The Eels.—We may here break the sequence from the Isospondyli to the other soft-rayed fishes, to interpolate a large group of uncertain origin, the series or subclass of eels.

The mass of apodal or eel-like fishes has been usually regarded as constituting a single order, the Apodes (ἄ, without; ποῦς, foot). The group as a whole is characterized by the almost universal separation of the shoulder-girdle from the skull, by the absence of the mesocoracoid arch on the shoulder-girdle, by the presence of more than five pectoral actinosts, as in the Ganoid fishes, by the presence of great numbers of undifferentiated vertebræ, giving the body a snake-like form, by the absence in all living forms of the ventral fins, and, in all living forms, by the absence of a separate caudal fin. These structures indicate a low organization. Some of them are certainly results of degeneration, and others are perhaps indications of primitive simplicity. Within the limits of the group are seen other features of degeneration, notably shown in the progressive loss of the bones of the upper jaw and the membrane-bones of the head and the degradation of the various fins. The symplectic bone is wanting, the notochord is more or less persistent, the vertebral centra always complete constricted cylinders, none coalesced. But, notwithstanding great differences in these regards, the forms have been usually left in a single order, the more degraded forms being regarded as descended from the types which approach nearest to the ordinary fishes. From this view Professor Cope dissents. He recognizes several orders of eels, claiming that we should not unite all these various fishes into a single order on account of the eel-like form. If we do so, we should place in another order those with the fish-like form. It is probable, though not absolutely certain, that the Apodes are related to each other. The loss among them, first, of the connection of the post-temporal with the skull; second, of the separate caudal fin and its hypural support; third, of the distinct maxillary and premaxillary; and fourth, of the pectoral fins, must be regarded as successive phases of a general line of degradation. The large number of actinosts, the persistence of the notochord, the absence of spines, and the large numbers of vertebræ seem to be traits of primitive simplicity. Special lines of degeneration are further shown by deep-sea forms. What the origin of the Apodes may have been is not known with any certainty. They are soft-rayed fishes, with the air-bladder connected by a tube with the œsophagus, and with the anterior vertebræ not modified. In so far they agree with the Isospondyli. In some other respects they resemble the lower Ostariophysi, especially the electric eel and the eel-like catfishes. But these resemblances, mainly superficial, may be wholly deceptive; we have no links which certainly connect the most fish-like Apodes with any of the other orders. Probably Woodward's suggestion that they may form a series parallel with the Isospondyli and independently descended from Tertiary Ganoids deserves serious consideration. Perhaps the most satisfactory arrangement of these fishes will be to regard them as constituting four distinct orders for which we may use the names Symbranchia (including Ichthyocephali and Holostomi), Apodes (including Enchelycephali and Colocephali), Carencheli, and Lyomeri.

Order Symbranchia.—The Symbranchia are distinguished by the development of the ordinary fish mouth, the maxillary and premaxillary being well developed. The gill-openings are very small, and usually confluent below. These fresh-water forms of the tropics, however eel-like in form, may have no real affinity with the true eels. In any event, they should not be placed in the same order with the latter.

The eels of the suborder Ichthyocephali (ιχθύς, fish; κεφαλή, head) have the head distinctly fish-like. The maxillary, premaxillary, and palatines are well developed, and the shoulder-girdle is joined by a post-temporal to the skull. The body is distinctly eel-like, the tail being very short and the fins inconspicuous. The number of vertebræ is unusually large. The order contains the single family Monopteridæ, the rice-field eels, one species, Monopterus albus, being excessively common in pools and ditches from China and southern Japan to India.

The eels of the suborder Holostomi (ὀλός, complete; στόμα, mouth) differ from these mainly in the separation of the shoulder-girdle from the skull, a step in the direction of the true eels. The Symbranchidæ are very close to the Monopteridæ in external appearance, small, dusky, eel-like inhabitants of sluggish ponds and rivers of tropical America and the East Indies. The gill-openings are confluent under the throat. Symbranchus marmoratus ranges northward as far as Vera Cruz, having much the habit of the rice-field eel of Japan and China. The Amphipnoidæ, with peculiar respiratory structures, abound in India. Amphipnous cuchia, according to Günther, has but three gill-arches, with rudimentary lamina and very narrow slits. To supplement this insufficient branchial apparatus, a lung-like sac is developed on each side of the body behind the head, opening between the hyoid and the first branchial arch. The interior of the sac is abundantly provided with blood-vessels, the arterial coming from the branchial arch, whilst those issuing from it unite to form the aorta. Amphipnous has rudimentary scales. The other Holostomi and Ichthyocephali are naked and all lack the pectoral fin.

The Chilobranchidæ are small sea-fishes from Australia, with the tail longer than the rest of the body, instead of much shorter as in the others.

No forms allied to Symbranchus or Monopterus are recorded as fossils.

Order Apodes, or True Eels.—In this group the shoulder-girdle is free from the skull, and the bones of the jaws are reduced in number, through coalescence of the parts.

Three well-marked suborders may be recognized, groups perhaps worthy of still higher rank: Archencheli, Enchelycephali, and Colocephali.

Suborder Archencheli.—The Archencheli, now entirely extinct, are apparently the parents of the eels, having, however, certain traits characteristic of the Isospondyli. They retain the separate caudal fin, with the ordinary hypural plate, and Professor Hay has recently found, in an example from the Cretaceous of Mount Lebanon, remains of distinct ventral fins. These traits seem to indicate an almost perfect transition from the Isospondyli to the Archencheli.

One family may be recognized at present, Urenchelyidæ.

The earliest known eel, Urenchelys avus, occurs in the upper Cretaceous at Mount Lebanon. It represents the family Urenchelyidæ, apparently allied to the Anguillidæ, but having a separate caudal fin. Its teeth are small, conical, blunt, in many series. There are more than 100 vertebræ, the last expanded in a hypural. Pectorals present. Scales rudimentary; dorsal arising at the occiput. Branchiostegals slender, not curved around the opercle. Urenchelys anglicus is another species, found in the chalk of England.

Suborder Enchelycephali.—The suborder Enchelycephali (ἔγχελυς, eel; κεφαλή, head) contains the typical eels, in which the shoulder-girdle is free from the skull, the palatopterygoid arch relatively complete, the premaxillaries wanting or rudimentary, the ethmoid and vomer coalesced, forming the front of the upper jaw, the maxillaries lateral, and the cranium with a single condyle. In most of the species pectoral fins are present, and the cranium lacks the combined degradation and specialization shown by the morays (Colocephali).

Family Anguillidæ.—The most primitive existing family is that of the typical eels, Anguillidæ, which have rudimentary scales oblong in form, and set separately in groups at right angles with one another. These fishes are found in the fresh and brackish waters of all parts of the world, excepting the Pacific coast of North America and the islands of the Pacific. In the upper Great Lakes and the upper Mississippi they are also absent unless introduced. The species usually spawn in the sea and ascend the rivers to feed. But some individuals certainly spawn in fresh water, and none go far into the sea, or where the water is entirely salt. The young eels sometimes ascend the brooks near the sea in incredible numbers, constituting what is known in England as "eel-fairs." They will pass through wet grass to surmount ordinary obstacles. Niagara Falls they cannot pass, and according to Professor Baird "in the spring and summer the visitor who enters under the sheet of water at the foot of the falls will be astonished at the enormous numbers of young eels crawling over the slippery rocks and squirming in the seething whirlpools. An estimate of hundreds of wagon-loads, as seen in the course of the perilous journey referred to, would hardly be considered excessive by those who have visited the spot at a suitable season of the year." "At other times large eels may be seen on their way down-stream, although naturally they are not as conspicuous then as are the hosts of the young on their way upstream. Nevertheless it is now a well-assured fact that the eels are catadromous, that is, that the old descend the watercourses to the salt water to spawn, and the young, at least of the female sex, ascend them to enjoy life in the fresh water."

Reproduction of the Eel.—Dr. Gill ("Riverside Natural History," p. 103) gives the following account of the reproduction of Anguilla:

"The generation of the eel was long involved in great mystery, and the knowledge thereof is one of the recent acquisitions of scientific investigation. So late, indeed, as 1880 it was declared that 'their mode of propagation is still unknown.' In want of positive knowledge the rein has been given to loose hypothesis and conjecture. It has been variously asserted that eels were generated from slime, from dew, and from the skins of old eels or of snakes. The statement that they come from horse-hairs is familiar to many country boys, and the origin of this belief is due simply to the fact that there are certain aquatic worms, known under the generic name Gordius, which are elongated and apparently smooth like the eel, and which may be found in the same waters. It was one of the ideas of the Greek to attribute their paternity, as of many other doubtful offspring, to the convenient Jupiter. The statement that they are viviparous has arisen from two causes: one the existence of intestinal worms, and the other from the confusion of the eel with an elongated and consequently eel-like but otherwise very different form, the Zoarces viviparus. The Zoarces is indeed, in Germany as well as in the Scandinavian countries, generally known as the Aal-mutter, or eel-mother, and thus in its name perpetuates the fancy. Even where eels are to be found in extreme abundance, and where they are the objects of a special culture, like erroneous opinions prevail. Thus, according to Jacoby, about the lagoon of Comacchio there is an 'ineradicable belief among the fishermen that the eel is born of other fishes; they point to special differences in color and especially in the common mullet, Mugil cephalus, as the causes of variation in color and form among eels. It is a very ancient belief, widely prevalent to the present day, that eels pair with water-snakes. In Sardinia the fishermen cling to the belief that a certain beetle, the so-called water-beetle, Dytiscus ræselii, is the progenitor of eels, and they therefore call this "mother of eels."' The assignment of such maternity to the water-beetle is doubtless due to the detection of the hair-worm, or Gordius, in the insect by sharp-sighted but unscientific observers, and, inasmuch as the beetle inhabits the same waters as the eel, a very illogical deduction has led to connect the two together.

"All such beliefs as have been thus recounted are due to the inconspicuous nature of the generative organs in eels found in fresh waters and at most seasons—a characteristic which is in strong contrast to the development of corresponding parts in fishes generally. Nevertheless the ovaries of the eel were discovered, as long ago as 1707, by Dr. Sancassini of Comacchio, and described by the celebrated Valisneri (after whom the plant Valisneria was named) in 1710, again by Mondini in 1777, and almost contemporaneously by O. J. Müller of Denmark. Later the illustrious Rathke (in 1824, 1838, and 1850) and also Hornbaum-Hornschuch published the results of special investigations, and figured the eggs. But it was only in 1873 (after several futile endeavors by others) that the male organ of the eel was recognized, also by an Italian naturalist, Dr. Syrski, in small individuals of the species, and a previous idea that the eel was hermaphroditic thereby dispelled. The sexual differences are correlated with external ones, and generally the males and females, when adult, can be told apart. Jacoby testifies that he examined large numbers with a view to solve this question. The most important differences relate to (1) size; (2) form of the snout; (3) color; (4) dorsal fin; and (5) size of the eyes. (1) The males rarely attain a length of more than seventeen to nineteen inches, while adult females are generally much larger; (2) the snout in the male is attenuated and rather pointed, while in the female it is comparatively broad and blunt; (3) the male is of a deep darkish green, or often a deep black with a shining luster and a whitish belly, while the female has a clearer color, usually of a greenish hue on the back and yellowish on the belly; (4) the dorsal fin is lower and less developed in the male than in the female; and (5) the eye of the male is large and that of the female, as a rule, comparatively small. These characters, however, do not always hold good. Jacoby remarked that 'special reference having been paid to the height and narrowness of the dorsal fin, much success has been met with in picking out, in the fish-market of Trieste, the eels which possessed the organ of Syrski (that is, the male organ); absolute certainty, however, in recognizing them cannot be guaranteed. If one is searching among living eels with no characters in mind,—with the exception of the first, that of length,—he will find in every ten eels, on an average, eight females and two with the supposed male organ; but if the selection is made with a careful reference to all these marks of difference, the proportion changes, and out of every ten examples about eight will be found with the supposed male organ.'

"According to Herr Benecke, 'it may be assumed with the greatest safety that the eel lays its eggs like most other fish, and that, like the lamprey, it spawns only once and then dies. All the eggs of a female show the same degree of maturity, while in the fish which spawn every year, besides the large eggs which are ready to be deposited at the next spawning period, there exist very many of much smaller size, which are destined to mature hereafter and be deposited in other years. It is very hard to understand how young eels could find room in the body of their mother if they were retained until they had gained any considerable size. The eel embryo can live and grow for a long time supported by the little yolk, but, when this is done, it can only obtain food outside of the body of its mother. The following circumstances lead us to believe that the spawning of the eel takes place only in the sea: (1) that the male eel is found only in the sea or brackish water, while female eels yearly undertake a pilgrimage from the inland waters to the sea, a circumstance which has been known since the time of Aristotle, and upon the knowledge of which the principal capture of eels by the use of fixed apparatus is dependent; (2) that the young eels, with the greatest regularity, ascend from the sea into the rivers and lakes.'"

All statements in opposition to this theory are untenable, since the young eels never find their way into landlocked ponds in the course of their wanderings, while eels planted in such isolated bodies of water thrive and grow rapidly, but never increase in numbers. Another still more convincing argument is the fact that in lakes which formerly contained many eels, but which, by the erection of impassable weirs, have been cut off from the sea, the supply of eels has diminished, and after a time only scattering individuals, old and of great size, are taken in them. An instance of this sort occurred in Lake Muskengorf in West Prussia. If an instance of the reproduction of the eel in fresh water could be found, such occurrences as these would be quite inexplicable.

In the upper stretches of long rivers the migration of the eels begins in April or in May; in their lower stretches and shorter streams, later in the season. In all running waters the eel-fishery depends upon the downward migrations; the eels press up the streams with occasional halts, remaining here and there for short periods, but always make their way above. They appear to make the most progress during dark nights, when the water is troubled and stormy, for at this time they are captured in the greatest numbers. It is probable that after the eels have once returned to the sea and there deposited their spawn, they never can return into fresh water, but remain there to die. A great migration of grown eels in spring or summer has never been reported, and it appears certain that all the female eels which have once found their way to the sea are lost to the fisherman.

Food of the Eel.—Eels, in the words of Mr. W. H. Ballou, are "among the most voracious of carnivorous fishes. They eat most inland fishes, except the garfish and the chub. Investigation of six hundred stomachs by Oswego fishermen showed that the latter bony fish never had a place in their bill of fare. They are particularly fond of game-fishes, and show the delicate taste of a connoisseur in their selection from choice trout, bass, pickerel, and shad. They fear not to attack any object when disposed, and their bite in human flesh shows even a vicious attitude towards man. On their hunting excursions they overturn huge and small stones alike, working for hours if necessary, beneath which they find species of shrimp and crayfish, of which they are exceedingly fond. Of shrimps they devour vast numbers. Their noses are poked into every imaginable hole in their search for food, to the terror of innumerable small fishes."

In the opinion of Mr. Ballou, too, "eels are to the water what the fishhawk is to the air. They are, perhaps, the most powerful and rapid of natatorians. Again, they hide in the mud beneath some log or overhanging rock, and dart out with tremendous fury at the unsuspecting prey. They attack the spawn of other fishes open-mouthed, and are even said to suck the eggs from an impaled female. They fearlessly and rapidly dive head-foremost in the mud, disappearing from view in the twinkling of a star. They are owl-like in their habits, committing many of their depredations at night.

"No fish is yet reported to utilize a full-grown eel as food. Pickerel, garfish, and bass, which are particularly numerous in these lakes, are supposed to literally devour the young fry. Mr. Sawyer describes the operation of the pickerel darting through a long column of young eels open-mouthed and devouring vast numbers of them."

Larva of the Eel.—The translucent band-shaped larva of the common eel has been very recently identified and described by Dr. Eigenmann. It is probable that all true eels, Enchelycephali, pass through a band-shaped or leptocephalous stage, as is the case with Albula and other Isospondyli. In the continued growth the body becomes firmer, and at the same time much shorter and thicker, gradually assuming the normal form of the species in question.

In a recent paper Dr. Carl H. Eigenmann has very fully reviewed the life-history of the eel. The common species live in fresh waters, migrating to the sea in the winter. They deposit in deep water minute eggs that float at the surface. The next year they develop into the band-shaped larva. The young eels enter the streams two years after their parents drop down to the sea. It is doubtful whether eels breed in fresh water. The male eel is much smaller than the female.

The eel is an excellent food-fish, the flesh being tender and oily, of agreeable flavor, better than that of any of its relatives. Eels often reach a large size, old individuals of five or six feet in length being sometimes taken.

Species of Eels.—The different species are very closely related. Not more than four or five of them are sharply defined, and these mostly in the South Seas and in the East Indies. The three abundant species of the north temperate zone, Anguilla anguilla of Europe, Anguilla chrisypa of the eastern United States, and Anguilla japonica of Japan, are scarcely distinguishable. In color, size, form, and value as food they are all alike.

Fossil species referred to the Anguillidæ are known from the early Tertiary. Anguilla leptoptera occurs in the Eocene of Monte Bolea, and Anguilla elegans in the Miocene of Œningen in Baden. Other fossil eels seem to belong to the Nettastomidæ and Myridæ.

Pug-nosed Eels.—Allied to the true eel is the pug-nosed eel, Simenchelys parasiticus, constituting the family of Simenchelyidæ. This species is scaled like a true eel, has a short, blunt nose, and burrows its way into the bodies of halibut and other large fishes. It has been found in Newfoundland and Madeira. Another family possessing rudimentary scales is that of the Synaphobranchidæ, slender eels of the ocean depths, widely distributed. In these forms the gill-openings are confluent. Synaphobranchus pinnatus is the best-known species.

Conger-eels.—The Leptocephalidæ, or conger-eels, are very similar to the fresh-water eels, but are without scales and with a somewhat different mouth, the dorsal beginning nearer to the head.

The principal genus is Leptocephalus, including the common conger-eel (Leptocephalus conger) of eastern America and Europe and numerous very similar species in the tropics of both continents. These fishes are strictly marine and, reaching the length of five or six feet, are much valued as food. The eggs are much larger than those of the eel and are produced in great numbers, so that the female almost bursts with their numbers. Dr. Hermes calculated that 3,300,000 were laid by one female in an aquarium.

These eggs hatch out into transparent band-like larva, with very small heads formerly known as Leptocephalus, an ancient name which is now taken for the genus of congers, having been first used for the larva of the common conger-eel. The loose watery tissues of these "ghost-fishes" grow more and more compact and they are finally transformed into young congers.

The Murænesocidæ are large eels remarkable for their strong knife-like teeth. Murænesox savanna occurs in the West Indies and in the Mediterranean, Murænesox cinereus in Japan, and Murænesox coniceps on the west coast of Mexico, all large and fierce, with teeth like shears. The Myridæ are small and worm-like eels closely allied to the congers, having the tail surrounded by a fin, but the nostrils labial. Myrus myrus is found in the Mediterranean. Species of Eomyrus, Rhynchorhinus, and Paranguilla apparently allied to Myrus occur in the Eocene. Other related families, mostly rare or living in the deep seas, are the Ilyophidæ, Heterocongridæ, and Dysommidæ.

The Snake-eels.—Most varied of the families of eels is the Ophichthyidæ, snake-like eels recognizable by the form of the tail, which protrudes beyond the fins. Of the many genera found in tropical waters several are remarkable for the sharply defined coloration, suggesting that of the snake. Characteristic species are Chlevastes colubrinus and Leiuranus semicinctus, two beautifully banded species of Polynesia, living in the same holes in the reefs and colored in the same fashion. Another is Callechelys melanotænia. The commonest species on the Atlantic coast is the plainly colored Ophichthus gomesi.

In the genus Sphagebranchus, very slender eels of the reefs, the fins are almost wanting.

Allied to the Congers is the small family of duck-billed eels (Nettastomidæ) inhabiting moderate depths of the sea. Nettastoma bolcense occurs in the Eocene of Monte Bolca. The produced snout forms a transition to the really extraordinary type of thread-eels or snipe-eels (Nemichthyidæ), of which numerous genera and species live in the oceanic depths. In Nemichthys the long, very slender, needle-like jaws are each curved backward so that the mouth cannot by any possibility be shut. The body is excessively slender and the fish swims with swift undulations, often near the surface, and when seen is usually taken for a snake. The best-known species is Nemichthys scolopaceus of the Atlantic and Pacific. Nemichthys avocetta, very much like it, has been twice taken in Puget Sound.

Suborder Colocephali, or Morays.—In the suborder Colocephali (κολός, deficient; κεφαλή, head) the palatopterygoid arch and the membrane-bones generally are very rudimentary. The skull is thus very narrow, the gill-structures are not well developed, and in the chief family there are no pectoral fins. This group is very closely related to the Enchelycephali, from which it is probably derived.

In the great family of morays (Murænidæ) the teeth are often very highly developed. The muscles are always very strong and the spines bite savagely, a live moray being often able to drive men out of a boat. The skin is thick and leathery, and the coloration is highly specialized, the pattern of color being often elaborate and brilliant. In Echidna zebra for example the body is wine-brown, with cross-stripes of golden yellow. In Muræna each nostril has a barbel. Muræna helena, the oldest moray known, is found in Europe. In Gymnothorax, the largest genus, only the anterior nostrils are thus provided. Gymnothorax mordax of California is a large food-fish, as are also the brown Gymnothorax funebris and the spotted Gymnothorax moringa in the West Indies. These and many other species may coil themselves in crevices in the reefs, whence they strike out at their prey like snakes, taking perhaps the head of a duck or the finger of a man.

In many of the morays the jaws are so curved and the mouth so filled with knife-like teeth that the jaws cannot be closed. This fact, however, renders no assistance to their prey, as the teeth are adapted for holding as well as for cutting.

In Enchelynassa bleekeri, a huge wine-colored eel of the South Seas, the teeth are larger than in any other species. Evenchelys (macrurus) is remarkable for its extraordinary length of tail, Echidna for its blunt teeth, and Scuticaria, Uropterygius, and Channomuræna for the almost complete absence of fins. In Anarchias (allardicei; knighti), the anal fin is absent. The flesh of the morays is rather agreeable in taste, but usually oily and not readily digestible, less wholesome than that of the true eels.

The Myrocongridæ are small morays with developed pectoral fins. The species are few and little known.

Family Moringuidæ.—Structurally one of the most peculiar of the groups of eels is the small family of Moringuidæ of the East and West Indies. In these very slender, almost worm-like fishes the heart is placed very far behind the gills and the tail is very short. The fins are very little developed, and some forms, as Gordiichthys irretitus of the Gulf of Mexico, the body as slender as a whiplash, possess a very great number of vertebræ. Moringua hawaiiensis occurs in Hawaii, M. edwardsi in the Bahamas. This family probably belongs with the morays to the group of Colocephali, although its real relationships are not wholly certain.

Order Carencheli, the Long-necked Eels.—Certain offshoots from the Apodes so widely diverging in structure that they must apparently be considered as distinct orders occur sparingly in the deep seas. One of these, Derichthys serpentinus, the long-necked eel, constitutes the sole known species of the suborder Carencheli (καρά, head; ἔγχελυς, eel). In this group the premaxillaries and maxillaries are present as in ordinary fishes, but united by suture and soldered to the cranium. As in true eels, the shoulder-girdle is remote from the skull. The head is set on a snake-like neck. The single species representing the family Derichthyidæ was found in the abysmal depths of the Gulf Stream.

Order Lyomeri, or Gulpers.—Still more aberrent and in many respects extraordinary are the eels of the order or suborder Lyomeri (λυός, loose; μέρος, part), known as "Gulpers." These are degenerate forms, possibly degraded from some conger-like type, but characterized by an extreme looseness of structure unique among fishes. The gill-arches are reduced to five small bars of bone, not attached to the skull, the palatopterygoid arch is wholly wanting, the premaxillaries are wanting, as in all true eels, and the maxillaries loosely joined to the skull. The symplectic bone is wanting, and the lower jaw is so hinged to the skull that it swings freely in various directions. In place of the lateral line are singular appendages. Dr. Gill says of these fishes: "The entire organization is peculiar to the extent of anomaly, and our old conceptions of the characteristics of a fish require to be modified in the light of our knowledge of such strange beings." Special features are the extraordinary size of the mouth, which has a cavity larger than that of the rest of the body, the insertion of the very small eye at the tip of the snout, and the relative length of the tail. The whole substance is excessively fragile as usual with animals living in great depths and the color is jet black. Three species have been described, and these have been placed in two families, Saccopharyngidæ, with the trunk (gill-opening to the vent) much longer than the head, and Eurypharyngidæ, with the trunk very short, much shorter than the head. The best-known species is the pelican eel (Eurypharynx pelacanoides), of the coast of Morocco, described by Vaillant in 1882. Gastrostomus bairdi, very much like it, occurs in the great depths under the Gulf Stream. So fragile and so easily distorted are these fishes that it is possible that all three are really the same species, for which the oldest name would be Saccopharynx ampullaceus. Of this form four specimens have been taken in the Atlantic, one of them six feet long, carried to the surface through having swallowed fishes too large to be controlled. To be carried above its depth in a struggle with its prey is one of the greatest dangers to which the abysmal fishes are subject.

Order Heteromi.—The order of Heteromi (ἑτερός, different; ὤμος, shoulder), or spiny eels, may be here noticed for want of a better place, as its affinities are very uncertain. Some writers have regarded it as allied to the eels; some have placed it among the Ganoids. Others have found affinities with the sticklebacks, and still others with the singular fresh-water fishes called Mastacembelus. The Heteromi agree with the eels, as well as with Mastacembelus, in having the scapular arch separate from the cranium. Unlike all the true eels, most of the species have true dorsal and anal spines, as in the Percesoces and Hemibranchii. The ventral fins, when present, are abdominal and each with several spines in front, a character not found among the Acanthopteri. There is no mesocoracoid.

The air-bladder has a duct, and the coracoids, much as in the Xenomi, are reduced to a single lamellar imperforate plate. The two groups have little else in common, however, and this trait is possibly primitive in both cases, more likely to have arisen through independent degeneration. The separation of the shoulder-girdle doubtless indicates no affinity with the eels, as the bones of the jaws are quite normal. Two families are known, both from the deep sea, besides an extinct family in which spines are not developed.

The Notacanthidæ are elongate, compressed, ending in a band-shaped, tapering tail; the back has numerous free spines and few or no soft rays, and the mouth is normal, provided with teeth. The species of Notacanthus are few and scantily preserved. Those of Macdonaldia are more abundant. Macdonaldia challengeri is from the North Pacific, being once taken off Tokio. The extinct family of Protonotacanthidæ differs in the total absence of dorsal spines and fin-rays; the single species, Pronotocanthus sahel-almæ, originally described as a primitive eel, occurs in the Cretaceous of Mount Lebanon.

The Lipogenyidæ have a round, sucker-like mouth, with imperfect lower jaw, but are otherwise similar. Lipogenys gilli was dredged in the Gulf Stream.

Dr. Boulenger has recently extended the group of Heteromi by the addition of the Dercetidæ, Halosauridæ (Lyopomi), and the Fierasferidæ. We can hardly suppose that all these forms are really allied to Notacanthus.