the several primitive types of shark we are stopped by an undetermined question of theory. What is the origin of paired limbs? Are these formed, like the unpaired fins, from the breaking up of a continuous fold of skin, in accordance with the view of Balfour and others? Or is the primitive limb, as supposed by Gegenbaur, a modification of the bony gill-arch? Or again, as supposed by Kerr, is it a modification of the hard axis of an external gill?

If we adopt the views of Gegenbaur or Kerr, the earliest type of limb is the jointed archipterygium, a series of consecutive rounded cartilagenous elements with a fringe of rays along its length. Sharks possessing this form of limb (Ichthyotomi) appear in the earliest rocks, and from these the Dipnoi, on the one hand, may be descended and, on the other, the true sharks and the Chimæras.

On the other hand, if we regard the paired fins as parts of a lateral fold of skin, we find primitive sharks to bear out our conclusions. In Cladoselache of the Subcarboniferous, the pectoral and the ventral fins are long and low, and arranged just as they might be if Balfour's theory were true. Acanthöessus, with a spine in each paired fin and no other rays, might be a specialization of this type or fin, and Climatius with rows of spines in place of pectorals and ventrals might be held to bear out the same idea. But in all these, the tail is less primitive than in the Ichthyotomi. On the whole, however, there is much to be said on the primitive nature of the Ichthyotomi, and Pleuracanthus, with the tapering tail and jointed pectoral fin of a dipnoan, with other traits of a shark, is as likely as Cladoselache to point directly to the origin of the shark-like forms.

Hasse finds this origin in a hypothetical group of Polyospondyli which have many vertebra undifferentiated and without calcareous material. These fishes are represented only by fin spines (Onchus), which may have belonged to something else. These gave rise to Ichthyotomi, with jointed fins, and through these to Dipnoi and a long series leading to the bony fishes on the one hand and on the other to the Amphibia, Reptiles and Higher Vertebrates.

The branch of higher sharks would lead to the Diplospondyli of Hasse's system, of which Cladoselache should be a primitive example. These sharks have the weakly ossified vertebræ joined together in pairs and there are six or seven gill openings. This primitive type called Notidani has persisted to our day, the frilled shark (Chlamydoselachus) and the genera Hexanchus and Heptranchias, still showing its archaic characters.

Here the sharks diverge into two groups, the one with the vertebræ better developed and its calcareous matter arranged star fashion. This forms Hasse's group of Asterospondyli, the typical sharks. The earliest forms (Heterodontidæ, Hybodontidæ) approach the Notidani, and one such ancient type Heterodontus, still persists. The others diverge to form the three chief groups of cat-sharks (Scyliorhinus, etc.), mackerelsharks (Lamna, etc.) and iron-sharks (Carcharhias, etc.).

In the other group the vertebræ have their calcareous matter arranged in rings, one or more about the notochordal center. In all these the anal fin is absent, and in the process of specialization, is formed the flattened body and broad fins of the ray. This group is called Tectospondyli. Hasse's Cyclospondyli (sharks with one ring of calcareous matter) constitute the most primitive extreme of a group representing continuous evolution.

From Cladoselache and Chlamydoselachus through the sharks to the rays we have an almost continuous series which reaches its highest development in the devil rays or mantas of the tropical seas, Manta and Mobula being the most specialized genera and among the very largest of the fishes. However different the rays and skates may appear in form and habit, they are structurally similar to the sharks and have sprung from the main shark stem.

The most ancient offshoot from the shark stem, perhaps dating before Silurian times and having its root in ancient Ichthyotomi, is the group of Holocephali or chimæras, shark-like in essentials, but differing widely in details. Of these there are but few living forms, and the fossil types are known only from dental plates and fin spines. The living forms are found in the deeper seas, the world over, the most primitive genus being the newly discovered Rhinochimæra. The fusion of the teeth into overlapping plates, the covering of the gills by a dermal flap, the complete union of the palatoquadrate apparatus or upper jaw with the skull and the development of a peculiar clasping spine on the forehead of the male are characteristic of the chimæras. The group is one of the most ancient, but with the chimæras it ends, for the species has nothing in common with modern fishes except what both have derived from their common ancestors the sharks.

The most important offshoot of the primitive sharks is not the chamæras, nor even the shark series itself, but the group of dipnoans or lung-fishes and the long chain of their descendants. With the dipnoan appears the lung or air-bladder, at first an outgrowth from the ventral side of the oesophagus, as it still is in all higher animals, but later turning over, among fishes, and springing from the dorsal side. At first an arrangement for breathing air, a sort of accessory gill—it becomes the sole organ of respiration in the higher forms, while in the bony fishes its respiratory function is lost altogether. The air-bladder is a degenerate gill. In the dipnoans the shoulder girdle moves forward to the skull, and the pectoral limb, a jointed and fringed archipterygium, apparently derived from ancestors of the type of Pleuracanthus is its characteristic appendage. The shark-like structure of the mouth remains.

The few living lung fishes resemble the salamanders almost as closely as they do fishes, and they may well be ranged as a class by themselves midway between the primitive sharks and the amphibians. The few living forms show these intermediate characters in the development of lungs and the primitive character of the pectoral and ventral limbs. Those now extant give but little idea of the great variety of extinct dipnoans, but the obvious suggestion that with the lung fish is the place of divergence of the higher vertebrates from the fish series may be the correct one. The living genera are three in number, Neoceratodus in Australian rivers, Lepidosiren in the Amazon and Protopterus in the Nile. These are all mud fishes, some of them living through most of the dry season encased in a cocoon of dried mud. Of these forms Neoceratodus is certainly the nearest to the ancient forms, but its embryology, owing to the shortening of its growth stages due to its environment, has thrown little light on the question of its ancestry.

From some branch of the dipnoans the ancestry of the amphibians and through them that of the reptiles, birds and mammals may be traced, although some reason exists for regarding the primitive Crossopterygium as the point of divergence. It may be that the Crossopterygian gave rise to Amphibian and Dipnoan alike.

In the process of development we next reach the characteristic fish mouth in which the upper jaw is formed of maxillary and premaxillary elements distinct from the skull. The upper jaw of the shark is part of the palate, the palate being fused with the quadrate bone which supports the lower jaw. That of the dipnoan is much the same. The development of a typical fish mouth is the next step in evolution and with its appearance we note the decline of the air-bladder in size and function.

The next great offshoot is the group of crossopterygians, fishes which still retain the old-fashioned the of pectoral and ventral fin, the archipterygium. In the archaic tail, enameled scales and cartilaginous skeleton the crossopterygian shows its affinity with its dipnoan ancestry. Thus these fishes unite in themselves traits of the shark, lung-fish and Ganoid. The few living crossopterygians, Polypterus, and Erpetoichthys are not very different from those which prevailed in Devonian times. The larvæ possess external gills with firm base and fringelike rays, suggesting a resemblance to the pectoral fin itself which develops from the shoulder-girdle just below it and would seem to give some force to Kerr's contention that the archipterygium is only a modified external gill. In Polypterus the archipterygium has become short and fan-shaped, its axis made of two diverging bones with flat cartilage between. From this type it is thought that the arm of the higher forms has been developed. The bony basis may be the humerus, from which diverge radius and ulna, the carpal bones being formed of the intervening cartilage.

From the crossopterygians springs the main branch of true fishes, known collectively as Actinopteri, those with ordinary rays on the paired fins instead of the jointed archipterygium. The transitional series of primitive Actinopteri is here called by the name of Ganoid. The ganoid differs from the Crossopterygian in having basal elements of the paired fins small and concealed within the flesh. But other associated characters of the Crossopterygii and Dipnoi are preserved in most of the species. Among these are the mailed head and body, the heterocercal tail, the cellular air-bladder, the presence of valves in the arterial bulb, the presence of a spiral valve in the intestine and of a chiasma in the optic nerves. All these characters are found in the earlier types so far as known, and all are more or less completely lost or altered in the teleosts or bony fishes. Among the existing ganoids the gar-pike (Lepisosteus) is the last of a long series of Mesozoic forms of the same general structure. The gar-pikes are cylindrical or arrow-shaped. Among these early types is every variety of form, some of them being almost as long as deep, and every intermediate form being represented. An offshoot from this line is the bow-fin (Amia calva), perhaps the closest living ally of the bony fishes, showing distinct affinities with the great group to which the herring and salmon belong. Near relatives of the bow-fin fiourished in the Mesozoic, among them some with a forked tail, and some with a very long one. From forms of this type the body of recent fishes may be descended.

Another branch of ganoids, widely divergent from both gar-fish and bow-fin and not recently from the same primitive stock, included the sturgeons (Acipenser, Scaphirhynchus, Kessleria) and the paddle-fishes (Polyodon and Psephurus). These differ widely from any other types, recent or fossil, showing features of degeneration as compared with their extinct ancestors, while again sturgeon and paddlefish differ widely from, each other. It has been suggested that the cat-fishes (Siluridæ) are descended from the sturgeons, but the resemblance vanishes the more closely the groups are compared, nor are we anywhere sure of the point where any part of the teleost series is joined to the ganoids. We can only say that the sturgeons are more or less degraded ganoids with cartilaginous skeletons, of unknown derivation and of unsettled relationships.

All other fishes have ossified instead of cartilaginous skeletons. The dipnoan and ganoid traits one by one are more or less completely lost. Through these the main line of fish development continues and the various groups are known collectively as bony fishes or teleosts.

The earliest of the true bony fishes or teleosts appear in Mesozoic times, the most primitive forms being soft-rayed fishes with unmodified vertebræ, allied more or less remotely to the herring of to-day. In these and other soft-rayed fishes the pelvis still retains its posterior insertion, the ventral fins being said to be abdominal. The next great stage in evolution brings the pelvis forward, attaching it to the shoulder girdle so that the ventral fins are now thoracic as in the perch and bass. If brought to a point in front of the pectoral fins, a feature of specialized degradation, they become jugular as in the cod-fish. In the abdominal fishes the air bladder still retains its rudimentary duct joining it to the oesophagus.

From the abdominal forms allied to the herring, the huge array of modern fishes, typified by the perch, the bass, the mackerel, the wrasse, the globe-fish, the sculpin, the seahorse and the cod descended in many diverging lines. The earliest of the spine-rayed fishes with thoracic fins belong to the type of Berycidæ, a group characterized by rough scales and the retention of the primitive larger number of ventral rays. These appear in the Cretaceous or chalk deposits, and show various attributes of transition from the abdominal to the thoracic type of ventrals.

Another line of descent apparently distinct from that of the herring and salmon extends through the characins to the loach, carps, cat-fishes and electric eel. The fishes of this series have the anterior vertebrae coossified and modified in connection with the hearing organ, a structure not appearing elsewhere among fishes. This group includes the majority of fresh-water fishes. Still another great group, the eels, have lost the ventral fins and the bones of the head have suffered much degradation.

The most highly developed fishes, all things considered, are doubtless the allies of the perch, bass and sculpin. These fishes have lost the air-duct and on the whole they show the greatest development of the greatest number of structures. But they do not represent an excessive degree of specialization. In other groups their traits one after another are carried to an extreme and these stages of extreme specialization give way one after another to phases of degeneration. The specialization of one organ usually involves degeneration of some other. Extreme specialization of any organ tends to render it useless under other conditions and may be one step toward its final degradation.

"We have thus seen, in hasty review, that the fish-like vertebrates spring from an unknown and possibly worm-like stock,—that from this stock, before it became vertebrate, degenerate branches have fallen off, represented to-day by the Tunicates and Balanoglossus. We have seen that the primitive vertebrate was headless and limbless without hard parts. The lancelet remains as a possible direct off-shoot from it; the cyclostome with brain and skull is a probable derivative from archaic lancelets. The earliest fishes leaving traces in the rocks were doubtless cyclostomes, limbless, naked lampreys and mailed ostracophores. The lampreys gave rise to sharks and chimæras. The sharks developed into rays in one right line, and into the highest sharks along another; while by a side branch through lost stages the primitive sharks passed into dipnoans or lung-fishes. All these types and others abound in the Devonian age and the early records were lost in the Silurian. From the lung-fishes or the ancestors or descendants by the specialization of the lung and limbs, the land animals, at first amphibians, after these reptiles, birds and mammals, arose.

In the sea, by a line still more direct, through the gradual emphasis of fish-like characters, we find developed the crossopterygians with archaic limbs and after these the ganoids with fish-like limbs but otherwise archaic; then the soft-rayed and finally the spiny-rayed bony fishes which culminate in specialized and often degraded types, as the anglers, globe-fishes, parrot-fishes and flying gurnards. Side branches are the ostracophores, perhaps from primitive lampreys; arthrognaths from a lost ancestry possibly; chimseras, from primitive sharks, and rays from sharks less primitive; dipnoan sturgeons, gar-pikes and perhaps carp and cat-fishes; and from each of the ultimate lines of descent radiate infinite branches till the sea and rivers are filled, and almost every body of water has fishes fitted to its environment.