CHAPTER XXXIV

THE CROSSOPTERYGII

=Class Teleostomi.=--We may unite the remaining groups of fishes into a single class, for which the name _Teleostomi_ (~teleos~, true; ~stoma~, mouth), proposed by Bonaparte in 1838, may be retained. The fishes of this class are characterized by the presence of a suspensorium to the mandible, by the existence of membrane-bones (opercles, suborbitals, etc.) on the head, by a single gill-opening leading to gill-arches bearing filamentous gills, and by the absence of claspers on the ventral fins. The skeleton is at least partly ossified in all the _Teleostomi_. More important as a primary character, distinguishing these fishes from the sharks, is the presence typically and primitively of the air-bladder. This is at first a lung, arising as a diverticulum from the ventral side of the oesophagus, but in later forms it becomes dorsal and is, by degrees, degraded into a swim-bladder, and in very many forms it is altogether lost with age.

This group comprises the vast majority of recent fishes, as well as a large percentage of those known only as fossils. In these the condition of the lung can be only guessed.

The _Teleostomi_ are doubtless derived from sharks, their relationship being possibly nearest to the _Ichthyotomi_ or to the primitive _Chimæras_. The Dipnoans among _Teleostomi_ retain the shark-like condition of the upper jaw, made of palatal elements, which may be, as in the _Chimæra_, fused with the cranium. In the lower forms also the primitive diphycercal or protocercal form of tail is retained, as also the archipterygium or jointed axis of the paired fins, fringed with rays on one or both sides.

We may divide the Teleostomes, or true fishes, into three subclasses: the _Crossopterygii_, or fringe-fins; the _Dipneusti_, or lung-fishes; _Actinopteri_, or ray-fins, including the _Ganoidei_ and the _Teleostei_, or bony fishes. Of these many recent writers are disposed to consider the _Crossopterygii_ as most primitive, and to derive from it by separate lines each of the remaining subclasses, as well as the higher vertebrates. The _Ganoidei_ and _Teleostei_ (constituting the _Actinopteri_) are very closely related, the ancient group passing by almost imperceptible degrees into the modern group of bony fishes.

=Subclass Crossopterygii.=--The earliest Teleostomes known belong to the subclass or group called after Huxley, _Crossopterygii_ (~krossos~, fringe; ~pteryx~, fin). A prominent character of the group lies in the retention of the jointed pectoral fin or archipterygium, its axis fringed by a series of soft rays. This character it shares with the _Ichthyotomi_ among sharks, and with the _Dipneusti_. From the latter it differs in the hyostylic cranium, the lower jaw being suspended from the hyomandibular, and by the presence of distinct premaxillary and maxillary elements in the upper jaw. In these characters it agrees with the ordinary fishes. In the living Crossopterygians the air-bladder is lung-like, attached by a duct to the ventral side of the oesophagus. The lung-sac, though specialized in structure, is simple, not cellular as in the Dipnoans. The skeleton is more or less perfectly ossified. Outside the cartilaginous skull is a bony coat of mail. The skin is covered with firm scales or bony plates, the tail is diphycercal, straight, and ending in a point, the shoulder-girdle attached to the cranium is cartilaginous but overlaid with bony plates, and the branchiostegals are represented by a pair of gular plates.

In the single family represented among living fishes the heart has a muscular arterial bulb with many series of valves on its inner edge, and the large air-bladder is divided into two lobes, having the functions of a lung, though not cellular as in the lung-fishes.

The fossil types are very closely allied to the lung-fishes, and the two groups have no doubt a common origin in Silurian times. It is now usually considered that the Crossopterygian is more primitive than the lung-fish, though at the same time more nearly related to the Ganoids, and through them to the ordinary fishes.

=Origin of Amphibians.=--From the primitive _Crossopterygii_ the step to the ancestral _Amphibia_, which are likewise mailed and semi-aquatic, seems a very short one. It is true that most writers until recently have regarded certain Dipneustans as the _Dipteridæ_ as representing the parents of the Amphibians. But the weight of recent authority, Gill, Pollard, Boulenger, Dollo, and others, seems to place the point of separation of the higher vertebrates with the Crossopterygians, and to regard the lobate pectoral member of _Polypterus_ as a possible source of the five-fingered arm of the frog. This view is still, however, extremely hypothetical and there is still much to be said in favor of the theory of the origin of Amphibia from Dipnoans and in favor of the view that the Dipnoans are also ancestors of the Crossopterygians.

In the true Amphibians the lungs are better developed than in the Crossopterygian or Dipnoan, although the lungs are finally lost in certain salamanders which breathe through epithelial cells. The gills lose, among the Amphibia, their primitive importance, although in _Proteus anguineus_ of Austria and _Necturus maculosus_, the American "mud-puppy" or water-dog, these persist through life. The archipterygium, or primitive fin, gives place to the chiropterygium, or fingered arm. In this the basal segment of the archipterygium gives place to the humerus, the diverging segments seen in the most specialized type of archipterygium (_Polypterus_) become perhaps radius and ulna, the intermediate quadrate mass of cartilage possibly becoming carpal bones, and from these spring the joints called metacarpals and phalanges. In the Amphibians and all higher forms the shoulder-girdle retains its primitive insertion at a distance from the head, and the posterior limbs remain abdominal.

The Amphibians are therefore primarily fishes with fingers and toes instead of the fringe-fins of their ancestors. Their relations are really with the fishes, as indicated by Huxley, who unites the amphibians and fishes in a primary group, _Ichthyopsida_, while reptiles and birds form the contrasting group of _Sauropsida_.

The reptiles differ from the Amphibians through acceleration of development, passing through the gill-bearing stages within the egg. The birds bear feathers instead of scales, and the mammals nourish their young by means of glandular secretions. Through a reptile-amphibian ancestry the birds and mammals may trace back their descent from palæozoic Crossopterygians. In the very young embryo of all higher vertebrates traces of double-breathing persist in all species, in the form of rudimentary gill-slits.

=The Fins of Crossopterygians.=--Dollo and Boulenger regard the heterocercal tail as a primitive form, the diphycercal form being a result of degradation, connected with its less extensive use as an organ of propulsion. Most writers who adopt the theory of Gegenbaur that the archipterygium is the primitive form of the pectoral fin are likely, however, to consider the diphycercal tail found associated with it in the _Ichthyotomi_, _Dipneusti_, _Crossopterygii_ as the more primitive form of the tail. From this form the heterocercal tail of the higher sharks and Ganoids may be derived, this giving way in the process of development to the imperfectly homocercal tail of the salmon, the homocercal tail of the perch, and the isocercal tail of the codfish and its allies, the gephyrocercal and the leptocercal tail, tapering or whip-like, representing various stages of degeneration. Boulenger draws a distinction between the protocercal tail, the one primitively straight, and the diphycercal tail modified, like the homocercal tail, from an heterocercal ancestry.

=Orders of Crossopterygians.=--Cope and Woodward divide the _Crossopterygia_ into four orders or suborders, _Haplistia_, _Rhipidistia_, _Actinistia_, and _Cladistia_. To the latter belong the existing species, or the family of _Polypteridæ_, alone. Boulenger unites the three extinct orders into one, which he calls _Osteolepida_. In all three of these the pectorals are narrow with a single basal bone, and the nostrils, as in the Dipneustans, are below the snout. The differences are apparently such as to justify Cope's division into three orders.

=Haplistia.=--In the _Haplistia_ the notochord is persistent, and the basal bones of dorsal and anal fins are in regular series, much fewer in number than the fin-rays. The single family _Tarrassiidæ_ is represented by _Tarrasius problematicus_, found by Traquair in Scotland. This is regarded as the lowest of the Crossopterygians, a small fish of the Lower Carboniferous, the head mailed, the body with small bony scales.

=Rhipidistia.=--In the _Rhipidistia_ the basal bones of the median fins ("axonosts and baseosts") are found in a single piece, not separate as in the _Haplistia_. Four families are recognized, _Holoptychiidæ_, _Megalichthyidæ_, _Osteolepidæ_, and _Onychodontidæ_, the first of these being considered as the nearest approach of the Crossopterygians to the Dipnoans.

The _Holoptychiidæ_ have the pectoral fins acute, the scales cycloid, enameled, and the teeth very complex. _Holoptychius nobilissimus_ is a very large fish from the Devonian. _Glyptolepis leptopterus_ from the Lower Devonian is also a notable species. _Dendrodus_ from the Devonian is known from detached teeth.

In the Ordovician rocks of Cañon City, Colorado, Dr. Walcott finds numerous bony scales with folded surfaces and stellate ornamentation, and which he refers with some doubt to a Crossopterygian fish of the family _Holoptychiidæ_. This fish he names _Eriptychius americanus_. If this identification proves correct, it will carry back the appearance of Crossopterygian fishes, the earliest of the Teleostome forms, to the beginning of the Silurian, these Cañon City shales being the oldest rocks in which remains of fishes are known to occur. In the same rocks are found plates of Ostracophores and other fragments still more doubtful. It is certain that our records in palæontology fall far short of disclosing the earliest sharks, as well as the earliest remains of Ostracophores, Arthrodires, or even Ganoids.

=Megalichthyidæ.=--The _Megalichthyidæ_ (wrongly called "_Rhizodontidæ_") have the pectoral fins obtuse, the teeth relatively simple, and the scales cycloid, enameled. There are numerous species in the Carboniferous rocks, largely known from fragments or from teeth. _Megalichthys_, _Strepsodus_, _Rhizodopsis_, _Gyroptychius_, _Tristichopterus_, _Eusthenopteron_, _Cricodus_, and _Sauripterus_ are the genera; _Rhizodopsis sauroides_ from the coal-measures of England being the best-known species.

The _Osteolepidæ_ differ from the _Megalichthyidæ_ mainly in the presence of enameled rhomboid scales, as in _Polypterus_ and _Lepisosteus_. In _Glyptopomus_ these scales are sculptured, in the others smooth. In _Osteolepis_, _Thursius_, _Diplopterus_, and _Glyptopomus_ a pineal foramen is present on the top of the head. This is wanting in _Parabatrachus_ (_Megalichthys_ of authors). In _Osteolepis_, _Thursius_, and _Parabatrachus_ the tail is heterocercal, while in _Diplopterus_ and _Glyptopomus_ it is diphycercal. _Osteolepis macrolepidotus_ and numerous other species occur in the Lower Devonian. _Diplopterus agassizii_ is common in the same horizon. _Megalichthys hibberti_ is found in the coal-measures, and _Glyptopomus minimus_ in the Upper Devonian. _Palæosteus_ is another genus recently described.

The _Onychodontidæ_ are known from a few fragments of _Onychodus sigmoides_ from the Lower Devonian of Ohio and _Onychodus anglicus_ from England.

=Order Actinistia.=--In the _Actinistia_ there is a single fin-ray to each basal bone, the axonosts of each ray fused in a single piece. The notochord is persistent, causing the back-bone in fossils to appear hollow, the cartilaginous material leaving no trace in the rocks. The genera and species are numerous, ranging from the Subcarboniferous to the Upper Cretaceous, many of them belonging to _Coelacanthus_, the chief genus of the single family _Coelacanthidæ_. In _Coelacanthus_ the fin-rays are without denticles. _Coelacanthus granulatus_ is found in the European Permian. _Coelacanthus elegans_ of the coal-measures is found in America also. In _Undina_ the anterior fin-rays are marked with tubercles. _Undina penicillata_ and _Undina gulo_ from the Triassic are well-preserved species. In _Macropoma_ (_lewesiensis_) the fin-rays are robust, long, and little articulated. Other genera are _Heptanema_, _Coccoderma_, _Libys_, _Diplurus_, and _Graphiurus_. _Diplurus longicaudatus_ was found by Newberry in the Triassic of New Jersey and Connecticut.

=Order Cladistia.=--In the _Cladistia_ the axis of the pectoral limb is fan-shaped, made of two diversified bones joined by cartilage. The notochord is restricted and replaced by ossified vertebræ. The axonosts of the dorsal and anal are in regular series, each bearing a fin-ray. The order contains the single family _Polypteridæ_. In this group the pectoral fin is formed differently from that of the other Crossopterygians, being broad, its base of two diverging bones with cartilage between. This structure, more specialized than in any other of the Crossopterygians or _Dipneusti_, has been regarded by Gill and others, as above stated, as the origin of the fingered hand (chiropterygium) of the frogs and higher vertebrates. The base of the diverging bones has been identified as the antecedent of the humerus, the bones themselves as radius and ulna, while the intervening non-ossified cartilage breaks up into carpal bones, from which metacarpals and digits ultimately diverge. This hypothesis is open to considerable doubt. The nostrils, as in true fishes, are superior. The body in these fishes is covered with rhombic enameled scales, as in the garpike; the head is similarly mailed, but, in distinction from the garpike, the anterior rays of the dorsal are developed as isolated spines.

The young have a bushy external gill with a broad scaly base. The air-bladder is double, not cellular, with a large air-duct joining the ventral surface of the oesophagus. The intestine has a spiral valve.

The cranium, according to Boulenger ("Poissons du Bassin du Congo," p. 11), is remarkable for its generalized form, this character forming a trait of union between the Ganoids and the primitive _Amphibia_ or _Stegocephali_. Without considering _Polypterus_, it is not possible to interpret the homologies of the cranium of the amphibians and the sharks.

The jaws are similar to those of the vertebrates higher than fishes. Tooth-bearing premaxillaries and dentaries are solidly joined at the front of the cranium, and united by a suture to the toothed maxillaries which form most of the edge of the mouth. Each half of the lower jaw consists of four elements, covering Meckel's cartilage, which is ossified at the symphysis. These are the articular, angular, dentary, and splenial (coronoid). Most of these bones are armed with teeth. The palato-suspensory consists of hyomandibular, quadrate, ectopterygoid, entopterygoid, metapterygoid, and palatine elements, the pterygoid elements bearing teeth. In _Erpetoichthys_ only the opercle is distinct among the gill-covers. In _Polypterus_ there is a subopercle also; the suborbital chain is represented by two small bones.

The gill-arches are four, but without lower pharyngeals. The teeth are conic and pointed, and in structure, according to Agassiz, they differ largely from those of bony fishes, approaching the teeth of reptiles.

The external gill of the young, first discovered by Steindachner in 1869, consists of a fleshy axis bordered above and below by secondary branches, themselves fringed. In form and structure this resembles the external gills of amphibians. It is inserted, not on the gill-arches, but on the hyoid arch. Its origin is from the external skin. It can therefore not be compared morphologically with the gills of other fishes, nor with the pseudobranchiæ, but rather with the external gills of larval sharks. The vertebræ are very numerous and biconcave as in ordinary fishes. Each of the peculiar dorsal spines is primitively a single spine, not a finlet of several pieces, as some have suggested. The enameled, rhomboid scales are in movable oblique whorls, each scale interlocked with its neighbors.

The shoulder-girdle, suspended from the cranium by post-temporal and supraclavicle, is covered by bony plates. To the small hypercoracoid and hypocoracoid the pectoral fin is attached. Its basal bones may be compared to those of the sharks, mesopterygium, propterygium, and metapterygium, which may with less certainty be again called humerus, radius, and ulna. These are covered by flesh and by small imbricated scales. The air-bladder resembles the lungs of terrestrial vertebrates. It consists of two cylindrical sacs, that on the right the longer, then uniting in front to form a short tube, which enters the oesophagus from below with a slit-like glottis. Unlike the lung of the _Dipneusti_, this air-bladder is not cellular, and it receives only arterial blood. Its function is to assist the respiration by gills without replacing it.

=The Polypteridæ.=--All the _Polypteridæ_ are natives of Africa. Two genera are known, no species having been found fossil. Of _Polypterus_, Boulenger, the latest authority, recognizes nine species: six in the Congo, _Polypterus congicus_, _P. delhezi_, _P. ornatipinnis_, _P. weeksi_, _P. palmas_, and _P. retropinnis_; one, _P. lapradei_, in the Niger; and two in the Nile, _Polypterus bichir_ and _P. endlicheri_. Of these the only one known until very recently was _Polypterus bichir_ of the Nile.

These fishes in many respects resemble the garpike in habits. They live close on the mud in the bottom of sluggish waters, moving the pectorals fan-fashion. If the water is foul, they rise to the surface to gulp air, a part of which escapes through the gill-openings, after which they descend like a flash. In the breeding season these fishes are very active, depositing their eggs in districts flooded in the spring. The eggs are very numerous, grass-green, and of the size of eggs of millet. The flesh is excellent as food.

The genus _Erpetoichthys_ contains a single species, _Erpetoichthys calabaricus_,[163] found also in the Senegal and Congo. This species is very slender, almost eel-like, extremely agile, and, as usual in wriggling or undulating fishes, it has lost its ventral fin. It lives in shallow waters among interlaced roots of palms. When disturbed it swims like a snake.

FOOTNOTES:

[163] This genus was first called _Erpetoichthys_, but the name was afterwards changed by its author, J. A. Smith, to _Calamoichthys_, because there is an earlier genus _Erpichthys_ among blennies, and a _Herpetoichthys_ among eels. But these two names, both wrongly spelled for _Herpetichthys_, are sufficiently different, and the earlier name should be retained. "A name in science is a name without necessary meaning" and without necessarily correct spelling. Furthermore, if names are spelled differently, they are different, whatever their meaning. The efforts of ornithologists, notably those of Dr. Coues, to spell correctly improperly formed generic names have shown that to do so consistently would throw nomenclature into utter confusion. It is well that generic names of classic origin should be correctly formed. It is vastly more important that they should be stable. Stability is the sole function of the law of priority.