Chapter XXIII (part 2 of 3)

Chapter XXIII (part 2 of 3)

On the other hand, the primitive dorsal vessel (dorsal aorta), retains not only its original position and relations, but also its primary function as the main channel for the distribution of arterialised blood to all parts of the body. The system of lateral and probably segmentally arranged vessels, by which the dorsal and subintestinal vessels were connected in the primitive Vertebrata, have undergone considerable modification in all existing Fishes, but nevertheless retain much of their original disposition and relations in the pharyngeal region of the alimentary canal, where they are represented by the afferent and efferent vessels of the gills.

A more detailed account of the condition of the vascular system in the Cyclostomata and Fishes will now be given.

THE VENOUS SYSTEM.—The Cyclostomata,[359] as might be expected, exhibit a more primitive condition of the venous system {316}in certain features than is the case in any other group. The precaudal portion of the subintestinal vein retains much of its original importance and runs in the rudimentary intestinal spiral valve as far as the liver, where it becomes the hepatic portal vein. From the liver the blood is collected into a single hepatic vein, and by it is conveyed to the sinus venosus. The caudal section of the subintestinal vein, now known as the caudal vein, bifurcates near the anus, and its two branches become directly continuous with the right and left posterior cardinals, without forming a renal portal system. In their forward course to the heart the posterior cardinals are situated directly beneath the notochord, and after receiving the blood from the kidneys and gonads, and from the numerous pairs of segmental veins of the body-wall, join the corresponding anterior cardinal veins, and form on each side a short transverse Cuvierian duct which opens into the sinus venosus. There is also a pair of inferior jugular veins which, however, unite opposite the fifth pair of gill-sacs to form a single trunk; this vessel is continued backwards, externally to the medio-ventral cartilage of the branchial basket, and finally opens directly into the sinus venosus.

In Elasmobranchs (e.g. Mustelus antarcticus)[360] the caudal vein (Fig. 186) lies in the haemal canal of the caudal portion of the vertebral column. On reaching the kidneys the vein divides into two renal portal veins, which, however, are not directly continuous with the posterior cardinal veins as in the Cyclostomata, but, on the contrary, after receiving the posterior segmental and oviducal veins, become continuous with the capillaries of the kidneys.

From the latter organs the blood is collected by a series of renal veins, and by them conveyed to the posterior cardinals, and thence to the Cuvierian ducts. In the adult, therefore, there is a well-developed renal portal system, but it is worthy of note, nevertheless, that this system is developed comparatively late in embryonic life, and that at an earlier stage the caudal vein is directly continuous with the two posterior cardinals, precisely as is the case in the Cyclostomata throughout life. The posterior cardinal veins are situated in the dorsal wall of the coelom (Fig. 187), beneath the vertebral column. For the hinder portion of {317}their extent they are embedded in the kidneys (Fig. 186); and in this region the two veins are in close relation in the median line, and here and there freely communicate with each other. More anteriorly, they enlarge so much that they present the appearance of cavernous sinuses. In addition to the anterior segmental and oviducal veins, the posterior cardinals receive the spermatic or the ovarian vein from the male or female gonad.

The precaudal section of the primitive subintestinal vein, now termed the internal intestinal vein (Figs. 186 and 187), traverses the spiral valve as it passes forwards to the liver, but from a physiological point of view is now merely one of the factors of the great hepatic portal vein, the principal tributaries of which are the veins from the stomach and intestine, including the rectal gland, and the pancreas and spleen. On entering the liver the hepatic portal vein divides into two principal branches for the right and left halves of the gland. From the liver the blood is conveyed by two hepatic veins to the sinus venosus.

The lateral veins (Fig. 186) are situated in the lateral walls {318}of the abdomen, immediately external to the peritoneum (Fig. 187). Each vein begins near the pelvic fin, where it is connected with its fellow across the dorsal face of the ischio-pubic cartilage, and thence runs forward towards the pectoral fin. At its origin the lateral vein receives a femoral vein from the pelvic fin and a cloacal vein, and also, near its anterior end, a brachial vein from the pectoral fin, finally joining the Cuvierian duct of its side.[361]

The anterior cardinal vein is situated directly above the gill-arches of its side of the head, and extends forwards from its junction behind with the Cuvierian duct to the outer side of the auditory capsule, where it communicates by a valvular orifice with a large sinus surrounding the eye-muscles (orbital sinus), and ventrally, by means of a similar aperture, with another large sinus, the hyoidean sinus, which lies on the outer face of the corresponding hyoid arch, and is continuous ventrally with its fellow of the opposite side. Into the orbital sinus open the anterior facial vein from the anterior and external regions of the head, and the anterior cerebral vein from the lateral half of the brain, and, into the hyoidean sinus, the nutrient veins from the hyoidean hemibranch.

The inferior jugular veins are situated beneath the branchial apparatus. Each vein begins anteriorly by communicating with {319}the hyoidean sinus of its side, and, after receiving the nutrient veins from the holobranchs of the first four branchial arches, opens into the corresponding Cuvierian duct.

The venous blood from the heart itself is collected into two coronary veins, which open into the sinus venosus.

In addition to the more important veins already described, there is also a series of median and lateral cutaneous veins communicating at different points with certain of the more deeply seated veins (Fig. 187).

Characteristic features in the venous system of Mustelus, as also of Elasmobranchs in general, are the development of transverse connexions between certain of the principal paired veins, and the tendency of many of the main veins to enlarge into more or less irregularly-shaped sinuses.

In its broad outlines the venous system of the Teleostomi agrees with that of Elasmobranchs, but is nevertheless characterised by several more or less important modifications, while at the same time exhibiting many differences in minor details.

A renal portal system is usually present, but is singularly variable in the source of its tributary veins, even in closely allied forms.[362] In the Sturgeon (Acipenser) and in some Teleosts, as in the Siluroid, Amiurus catus, it resembles that of Elasmobranchs. In other Teleosts, on the contrary, the renal portal system presents various grades of degeneration, or, possibly, of imperfect evolution, as will be seen from the following illustrations of its condition in different genera.

In Amiurus the caudal vein, after giving off right and left renal portal veins to the renal capillaries, emerges from the ventral surface of the kidneys, and is then continued forwards between the gonads, the veins from which it receives, as the radicle of the hepatic portal vein.

In the Eel (Anguilla vulgaris) the caudal vein (Fig. 188) traverses the fused hinder portions of the kidneys, receiving several segmental veins from the body-wall and also giving off from each side numerous renal portal branches. More anteriorly, where the two kidneys become distinct, the caudal vein also divides into two renal portal veins and, as each vein extends {320}forwards along the outer border of the kidney of its side, it receives a number of segmental veins, and, at the same time, gives off branches to the renal capillaries. In addition, each renal portal vein is connected with the hepatic portal vein by a series of singular arch-like vessels into which the ovarian or spermatic veins open.

It is obvious, therefore, that in both Amiurus and Anguilla the primitive direct continuity of the caudal and posterior cardinal veins has been interrupted by the formation of a well-developed renal portal system, and further, that the residue of the caudal venous blood finds its way to the liver through the hepatic portal vein; hence it follows that, as in so many of the lower air-breathing Vertebrates, the whole of the venous blood from the tail is distributed either to the kidneys or liver in the course of its return journey to the heart.

The Tench (Tinca vulgaris) exhibits the interesting anomaly of possessing two caudal veins, a dorsal and a ventral (Fig. 189). The dorsal vein is directly continuous with the right posterior cardinal, while the ventral one divides into three branches, two forming right and left renal portal veins and receiving numerous segmental veins, and the third becoming one of the affluents of the hepatic portal vein. In this Teleost it is clear that a portion of the caudal blood passes directly to the heart through the right posterior cardinal without traversing either the renal portal or hepatic portal system.

In the Cod (Gadus morrhua) the caudal vein divides into two branches. The larger right vein retains its direct continuity {321}with the corresponding posterior cardinal; the left, on the contrary, has ceased to be continuous with the greatly reduced left posterior cardinal and forms a renal portal vein, the distribution of which is, however, restricted to the hinder portion of the left kidney (Fig. 190). As in Amiurus, a branch of the caudal vein forms one of the tributaries of the hepatic portal vein. In the Cod it would therefore seem that only a relatively small proportion of the caudal blood flows through the imperfectly developed renal portal system, the bulk of it traversing the right posterior cardinal and passing directly to the heart, leaving, nevertheless, a modicum for transmission to the liver. Finally, it may be mentioned that in some Teleosts the caudal vein retains its embryonic continuity with one, usually the right, posterior cardinal, without giving off a renal portal affluent, as in the Perch (Perca fluviatilis); or, after division, with both posterior cardinals, as in the Lump-sucker (Cyclopterus lumpus). In such instances as these no portion of the caudal blood traverses the kidneys, and if a renal portal system exists at all, the only true renal portal veins are the adjacent segmental veins, which transmit venous blood directly to the kidneys, instead of first uniting with renal portal branches of the caudal vein as in the Tench and the Eel.

Whatever may be the condition of the renal portal system, all the renal blood is eventually collected by renal veins and conveyed to the posterior cardinals, which are often connected by one or {322}by several transverse anastomoses (Fig. 190). In the region of the heart each posterior cardinal joins the corresponding anterior cardinal to form a short but wide Cuvierian duct, which finally opens into the sinus venosus.

A subintestinal vein is present in the embryo (e.g. Lepidosteus, Acipenser, and some Teleosts),[363] but in the adult Teleostome its precaudal section is usually absorbed, or at all events ceases to be recognisable except, perhaps, as one of the minor tributaries of the hepatic portal vein.[364]

The hepatic portal vein is formed as in Elasmobranchs, but in different Teleostomi it may also receive the veins from the pyloric caeca, from a portion of the air-bladder, the gonads, and, as previously mentioned, a tributary from the caudal vein. There are usually two hepatic veins opening into the sinus venosus, and generally of equal size (Fig. 190).

Most of the veins from the air-bladder join the hepatic portal {323}vein, as already mentioned (Fig. 190), but more or fewer of them, especially those from the dorsal wall of the organ, open into the posterior cardinals. They may, as in Polypterus, even join the hepatic veins.[365]

The veins from the gonads are very variable in their destination, sometimes joining the posterior cardinals, as in the Salmon (Salmo salar); or the hepatic portal vein, as in Amiurus; or, as in the Perch (Perca fluviatilis), forming by their union a single trunk, which communicates directly with the left Cuvierian duct.

Representatives of the great lateral veins of Elasmobranchs appear to be absent in the Teleostomi, the veins from the pectoral and pelvic limbs joining the Cuvierian duct and the posterior cardinal veins respectively.

The two large anterior cardinal veins, which collect the blood from the head and brain, occupy their usual position directly above the branchial apparatus, and are sometimes connected by transverse anastomoses as they pass backwards to join the Cuvierian ducts. The inferior jugular vein is either single (e.g. Gadus); or paired, as in Perca (Fig. 190).

In the Dipnoi the venous system is distinguished by an interesting combination of characters, some of which are either primitive or peculiar to the group, while others exhibit a distinct transition to the embryonic or the adult condition of the lower air-breathing Vertebrates.

In Neoceratodus[366] (Fig. 191) the renal portal system is unusually complex, the veins distributing venous blood to the kidneys being derived from several sources, as follows: (1) from each of the two branches into which the caudal vein divides on its exit from the haemal canal (af.r.v); (2) from a common trunk (pt.v) which, on each side, is formed by the union of segmental veins from certain of the post-cloacal myotomes and is united with its fellow by a transverse anastomosis; (3) from more anteriorly situated intercostal or segmental veins (i.c.v) which enter each kidney directly; and (4) from a vein on each side corresponding to the renal portal vein of Amphibia. The latter vein (rp.v) is formed by one of the two branches of the iliac or femoral vein, and joins {324}the corresponding vein from the caudal myotomes; from the common trunk numerous branches enter the kidney.

In the derivation of renal portal veins from each of the two veins into which the caudal vein divides, Neoceratodus approaches the Elasmobranchs. On the other hand, the utilisation of ordinary segmental veins from the caudal and pre-caudal myotomes, some of which directly enter the kidney, is a feature which has already been remarked in some Teleosts; while the formation of a renal portal affluent by a branch of the femoral vein is an even more striking Amphibian characteristic.

The efferent renal veins[367] join the root of the left posterior cardinal and the adjacent portion of the caudal vein.

Of the two great venous trunks into which the caudal vein divides, the right is much the larger and behaves somewhat differently to the left. The former (i.v.c) passes forwards in relation with the right kidney, receiving in its course the spermatic or ovarian veins from the gonad of its side, and then traverses the liver, finally opening into the median portion of the sinus venosus, between the orifices of the two hepatic veins. The left branch of the caudal vein (l.p.c) also passes forwards in relation with the left {325}kidney and receives veins from the corresponding gonad; but, instead of traversing the liver, it passes above that organ, and finally opens into the left Cuvierian duct. The course of the left vein, and the relations of the vessel to the caudal vein and the left Cuvierian duct, point to the conclusion that it represents the left posterior cardinal of other Fishes. From its continuity with the caudal vein it is also obvious that the hinder or renal portion of the right trunk is a remnant of the right posterior cardinal; but the more anterior section so closely resembles the postcaval vein, or inferior vena cava of the higher Vertebrates, in its relations to the liver, the hepatic veins, and the sinus venosus, that its identity as such seems beyond doubt, and this interpretation is supported by well-known observations[368] on the mode of origin of the inferior vena cava in Amphibia, and especially the union of the independently formed inferior vena cava with the posterior or inter-renal portion of the embryonic right posterior cardinal vein, combined with the atrophy of the anterior portion of the latter vein.[369] The singular connexions and relations of these two great veins afford an additional illustration of the significant transitional condition of the venous system in the Dipnoi. On the other hand, the direct continuity of the caudal vein with vessels which, wholly or in part, represent the two posterior cardinals, is a feature alike characteristic of the adult Cyclostome and the embryonic Elasmobranch, Teleost, and Amphibian.

As in the Cyclostomes and Elasmobranchs, the precaudal section of the embryonic subintestinal vein is represented in the adult by an intra-intestinal vein which traverses the spiral valve near its free edge and is a tributary of the hepatic portal vein.

The two veins from the undivided air-bladder unite to form a single vessel, which, instead of joining the hepatic portal or posterior cardinal veins as in other Fishes, opens into the left auricle, like the pulmonary veins of the Amphibia.

A further resemblance to the Amphibia is to be found in the presence of an anterior abdominal vein. After leaving the pelvic {326}limb each femoral vein divides into two branches; one of these forms a renal portal vein as previously described; the other, which may rightly be termed a pelvic vein (pv.v), unites with its fellow to form a median anterior abdominal vein (a.ab). Pursuing its course forwards in the ventral abdominal wall, the vein eventually reaches the heart and opens into the sinus venosus. The direct connexion of the anterior abdominal vein with the heart is yet another example of the retention in the adult Neoceratodus of a transitory embryonic feature in the developing Amphibian.[370]

As in other Fishes, the blood from the head is conveyed to the Cuvierian ducts by an anterior cardinal and an inferior jugular on each side. There are no lateral veins, the blood from the pelvic fins flowing into the renal portal system or into the anterior abdominal vein, and that from the pectoral fin through subscapular and brachial veins into the Cuvierian ducts. Lateral cutaneous veins are, however, present; and, as in Elasmobranchs (e.g. Mustelus antarcticus), anastomose anteriorly with the subscapular vein and behind with the caudal vein.

{327}Less is known of the venous system of Protopterus,[371] but it is certain, nevertheless, that it presents a more advanced grade of evolution than in Neoceratodus, and, except for the doubt as to the existence of an anterior abdominal vein, it is essentially similar to that of a Urodele Amphibian in which the right posterior cardinal vein has aborted.

The caudal vein (Fig. 192) divides into right and left renal portal branches, neither of which, however, is directly continuous with the inferior vena cava or the left posterior cardinal; on the contrary, each renal portal vein is joined by the corresponding iliac or femoral vein, and also by numerous segmental veins, and then distributes the whole of its venous blood to the kidney. The radicles of the inferior vena cava and the left posterior cardinal are formed by the renal veins from the two kidneys, and in their forward course to the heart both veins receive in addition genital and segmental veins. In its course through the liver the inferior vena cava receives several hepatic veins, and finally opens into the sinus venosus, while the left posterior cardinal vein joins the corresponding Cuvierian duct, which also receives anterior cardinal, inferior jugular, and subclavian veins. There is an intra-intestinal vein as in Neoceratodus, but an anterior abdominal vein has yet to be discovered. The two pulmonary veins from the double air-bladder form a single trunk before communicating with the left auricle.

With the exception of certain doubtful details which need further investigation, the venous system of Lepidosiren[372] seems to resemble that of Protopterus.

THE HEART.—The heart is more anteriorly placed than in other Vertebrates, being situated directly behind and beneath the last pair of branchial clefts and internal to the ventral portion of the pectoral girdle. The organ is enclosed in a pericardial cavity, which, in the adult, is separated from the abdominal portion of the coelom by a transverse pericardio-peritoneal septum, and in the Lamprey (Petromyzon) is partially enclosed within a cartilaginous, cup-like modification of the hinder part of the branchial basket. In the Ammocoetes-stage of the Lamprey the pericardium is in communication behind with the general coelom, but the connexion is lost in the adult. In Elasmobranchs the {328}two cavities are connected by a single pericardio-peritoneal canal, or by two such canals; and in Chimaera, and in the Sturgeon (Acipenser) and Polyodon, by a single canal.

The heart consists of at least three chambers, a sinus venosus which receives the venous blood from the body, an auricle and a ventricle, to which is added a conus arteriosus in the Elasmobranchs, certain Teleostomi (Crossopterygii, Chondrostei, and Holostei), and in the Dipnoi. Through these cardiac chambers the blood is forced in the order mentioned. In the Dipnoi the auricle is subdivided by a more or less complete interauricular septum into a right and left auricle,[373] the former receiving the venous blood from the sinus venosus, and the latter the aerated blood from the lung-like air-bladder.

The sinus venosus and the auricle have very thin walls; the ventricular walls, on the contrary, are very thick and in great measure are composed of a sponge-like network of muscular bundles which generally encroaches considerably on the ventricular cavity. Membranous valves, the sinu-auricular, and the auriculo-ventricular valves, are developed at the junctions of the sinus venosus with the auricle, and the auricle with the ventricle respectively. The conus arteriosus is muscular and contractile, and is interposed between the ventricle and the root of the ventral aorta. Internally, the conus is provided with several transverse rows of pocket-shaped or semilunar valves. In Teleosts the conus is non-muscular and vestigial, and has but a single row of valves, corresponding to the most anterior of the multiple rows of valves in the Elasmobranchs. In these {329}Fishes the vestigial conus is succeeded by a non-contractile, bulb-like dilatation, or bulbus aortae, of the root of the ventral aorta. In only a single Teleost, viz. Albula, one of the Albulidae, is the vestigial conus muscular, and at the same time provided with two rows of valves.[374] In the Cyclostomata there is a bulbus with a single row of two valves, but no true conus.

In the Dipnoi (e.g. Protopterus) the heart, like the rest of the vascular system, exhibits certain interesting resemblances to the Amphibian heart. In addition to a more or less complete interauricular septum separating right and left auricles, there is a median longitudinal ridge, partly muscular and partly fibrous, which incompletely subdivides the cavity of the ventricle. The spirally-twisted conus arteriosus is furnished with several transverse rows of valves, certain of which coalesce longitudinally to form a complete septum dividing the cavity of the conus into two distinct lateral channels: with this septum there coalesces another septum, which cuts off the origins of the anterior two pairs from the remaining afferent branchial arteries. The formation of these septa has the physiological effect of subdividing the series of cardiac cavities into two parallel channels, of which one has its origin behind in the sinus venosus and transmits venous blood to the posterior afferent branchial vessels; while the other, commencing with the left auricle, conveys arterial blood to the first two pairs of afferent branchial arteries.[375] In Neoceratodus, however, the longitudinal septum in the conus is incomplete, and hence the blood which is sent to the anterior afferent vessels is mixed.[376]

THE ARTERIAL SYSTEM.—The ventral aorta is a median artery situated beneath the floor of the pharynx, and having its origin, behind, either directly from the ventricle or from the conus arteriosus.

In the Cyclostomata[377] (e.g. Petromyzon) the ventral aorta (Fig. 194) is continued forwards from the heart as a single vessel to the fourth pair of gill-sacs, where it divides into right and left branches which extend as far as the anterior walls of the first pair of gill-sacs. Eight pairs of afferent branchial arteries arise from the ventral aorta and its two branches, of {330}which the first and last supply the anterior walls of the first pair of sacs and the posterior walls of the last pair respectively. Each of the remaining afferent vessels extends into an interbranchial septum, and supplies the gill-lamellae of the posterior wall of one sac and those of the anterior wall of the next sac behind. The corresponding efferent branchial vessels have a similar distribution, and unite dorsally to form a median dorsal aorta. Beneath the base of the skull the latter vessel divides into two branches which, after receiving the first pair of efferent branchial vessels, pursue a divergent course forwards, but subsequently converge and unite to form a "circulus cephalicus," as in Teleosts. From the cephalic circle are given off on each side (1) an "internal carotid" artery for the brain and eye; (2) an "external carotid" for the lateral and ventral walls of the head; and (3) a large ventral branch which supplies the lingual apparatus; while from the abdominal portion of the dorsal aorta are derived, first, a coeliaco-mesenteric artery for the liver and alimentary canal, and subsequently branches for the myotomes, kidneys, and the gonad. The terminal portion of the aorta then enters the tail and forms the caudal artery.

In Elasmobranchs[378] (e.g. Mustelus antarcticus) the undivided ventral aorta gives off five pairs of afferent branchial arteries which, on each side, ascend in succession the outer convex sides of the hyoid and first four branchial arches (Fig. 195).

{331}[Illustration: FIG. 195.—The branchial arterial system of Mustelus antarcticus. Left lateral view. The ventral aorta and afferent branchial vessels are in solid black, the efferent arteries and their branches have double contours. The branchial clefts have fringed borders to indicate their hemibranchs, and the arches are in simple outline. a.c.a, Anterior carotid; a.d.a, anterior dorsal aorta; af.b.a, afferent branchial artery; br.a, brachial artery; c.m.a, coeliaco-mesenteric; d.a, dorsal aorta; E, eye; ep.a, epibranchial artery; H, heart; h.b.a, hypobranchial artery; hy.a, afferent pseudobranchial or hyoidean artery; md.a, mandibular artery; op.a, ophthalmic artery; p.c.a, posterior carotid; sb.a, subclavian; sp, spiracle; v.a, ventral aorta; 1-5, the hyobranchial and four succeeding branchial clefts. The hypobranchial artery is seen immediately beneath the ventral aorta. (After T. Jeffery Parker, diagrammatic.)]

The first or most anterior of these arteries supplies the hyoidean hemibranch, while the succeeding four supply the holobranchs of the four branchial arches. The blood is collected from the capillaries of the branchial lamellae by a series of efferent branchial vessels, a pair for the two hemibranchs of each branchial arch and a single vessel for the hyoidean hemibranch, which unite with one another in a somewhat singular fashion. The efferent arteries from the anterior and posterior hemibranchs of each branchial cleft unite above and below each cleft in such a way as to form a series of complete vascular loops round the hyoidean cleft and the three succeeding branchial clefts, which are connected by short longitudinal trunks in each arch and also by a longitudinal commissural vessel between their ventral extremities. As the fifth arch is gill-less, there is no complete loop round the fifth cleft, the blood collected by the efferent vessel of the posterior hemibranch of the fourth arch being conveyed to the corresponding vessel of the anterior hemibranch of the same arch by one of the short longitudinal vessels above mentioned. Dorsally, each arterial loop is continuous with an epibranchial artery; and by the dorsal union of the four {332}epibranchial arteries of the two sides the dorsal aorta is formed. It may be pointed out that the anterior efferent vessel of each arch, which is usually larger than the posterior one, is to be regarded as the primary efferent artery of the corresponding holobranch, and as such is directly continuous with an epibranchial artery, the posterior efferent artery being a secondary vessel which opens not into the primary trunk of its own branchial arch, but into that of the succeeding arch.[379] The principal arteries which supply the various parts of the head with blood are derived from the first efferent branchial vessel. From the ventral end of this artery a mandibular artery is given off, which subdivides into branches for the muscles of the lower jaw as well as into nutrient vessels for the hyoidean hemibranch. At about the middle of its length the same artery gives off an afferent pseudobranchial or hyoidean artery, to the spiracular or mandibular pseudobranch. From the latter organ the blood is collected by an anterior carotid artery which, after giving off an ophthalmic branch to the eye, perforates the orbital wall and enters the cranial cavity, where it is joined by an anastomotic trunk from the posterior carotid of the opposite side; finally, the anterior carotid divides into anterior and posterior cerebral arteries for the brain. The third and last of the cephalic arteries is the posterior carotid; this artery arises from the dorsal extremity of the first efferent branchial vessel, and, on entering the orbit, gives off the anastomotic trunk previously mentioned. The latter vessel enters the cranial cavity, and, after crossing its fellow, joins the anterior carotid of the opposite side, as described above. The main trunk is then continued forwards in the orbit, and its various branches eventually supply the eye-muscles, the mandibular adductor muscle, and some other parts of the head.

It is worthy of note that the median dorsal aorta is prolonged forwards in front of the first pair of epibranchial arteries as a slender median vessel (a.d.a), which ultimately divides into two branches, each branch uniting with the posterior carotid of its side.

A remarkable system of arteries for the supply of nutrient blood to the gills and heart has its origin in the following {333}manner. On each side, the longitudinal commissural vessel, which connects the ventral ends of the arterial loops surrounding the different gill-clefts, gives origin to a series of pairs of short transverse vessels, and by their union these combine to form a median longitudinal hypobranchial artery which lies beneath the ventral aorta. From the hypobranchial artery are derived the coronary arteries for the heart; and from the same artery, or from its lateral connexions with the longitudinal commissural artery, and, in the case of the hyoidean hemibranch, from the mandibular artery, are derived the various nutrient vessels for the gills.

The arteries for the trunk, and for the pectoral and pelvic limbs, arise in succession from the dorsal aorta. The first of the series is the subclavian artery, which has its origin from the aorta close to the dorsal extremities of the fourth pair of epibranchial arteries. Each subclavian artery gives off a brachial artery to the pectoral fin, and is then continued forwards as a lateral hypobranchial artery, which, with its fellow of the opposite side, eventually becomes continuous with the hinder end of the median hypobranchial artery. Behind the subclavian artery there is a median coeliaco-mesenteric artery, the various branches of which are distributed to the liver, stomach, and intestine. A lieno-gastric artery supplies the pancreas and spleen, and also sends branches to the stomach. In addition, there are also arteries for the gonads, numerous segmental arteries for the myotomes, and renal arteries for the kidneys. Finally, the aorta gives off a pair of iliac arteries for the pelvic fins, and then enters the haemal canal as the caudal artery.

The more important differences in the arterial system of the Holocephali and the Teleostomi relate to (1) the absence of the posterior efferent branchial artery in each branchial arch; (2) modifications dependent on the condition of the spiracular and hyoidean hemibranchs, and the mode of origin and the course of their afferent and efferent vessels; and (3) the source from whence the air-bladder derives its blood when that organ is present.

(1) The branchial arterial system is somewhat more primitive than in the generality of Elasmobranchs.[380] There are no complete vascular loops round the gill-clefts, and the blood from the two {334}hemibranchs of each branchial arch is conveyed to the dorsal aorta by a single efferent vessel which corresponds to the more anterior of the two in Mustelus antarcticus.[381]

(2) In Callorhynchus[382] among the Holocephali, where the spiracle is absent but the hyoidean hemibranchi is still a true gill, the latter organ is supplied with venous blood by a branch from the ventral aorta, the corresponding efferent vessel joining the dorsal aorta (Fig. 196). In the absence of a spiracular pseudobranch the anterior carotid may be regarded as continuous with the hyoidean artery,[383] and as having its origin directly from the efferent artery of the hyoidean hemibranch (Fig. 196). At its origin the anterior carotid anastomoses with the mandibular artery.

The Sturgeon more closely resembles the Elasmobranchs. The hyoidean gill is supplied by an afferent branchial artery from the ventral aorta, and its efferent vessel joins the corresponding trunk from the holobranch of the first branchial arch. A hyoidean artery supplies the spiracular pseudobranch, the efferent vessel of which contributes to the blood-supply of the brain and the eye, and probably represents an anterior carotid.

Lepidosteus[384] offers a singularly interesting transition from the {335}Elasmobranch to the Teleost. As indicated in the preceding chapter, this Fish possesses both a hyoidean gill and a spiracular pseudobranch (Figs. 197 and 198). The hyoidean gill is supplied by an afferent artery direct from the ventral aorta, but the proper efferent vessel of the gill, which primitively joined the dorsal aorta, is suppressed, and the blood is collected into a vessel, which, like the hyoidean artery in Elasmobranchs, becomes the afferent artery of the spiracular pseudobranch. The latter artery unites, however, with a second hyoidean artery derived from the efferent branchial vessel of the first branchial arch, and represents the artery termed "hyoidean" in Teleosts. The efferent vessel from the spiracular pseudobranch joins an internal branch from the carotid artery, and then distributes its blood both to the eye and the brain.

In Teleosts, as already mentioned in a preceding chapter, it is probable that the hyoidean hemibranch is suppressed, the so-called hyoidean pseudobranch being a spiracular pseudobranch. The latter is now supplied by a "hyoidean" artery, which has its origin from the ventral end of the efferent {336}branchial artery of the first branchial arch, the corresponding efferent trunk forming an ophthalmic artery, and passing to the choroid gland of the eye (Fig. 199). Both the proper afferent and efferent arteries of the hyoidean hemibranch either disappear or, as in the Cod (Gadus morrhua), the efferent artery may be represented on each side by an anastomosis between the hyoidean artery and the cephalic circle. Hence, the "hyoidean" artery of Teleosts corresponds to the one which has a similar origin in Lepidosteus.

A brief description of the remaining efferent branchial arteries and their derivatives in the Cod (Gadus morrhua) will illustrate the condition of these structures in a well-known Teleost.

In this Fish the efferent branchial vessels open dorsally into right and left suprabranchial arteries,[385] which unite behind to form a median dorsal aorta (Fig. 199). Anteriorly, the paired suprabranchial arteries extend towards the base of the skull as the so-called "carotid" arteries. The two carotids enter the cranial cavity, and there unite in the median line, as in the Cyclostomes. By the union of these arteries in front, and of the {337}right and left suprabranchial arteries behind, the characteristic "circulus cephalicus" of Teleosts is completed.[386] From the anterior part of the cephalic circle are derived two internal carotid arteries[387] for the brain, and also a pair of orbito-nasal arteries for the eye-muscles and the nasal sacs, while more posteriorly an external carotid has its origin from each suprabranchial artery.

(3) In most Teleostomi the air-bladder is supplied with blood by branches of the coeliac artery, with the addition of small branches arising directly from the dorsal aorta. Polypterus and Amia[388] are, however, exceptional, inasmuch as the arteries for the air-bladder are derived from the last or fourth pair of efferent branchial vessels, and in this respect, but not in the destination {338}of the corresponding veins, the two genera exhibit a significant resemblance to the Dipnoi.

In the Dipnoi the ventral aorta is so short that the afferent branchial arteries arise almost directly from the conus arteriosus with their roots in close contiguity to one another (Fig. 200).

In Neoceratodus (Fig. 200),[389] there are two efferent vessels to each gill-bearing branchial arch, which unite above to form an epibranchial artery, and by the successive union of the four epibranchial arteries a short common trunk is formed on each side. Posteriorly, the two trunks unite to form a median dorsal aorta. Immediately above the gill-clefts each efferent vessel gives off a branch which, passing either forwards or backwards, unites with the corresponding branch of the efferent vessel in front or behind as the case may be. A hyoidean artery arises from the ventral extremity of the anterior efferent artery of the first branchial arch, and, after giving off a lingual artery, ascends the hyoid arch and supplies the hyoidean pseudobranch. The efferent vessel of the pseudobranch (a.c.a) or anterior {339}carotid artery, eventually enters the cranial cavity and subdivides into anterior and posterior cerebral arteries for the brain, also giving off a branch which unites with its fellow of the opposite side directly behind the infundibulum. A posterior carotid springs from the epibranchial of the first branchial arch and divides into palatine, orbital, and ocular branches; and from the ventral end of the anterior efferent vessel of the second branchial arch is derived a hypobranchial artery for the heart and pericardium. The pulmonary arteries for the lung-like air-bladder have their origin from the fourth pair of epibranchial arteries.

As in so many other details of its anatomy, Neoceratodus exhibits in its arterial system abundant evidence of the wide-spreading affinities of the group to which it belongs. In its branchial arterial system Neoceratodus presents a singular combination of features which, individually, are characteristic of Amphibia and Elasmobranchs. Special Amphibian features may be noted in the origin of the afferent branchial arteries almost simultaneously from the anterior end of the conus arteriosus; in the mode of union of the epibranchial arteries to form the dorsal aortae; in the origin of a lingual artery from the efferent vessel of the first branchial arch; and in the derivation on either side of a pulmonary artery from the fourth epibranchial artery. Agreement with Elasmobranchs is to be found in the presence of two efferent branchial vessels in each branchial arch, although the relations of these arteries are more primitive than in most adult Elasmobranchs, inasmuch as the two efferent vessels of the same arch unite to form an epibranchial artery; and also in the origin and distribution of the anterior and posterior carotids. Lastly may be mentioned the fact that Neoceratodus agrees not only with the Amphibia but also with those generalised Teleostomi, Polypterus and Amia, in the mode of origin of the great arteries for the air-bladder.

Of the two remaining Dipnoi, the arterial system of Protopterus[390] is better known than that of Lepidosiren, but in both cases further research is needed before a satisfactory comparison can be made with Neoceratodus and other Vertebrates. It is evident, nevertheless, that both genera differ from Neoceratodus in approximating more closely to the Amphibia than to the {340}lower Fishes, in so far as the branchial part of the arterial system is concerned.

In their origin from the conus the four afferent branchial arteries of Protopterus resemble those of Neoceratodus, but their relations to the branchial clefts are somewhat different (Fig. 201). The first or hyoidean cleft is closed, and the first afferent vessel lies between the second cleft and the third, and is therefore in relation with the second branchial arch. The remaining afferent arteries are disposed between the succeeding clefts and are related to the corresponding arches. As the second and third arches, like the vestigial first arch, bear no gill-lamellae, their afferent arteries are directly continuous with the corresponding efferent vessels, as in those Teleosts in which certain arches are gill-less, as well as in the Tadpole-stage of the tailless Amphibia when the internal gills begin to degenerate; and they apparently transmit arterial blood directly to the dorsal aorta.[391] The third and fourth afferent arteries, on the contrary, supply venous blood to the two hemibranchs which are borne by each of the two corresponding arches, viz.: the fourth and fifth, and from each pair of hemibranchs the blood is collected into two efferent vessels which unite dorsally {341}to form an epibranchial artery. From the dorsal end of the fourth afferent artery there arises a recurrent branch which curves round the upper margin of the sixth cleft and supplies the gill-lamellae on the posterior margin of that cleft, a fact which lends support to the view that these lamellae are "emigrants" from the anterior margin of the cleft; the efferent vessel from the "emigrant" lamellae joins the fourth epibranchial artery. The blood-supply of the external or cutaneous gills is derived from the dorsal extremities of the second, third, and fourth afferent arteries, while the efferent vessels from these organs join the corresponding epibranchial arteries; in this respect there is a close resemblance between Protopterus and those larval Amphibians which possess similar cutaneous gills. All four epibranchial arteries unite together at about the same point to form a short common trunk, the right or left dorsal aorta, which subsequently unites with its fellow to form the median dorsal aorta.

There is a so-called "hyoidean" artery, which, however, has its origin, not from an anterior efferent branchial vessel as in Neoceratodus, but from the first afferent branchial artery. After giving off a submaxillary or lingual artery, the "hyoidean" artery (af) becomes the afferent vessel for the "opercular gill" or "hyoidean pseudobranch,"[392] and supplies the latter with arterial blood. The efferent vessel (ef) from the pseudobranch unites with the four epibranchial arteries in forming the right or left dorsal aorta. A "carotid" artery arises from the efferent vessel of the "hyoidean pseudobranch," and a pulmonary artery has its origin from the root of the dorsal aorta of its side, and not from the fourth epibranchial artery as in Neoceratodus.

THE BLOOD.—The blood consists of a nutritive fluid plasma in which float red corpuscles and leucocytes. In the Cyclostomata (e.g. Petromyzon) the red corpuscles are circular, but in Myxine they have the usual oval shape. In Fishes the red corpuscles are almost invariably flat, oval, biconvex, and nucleated, and owe their colour to the presence of the characteristic oxygen-absorbing, iron-containing pigment, haemoglobin. They are unusually large in the Dipnoi and are only exceeded in size by those of certain Urodele Amphibians. The leucocytes are much less numerous than the red corpuscles, although their relative proportions are very variable, even in the same species under different {342}conditions. They appear to be more numerous in the Dipnoi (e.g. Protopterus) than in any other Vertebrates, except under pathological conditions.[393]

THE LYMPHATIC SYSTEM.—In addition to blood-vessels, Fishes possess a lymphatic system, consisting of smaller vessels, lymph-capillaries or lymph-spaces, distributed in the connective tissue of different parts of the body, and by their union ultimately forming larger lymph-vessels or sinuses which communicate with certain of the principal veins, the whole forming a series of channels for the collection of the blood-plasma which has exuded from the blood-capillaries for the nutrition of the tissues, and for its conveyance to the general venous system. The fluid in the lymphatics, or lymph, consists of dilute blood-plasma containing leucocytes but devoid of red corpuscles. At the points where the larger lymphatics open into the veins, lymph-hearts may be developed. In the Eel (Anguilla vulgaris) there is a lymph-heart in the tail, which communicates by a valvular orifice with the smaller of the two caudal veins, and by its rhythmical pulsations propels the lymph into the vein. In Silurus there are two caudal lymph-hearts. Apart from the lymphoid tissue, which is so abundantly present in certain parts of the body, Fishes appear to be devoid of the special "lymphatic glands" of the higher Vertebrates.

THE DUCTLESS OR BLOOD-GLANDS.—All the important blood-glands of other Vertebrates have their representatives in Fishes. Nothing is certainly known of the function of these organs in Fishes, but from the general structural resemblance which they present to their equivalents in the higher Vertebrates, it is perhaps not unreasonable to infer that they are similar in function. If this be so, the blood-glands of Fishes are organs for leucocyte-formation and phagocytosis, involving the destruction and removal of effete red blood-corpuscles; in addition, they may also be concerned with certain obscure chemical changes in the composition of the blood, which have an important relation to general or local nutrition.

THE SPLEEN.—This lymphoid organ is the largest of all the blood-glands, and, in the form of a compact or more or less lobulated body, is present in all Fishes, and possibly in Cyclostomes. In position the spleen is usually in close proximity to the stomach, to which it is attached by an extension round it {343}of the peritoneal investment of that organ. Thus, in the Dog-Fish (Scyllium), the spleen is a large reddish body attached to the convexity of the U-shaped stomach, and, in addition, sends a long narrow lobe between the distal limb and the valvate portion of the intestine (Fig. 153, spl). In the Sturgeon (Acipenser), the organ is also large, but is attached to the left side of the commencement of the intestine. In the Cod (Gadus) among Teleosts the spleen is much elongated and is situated on the dorsal side of the stomach. In the Dipnoi (e.g. Protopterus)[394] the organ is probably represented by a large compact lymphoid mass, closely connected with the dorsal and lateral walls of the stomach (Fig. 154, A, s).

THE THYROID GLAND.—This organ[395] usually arises in the form of a small median evagination of the hypoblastic epithelium of the ventral wall of the pharynx, in the region of the second visceral arch. Later it becomes detached from the place of origin and converted into a solid spherical body. Eventually the component cells form the limiting epithelium of a series of follicles or vesicles embedded in a matrix of connective tissue and blood-vessels, and the characteristic adult structure is attained.

Among the Cyclostomata the evagination is relatively large in the young Lamprey (Petromyzon fluviatilis), as also is the orifice of communication with the pharynx (Fig. 202, th).[396] The aperture soon becomes reduced to a mere pore, and finally disappears. During the larval or Ammocoetes-stage the organ consists of a median ciliated portion, communicating with a pair of laterally placed glandular sacs, but in the adult it is much smaller, and acquires the usual follicular structure. In adult Elasmobranchs the thyroid is represented by a moderately large compact organ, situated near the anterior end of the ventral aorta. In Teleostomi the organ may be paired, or, as in the Perch (Perca), more diffuse, consisting of masses of reddish lobules lying beneath the aorta, and also scattered for a variable distance along the course of the afferent branchial arteries.

In the Dipnoi (e.g. Protopterus)[397] the thyroid is small, {344}consisting of two lateral lobes connected by a constricted median portion, and situated beneath the epithelium of the tongue, immediately above the hyoidean symphysis. A similar structure has been described by Bischoff[398] in Lepidosiren, and was regarded by him as a salivary gland.

As in Reptiles, Birds, and Mammals, paired or accessory thyroid bodies ("supra-pericardial organs")[399] are present in many Fishes, and appear to be similar in structure to the median thyroid. In Elasmobranchs these bodies originate as a pair of outgrowths from the epithelium of the pharynx behind the last pair of branchial arches (Fig. 202, B, a.th). Subsequently they become detached from the pharynx, and in the adult are situated on the dorsal side of the pericardium, remote from the median thyroid.

According to Dohrn the median thyroid is to be regarded as the vestige of a gill-cleft which primitively existed between the hyomandibular cartilage and the hyoidean arch. This conclusion seems, however, to be less in harmony with the facts of development than the view[400] that the organ is derived from the characteristic hypobranchial groove or "endostyle" of Ascidians {345}and Amphioxus, which has undergone a change of function from a mucus-conveying groove to a blood-gland. On the other hand, the mode of origin of the paired thyroids certainly favours the suggestion that they represent a posterior pair of vestigial gill-clefts, a view which derives some support from the fact that in Notidanus, where additional branchial arches and clefts are present, the paired thyroids are absent.

THE THYMUS.—In the embryo Elasmobranch and Teleost[401] the thymus has a multiple origin, being derived from a series of distinct epithelial thickenings, one of which is developed at the dorsal extremity of each of the gill-clefts except of the spiracle. These rudiments subsequently detach themselves from the epithelial surface and sink inwards, eventually fusing together on each side to form a single independent structure. Later, the epithelial mass thus formed becomes invaded by connective tissue, and by leucocytes which form lymph follicles, and the thymus gradually assumes the structure of a lymphoid organ. From its mode of development it has been suggested that the thymus owes its evolution to the metamorphosis and ingrowth of branchial filaments,[402] but it is also noteworthy that each embryonic rudiment of the organ closely resembles, both in position and origin, one of the developing branchial tongue-bars of Amphioxus.[403] The abundance of leucocytes which it contains has also prompted the further suggestion that the origin of the thymus may be due to the necessity of providing for the phagocytic protection of the gills themselves from the ravages of harmful micro-organisms, fungoid spores, etc., as well as to aid in the removal of such portions of the gills as may have been injured.[404]

A thymus is probably present in all Fishes, if not in the adult at all events in the embryo, but is always relatively small in size. In Elasmobranchs the organ lies on each side above the branchial arches and beneath the dorsal musculature; and in Teleostomi at the dorsal extremity of the last branchial arch, in close proximity to the mucous membrane of the branchial cavity. In a similar position in the Dipnoi (e.g. Protopterus)[405] there are, on each side, {346}two contiguous lobes of lymphoid tissue which apparently represent a thymus.

THE SUPRA-RENAL BODIES.—The supra-renal bodies are organs of problematic function, which are present in the Cyclostomata, and probably in all Fishes, and situated in close proximity to the kidneys.

In the Cyclostomata (Petromyzon) these bodies are represented by lobules of cells along the posterior cardinal veins, and also by masses of peculiar cells ("chromaffin cells") along the sides of the aorta and segmental arteries.[406] In Elasmobranchs there are two distinct structures, the paired supra-renals and the inter-renals (Fig. 203, A). The former are a series of pairs of segmentally arranged bodies, situated on the successive pairs of segmental arteries given off from the dorsal aorta. The two bodies which form the first pair are much larger than any of the others, and were formerly spoken of as "axillary hearts." The inter-renal is usually a thin elongated "ochre-yellow" body, from which one or two lobes may be detached in front, and extends for a variable distance in {347}the median line between the two kidneys, or is unsymmetrically placed on the ventral surface of either kidney.[407] Sometimes (e.g. in Raia) the inter-renals are paired, in which case they are applied to the inner and hinder margins of the kidneys. In the Sturgeon (Acipenser sturio) the "supra-renals" appear as numerous "ochre-yellow" bodies, variable in size and distribution (Fig. 203, B). Some of them are visible on the surface of the kidneys, while others are scattered about in their substance, but on the whole are more anteriorly placed than in Teleosts. In the latter group the "supra-renals" are usually two in number (Fig. 203, C), but may be as many as five or reduced to one. They are disposed either on the ventral or the dorsal surface of the kidneys, generally near their hinder extremities, or more or less deeply embedded in their substance. Besides these bodies there are also chromaffin cells in the walls of the anterior cardinal veins.[408]

Histologically, the paired segmentally arranged bodies of Elasmobranchs differ considerably in structure from the inter-renal bodies, the former resembling the "medulla," while the inter-renals, as well as the so-called supra-renals of Acipenser, exhibit a striking resemblance to the alveolar "cortical" substance of the Mammalian supra-renals.[409] In Cyclostomes the cortex is apparently represented by the lobules of cells along the posterior cardinal veins and the medulla by the "chromaffin" cells, while in Teleosts the cortex and the medulla have their respective counterparts in the supra-renals and the "chromaffin" cells in the walls of the anterior cardinal veins. It may be concluded, therefore, that Elasmobranchs, Cyclostomes, and Teleosts possess anatomically distinct representatives of both the "medulla" and "cortex" of Mammalia, although the Sturgeon is at present only known to possess the equivalent of the "cortex." In Amphibia, Reptilia, and Aves both "cortex" and "medulla" are present, and in the varying intimacy of their relations offer a transition to the Mammalian arrangement of a central medulla closely invested by a sheath of cortical substance. A more or less intimate connexion exists between the paired supra-renals of Elasmobranchs and the sympathetic nervous {348}system. The former are usually well supplied with sympathetic nerve fibres, and contain ganglion-cells in their substance.

The primitive origin of these organs is very obscure, and as regards their development there is much diversity of opinion. It seems certain, however, that the cortex and medulla of the higher Vertebrates, including their equivalents in the Elasmobranchs, have independent origins, and the balance of opinion seems to point to the derivation of the cortex from some portion of the germinal coelomic epithelium, while the medulla is derived from the embryonic nerve cells of the sympathetic ganglia.

LYMPHOID TISSUE.—In addition to certain of the ductless glands, and the local or diffused masses of their characteristic tissue already mentioned in connexion with the alimentary canal, lymphoid tissue is often abundantly present in other parts of the body. There is, for example, a mass of this tissue on the heart of the Sturgeon (Acipenser). The anterior enlarged portion of the mesonephros, commonly termed the "head-kidney" of the Teleostomi (Fig. 203, B, C), is almost entirely composed of lymphoid tissue,[410] which has replaced, wholly or partially, the proper renal structure; and from the presence of free red blood-corpuscles and of crystals of oxy-haemoglobin and other derivatives of haemoglobin, it may be inferred that the "head-kidney," in common with the more orthodox blood-glands, performs a blood-destroying function.[411] On the other hand, the example of the spleen, which is alike the seat of leucocyte-formation and of blood-destruction, renders it unnecessary to reject the view that the "head-kidney" is an organ in which leucocytes or blood-corpuscles are formed. In but few Teleostomi is a purely lymphoid "head-kidney" entirely wanting, as, for example, in the Sun-Fish (Orthagoriscus mola).[412] As previously mentioned the Dipnoi are remarkable for the extraordinary development of lymphoid tissue, inasmuch as it forms a thick investing mass round the kidneys and gonads in addition to its exceptional abundance in the walls of the alimentary canal.

The absence of ordinary lymphatic glands in Fishes is well known, and it is at least probable that, functionally, the want of these lymphoid organs may be compensated for by the superabundance of lymphoid tissue in other parts of the body.[413]

{349}CHAPTER XIII

MUSCULAR SYSTEM—LOCOMOTION—SOUND-PRODUCING ORGANS—ELECTRIC ORGANS

MUSCULAR SYSTEM.—The various muscles of the body may be arranged in two systems: (i.) the somatic or parietal, composed of striated or voluntary muscle-fibres; and (ii.) the splanchnic or visceral, consisting for the most part of unstriated or involuntary fibres. Somatic muscles form the great lateral longitudinal muscles of the trunk and tail, which retain the primitive embryonic metamerism to a greater extent in Fishes than in any other Vertebrates, and are the principal muscles associated with locomotion. The lateral muscles are composed of a series of transverse muscle-segments or myotomes, which are >-shaped, or S-shaped, or they even take a zigzag course from above downward. The myotomes are disposed in pairs, and they are separated from one another by fibrous septa or myocommata. Each myotome is divided into a dorsal or epiaxial portion, and a ventral or hypaxial portion, by a longitudinal, horizontal, fibrous septum extending outwards from the vertebral centra to the skin. The muscles of the pectoral and pelvic fins are derivatives from more or fewer of the adjacent myotomes. The splanchnic muscles include the musculature of the walls of the alimentary canal, as well as those specialised portions of the visceral system which are represented by the muscles of the branchial arches and the jaws, and are composed of striated fibres.

LOCOMOTION.—A Fish and a Bird are equally remarkable for the many and various ways in which they are adapted for locomotion in the particular medium in which they live. In its shape the Fish is admirably adapted for cleaving the water. Spindle-like in shape, but thicker in front than behind, a Fish resembles {350}a double wedge, the thick part of which is represented by the head and one of the thin edges by the free hinder margin of the caudal fin. The body is bounded by smooth flowing contour lines, unbroken by any sharp separation of the body regions from one another, and with no points of resistance to its forward motion through the water. The body being thicker in front than behind, and, as seen in transverse section, broader above than below, it follows that its centre of gravity will be nearer the head than the tail, and nearer the dorsal than the ventral surface. The dorsal position of the centre of gravity necessarily renders the equilibrium of the body unstable, and were it not for the balancing action of the paired fins the Fish would float belly upwards, as is always the case after death. Most Fishes are provided with a membranous gas-containing sac, the air-bladder, the principal function of which is to render the Fish, bulk for bulk, of the same weight as the water, so that in this position of equilibrium, or plane of least effort, the animal can execute its various locomotor movements with a minimum expenditure of muscular effort—an advantage which no other animal possesses.[414] To give stability to the body, and to steady its course when swimming, the Fish has a dorsal and a ventral keel, formed by the anal and dorsal fins, which, like the sliding keel of a yacht, can be raised or lowered as occasion requires. When these fins are removed the course of the Fish becomes zigzag, and the animal wobbles.

The organs more directly concerned with swimming are the tail and the caudal fin, and the pectoral and pelvic fins, but the relative share which these structures take in the actual propulsion of the Fish differs greatly. The principal organ of locomotion in the typical Fish is the powerful muscular tail, which, in swimming, is lashed from side to side by the alternating contraction of the great longitudinal muscles on opposite sides of the vertebral column.[415] In such movements the tail is first flexed or bent, say to the right side: this stroke has been termed the non-effective or back stroke. By a stroke in the reverse direction the tail is then extended and straightened, that is to say, the Fish makes the forward or effective stroke. By a rapid succession of such strokes to the right and left sides alternately the Fish is {351}forced through the water. It is obvious, however, that the extension or effective stroke must have a considerable surplus of power over the flexion or non-effective stroke, and how this result is achieved will now be briefly considered. Experiment, and the observation of Fishes like the Sturgeon, which habitually move with sufficient slowness to allow the phases of their swimming movements to be followed without much difficulty, show that in swimming a Fish throws its body into two opposite and complementary curves, a cephalic curve formed by the anterior half of the body and a caudal curve by the tail. The double curve enables the Fish always to present a convex, less resisting or non-biting surface to the water during the flexion of the tail to the right or left as the case may be, and a concave or biting surface during extension, that is when the tail is straightening itself during the effective stroke.

Fig. 204, which represents a Fish in two successive positions while swimming, will serve to illustrate these conclusions. A Fish in the position A has its body thrown into a cephalic concavity directed to the right and a caudal concave surface facing the left. The tail is bent to the right of the line a b, which corresponds to the axis of the Fish when at rest and to the course pursued by the animal when swimming, and is in the position which it assumes during a flexion stroke, with its convex non-biting surface directed outwards and its concave biting surface inwards. The tail is now ready for an extension stroke, and while this is in progress it is clear that the concave biting surface of the tail will meet the water, while at the conclusion of the stroke the tail will be in a line with a b. At the same time the cephalic curve has so far diminished that the long axis of the body for a momentary period will also coincide with a b, and the Fish is free to advance without impediment. The tail, {352}however, continues its movement to the left, but now as a flexion stroke, and assumes the curvature and position indicated in B, with a reversal in the direction of both the cephalic and caudal curves, but in the meantime the force of the preceding extension stroke has forced the Fish along the line a b to the new position indicated by B. By a rapid succession of alternating flexions and extensions, during which the tail describes figure-of-8 curves, the Fish travels in an undulating forward course with a maximum of propelling power and a minimum of "slip." In short, the action of the tail precisely resembles the action of the stern-oar in the operation of sculling a boat.

There are also other considerations which add to the surplus power of the extension stroke by lessening the resistance of the water to the flexion or non-effective stroke. During the flexion stroke the tail fin is less expanded and its area diminished, and by the rotation of the Fish on its long axis the surface of the tail strikes the water obliquely, and further, the tail moves with less rapidity. On the contrary, when the extension stroke is made these conditions are reversed. The caudal fin is expanded, the stroke is more rapid, and by the reverse rotation of the Fish the tail now strikes the water with its flat surface. In other words, the action of the tail during the two strokes may be compared to the "feathering" of an oar in rowing. Nor is this all. A Fish in motion through the water produces a suction current behind it. The current offers but little resistance to the flexion stroke, inasmuch as the direction of the two coincide, but during the extension stroke the tail meets the full force of the current, and consequently its grip and propelling power are greatly enhanced. There is a striking analogy between the movements of a Fish's tail in swimming and the action of the screw of a steamer, but as a propelling organ the former is far superior to the latter. As we have seen, the tail of a living Fish can so adjust its shape and surface that it alternately eludes and grips the water in accordance with the needs of particular strokes.

The curves into which the body of a Fish is thrown when swimming are never less than two, but in long-bodied Fishes, such as the Eels, the number may be increased, and in every case the curves occur in pairs and are complementary to one another.

{353}Many Fishes can jump out of the water, either in pursuit of insect food, like the Trout, or to enable them to escape the pursuit of their foes, like the Flying-Fish (Exocoetus), by means of a single forcible stroke of the tail, when the Fish is in a nearly vertical position close to the surface of the water. It is thus that the Salmon executes its remarkable leaps over weirs or up salmon-ladders when ascending rivers for spawning.

The tail is also used for steering. If kept bent to one side when the Fish is moving the tail acts like a rudder, and the course of the Fish is deflected to that side; or the direction may be altered by single strokes of the tail to the right or left, according to the course which the Fish desires to pursue.

In the majority of Fishes the paired fins are probably of little use for propulsion, and their action in this as in other functions is not always clear. In the Sharks and Dog-Fishes as well as in some Teleosts their planes are nearly horizontal when the fins are extended from the body; in others they are more oblique, so that the surfaces of the fins look upwards and backwards, and downwards and forwards; and in others again their surfaces are so nearly vertical that their strokes will be backwards and forwards. The pectoral fins also vary in their position on the sides of the body, being much more dorsal in some Fishes than in others. The paired fins may act as lateral keels in steadying the course of the Fish especially when the fins are extended and their planes are horizontal. They certainly seem to act as balancers in keeping the Fish on an even keel, and in counteracting the tendency of the Fish to turn belly upwards—a result which is attained by a slight upward and downward movement of the fins, and particularly of the pectoral fins. A Fish deprived of its pectoral members sinks downwards at the head and assumes an oblique position in the water. Removal of both the pectoral and pelvic fins of one side causes the Fish to roll over to that side; and if the fins are removed from both sides the animal turns belly upwards like a dead Fish. The pectoral fins may also be used for steering: a backward stroke of one fin while the other is kept folded back against the body will wheel the Fish round to the opposite side. From the ventral position of its mouth a Shark is forced to turn over to one side in order to seize its prey, and this movement of rotation is probably produced by the down strokes of the pectoral fin of one side. In {354}some Fishes it would seem that the pectoral fins may assist locomotion by acting as paddles. The 15-spined Stickleback (Gastrosteus spinosus) frequently progresses by their aid alone; and, as their action can be reversed at pleasure, it is not unusual to see this Fish move backwards. The fins appear to be rotated or twisted in spiral movements like the tail when used for swimming, or like the wings of Insects in flying.

It has been mentioned that the function of the median fins (dorsal and anal) is to give stability to the Fish by acting as dorsal and ventral keels. This is certainly the case in the generality of Fishes. Nevertheless, there are exceptional instances in which one, or even both, of these fins are important swimming organs, acting either as a substitute for a tail which has become adapted for other uses, or as supplementary to that organ. Thus, in some of the Syngnathidae (Pipe-Fishes and Sea-Horses) the small size or absence of the caudal fin, and its use as a prehensile organ, renders the tail of little or no value as a propelling organ: hence it is that these Fishes swim by a lateral undulating movement of the dorsal fin. To enable them to do this the supporting skeleton presents certain interesting modifications. In the majority of Teleosts the arrangement of the fin-muscles, and the nature of the articulation between the dermal fin-rays and their basal radial supports, which is generally some form of a hinge-joint, are such as to limit the motion of the rays to simple elevation or depression in the vertical plane, and no lateral motion of the fin is possible. But in the Syngnathidae, as in the Pipe-Fish (Siphonostoma typhle), there is an exceptionally mobile articulation between the dermal fin-rays and the distal radial nodules which their cleft bases embrace and the bony proximal or basal radials, so that the fin can be flexed or bent to the right or to the left. In addition to this, by a change in the insertion of their tendons, the muscles corresponding to the ordinary elevator and depressor muscles of the fin-rays in other Fishes are capable of producing extensive lateral movements of the fin, or, by contracting in orderly sequence, of bringing about the characteristic undulating motion of the fin. A similar mechanism exists in many Plectognathi (e.g. species of Balistes, Monacanthus, Diodon, Tetrodon and Orthagoriscus)[416] in connexion with both the dorsal and anal fins, but in these Fishes the {355}action of the median fins in swimming must be regarded as supplementary to that of the tail.

Swimming is by no means the only form of locomotion in vogue amongst Fishes. A few, like the Angler-Fishes (Lophius), habitually use the pectoral fins for crawling about the sea-bottom. The East Indian Goby, Periophthalmus, uses its pectoral fins, which are bent at an angle like an elbow-joint, for hopping over sandy flats left bare by the retreating tide. The Flying-Fish (Exocoetus), when projected from the water by a stroke of its powerful tail, expands its large pectoral fins, and, using them after the fashion of a parachute, floats through the air for considerable distances before returning to its natural medium. The "Flying Gurnards" (Dactylopterus) are also capable of short aerial excursions in a similar fashion. Nor is tree-climbing beyond the province of a Fish, if credit be given to the assertion that the Indian "Climbing-Perch" (Anabas scandens) uses its opercular spines for ascending trees. Many freshwater Fishes are known to migrate across land from one pool or river to another, usually during the night. Eels do so by a serpentine or wriggling motion of their long bodies, but in others the pectoral fins seem to be the principal organs used for the purpose, aided, it may be, by a perverted use of the tail.

SOUND-PRODUCING ORGANS.—Contrary to popular belief sound-producing or vocal organs are by no means uncommon in Fishes, especially in certain families of Teleosts. It is not always easy, however, to discriminate between involuntary, abnormal, or accidental sounds, and those due to the action of special vocal organs. There are, moreover, some Fishes which observations have shown to utter highly characteristic sounds, although the precise nature of the sound-producing mechanism is at present unknown; while other Fishes appear to possess organs which, on anatomical grounds, are perhaps vocal in function, although nothing is known of the nature of the sounds they emit. Here those organs only will be considered which, either with certainty or with some degree of probability, may be regarded as vocal structures. For most of our knowledge of these interesting structures we are indebted to the researches of Sörensen and Dufossé.[417]

{356}[Illustration: FIG. 205.—Stridulating apparatus of Callomystax gagata. is^1, The first interspinous bone, the lower part of which forms the double file and fits into the interval between the cleft neural spines ns^4 and ns^5; is^2, is^3, second and third interspinous bones; ns^3, ns^4, ns^5, neural spines of the third, fourth, and fifth vertebrae; s^1, s^2, spine-like rays of the dorsal fin; so, supra-occipital. (After Haddon.)]

(a) Stridulation.—Stridulation as a method of sound-production has been recorded in many Teleosts, and one of the most interesting examples occurs in the singular Indian Siluroid, (Callomystax gagata).[418] In this Fish (Fig. 205) the first five vertebrae are rigidly connected with one another and with the skull, mainly through the union of the neural spines of the third, fourth, and fifth vertebrae, and their articulation with the supra-occipital bone. The united spines together form a high, laterally-compressed lamina of bone, the hinder portion of which is vertically cleft into two thin plates separated by an interval sufficiently wide to receive the first interspinous bone of the dorsal fin. The inner surface of each of the two plates is traversed by a series of about thirty parallel, close-set, vertical ridges, while the first interspinous bone is similarly ridged on both its faces like a double file. Lastly, it may be mentioned that owing to the width of the intervertebral ligament between them the fifth and sixth vertebral centra are articulated by a joint of unusual mobility. The action of the mechanism is simple. By the vertical movements of the sixth and succeeding trunk vertebrae, with the interspinous bones which they support, on the rigid structure formed by the head and first five vertebrae, the file-like first interspinous bone moves backwards and forwards, and, by scraping against the ridges on the inner surfaces of the cleft neural spines, gives rise to a harsh grating noise, which is particularly unpleasant when artificially produced. The lateral movements of the trunk in ordinary locomotion do not affect the mechanism: it is only when the trunk is alternately flexed and extended in the vertical plane that the mechanism comes into play and a noise is {357}produced. In the Bull-head (Cottus scorpius) the preoperculum is modified for stridulation, and in Dactylopterus the hyomandibular bone; in other Fishes, as in some Siluroids (e.g. species of Doras), stridulation takes place between a basal process from the great spine of the pectoral fin and the wall of a socket in the cleithrum into which the process is received, or between the small first spine of the dorsal fin and a roof-like process at the upper extremity of the first interspinous bone; also, in a somewhat similar fashion in the anterior dorsal fin of such widely different Fishes as certain Trigger-Fishes (Sclerodermi) pertaining to the genera Balistes, Monacanthus, and Triacanthus, Acanthurus chirurgus (Acanthuridae), the Boar-Fish (Capros aper), Centriscus scolopax (Centriscidae), and the Three-spined Stickleback (Gastrosteus aculeatus); and even between the spinose ray of the pelvic fin and the basipterygium in Triacanthus, Capros, and Gastrosteus.

In the "Drumming" Trigger-Fish (Balistes aculeatus),[419] which frequents the coral-reefs off the Island of Mauritius, stridulation takes place between the postclavicles and a longitudinally grooved area on the inner surface of each cleithrum. Both the cleithra and postclavicles are in intimate relation with the air-bladder, and the sound produced by friction is apparently strengthened by the transference of the vibrations to the walls and gaseous contents of that organ. The passage of the sound-vibrations to the surrounding medium is facilitated by the fact that for a portion of their extent the lateral walls of the air-bladder are in contact with the superficial skin, which visibly shares in the vibratory movement of the bladder when the characteristic drumming sounds of Balistes are being emitted.

Stridulating sounds may also be produced by the friction of the upper and lower pharyngeal teeth, as in a species of Mackerel (Scomber brachyurus). By the grating of its teeth the Sun-Fish (Orthagoriscus mola) is said to emit sounds similar to those produced by the grinding of the teeth in Pigs and Ruminants; and Moseley[420] has remarked of a species of Balistes that the "living Fish when held in the hand makes a curious metallic clicking noise by grating its teeth."

(b) Breathing sounds.—Characteristic breathing or murmuring sounds, or "bruits de souffle" as Dufossé terms them, are {358}produced by a few Teleosts, among which may be mentioned the Eels, certain Cyprinidae, as, for example, the Carp (Cyprinus carpio), several species of Loaches (e.g. Misgurnus fossilis and Cobitis taenia), and the European Siluroid, Silurus glanis. According to Dufossé these sounds originate in some cases from the expulsion of gas from the air-bladder through the ductus pneumaticus and mouth, and in others, as in Misgurnus fossilis, they are produced by the rapid ejection through the anus of bubbles of air previously taken in at the mouth.

(c) Sounds produced through the agency of muscles connected with the air-bladder.—In addition to its usual function as a hydrostatic organ or "float" the air-bladder is often modified in various ways in different Teleosts, and adapted for use as a sound-producing organ.

In the South American Siluroid, Auchenipterus nodosus, the transverse processes of the fourth vertebra are bent downwards and backwards, and at the same time become converted into flexible and highly elastic springs (Fig. 206, B). Their distal extremities expand into oval bony plates which are imbedded in the anterior wall of the air-bladder, and often cause the latter to bulge inwards (Fig. 206, A). From the occipital region of the skull arise two powerful muscles which pass backwards to {359}their insertion into the anterior faces of the two springs. By the contraction of these muscles the springs, and consequently also the front wall of the bladder, are drawn forwards; but directly the muscles relax, the elasticity of the springs causes them to move backwards to their former position, carrying with them the wall of the air-bladder. Hence it follows that the rapid alternating contraction and relaxation of the muscles will impart a vibratory movement to the anterior wall of the bladder and to the gaseous contents of that organ, with the result that a sound is produced. As a rule, those Fishes in which an elastic-spring-mechanism is present have the air-bladder subdivided by internal septa into a series of chambers freely communicating with one another; and no doubt the intensity of the sound is greatly increased by the vibratory movements of the gases across the free edges of the septa, and from one chamber to another. The elastic-spring type of vocal organ is apparently restricted to the Siluridae,[421] and besides occurring in Auchenipterus is found also in the South American genera Doras, Oxydoras, Rhinodoras, and Euanemus; in the African genera Synodontis and Malopterurus; and in at least four species of the Indian genus Pangasius.[422] There are also a few Teleosts in which the air-bladder is provided with special muscles, but, instead of being connected with elastic springs, the muscles extend from the skull, and are inserted directly into the wall of the bladder (Fig. 207); or, without being in any way attached to the skeleton, the muscles simply invest some portion of the surface of the air-bladder. In other Fishes the air-bladder, without possessing special muscles of its own, may, nevertheless, be partially invested by tendinous, or partly muscular and partly tendinous, extensions from the muscles of the body-wall (Fig. 208), or may be intimately related to certain muscles connected with the pectoral girdle.

{360}[Illustration: FIG. 207.—Ventral view of the air-bladder and its extrinsic muscle in Platystoma. a.b, Air-bladder; a.l.c, left antero-lateral caecum of the bladder; b.o, basioccipital; b.w, body-wall in contact with the lateral wall of the bladder; c^1, centrum of the first vertebra; cl, clavicle; d.p, ductus pneumaticus; m^1 and m^2, extrinsic muscles of the bladder; pt.i, post-temporal. (From Bridge and Haddon.)]

Whatever the precise relation of the air-bladder to its muscles it is probable that the physiological effect is in most cases the same. By the rapid alternating contraction and relaxation of the muscles, some part of the wall of the bladder becomes alternately compressed and relaxed in such a way as to initiate a series of vibratory movements in the gases of that organ, and so produce definite sounds. In not a few of the Fishes the cavity of the bladder is subdivided by external constrictions or by internal septa, or is complicated by the development of lateral, tubular, caecal branches; and hence the vibratory movements of the gases will be greatly strengthened by their passage across the edges of the septa, or the apertures of the caeca, and the intensity of the resultant sounds also increased. It will be readily understood that the nature and quality of the sounds emitted by different Fishes will necessarily vary with the shape of the air-bladder, the number and arrangement of the internal septa and the caeca, and the strength and disposition of the contracting muscles. In a few Teleosts (Triglidae and Zeidae) sounds are said to be produced by the rapid vibration of an annular, or centrally-perforated, muscular diaphragm, which stretches across the cavity of the air-bladder.[423] Nevertheless, it must be strongly emphasised that, while in some Fishes the air-bladder and its muscles {361}undoubtedly constitute a vocal organ, there are many others in which the bladder can only be inferred to be sound-producing from its general agreement in anatomical structure with the same organ in Fishes where its vocal function has been clearly proved.

By one or other of these various methods the air-bladder is either known to be sound-producing, or is believed with good reason to be such, in the following Teleosts,[424] and many others:—Certain species of the South-American genera of Siluridae, Pimelodus, Sorubim, Platystoma, Piratinga, Centromochlus, and Trachelyopterus; species of the South-American family Characinidae; Amblyopsis spelaea, the blind Fish from the Mammoth Cave of Kentucky (Amblyopsidae); among the Syngnathidae, the short-snouted Sea-Horse (Hippocampus brevirostris) of the British Coasts; certain Sclerodermi, such as the Trigger-Fishes, Batistes vetula, Triacanthus brevirostris, T. biaculeatus, and Monacanthus pardalis, and also some "Coffer Fishes" (e.g. species of Ostracion); some Gymnodontes (species of Diodon and Tetrodon); a few Serranidae (e.g. species of Therapon and Pristipoma); species of Holacanthus (Chaetodontidae) and in Holocentrum sogho (Berycidae); such Sciaenidae as the "Drum" (Pogonias chromis), the "Maigre" (Sciaena aquila), which has sometimes been taken in British waters, Umbrina cirrhosa, Otolithus regalis, and Micropogon undulatus, and, with more or less probability, many other species of the same family; one species of Zeidae, the John Dory (Zeus faber); Batrachus tau among the Batrachidae; several species of Gurnards (Triglidae) belonging to the genera Prionotus and Trigla; the so-called Flying Gurnard, Dactylopterus volitans (Dactylopteridae); the Indian species Ophiocephalus marulius and O. gachua (Ophiocephalidae); amongst the Gadidae, the Cod (Gadus morrhua) and the Haddock (G. aeglefinus); in such Zoarcidae as the blind Fish (Lucifuga subterranea) from the subterranean waters of the caves of Cuba, and also in some Ophidiidae (e.g. species of Ophidium).

In Fishes other than Teleosts, instances of normal sound-production by special vocal structures are rare. No recorded instances are known in the Cyclostomes or the Elasmobranchs,[425] {362}but there is evidence that sounds are emitted by Polypterus among the Crossopterygii, and by the Dipnoids Neoceratodus,[426] Protopterus, and Lepidosiren, although it is not certainly known how they are produced, or that they may not be the accidental concomitants of the inspiratory or expiratory action of the lungs in breathing.

As to the nature of the sounds produced by the air-bladder and its muscles in different Teleosts, a few examples may be given.

The sound produced by the elastic-spring-apparatus of a recently caught Doras maculatus, has been described as a "deep growling tone," which may be distinctly heard at a distance of 100 feet when the Fish is out of the water. Under like conditions the air-bladder and its muscles, in a species of Platystoma, emit a similar sound. On the other hand, the sound produced by the elastic springs of the Electric Siluroid (Malopterurus electricus) has been compared to the hissing of a cat. The Sea-Horse (Hippocampus brevirostris) utters a monotonous sound analogous to that of a tambour, which is characteristic of both sexes, but is more intense and frequent in the breeding season. The "Coffer Fish" (Ostracion trigonus) emits a growling sound, as also does the "Globe Fish" (Tetrodon honckenii) when taken out of the water.[427] The air-bladder and its muscles in the "Drum" (Pogonias chromis), constitute the most powerful sound-producing organ yet found in any Fish. The sounds emitted by the "Drum" are better expressed by the word drumming than by any other, and have frequently been heard by persons in vessels lying at anchor on the coasts of the United States, where these Fishes abound.[428] The "Drum" begins its drumming noise in the spawning season in April, but is rarely heard afterwards. The "Maigre" (Sciaena aquila), whose musical performances are perhaps responsible for the Homeric fable of the song of the Sirens, is remarkable among Fishes for the variety of its sounds, which have been compared to bellowing, purring, buzzing, and whistling.[429] The sound is often so intense that it may be heard when the Fish is at a depth of 18 metres, and the {363}ear of the observer two metres above the water; and it has been recorded that by listening for these sounds, shoals of Maigres have been successfully netted. They rarely emit sounds when isolated; but in shoals, during the breeding season, they do not cease to make sounds with a vigour and a persistency which apparently must soon wear out their strength. One of the Indian Horse-Mackerels (Caranx hippos) grunts like a young Pig when captured, and the sound is repeated whenever it is moved, as long as vitality remains. A West Indian species of the same family (Argyriosus vomer) has been observed to produce a like sound, while an Egyptian Caranx (C. rhonchus) is known to the Arabs as the "Chakoura" or "Snorter."[430] The sounds produced by the different British Gurnards, such as the Grey Gurnard (Trigla gurnardus), the Piper (T. lyra), the Elleck or Cuckoo Gurnard (T. cuculus), and the Tub-Fish (T. hirundo), have been compared to snoring, a sonorous and prolonged grunting, crooning (whence, perhaps, the term "crooner," by which the Grey Gurnard is known in Ireland), and croaking. The John Dory (Zeus faber)[431] also utters sounds analogous to those of the Gurnards. Among the Dipnoi Lepidosiren is said to make a growling sound, and Neoceratodus a grunting noise which may be heard at night for some distance.

Whatever the nature of the vocal mechanism, it is highly probable that the sounds produced by Fishes travel to considerable distances in the water, inasmuch as the latter medium is a far better conductor of sound than air, and, moreover, the transmission of sound-vibrations from the air-bladder to the water is facilitated in many Fishes by the fact that, for a portion of its extent on each side the bladder is in direct contact with the superficial skin behind the pectoral girdle.

From the by no means exhaustive list of examples given above, it is obvious that in some form or other vocal organs are present in a considerable number of Fishes, both freshwater and marine, belonging to widely different groups; and further, that even in the same species (e.g. Doras maculatus and other Siluridae), both stridulation and the action of extrinsic muscles on the air-bladder may be utilised as a means of sound-production. Certain Teleostean families like the Siluridae, the Sciaenidae, and the Triglidae, seem to be distinguished above all others by the {364}prevalence of some form of vocal organ. According to Sörensen, the first mentioned of the three families includes no less than 68 species, which utilise the air-bladder alone as a sound-producing organ. Nevertheless, there still remain many Teleostean families, rich in genera and species, and with an almost world-wide geographical distribution, in which such organs have not yet been found.

The advantages which Fishes derive from the possession of sound-producing organs are sufficiently obvious.

A characteristic feature in the reproduction of most Fishes is the general absence of any process of conjugation between the sexes, the eggs being fertilised in the water after their extrusion from the body of the female, and, consequently, any device which will facilitate the formation of shoals during the breeding season must be of great advantage to the species by largely increasing the chances that the ova will be fertilised, and thus secure the more successful propagation of the race. Hence it may be concluded that the vocal organs of Fishes are a means to this end, and that the sounds they produce are in fact recognition-sounds which enable Fishes of the same species to congregate together at periods when reproductive activity is greatest. This view is in harmony with much that is known of the habits of these Fishes, especially with the fact that particular sounds are often characteristic of particular species, and that the sounds are produced most frequently and with greater intensity during the breeding season than at any other time. While useful to all Fishes that possess them, vocal organs are, no doubt, specially serviceable to those Fishes which, from the nature of their habitat, can make but little use of their eyes; and this fact may perhaps explain the prevalence of such organs in the Siluridae, which are frequently bottom- or ground-feeding Fishes, and often live in muddy waters.

The sounds emitted by Fishes may also, in some instances at least, be warning sounds. Many of the sound-producing Fishes are provided with exceptionally strong spines either in connexion with the median and paired fins, as in many Siluridae, or on the general surface of the body, as in Diodon hystrix. Such spines are very effective weapons for offensive or defensive purposes, and are capable of inflicting very severe wounds. The natural enemies of these Fishes learn by experience or instinct to {365}associate particular sounds with the possession of dangerous spines, and warned by the sounds, they refrain from attacking the owner of the spines, to the mutual advantage of both.

ELECTRIC ORGANS.—Electric organs capable of generating more or less powerful electric discharges are present in certain Fishes, both marine and freshwater. They occur in a few Elasmobranchs (species of Raia, Torpedo, and Hypnos), in such Teleosts as the African Silurid Malopterurus the "Electric Eel" (Gymnotus), and in species of Mormyridae (e.g. Mormyrus). With one exception electric organs are composed of metamorphosed muscular fibres, and their nerve-endings or motor end-plates. The species of Raia have two small electric organs, one on each side of the terminal portion of the tail.[432] In Gymnotus[433] the {366}organs are much larger, and extend the whole length of the tail, which is fully four-fifths of the total length of the Fish. The Mormyridae also have their feeble electric organs in the caudal region. In all these Fishes the electric organs are modified portions of the caudal muscles. In the Torpedo, however, these organs are two large oval masses, one on each side of the head, between the gills and the cephalic prolongation of the pectoral fin (Fig. 209). Malopterurus[434] is exceptional in possessing an electric organ derived from the epidermis and not from the muscular system. In this Fish the organ envelops nearly the whole body like a mantle, between the skin and the subjacent muscles of the trunk and tail. An electric organ is composed of an immense number of "electric plates" (modified motor end-plates), abundantly supplied with nerves on one of their surfaces, and disposed in a series of vertical (Torpedo) or longitudinal (Gymnotus) columns, separated by septa of connective tissue. In the active state of the organ in the Torpedo[435] the ventral surfaces of the plates, on which the nerves are distributed, become negative to the dorsal, and "the effect in all the plates of a column when summed up is, therefore, such that the dorsal end of a column becomes positive to the ventral end."[436] Hence the current in the form of a succession of shocks passes from the ventral to the dorsal surface of the head. In Gymnotus, where the columns are longitudinally arranged, it is the anterior and posterior surfaces which become oppositely electrified, and the current passes from the tail to the head. The shock imparted by an electric discharge is most powerful in Gymnotus,[437] Malopterurus, and Torpedo, in the order named, and relatively weak in the remaining genera. The strength of the shock increases with the number of electric plates included in the circuit. Thus in Gymnotus the maximum shock is given when the body of the Fish is so curved that the head and the tail are in contact with different points on the surface of some other Fish. The discharge may be reflex or voluntary. Repeated discharges induce fatigue and weaken the shocks. Electric organs are powerful offensive or defensive structures, enabling the Fish to repel the attacks of enemies, or to stun or kill their prey.

{367}CHAPTER XIV

NERVOUS SYSTEM AND ORGANS OF SPECIAL SENSE

The nervous system consists of the brain and the spinal cord, and of the cranial and spinal nerves. The rudiment of the future brain and spinal cord first appears in the embryos of some Cyclostomes (e.g. Bdellostoma), of Elasmobranchs, and of Chondrostei (e.g. Acipenser), and of Neoceratodus among the Dipnoi, in the form of a tubular medullary canal pinched off from the epiblast of the dorsal surface of the body. By a somewhat different method, but with the same final result, a medullary canal is formed in other Cyclostomes (e.g. Petromyzon), in the Holostei and Teleostei, and in Lepidosiren,[438] from a solid ingrowing keel of epiblast which subsequently becomes tubular. Later, the medullary canal in the head enlarges, and becomes divided by two transverse constrictions into three vesicles, the primary fore-, mid-, and hind-brain, leaving the rest of the canal to form the spinal cord.

THE SPINAL CORD.—This portion of the medullary canal retains a simpler and more uniform cylindrical structure. Its walls thicken and their component cells become converted into nerve cells and nerve fibres, but a remnant of the original cavity remains in the adult as a minute axial canal, with a ciliated epithelial lining, the central canal of the spinal cord or myelocoele. In most Fishes the spinal cord extends the whole length of the body, but in some Teleosts, especially in certain Plectognathi, it is remarkably short. In a Sun-Fish (Orthagoriscus), 2½ metres long, and weighing about a ton and a half, the cord was only 15 mm. in length, or shorter than the brain.

THE BRAIN.—At an early stage in its embryonic history the {368}brain consists of three simple vesicles, the fore-, the mid-, and the hind-brain, the first of which lies in front of the anterior end of the notochord and is therefore pre-chordal in position. As development proceeds the walls of the vesicles undergo local thickenings, or they give rise to hollow paired or median outgrowths, and by one or other of these methods the different parts of the complex adult brain are evolved, while the original cavities of the vesicles or of their outgrowths persist as a continuous system of epithelium-lined spaces or "ventricles."[439] The fore-brain is remarkable for the number and importance of the parts to which it gives rise. First, it bulges out in front into a hollow vesicle, the prosencephalon, leaving the rest of the fore-brain as the thalamencephalon or diencephalon (Fig. 210). The cavity of the prosencephalon is the prosocoele, and a pair of thickenings in its floor form two basal ganglia or corpora striata. In many Fishes the prosencephalon retains this simple vesicular condition, in which case the roof or pallium is usually epithelial and non-nervous; but in others two hollow lobes grow out from it in front and give rise to two cerebral hemispheres or parencephala.[440] Both contain extensions of the prosocoele, the paracoeles or lateral ventricles, from the floor of which the corpora striata now project. The prolongation of the pallium forming the roof of the lateral ventricles either remains partially epithelial, or it may acquire a wholly nervous structure and thicken to an extent which differs greatly in different Fishes. With the formation of the hemispheres the prosencephalon and its prosocoele become of secondary importance, and may cease to be recognisable as distinct from the thalamencephalon and its ventricle. The lateral ventricles then appear to communicate directly with the third ventricle by two apertures, the foramina of Munro. The forward growth of the brain is completed by the development of two hollow lobes, the olfactory lobes or rhinencephala, each of which contains a ventricle or rhinocoele communicating behind with the prosocoele, or, if hemispheres are present, with the corresponding lateral ventricle.

{369}[Illustration: FIG. 210.—Diagram of the general structure of the brain in Craniates. A, vertical longitudinal section; B, dorsal view showing the brain cavities on the right side. c, Cerebellum; c.c, central canal of the spinal cord; c.h, cerebral hemispheres; c.s, corpus striatum; F.B, fore-brain; f.m, foramen of Munro; H.B, hind-brain; in, infundibulum; l.v, lateral ventricle; m, mesocoele; M.B, mid-brain; m.o, medulla oblongata; o.l, olfactory lobe; op.l, optic lobe; op.t, optic thalamus; p, paraphysis; pc, prosocoele; pn.o, pineal organ; p.o, parietal organ; pr, prosencephalon; pt, pituitary body; rh, rhinocoele; sp.c, spinal cord; s.v, saccus vasculosus; th, thalamencephalon; iii, iv, third and fourth ventricles. (After Parker and Haswell.)]

Scarcely less complicated, and perhaps even more interesting from a morphological standpoint, are the structures arising out of the thalamencephalon. By thickenings of its lateral walls two large ganglia, the optic thalami, are formed, and on the inner or dorsal aspect of each of these a ganglion habenulae is developed. From the sides of the thalamencephalon the primary optic vesicles are derived, which later become transformed into the retinal parts of the paired eyes and the optic nerves. Besides the optic vesicles there is a second pair of embryonic outgrowths which arise from the roof of the thalamencephalon. These outgrowths form stalked vesicles and represent a pair of degenerate visual {370}organs. Usually they become so displaced that the left one lies in front of the right, and they appear as if median. The subsequent fate of the vesicles differs greatly in different Craniates. Both persist in the Lamprey, the right vesicle to some extent retaining its primitive visual function as a parietal eye and directly overlying the left or pineal vesicle. In Elasmobranchs the two unite to form a glandular organ, the so-called pineal body of the adult, and in Teleosts the left vesicle disappears, leaving the right as a pineal body.[441] There is also an embryonic median outgrowth from the roof of the prosencephalon, the paraphysis, which soon disappears and whose significance is not known. A median hollow downgrowth from the floor of the thalamencephalon forms the infundibulum, which becomes attached to a caecal diverticulum from the roof of the mouth. With rare exceptions the diverticulum loses all connexion with the mouth, and, as the pituitary body or hypophysis, it appears as an appendage to the extremity of the infundibulum. In the Crossopterygii the connexion is retained even in the adult by means of a slender canal extending from the pituitary body and opening into the oral cavity. Laterally, the base of the infundibulum grows out into a pair of rounded lobes, the lobi inferiores, and distally into a thin-walled glandular sac, the saccus vasculosus, which lies just behind the pituitary body. The cavity of the thalamencephalon persists as the third ventricle or diacoele. The parts of the brain developed from the mid-brain and the hind-brain are much less complicated, and, except for variations in size, they present a fairly uniform character in most Fishes.

In the mid-brain the roof bulges out into a pair of optic lobes, and by the growth of lateral thickenings in its floor two thick strands of longitudinally disposed nerve fibres, the crura cerebri, are formed. The cavity of the mid-brain remains as the mesocoele, and from it an extension may be prolonged into each optic lobe.

From the hind-brain are formed the cerebellum or epencephalon and the medulla oblongata or metencephalon, the former as a dorsal bulging, the latter as a ventral thickening. Except where the cerebellum is developed the dorsal wall remains epithelial, and forms the roof of the persistent cavity of the {371}hind-brain, the fourth ventricle or metacoele, which retains its primitive continuity with the central canal of the spinal cord. Lateral lobe-like outgrowths from the dorsal columns of the medulla are conspicuous structures in some Fishes, and are known as corpora restiformia. The paired portions of the brain are connected across the middle line by a series of transverse commissures. The more important modifications of the brain in Cyclostomes and Fishes will now be briefly dealt with.

{372}[Illustration: FIG. 212.—Dorsal view of the brain of Myxine. c.r, Corpora restiformia; m.o, medulla oblongata; n.p, naso-pituitary canal; ol.o, olfactory organ enclosed in its fenestrated cartilaginous capsule; op.l, optic lobes; pr, prosencephalon; s, s, dorsal roots of spinal nerves; sp.c, spinal cord; th, thalamencephalon. (From Wiedersheim, after Retzius.)]

In the Cyclostome Petromyzon there is a small prosencephalon with an undivided prosocoele, and on each side of it a small cerebral hemisphere which appears as a mere appendage to the much larger olfactory lobe (Fig. 211). The prosocoele divides in front into two outwardly directed branches, and of the two diverticula into which each branch divides one extends as a lateral ventricle into the hemisphere of its side, and the other as a rhinocoele into the corresponding olfactory lobe. The ganglia habenulae are unusually large, the right one being larger than the left. The optic lobes are large, but not obviously double. So small is the cerebellum that it seems to be little more than a narrow transverse band crossing the fore-part of the fourth ventricle. The roof of the brain is largely epithelial, especially in the prosencephalon, the thalamencephalon, and the hind-brain. Over these epithelial areas the pia mater is unusually vascular and forms a series of "choroid plexuses." The ventricular system is complete and continuous. By contrast with the Lamprey the brain of Myxine[442] is very primitive, more so perhaps than in any other Craniate (Fig. 212). In a dorsal view the brain is divided into four pairs of laterally expanded and longitudinally compressed lobes by a median longitudinal fissure and three transverse fissures. The two anterior lobes are little more than the thickened anterior wall of the thalamencephalon, although, judging from their histological structure, they represent a very imperfectly differentiated prosencephalon and olfactory lobes. The second and largest pair constitute the thalamencephalon. The last two pairs of lobes represent a transversely divided pair of optic lobes, or "corpora quadrigemina." There is a large medulla oblongata with a pair of corpora restiformia, but the {373}cerebellum is entirely absent. The ventricles are subject to some individual variation. Third and fourth ventricles are generally recognisable, either as isolated cavities or connected by a remnant of the mesocoele. In the feeble development of the prosencephalon, in the striking preponderance of the mid-brain over the rest of the brain, and in the absence of a cerebellum, Myxine is unique amongst Craniates.

In Elasmobranchs among Fishes the brain attains a much higher grade of structure. In Scyllium (Fig. 213) there is a large prosencephalon, and directly in front of it a pair of imperfectly differentiated cerebral hemispheres, while from its antero-lateral regions the large olfactory lobes arise. The prosocoele divides in front into four diverticula, of which the two {374}inner ones extend into the hemispheres as lateral ventricles, and the two outer as rhinocoeles into the olfactory lobes (Fig. 214). In connexion with the infundibulum there is a pair of sacci vasculosi, consisting mainly of gland-tubules, opening into the infundibular cavity.[443] The cerebellum is exceptionally large, but it does not form a "valvula cerebelli." Large ear-like corpora restiformia are present. The third and fourth ventricles alone retain an epithelial roof in relation with choroid plexuses.

In all essentials the brain of the Holocephali is a repetition of the Elasmobranch type, more especially of the elongated form seen in Notidanus. Indications of a higher grade of structure are, however, to be seen in the reduction of the prosencephalon which, with its prosocoele, is now scarcely distinguishable from the thalamencephalon and its ventricle; and in the more complete differentiation of the cerebral hemispheres from one another and from the rest of the brain. Large frilled corpora restiformia are conspicuous structures on each side of the medulla oblongata. Besides the usual intra-cranial pituitary body, there is also a separate extra-cranial portion lodged in a pit on the ventral surface of the basis cranii: in the embryo the two are continuous.

In the Teleostomi the brain is distinctly of a more primitive type than in any other Fishes (Fig. 215).[444] The most striking feature is the absence of cerebral hemispheres, the evolution of the primary fore-brain proceeding no farther than the formation of an undivided prosencephalon with a non-nervous roof, and a prosocoele which forms a continuous cavity with the third ventricle, or at the most is only separated from it by an infolding of the epithelial roof or velum transversum.

{375}[Illustration: FIG. 215.—A, dorsal view of the brain of a Trout (Salmo fario); B, a vertical longitudinal section. c.il, Commissura interlobularis; g.h, ganglion habenulae; h.c, habenular commissure; i.c, inferior commissure; l.i, lobus inferior; myc, myelocoele; p.c, posterior commissure; v.o, valvula cerebelli; v.t, velum transversum; ii., optic nerve; v.iii., v.iv., third and fourth ventricles; v, vii, viii, ix, x, fifth, seventh, eighth, ninth, and tenth cranial nerves; remaining reference letters as in Fig. 210. (A, From Wiedersheim; B, after Haller.)]

Amongst other diagnostic characters may be mentioned the predominance of the mid-brain over the other divisions, the anterior extension of the large cerebellum into the mesocoele as a "valvula cerebelli," {376}and the absence of corpora restiformia. This type of brain is most strongly marked in the Teleostei, but in other Teleostomes some, like Acipenser,[445] are typically Teleostean in this respect (Fig. 216), while others, such as Lepidosteus, have small cerebral hemispheres with lateral ventricles as well as a prosencephalon.

The most obvious feature in the brain of the Dipnoi is the great development of the cerebral hemispheres. In this respect these Fishes approach the Amphibia, but in other features of brain-structure they present points of agreement with most other groups of Fishes without being closely related to any one of them. In Protopterus[446] (Fig. 217) the hemispheres are quite distinct except behind, and the walls of their spacious lateral ventricles are entirely nervous. Olfactory lobes are sessile on their anterior extremities, and behind and below they enlarge into ventral lobes which probably represent the hippocampal lobes of the higher Vertebrates. A vesicular pineal body at the end of a slender stalk overlies a singular conical projection from the roof of the thalamencephalon or "pineal pillow."

{377}[Illustration: FIG. 217.—Dorsal (A), ventral (B), and lateral (C) views of the brain of Protopterus annectens. C, Cerebellum; C.H, cerebral hemisphere; D.S.E, branches of the sinus endolymphaticus; In, infundibulum; L.I,, lobi inferiores; M.O, medulla oblongata; O.L, olfactory lobe; Op.L, optic lobe; P, pituitary body; P.B, "pineal pillow"; S.E, sinus endolymphaticus; Sp.c, spinal cord; Sp.n, spinal nerve; Vel, velum transversum; Z, pineal body; IV.V., fourth ventricle; ii., iii., iv., v., vi., vii., viii.1, viii.2, viii.3.4, ix., and x., roots of the cranial nerves. (From Burckhardt.)]

The optic lobes form a single oval body, and, as in Petromyzon and the Amphibia, the cerebellum is very small. A posterior choroid plexus covers the roof of the fourth ventricle, and an anterior plexus in connexion with the roof of the thalamencephalon projects downwards into the third ventricle, and is also prolonged forwards into each lateral ventricle. In Neoceratodus[447] the brain is certainly more primitive and distinctly less Amphibian. As compared with Protopterus the olfactory lobes and the cerebellum are larger, and the optic lobes are paired. The smaller hemispheres are {378}non-nervous dorsally and medianly, the roof and inner wall of each being formed by an extension of the thick, glandular choroid plexus which forms the roof of the thalamencephalon.

THE SPINAL NERVES.—The spinal nerves of Cyclostomes (e.g. Petromyzon) consist of a series of dorsal nerves arising on each side from the dorsal surface of the spinal cord, and of a similar double series arising from the ventral surface, the dorsal nerves regularly alternating with the ventral nerves. Each myotome is supplied by a dorsal and a ventral nerve which pass separately to their peripheral distribution in the skin and muscles. In Fishes, as in the higher Vertebrates, each dorsal nerve, now termed a dorsal root, enlarges into a ganglion and then unites, either before or directly after issuing from the neural canal, with the next ventral nerve or ventral root in front to form a main spinal nerve. At the same time the spinal nerves of opposite sides tend to form pairs in the same transverse plane. After the union of the two roots the spinal nerve divides into three typical branches: a dorsal nerve (ramus dorsalis), and a ventral nerve (ramus ventralis), both of which include somatic sensory or afferent fibres, and somatic motor or efferent fibres, for the innervation of the skin and muscles of the dorsal and lateral portions of a myotome; and a visceral branch (ramus visceralis), composed of afferent and efferent visceral fibres, which supplies the adjacent viscera (alimentary canal and its glands and blood-vessels), and helps to form the sympathetic nervous system.[448] The somatic afferent and the visceral afferent fibres enter the spinal cord by the dorsal roots, the somatic efferent leaving the cord through the ventral roots, although the visceral efferent fibres traverse both roots. In the region of the paired fins more or fewer of the rami ventrales unite to form a plexus, the brachial or the pelvic plexus, from which the nerves to the fins take their origin.

THE CRANIAL NERVES.—It is usual to describe the cranial nerves of Cyclostomes and Fishes as consisting of ten serially disposed pairs, viz.: the olfactory (i.), optic (ii.), oculomotor (iii.), trochlear (iv.), trigeminal (v.), abducens (vi.), facial (vii.), auditory (viii.), glossopharyngeal (ix.), and the vagus (x.) Like the spinal nerves, the cranial nerves collectively include somatic sensory (general cutaneous) and motor fibres, and also visceral sensory {379}and motor fibres, all of which have their own special centres in the brain, but the proportions of these nerve components differ greatly in different nerves. Certain preoral nerves (iii., iv., and vi.) are exclusively somatic motor; others (i. and ii.) are special sensory nerves for the olfactory and visual organs; but most of the other cranial nerves include several components, and are therefore "mixed" nerves. Besides these components some cranial nerves include also a quasi-independent system of nerve-fibres, which converge from certain cutaneous sense-organs to an independent centre in the medulla oblongata, the tuber acusticum,[449] and is probably derived from the general cutaneous system of nerve components. Such nerve fibres, including also the auditory nerve, which has its origin from the same centre, constitute the lateralis system. Perhaps the most striking feature in the postoral cranial nerves is the predominance of the visceralis or sympathetic system over the somatic. Omitting the lateralis fibres and a relatively few somatic sensory fibres, visceral fibres, sensory and motor, are the principal components of all these nerves, including v. but excluding viii. The reason for this is to be found in the fact that splanchnic or visceral muscles in relation with the jaws and branchial arches have usurped the place of somatic muscles in the muscular system of the head. For developmental and other reasons the olfactory and optic nerves stand in a category of their own, and the same may be said of the third, fourth, and sixth nerves, which innervate the muscles of the eyeball. The remaining nerves, all of which have their origin in the medulla oblongata, possess certain features in common, and as they are related to the gill-clefts in such a way that each forks over a cleft, they may be conveniently distinguished as "branchial" or "branchiomeric nerves." A typical branchial nerve consists of (1) a principal ganglion near the origin of the nerve from the brain; (2) a main trunk which gives off (3) a somatic sensory branch or dorsal nerve to the skin; (4) a palatine nerve (visceral sensory) to the oral or pharyngeal mucous membrane; (5) an epibranchial ganglion which is associated with a transitory embryonic epibranchial sensory organ at the dorsal border of a branchial cleft; (6) a pre-branchial nerve (visceral sensory), skirting the anterior margin of a cleft in its ventral course; and (7) a post-branchial branch (visceral motor) similarly related to the hinder margin.

{380}[Illustration: FIG. 218.—Diagram showing the principal branches of the cranial nerves in a Fish, mk.c, Meckel's cartilage; ol.o, olfactory organ; p.q, palato-quadrate; s, spiracle; i-v, branchial clefts; i, ii, iii, iv, vi, the first, second, third, fourth, and sixth cranial nerves. The remaining nerves are differently shaded. Black.—The trigeminal nerve: g.g, Gasserian ganglion; md, mandibularis; mx, maxillaris; op.p, ophthalmicus profundus. Oblique shading.—The lateralis system and its centre (t.a), the tuber acusticum: bucc.vii, buccalis branch of vii; md.ex, external mandibular branches of vii; _l.n._x, lateralis nerve, with its supra-temporal branch (s.t), and its commissural connexion (c) with _op.s._vii, the ophthalmicus superficialis of vii; viii, auditory nerve. Dotted.—The facialis proper, including c.t, chorda tympani; gn.g, geniculate ganglion; hy, hyomandibularis, with its motor branches m, m, m; _p.n._vii., palatine. Dark grey.—The glossopharyngeal (ix), with its pre- and post-branchial branches and its palatine nerve, _p.n._ix; anastomosing with the palatine branch of vii (Jacobson's anastomosis). White.—The vagus: x^{1-4}, the branchial nerves, ganglionated and forking over clefts ii-v; _v.n._x, visceral nerve; oc, occipito-spinal nerves; d.r and v.r, the dorsal and ventral roots of the first two spinal nerves. (Slightly modified after Wiedersheim.)]

{381}The first six cranial nerves resemble those of the higher Craniates in their mode of origin from the brain, in the physiological nature of their component fibres, and in their peripheral distribution, and therefore they need not be specially referred to here. The principal branches of the fifth or trigeminal nerve are shown in Fig. 218. Comparing this nerve with a typical branchial nerve it would seem that the profundus and superficialis ophthalmic nerves are dorsal nerves; the maxillaris and mandibularis, pre- and post-branchial branches, respectively, in relation with the modified gill-cleft which forms the mouth, while the branch to the oral surface represents a palatine nerve. The most important of the distinctive features in the cranial nerves of Fishes are to be found in the relations of nerves vii., ix., and x. to branchial clefts, and in the lateralis system of nerve components and its association with the lateral line sensory organs. The seventh or facial nerve is an exceptionally interesting nerve. Besides the usual components of a typical branchial nerve certain of its so-called branches are wholly or largely derived from the lateralis system. For this reason the nerve may be said to consist of two portions, the facial proper, or those fibres which constitute the facial nerve in air-breathing Craniates, and the lateralis branches which solely innervate lateral line sense-organs, and are therefore peculiar to aquatic forms. The facial proper has a ganglion (the facial or geniculate ganglion) on its root, and on entering the orbit after traversing the cranial wall it gives off a palatine nerve. Just over the spiracle a pre-branchial nerve, the representative of the chorda tympani of Mammals, leaves the main trunk, and passes ventrally in relation with the anterior wall of the spiracle to its ultimate distribution in the walls of the mouth cavity. The main trunk, now called the ramus hyomandibularis, then pursues a ventral course behind the spiracle as a post-branchial nerve, and certain of its mainly motor branches which pass downwards in connexion with the hyoid arch supply the muscles of that arch, and, if an operculum is present, the opercular muscles as well. The lateralis portion of the facial includes the following principal branches, {382}each of which may have a ganglion on its root: (1) an ophthalmicus superficialis; (2) a buccalis nerve with its ramus oticus; and (3) external mandibular nerves which course in the ramus hyomandibularis. The addition of the great lateralis nerve, which is usually described as the lateral branch of the tenth nerve, and of the eighth or auditory nerve which supplies the auditory organ, completes the enumeration of the main factors of the lateralis system. The ninth or glosso-pharyngeal nerve, perhaps the most typical of all the branchial nerves, has pre- and post-branchial branches which enclose the hyo-branchial cleft. Its palatine nerve usually extends forwards and anastomoses with the corresponding branch of the seventh, thus forming a connexion (Jacobson's anastomosis) between the two cranial nerves. In some Elasmobranchs and Teleosts fibres derived from the dorsal branch of the ninth nerve innervate a few sense-organs of the lateral sensory canal of the head, and hence that nerve sometimes contains lateralis fibres. The tenth or vagus is a compound nerve. Besides the great lateralis nerve generally associated with it, the vagus includes as many typical branchial nerves as there are branchial clefts behind the hyo-branchial cleft, and in Elasmobranchs and in Chimaera these nerves have independent origins from the medulla oblongata. Each nerve has the typical structure, a ganglionated trunk which forks over a gill-cleft into the usual pre- and post-branchial branches, and palatine branches to the pharyngeal walls. In the Dipnoi the lateralis nerve is connected with the superficial ophthalmic branch of the seventh nerve by a commisural nerve which curves across the outer face of the auditory capsule. A somewhat similar anastomosis is also present in Petromyzon. The vagus also includes a large ramus intestinalis, which in Elasmobranchs, at all events, has a distinct ganglionated root. The nerve forms characteristic plexuses on the oesophagus and stomach, and in Cyclostomes its branches may extend nearly the whole length of the intestine. In Ganoids and Teleosts there is an interesting nerve known as the "lateralis accessorius." It is a compound nerve, and owes its formation to the union of somatic sensory fibres derived in succession from dorsal branches of the v., vii., ix., and x. nerves, and also from the corresponding branches of a variable number of spinal nerves. The finer branches of the nerve are distributed to the skin of one or more of the fins, or {383}even, as in Gadus, to all the fins, especially to the numerous "end-buds" which are present on those organs. In many Fishes a variable number of the anterior spinal nerves (spino-occipital) perforate the occipital region of the skull. They probably represent the ventral roots only of the ordinary spinal nerves of this region.

SENSE-ORGANS

THE CUTANEOUS SENSE-ORGANS.—These organs, the most remarkable and certainly the most characteristic of the sense-organs of Cyclostomes and Fishes, are bud-like groups of epidermic cells in relation with the ends of sensory nerve fibres. Each consists of a central core of sensory cells, provided with terminal cuticular sensory hairs, and surrounded by a zone of supporting and mucus-secreting cells which leave the hairs exposed at the apex of the bud. Two kinds of these organs can be distinguished, which differ in their innervation and in their position in the skin. Of the two, the so-called end-buds are the more primitive. They occupy a superficial position in the epidermis, and their sense-cells are as long as the supporting cells. They are present in Cyclostomes and Elasmobranchs, and especially in Teleosts, where they are irregularly distributed over the surface of the body, on the fins, lips, and barbels, and also in the epithelium of the mouth and pharynx. In the Dipnoi they are limited to the oral cavity, and in the higher Craniates they become taste-buds.[450] Their somatic sensory nerves[451] are derived from the vii., ix., and x. cranial nerves, and the lateralis accessorius. In the second type, usually called "nerve-eminences," the sensory cells are shorter than the supporting cells, and they are always innervated by the lateralis system. When first developed in the embryo they are quite superficial, like end-buds, but later the epidermis in which they lie sinks inwards so as to line a series of pits, closed sacs, tubes, open grooves, or closed canals. Pit-organs, so abundant on the head and trunk of Teleosts (Fig. 220), are simple epidermic pits with insunken nerve-eminences, disposed in groups or in {384}lines (accessory lateral lines) or irregularly distributed. The "Spalt-papillen" of Elasmobranchs are pit-organs in which the orifice of the pit is reduced to a slit. The more deeply-seated Savi's vesicles on the ventral surface of the Torpedo, and the nerve-sacs of Ganoids, are similar organs converted into closed sacs and pinched off from the rest of the epidermis. Lorenzini's ampullae or mucus canals, which are found in definitely located groups on the lateral and upper surfaces of the head in Elasmobranchs, may perhaps be compared to pit-organs prolonged inwards to form subcutaneous tubes, each of which terminates in a radially-septate, chambered dilatation or ampulla, containing groups of sensory cells.

Besides the more diffusely scattered sense-organs there are others which become disposed in definite lines along the sides of the body and on the head, and, enclosed in grooves or closed canals, constitute the highly characteristic lateral line system of Cyclostomes and Fishes.[452] The auditory organ must also be included as a specialised portion of this system. Both organs are innervated by the lateralis system, and both arise from a common rudiment in the embryonic epidermis in the position of the future auditory organ. This rudiment grows backwards along the side of the body in the form of a cord of cells differentiated from the epidermis, and also forwards, where it soon divides into the rudiments of future supra-orbital and infra-orbital canals. Sense-organs are differentiated at intervals along the line of the cord; and in the body, but not on the head, they frequently exhibit a segmental disposition. Each sensory organ then sinks down into a short epidermic groove, which by the subsequent meeting of its lips becomes a canal detached from the epidermis. The short canals then become continuous, leaving, however, an externally opening primary pore between every two consecutive canals, and the result is a continuous canal having sense-organs imbedded in its epidermic lining and connected with the exterior by pores at intervals (Fig. 219).[453] The enclosure of the canals in the scales of the lateral line of the trunk or in special drain-pipe ossicles on the head, and the dichotomous subdivision of the primary pores into groups of surface-pores, complete the evolution of the system in its more advanced condition.

{385}[Illustration: FIG. 219.—Vertical longitudinal section through the lateral canal of Amia calva. l.n, Lateralis nerve with its branches, n, n, to the sensory organs, s.o, s.o; p, p, p, external pores; s.c, sensory canal; s, s, scales of the lateral line. (From Wiedersheim, after Allis.)]

Typically, the lateral line system consists of certain canals or grooves, usually but not invariably continuous, and defined by their innervation, (i.) a lateral canal extending along the side of the body and the hinder part of the head, and having its sensory organs supplied by the great lateralis nerve (Fig. 220); (ii.) a supra-orbital canal passing forwards over the eye and innervated by the superficial ophthalmic branch of the facial nerve; (iii.) an infra-orbital canal supplied by the buccalis and otic branches of the same nerve; and (iv.) a hyo-mandibular or operculo-mandibular canal, situated on the outer side of the hyoid region, and thence prolonged downward and forward in relation with the lower jaw, and innervated by the external mandibular branches of the facial nerve. The hyo-mandibular canal is sometimes distinct from the other canals, as in Elasmobranchs and some Teleosts (Fig. 220); and in certain North American Siluroids the same may be said of the supra-orbital. But, as a rule, the infra-orbital is continuous behind both with the lateral and the supra-orbital canals, while the hyo-mandibular canal joins the infra-orbital, or, exceptionally, {386}the supra-orbital canal. Transverse commissural canals often connect the lateral and supra-orbital canals of opposite sides across the dorsal surface of the head, and the corresponding infra-orbital and hyo-mandibular canals may also be continuous at the extremity of the snout or at the mandibular symphysis.

Throughout their extent the canals communicate with the exterior by pores, or short canals terminating in pores, or by branched canals ending in groups of pores. In Cyclostomes[454] the lateral line system is represented by pit-organs disposed as in Fishes, and innervated by a true lateralis nerve. Some Elasmobranchs have the lateral canal of the trunk represented by an open groove protected by marginal denticles. Chimaera is more primitive still in this respect, for on the head as well as on the body the sensory organs are in open grooves. Amongst Fishes these organs are most primitive in the Dipnoi, where they retain their superficial position in the epidermis. In Teleostomes the lateral canals perforate the scales of the lateral line, and at intervals they open externally by simple or multiple pores which perforate the scales. On the head they are more or less completely enclosed in special ossicles which either remain distinct or fuse with certain of the adjacent dermal or cartilage bones of the skull. The use of the lateral line organs is not certainly known. They occur only in Fishes and Amphibia, and as blind Fishes are able to avoid obstacles with the greatest ease when swimming, it is possible that these organs enable their possessors {387}to appreciate undulatory movements in water in the shape of reflex waves from contiguous surfaces or objects.[455] Their great antiquity is shown by their existence in most of the Heterostraci, and in the Antiarchi and Arthrodira, although they have not yet been discovered in the Osteostraci.

THE AUDITORY ORGANS.—In its more typical condition each auditory organ consists of a membranous sac or vestibule, partially constricted into an upper portion or utriculus and a lower or sacculus (Fig. 221, A). Three semicircular canals are connected with the utriculus, of which two are vertical and at right angles to one another, and the third is horizontal. One end of each canal is dilated into an ampulla. A slender tube, the ductus endolymphaticus, leaves the sacculus, and ends in a sac-like swelling, the sinus endolymphaticus, which apparently represents a portion of the embryonic epidermic involution from which the auditory organ is formed. A smaller sac-like outgrowth from the sacculus, the lagena, corresponds to the cochlea of the higher Vertebrates. The epidermic lining of this system of cavities is differentiated into patches or ridges of sense-cells (maculae or cristae), separated by supporting cells and innervated by the terminal branches of the auditory nerve. There is a crista acustica in each ampulla; and maculae acusticae are present in the utriculus, sacculus, and lagena. A fluid, the endolymph, fills all the cavities, and a similar fluid or perilymph occupies the spaces in the periotic capsule in which the various chambers are lodged. Among the more notable deviations from this type of auditory organ the Cyclostome Myxine, apparently, has but a single semicircular canal with an ampulla at each end, and the vestibule is a simple sac (Fig. 221, B). Petromyzon has two canals, but lacks the horizontal canal. In Elasmobranchs, including Chimaera (C), the ductus endolymphaticus retains its primitive connexion with the exterior by means of a pore on the dorsal surface of the head. In the Dipnoi (e.g. Protopterus) the paired endolymphatic sinuses divide into a number of caecal branches containing otoliths, which meet and interlace over the fourth ventricle (Fig. 217).[456] Otoliths, either in the form of fine, {388}mucus-connected, calcareous particles, as in Elasmobranchs, or as massive solid concretions in Teleosts, are present in relation with the sensory areas of the utriculus, sacculus, and lagena.

In a few marine and in a large number of freshwater Teleosts the auditory organ enters into a more or less intimate connexion with the air-bladder by one of three different methods.

{389}[Illustration: FIG. 222.—Cavity of the air-bladder of a Siluroid (Macrones nemurus) exposed by the removal of its ventral wall. a.c, Anterior chamber; b.o, basioccipital; b.w, body wall, here reduced to the external skin; cl, clavicle; l.c, lateral chamber; l.s, longitudinal septum; pt, post-temporal; tr.a, anterior portion of the tripus; tr.c, crescentic portion of the tripus; t.s, transverse septum; t.s', shorter transverse septum. (From Bridge and Haddon.)]

The first and simplest is by the apposition of the extremities of a pair of caecal tubular prolongations from the air-bladder to the outer surfaces of the fibrous membranes which close a pair of vacuities in the outer bony walls of the periotic capsules, the inner surfaces being bathed by the perilymph surrounding the auditory organs. This method is characteristic of certain Serranidae, Berycidae, Sparidae, Gadidae, and Notopteridae,[457] and probably in the Hyodontidae. In the second method, of which several Clupeidae (e.g. Herring, Pilchard, etc.) furnish examples, the periotic vacuities are open instead of closed, and the sac-like ends of the tubular extensions from the air-bladder are in actual contact with protruding outgrowths from the utriculus.[458] The third method, by far the most elaborate, is by the intervention of a series of movably connected "Weberian" ossicles, of which the most posterior on each side (the tripus) is inserted into the dorsal wall of the air-bladder (Fig. 223), while the anterior one (scaphium) forms the outer wall of a median backward prolongation (sinus impar) of the perilymph-containing spaces surrounding the two auditory organs. This in turn encloses a similar median prolongation (sinus endolymphaticus) from the two sub-cerebrally united endolymphatic ducts (Fig. 223).[459] This complex mechanism is present in the Cyprinidae, Siluridae, Characinidae, and Gymnotidae; and hence the term "Ostariophysi"[460] as a collective name for these families.[461] {390}The physiological raison d'être of the connexion between the air-bladder and the auditory organ cannot yet be regarded as satisfactorily determined. It is possible, as Weber thought, that it may be an auxiliary to the function of hearing by transmitting to the ear sound-waves impinging on the surface of the body and affecting the gases in the air-bladder.[462] On the other hand, it may be urged with perhaps greater probability that the connexion exists for the purpose of conveying to the ear stimuli due to the varying degrees of distension of the air-bladder, such as, it may be presumed, are naturally brought about by the variations of hydrostatic pressure which a Fish encounters in the course of its ascent or descent in the water.[463] Whether regarded as an accessory to hearing, or as a means of regulating the distension of the air-bladder, the physiological value of the connexion must be considerable, and on this point it is at least significant that the Weberian mechanism is characteristic of the dominant families of freshwater Teleosts at the present day.[464]

THE OLFACTORY ORGANS.—These organs are essentially a pair of pit-like inpushings of the skin of the ventral side of the head in front of the mouth, with their lining epidermis differentiated into sensory cells separated by supporting cells, and connected with the olfactory lobes of the brain by olfactory nerves.

{391}[Illustration: FIG. 224.—Two stages in the development of the olfactory organ and the pituitary involution in Petromyzon. A is the earlier, B a much later stage. br, Brain; in, infundibulum; l.lp, lower lip; ms, mesenteron; n, notochord; ol.o, olfactory organ; pn, pineal body; pt.s, pituitary sac; st, stomodaeum; u.lp, upper lip. (From Parker and Haswell, after Dohrn.)]

The Cyclostomata are unique amongst Craniates in the apparently unpaired condition of the olfactory organ, and in its remarkable relation to the pituitary involution. In the embryo Lamprey the median and ventral olfactory pit is carried inwards with the pituitary invagination, so that the former appears as a dorsal outgrowth from the latter, and the two have a common external opening, the naso-pituitary aperture (Fig. 224). Later the extraordinary forward growth of the upper lip to form the roof of the buccal funnel has the effect of shifting the naso-pituitary involution and its aperture to a final position on the dorsal side of the head. It is due to this dorsal displacement that, as we shall see, the pituitary caecum reaches the ventral surface of the brain by perforating the basis cranii from above, instead of from below as in all other Craniates. The pituitary body is pinched off from the dorsal side of the naso-pituitary involution. In the adult Lamprey the olfactory organ appears as a round sac divided by a median septum into two lateral chambers (Fig. 225), the lining epithelium of which is raised into prominent ridges. Behind the sac the pituitary involution is prolonged backwards beneath the brain, and, after traversing the basi-cranial fontanelle, it widens out into a spacious cul-de-sac and terminates on the dorsal side of the pharynx, beneath the anterior end of the notochord. In Myxine the pituitary involution ends by opening into the pharynx.

{392}[Illustration: FIG. 225.—Side view of the brain of Petromyzon, with the olfactory organ and the pituitary caecum in section. cblm, Cerebellum; crb.h, cerebral hemisphere; dien, thalamencephalon; f, fold in the nasal tube; gl, nasal glands; inf, infundibulum; l.gn.hb, left ganglion habenulae; med.obl, medulla oblongata; na.ap, naso-pituitary aperture; n.ch, notochord; Nv^1-nv^{10}, cranial nerves; Nv^{12}, first ventral spinal nerve; olf.cp, olfactory capsule; olf.l, olfactory lobe; olf.m.m, olfactory mucous membrane; opt.l, optic lobe; pn, pineal body; pn′, inferior pineal body; pn.e, parietal eye; pty.b, pituitary body; pty.p, pituitary cul-de-sac; sp, median septum of the olfactory sac; sp^1, first dorsal spinal nerve. (From Parker and Haswell, after Ahlborn and Kaenische.)]

The apparently monorhinal condition of the Cyclostomes is probably a secondary acquisition. At the earliest embryonic stage at which any trace of an olfactory organ is apparent, there is a median thickening of the epidermis, possibly a vestige of some older sensory organ comparable, it may be, to the so-called olfactory organ of Amphioxus; on each side of it there is a lateral thickening, the rudiments of the paired organs.[465] The three thickenings, or "plakodes," then sink inwards to form an olfactory pit. The partial subdivision of the adult organ by a vertical septum, and the presence of two olfactory nerves, point to the same conclusion.[466] All Fishes possess olfactory organs which are obviously paired. In Elasmobranchs and Dipnoi they retain their primitive ventral position. Many Sharks and Dog-Fishes possess an oro-nasal groove leading from each olfactory organ to the corresponding angle of the mouth. The Dipnoi proceed a stage farther, and, by the conversion of the grooves into short canals, the olfactory pits communicate with the mouth by true internal nostrils, as in the higher {393}Vertebrates. In the adults of existing Teleostomi the orifice of each organ is usually divided into two by the growth of a fold of skin across it, and the two apertures rotate outwards and upwards on to the lateral or the upper surface of the snout. Of the two nostrils the posterior one probably corresponds to an external nostril, and the anterior one to the internal nostril. Occasionally each olfactory organ has only a single orifice. In the Crossopterygii and in some Teleostei the nostrils become tubular. The lining epithelium of the olfactory pits is usually produced into ridges, disposed longitudinally or transversely, or in the form of radii from a central point in the roof. Many Teleosts have each olfactory organ prolonged backwards into one or two sacs, the nasal sacs, which are either simple reservoirs, or glandular and mucus-secreting. In a species of Chinese Sole (Cynoglossus semilaevis) the two sacs, one from each olfactory organ, unite over the roof of the mouth in a common median sac, and in one unique specimen the latter communicated with the mouth by a large naso-pharyngeal aperture.[467]

THE EYES.—In essential structure the eyes of Cyclostomes and Fishes resemble those of the higher Craniates. As a rule, in Fishes they are relatively larger, however, and the lens is globular and the cornea somewhat flatter. Ciliary processes and ciliary muscles are absent. As the eyes are nearly always lateral in position it is probable that monocular vision is the rule. In Teleosts and in Amia a "choroid gland," consisting of a mass of capillary blood-vessels, surrounds the optic nerve externally to the retina, and derives its blood from the efferent artery of the pseudobranch (Fig. 226). In most Teleostomi, but not in Cyclostomes, Elasmobranchs, and Dipnoi, there is a singular prolongation of the choroid coat, known as the "processus falciformis," which extends across the vitreous humour to the inner face of the lens, where it ends in an expansion, the "campanula Halleri" (Fig. 226). Accommodation to vision at different distances is not effected by alterations in the convexity of the lens, but by a change in its position with regard to the retina, apparently brought about by the contraction of a special retractor muscle.[468] Some oceanic pelagic Teleosts are remarkable for their curious telescopic eyes in the shape of short protruding {394}cylinders, each terminating in a strongly convex cornea (Fig. 227).[469] The eyes are directed either upwards or forwards, and, as their long axes are parallel in either position, it is probable that these Fishes are capable of binocular vision. In the young of certain Teleosts occurring in the Antarctic and Indian Oceans the large eyes are situated at the extremities of extraordinary long stalks extending from the sides of the head.

In the quasi-parasitic Cyclostome, Myxine, and in many Teleosts belonging to widely different families, which live at great depths in the sea or inhabit subterranean waters, the eyes suffer from disuse and degenerate in structure. The influence of a deep-sea habitat on the eyes of Fishes is somewhat varied. The eyes are often small. A few abyssal Fishes are totally blind, and no external trace of eyes can be seen (Fig. 430). In such Fishes compensation is often afforded by an extraordinary development of tactile organs in the form of long barbels, or of trailing filaments derived from the median or the paired fins (Fig. 371, B). Many deep-sea forms possess eyes of the normal size, or even exceptionally large eyes, probably because either they occasionally migrate towards the surface, or else they possess phosphorescent organs and are able to see by the aid of the light they themselves emit. A blind Siluroid (Amiurus nigrilabris) frequents the cave streams of Pennsylvania, and many members of the same family which live in muddy waters have very small or even minute eyes. One of the Gobies (Typhlogobius),[470] which buries itself in the sand, or is found under stones in the holes of a burrowing Crab on the coast of California, is also {395}blind. Amongst other blind Fishes Amblyopsis and Typhlichthys (Amblyopsidae)[471] and Lucifuga (Zoarcidae) may be mentioned, the first two inhabiting the cave streams of North America, while the third has a similar habitat in Cuba. When the eyes degenerate they dwindle in size and recede from the surface. The lens and the iris wholly or partially disappear, and although it is generally recognisable the retina loses certain of its characteristic layers, or the latter are but imperfectly formed. In Myxine even the eye-muscles are absent.

The eyelids of Fishes are little more than marginal folds of skin, capable of little if any movement, and leave the eyes largely uncovered. Some Sharks have a third eyelid or "membrana nictitans" at the anterior corner of the eye. Lachrymal glands are unknown.

THE PARIETAL EYE.—It is only in the Cyclostomes that this structure can have any claim to be regarded as a visual organ. In the Lamprey (Fig. 228) the parietal eye is a slightly flattened vesicle lying directly over the pineal vesicle, and connected by a slender stalk or nerve with the right ganglion habenulae. The dorsal or more external half of the vesicle is bi-convex, and forms the "pellucida," while the inner half or retina is said to consist of supporting cells with interspersed deeply pigmented sense-cells and ganglion cells.[472] The external skin over the parietal eye is partially transparent in the living animal.

In many of the oldest known Fishes, such as the Ostracodermi, the Antiarchi, and the Crossopterygian Osteolepida, there are indications of the existence either of one or of two median sense-organs on the upper surface of the skull, in the shape of one or two foramina, or hollow protuberances, or pit-like grooves or {396}depressions, but, as a rule, when one of them is present the other is absent. It is probable that both these structures were associated with sensory organs, of which one may have been a parietal eye and the other a pineal eye. Some Teleosts (e.g. many deep-sea Scopelidae) have a transparent, convex, cornea-like prominence on the upper surface of the head which may be related to one of these singular organs.[473]

{397}CHAPTER XV

THE KIDNEYS AND THE REPRODUCTIVE ORGANS—BREEDING

The kidneys and the reproductive organs are so intimately connected that it is necessary to deal with them together. Both organs are specialised portions of the coelom and its epithelial lining. The KIDNEYS are essentially a series of tubular and at first segmentally-disposed outgrowths from the coelom (urocoeles) which acquire a connexion with the exterior, while the gonads have their origin from local modifications of the coelomic epithelium. At a very early embryonic stage each lateral half of the coelom presents three well-marked divisions: (1) a series of dorsal portions ("myocoeles"), the cavities of the myotomes or muscle-segments; (2) a longitudinally continuous unsegmented portion extending round the alimentary canal, the "ventral coelom"; and (3) a series of intermediate tubular portions or "nephrotomes," each of which leads from a myocoele to the ventral coelom (Fig. 229, A). The essential components of the kidneys, the urocoeles or renal tubules, are derived from the nephrotomes. In its typical condition each kidney consists of three portions, which, in accordance with their embryological and evolutionary sequence, are termed the "pronephros," the "mesonephros," and the "metanephros." The pronephros, the larval or provisional kidney, is formed from a limited number of the nephrotomes immediately behind the head. From each nephrotome a hollow tubular outgrowth is formed, which grows towards the lateral surface of the body, and then unites with its fellows of the same side to form a main longitudinal duct—the "archinephric" or "pronephric duct" (Fig. 229, A, Fig. 230, A). This duct grows backwards until it opens into the cloaca.[474]

{398}[Illustration: FIG. 229.—Diagrammatic transverse sections through an embryo Craniate to show the mode of development of the pronephros (A) and of the mesonephros (B). The right side of each figure shows an earlier stage than the left. In B (left side) the connexion of a vas efferens with a mesonephric tubule, and the division of the archinephric duct into Müllerian and mesonephric ducts are shown, a, Aorta; a.c, alimentary canal; a.d, archinephric duct; g, glomus; gl, glomerulus; i.n, inner nephrostome; mb, Malpighian body; md, Müllerian duct; mnd, mesonephric duct; mnt, mesonephric tubule; myc, myocoele; myt, myotome; n, notochord; np, nephrotome; nt, nephrostome; o.n, outer nephrostome; pn.t, pronephric tubule; s.c, spinal cord; t, testes; v.c, ventral coelom; v.ef, vas efferens. (After Kingsley and Semon.)]

At the same time the nephrotomes lose their connexion with the myocoeles, although they still retain their "nephrostomes" or apertures through which they communicate with the ventral coelom. When fully developed the pronephros consists of a few tubules, more or less convoluted, opening at their inner extremities into the coelom by means of their ciliated nephrostomes, and at their outer ends communicating with the exterior through the archinephric duct. In relation with the pronephros a branch from the dorsal aorta forms a tuft of capillary blood-vessels or "glomus," opposite the nephrostomes, which projects into the ventral coelom on each side. Later, a second series of much more numerous tubules is formed behind the pronephros, which constitute the mesonephros. In forming mesonephric tubules the nephrotomes become disconnected from the myotomes and their myocoeles, and curving outwards they {399}come to open into the archinephric duct, although they do not in any way contribute to its formation (Fig. 229, B). Segmentally-arranged twigs from the dorsal aorta end in tufts of capillaries or glomeruli, each of which projects into a small sac-like enlargement of a mesonephric tubule, pushing before it the wall of the sac. In this way a double-walled "Malpighian body," containing a "glomerulus," is formed in connexion with each tubule. Subsequently, the mesonephric tubules increase in number by budding. New nephrostomes and Malpighian bodies are developed on the secondary branches, and the original segmental arrangement of the tubules becomes obscured. With the growth of new tubules, and the formation of blood-vessels and of connective and lymphoid tissues between them, each mesonephros finally assumes the condition of a compact gland imbedded in the dorsal wall of the coelom, with its ventral surface invested by the peritoneum. A "metanephros," which in the higher Vertebrates replaces the mesonephros as the functional kidney, is perhaps not represented in Fishes.

A more or less well-developed pronephros is present in the embryos or larvae of the Cyclostomes and of all Fishes, but as a rule it completely disappears at an early period and is replaced by the mesonephros. It is retained throughout life, however, in the Myxinoid Cyclostomes (Fig. 230, B), and has its persistent nephrostomes opening into the pericardial cavity.[475] In a few Teleosts the pronephros is also persistent, as in Fierasfer and Dactylopterus, and in others the organ may not completely disappear until the approach of sexual maturity. But with these exceptions the mesonephros is the sole functional kidney in the adults of the Cyclostomes and of all Fishes. As regards the nature of the duct by which the excretion of the mesonephros is conveyed outwards, there are notable differences in different Craniates. The Cyclostomes and the Teleostomi retain that part of the archinephric duct into which the mesonephric tubules open, and which remains after the atrophy of the pronephros (Fig. 230, B, E, F). In Elasmobranchs, and probably also in the Dipnoi, a special mesonephric duct is developed in a way which will be described later (Fig. 230, C, D).

{400}[Illustration: FIG. 230.—Showing the principal modifications of the kidneys and reproductive organs in Cyclostomes and Fishes. A, The pronephros and its duct in the embryo; B, the kidneys and genital pores in Petromyzon, the vestigial pronephros represented as in Myxine; C and D, the urinogenital organs of a male and female Elasmobranch; E, of a male or female Teleost, or a male Lepidosteus; F, of a female Polypterus, Acipenser, Amia, or Osmerus. a, Anus; a.d, archinephric duct; c, cloaca; c.a, the coelomic aperture of the Müllerian duct; c.p, cutaneous pit; g, gonad; gd, gonoduct; g.p, genital pore; i, intestine; m, Malpighian body; m.d, Müllerian duct; mn, mesonephros; mn^1, vestigial mesonephros; mn^2, excretory portion of the mesonephros ("metanephros"); mn^3, genital portion of mesonephros; mn.d, mesonephric duct; mtn.d, metanephric duct; n, nephrostome; ov, ovary; p.a, abdominal pore; p.f, peritoneal funnel; pn, pronephros; pn′, vestigial pronephros; s.g, shell gland; s.s, sperm sac; t, testis; ug.s, u.s, urinogenital sinus; v.ef, vasa efferentia; v.s, vesicula seminalis.]

{401}In the males of Elasmobranchs some of the hinder mesonephric tubules unite to form a single main duct opening into the terminal part of the mesonephric duct, and these tubules and their separate duct are sometimes regarded as a metanephros and a metanephric duct. The mesonephric nephrostomes are persistent throughout life in a few Elasmobranchs (e.g. Notidanidae, Heterodontidae, Rhinidae, and some Scylliidae), and also in Amia:[476] in all other Fishes as well as in the Cyclostomes they become closed in early life.

In many Fishes the hinder extremity of the coelom communicates directly with the exterior through "abdominal pores," of which there is usually a pair, rarely a single pore, situated close to the cloacal or the anal aperture.[477] Elasmobranchs usually have a pair, often at the extremities of a pair of cloacal papillae (Fig. 231), but they are absent in some families (e.g. Heterodontidae and Rhinidae); and in some Scylliidae (e.g. Scyllium canicula) they are very variable, being either present or absent on both sides, or an open pore is present on one side only. Pores are present and paired in the Crossopterygii, the Chondrostei, and the Holostei. Amongst the Dipnoi Neoceratodus has a pair of pores. Protopterus sometimes has two pores opening into the cloaca, but as a rule the two become confluent and have a single external aperture. In Lepidosiren pores are wanting. Abdominal pores are rarely present in Teleostei. They exist, however, in the Mormyridae (Gymnarchus and several species of Mormyrus), and also in the {402}Salmonidae,[478] where they are as singularly variable in different species and individuals as in the Elasmobranch Scylliidae. The use of abdominal pores is not certainly known, unless the coelom of those Fishes which possess them continues to retain some measure of its primitive excretory function, and the pores act as excretory ducts. That the nephrostomes are excretory organs has been shown by experiment, and it is worthy of note that there exists a reciprocal relation between these structures and abdominal pores, to the extent that while there are a few Fishes (e.g. certain Elasmobranchs and Amia) in which both coexist, there are many others in which the presence of nephrostomes is correlated with the absence of pores and vice versâ.

The male and female GONADS, testes and ovaries, are derived from the coelomic epithelium near the inner or median aspect of the nephrotomes (Fig. 229, B). Here the epithelium remains columnar, and soon projects into the ventral coelom as a continuous longitudinal ridge. It is probable that at first the modified epithelium is segmented as a series of "gonotomes," but if so, the latter must soon coalesce into a continuous ridge. Some of the epithelial cells enlarge to form the primitive sex-cells. In the development of an ovary, portions of the epithelium sink inwards, carrying with them the primitive ova. Certain of the cells form the epithelial walls of a number of ovisacs, each of which encloses an ovum. As the ovisacs increase in number and size the germinal ridges project more and more into the coelom until, as ripe ovaries, they become suspended from its dorsal wall by a double peritoneal fold, the "mesovarium" (Fig. 156). The testes develop in a similar fashion except that the primitive sex-cells, which later give rise to spermatozoa, form the lining of a number of simple or ampulla-like tubules, the seminiferous tubules, and the suspensory fold is termed the "mesorchium."

The Cyclostomes have gonads in the shape of unpaired organs extending nearly the whole length of the coelom, but in all Fishes the organs are primarily paired, although by fusion, or by the absorption of one gonad, the ovaries or the testes sometimes appear as if single. The ovaries may either be naked, as in Elasmobranchs, Dipnoi, Crossopterygii, and Chondrostei, and in Amia amongst the Holostei; or, as in Lepidosteus and most Teleosts, they become enclosed in coelomic sacs. The {403}former, or "gymnoarian," condition is primitive; the latter, or "cystoarian," is secondary, and is brought about by the growth of two peritoneal folds round the ovary and the union of their margins. Into these coelomic sacs the egg-bearing or real ovarian tissue projects either in the form of processes or of transversely- or longitudinally-arranged plates or folds (Fig. 232, B). The testes are composed of seminal ampullae, as in Elasmobranchs, or of radially-arranged and sometimes plexiform tubules opening into the gonoduct, as in nearly all other Fishes (Fig. 232, A).

In the Cyclostomes (e.g. Petromyzon) the eggs and spermatozoa are discharged from the gonads into the coelom, whence they reach the exterior through a pair of "genital pores" leading from the hinder end of the coelom into a urinogenital sinus formed by the united extremities of the two archinephric ducts.[479] Myxine has, however, but a single median pore, opening into an integumentary cloaca, which also receives the rectal and urinary orifices. Bdellostoma has two such pores communicating with a similar cloaca.[480]

{404}The nature and homologies of the genital ducts in the different groups of Fishes are amongst the most puzzling of the many problems which vex the soul of the Vertebrate morphologist, and although there is a fairly general agreement on some points, there are others of great importance of which it may be said quot homines, tot sententiae.

Broadly speaking, there are two types of genital ducts in Fishes: (1) those which are obviously derived from some part of the kidney system; and (2) those which are special ducts and appear to have no connexion with kidney-ducts.

The Elasmobranchs offer a typical example of gonoducts of the first kind. At an early embryonic period in both sexes each archinephric duct becomes longitudinally split into two ducts, of which one continues to receive the openings of the mesonephric tubules and remains as a mesonephric duct (Fig. 229, B).[481] The other, which has no connexion with the mesonephros, opens anteriorly into the coelom by means of the united nephrostomes of the pronephros, and is known as the "Müllerian duct" (Fig. 230, C and D). In the adult male the Müllerian ducts are useless vestiges, but in the female they persist and act as oviducts, receiving the eggs set free from the ovarian ovisacs through their coelomic apertures, and thence conveying them to the cloaca. In the male, certain of the anterior mesonephric tubules become connected with the testicular ampullae by means of a network of slender tubules, the "vasa efferentia" or testicular network, and through the latter the spermatozoa pass from the testes to the mesonephric duct (Fig. 230, C). Consequently, the mesonephric duct conveys both spermatozoa and the kidney excretion to the cloaca. It is obvious, therefore, that both the male and female gonoducts are derived from kidney-ducts.

The Teleostei afford an equally typical illustration of the second type. Each female gonoduct (oviduct) is formed by a backward growth of the same two peritoneal folds which enclose the ovary; these are converted into a "peritoneal tube" or canal by the union of their margins. The male gonoducts are also formed in continuity with the testes, that is, as backward prolongations from the latter. Each duct, male or {405}female, seems to be a duct sui generis and to have no connexion whatever with the kidney system (Fig. 230, E). In the Salmonidae, Anguillidae, Galaxiidae, Hyodontidae, Notopteridae, and Osteoglossidae, and also in Misgurnus, the oviducts lose their continuity with ovaries and degenerate to an extent which differs greatly in different families. Thus in some Salmonidae, as in the Smelt (Osmerus eperlanus),[482] the oviducts end anteriorly in wide funnel-like coelomic apertures after the fashion of Müllerian ducts, and do not embrace the ovaries: hence the ovaries are naked and not cystoarian, and their ducts are not peritoneal tubes but "peritoneal funnels" (Fig. 230, F). In other Salmonidae and in the Anguillidae the oviducts appear to have so far degenerated that they are represented either by a pair of very short funnels or by a pair of genital pores, which, as in the Salmon, have a common external aperture behind the anus and in front of the single orifice of the united archinephric ducts (Fig. 233, A). In all such instances the eggs are set free from the ovaries into the coelom, from whence they escape through the peritoneal funnels or genital pores. In the Eels the male gonoducts also degenerate, and, losing all connexion with the testes, they become reduced to genital pores as in the female.

The Holocephali and probably the Dipnoi conform to the Elasmobranch type in the nature of their male and female gonoducts. In the Crossopterygii[483] each testis has its own proper duct, which has no connexion with the kidney system and apparently belongs to the Teleostean type, while the oviduct, which is almost certainly not a Müllerian duct, is probably a peritoneal funnel. On the other hand, the Chondrostei and the Holostei are in the interesting transitional condition of possessing male ducts of the Elasmobranch type and female ducts of the Teleostean type, the latter being either ducts directly continuous with the ovaries, as in Lepidosteus, or of the nature of peritoneal funnels, as in Acipenser, Polyodon, and Amia (Fig. 230, E and F).

How far the distinction between the two types of gonoduct holds good in the case of the male is not quite clear, and it has recently been argued that the Dipnoi offer a connecting link between the two.[484]

{406}[Illustration: FIG. 233.—Diagram to show the kidneys and gonoducts of a female Salmon (A), and of a male Protopterus (B). md^1 and md^2, Anterior and posterior vestiges of the Müllerian duct; t.t, tubular posterior portion of the testis (t). Other reference letters as in Fig. 230. (B, after Graham Kerr.)]

In Protopterus each testis is divided into an anterior sperm-producing part and a posterior tubular portion which has lost the capacity of producing sex-cells. The testicular network is greatly reduced, and forms but a limited connexion between the tubular portion of the testes and the mesonephric duct (Fig. 233, B). If it be supposed that the testicular network became still further reduced so that the connexion between the testes and the kidney-duct took place directly through a single channel instead of through several, the result would be a gonoduct essentially similar to the male duct of an ordinary Teleost. Should this view prove to be correct, it will follow that the male gonoducts of all Fishes are differently-modified examples of the Elasmobranch type. But there will still remain the female gonoducts of Ganoids and Teleosts, which must be regarded as distinct from Müllerian ducts unless it can be shown that their different methods of development are not necessarily fatal to their homology with Müllerian ducts, or that both types of gonoduct can be derived from some intermediate type. Assuming that some Fishes do possess male or female ducts which have not been derived from the kidney system, but have been independently acquired, there is still the question, which of the two types is the more primitive, or, in other words, has the Elasmobranch type superseded the Teleostean, or vice {407}versa? To this question no decisive answer can at present be given.

The terminal relations of the kidney-ducts and the gonoducts, and the presence of accessory or of vestigial organs in connexion with them, will now be briefly dealt with. In the males of the Elasmobranchs the mesonephric ducts which, as already pointed out, act both as kidney-ducts and gonoducts, dilate posteriorly to form a pair of vesiculae seminales, and then unite to form a urinogenital sinus, opening into the cloaca at the extremity of a median papilla (Fig. 230, C). The sinus also receives ducts from the hinder part of the mesonephros, either separately, as in the female, or by a common duct on each side—the so-called metanephric duct—as in the male. Two tubular caecal outgrowths from the sinus form two sperm sacs. Only the anterior portions of the Müllerian ducts with their coelomic apertures are retained in the adult. In the female the mesonephric ducts are purely excretory, but otherwise they are similar, and the oviducts (Müllerian ducts) open into the cloaca separately or by a common orifice (Fig. 230, D). A glandular dilatation of each oviduct forms the oviducal or shell gland by which the horny egg-cases are secreted. In the males of the Holocephali the gonoducts open into a urinogenital sinus with an external orifice distinct from and behind the anus; but the female has separate apertures for the rectum, the conjoined oviducts, and the united mesonephric ducts. Both sexes have complete Müllerian ducts communicating with the coelom in front, and behind with the exterior. The Dipnoi of both sexes essentially resemble the Elasmobranchs in the general relations of their ducts, but the Müllerian ducts of the male exhibit marked differences in the three genera.[485] In Neoceratodus the ducts are as complete as their functional representatives in the female. Protopterus retains anterior vestiges and the coelomic apertures, and also vestiges of the hinder portions which unite and end blindly in the urinogenital papilla, but the middle sections of the two ducts are suppressed (Fig. 233, B). In the Teleostomi there is a general similarity in the terminal relations of the gonoducts and kidney-ducts. In the Ganoids the archinephric ducts unite and then expand into a urinary sinus or bladder, and the gonoducts of the female, or of both sexes in Lepidosteus, open either into the archinephric {408}ducts or into the common sinus, and therefore both ducts communicate with the exterior by a urinogenital orifice behind the anus. Peritoneal funnels, similar to the functional oviducts of the female, are present in the males of the Chrondrostei and of Amia. In Teleosts the terminal connexions of the ducts tend to become less intimate. The archinephric ducts often dilate into a urinary bladder either before or after their union, and the common duct joins the united gonoducts to form a short urinogenital sinus which opens externally, or the confluent gonoducts have an independent genital orifice between the anus and the urinary aperture. Not rarely the genital or the urinogenital orifice is prolonged into a tubular papilla, which in the male acts as an intromittent organ, or, as in the females of the Cyprinoid Rhodeus amarus, the long oviducal tube serves the purpose of an ovipositor. The males and females of the Siluroid Plotosus have a remarkable vascular and glandular arborescent appendage just behind the urinogenital papilla, the use of which is unknown.[486]

The EGGS of different Fishes[487] exhibit considerable diversity in size and shape as well as in the nature of their external coverings and their mode of deposition.[488] The size of the eggs largely depends on the quantity of food-yolk stored up in their substance for the nutrition of the embryo: hence the eggs of Elasmobranchs, which resemble Fowls' eggs in the superabundance of their yolk, are by far the largest. Teleostomi have much smaller eggs. The largest Teleostean ova are those which are heavy and sink (demersal ova); the smallest, those which are buoyant and float (pelagic ova). Of the former, the eggs of Gymnarchus are about 10 mm. in diameter; those of the Salmon about 5 mm.; and those of some species of Arius, 5 to 10 mm. The eggs of the Wolf-Fish (Anarrhichas lupus) are about 6 mm. Smaller demersal ova are those of the Lump-sucker (Cyclopterus) and Heterotis, which are 2.6 and 2.5 mm. respectively. Pelagic {409}eggs are very small, those of the Plaice, which are exceptionally large, varying from 1.65 to 1.95 mm.

An egg-cell consists of living protoplasm and a nucleus, a variable quantity of non-living food-yolk, and of certain enveloping and protective egg-membranes. The ova of Fishes differ principally in the amount and disposition of the food-yolk, in the character of the egg-membranes, and in the presence or absence of special perforations in the egg-membranes for the entrance of spermatozoa into the eggs. In the small ova of some of the lower Chordata (e.g. Amphioxus), where the very small quantity of food-yolk is uniformly distributed, and its presence affects all parts of the egg alike, the process of segmentation which follows fertilisation results in the transformation of the entire egg into a mass of approximately equal-sized cells or blastomeres (Fig. 82). {410}The eggs are therefore described as "alecithal," and the segmentation as being "holoblastic" and "equal." On the other hand, all Fishes possess "telolecithal" eggs, that is, ova in which the food-yolk is more or less abundant, and tends to accumulate at one pole of the egg ("vegetative pole"), while the opposite or "animal pole" consists of protoplasm, comparatively free from yolk granules and containing the nucleus (Fig. 234, A). The term telolecithal is, however, a somewhat comprehensive one, and covers important variations in the relations of the inert food-yolk and the living protoplasm in different Fishes, which greatly modify the process of segmentation. Thus there are some Fishes in which the amount of food-yolk at the vegetative pole is sufficient to retard segmentation in that part of the egg without actually preventing it, and consequently segmentation begins in the animal pole, and takes place more rapidly there than it does when it extends into the vegetative pole. Hence it follows that although the entire egg is segmented the blastomeres are of unequal size, the animal pole giving rise to a large number of small cells or micromeres, and the vegetative pole to a smaller number of much larger cells or macromeres. The segmentation of such an egg is said to be holoblastic but unequal (Fig. 234, B and C). This type of egg is characteristic of the Chondrostei, the Holostei, and the Dipnoi. In other Fishes, like the Elasmobranchs and the Teleostei, the food-yolk so greatly preponderates that it entirely prevents segmentation in the vegetative part of the egg, and segmentation is restricted to the small mass of protoplasm (germinal disc) at the animal pole, in which the nucleus is situated (Fig. 234, D). Eggs undergoing partial segmentation in this way are termed "meroblastic." No hard and fast line can be drawn between the two types, and in the Chondrostei and Holostei an interesting transition between the holoblastic and meroblastic ova may be observed. The egg-membranes are formed either by the egg itself or by the epithelium of the ovarian ovisacs, and, as will shortly be seen, the character of the outer egg-membrane greatly influences the mode of deposition of the eggs and their location afterwards. In Elasmobranchs the egg is enclosed in a stout horny egg-shell, secreted by the oviducal shell gland.[489] In many Fishes, as in the Chondrostei, Holostei, and Teleostei, the egg-membranes {411}are perforated at the animal pole of the egg by a small aperture or "micropyle," which is only large enough to admit of the entrance of a single spermatozoon at a time (Fig. 235). Generally, there is only a single micropyle, but, according to Salensky, the Sturgeon (A. sturio) has from 3 to 9, and the Sterlet (A. ruthenus) from 5 to 13.

An important distinction may be made between the ova of different Teleostomi as regards their location after extrusion from the female. From this point of view two types of ova can be distinguished, demersal and pelagic ova. Demersal eggs are characterised by their larger size and greater weight, so that they always sink after extrusion; and by their opacity. They may either have an outer egg-membrane which is viscid and adhesive, so that the eggs readily adhere to one another or to foreign objects, or the membrane is smooth and non-adhesive. The Salmonidae, for example, produce non-adhesive demersal eggs, which remain separate after being deposited on the gravelly bed of a stream. Most freshwater and many marine shore Fishes have adhesive demersal eggs, which are deposited at the bottom of the water, generally adhering to one another in larger or smaller clumps, masses, or sheets, and attached to rocks, stones, or empty shells, like the eggs of many shore Fishes, or to aquatic plants after the fashion of the eggs of the Carp, Perch, and Pike, or even to branching zoophytes, as is the case with the eggs of the Sea-snail (Liparis). In some adhesive eggs the external egg-membrane forms threads for their attachment. The eggs of the Gar-Fish (Belone), and those of the Saury Pike (Scombresox) and of the Flying Fishes (Exocoetus), have viscid threads developed from opposite points on the surface, which are either attached to foreign objects or they become entangled with those of other eggs of the same species. The oval eggs of some of the Gobies have a bunch of fibres at one pole which serves to attach them. In the Smelt (Osmerus eperlanus) a portion of the outer egg-membrane breaks away from the rest and becomes turned back, inside out, but remains attached to the egg at one point. By means of this membrane the egg is attached to rocks or stones. Pelagic eggs are distinguished by their lightness and buoyancy, so that they always float near the surface of the water, and by their smaller size and remarkable transparency (Fig. 235). A conspicuous feature in many of them is the presence of a single {412}large oil globule on the surface of the yolk, and not infrequently the yolk becomes partially or completely broken up into small masses. Pelagic eggs are always non-adhesive and free, and they invariably belong to marine Fishes. Amongst the British food Fishes which produce pelagic ova may be mentioned the Gadidae (e.g. Cod, Whiting, Hake, Ling), the Pleuronectidae (e.g. Turbot, Brill, Sole, Plaice), Scombridae (e.g. Mackerel), Triglidae (e.g. the Gurnards), Percidae (e.g. the Bass), and Clupeidae like the Pilchard and Sprat, but not the Herring, whose adhesive demersal eggs are deposited in clumps on shingly banks in the sea at varying distances from the shore.

The eggs of Elasmobranchs are deposited singly or in pairs at considerable intervals, and the period of egg-laying is prolonged over a considerable part of the year. In most other Fishes, as in Teleosts, the period of spawning is limited to a few months, usually in the spring and summer in temperate latitudes; and in the case of a single Fish it may last only a few days or weeks, but the number of eggs produced is often enormous. Thus, in a Ling 61 inches long and weighing 54 pounds the ovaries contained 28,361,000 eggs. A Turbot, 17 pounds in weight, had 9,161,000 eggs; and a Cod of 21½ pounds 6,652,000. The least prolific of the British food Fishes is the Herring, in which the number of ovarian eggs varied from 21,000 to 47,000 in four specimens examined.[490] The extraordinary fecundity of many Fishes seems to bear no relation to the relative abundance of the Fishes themselves, but rather it is to be associated with certain {413}disadvantages attendant on the sexual relations of Fishes, involving a considerable waste of the sex-cells, while in many Fishes it no doubt helps to compensate for any subsequent mortality among the larvae, which may result from an uncertain and precarious food supply and from the attacks of enemies. Whenever internal fertilisation is the rule, or when, as in nest-building and marsupial Fishes, the propinquity of the sexes in the breeding season ensures the fertilisation of a larger proportion of the eggs and the protection of the young, the number of eggs produced is small.

The male sex-cells or spermatozoa are essentially similar to those of other Vertebrates, although in different Fishes they may vary in such details as length, and the shape and size of the head, which may be rod-like and wavy, elliptical or globular.

As a rule, in Fishes females are more numerous than males, and generally they are larger, but to both statements there are notable exceptions. The relations of the sexes in the breeding season are usually very promiscuous, especially in those Teleosts which discharge their sex-cells while swimming together in shoals. A female may, however, consort with several males (polyandry), or a male with several females (polygamy); or, as in some of the nest-building Fishes (e.g. Gastrosteus), there are not wanting examples of the pairing of one male with one female (monogamy).

Fishes often migrate at the commencement of the breeding-season to localities most suitable for the deposition of the eggs. Many marine species seek banks or shallower water near the shore, and some, like the Salmon and the Sturgeon, are anadromous, and ascend rivers for long distances to deposit their spawn.

In all Fishes except the Elasmobranchs and a few Teleosts the fertilisation of the eggs takes place in the water after their extrusion, the male depositing its seminal fluid over the eggs or in their neighbourhood. The waste of the sex-cells is often, no doubt, very considerable, especially when the eggs are adhesive and fixed, and the seminal fluid is liable to drift at the mercy of tides and currents. With pelagic ova the waste is perhaps not so great, inasmuch as the eggs as well as the spermatozoa would probably drift at the same rate and in the same direction. Liability to waste must also be greatly diminished in many Fishes by their habit of living in shoals, or of congregating {414}together in the breeding season, in which they are sometimes aided by their power of emitting characteristic sounds, and in the case of nest-building Fishes by the still more intimate relations of the sexes. Even when the liability to waste is very great, compensation may be afforded by exceptional fecundity. The copulation of the sexes and the internal fertilisation of the eggs occur only in Elasmobranchs and some Teleosts. The copulatory organs of Elasmobranchs are the so-called "claspers" with which the males are provided. Some form of copulation is probably the rule in the viviparous Teleosts, where the eggs are fertilised in the oviducts, or even while they are still in the ovaries, and the young are born alive. As mentioned above, an intromittent organ is often formed by the prolongation of the genital or the urinogenital orifice into a papilla, or a longer or shorter tube.[491] Some Cyprinodontidae[492] (e.g. Anableps) have the anterior part of the anal fin modified in the male to form an intromittent organ along which the urinogenital canal runs (Fig. 374). In the females the genital aperture is covered by a special scale, which is free on one side and not on the other. "The male organ in some individuals is turned to the right, in others to the left, and in some females the opening beneath the special scale is to the right, in others to the left. Copulation thus takes place sideways, a left-sided male pairing with a right-sided female, and vice versa."[493] The anal fin also forms an intromittent organ in the "Half-beak" (Hemirhamphus). In a genus (Girardinus) of the same family the anal fin is modified to form an apparatus for holding the female during sexual congress.[494] The singular method of fertilisation practised by the males and females of Callichthys paleatus is referred to elsewhere.[495]

With the exception of the pelagic Antennarius, which builds its nest in the Sargasso weed in mid-ocean, nest-building and parental solicitude for the young are confined to freshwater Fishes and to marine forms with demersal ova. Pelagic ova must necessarily be beyond the scope of parental care. As a rule it is the male which acts as guardian nurse, the female troubling herself but little about the fate of her eggs or her offspring.

{415}[Illustration: FIG. 236.—The Butter-Fish (Pholis gunnellus) coiling round a mass of eggs. (From Cunningham, after Holt.)]

Perhaps the more primitive form of parental foresight is exhibited by those Fishes which, like the females of the Salmonidae, make a furrow in the gravelly bottom of a running stream for the reception of the eggs, and then cover them over with a layer of gravel, or like the Siluroid Arius australis, of the Burnett river in Queensland, which deposits its eggs in circular excavations in the sandy bed of the river and covers them with layers of large stones. But in neither case does it appear that either the male or the female takes any further interest in the eggs or in the young when hatched. Without actual nest-building, or even the preparation of a place for their reception, the eggs may be protected in various ways by the male. The common British Gunnel or Butter-Fish (Pholis gunnellus) rolls its eggs into a rounded mass by coiling its body round them, the male and female taking possession of them alternately. The little clumps of eggs are then deposited in holes made by the boring Mollusc, Pholas. Some British Blennies attach their eggs in a single layer to the sides of cavities in rocks, or between stones, where they are watched over by the male parent. The eggs of the Lump-Sucker (Cyclopterus lumpus) are attached in masses to rocks or to piles and guarded by the male, who aerates them by keeping up a flow of water over the spawn through the action of his pectoral fins. When hatched, the young fry cling to the body of their watchful parent by their suckers. A more decided approach to nest-building is exhibited by the Sand Goby (Gobius minutus). In this species the male scoops out the sand from beneath an empty shell, generally that of a Pecten, and the female deposits her adhesive eggs on the under surface of the shell.

{416}[Illustration: FIG. 237.—Showing the embryos of Rhodeus amarus in the gill-cavities of Unio. e, Embryos; g, inter-lamellar cavities; i.l.j, an inter-lamellar junction. (From Olt.)]

The male remains on guard, and by the movements of its pectoral fins promotes the aeration of this rude form of nest. References to some of the more striking examples of true nest-building in Fishes will be found in the systematic part of this volume, especially in those chapters treating of the Dipnoi and Amiidae, and such Teleosts as the Mormyridae, Osteoglossidae, Siluridae, Gastrosteidae, Centrarchidae, Osphromenidae, Labridae, and Antennariidae. Other illustrations of parental care are to be found in the development of marsupial pouches or grooves for the reception of the eggs in the males of the Syngnathidae (Fig. 387) and the females of the Solenostomidae, and the use of the oral cavity for a similar purpose in the males, rarely in the females, of some Siluridae, and the males or females, according to the species, of the Cichlidae. The singular method by which the female Aspredo safeguards both her eggs and her progeny is referred to on p. 596. The Cyprinoid, Rhodeus amarus (the "Bitterling" of Central Europe), is unique in the means which it adopts to {417}secure the same result.[496] By means of its long ovipositor the female Fish deposits its eggs in the mantle cavity of a Unio, or of an Anodon. Here they are fertilised by spermatozoa carried in through the inhalent siphon of the Mollusc with the inspiratory water current, and they complete their development in the gill-cavities (Fig. 237).[497]

The time which elapses between the fertilisation of the egg and the hatching out of the young Fish varies greatly in different Teleosts. The eggs of some Clupeidae hatch in a very short time, two to three days in the Anchovy, and three to four days in the Sprat. In most of the British marine food Fishes the period rarely exceeds twelve to fourteen days. The larger demersal eggs with much food-yolk are longer in hatching; in the Salmon the time ranging from thirty-five to one hundred and forty-eight days. A low temperature lengthens the time. The eggs of the Herring which hatched in eight to nine days at a temperature of 52° to 58° F. took forty-seven days in water at 32° F.

The extent to which the development of the embryo proceeds while it is still enclosed in the egg-membranes, and consequently the condition of the embryo when hatched, depends largely but not exclusively on the quantity of food-yolk which is present in the egg and available for the nutrition of the embryo during its earlier stages. Embryos hatched from pelagic ova are very small and imperfectly developed. The mouth is usually not yet formed. The median fins, which later become isolated, are continuous, and the caudal fin is diphycercal, although it subsequently becomes homocercal after passing through a heterocercal stage. The blood is colourless, and even the gill-clefts may at first be lacking. In this condition the newly-hatched Fish is nourished at the expense of the residual food-yolk, which is enclosed in a yolk-sac projecting from the ventral surface of the body (Fig. 238). As the yolk is gradually used up the mouth is formed, and the young Fish feeds on the minute organisms of various kinds living in the sea, and by degrees the form, proportions, and structure of the more mature Fish are acquired. In the case of the larger demersal eggs the young are not only longer in hatching, but when hatched they are larger and more advanced in development. The young of many Fishes are {418}provided with larval or provisional organs, and they may be so unlike the adult in other respects that their subsequent development assumes the form of a more or less striking metamorphosis. As examples of larval organs, mention may be made of the adhesive or cement organs of the larval Chondrostei and Holostei, and of the Dipnoi (e.g. Protopterus and Lepidosiren), and also of a Teleost, probably the Mormyrid (Hyperopisus bebe, Lacép);[498] the cutaneous gills of the Crossopterygii and some Dipnoi; the so-called external gills of such Teleosts as Cobitis, Gymnarchus (Fig. 239), and Heterotis, which are singularly like those of Elasmobranchs; and the defensive spines which are developed on the scales or scutes of the trunk, and the dermal bones of the skull, in the young of some Plectognathi. The most striking metamorphosis to be found in Fishes occurs in the Flat-Fishes and in the Eels, an account of which will be found in other parts of this volume (pp. 685, 602).

The only examples of viviparous Fishes occur in certain families of Elasmobranchs,[499] and in five families of Teleosts, viz. the Blenniidae, the Cyprinodontidae, the Scorpaenidae, the Comephoridae, and the Embiotocidae.[500] In the Teleosts mentioned the eggs are fertilised while they are still either in the ovarian ovisacs or in the cavity of the ovary, and their development may take place in either position. In such Cyprinodonts as Gambusia and Anableps the embryos are developed in the ovisacs, but as a rule both fertilisation and development occur in the ovarian cavity. During a prolonged gestation the young are nourished partly by the {419}food-yolk present in the eggs, and partly by a nutritive secretion derived from the ovarian walls or from the epithelial wall of the ovisacs as the case may be. In Anableps the secretion of the walls of the ovisacs is absorbed by papillae developed on the surface of the yolk-sac of the embryo along the course of its blood-vessels. The eggs of the Embiotocidae have little food-yolk, and the embryos are mainly nourished by the secretion of the ovarian walls, which is swallowed by the embryo and absorbed by villi on the inner surface of the intestine. The number of young produced varies considerably. In the Embiotocidae the ovarian cavity contains 40 to 50 young. The viviparous Scorpaenid, Sebastes norvegicus of Northern Europe, produces, it has been estimated, about 1000 young, while the Blenny (Zoarces viviparus), the only other European viviparous Teleost, produces from 20 to 300 or more, according to the size of the female. In the Blenny the eggs are hatched in about twenty days, but the young are not born until about four months after fertilisation, when they are about an inch and a half long, and in every outward respect similar to the adult Fish.

Besides the distinction between the sexes resulting from the different nature of their gonads and sex-cells, the males and females are often distinguished by secondary sexual characters {420}("sexual dimorphism"[501]). As mentioned above, females are usually larger as well as more numerous than the males, although in one or both respects the reverse may be the case. Secondary sexual characters are best marked in Teleosts, where they are generally related to the special rôle which each sex takes in the deposition and fertilisation of the eggs, and the nurture and protection of the young, of which examples have already been given. To a more limited extent they may be associated with the struggle of the males for the females, and in at least a few Teleosts the exuberant coloration of the males in the breeding season suggests that instances of courtship and sexual selection are not altogether wanting.[502]

Although the vast majority of Fishes are dioecious, instances of functional hermaphroditism are not unknown in a few Teleosts.[503] Species of the Percoid genus Serranus (e.g. S. cabrilla, S. hepatus, and S. scriba) are invariably hermaphrodite and self-fertilising. Chrysophrys auratus is an example of successive hermaphroditism, the male and female sex-cells ripening alternately. As an occasional variation hermaphroditism has been recorded in several other Teleosts, including amongst others such well-known Fishes as the Cod, the Mackerel, and the Herring. The relations of the gonads in hermaphrodites is subject to much variation. In the Cod, for example, the testes may be double, each being continuous with the hinder end of the ovary of its side, or there may be only a single testis confluent with the anterior or the posterior portion, or with some other part of the surface, of either the right or left ovary. In other Teleosts individuals occasionally present themselves with a testis and an ovary on opposite sides.

{421}CHAPTER XVI

CYCLOSTOMATA (SYSTEMATIC)

CLASS I. CYCLOSTOMATA

The Cyclostomata, or, as they are sometimes called, the Marsipobranchii, from the pouch-like, or rather sac-like, shape of their branchial clefts, are divided into two orders, the first comprising the "Hag-Fishes" or "Borers," and the second the Lampreys.

ORDER I. MYXINOIDES.

The Hag-Fishes are probably the most primitive of all existing Craniates. The mouth is nearly terminal, and there is no buccal funnel. The naso-pituitary involution communicates behind with the oral cavity and functions as a channel for the in-streaming water-current to the gills. Four pairs of short tentacles, supported by a special tentacular skeleton, are present in relation with the mouth and the terminally-placed naso-pituitary orifice. The gill-sacs open directly into the pharynx. The branchial basket is but feebly developed, and at the most it is only represented by small isolated cartilages in relation with the external branchial apertures. The lingual apparatus is remarkably developed. Besides the lingual teeth there is only a single dorsal tooth in the roof of the mouth. The dorsal arcualia are restricted to the tail, or they extend for a short distance only into the trunk. A spiral valve is absent. There is a row of mucus-secreting sacs along each side of the body. The brain has no obvious cerebral hemispheres, nor a cerebellum. Only one semicircular canal is present in the auditory organ. The eyes are degenerate, and the usual eye-muscles with the cranial nerves {422}supplying them have atrophied. The embryonic pronephros is retained in the adult. The eggs are large; segmentation is meroblastic; and development is direct, without a larval metamorphosis. Two families can be distinguished.

FAM. 1. MYXINIDAE.—Gill-sacs not exceeding six pairs, with a common external aperture on each side of the body.

The family includes a single genus, Myxine, of which the common Hag (M. glutinosa) from the North Atlantic is the best known species (Figs. 92, A, and 240). This Hag-Fish occurs off the coasts of Northern Europe, including the British Isles, as well as on the Atlantic sea-board of North America,[504] southwards to Cape Cod. Other species are found off the coasts of Chili and Japan. Myxine is quasi-parasitic in its habits, boring its way into the bodies of large Fishes. By means of its rasping "tongue" it devours all the soft parts of its prey, leaving little more than a mere shell of skin and bones. The Fishes usually attacked are the Cod and other Gadoids, but the Sturgeon is not immune, and the presence of a Hag in the abdominal cavity of a Shark (Lamna cornubica) has been recorded. Myxine has the reputation of being very destructive to Fishes caught on lines, and it is said that whole "catches" have been destroyed by its depredations, so that North Sea fishermen have been forced to change their fishing-ground. To what extent the Hags attack Fishes which are living and free is somewhat uncertain, but the little evidence obtainable seems to point to the conclusion that, as a rule, they only prey on Fishes when the latter are hooked or netted, or injured or dead. When not seeking food the Hag lives {423}in the mud of the sea-bottom at depths ranging to nearly 350 fathoms. They are able to swim very rapidly in an undulatory eel-like fashion. M. glutinosa may grow to a length of nearly two feet. The Hag has been described as a protandrous hermaphrodite, that is, it is first a male and then a female, the gonad of the young first producing spermatozoa, and at a later period becoming an ovary and giving rise to eggs. This view has hitherto met with general acceptance, but it has recently been urged with some force that the presence of the two kinds of sex-cells in a young animal is no proof of functional hermaphroditism, since it is not uncommon "to find immature eggs in the testis of many Vertebrates (Teleosts, Petromyzon, Amphibia), where the assumption of hermaphroditism, to say nothing of its protandric form, is entirely unwarranted."[505] Myxine produces eggs similar to those of Bdellostoma. Nothing is known of its breeding habits, or of its embryology.

FAM. 2. BDELLOSTOMATIDAE.—Gill-sacs 6-14 pairs, all with separate external orifices. Bdellostoma (Fig. 92, B) is found on the Pacific sea-board of both North and South America, at the Cape of Good Hope, and on the coasts of New Zealand. The numerical variation of the gill-sacs in different species, and in different individuals of the same species, and even on opposite sides of the same individual, is very remarkable. Out of 354 examples of the Californian species (B. stouti) examined by Dr. Ayres,[506] 101 had 11 gill-sacs on each side; 26 had 11 on one side and 12 on the other; 208 had 12 on each side; 11 had 12 on one side and 13 on the other; and 8 had 13 on each side. Occasional specimens may have 14 gill-sacs on each side. The variations are apparently quite independent of size, age, or sex; and when the gill-sacs are asymmetrically developed, the additional sac may be either on the right side or on the left. In the Chilian species there are 10 gill-sacs on each side, but in the species from the Cape of Good Hope the number is reduced to 6 or 7. Bdellostoma closely resembles Myxine in its habits and mode of feeding. The Californian species attaches itself to the gills or to the isthmus of large Fishes, and then rapidly bores its way into the body, devouring the viscera and muscles but leaving the skin intact. It usually attacks large {424}Flounders and species of Sebastodes, and it is especially destructive to Fishes taken in gill-nets. At Monterey every net in the summer contains the empty shells of eviscerated Fishes, and when these are taken out of the water the Hag scrambles out with great alacrity. Large fishes of even 30 pounds weight are often captured without either flesh or viscera, and it cannot be supposed that they entered the net in this condition.[507] The species lives on the sea-bottom most abundantly at a depth of 10-20 fathoms, but becomes rarer as the water deepens or becomes shallower.

The eggs of the Californian Bdellostoma are large, varying in size from 14.3-29 mm. in length, and from 6.8-10.5 mm. in width, and each egg is enclosed in a horny egg-case secreted by the epithelium of its ovarian ovisac[508] (Fig. 241). At each pole of the egg-case there is a tuft of numerous horny filaments which end in 2- 3- or 4-hooked, anchor-like extremities. In the centre of the tuft of filaments at the animal pole of the egg the egg-case is perforated by a micropyle, and a little below this {425}point the case is encircled by an opercular groove, which enables the polar portion to be thrown off like a cap at the time of hatching, so as to allow the young Bdellostoma to make its escape. The large size of the egg, which almost completely fills the cavity of the egg-case, is due to the fact that it consists mainly of food yolk, the germinal protoplasm containing the nucleus forming only a small hillock near the inner extremity of the micropyle. Bdellostoma spawns during the greater part of the year, but chiefly in the early summer, and probably about 20 eggs are deposited at one time, generally on a shelly or rocky bottom. After deposition the eggs become connected together in long chains or clusters by the interlocking of their polar hooks. Fertilisation takes place after extrusion, and the segmentation is meroblastic and discoidal, much as in Teleosts. The embryo completes its development within the egg, and when hatched it is a miniature of the adult.

ORDER II. PETROMYZONTES.

In the Lampreys there is a large suctorial buccal funnel leading behind and above into the mouth, which is supported by special cartilages, and furnished with a marginal fringe of small cirri. Numerous horny teeth are present on the inner surface of the funnel as well as on the tongue. The naso-pituitary involution forms a caecum and does not communicate with the mouth. The gill-sacs, seven in number, open externally by separate orifices, but internally they open into a median branchial canal, situated below the oesophagus and opening into the mouth in front. There is a well-developed branchial basket. Dorsal arcualia are present throughout the precaudal as well as in the caudal region. A rudimentary spiral valve is present. The brain consists of parts usually present in other Craniates, including cerebral hemispheres and a cerebellum. The auditory organ {426}has two semicircular canals, and the eyes are not degenerate. The pronephros is suppressed in the adult. The eggs are small; the segmentation is holoblastic; and there is a larval metamorphosis. There is but one family.

FAM. 1. PETROMYZONTIDAE.—The family has a nearly world-wide distribution. Most Lampreys are marine, although to a greater extent in some species than in others, but all of them seem to ascend rivers for spawning. The genus Petromyzon is characteristic of the northern hemisphere, where it is represented by various species on the coasts and in the rivers of Europe, West Africa, Japan, and North America. Three species, widely distributed in Europe, occur in the British Isles, viz.:—the Sea-Lamprey (Petromyzon marinus), which may reach or even exceed three feet in length, and is also found on the west coast of Africa and on the Atlantic coast of North America; the "Lampern" or fresh-water Lamprey (P. fluviatilis), about 18 inches long; and the Sand-Pride, Sand-Piper, or lesser freshwater Lamprey (P. planeri), usually less than a foot in length. Ichthyomyzon, Bathymyzon, Entersphenus, and Lampetra are also northern forms, collectively distributed along the Atlantic and Pacific coasts and in the rivers and great lakes of North America.[509] Other Lampreys occur only in the southern hemisphere. Geotria is common in the rivers of Chili, Australia, and New Zealand; and another genus, Mordacia, has a parallel distribution, being found on the coasts of Chili and Tasmania. A new genus and species from Chili has been recently described under the name of Macrophthalmia chilensis.[510] This Lamprey, which is only 107 mm. in length, has remarkably large eyes (2.5 mm. in diameter), vertically compressed gill-clefts, and a simple dentition resembling that of Myxine. All Lampreys are carnivorous. They feed by attaching themselves to the bodies of Fishes by their suctoral buccal funnels, and then rasping off the flesh with their lingual teeth. While thus engaged they are carried about by their victims. Salmon have been captured in the Rhone with the marine Lamprey attached to them. The Lamprey usually keeps near the bottom, either swimming with a graceful serpentine movement, or attached to stones by the buccal funnel.

{427}[Illustration: FIG. 243.—Spawning of the Brook-Lamprey (P. wilderi). On the right side of the figure a male is attached to the head of a female. (From Bashford Dean and F. B. Sumner.)]

In the spring the Sea-Lamprey ascends the rivers to spawn, and, after depositing its eggs in furrows which it excavates in the river-bottom, it returns to the sea. The river-Lampreys spawn in the smaller streams and brooks. The North American Brook-Lamprey, Petromyzon (Lampetra) wilderi, which is found in the neighbourhood of New York, deposits its eggs on the gravelly bottom of a brook, in a small gravel-filled hole lying between a number of large rounded stones[511] (Fig. 243). In the vicinity of the "nest" some ten to twelve Lampreys congregate, the males, however, being much more numerous (five to one) than the females.

{428}[Illustration: FIG. 244.—Head of the Ammocoetes of P. fluviatilis. A, ventral view; B, side view. br.1, First branchial aperture; eye, eye; l.l, lower lip; na.ap, naso-pituitary aperture; u.l, upper lip. (From Parker and Haswell, after W. K. Parker.)]

Much energy is spent by both sexes in moving stones by lifting them with the buccal funnel, but it is not always clear that this is done to circumscribe the nest, or to remove impeding obstacles. Eventually, a male attaches himself to the back of the head of a female, who at the same time is holding fast to a stone. The male then rotates its body so that the urino-genital papilla is brought near the genital orifice of the female, and the simultaneous extrusion of eggs and spermatozoa at once follows. Owing to the small amount of food-yolk which they contain the eggs of the Lamprey (e.g. P. planeri) are small, measuring about 1.1-1.2 mm. in length, and from 0.9-1.0 mm. in width. There is a micropyle at the animal pole of the egg, but the characteristic horny egg-case and the polar hooks of the Myxinoids are both wanting. The embryo hatches out as a larva known as the "Ammocoetes." At this stage of its development the larva lacks several of the most striking features which characterise the adult, and it is highly probable that the Ammocoetes represents a stage in the evolution of Vertebrates in some respects intermediate between Amphioxus and a very primitive Craniate. The mouth of Ammocoetes is bounded laterally and in front by a curious hood-like upper lip, and behind by a short transverse lower lip (Fig. 244). The eyes are deeply seated and rudimentary, and as visual organs they are useless, but the parietal eye is well developed. As in the adult, there are seven pairs of gill-sacs, but they open internally into a pharynx, directly continuous behind with the rest of the alimentary canal, and there is no dorsal oesophagus. Like the skull, the branchial basket is still very rudimentary. The dorsal and caudal fins are {429}continuous. A gall-bladder is present, and also a bile duct opening into the gut. In its mode of life, and especially in the manner in which it obtains its food, Ammocoetes presents a most remarkable resemblance to Amphioxus and the Ascidians. In the median line of the pharyngeal floor there is an open groove, the hypopharyngeal groove or endostyle, and a tract of ciliated cells along the dorsal wall represents a hyperpharyngeal groove: connecting the two in front there is a peripharyngeal ciliated groove.[512] The Ammocoetes feeds on small food particles carried through the mouth into the pharynx by currents of water produced by ciliary action. The food becomes entangled in strings of mucus probably secreted by the cells lining the endostylar groove. The mucus is then swept upwards in the pharyngeal groove, and finally wafted backwards to the stomach and intestine by the cilia of the hyperpharyngeal band. The skin exhibits the remarkable peculiarity of containing a peptic ferment capable of digesting proteids in a .2 per cent solution of hydrochloric acid. As the larva lives buried in the mud, the epidermic secretion probably helps to keep the skin free from bacteria, microscopic spores, and fungoid, or other parasitic growths.[513] The young Lamprey lives as an Ammocoetes from 3-4 years, and then in the course of a few weeks in the winter it undergoes a metamorphosis, losing its larval characters and acquiring the structure and habits of the adult. During this period the buccal funnel is completed and teeth are developed. The eyes approach the surface and become functional. The continuity of the median fins becomes interrupted. The endostylar groove becomes transformed into a thyroid gland, the gall-bladder disappears, and the bile duct becomes obliterated and changed into a mass of small follicles. The skull and branchial basket complete their development. At the same time the pharynx loses its connection with the rest of the alimentary canal and remains as the branchial canal. The so-called oesophagus of the adult is apparently a new formation which grows forwards and acquires a connection with the mouth. It is probable that it represents a hyperpharyngeal groove constricted off from the dorsal wall of the pharynx.

Both the marine Lamprey and the "Lampern" are captured {430}for food, either by nets or wicker traps. Formerly the Lampern was taken in enormous numbers in several British and Irish rivers, especially in the Severn from February to May, and in the Thames during May and June, but for various reasons the supply has much diminished in recent years. The Lampern makes excellent bait for Cod and Turbot, and for this purpose large numbers used to be taken in the Trent and Thames for despatch to Grimsby and other fishing ports.[514]

{431}CHAPTER XVII

ELASMOBRANCHII: GENERAL CHARACTERS—PLEUROPTERYGII—ICHTHYOTOMI—ACANTHODEI— PLAGIOSTOMI—SELACHII—BATOIDEI—HOLOCEPHALI

CLASS II. PISCES.

SUB-CLASS I. ELASMOBRANCHII.

In both the ancient and the modern Sharks, Dog-Fishes, and Rays, the exoskeleton takes the form of a more or less uniform investment of dermal denticles or "shagreen." The endoskeleton is wholly cartilaginous or partially calcified, and there are neither cartilage- nor membrane-bones. The vertebral column is acentrous or chordacentrous, generally with alternating basi- and inter-dorsal elements, and terminating in a heterocercal tail. The skull is usually hyostylic, very rarely amphistylic or autostylic, and the lateral halves of the primary upper jaw (palato-quadrate cartilages) usually meet in a highly characteristic median symphysis beneath the base of the skull. Branchial arches and clefts are five to seven in number, and the clefts are separated by complete interbranchial septa, which, as a rule, are continuous externally with the skin. An operculum is developed only in the Holocephali. A pelvic girdle is present. With rare exceptions the pectoral fin is uniserial. The pelvic fin is invariably uniserial. The exoskeletal supports of all the fins consist of ceratotrichia, and, when present, the fin-spines are invested by enamel. Claspers are generally present in the males.

In the surviving members of the group the nostrils retain their primitive ventral position. There is a conus arteriosus with several rows of valves. A spiracle, often furnished with a spiracular pseudobranch, is generally present, and, as a rule, {432}there is a hyoidean hemibranch supplied with venous blood from the ventral aorta. The gill-filaments are attached throughout their length to the interbranchial septa. There is an optic chiasma. An air-bladder is not developed. The intestine has a spiral valve, and there is a cloaca. The gonoducts in both sexes are derived from the kidney system. The ova are large, few in number, and enclosed in horny egg-cases, and they are fertilised before extrusion. The segmentation is meroblastic, and the embryo is furnished with long external gills.

The Elasmobranchs are for the most part active predaceous Fishes, living at different depths in the sea, from the surface to nearly a thousand fathoms, and ranging from mid-ocean to the shallower waters round the coasts in almost every part of the world. Although typically marine, they sometimes ascend rivers beyond the reach of tides, and a few are permanent inhabitants of fresh water. They are most abundant in tropical and subtropical areas, where they also attain their greatest size, and are numerous in temperate regions, but there are some species which are typically Arctic. None of them are small, and some of the Sharks are the largest of living Fishes. All are carnivorous, but so diversified is their food that in different species it may range from other Fishes of no mean size to Molluscs, Crustaceans and other Invertebrates, or even to plankton. In their breeding habits the Sharks and Dog-Fishes present many interesting features. Unlike the generality of Fishes, the eggs are fertilised internally as a sequel to the copulation of the sexes. For this purpose the males are furnished with special intromittent organs, the myxopterygia or so-called claspers, which are developed as modifications of the hinder portions of the pelvic fins.[515] Each clasper is supported by an internal skeleton, consisting of several cartilages derived from the radialia of the fins, and is traversed along its inner aspect by a groove. When sexual congress takes place the claspers are thrust through the cloaca of the female into the oviducal orifices, and in some instances it is probable that they are retained in this position by hook-like denticles developed at their free extremities. The seminal fluid then flows along these conduits into the oviducts, in the upper portions of which it meets and impregnates the eggs. After fertilisation the egg is enclosed in a dark brown horny egg-case, secreted by the oviducal gland.

{433}[Illustration: FIG. 245.—Egg-case of Heterodontus (Cestracion) galeatus. (From Parker and Haswell, after Waite.)]

As a rule each egg-case has but a single egg, but in Rhinobatus and Trygonorhina (Batoidei), both of which are viviparous, each case contains three to four eggs. Generally the egg-cases are somewhat quadrangular in shape, with the four angles, two at each end, prolonged either into short horns, or into long tapering tendrils (Fig. 246). The oval egg-cases of the Heterodontidae are remarkable not only for their size, but also for the presence of a broad spiral lamina winding round the exterior of the case from one pole to the other (Fig. 245). The majority of the Sharks, Dog-Fishes, and Rays are viviparous, that is, the young are born alive; the rest, like the Scylliidae (e.g. the common British Dog-Fishes, Scyllium canicula and S. catulus), the Heterodontidae, and the Raiidae are oviparous, that is, the young are hatched out after the extrusion of the eggs. In the oviparous species the eggs are extruded either singly or in pairs, and generally deposited on the sea-bottom. When, however, the egg-cases are provided with tendrils, as, for example, in the two British Dog-Fishes just mentioned, these organs act as anchoring filaments. When extruding an egg, the female swims round and round some piece of upright seaweed, and the curling tendrils become entwined round it in such a way that the egg becomes securely attached thereto (Fig. 246).[516] The embryos are long in developing, and in Scyllium it may be several months after fertilisation (200 to 275 days) before they are hatched, the young Fish finally escaping through a rupture in the egg-case.

{434}[Illustration: FIG. 246.—Egg of the Spotted Dog-Fish (Scyllium canicula), showing its mode of attachment after extrusion. (From Hertwig, after Kopsch.)]

In the oviparous species the nutritive food-yolk stored up, first in the egg and subsequently in the yolk-sac (Fig. 248), suffices for the nourishment of the embryo until the period of hatching, but in viviparous forms, whose embryonic development is completed within special uterine dilatations of the oviducts, additional means of nutrition are provided for the young. Such Elasmobranchs as Spinax, Acanthias, Centrina, Scymnus, Trygon, Torpedo, and Myliobatis have long filaments (villi or trophonemata) developed from the inner surface of the uterus, which secrete a nutritive fluid, and this fluid is either absorbed by the blood-vessels of the embryonic yolk-sac, or it is taken up by the embryo in some more direct manner. In some of the Trygonidae and Myliobatidae of the Indian Ocean it seems probable that the secretion is taken into the alimentary canal of the embryo either through the mouth or through the open spiracles.[517] One species, Pteroplatea micrura, has its long and highly vascular and glandular trophonemata gathered into two bundles, which are thrust through the huge spiracles into the pharynx of the embryos, of which there may be from one to three, and the nutritive secretion is apparently digested in the alimentary canal of the embryo and absorbed by the foetal blood-vessels (Fig. 247). A few Sharks, like most species of Mustelus, develop a placenta when the food-yolk in the yolk-sac of the embryo is nearly used up. Folds or projections from the highly vascular wall of the yolk-sac interlock with similar vascular folds of the lining membrane of the uterus, and a diffusion of nutrient material takes place from the maternal blood in the uterine blood-vessels to the foetal blood in the {435}vessels of the yolk-sac.[518] Each embryo has its own placenta, and in Mustelus antarcticus the uterine portion of the oviduct is divided by septa into several chambers, each containing a single embryo.[519] It is worthy of note that in the viviparous species a distinct but very thin, delicate egg-case is formed, occasionally even with the rudiments of tendrils, which may either be retained or thrown off in the uterus. The Greenland Shark (Laemargus borealis) is unique amongst Elasmobranchs. Its eggs are small and unprotected by egg-cases, and their fertilisation is said to be effected in the water after deposition, as in the generality of Fishes.

Fossil remains of Elasmobranchs in the shape of fin-spines (ichthyodorulites) and dermal denticles, associated with various Ostracodermi (Coelolepidae, Pteraspidae, and Cephalaspidae), are amongst the earliest undoubted indications of Vertebrate life. They first appear in the Upper Ludlow Bone Bed and in Silurian rocks in other parts of Europe, and in North America; and in greater or less abundance the group is represented in almost every subsequent geological period. It cannot be said that the group shows signs of decadence, for Elasmobranchs still survive in apparently undiminished numbers and variety in the marine fauna of the present day.

The Elasmobranchs are certainly a very primitive race of Fishes. Their earliest representatives of whose structure we have any precise knowledge (e.g. Cladoselache and Pleuracanthus) are in many respects the most archaic of known gnathostomatous {436}Craniates, and from such types as these, amongst others, we may reasonably look for the ancestors of all or most of the remaining groups of Fishes. It has been well said of Pleuracanthus that "it is a form of Fish which might with little modification become either a Selachian, Dipnoan, or Crossopterygian,"[520] while the condition of the primary upper jaw in the Chondrostean Polyodon suggests that even the more primitive Actinopterygii had an Elasmobranch origin. The important researches of Dr. Traquair render it also highly probable that the ancient Ostracodermi may claim kinship through their Coelolepid ancestors with some primitive type of Elasmobranch; and within the limits of the group there is ample evidence that differentiation has taken place on many divergent lines, of which we have notable examples in such specialised offshoots as the Acanthodei and the Holocephali, to say nothing of several highly specialised families which became extinct at successive periods in the history of the group.

ORDER I. PLEUROPTERYGII.

The only certain representative of this group is the Palaeozoic form Cladoselache, probably the most primitive Elasmobranch at present known (Fig. 249). Elongated and somewhat cylindrical in shape, Cladoselache[521] has a terminal mouth, five or possibly seven pairs of branchial clefts, and a pair of olfactory organs, lateral in position near the extremity of the snout. Wide-based, triangular pectoral and pelvic fins, a low anterior and a posterior dorsal fin, devoid of spines, and a heterocercal caudal fin with homocercal tendencies, are present, but no anal fin has yet been detected.

{437}[Illustration: FIG. 249.—Restoration of Cladoselache fyleri. Lateral and ventral views. (From Parker and Haswell, after Dean.)]

The exoskeleton consists of minute lozenge-shaped denticles, which invest the body and extend on to the surfaces of the fins, and there is also a circumorbital ring of several concentric rows of small square plates. A lateral line, in the form of a groove between two rows of denticles, extends along each side of the body. The notochord is persistent. Calcified neural and haemal arches (basidorsals and basiventrals) have been observed in the caudal region, where they correspond numerically with the remains of the myotomes, but interdorsal or intercalary arcualia seem to be absent. The upper and lower jaws, similar in size and shape, are apparently supported by a hyomandibular cartilage; hence the skull is hyostylic. The endoskeletal supports of the pectoral, and especially those of the pelvic fins, exhibit a more primitive disposition than in any other Fishes. They extend nearly to the distal margins of the fins, where they seem to interdigitate with the proximal ends of feebly-developed ceratotrichia (Fig. 145). The extension of the fins in the horizontal plane, the gradual shading off of their broad bases into the sides of the body, and the resemblance between their radialia and those supporting median fins, are very suggestive of the origin of the paired fins from continuous lateral fin-folds. Claspers are absent. The dentition is well developed, and several rows of teeth seem to be functional at the same time. {438}Each tooth consists of a broad base, supporting a long pointed central cusp and a variable number of similarly shaped but much shorter lateral cusps. The teeth in the various transverse rows from without inwards are closely wedged together by the interlocking or overlapping of their bases.

FAM. 1. CLADOSELACHIDAE.—Several species of Cladoselache, varying from 2 to 5 feet in length, have been found in the Cleveland Shale (Upper Devonian) of Ohio. Isolated teeth similar to those of Cladoselache occur in the Lower Carboniferous of Europe, India, and North America, and have been referred to various species of the genus Cladodus, but with one exception nothing more is known of the structure of these Fishes, and consequently their relationship to Cladoselache is doubtful. C. neilsoni,[522] from the Lower Carboniferous (Calciferous Sandstones) of Kilbride in Scotland, has a very different type of pectoral fin, which appears to be distinctly uniserial, but intermediate in structure between the biserial fin of Pleuracanthus and that of the modern sharks. There are several other genera from the Devonian and Lower Carboniferous whose claims to inclusion in this group rest on no better foundation.

ORDER II. ICHTHYOTOMI.

While more specialised than the Pleuropterygii the Fishes included in this group represent an extremely generalised type of Elasmobranch, which, as already indicated, may easily have been the ancestor of more than one group of Fishes. In the typical genus Pleuracanthus[523] (Fig. 250)[524] the body is elongate, but slightly depressed, with a terminal mouth, and a tapering diphycercal tail fringed above and below by a continuous caudal fin. A long dorsal fin, two small anal fins, and well-developed paired fins with contracted bases, are present. The head is armed with a prominent, serrated, dorsal spine, but it is doubtful if dermal denticles (shagreen) are present. The vertebral column {439}is acentrous, and the persistent notochord supports a series of basidorsal cartilages, which alternate with small interdorsals, a series of basiventrals supporting small ribs, and in the caudal region well-developed haemal arches. The dorsal fin is supported by slender, tri-segmented radialia, which appear to be twice as numerous as the neural arches in the trunk; but in the dorsal lobe of the caudal fin the two structures agree in number. Ventrally-prolonged haemal spines are the sole endoskeletal supports of the inferior lobe of the caudal. The coraco-scapular cartilages of opposite sides remain distinct, and each supports a biserial fin. The pelvic girdle is represented by a pair of small cartilages supporting the basipterygia. The pelvic fins are uniserial, with post-axial skeletal supports for claspers in the males. Both the median and the paired fins are provided with marginal ceratotrichia. The skull is probably amphistylic. Five, possibly six or seven, branchial arches, bearing clusters of minute denticles, are present. Circumorbital plates are wanting. All the endoskeletal structures are partially calcified. The teeth are tricuspid, each with two long divergent lateral cusps and a minute median cusp; the broad bases of the teeth overlap and articulate with one another by means of facets.

{440}FAM. 1. PLEURACANTHIDAE.—The single family included in the group ranges from the Lower Carboniferous to the Lower Permian. Within these limits the family is widely distributed in different formations in Great Britain, Continental Europe, New South Wales (Lower Hawkesbury Formation), and North America. Pleuracanthus, of which complete skeletons and skulls have been found, is the best known genus.

ORDER III. ACANTHODEI.

The Fishes comprising the Acanthodei[525] may be regarded as a highly specialised and terminal offshoot from some primitive race of early Elasmobranchs. The Elasmobranch kinship of the Acanthodei is indicated by their exoskeleton of shagreen tubercles; the completely heterocercal tail; the absence of an operculum, the external gill-clefts apparently being exposed; the position of the lateral line of the trunk between two rows of shagreen denticles; the nature of the powerful spines in connexion with the dorsal and anal, and the pectoral and pelvic fins; and the formation of the hard parts of the skeleton, not by ossification involving the presence of bone-cells, but by the calcification of cartilage, or of more superficial membranous or fibrous tracts. On the other hand, it may be noted that the Acanthodei appear to have undergone much specialisation on lines in some respects parallel to those which have marked the evolution of the Teleostomi, but by methods which are simply an exaggeration of features normally characteristic of Elasmobranchs. Perhaps the most striking illustration of this is to be seen in the development of a species of secondary skull by an extension of a process of calcification as distinguished from ossification. Hence the presence of membrane-calcifications in relation with the upper and lower jaws, whose development is proportional to the size of the teeth they support, and of smaller investing plates of the cranial roof. Similar exoskeletal calcifications, when most completely developed (e.g. Diplacanthus), form a dorsally incomplete arch, apparently corresponding to a secondary pectoral girdle for the support of the stout pectoral spines, in which elements {441}analogous to clavicles or cleithra and infra-clavicles can be recognised. Each pectoral spine forms the preaxial margin of the fin, and behind it there is a series of ceratotrichia. Nothing is known of the endoskeletal supports, but having regard to the nature and proportions of the pectoral spines it may be inferred that the exoskeletal elements of the fins predominate over the former to an extent which is only paralleled elsewhere in the Teleostei.

Apparently the notochord is persistent, and there are long and slender neural and haemal arches, but no ribs. The dermal denticles are uniform in size, and so small as to give a granular appearance to the skin. In structure they are thick, with a flat, enamelled, often sculptured, external surface, quadrate or rhombic in shape, and fitting closely together. Teeth are either absent or very minute, but sometimes (e.g. Acanthodopsis and Ischnacanthus) they are few in number and large, conical in shape, occasionally with minute cusps between the larger teeth. Claspers are absent. The Acanthodei are small Fishes, most of them being less than .3 m. in length, and ranging from the Upper Silurian to the Lower Permian inclusive. Two families are recognised.

FAM. 1. DIPLACANTHIDAE.—Two dorsal fins are present. Usually there is a row of lateral spines extending along each side of the body between the pectoral and pelvic fins. Exclusively Upper Silurian and Devonian.

The genera Diplacanthus, Climatius, Parexus, Euthacanthus, and Ischnacanthus are all found in the Lower Old Red Sandstone of Scotland. Climatius and Diplacanthus are also represented in the Devonian of Canada.

FAM. 2. ACANTHODIDAE.—A single dorsal fin; lateral spines vestigial or absent. Lower Devonian to the Lower Permian.

{442}The widely-distributed genus Acanthodes (Fig. 251) is represented in the Lower Old Red of Scotland, the Devonian of Siberia and Canada, the Carboniferous of England and Scotland, and the Lower Permian of France, Germany, and Bohemia. Acanthodopsis (Coal Measures), and Mesacanthus and Cheiracanthus (Lower Old Red) are the remaining genera.

ORDER IV. PLAGIOSTOMI.

Head prolonged in front of the ventrally-situated mouth as a more or less prominent preoral rostrum, vertebral column consisting of alternating basi- and inter-dorsal cartilages, generally supported by more or less well-developed chorda-centra. Pectoral and pelvic fins uniserial. Pelvic girdle and claspers present. Except in two families the branchial arches and clefts are invariably five in number. An operculum is not developed.[526]

SUB-ORDER 1. SELACHII.

Body elongate or fusiform, shading imperceptibly into a powerful swimming tail. Pectoral fins of moderate size, with contracted bases; not confluent with the sides of the head. Branchial clefts lateral in position. Vertebral centra generally asterospondylic or cyclospondylic.

This sub-order includes such typical Elasmobranchs as the modern Sharks and Dog-Fishes as well as numerous fossil representatives ranging from the Carboniferous, and probably from still earlier periods, to the present day.

FAM. 1. NOTIDANIDAE.—Body moderately elongate, the spineless dorsal fin opposite the anal. Mouth ventral; nostrils ventral, near the extremity of the snout, without oro-nasal grooves. Branchial arches and clefts six or seven. Interbranchial septa devoid of marginal frills. Notochord persistent and continuous, partially constricted by simple chorda-centra, each consisting of two distinct rings, without either concentric or radial lamellae, except {443}in one species (Notidanus cinereus), which exhibits a feeble asterospondylism in the caudal vertebrae. Skull amphistylic. Teeth unlike in the two jaws; those in the upper jaw usually with a large central cusp and smaller lateral cusps; those in the lower jaw comb-like, each consisting of numerous graduated pointed cusps inclining in the same direction, and supported on a long basal plate.

The very few species included in this family are widely distributed in the tropical and subtropical regions of the Atlantic and Pacific Oceans. Notidanus (Heptanchus) cinereus, which has seven branchial arches and clefts, inhabits the Mediterranean and Atlantic. N. (Hexanchus) griseus, with six branchial arches and clefts, has a similar distribution, but besides being an occasional visitant to the British coasts, it is not uncommon at Cuba in the West Indies. It is said to grow to a length of 26 feet.

Fossil remains of Notidanus, principally teeth, occur in the Middle and Upper Jurassic, in the Cretaceous, and in the Eocene and Pliocene of England and the Continent.

FAM. 2. CHLAMYDOSELACHIDAE (Frilled Sharks).—Body much elongate. Median fins as in Notidanus. Mouth nearly terminal. Nostrils lateral, nearly terminal, and without oro-nasal grooves. Branchial arches and clefts six. The outer margins of the interbranchial septa are produced into overlapping cutaneous frills, the first of which is developed from the hyoid arch and overlaps the hyobranchial cleft, like a rudimentary operculum. Vertebral column as in the preceding family, but in the hinder part of the trunk the notochord is unconstricted and uniform in diameter, centra being absent. Skull hyostylic. Lateral line an open groove. Teeth alike in both jaws, each consisting of a broad basal plate supporting three slender curved cusps, separated by a pair of much smaller cusps.

The only living species known is Chlamydoselachus anguineus (Fig. 252),[527] which occurs in the Pacific near Japan, in deep water off Madeira, and also off the Azores and the coast of Norway. It reaches a length of 4 to 5 feet. Teeth from the Pliocene deposits of Tuscany have been referred to an extinct species, C. lawleyi.

{444}Scarcely anything is known of the habits of the Notidanidae and the Chlamydoselachidae. It is evident that they are closely-related forms, and from the unusual number of their gill-clefts and branchial arches, and the condition of the vertebral column, it is also obvious that they are the most archaic of modern Selachians.

FAM. 3. HETERODONTIDAE (Bullhead Sharks).—Head large and high, with a blunt snout projecting but little in front of the small and almost terminal mouth, and with prominent supraorbital crests. Trunk thick-set and somewhat trihedral, covered with fine shagreen. Nostrils ventral but nearly terminal, with oro-nasal grooves. Spiracles small, beneath the eyes. Two dorsal fins, each with a spine in front, the first opposite the interval between the pectorals and pelvics, the second in front of the anal. Vertebral centra asterospondylic when fully developed. Palato-quadrate cartilages with an extensive articulation with the sides of the preorbital regions of the cranium, the normal suspensoria of a hyostylic skull (hyomandibular cartilages) taking little share in their support. Dentition similar in both jaws. Teeth at the symphyses numerous, small, and conical, furnished with three to five cusps in the young; those behind broad and pad-like, arranged in oblique rows, the teeth forming the two middle rows being much larger than those in the front or behind. Living species, oviparous. Egg-cases large, with an external spiral lamina (Fig. 245).

About four species belonging to one genus, Heterodontus (= Cestracion) (Fig. 253), or possibly to two, represent this dwindling family. All are inhabitants of the Pacific Ocean (Japan, Amboyna, Australia, the Galapagos, and the Californian coast of North America). Little is known of their habits. They feed {445}principally on Molluscs, the shells of which are crushed by their massive grinding teeth. The different species vary in size from 2 to 5 feet.

The Heterodontidae were the most characteristic and abundant Sharks of the Mesozoic period. Amongst extinct genera Hybodus ranges from the Middle Trias to the Lower Cretaceous (Wealden); an allied genus, Acrodus, from the Middle Trias to the Upper Cretaceous (Gault). Palaeospinax occurs in the Lias and possibly in the Upper Trias. Synechodus is a Cretaceous genus, and Asteracanthus, which has large hooked spines on the head, is characteristic of the Middle and Upper Jurassic. An even greater antiquity may be claimed for the Heterodontidae if, as is not improbable, such Palaeozoic Sharks as Orodus, Sphenacanthus, Tristychius (Carboniferous), and Wodnika (Permian) belong to this family. Many ichthyodorulites are probably the spines of various extinct Heterodontidae.

FAM. 4. COCHLIODONTIDAE.[528]—This Palaeozoic family includes a number of Sharks probably related to the Heterodontidae, but of which little is known except their dentition. The teeth are in some respects similar to those of Heterodontus, except that those which appear to correspond to one or both of the middle rows of the latter genus tend to fuse and form a few large, convex, and often scroll-like plates. The typical Cochliodonts are exclusively Carboniferous (Europe and North America). Psephodus, Pleuroplax, Deltodus, Poecilodus, Cochliodus, Deltoptychius, Helodus, and Menaspis (Permian) are characteristic genera. {446}Probably some ichthyodorulites described under various generic names belong to this family.

FAM. 5. PSAMMODONTIDAE.—Teeth large, flat or slightly arched, oblong or quadrate, and arranged in one, two, or more longitudinal rows. Only the teeth are known, and from differences in their shape, size, and surface markings, the genera Psammodus, Archeobatis, and Copodus have been recognised. The family is confined to the Lower Carboniferous of Great Britain and Ireland, Russia, Belgium, and North America.

FAM. 6. PETALODONTIDAE.—Teeth transversely elongated, with a blunt or a sharply-ridged crown, separated from a single or multiple root by a constricted neck, and disposed in transverse and longitudinal pavement-like rows; exoskeleton of smooth, oval, rounded or quadrate shagreen denticles. Only the teeth, and in some genera the dermal denticles, are known, except in Janessa, which has a Ray-shaped body, with large pectoral fins prolonged towards the head. The family is mainly confined to the Carboniferous formations of Great Britain, Europe, and North America. Petalodus, Janessa (also represented in the Permian), Glossodus, Polyrhizodus, and Callopristodus are characteristic genera.

FAM. 7. SCYLLIIDAE (Dog-Fishes).—Dorsal fins two in number, small, and without spines, the first above or behind the pelvic fins, the second usually behind the anal. Tail not bent upwards or but slightly so, without lateral keels. Spiracles present. Nictitating membranes absent. Vertebrae asterospondylic. Teeth small, each with a median cusp, and one to four small cusps on each side. Oviparous. Egg-cases (Fig. 246) large, quadrate, with long twining tendrils at the angles for attachment.

The genus Scyllium includes the true Dog-Fishes (Fig. 254). The species are coast Fishes of small or moderate size, and are widely distributed in temperate and tropical seas, at depths not as a rule exceeding 400 fathoms. Two species, S. canicula and S. catulus, are common on the British coasts, living near the bottom and feeding on Crustaceans and Molluscs. An allied form, Pristiurus, is also common in European and British waters. Chiloscyllium is a widely-distributed genus ranging from the Cape of Good Hope through the Indian Ocean to the coasts of Australia, China, and Japan. Stegostoma tigrinum of the {447}Indian Ocean attains a length of 10 to 15 feet, and is remarkable for its handsome coloration of dark bands on a yellow ground, which has suggested the name of Tiger- or Zebra-Shark. The pelagic genus Ginglymostoma has the terminal portion of the tail bent upwards, and grows to a length of 6 to 12 feet. It is represented by species in the Indian Ocean and the tropical parts of the Atlantic (West Indies and the west coast of Mexico). Crossorhinus includes species of large size, some of which are 10 feet long. They are ground-sharks, frequenting the coasts of Australia and Japan, which lie on the bottom watching for their prey, and in accordance with this habit their coloration closely resembles that of their surroundings.[529] A large North Atlantic Shark (Pseudotriakis microdon), of which only two specimens are known, one taken on the Portuguese coast, and the other, 10 feet in length, off Long Island, on the Atlantic coast of North America, has the general characters of the Scylliidae, except that the first dorsal fin is opposite the interval between the pectoral and pelvic fins. Some Scylliidae live at great depths, Scyllium (Scylliorhinus) profundorum having been obtained from a depth of 816 fathoms in the North Atlantic.[530]

Most of the fossil Scylliidae belong to existing genera. The earliest known representatives of the family occur in the Upper Jurassic (Lithographic Stone of Bavaria), where the extinct genus Palaeoscyllium, a near ally of the existing Scyllium, and Pristiurus, are found, nearly complete. Scyllium itself ranges from the Cretaceous through the different Tertiary formations. A species of Chiloscyllium has been recorded from the Miocene Tertiaries, and detached teeth of Ginglymostoma from the Eocene of Belgium and North America. An extinct genus (Mesiteia), which is found in the Upper Chalk of Mount Lebanon and the Upper Eocene of Monte Bolca, is remarkable for the enclosure of {448}its lateral sensory canals in a series of incomplete calcified rings, as in the Holocephali.

FAM. 8. CARCHARIIDAE.—Sharks with two dorsal fins, the first in front of the pelvic fins and the second opposite the anal fin, both devoid of spines. Tail without lateral keels. Preoral rostrum elongated. Mouth crescentic. Eyes with nictitating membranes. Spiracles small or absent. Vertebrae asterospondylic. Teeth usually consisting of a single triangular cusp, with smooth, trenchant, or serrated margins, rarely with basal cusps; generally with an axial cavity when fully developed. Viviparous. The family comprises about twenty genera, and approximately sixty species; found in all seas, often in mid-ocean. Amongst the more important genera may be mentioned Carcharias (Carcharhinus), Galeocerdo, Triakis, Thalassorhinus, Galeus, Mustelus and Scylliogaleus.

Species of Carcharias are found in nearly all tropical and subtropical seas. The genus is a somewhat comprehensive one, and groups of its species have been distinguished as sub-genera under the names of Prionodon, Hypoprion, Scoliodon, Aprionodon,[531] etc. One of the most widely distributed of the thirty to forty species is the Blue Shark, C. (Prionodon) glaucus (Fig. 255), of the Atlantic and Pacific Oceans, which may grow to a length of 25 feet, although the young forms not infrequently captured in British waters do not exceed 6 to 8 feet. It is a slender, swift, pelagic Shark, of a slaty-blue colour above and white underneath, and a voracious hunter of other Fishes. C. nicaraguensis, a Shark about 7 feet long, is confined to Lake Nicaragua and its outlet the Rio San Juan, and is one of the very rare strictly freshwater Sharks. Galeocerdo is a large Shark found in temperate and tropical waters, but one species, G. arcticus, is confined to Arctic seas. The variegated G. tigrinus, or West Indian Tiger-Shark, is said {449}to reach a length of 15 to 20 feet. The genus Galeus includes the small Sharks commonly known as "Topes," which are common in nearly all tropical and temperate seas. The British species, G. canis, which ranges from 4 to 6 feet in length, is a bottom-feeding Fish, preying on Molluscs, Crustacea, Star-Fish, and small Fishes. The various species of Mustelus, or "Hounds," resemble the Topes in their habits and distribution. Living principally on Molluscs and Crustaceans, the dentition has lost the trenchant, unicuspidate type characteristic of most other Carchariidae, and is adapted for crushing and grinding, the teeth being flat, without cusps, and arranged in pavement-like rows. Two species, M. vulgaris and M. laevis, are abundant on the coasts of Europe and the British Isles. Scylliogaleus, which combines the general characters of Mustelus with nostrils similar to those of a Scyllium, is known only from a single specimen from the coast of Natal.[532]

The Carchariidae are comparatively modern Sharks. No undoubted remains are known earlier than the Eocene, in which, as in the succeeding Miocene and Pliocene deposits, they are represented principally by their characteristic teeth. The extinct fossil genera are few in number, and so far as their dentition is concerned they differ but little from their living allies.

FAM. 9. SPHYRNIDAE (Hammer-head Sharks).—In their general characters the Hammer-head Sharks agree with the Carchariidae. They are distinguished, however, by the remarkable shape of the head, which is prolonged into two conspicuous lateral lobes, supported internally by corresponding cartilaginous outgrowths from the post-orbital and the lateral ethmoidal or nasal regions of the skull, with the eyes at their distal extremities, and the nostrils in relation with their anterior margins. One genus and five species.

The Sphyrnidae are denizens of nearly all tropical and subtropical seas. Sphyrna (Zygaena) tudes occurs in the Mediterranean, and S. zygaena is a very rare visitant to the British coasts. A specimen over 13 feet in length was captured at Ilfracombe in 1865, and other examples have been taken off Banffshire, at Newlyn in Cornwall, at Yarmouth, and in Carmarthen Bay.[533] The shape of the head differs in different species, and in young {450}forms the peculiarities of the adult are less marked. In the Bonnet Shark (S. tiburo) (Fig. 256, A), the head is crescentic or kidney-shaped, with prominent postero-lateral angles, and between this type of head and the more pronounced "hammer" of S. zygaena (Fig. 256, B) an almost perfect gradation is supplied by other species. The Hammer-heads are voracious Sharks, usually living in deep water, and they may grow to a length of 15 feet. As many as thirty-seven embryos have been taken from the oviducts of a female nearly 11 feet in length.[534]

Teeth assigned with more or less probability to Sphyrna are found in the Miocene of Europe and North America.

FAM. 10. LAMNIDAE (Porbeagle Sharks).—Large, stout-bodied Sharks with two dorsal fins, the first just behind the pectoral fins, the second, which is small, opposite the small anal fin; both {451}without spines. Tail with a prominent lateral keel on each side. Nictitating membranes absent. Spiracles minute or wanting. Branchial clefts very wide. No oro-nasal grooves. Vertebrae asterospondylic. When fully developed the teeth are solid.

In the genus Lamna, which includes the Porbeagle Sharks, the teeth are large, each consisting of a long narrow central cusp, usually with smaller cusps at the base. The common Porbeagle (L. cornubica), a fierce pelagic Shark, which may reach a length of 10 feet, frequents the North Atlantic and the North Pacific (Fig. 257). It has often been captured off the coasts of Great Britain and Ireland in Mackerel or Salmon nets, or by lines laid for food Fishes. An allied genus, Isurus, is represented by species on the Atlantic coast of North America, in the Mediterranean and the neighbouring parts of the Atlantic, and also in Asiatic seas. Carcharodon rondeletii[535] is a pelagic Shark with large, triangular, finely-serrated teeth, without basal cusps, and is found in all tropical and subtropical seas from the Mediterranean to Australia and New Zealand. It is one of the largest and most formidable of Sharks, and it is said to grow to a length of 40 feet. Nothing is known of its breeding habits. Odontaspis, which has minute pore-like spiracles, but no lateral caudal keels, is a Shark of moderate size, chiefly inhabiting the Atlantic, but found also in the Mediterranean and the Southern Pacific. Its teeth are long and awl-like, with small basal cusps.

The Thresher or Fox Shark (Alopecias vulpes) is remarkable for the extraordinary length of the upper lobe of the caudal fin, {452}which is as long as the rest of the body (Fig. 258). Its teeth are of moderate size, triangular in shape, and without serrations. The "Thresher" has a wide distribution, being abundant in the Atlantic and Pacific Oceans, besides being the commonest of the larger Sharks frequenting the British coasts. It grows to a length of 15 feet, of which the tail forms at least one-half. Quite inoffensive to man, the Thresher feeds on the shoals of smaller Teleosts, such as Pilchards, Herrings, and Sprats. When feeding it swims in gradually diminishing circles round the shoal, splashing the water with its long tail, and keeping its victims so crowded together that they become an easy prey. A remarkable Lamnoid Shark (Mitsukurina owstoni),[536] which has the snout produced into a "long, flat, flexible, leaf-like blade," somewhat resembling that of Polyodon, but narrower and more pointed, and has protractile jaws and large spiracles, is found in deep water near Yokohama, and may prove to be generically identical with the Cretaceous Shark Scapanorhynchus.[537]

Lamnoid Sharks are not certainly known to have existed until the Upper Cretaceous formations, in which, as well as in different Tertiary deposits, teeth indistinguishable from those of the existing genera Lamna, Odontaspis, and Carcharodon are found. The interesting genus Carcharodon has one extinct species in the Cretaceous and several others distributed in Tertiary formations in nearly every part of the world. The teeth of some of the Tertiary species measure 5 inches along the margin and 4 inches across the base, and it is evident that they belonged to Sharks so gigantic as completely to dwarf the existing species. That these giant Lamnidae have only recently {453}become extinct is proved by the fact that similar teeth have been dredged from the bottom of the Pacific. Teeth and detached vertebrae from various Tertiary deposits have been referred to species of Alopecias. Entire Fishes, with an elongated rostrum and an extensive anal fin, from the Cretaceous of Mount Lebanon, have been assigned to an extinct genus, Scapanorhynchus.

FAM. 11. CETORHINIDAE (Basking Sharks).—Two dorsal fins, without spines, the anterior midway between the pectoral and pelvic fins. Tail without lateral keels. Nictitating membranes absent. Spiracles small, situated just above the angles of the mouth. Branchial clefts wide and of great vertical extent, extending from the dorsal to the ventral surface. Teeth small, very numerous, conical in shape, without serrations. Claspers of the male provided with horn-like denticles.

The single species included in this family, the Basking Shark, (Cetorhinus (Selache) maximus), is one of the largest of living Fishes, reaching a length of 40 feet (Fig. 259). It is a pelagic Shark, inhabiting the Arctic seas, but wandering as far south on opposite sides of the Atlantic as the Mediterranean, the coasts of Portugal and Virginia, and in the Pacific to the Californian coast. Although generally described as a northern form, Cetorhinus is known to occur in Australian waters.[538] It is fairly common off the coasts of Scotland, and it has been seen or captured at various points on the western coast of Ireland, and {454}the eastern and southern coasts of England. The Fish is gregarious in its habits, often swimming in shoals near the surface. The name "Basking Shark" has been suggested by its habit of lying motionless on the surface in warm or calm weather, as if basking in the sun, with its dorsal fin protruding from the water. Unless attacked, this Shark is quiet and inoffensive. It derives its food-supply from small pelagic Fishes, and also from marine Invertebrates, which are strained from the water by the fringes of long, slender gill-rakers with which the branchial arches are provided. At one time harpooned and caught off the Irish, Scotch, and Norwegian coasts for the sake of the oil obtained from its liver, the Fish is now of little economic importance. Nothing is known of its mode of reproduction.

Extinct species of Cetorhinus have been founded on detached vertebrae and isolated teeth from deposits of Pliocene age in Belgium and Italy, and possibly from still earlier Tertiary formations. Dermal spines similar to those found on the claspers of the males in the existing species occur in the Antwerp Crag, and in the Red Crag of Suffolk.

FAM. 12. RHINODONTIDAE.—Two dorsal fins, without spines, the anterior a little in front of the pelvic fins, the second opposite the anal. Tail with lateral keels and a pit at its root. Spiracles small. Nictitating membranes absent. Mouth and nostrils nearly terminal. Teeth very minute, numerous, and conical in shape.

One genus, Rhinodon, with one or two species, is known. These Sharks are very widely distributed, specimens having been seen or captured in the neighbourhood of Ceylon, at the Seychelles, the Cape of Good Hope, Callao on the Peruvian coast, in the Gulf of California, and off the coast of Florida. Rhinodon is probably the largest known Shark. It is stated to exceed 50 feet in length, but to be quite harmless. Scarcely anything is known of its habits, but the small size of the teeth, and the length of the gill-rakers, which resemble those of the Basking Shark, suggest a similar kind of food.

FAM. 13. SPINACIDAE.—Two dorsal fins, the first in advance of the pelvic fins. Anal fin absent. Nictitating membrane absent. Spiracles rather large. Vertebrae cyclospondylic. Teeth variously modified in different genera.

{455}The more typical representatives of this family are the Spiny Dog-Fishes, which are distinguished by the presence of a strong spine in front of each dorsal fin. They are more abundant in temperate regions than in the intervening tropics. The more important genera are Acanthias, Centrina, Centrophorus, Spinax, and Centroscyllium. Acanthias vulgaris, the Picked or Piked Dog-Fish, is a gregarious, voracious Shark, about 3 to 4 feet in length, and is frequently seen in huge shoals all round the British coasts, especially during the summer months. It is very destructive to food Fishes, and its ravages result in serious loss to fishermen. Acanthias is viviparous. Centrina salviani is a much smaller Shark, which frequents the Mediterranean and the Bay of Biscay; on rare occasions it has been taken off the southern coast of England. Centrophorus occurs in deep water in the Mediterranean and adjacent portions of the Atlantic, and off the coasts of Japan. Centroscyllium is found on opposite sides of the North Atlantic (Greenland and Massachusetts), and in the opposite hemisphere at the Falkland Isles. A deep-water form, Paracentroscyllium, has been obtained in the Bay of Bengal at depths from 285 to 405 fathoms.[539]

Three remaining genera (Scymnus, Laemargus, and Echinorhinus) differ from the preceding in the absence of dorsal spines.

Scymnus lichia is common in the Mediterranean and the neighbouring parts of the Atlantic. The Greenland Shark (Laemargus borealis) (Fig. 260) is an inhabitant of the Arctic regions, wandering as far southwards on opposite sides of the Atlantic as the French coast and Cape Cod. It is a huge, clumsy shark, reaching a length of 26 feet. Numerous instances are recorded of its capture off the coasts of Great Britain, especially in northern waters. The Greenland Shark is said to be a determined foe to {456}the Right Whale, which it attacks, biting pieces out of its body. Scymnus is viviparous, Laemargus oviparous, and the latter is unique among Sharks in producing eggs devoid of a horny shell, which are deposited on the sea-bottom. Echinorhinus has dermal denticles in the form of relatively large rounded tubercles, each surmounted by a tuft of fine spines. One species only is known, E. spinosus, a large Shark attaining a length of 10 feet, and frequenting deep water off the Atlantic coasts of Europe and Africa from the North Sea to the Cape of Good Hope. A single specimen has been taken at Cape Cod on the eastern coast of the United States, and another off Dunedin, New Zealand. The capture of thirty examples in British waters since 1828 has been recorded,[540] the largest a female 9 feet in length.

Most of the existing genera of Spinacidae are represented by teeth or detached spines in the later Tertiary deposits, but none are certainly known to occur earlier than the Pliocene.

FAM. 14. RHINIDAE (Angel-Sharks).—Ray-like Sharks with a flattened head and body, and nearly terminal mouth and nostrils. Pectoral fins very large, horizontally expanded, but constricted at the base and not adherent to the sides of the head or trunk. Two dorsal fins, both small, without spines, and situated on the tail behind the pelvic fins. Anal fin absent. Spiracles large {457}and crescentic. Vertebrae tectospondylic. Teeth conical and pointed. A single species only is known.

Rhina squatina, the Angel-Shark or Monk-Fish (Fig. 261), is intermediate between the ordinary Sharks and the Skates and Rays, both in external appearance and internal structure, but is more Ray-like than Shark-like in its habits. Within the temperate and tropical regions of both hemispheres it is almost cosmopolitan in its distribution, frequenting the coasts of Europe, including the British Isles, the Atlantic and Pacific coasts of North America, and the shores of South Australia and Japan. The Angel-Shark is viviparous, producing about twenty young at a time. Not rarely it grows to a length of 5 feet.

The family ranges from the Upper Jurassic to the present time. Species of Rhina are represented by more or less complete skeletons in the Lithographic Stone of Bavaria, and in the Upper Cretaceous of Westphalia and Mount Lebanon, and by teeth and vertebrae in the English Chalk, as well as in different European Tertiary formations.

FAM. 15. PRISTIOPHORIDAE.—Prenasal portion of the head and cranium produced into a long flattened rostrum, furnished with a pair of long tentacles on its under surface, and, as in Saw-Fishes, with a series of large, tooth-like, dermal denticles, of equal or unequal size, along each of its lateral margins. Two dorsal fins, without spines, the first in front of the pelvics. No anal fin. Pectoral fins large, distinct from the head and trunk, with a contracted base. Spiracles large and crescentic. Teeth small, with a conical cusp and a broad base.

These singular Sharks closely resemble the true Saw-Fishes (Pristidae), but they differ in the lateral position of their gill-clefts, the presence of rostral tentacles, and their smaller size. The few species known belong to the genus Pristiophorus, and are confined to the Australian and Japanese seas.

Pristiophorus is represented in the Upper Cretaceous of Mount Lebanon, and in the Miocene deposits.

SUB-ORDER 2. BATOIDEI.

Body generally discoidal or rhombic in shape, the axial portion being formed by the flattened head and trunk, and the lateral portions by the enormously expanded pectoral fins, which {458}are usually confluent with the sides of the head. Tail slender, sharply marked off from the trunk, to which it usually appears as a mere appendage. Dorsal fins, when present, on the tail. Anal fin absent. Branchial clefts ventral in position. Spiracles large, usually crescentic. Vertebrae tectospondylic.

For the most part the Batoidei are sluggish ground-Fishes, slowly moving over the sea-bottom by the gentle undulatory vibrations of the margins of their huge pectoral fins, the tail being of little use in locomotion. They feed principally on Crustacea, Molluscs, and the smaller Teleosts. As with other Fishes of similar habits, the coloration of the dorsal surface harmonises with that of the sea-bottom, while the ventral surface is either deficient in pigment or white. The majority of them are coast Fishes, rarely descending to a greater depth than 500 fathoms, but some are pelagic. The Batoidei are a relatively modern race, first appearing towards the middle of the Mesozoic period, and evidently representing an assemblage of specialised Elasmobranchs adapted for a bottom-living existence. As remarked by Smith Woodward, the three families, Rhinobatidae, Raiidae, and Trygonidae, are not so clearly differentiated before the close of the Cretaceous period as they subsequently become.[541]

The first two families, the Pristidae and the Rhinobatidae, are interesting connecting-links between such Selachii as the Rhinidae and the Pristiophoridae and the more specialised Batoidei like the Skates, Rays, and Trygons. While they agree with the latter in the ventral position of the gill-clefts, the absence of an anal fin, and the caudal position of the dorsal fins, the body still retains an elongated and somewhat Shark-like shape, and shades off imperceptibly into a powerful swimming tail, and in the Pristidae at all events the pectoral fins are of moderate size and free from any fusion with the sides of the head. It must be admitted that the institution of the two sub-orders introduces a somewhat arbitrary distinction between certain families of Plagiostomes which has little to recommend it except custom and some measure of convenience. The two series of Fishes shade almost imperceptibly into one another, and the importance of the ventral position of the gill-clefts has probably been overestimated. Primitively, the gill-clefts are lateral, and lie wholly in front of the pectoral fins, a position which is retained in many {459}Selachii. In others, however, the hinder gill-clefts tend to extend backwards above the base of the pectoral fins, while in some the clefts assume a more ventral position, and extend beneath the pectoral fin; hence, even within the limits of the Selachii the position of the gill-clefts varies to the extent that these structures may be lateral, or they may tend to become either dorsal or ventral.[542] On the score of convenience the customary usage is adopted here.

FAM. 1. PRISTIDAE (True Saw-Fishes).—Although somewhat depressed, the body is still elongate and Shark-like, with a well-developed tail terminating in a heterocercal caudal fin. Dorsal fins large, the first opposite the pelvic fins. Head and skull prolonged into a long flattened rostrum, the lateral margins of which are armed with a series of strong tooth-like denticles, firmly implanted in sockets in the calcified rostral cartilage. No rostral tentacles. Teeth in the jaws minute and obtuse. One genus and about four or five species are known, all inhabitants of tropical and subtropical seas.

Some of the true Saw-Fishes attain a considerable size, 10 to 20 feet or even longer, and "saws" 6 feet long and a foot in width across the base are not uncommon. By means of powerful lateral strokes of its saw the Fish is capable of lacerating the bodies of other animals and tearing off pieces of flesh, which it then devours. Indian species are known to ascend rivers beyond tidal influence, and an American species, ranging northwards to the West Indies and the Gulf of Mexico, where it is abundant, enters the lower Mississippi. P. antiquorum occurs in the Mediterranean and the Atlantic, but does not extend so far northward as the British coasts.

The earliest known representative of the family is the {460}extinct genus Sclerorhynchus from the Upper Chalk of Mount Lebanon, in which the smaller size and more superficial position of the rostral "teeth," and the absence of sockets in the rostral cartilage, prove that the "teeth" approximate more to ordinary dermal spines in this genus than in any of the more recent Saw-Fishes. An extinct genus Propristis, from the Upper Eocene of Egypt, with non-socketed teeth, and species of the existing genus Pristis from the English Middle Eocene, are also known.

FAM. 2. RHINOBATIDAE.—Owing to the increased expansion of the pectoral fins and the forward growth of their anterior cutaneous portions along the sides of the head, as well as backwards along the trunk, the body now assumes a sub-rhombic shape, and approximates to the disc of the more typical Batoidei, but the tail with its dorsal and caudal fins is still strongly developed, and blends imperceptibly with the trunk in front. Teeth very obtuse. No electric organs. About five genera and twenty species are known, distributed in most tropical and subtropical seas.

The cosmopolitan Rhinobatus is represented by species from the Mediterranean, the Red Sea, the west coast of Africa, the Indian Ocean, Australia and China, as well as from the Atlantic and Pacific coasts of America, and the Galapagos. Rhynchobatus ranges from the Red Sea through the Indian Ocean to China, Zapteryx occurs at San Diego and Panama, and Platyrhinoidis on the Californian coast. Trygonorhina is an Australian genus.

The family dates from the Upper Jurassic. Rhinobatus is {461}represented by complete skeletons in the Lithographic Stone of Bavaria, the Upper Cretaceous of Mount Lebanon, and the Upper Eocene of Monte Bolca. Trygonorhina occurs in the Eocene.

FAM. 3. RAIIDAE (Skates or Rays).—The endoskeletally supported portions of the large pectoral fins extend along the lateral margins of the trunk and head from the pelvic fins to the snout, and are confluent therewith, forming the lateral portions of a large rhombic disc. The tail is slender, and sharply marked off from the trunk. Usually two small dorsal fins on the tail. Caudal fin small or absent. No serrated spine on the tail. Caudal electric organs are often present. Larger or smaller denticles or spines are generally present on the skin. Oviparous. Egg-cases four-horned, without tendrils. Four genera and from thirty to forty species. Found in all temperate seas, a few ranging into deep water.

The great majority of the species belong to the genus Raia (Fig. 264), which chiefly inhabits temperate seas, but is more abundant in the northern than in the southern hemisphere, and approaches nearer to the Arctic and Antarctic regions than any other Batoidei. The colour of the upper surface of the body is closely assimilated to that of the sandy or gravelly bottom on {462}which they live, and thus concealed, small Fishes, Crustaceans, and other organisms are lured unsuspectingly within the reach of the comparatively inactive and sluggish Ray. From the ventral position of the mouth the Ray cannot at once seize its prey, but the Fish darts over its victim and covers it with its body, and then readily devours it. The sexes are usually distinguished by secondary sexual characters, which take the form of differences in size and coloration, in the dentition, and also in the presence and position of patches or rows of specially modified dermal spines on the dorsal surface (Fig. 264). Some of the larger species reach a great size, the disc measuring 7 to 8 feet in width. A few species range into deep water. R. mamillidens, a uniformly jet-black species, has been obtained from a depth of 597 fathoms in the Bay of Bengal,[543] and R. abyssicola from 1588 fathoms off Queen Charlotte Islands, British Columbia.[544] The following are British species: the Thornback (R. clavata); the Spotted Ray (R. maculata); the Painted Ray (R. microcellata); the Starry Ray (R. radiata); the Cuckoo or Sandy Ray (R. circularis); the Skate (R. batis); the Flapper Skate (R. macrorhynchus); the White Skate (R. alba); the Long-nosed Skate (R. oxyrhynchus); and the Shagreen Ray (R. fullonica).[545] Most of the species are of some economic value as food Fishes. Psammobatis, with a circular disc, frequents the southern coasts of South America, and Platyrhina the coasts of India, China, and Japan.

The family ranges from the Upper Cretaceous, in which, as well as in different Tertiary deposits, it is represented by species of Raia. An extinct genus, Cyclobatis, with a circular or oval disc, occurs in the Upper Cretaceous of Mount Lebanon.

FAM. 4. TAMIOBATIDAE.—The systematic position of the only representative of this family, Tamiobatis vetustus,[546] from the Devonian or Lower Carboniferous of Kentucky, is very uncertain, but in some respects this unique type seems to be intermediate between the modern Sharks and the Rays.

FAM. 5. TORPEDINIDAE (Electric Rays).—A disc is formed as in the Raiidae, but it is sub-circular in shape rather than rhombic, and in the nature of its endoskeletal supports it is in some respects unique. Its semicircular anterior margin is supported {463}in the centre by a branched prenasal rostrum, and laterally by the curiously branched preorbital cartilages, each of which radiates outwards and forwards from a common basal articulation with the lateral ethmoid regions of the skull. Tail relatively short and thick, with two dorsal fins, a caudal fin, and two lateral longitudinal folds. Skin smooth, without denticles. Mouth transverse and ventral. A characteristic quadrangular naso-frontal lobe, with a free hinder margin, which forms the anterior lip, is enclosed by the two nasal organs and the oro-nasal grooves leading from them to the corresponding angles of the mouth. A pair of large electric organs between the pectoral fins and the head. Seven genera and about fifteen species. Inhabitants of most warm seas.

The well-known genus Torpedo (Fig. 265) is represented by species in the Mediterranean (T. marmorata, T. narce, T. hebetans), the Red Sea, and the Atlantic and Pacific Oceans. T. hebetans {464}has been taken at several places in British waters. An American Torpedo (Tetronarce) is represented by species on the Atlantic and Pacific coasts. Narcine is a very widely distributed genus, species having been recorded from the East Indies, Tasmania, China, Japan, South Africa, and the Atlantic coasts of North and South America. Discopyge is an eastern Pacific genus (Peru and Panama). Hypnos frequents the Australian seas.

The family seems to be exclusively Tertiary, and its earliest fossil representatives are from the Upper Eocene of Monte Bolca.

FAM. 6. TRYGONIDAE (Sting- or Whip-tailed Rays).—Disc sub-rhombic, broader than long. Pectoral fins confluent with the sides of the head, their preaxial endoskeletal radialia meeting in front of the skull along the lateral margins of a slender prenasal rostral cartilage. Tail usually whip-like, terminating in a small caudal fin, and generally armed with a sharp, serrated spine, which takes the place of a dorsal fin. Skin smooth or spinose. A rectangular naso-frontal flap in front of the mouth. About ten genera and fifty species. Found in nearly all tropical and subtropical seas.

Of the more important genera, Trygon (Dasyatis) is represented by numerous species in the tropical parts of the Atlantic and Pacific Oceans, including the Pacific coasts of North and South America. Two species occur in the Mediterranean, and one of them (T. pastinaca), ranges from the coasts of Norway and the British Isles through the Atlantic and Indian Oceans to Japan. Urogymnus frequents the Red Sea and the Indian Ocean. Urolophus includes a few species of small size, distributed along the Atlantic and Pacific coasts of Central and North America, and in Australian seas. Pteroplatea comprises rather large species, and is almost cosmopolitan in its distribution, being represented by species on the Atlantic and Pacific coasts of North and South America, in the Mediterranean and the Red Sea, the Indian Ocean, the Malay Archipelago, and on the coasts of China and Japan. The caudal spines, which may be 8 to 9 inches long in some of the larger species, are capable of inflicting very severe wounds, the danger of which is greatly increased by the apparently poisonous cutaneous mucus introduced into the wound. As the spines become lost they are replaced by others developed from behind. Some Trygonidae live in fresh waters. Trygon (Dasyatis) sabina frequents the streams and estuaries of Florida {465}as well as on the adjacent coasts, and specimens have been obtained from Lake Munroe at some distance from salt water.[547] Ellipesurus and Paratrygon are freshwater genera, found in Colombia, Venezuela, and Guiana.

Fossil remains of undoubted Trygonidae appear to be confined to the Tertiary period.

FAM. 7. MYLIOBATIDAE (Eagle-Rays).—Disc much broader than long, and rhombic in shape. The huge pectoral fins are not continued to the extremity of the snout, but cease on the sides of the head, and reappear in front of the snout as a pair of distinct folds, the so-called cephalic fins. The head projects above the level of the disc, and consequently the eyes and spiracles are lateral in position. Tail long, slender, and whip-like, with a single dorsal fin near the root, and usually one or two serrated spines behind the fin. A rectangular naso-frontal fold is present. The dentition consists of flat, hexagonal, pavement-like crushing teeth arranged from before backward in arched rows in both jaws, and there is either a single median row of large teeth, with (e.g. Myliobatis) or without (e.g. Aëtobatis) the addition of several rows of much smaller teeth on each side, or there are numerous rows, the teeth then decreasing in size from the middle line laterally (e.g. Rhinoptera). Skin smooth. Sexes similar. Five genera and about twenty-seven species are known; all inhabitants of tropical and subtropical seas.

Myliobatis is represented in the Mediterranean by two species, and one of them, the almost cosmopolitan M. aquila (Fig. 266), has been taken at various points on the eastern and southern coasts of England. Aëtobatis is also widely distributed in tropical seas, but is unknown in European waters. Rhinoptera has one species in the Mediterranean, while others have been recorded from Brazil, the Atlantic and Pacific coasts of North America, and the East Indies. The two tropical genera Dicerobatis and Ceratoptera have the cephalic fins prolonged anteriorly into a pair of horn-like appendages, which are said to be used in conveying food to the mouth. The teeth are small, flat or tubercular, and are arranged in numerous rows. In Ceratoptera they are wanting in the upper jaw. The Eagle-Rays feed principally on Molluscs, the shells of which they crush with their large grinding-teeth. Some of them attain {466}an enormous size, and are among the largest of Fishes. Ceratoptera vampyrus of the West Indies, for example, grows to a width of 20 feet, and an embryo extracted from the oviduct of a gravid female 15 feet wide, and from 3 to 4 feet in thickness, measured 5 feet across the disc and weighed twenty pounds.[548] This Fish is much dreaded by the divers engaged in the pearl fisheries near Panama, whom it is said to devour after enveloping them with its vast wings.[549]

The family is exclusively Tertiary, and with the exception of an extinct genus, Promyliobatis, from the Eocene of Monte Bolca, all the fossil species belong to the existing genera Myliobatis, Rhinoptera, and Aëtobatis.

ORDER V. HOLOCEPHALI

The propriety of including the Holocephali in the sub-class Elasmobranchii is scarcely open to doubt. Like the Acanthodei {467}they seem to represent a divergent and specialised offshoot from some primitive Elasmobranch type, and while retaining most of the essentially distinctive features of their ancestors, they have acquired, perhaps independently, certain characters distinctive of the Teleostomi, combined with others peculiar to themselves. In the few surviving genera agreement with the Elasmobranchs is to be seen in the wholly cartilaginous condition of the endoskeleton and the complete absence of cartilage- and membrane-bones. The vertebral column is acentrous and ribless, and the notochord is persistent; the dorsal arcualia include supradorsals and regularly alternating basi- and inter-dorsals. The limbs and limb-girdles are essentially Elasmobranch. Dermal denticles are present, either locally, or, as in some of the fossil types, in the form of a general investment. The brain and the reproductive organs agree more closely with the corresponding structures in the Elasmobranchs than with those of any other Fishes, and the agreement extends to the large size of the eggs and their enclosure in horny egg-cases. In both groups the nostrils are connected with the mouth by oro-nasal grooves; the hyoidean hemibranch is a true gill, and there is no air-bladder. The Holocephali also agree with the Elasmobranchs in retaining such primitive features as an intestinal spiral valve and a conus arteriosus. On the other hand, indications of specialisation in the Teleostome direction are to be noticed in the tendency to the concentration of the branchial arches towards and beneath the skull; the reduction of the interbranchial septa to the extent that they are no longer continuous with the skin, and the gill-filaments project beyond their outer margins; the presence of an operculum; the suppression of the spiracles; and the absence of a cloaca, the rectum opening externally by an anus in front of the urino-genital apertures. Among the more notable features evolved within the limits of the group mention may be made of the autostylic condition of the skull, probably an adaptive modification induced by the large size of the crushing dental plates which have taken the place of ordinary teeth; and the singular development of anterior and frontal "claspers."

The group is one of great antiquity. Apart from the isolated spines or "ichthyodorulites" common in Devonian and Carboniferous strata, some of which are probably the frontal or the {468}fin-spines of ancient Holocephali, dental plates, closely resembling those of modern Chimaeroids and referred to the Ptychodontidae, are probably the earliest indications of the existence of the group. The Holocephali become more abundant in the Mesozoic period, but of the four families usually recognised, only one, the Chimaeridae, has survived.

FAM. 1. PTYCTODONTIDAE.—This Palaeozoic family is known only by the dental plates, of which there is a single pair in each jaw, meeting at the symphysis. Ptyctodus[550] and Rhynchodus occur in the Devonian of either Russia or Germany, and in North America, and Palaeomylus only in the Devonian of North America.

FAM. 2. SQUALORAIIDAE.—General shape of the body similar to the existing Harriotta. There is a long, depressed, preoral rostrum, and in the male the head carries a long slender frontal spine. Conical denticles are sparsely present on the head and body. No dorsal fin-spine. Dental plates similar to those of the living Chimaeroids, but thinner, the tritoral areas being less well defined. The only genus is Squaloraia from the English Lias, of which nearly complete skeletons are known.[551]

FAM. 3. MYRIACANTHIDAE.[552]—Body elongate, but less depressed. A dorsal fin-spine is present, and in the males a frontal spine. The dentition consists of a median incisor-like tooth at the symphysis of the lower jaw, in addition to dental plates similar to those of Squaloraia. There is a symmetrical series of tuberculated dermal plates on the lateral surfaces of the head, which probably represent groups of fused denticles. One species (Myriacanthus granulatus) has its rostrum terminating in a cutaneous flap, as in Callorhynchus. Myriacanthus, from the Lower Lias of Lyme Regis, and Chimaeropsis, from the Lithographic Stone of Bavaria, are the only two genera.

FAM. 4. CHIMAERIDAE.—Body elongate and shark-like in form, but the head is compressed and the mouth is small. Pectoral and pelvic fins large, especially the former, which are somewhat ventrally placed. Two dorsal fins, the anterior over the pectorals, with a stout spine in front; and a small anal fin. Dermal denticles restricted to the claspers, and to localised areas {469}on the dorsal surface in young forms. Dental plates large and thick, including a single pair in the lower jaw and two pairs, vomerine and palatine teeth, above, which combine trenchant edges with well-marked grinding areas. Three genera are known.

In Chimaera (Fig. 267) the mouth and nostrils are ventral, posterior to a bluntly conical snout. Head surmounted in the males by a club-shaped appendage armed with a pad of recurved denticles, the frontal clasper; there is also an anterior clasper armed with similar denticles and retractile into a shallow glandular pouch in front of each pelvic fin, in addition to the ordinary clasper behind the fin. The caudal fin consists of nearly equal-sized dorsal and ventral lobes, between which the slightly up-tilted caudal axis is prolonged as a long tapering filament: hence the tail appears to be nearly diphycercal. C. monstrosa occurs off the coasts of Europe from Norway to Portugal, including the Mediterranean, and also in the neighbourhood of the Azores, as far south as the Cape of Good Hope, and eastwards off the coast of Japan. It is the largest of the living species, reaching a length of 3 feet. C. affinis was first taken off the coast of Portugal, and subsequently on the North American side of the Atlantic, at depths ranging from 200 to 1200 fathoms. C. (Hydrolagus) colliei is restricted to the North Pacific, and is especially plentiful off South-eastern Alaska, and about the wharves at Esquimalt. Unlike most other Chimaeroids this species swims at the surface, and there is no evidence that it is a deep-sea form. In its breeding habits, and in the mode in which its eggs are fertilised, Chimaera probably resembles the oviparous Sharks and Dog-Fishes.

{470}[Illustration: FIG. 268.—Egg-case of a species of Chimaera. a, Transverse section across the case at x, showing the lateral valvular slits; b, similar section across x′, showing the vertical ridges. (From Günther.)]

The eggs appear to be deposited on the sea-bottom in deep water, but they are very rarely obtained. An egg-case dredged up off the south-west coast of Ireland, at a depth of 315 fathoms, and about 6½ inches in length, is shown in Fig. 268.[553] It consisted of a broad, somewhat oval, flattened portion which contained the egg, and terminated at one end in a truncated margin, while at the other it was produced into a long tapering styliform process, traversed by dorsal, ventral, and lateral ridges. The cavity of the egg-case was open in front, and also along each side, where linear, slit-like valvular apertures freely admitted sea-water into the central cavity. A similar egg-case from Japan, measuring 9 inches in length, had its surface traversed by longitudinal and {471}transverse ridges, and no doubt belonged to a Japanese Chimaera.[554] In neither egg-case was there any trace of tendrils. The eggs probably lie on the sea-bottom, or, when the cases have styliform prolongations, it is possible that they are implanted in the ooze.

Callorhynchus (Fig. 269) is distinguished by a singular prolongation of the rostrum, which terminates in a downwardly-directed cutaneous flap, evidently from its abundant nerve-supply an important tactile organ. A frontal clasper is present in the male. The prolonged caudal axis is up-tilted, and the tail is more distinctly heterocercal than in Chimaera. The only species, C. antarcticus, is confined to the Antarctic basin and the South Pacific. The egg-cases of Callorhynchus differ considerably from those of Chimaera, and so large are they that one may measure 25 cm. in length, or nearly as long as the abdominal cavity of the Fish. Each case is ovoid in shape, surrounded by a wide flat margin which is covered on one side with yellow hair-like fibres, thus giving to the case a protective resemblance to a mass of seaweed (Fig. 270). In the central part of the case there is a pear-shaped cavity in which the egg or the embryo is contained. From one end of this cavity a passage, guarded by a valve, leads to the exterior, and provides for the escape of the young. While in the egg-case the nearly ripe embryo has long external gills, and its body is nearly sessile on a large and singularly lobed yolk-sac.

{472}[Illustration: FIG. 271.—Harriotta raleighana. A, lateral view; B, ventral view of a male. (From Goode and Bean.)]

The third genus, Harriotta (Fig. 271),[555] is remarkable for its elongated, tapering, and depressed rostrum, and for the large size and wing-like appearance of the pectoral fins. There is no frontal clasper, and the ordinary claspers in the young male examined {473}were very small and simple. The caudal filament, which is longer in older specimens than in the younger, and is not developed at all in the youngest examples at present known (Fig. 272, A), is not uptilted, although the lower lobe of the caudal fin is much larger than the upper. Young forms have a double row of stout spine-like denticles in front of the second dorsal fin, and also in the interval between the latter and the upper caudal lobe. Similar denticles are also present on the upper surface of the head between the orbits (Fig. 272). H. raleighana is found in the North Atlantic. Individuals varying in length from 4 to 25 inches have been taken at depths ranging from 707 to 1081 fathoms. A species of Harriotta has also been recorded as occurring in Japanese waters.[556]

With the probable exception of Chimaera colliei the surviving Holocephali are denizens of deep water; hence their comparative rarity and our almost complete ignorance of their habits. Young forms of C. monstrosa, 1½ to 5 inches in length, have been dredged in the Färoe Channel at depths from 505 to 555 fathoms;[557] and the youngest specimen of Harriotta was obtained from 991 fathoms. Egg-cases are rarely obtained, and then only from considerable depths. It is therefore reasonable to {474}infer that these Fishes breed in deep water. As might be expected, little is known of the embryology of any of the Holocephali, but that little adds further proof of the Elasmobranch relationship of the group. The segmentation of the egg of Chimaera and the overgrowth of the yolk by a circular blastoderm are essentially as in Elasmobranchs. The early embryos are said to be shark-like, and to possess both spiracles and "external gills," and the primary upper jaw is less completely confluent with the skull than in the adult. It is also said that the palatine dental plates are represented at an early stage by series of small, more or less conical elements, which, outwardly at least, resemble the rudiments of the grinding teeth of the Cestraciont Sharks.[558]

The Chimaeridae first appear in the Lower Oolites, and attain their maximum development in the Cretaceous and the Eocene.[559] Ganodus is an Oolitic genus. Ischyodus ranges from the Lower Oolites to the Lower Cretaceous. Edaphodon is Cretaceous and Eocene, extending, however, into the Miocene, and Elasmodus ranges from the Upper Cretaceous into the Eocene. Teeth of the existing genus Callorhynchus occur in the Cretaceous of New Zealand, and of Chimaera in the Upper Tertiary of Europe and Java. The fossil Holocephali afford little evidence of the origin of the group from more typical or more primitive Elasmobranchs. So far as their structure is known, they all possess the essentially distinctive features of their modern representatives, and offer little evidence of transitional forms. The surviving Chimaeroids seem to have acquired a more specialised dentition, but in other respects they are either more primitive, or possibly somewhat degenerate.

{475}CHAPTER XVIII

TELEOSTOMI: GENERAL CHARACTERS—CROSSOPTERYGII—CHONDROSTEI—HOLOSTEI

SUB-CLASS II. TELEOSTOMI.

In this group of Fishes the primary upper and lower jaws (palato-quadrate and Meckelian cartilages) are supplemented by the addition of certain tooth-bearing membrane bones which form secondary jaws corresponding to the functional jaws of the higher Craniates.[560] The chondrocranium and the primary jaws are usually more or less completely ossified by cartilage bones, and there is always a secondary cranium of dermal bones, of which paired parietals and frontals above, and a median vomer and a parasphenoid below, are amongst the most constant. The skull is hyostylic. An operculum covering the gill-clefts and supported by a special opercular skeleton is a constant feature. The vertebral column is often acentrous, and when centra are present they are invariably arch-centra. There is a well-developed secondary pectoral girdle, connected dorsally with the hinder part of the skull. As a rule the pelvic girdle is absent altogether, and when present it is rarely more than a rudiment or a vestige. The endoskeletal supports of the paired fins are uniserial. The dermal fin-rays of the paired and median fins are probably modified scales or lepidotrichia. In the median fins the fin-rays are at first more numerous than their supporting radials, but in the more specialised Teleostomes they ultimately equal them in number. The body is usually invested by an exoskeleton of articulated rhombic or imbricated cycloid scales. Claspers are unknown. In the surviving members of the group there is usually an {476}air-bladder. The gill-filaments project freely beyond the outer edges of the greatly reduced interbranchial septa. The external opening of each nasal sac is usually divided into two distinct apertures, and there is no oro-nasal groove leading from the sac to the mouth. The brain has no proper cerebral hemispheres, but retains an undivided prosencephalon with a non-nervous roof. A cloaca is not developed, the rectum opening externally by an anus in front of, and distinct from, the separate or united urino-genital apertures. The ova are small and numerous, and the segmentation is either holoblastic and unequal, or meroblastic. Besides a large number of fossil forms the group includes the vast majority of living Fishes.

The Teleostomi include four "Orders," the CROSSOPTERYGII, the CHONDROSTEI, the HOLOSTEI, and the TELEOSTEI. Of these the Crossopterygii occupy a remarkably central position. Remotely connected with the Elasmobranchs on the one hand, and more intimately related to the Holostei and Teleostei on the other, they also probably represent the ancestral stock from which the Stegocephalan Amphibia and the Dipneusti have had their origin. Of the three remaining groups, often collectively spoken of as "Actinopterygii," the Chondrostei are the oldest and most primitive. Like the Crossopterygii, they are not without evidence of a remote kinship with the Elasmobranchs, but in a broad general sense they also represent the initial stages in a sequence of structural modifications, of which the Teleostei, the dominant Fishes of the present day, are the final outcome.

ORDER I. CROSSOPTERYGII.

Pectoral fins obtusely lobate and probably uniserial, or acutely lobate and probably biserial. Pelvic fins abdominal in position, uniserial, non-lobate, or obtusely lobate. Scales rhombic or cycloid, and, like the dermal cranial bones, they are generally invested by a layer of enamel-like ganoin. Tail heterocercal, or apparently diphycercal or gephyrocercal. Vertebral column acentrous, or with ring-like centra, or even with complete bony amphicoelous centra. Lower jaw with dentigerous splenials. As a rule, the opercular series includes an operculum and a suboperculum. Branchiostegal rays absent, their place being taken by a remarkable armature of jugular plates (Fig. 274). Secondary pectoral girdle {477}complete, including a pair of infra-clavicles. With rare exceptions the fin-rays of the median fins retain their numerical preponderance over the supporting radials. The group is divisible into two "sub-orders," the OSTEOLEPIDA and the CLADISTIA.[561]

SUB-ORDER 1. OSTEOLEPIDA.

The obtusely or acutely lobate pectoral fins articulate with the pectoral girdle by a single basal endoskeletal element. Nostrils on the ventral surface of the snout. Two dorsal fins and an anal fin. Dermal bones of the ethmoid region often fused with one another and with the premaxillae in front and the frontals behind to form a continuous rostral shield. Infra-dentary bones may be present. A series of lateral jugular plates often present in addition to the pair of principal plates. The Osteolepida first make their appearance in the Old Red Sandstone and Devonian formations, where they become abundant. They are also well represented in the Carboniferous, but only one family survived to the Mesozoic period, finally becoming extinct in the Upper Cretaceous. The following are the more important families:—

FAM. 1. OSTEOLEPIDAE.—Scales rhombic and thickly enamelled. Pectoral and pelvic fins obtusely lobate. Tail heterocercal. Teeth simple, not complicated by surface infoldings except quite at the base. Genera:—Osteolepis (Fig. 273), Thursius, Diplopterus (Middle Old Red Sandstone, Scotland), Glyptopomus (Upper Old Red Sandstone, Scotland), {478}Megalichthys (Carboniferous and Lower Permian of Europe and North America).

FAM. 2. RHIZODONTIDAE.—Scales cycloid and overlapping. Paired fins obtusely lobate. Tail heterocercal, sometimes apparently gephyrocercal. Teeth with the external enamelled layer of dentine infolded towards the axis in the form of radially arranged folds. In some genera ring-like vertebral centra have been recognised and also a preoperculum. Genera:—Rhizodus, Lower Carboniferous of Scotland and Northumberland; Tristichopterus[562] (Fig. 275), Old Red Sandstone of Scotland; Eusthenopteron[563] (Fig. 276), Upper Devonian of Scaumenac {479}Bay, Canada; Gyroptychius, Old Red Sandstone, Scotland; Rhizodopsis[564] (Fig. 274), Carboniferous of England, Scotland, Silesia, and North America; Strepsodus, Carboniferous of Great Britain, Ireland, and North America.

FAM. 3. HOLOPTYCHIDAE (Dendrodontidae).—Scales cycloid. Pectoral fins acutely lobate; pelvic fins short and somewhat obtusely lobate. Tail heterocercal. Teeth similar to those of the Rhizodontidae but more specialised, the enamelled dentine infoldings being much more complicated, presenting a radiating {480}arborescent appearance in transverse sections. Vertebral column acentrous. Genera:—Holoptychius[565] (Fig. 277), Old Red Sandstone of Scotland; Devonian of Belgium, Russia, North America, and East Greenland. Glyptolepis has a similar range.

FAM. 4. COELACANTHIDAE.[566]—Scales cycloid. Paired fins obtusely lobate. Tail symmetrical but apparently gephyrocercal, usually with a protruding axial vestige of the disappearing terminal part of the tail and of the proper caudal fin. Radialia of the functional caudal lobes agree in number with the contiguous neural and haemal arches and dermal fin-rays, the diagnostic feature of Smith Woodward's Actinistia. Proximal radials of the dorsal and anal fins fused into a single, internally-forked basipterygium in each fin. Teeth simple. Vertebral column acentrous. The skull presents several interesting features. The hyomandibular and the palato-quadrate bar, for example, are fused on each side into a continuous triangular bone, articulating with the cranium above and with the lower jaw below. The opercular skeleton is reduced to an operculum and two jugular plates. A very singular feature in these Fishes is the ossification of the walls of the air-bladder (Fig. 278), a structural modification which has no parallel in Fishes, except in certain Teleosts (Siluridae and Cyprinidae)[567] {481}in which the organ becomes encapsuled by bone owing to the partial ossification of its walls.

From their first appearance in the Lower Carboniferous the Coelacanthidae range, practically unchanged, through the intervening formations to the Upper Cretaceous. Coelacanthus itself occurs in the Carboniferous and Permian of England, Scotland, and Germany, and in the Carboniferous of North America. Undina[568] (Fig. 278) is a Jurassic genus. Diplurus is found in the Trias of North America, and Macropoma is a well-known form from the Middle and Upper Cretaceous beds of England, and other parts of Europe.

SUB-ORDER 2. CLADISTIA.

Pectoral fins uniserial and abbreviate, with three basal endoskeletal elements. Nostrils on the upper surface of the snout. Entire skeleton well ossified. Notochord replaced by bony, amphicoelous vertebral centra. Bones of the ethmoid region not fused to form a rostral shield. Infra-dentary bones absent. Jugular plates reduced to a single pair of large plates. As this group includes the only Crossopterygii which have survived to the present day, it is noteworthy that they retain certain primitive features indicative of their remote origin. The spiracles are persistent; the intestine has a spiral valve; and the conus arteriosus is furnished with several rows of valves. Amongst other characters of contrary significance, the air-bladder is double; its oesophageal aperture is ventral; and its afferent arteries are pulmonary arteries derived from a posterior aortic arch.

FAM. 5. POLYPTERIDAE.[569]—Pectoral fins obtusely lobate. Pelvic fins non-lobate. Scales rhombic and thickly enamelled. Dorsal fin in the form of a series of isolated finlets, each consisting of a stout spine-like[570] fulcral scale supporting a single soft ray, or a fringe of several rays, along its hinder margin. Tail symmetrical, apparently gephyrocercal. Teeth simple. Nostrils tubular.

The only representatives of the sub-order and the sole {482}surviving family of Crossopterygii, the Polypteridae, are restricted to the Nile and to the river basins of tropical Africa which drain into the Atlantic (Fig. 280). Only two genera are known, Polypterus and Calamichthys, neither of which has yet been discovered in any geological deposits, ancient or recent.

In Polypterus each of the spines of the dorsal fin supports several soft rays. Pelvic fins and a suboperculum are present. Ten species are known, of which six pertain to the Congo and its tributaries.[571] P. bichir is said to attain a length of four feet.

Until recently little was known of the habits of Polypterus, but the observations of Budgett[572] on the widely distributed P. senegalus and those of Harrington[573] on P. bichir, have brought to light many interesting facts about these most interesting Fishes.

P. bichir haunts the deeper holes and depressions of the muddy bed of the Nile, although it is "not essentially a bottom-liver or a mud-fish." It is most active at night when in search of food, and then it may readily be taken by trawl lines. The lobate pectoral fins are used for progression, but their primary function is to act as balancers, and they exhibit the characteristic trembling movements so often seen in the balancing fins of Teleosts. Polypterus does not readily live out of water, rarely longer than three to four hours, and then only when covered with damp grass or weeds. P. bichir is said to feed on small Teleosts, which it swallows whole, and to these there may be added in other species, Batrachians and Crustaceans. The observations of Budgett show that in captivity Polypterus often remains motionless for a long time at the bottom of the water, the anterior part of the body resting upon the tips of the {483}pectoral fins. According to the same observer, the air-bladder is an accessory respiratory organ, supplementary to the gills, rather than a hydrostatic organ.

In P. bichir the eggs ripen from June to September, inclusive, and, as in most other Nile Fishes, the breeding season is during or just after the period of inundation. P. senegalus and P. lapradei spawn during the rainy season in the months of July, August, and September, but nothing is certainly known as to the place or mode of deposition of the eggs. During the breeding season Polypterus is unusually active and excitable, and at this period the anal fin of the male becomes greatly thickened and enlarged, and has its surface thrown into deep folds between the successive fin-rays.[574] The use of the modified fin is not known. During his stay at McCarthy Island, about 160 miles up the River Gambia, Budgett[575] was fortunate in securing a larva of P. senegalus, 1 to 1¼ inches in length, or only about one-third the length of any larval Polypterus previously known (Fig. 281). The larva is described as a most beautiful object, "marked with black stripes on a golden ground, with a conspicuous golden stripe on each side above the eye, across the spiracle, and along the dorsal surface of the external gill." The pinnate external or cutaneous gills were relatively of much greater size than in the considerably more advanced stage figured elsewhere,[576] and reached half-way to the tail. The dorsal fin is not divided into finlets, and behind it is continuous with the caudal, while the anal fin is scarcely distinct from the {484}lower lobe of the caudal. The fin-rays which support the ventral portion of the caudal fin are more numerous and longer than those in relation with the dorsal lobe, and hence at this stage the tail is really heterocercal.

In the genus Calamichthys the body is greatly elongate and Eel-like in shape. Pelvic fins are absent, and normally there is no suboperculum. The dorsal finlets are more isolated than in Polypterus, and each spine supports but a single soft ray. Only a single species is known, C. calabaricus[577] (Fig. 282).

Calamichthys has a more restricted distribution than Polypterus, and is confined to certain rivers of West Africa. First obtained at Creek Town on the Old Calabar river, it is now known to occur in the delta of the Niger, on the coast of Cameroon, and as far south as the river Chiloango, frequenting the smaller muddy rivers opening into the estuaries.[578] It is a {485}very agile Fish, swimming like a snake, and subsisting on insects and crustaceans. The anal fin is enlarged in the male, and the young are provided with cutaneous gills. Calamichthys may attain a length of nearly 40 cm.

* * * * *

In the remaining Teleostomi (ACTINOPTERYGII) the paired fins are invariably non-lobate, with abbreviate, multibasal endoskeletal supports. Fin-rays are the main support of both the median and paired fins. Jugular plates are usually replaced by branchiostegal rays, but both may co-exist. The Actinopterygii are the successors of the Crossopterygii in palaeontological sequence, and when the latter began to decline in Carboniferous and Permian times, the former, mainly represented by the earlier Chondrostei, had already become the dominant Fishes of the period.

ORDER II. CHONDROSTEI (ACIPENSEROIDEI).

In these Fishes, the oldest and the most primitive of the Actinopterygii, the fin-rays of the median fins still continue to retain their primitive numerical superiority over the radials, and the tail is heterocercal. There is a single dorsal and an anal fin, which, like the upper lobe of the caudal fin, are generally provided with fulcra. Pelvic fins abdominal. Squamation typically rhombic and ganoid. Vertebral column acentrous. So far as is known the chondrocranium is but little ossified, and the cranial bones are mainly dermal. The secondary pectoral girdle still includes a pair of infra-clavicles.

The Chondrostei are first represented in the Lower Devonian by the solitary Palaeoniscid genus Cheirolepis, a contemporary of the earliest Crossopterygii. They occur throughout the Mesozoic period, except in the Cretaceous, and also in the Eocene, and while steadily diminishing in number and variety they gradually approximate to their degenerate and in some respects highly specialised descendants, the Sturgeons and Paddle-Fishes of the existing Fish fauna. Of the seven families included in the group the Palaeoniscidae are the oldest and the most generalised. The Platysomidae are a specialised offshoot from the Palaeoniscidae, and, if they are rightly to be considered as Chondrostei, perhaps the same may be said of the problematic Belonorhynchidae. On {486}the other hand, there are certain features in the Catopteridae which indicate an approach to Fishes of an altogether more modern type. Finally, the Chondrosteidae represent a stage in a career of degeneration, the climax of which is reached by the modern Polyodontidae and Acipenseridae.

FAM. 1. PALAEONISCIDAE.[579]—Fishes with fusiform bodies, short dorsal and anal fins, and usually with a complete investment of articulating rhombic, rarely cycloid, ganoid scales (Fig. 283). Fulcra generally present at the bases of the median fins, and especially along the dorsal border of the upper caudal lobe. Ribs are not known to be present. Skull invested by a very complete series of paired dermal bones, which in number and disposition conform to the normal Teleostome type (Fig. 284). The secondary upper jaw includes both premaxillae and large maxillae; and, as a rule, both the dentary and splenial bones {487}of the lower jaw are dentigerous. Except for the absence of an interoperculum, the opercular series of bones is complete, including numerous branchiostegal rays. There is a single small median jugular plate.

The Palaeoniscidae are remarkable both for their individual and specific abundance and for their extensive range in time. Represented only by Cheirolepis in the Middle Old Red Sandstone and Devonian, the family attained its maximum development in the later Palaeozoic rocks (Carboniferous and Lower Permian), became rare in the Mesozoic, finally dwindling away at the close of the Jurassic period. Their geographical distribution in the past is hardly less remarkable. In various geological formations they have been found in Great Britain and Ireland, in widely remote parts of continental Europe, and in North America, South Africa, and Australia. Cheirolepis, Amblypterus, Canobius, Phanerosteon, Elonichthys, Cryphiolepis, Palaeoniscus, and Trissolepis are Palaeozoic genera. Gyrolepis, Urolepis, Coccolepis, Oxygnathus, and Centrolepis are characteristic Mesozoic forms.

FAM. 2. PLATYSOMIDAE.[580]—More or less deep-bodied Fishes, with elongated dorsal and anal fins, a high head, short jaws, usually armed with bluntly conical tritoral teeth, and a complete investment of high, narrow, rhombic scales. They agree with the Palaeoniscidae in their osteology and in most other essential {488}features, and they flourished in large numbers during the Carboniferous and Permian periods. Platysomus ranges from the Lower Carboniferous to the Upper Permian in Great Britain and continental Europe, and also occurs in the Carboniferous of North America. Eurynotus (Fig. 285), and the singularly deep-bodied Cheirodus (Fig. 286), in which pelvic fins are unknown, are British Carboniferous genera.

FAM. 3. BELONORHYNCHIDAE.—The systematic position of these Triassic forms is very doubtful, and it is by no means clear that they are Chondrostei at all.

FAM. 4. CATOPTERIDAE.—It is very probable that this widely-distributed Triassic family is an offshoot from the Palaeoniscidae. It agrees with the latter in the general character of the head and pectoral girdle and in the rhombic squamation, but differs from its progenitors and approaches the more modern Holostei in the semi-heterocercal condition of the tail, and in the approximate numerical agreement between the fin-rays and radialia of the dorsal and anal fins.[581]

{489}FAM. 5. CHONDROSTEIDAE.—This family affords an interesting annectant link between the Palaeoniscidae and their degenerate living representatives the Polyodontidae and Acipenseridae. They agree with the latter in the general shape of the body, the growth of a preoral rostrum, and in the relatively small size of their ventrally-placed and probably protrusible mouth (Fig. 287). The skin is entirely scaleless, except on the upper lobe of the caudal fin, where, as in Polyodon and Acipenser, the primitive rhombic squamation and a series of fulcra are retained.

On the other hand, their relationship to the Palaeoniscidae is indicated by the general disposition of the dermal bones of the cranial roof, and the presence of a transverse row of supra-temporals and of an extensive series of branchiostegal rays (Fig. 288). The family is represented by Chondrosteus[582] from the Lower Lias of Dorset and Leicestershire, and Gyrosteus from the Upper Lias of Yorkshire. From an evolutionary point of view it is significant that the Chondrosteidae do not make their appearance until the Palaeoniscidae are approaching extinction.

The two remaining families, the Polyodontidae and the Acipenseridae, agree in presenting a remarkable leaven of characters otherwise distinctive of the typical Elasmobranch, associated with certain primitive features which they have doubtless inherited from some remote ancestral stock common both to existing Elasmobranchs and to the other primary groups of Fishes, and also with others obviously due to degeneration.

{490}[Illustration: FIG. 288.—Lateral view of a restored skull and pectoral girdle of Chondrosteus acipenseroides. a, Angular; br, branchiostegal rays; c.h, cerato-hyal; h.m, hyomandibular; j, jugal; p.f, post-frontal; s.op, suboperculum; s.t, supra-temporal; other reference letters as in Fig. 284. (After Traquair.)]

The most interesting illustration of the first point is to be found in the condition of the primitive upper jaw which, especially in the Polyodontidae, is typically Elasmobranch in the median union of the palato-quadrate bars beneath the basis cranii, but Teleostome in the presence of a secondary upper jaw formed by two maxillae. Both families also agree in possessing an acentrous vertebral column which, if it does so far resemble that of Teleostomes in being potentially arco-centrous, nevertheless has a better developed series of distinct inter-dorsal and inter-ventral cartilages, regularly alternating with only partially bony basi-dorsals and basi-ventrals, than is to be met with in any other adult Fishes except Elasmobranchs. Primitive features are apparent in the presence of spiracles, sometimes associated with pseudobranchs; the presence in one family (Acipenseridae) of a hyoidean hemibranch supplied with blood directly from the ventral aorta, and the existence of a multi-valvular conus arteriosus and an intestinal spiral valve. Finally, the massive growth of the chondrocranium wholly devoid of cartilage bones, except in so far as they may be represented by splint-like membrane bones, the fragmentation of the investing dermal bones, the degeneration of the opercular skeleton and the loss of branchiostegal rays, and the almost complete disappearance of the primitive rhombic squamation, are probably to be regarded {491}as the outcome of a long-continued career of degeneration from some remote Palaeoniscid ancestor.

FAM. 6. POLYODONTIDAE.—The Polyodontidae are more generalised, and in some features decidedly more Selachioid than the Acipenseridae. Body fusiform and apparently scaleless, but the primitive squamation is still represented by isolated vestigial scales imbedded in the otherwise soft skin, and by a continuous series of rhombic scales on the upper caudal lobe, which also has a dorsal fringe of large fulcra.[583] Rostrum exceptionally long, spatulate or somewhat conical, with a rigid axis and thinner and more flexible margins. Barbels absent. Mouth wide, not spout-like. Pectoral fins devoid of spines. Two pairs of membrane-closed vacuities separate the paired dermal bones of the cranial roof (possibly parietals and frontals) from the more laterally-placed post-temporals and squamosals, and there are no median plates posterior to the orbits, nor any representatives of supra-temporals. A feeble suboperculum is retained in addition to a small rayed operculum. Hyoidean hemibranch completely suppressed. Two genera only are known, each with a single species.

The Paddle-Fish or Spoon-Bill, Polyodon folium (Fig. 289) inhabits the rivers of the Southern States of North America, the Mississippi, Ohio, and Missouri, and their numerous tributary rivers and streams. A Fish of sluggish habits, Polyodon feeds chiefly on mud and the minute organisms it contains, the exceptionally long gill-rakers probably forming an efficient filter to prevent the food particles escaping through the gill-clefts with the expiratory water current. The singular rostrum is apparently used for stirring up the mud when feeding, but in view of the muddy waters the Fish frequents, and the very small size of {492}the eyes, its value as a tactile organ must not be overlooked. Polyodon may attain a length of 5 to 6 feet. The time of spawning varies, according to locality, from March to June. Nothing is known of the development of Polyodon. Young less than 6 to 8 inches in length are unknown, and specimens of this size are very rarely seen. The jaws are furnished with minute teeth until the Fish is about half-grown, when they become edentalous. Caviare is made from the eggs, and the centres at which this industry is carried on are chiefly situated along the course of the Mississippi. The second species, Psephurus gladius, inhabits the Yang-tse-Kiang and Hoangho rivers of China, and differs from Polyodon in the conical shape of its rostrum and the smaller number and larger size of its fulcra. Psephurus is stated to reach a length of 20 feet. The family is represented in the Eocene of Wyoming by the genus Crossopholis, which is note-worthy for the retention of trunk scales in the form of small, somewhat quadrate denticulated discs, arranged in oblique rows.

FAM. 7. ACIPENSERIDAE.—In the Sturgeon family the body is elongate, cylindrical, and somewhat bulky. Rostrum well developed and often massive, with a transverse row of simple or branched preoral barbels on its ventral surface. Mouth small and remarkably protrusible. Jaws devoid of teeth except in the larvae. As in the preceding family, the primitive rhombic squamation is confined to the upper lobe of the tail, which, like the dorsal and anal fins, is furnished with fulcra. Elsewhere the scales are represented by five longitudinal rows of large bony scutes and by intervening small scattered ossifications. The anterior dermal ray of the pectoral fin is stout and spine-like. The dermal bones of the cranial roof suturally articulate with one another to form a continuous shield, uninterrupted by lateral vacuities. A median dermal bone in the occipital region transmits the occipital sensory canal. The opercular series is represented only by an opercular bone.

The family includes but two genera, Acipenser (Fig. 290) and Scaphirhynchus, and about twenty species, confined to the seas, estuaries, and rivers of the temperate and north temperate regions of the northern hemisphere. Acipenser includes the more typical Sturgeons, and is distinguished by the presence of spiracles, and by the fact that the longitudinal rows of scutes remain distinct to the base of the caudal fin.

{493}[Illustration: FIG. 290.—The Sterlet (Acipenser ruthenus). o, Barbels; c.f, caudal fin; d.f, dorsal fin; pct.f, pectoral fin; pv.f, pelvic fin; sc, scutes; v.f, ventral or anal fin. (From Parker and Haswell, after Cuvier.)]

There are probably about fifteen species, but the exact number is uncertain. Sturgeons are abundant in the Black Sea, the Sea of Azov, the Caspian, and their tributary rivers, notably the Danube, Don, Dnieper, Ural, and Volga. They are also present in the rivers and on the coasts of Northern Europe and of China. Five species occur in North America, on the Atlantic and Pacific coasts, and in the rivers of these regions as well as in the Great Lakes.[584] One or two species are almost exclusively fresh-water, but most Sturgeons are migratory Fishes, living in the sea, but ascending rivers for spawning. Their food consists of worms, molluscs, the smaller Fishes and aquatic plants; and in feeding the mouth is protruded downwards in the form of a cylindrical, spout-like structure and thrust into the mud. The only species certainly known to frequent the British coasts is the common Sturgeon (A. sturio), which is also found in the Black Sea and the Mediterranean, and is abundant on the Atlantic coast of North America from Maine to South Carolina. The species occurs all round our coasts, more plentifully, perhaps, on the northern and eastern shores. In the spring and summer the Fish ascends the rivers, often to a considerable distance. Its presence has been recorded in the Severn, near Shrewsbury; in the Trent at Nottingham, and also, but not in recent years, in the Thames above London Bridge.[585] In this country the species is a "Royal Fish," and by an unrepealed Act of Edward II. it is enacted that "the King shall have the wreck of the sea throughout the realm, Whales and Great Sturgeons, except in certain places privileged by the King."[586] If not so large as some of its Russian relatives, A. sturio often attains a great size. Even on {494}our own coasts the capture of individuals 8 to 10 feet in length has been recorded. The great Russian Sturgeon (A. huso), which is common in the Black Sea, the Sea of Azov and the Caspian, and in the rivers flowing into them, is the largest of all the Sturgeons, individuals weighing 2760 and 3200 pounds having been captured. The Sterlet (A. ruthenus), similarly distributed and often ascending the Danube to Vienna, is much smaller, rarely exceeding a length of three feet.

In Europe A. sturio spawns about July, but in North America (Delaware river) during May. Small in size, the eggs are produced in enormous numbers, a single female, it is said, producing about 3,000,000 in one season. They are invested by a gelatinous sheath, so that they readily stick to one another or to other objects, and, when deposited, they adhere in streaks or sheet-like masses to the bed of the river. The young are hatched very early, about the third or fourth day in A. sturio, and in the Sterlet between the ninth and twelfth, the length of the larva then varying from 7 to 10 mm. When they are a few days old the larvae closely resemble those of existing Holostei except that the small opercular folds leave the gills freely exposed (Fig. 291). A shallow pigmented groove in front of the mouth apparently represents the sucker of the young Amia and Lepidosteus. Although toothless in the adult, both the Sturgeon and Sterlet possess vestigial rudimentary, uncalcified, larval teeth, which in shape resemble the teeth of a Dog-Fish, consisting of a broad base and a sharp spine.

The Sturgeon is a Fish of considerable economic importance. The flesh is an article of food, and from the ovaries of certain Russian and American species thousands of hundredweights of {495}caviare are prepared annually. Large quantities of isinglass are obtained from the air-bladders, in the United States and in Russia. The organ is split open and washed; the inner lining is then stripped off and the bladder dried as rough isinglass.

The second genus, Scaphirhynchus, which includes the Shovel-nosed Sturgeons, differs from Acipenser in the long, flattened, and almost spatulate shape of the rostrum, the suppression of the spiracles, and the union of the longitudinal rows of scutes beneath the dorsal fin to form a scaly armature completely investing the tail. The distribution of the genus affords an interesting parallel to that of the Polyodontidae. Of the four species, one (S. platyrhynchus) is common in the Mississippi valley and in the rivers of the Western and Southern States of North America, while the remaining species, also exclusively fresh-water, frequent the rivers of Tartary.

The Acipenseridæ are not known to occur earlier than the Tertiary. Scutes, pectoral spines and fragmentary bones, indistinguishable from the corresponding parts of existing species, have been recorded from the London Clay of the Isle of Sheppey (Lower Eocene), and from later Eocene deposits in the Isle of Wight and Hampshire; and also from the Pliocene of England (Red Crag of Suffolk) and Virginia.

ORDER III. HOLOSTEI (LEPIDOSTEOIDEI).

The Holostei include a large and somewhat heterogeneous assemblage of Fishes, most of which are now extinct. As a group they are by no means easy to define or delimit. Widely separated from the Chondrostei, there is little evidence of the existence of connecting links between the two groups, although in some respects the Catopteridae may be regarded as transitional. On the other side, however, the Holostei shade off almost imperceptibly into the Malacopterygian Teleostei. In different fossil and recent Holostei there may be traced the gradual acquisition of the more special Teleostean characters and the elimination of the more archaic features of their remote Teleostome ancestors; and in a general sense this may be taken as the key to the more salient attributes of the group. It is not suggested that all the families of Holostei are on the direct lines of Teleostean descent. Some families, like the Eugnathidae and {496}Amiidae, may possibly occupy this position, but others, such as the Pycnodonts, for example, seem to be highly specialised and terminal offshoots which have left no descendants. Of the more generalised features which different Holostei retain, mention may be made of the prevalence of rhombic scales which, like the dermal cranial bones, are generally invested by a variously ornamented coat of ganoin; the presence of fulcra, cheek-plates, post- or sub-orbital ossicles, and of a complex lower jaw, which includes dentigerous splenials; and the abdominal position of the pelvic fins. On the other hand, indication of advancing specialisation in the Teleostean direction are to be noted in the numerical agreement between the dermal fin-rays of the median fins and their supporting radialia, and in the character of the vertebral column. Some Holostei, especially the earlier forms, are acentrous, but between this primitive condition and the possession of well-ossified centra, associated with equally bony arcualia, almost every gradation is to be found. The chondrocranium is more or less completely replaced by cartilage bones corresponding to those generally present in Teleosts, while the palato-pterygoid cartilages, likewise modified by the growth of cartilage bones, separately articulate with the lateral ethmoid regions instead of meeting in a ventral symphysis beneath the basis cranii. With rare exceptions (e.g. certain Pycnodonts) the opercular skeleton is complete, and includes branchiostegal rays; and although a single gular plate is often present, it may be absent in entire families. Like so many other structures, the tail is in a transitional state: really heterocercal, but incipiently homocercal, it may be described as semi-heterocercal. Infra-clavicular plates no longer form part of the secondary pectoral girdle, their place being taken by cleithra which, as in most Teleosts, meet in a ventral symphysis.

Indications of transition are not wanting in the squamation in certain families, and may be seen in the partial or complete replacement of the rhombic type by thin, imbricated, cycloid scales. Lastly, the soft parts of the two surviving genera are not without features of similar significance. A multivalvular conus arteriosus, it is true, is still retained, but the spiral valve is vestigial, the spiracles are closed, and in the female of one genus (Lepidosteus) the gonoducts are peritoneal tubes, continuous, as in most Teleosts, with the investments of the ovaries.

{497}The Fishes here included in the Holostei constitute the Protospondyli and Aetheospondyli of Smith Woodward.[587] In the former group vertebral centra are either entirely absent, or, if present, their components in the form of alternating hypo- and pleuro-centra invariably remain distinct in the tail. The latter group has been instituted for the provisional reception of two highly specialised families of uncertain relationships, which differ from the Protospondyli in their higher grade of vertebral structure, the centra always being complete without any indication of distinct hypo- and pleuro-centra.

The Holostei first appear in the Permian, where they are represented by a single genus (Acentrophorus). During the Mesozoic period they were abundant in the Trias, reaching their maximum development and becoming the dominant Fishes of the period in the Jurassic. In the Cretaceous they began to decline, and in the Tertiaries became reduced to the two families which at the present day are the sole survivors of the group.

Of the six families of Protospondyli the Semionotidae are the oldest and most generalised, and the Macrosemiidae a closely allied group. The Pycnodontidae are a highly specialised and terminal offshoot. The Eugnathidae obviously lead to the Amiidae, and from the same stock it is probable that the Pachycormidae have been derived. The relations of the Aspidorhynchidae and Lepidosteidae (Aetheospondyli) are extremely doubtful. That the two families are allied seems probable, but beyond the possibility of a remote connection with the Protospondyli there is no clue to their ancestry.

FAM. 1. SEMIONOTIDAE.—Small-mouthed, fusiform or {498}deep-bodied Holosteans with rhombic scales, rarely, as in Aetheolepis, cycloid in the caudal region. All the fins possess fulcra. Teeth more or less conical, with a tendency to become tritoral in certain genera. Jugular plate present or absent. Acentrophorus (Upper Permian); Semionotus (Trias of England, Germany, S. Africa, and N. America); Lepidotus (Fig. 292) (Trias of Germany, Jurassic of Europe and India, Cretaceous of Brazil); the deep-bodied Dapedius (Lias of Dorset, Fig. 293), and Aetheolepis (Jurassic of New South Wales) are characteristic genera.

FAM. 2. MACROSEMIIDAE.—Closely related to the Semionotidae, but with a more extended dorsal fin. Macrosemius (Upper Jurassic of England, Germany, France); Notagogus (Upper Jurassic of Naples, Bavaria, France); Petalopteryx (Upper Cretaceous of Syria).

FAM. 3. PYCNODONTIDAE.—Highly specialised deep-bodied Fishes, with a small mouth and characteristic grinding or tritoral teeth. Scales rhombic. Fulcra absent. Dorsal and anal fins long. There is no jugular plate. The family ranges from the Lower Lias to the Lower Eocene, inclusive. Mesodon, Mesturus, Gyrodus, and Microdon are Jurassic genera. Coccodus and Xenopholis occur in the Upper Cretaceous of Syria (Mount Lebanon), and Pycnodus in various European Eocene formations.

FAM. 4. EUGNATHIDAE.—Large-mouthed, elongate fusiform {499}predaceous Fishes, with pointed teeth, rhombic scales, short dorsal and anal fins, a single jugular plate and prominent fulcra. The vertebral centra are represented by distinct hypo- and pleuro-centra, which may form complete alternating rings in the tail.

The family first appears in the Trias and ranges throughout the Jurassic period. Eugnathus (Jurassic) and Eurycormus (Upper Jurassic). Caturus (Fig. 294) has a more extensive range, occurring in the Upper Trias of the Tyrol and in the Upper Jurassic of England and Bavaria. Caturus and Eurycormus, with their relatively thin, imbricated, cycloid scales, which have lost the peg-and-socket articulation, form connecting links between the more typical Eugnathus and the Amiidae.

FAM. 5. AMIIDAE.—Body fusiform and somewhat compressed. Scales uniformly thin, cycloid, and imbricated. Single dorsal fin long and low. Anal fin short. Tail nearly homocercal, with a rounded hinder margin. Fulcra absent from all the fins. Moderately large conical teeth are present on the premaxillae, maxillae, palatines and dentaries, and smaller teeth on the vomers, pterygoids, splenials and parasphenoid. Pre- and post-centra fused in the trunk, forming complete bony amphicoelous centra, but distinct in the tail. A single large jugular plate is present. In the solitary living species the air-bladder is cellular, and its afferent arteries are derived from a posterior aortic arch. Pyloric caeca absent. Two peculiar comb-like structures are present on the throat.

The Bow-Fin (Amia calva), the sole existing representative of the family, is abundant in the rivers and lakes of Central and Southern North America, including the great lakes Huron and {500}Erie. It is a voracious, carnivorous Fish, preying upon other Fish as well as upon fresh-water Crustaceans and Insects, very tenacious of life, and of no economic value. The male is smaller than the female, about 18 inches in length, and is distinguished by the presence of a round black spot, encircled by a margin of orange, at the base of the caudal fin (Fig. 295). The female may exceed 24 inches.

Amia frequently rises to the surface, especially when the water is foul, and takes in large mouthfuls of air, and it is probable that the air is subsequently passed into the spacious cellular air-bladder which acts as a lung. The breeding season, during which the coloration of the Fish is more brilliant than at other times, lasts from the beginning of May to June, but it may begin and end somewhat earlier if the temperature be favourable. The Fish makes its way from the deeper water, where it has remained sluggish during the winter, to the spawning ground. This is usually at the swampy end of a lake where there is an abundance of aquatic herbage intersected by channels of clear water. There the Fish is said to circle round until the soft weeds and rootlets are bent and crushed aside, so as to leave an area having the appearance of a crude form of nest,[588] in which the eggs are deposited. They may be found in enormous numbers adhering to the leaves and rootlets of the weedy home. After oviposition the male remains on guard until the young are hatched out, when they appear to leave the nest in a body, still under the protection of their watchful parent. At all events a little later the male has been observed to be accompanied by a swarm of young fry, which he keeps together by circling round them. The development of the eggs is remarkably rapid. From the first cleavage of the egg to the hatching of the embryo the whole {501}process may be completed within from 4 to 8 days. When hatched the larvae are about 5 to 6 mm. long. They possess a large yolk sac and a preoral sucker for attachment. The pectoral fins are conspicuous structures before there is any trace of the pelvic fins.

The Amiidae, represented by Megalurus,[589] first appear in the Upper Jurassic of Dorset, France, and Bavaria. In the Cretaceous period the family is represented by Amiopsis. Species of Amia occur in the Eocene of Europe and North America. In the former area the genus became extinct at the close of the Lower Miocene period, but in the latter Amia calva still survives.

FAM. 6. PACHYCORMIDAE.—Large-mouthed, predaceous Amioid Fishes with a more or less prominent snout and a short dorsal fin. Scales rhombic but thin, rounded behind, and overlapping, sometimes absent. A single large jugular plate.

In the earlier forms (e.g. Pachycormus, Lias) the snout is but slightly produced, but in Hypsocormus (Upper Jurassic), and {502}especially in Protosphyraena (Cambridge Upper Greensand and the Cretaceous of Europe and North America), it becomes greatly elongated and associated with an exceptionally strong dentition.

FAM. 7. ASPIDORHYNCHIDAE.—Long-bodied Fishes, with a pointed preoral rostrum, sharp teeth, and deep rhombic scales. Fins small, the dorsal and anal being remote from the pelvic fins. Fulcra vestigial or absent. Jugular plates not known.

Two genera only are known. Aspidorhynchus is a Jurassic form. Belonostomus is Upper Jurassic and Cretaceous. Species of the latter genus have a very wide distribution (Europe, North and South America, and Australia).

FAM. 8. LEPIDOSTEIDAE.—Body elongate, with a relatively short caudal region. Tail semi-heterocercal. Scales rhombic, thick, ganoin-coated and articulated, not vertically elongated on the sides of the body. Dorsal and anal fins short and remote from the pelvic fins. Median fins with fulcra. Both the upper and lower jaws more or less elongated, forming a broad and depressed or a long tapering beak, near the anterior end of which the nostrils are placed. Eyes small. Vertebral centra well {503}ossified, opisthocoelous and fused with the neural arches. The metapterygoid bones have a secondary articulation with the skull.[590] Maxillae segmented into numerous pieces. Jugular plates absent. Branchiostegal rays reduced to three on each side. Teeth numerous, slender, and of unequal size. In the larger teeth the dentine is intricately folded. Pyloric caeca branched and compacted together into a gland-like mass. Air-bladder cellular, but its blood is not derived from a posterior aortic arch.

The only known genus is Lepidosteus, the existing species of which frequent the fresh waters of North America.[591] The common or Long-nosed Gar-Pike (L. osseus), remarkable for its long and slender beak, is generally abundant in the rivers and lakes of the United States from Vermont to the Rio Grande, and it may reach a length of five feet. The "Short-nosed Gar" (L. platystomus, Günther) has a much shorter and broader beak, and a similar distribution (Fig. 299). The "Great" or "Alligator Gar" (L. viridis, Günther) has a more southerly habitat, frequenting the rivers of the Southern States, Northern Mexico, and Cuba. It is by far the largest species, sometimes reaching a length of 8 to 10 feet.

Lepidosteus is a voracious Fish, preying upon smaller Fishes, and, except in the breeding season, it frequents the deeper parts of the rivers or lakes. The Fish is constantly in the habit of rising to the surface and emitting bubbles of gas, either through the mouth or by the branchial clefts, and it is probable that this gas is air which has been previously swallowed at the surface and passed into the air-bladder. About May Lepidosteus resorts in large numbers to shallower water, where the temperature is {504}higher, for the purpose of spawning, each female being attended by from one to four males.[592] During brief recurring periods of excitement, accompanied by convulsive lashing movements, the eggs and sperm are emitted. The eggs are extremely sticky, and adhere tenaciously to the rocks and stones on which they are deposited. In a few days the embryos hatch out, and at this stage the larva has a huge mouth surmounted by a terminal preoral disc, fringed with a row of marginal wart-like suckers (Fig. 300). The yolk sac is so large as greatly to hamper the movements of the larva; hence, by means of its suckers, the young Lepidosteus attaches itself to surrounding objects, and remains almost entirely motionless for some little time after hatching. Later, about a fortnight after escaping from the egg, the yolk becomes completely absorbed, the suckers degenerate and eventually disappear, and the larva, freed from its load of nutritive reserve, assumes a more active life. After the absorption of the yolk the larvae of Mosquitos appears to form the exclusive diet of the young Lepidosteus for some time, but very soon young Fishes are readily devoured.[593]

Lepidosteus seems to have been abundant in Europe during the Eocene and Miocene periods, but became extinct before the Pliocene. In North America, also, the genus dates from the Eocene, and still survives.

ORDER IV. TELEOSTEI.

[For the account of this Order, see pp. 541 f.]

{505}CHAPTER XIX

DIPNEUSTI

SUB-CLASS III. DIPNEUSTI (DIPNOI).

These singularly interesting Fishes are distinguished by their more or less acutely lobate paired fins and their overlapping cycloid scales, and by the fact that the bony dermal fin-rays of the median fins are much more numerous than their supporting radialia. Tail heterocercal or apparently diphycercal. Nostrils inferior. Vertebral column acentrous. The radialia of the median fins articulate with the contiguous neural or haemal spines and agree with them numerically. Skull autostylic. Premaxillae and maxillae absent, but a secondary lower jaw is represented by certain dermal bones of which tooth-bearing splenials are the most important, the dentary bones being absent altogether, or, if present, toothless and small. The cranial dermal bones include median as well as paired lateral plates, but their relations to those of other Fishes are very obscure. Two opercular bones are always present, but branchiostegal rays are unknown. One of the most important diagnostic features is the dentition. All Dipneusti agree in possessing large tritoral dental plates supported by the palato-pterygoid and splenial bones. The secondary pectoral girdle includes only cleithra and infraclavicles (clavicles). There is a pelvic girdle. Claspers absent. Of the four families of Dipneusti, two, the Ctenodontidae and the Uronemidae, are exclusively Palaeozoic. The third, the Ceratodontidae, is Mesozoic, and still survives. The fourth, the Lepidosirenidae, is known only by two existing genera.

{506}[Illustration: FIG. 301.—Restoration of Dipterus valenciennesi. × ⅕. (From Traquair.)]

FAM. 1. CTENODONTIDAE.—Body fusiform. Tail heterocercal or apparently diphycercal. Excluding the anal fin, which is always distinct, the remaining median fins are either distinct or continuous. Dental plates traversed by radiating transverse ridges terminating in rows of conical denticles (ctenodont dentition, Fig. 303). Vomerine teeth not known. Cranial bones numerous and small, and, like the squamation, with or without an investment of ganoin. Jugular plates present or absent.[594] The oldest genus is Dipterus[595] from the Old Red Sandstone of Scotland, where it is contemporaneous with the earliest Crossopterygii and also with the oldest known Actinopterygii (Cheirolepis). Phaneropleuron (Old Red Sandstone of Scotland, Upper Devonian of Canada, Fig. 302), Scaumenacia (Upper Devonian of Canada), Ctenodus (Carboniferous of Great Britain and North {507}America), and Sagenodus (Carboniferous of Great Britain and Lower Permian of Bohemia) belong to the same family.

FAM. 2. URONEMIDAE.—Body fusiform. Dentition non-ctenodont, consisting of patches of distinct rounded denticles with a row of basally-confluent conical denticles along the outer margin of each. Scales thin. All the median fins are continuous. Tail apparently diphycercal. Cranial dermal bones as in Dipterus. Uronemus[596] (Lower Carboniferous of Scotland), and perhaps Conchopoma[597] (Lower Permian of Prussia), are the only known genera.

The two remaining families possess certain features which cannot be affirmed to have existed in their extinct allies. Thus, both agree in exhibiting those striking and, so far as Fishes are concerned, unique modifications of the air-bladder and vascular system, and the olfactory organs,[598] which are more or less closely associated with air-breathing habits and indicate a marked convergence towards the Amphibia. Side by side with such indications of advancing specialisation in certain directions, ample evidence of a remote ancestry is to be seen in such primitive features as the presence of a spiral valve and a multi-valvular conus arteriosus, and in the short and simple alimentary canal. Of other points of agreement mention may be made of the absence of jugular plates, the presence of vomerine teeth, the continuity of all the median fins, and the apparently diphycercal but probably gephyrocercal character of the tail.

FAM. 3. CERATODONTIDAE.—Body elongated and compressed. Scales large, thin, non-ganoid, and partially enclosed in dermal pouches. Paired fins biserial. Chondrocranium complete. Dermal bones wholly devoid of ganoin, reduced in number but increased in size. Circumorbital bones present. Dental plates oval, crescentic or triangular, traversed by several radiating enamelled ridges, terminating in smooth or feebly denticulated biting margins. Lower jaw with a small toothless dentary on each side. The hyoid arch includes a small hyomandibular and a hypo-hyal in addition to a cerato-hyal. Branchial arches five in number and bisegmented. The gills exhibit little evidence of {508}degeneration. Hyo-branchial cleft open, and associated with a pseudobranch. The first four branchial arches carry holobranchs. Air-bladder single. Young not provided with cutaneous gills. Two genera only are known, the Mesozoic Ceratodus and the still living Neoceratodus. The former genus includes numerous species, for the most part known only by their dental plates, and has a remarkably wide distribution in different geological formations. Species occur in the Trias of England, Germany, India, South Africa (Upper Karoo strata), and also, but more rarely, in certain Jurassic deposits in England and in Colorado.[599] Neoceratodus is represented by a solitary species, N. forsteri[600] (Fig. 304, A), which is now restricted to the Burnett and Mary rivers in Queensland. A somewhat wider distribution of the species in recent times is indicated by the presence of teeth in the later Tertiary (alluvial) deposits of Darling Downs, near the borders of New South Wales.

The Neoceratodus[601] of the Burnett frequents the comparatively stagnant pools or water-holes which alternate with shallow runs and are usually full of water all the year round. In these pools, filled with a rich growth of aquatic vegetation, and often the favourite haunt of the Platypus (Ornithorhynchus), the Fish is fairly abundant. Inactive and sluggish in its habits, usually lying motionless on the bottom, the Fish is easily captured by the natives with hand-nets or baited hooks. Neoceratodus lives on fresh-water Crustaceans, worms, and molluscs, and to obtain them it crops the luxuriant vegetation of the water-holes much in the same way that a Polychaet or a Holothurian swallows sand for the sake of the included nutrient particles. Apparently the air-bladder is a functional lung at all times, acting in conjunction with the gills. At irregular intervals the Fish rises to the surface and protrudes its snout in order to empty its lung and take in fresh air. While doing so the animal makes a peculiar grunting noise, "spouting" as the local fishermen call it, which may be heard at night for some distance, and is probably caused by the forcible expulsion of air through the mouth. Useful as the lung is as a breathing organ under normal conditions, there can be little doubt that its value as such is much greater whenever gill-breathing becomes difficult or impossible.

{509}[Illustration: FIG. 304.—A, Neoceratodus forsteri, Queensland; B, Protopterus annectens, Gambia. C, Lepidosiren paradoxa, Paraguay. The lozenge-shaped markings on the surface of B do not represent scales but areas of the skin outlined by pigment cells. In a fresh specimen the scales are as completely invisible as in C. (A, from Günther; B and C, from Lankester.)]

{510}[Illustration: FIG. 305.—A young Neoceratodus four weeks after hatching. c, Cloacal aperture; l.l, lateral line; m, mouth; op, operculum; p.f, pectoral fin. (From Semon.)]

This seems to be the case during the hot season, when the water becomes foul from the presence of decomposing animal or vegetable matter. Semon records a striking illustration of this in the case of a partially dried-up water-hole, in which the water had become so foul that it was full of dead fishes of various kinds. Fatal as these conditions were to ordinary Fishes, Neoceratodus not only survived but seemed to be quite healthy and fresh. Such observations are of exceptional interest. Not only do they afford a clue to the conditions of life which, in the course of time, probably led to lung-breathing in Neoceratodus, but they also suggest the possibility that a similar environment has been conducive to the evolution of air-breathing Vertebrates from gill-breathing and Fish-like progenitors. In spite of its pulmonary respiration, Neoceratodus more closely resembles the typical Fishes in its habits than any other Dipneusti. It lives all the year round in the water. There is no evidence that it ever becomes dried up in the mud, or passes into a summer sleep in a cocoon, and the well-developed condition of its gills suggest that these organs play a more important rôle in breathing than in either Protopterus or Lepidosiren. The Fish is not known to leave the water, and the paired fins, useful no doubt as paddles, are quite incapable of supporting the bulky body on terra firma. In fact, when Neoceratodus is taken out of its natural element it seems to be more helpless than most other Fishes, and, in spite of its capacity for lung-breathing, soon dies unless kept moist by artificial means. Spawning takes place from April to November, principally in September and October. The eggs, invested by a jelly-like coat, secreted by the oviducal walls, are deposited {511}not in a nest, but singly amongst aquatic vegetation, and, as they are not adherent, it is probable that they finally rest on the mud. The early developmental stages exhibit a general resemblance to those of Amphibia. There is no larval metamorphosis, and at no period does the young Neoceratodus (Fig. 305) possess cutaneous gills or a cement organ. The tail is apparently diphycercal from the first, and the pelvic limbs do not appear until about six weeks after the pectoral members. It is interesting to note that the dental plates are first represented by lines or patches of separate denticles (non-ctenodont), which subsequently fuse basally (ctenodont) before the adult condition is reached.[602] Neoceratodus is stated to grow to a length of 5 to 6 feet.

FAM. 4. LEPIDOSIRENIDAE.—Body elongate, cylindrical and more or less Eel-like, with small cycloid scales completely enclosed in the skin. Paired fins so acutely lobate as to present the appearance of tapering cylindrical filaments, equally devoid of scales and fin-rays. In a general way the cranial dermal bones correspond with those of Neoceratodus, but the place of the posterior median bone is taken by a large, gable-like fronto-parietal bone, situated internal to the head muscles, and in direct relation with the chondrocranium, which is largely aborted in the interorbital region. Circumorbital bones absent. Opercular bones much reduced. Lower jaw without dentary plates. Palatine and splenial dental plates with three non-denticulate, trenchant ridges. Hyoid arch consists of cerato-hyals only. Hyoidean cleft closed. Certain of the anterior branchial arches devoid of branchial filaments; when present the latter are leaf-like and free. Air-bladder a double lung. There is a larval metamorphosis, and the young possess cutaneous gills. The family includes two genera, Protopterus and Lepidosiren. In the former genus the paired fins are either uniserial or they consist of axial mesomeres only; there are six branchial arches and five clefts; and the larval gills are usually retained as vestiges throughout life. In Lepidosiren the paired fins are reduced to the segmented axis, without pre- or post-axial radials. There are five branchial arches and four clefts, and the cutaneous gills disappear soon after the larval metamorphosis.

{512}[Illustration: FIG. 306.—Map showing the distribution of the surviving Dipneusti.]

Protopterus has a wide distribution over the middle portion of the great African continent, ranging from the river Senegal and the White Nile on the north to the Congo basin, Lake Tanganyika, and the Zambesi on the south. Three species are known, P. annectens (Fig. 304), P. aethiopicus, and P. dolloi. Protopterus[603] is usually found in marshes in the vicinity of rivers. Voracious in its habits the Fish is mainly carnivorous, subsisting principally on Frogs, worms, insects, and crustaceans. It is by no means averse to preying upon its own kind, and if several of these Fishes are confined in the same aquarium they are apt to give free vent to their cannibal instincts by biting off the tails or limbs of their fellows. The missing parts are soon regenerated, but the new members are usually somewhat abnormal, the tail, for instance, never regaining its original length, while a new pectoral limb may be bifid or even trifid.[604] The tail is the principal organ of locomotion, and by its means the Fish is capable of remarkably quick, agile movements. When slowly moving over the bottom of an aquarium the paired limbs are observed to move to and fro on opposite sides alternately in a somewhat bipedal fashion. The limbs are useless for swimming, although it is possible that they may be helpful in creeping over the bottom, or in balancing, or as tactile organs. Protopterus is said to breathe by its lungs as well as by its gills, and to rise to the surface at short intervals to take in fresh air.

{513}[Illustration: FIG. 307.—Diagram of a torpid Protopterus, in situ. c, Cocoon; e, earth; f, funnel leading to the mouth of the Fish; l, lid; m, mouth; m.b, mouth of the burrow; t, tail. (From Newton Parker.)]

In the dry seasons the marshes in which Protopterus lives become dried up, and to meet this adverse change in its surroundings the Fish hibernates, or passes into a summer sleep, until the next rainy season brings about conditions more favourable to active life. Preparatory to this summer sleep, and before the ground becomes too hard, the Fish makes its way into the mud to a depth of about 18 inches, and there coils itself up in a flask-like enlargement (Fig. 307) at the bottom of the burrow, which is lined by a capsule of hardened mucus secreted by the glands of the skin.[605] The mouth of the flask is closed by the capsular wall or lid, which is perforated by a small aperture. The margins of this aperture are pushed inwards, so as to form a tubular funnel for insertion between the lips of the Fish. While encapsuled in its cocoon the Fish is surrounded by a soft slimy mucus, no doubt for the purpose of keeping the skin moist, and its lungs are the sole breathing organs, the air passing from the open mouth of the burrow through the hole in the lid directly to the mouth of the animal. The nutrition of the dormant Fish is effected by the absorption of the fat stored up about the kidneys and gonads, somewhat after a fashion not unknown in the fat-bodies of Insects and the hibernating glands of Rodents. Even portions of the caudal muscles undergo fatty degeneration, and thus, in a way which recalls the mode of nutrition of the Salmon during the breeding season, and of the Tadpole during its metamorphosis, a further store of nutritive material becomes available for the sustenance of the Fish during {514}its long summer nap. It is highly probable that the exceptionally numerous leucocytes act as carriers in the work of transporting the fatty particles to the different organs and tissues of the body. The length of the summer sleep naturally varies with the duration of the dry season, and probably it lasts on an average nearly half the year (August to December). The cocoons, imbedded in an outward casing of hardened mud, have often been brought to Europe, and when placed in water of suitable temperature the long torpid Protopterus escapes from its prison in a perfectly healthy condition, and resumes its partly branchial and partly pulmonary mode of breathing. The negroes of the West Coast of Africa are very partial to these Fishes, which they dig out of the dried marshes and preserve in their clumps of mud for food. With the advent of the rainy season, when the marshes become flooded, the Protopterus emerges from its cocoon, and returning to its former active life, soon enters upon the task of reproducing its kind. The important observations of Budgett[606] have thrown much light on the curious breeding habits and development of these Fishes. The Fish makes a nest near the edge of a swamp. The nest is simply a hole of irregular shape, about a foot in depth, filled with water and surrounded by long grass (Fig. 308). There is no lining to the nest, and the eggs are deposited on the bare mud. Until the eggs are hatched, which occurs about the eighth day, and while the larvae are in the nest, the male remains on guard, and is apt to bite severely an incautious intruder. Probably with the view of aerating the eggs the water is continually lashed about by the tail of the guardian parent. The male has no trace of the peculiar vascular filaments which adorn the pelvic limbs of the male Lepidosiren during the breeding season. The early developmental stages are similar in their main outlines to those of Neoceratodus, but the young are very different. When the young Protopterus (Fig. 309) is hatched it is provided with a crescentic glandular sucker or cement-organ, situated on the under side of the head behind the mouth, by means of which the larva attaches itself to the sides of the nest, or of the vessel in which it is confined, much in the same way as the young Lepidosteus, and probably for the same reason. It may be remarked that the sucker agrees in structure, position, and function with that found in Amphibian tadpoles, but it differs both in position and structure from its preoral analogue in the young of Acipenser, Amia, and Lepidosteus.

{515}[Illustration: FIG. 308.—Nest of the Protopterus of the Gambia. (From Budgett.)]

A month-old larva has much the aspect of a larval Newt. It has four pairs of vascular plumose cutaneous gills (Fig. 309), which are retained as vestiges for a long time or even throughout life, and two pairs of synchronously-developed limbs. As an interesting instance of a nocturnal and protective change of colour, it may be mentioned that the dark chromatophores of the skin of the larva expand in the day-time and the young Fish becomes darker in colour, and therefore less conspicuous when seen against a background of black mud or soil. At night the contraction of the colour-sacs renders the larva more transparent and probably less easily {516}visible than if opaque. The commencement of pulmonary respiration is coincident with the degeneration of the cutaneous gills, which takes place about seven weeks after the deposition of the eggs, and about a month after the larvae leave the nest. Protopterus is said to attain a length of six feet.

Lepidosiren paradoxa,[607] probably the only species of the genus, is confined to South America. It occurs along the course of the main Amazon river, entering some of its larger affluents, such as the Ucayale, the Madeira, the Rio Negro, and the Tapajóz, and also in the Chaco Boreal to the west of the Upper Paraguay river. The home of the Lepidosiren (or "Lolach," as the natives call the Fish) of the Chaco country is to be found in the wide-spreading marshes and swamps, which for a great part of the year are almost choked by a luxuriant growth of their own peculiar vegetation and covered by a floating carpet of surface weeds, with here and there deeper and clearer water and slow-flowing streams. In the dry season the water gradually shrinks and the swamps eventually become dried up. Of sluggish habits, the Fish wriggles slowly about at the bottom of the swamp like an Eel, using its hind limbs in an irregular bipedal fashion as it wends its way through the dense network of subaqueous plants. Lepidosiren is not exclusively carnivorous. The large fresh-water snail, Ampullaria, which lives in the swamps in enormous numbers, seems to be its favourite food; but masses of confervoid Algae are also eaten, and in its earlier stages it is probable that the Fish is more herbivorous than carnivorous. The Jacare (Caiman sclerops) feeds on Lepidosiren, and this fact, and probably also the cannibal habits of the Fish itself, may explain the capture of specimens with mutilated tails and regenerated, branched, pectoral limbs. Like other living Dipneusti, Lepidosiren rises to the surface to breathe. The intervals are, however, very variable, and no doubt depend on the relative purity or impurity of the water. Both expiration and inspiration are said to take place through the mouth. The snout is protruded on the surface, and the creature expires. After being withdrawn for a moment the head is again projected, and inspiration takes place through the partially open lips.

{517}[Illustration: FIG. 310.—Pelvic limb of the male Lepidosiren during the breeding season. (From Graham Kerr.)]

When the Fish finally sinks a few bubbles of surplus air escape through the gill-clefts. A nocturnal and protective change of colour, similar to that described in Protopterus, has been observed, and although most strikingly manifest in the larvae, it also occurs in individuals of older growth. The flesh is much esteemed as food by the Indians, who wade into the swamps and transfix the Fishes with spears. During the rainy season the Lepidosiren eats voraciously, and a reserve of fat is stored up in the tissues. Like its African relative, the Fish ceases to feed on the approach of the dry season, and eventually hibernates at the dilated extremity of a deep tubular burrow, the entrance to which is plugged by a small lump of clay perforated by several round holes. On the rising of the water at the next rainy season the Lepidosiren pushes out the plug and soon emerges from its burrow.[608] The breeding season begins soon after the escape of the Fish. The eggs are deposited in nests in the form of underground burrows excavated in the black peaty soil at the bottom of the swamp, with an entrance about 4-5 inches wide. At a depth of about a foot the burrow takes a horizontal course, its total length varying from 2-5 feet. After the eggs are laid the male remains to guard them. During the breeding season the pelvic limbs of the male enlarge and become covered by a rich growth of highly vascular, blood-red filaments 2-3 inches in length[609] (Fig. 310). The use of these curious structures is uncertain, but it is not improbable that they act as accessory gills to enable the male to guard the eggs in the nest without being forced to resort to the surface to breathe air. The development is essentially similar to that of Protopterus. The larva (Fig. 311) has four pairs of cutaneous gills in relation with the first, second, third, and fourth branchial arches, inclusive, the first three pairs being the homologues of the cutaneous gills of the tailed Amphibia; and also a cement-organ {518}which disappears shortly before the larval metamorphosis. At that period the circulation in the cutaneous gills becomes sluggish, and very soon these organs completely atrophy. About the same time the hyo-branchial cleft closes up, as in Protopterus. The young Lepidosiren soon begins to breathe air and to become more active and lively in its habits.[610] The adult may attain a length of four feet.

The relations of the different genera of Dipneusti to one another has been discussed by Dollo in a remarkably suggestive paper.[611] Until the publication of this treatise it was generally believed that the modern Dipneusti, Neoceratodus, Protopterus, and Lepidosiren, especially the first mentioned, were the most primitive and the more nearly related to the ancestral stock, while the older types, such as Dipterus, were regarded in the light of highly specialised offshoots. The continuity of the median fins, the apparently diphycercal character of the tail, and the wholly cartilaginous condition of the chondrocranium in the modern Dipneusti, were contrasted with the divided median fins, the heterocercal tail, and the more extensively ossified chondrocranium of the Palaeozoic forms, and the belief seemed inevitable. Dollo has shown, however, that there is good reason for the view that the evolution of the group has taken place in exactly the opposite direction; that, in fact, the older Dipneusti are the more archaic, and that their modern representatives have been derived from them by a sequence of retrogressive changes; or, in other words, the latter have much the same relation to the former as the degenerate Sturgeons and Paddle-Fishes to their Palaeozoic ancestors, the Palaeoniscidae. Taking Dipterus, the {519}most ancient of all the known Dipneusti, as a starting-point, it is possible to select a series of genera which illustrate the evolution of the group both in structure and in palaeontological sequence.[612] The series is as follows:—Dipterus, Scaumenacia, Phaneropleuron, Uronemus, Ceratodus (Neoceratodus), Protopterus and Lepidosiren. Briefly, the more important structural modifications observable in the transition from the older to the recent genera are (a) the gradual union of isolated median fins to form a continuous fin[613]; (b) the substitution of a gephyrocercal tail for a heterocercal[613]; (c) the degeneration of the squamation, the thick ganoid scales of the earlier types being replaced by thin, non-ganoid scales; (d) a reduction in the number of cranial dermal bones and the loss of their original ganoid investment; (e) the suppression of the jugular plates; and (f) a reduction in the size of the opercular bones. In the last two genera of the series, in which specialisation in some respects and degeneration in others have reached their maximum, the body no longer retains the fusiform and more typically Fish-like shape of the older genera, but, in accordance with Eel-like habits and mode of progression, has become more or less Eel-like in form.[614] The paired fins are almost vestigial, while the scales, so deeply insunken in the skin as to be externally invisible, suggest that the modern Dipneusti are approximating to a final scaleless as well as to an ultimately limbless condition. As to the origin of the Dipneusti as a group, it seems reasonable to look for their ancestors in the early Devonian Crossopterygii with acutely lobate fins, or, with greater probability, to some still more primitive Crossopterygian with simple, non-rhizodont teeth, capable by fusion of giving rise to massive tritoral plates, and involving as a consequence the substitution of an autostylic for an originally hyostylic skull, and the suppression of the secondary upper jaw. In fact, when our knowledge of the development of the surviving Dipneusti and Crossopterygii is more complete, it is not improbable that the inclusion of the two series of Fishes in subordinate divisions of the Teleostomi will prove to be amply justified. The relations of the Dipneusti to the Amphibia are {520}somewhat deceptive, and it seems improbable that the former group stands in the direct line of Amphibian descent. In most of their structural features not directly or remotely associated with air-breathing the Dipneusti are true Fishes, and the striking resemblances which they present to the Amphibians in the vascular system and lungs seem to be rather the outcome of physiological convergence, associated with adaptive and parallel modifications in structure, and due to the influence of a similar environment, than indicative of direct ancestral relations. With more reason it may be inferred that both the Dipneusti and the Amphibia have been derived from some primitive Crossopterygian ancestor with Elasmobranch tendencies, and subsequently became modified in certain respects on parallel lines.

{521}CHAPTER XX

APPENDIX TO THE FISHES: PALAEOSPONDYLIDAE—OSTRACODERMI—HETEROSTRACI— OSTEOSTRACI—ANASPIDA—ANTIARCHI—ARTHRODIRA.

In this chapter it is proposed to treat of certain fossil "Fishes" which, from our ignorance of much that is essential to a proper estimate of their true relationships, cannot at present be referred to any of the recognised primary groups of Fishes.

I. PALAEOSPONDYLIDAE.

The interesting little fossil, Palaeospondylus gunni,[615] discovered in the Lower Old Red Sandstone of Caithness, and first described by Traquair, represents the calcified endoskeleton of an elongated fish-like organism about an inch, or not exceeding two inches, in length. The vertebral column consists of a series of broad, calcified ring-like centra, destitute of ribs, but possessing neural arches and spines, and in the caudal region haemal arches and spines in addition. The skull, of which only the ventral surface is known, has a complete basis cranii, laterally expanded behind by periotic capsules, and in front by what seem to be bulging olfactory capsules. Anteriorly, the skull terminates in a ring of calcified cirri. Behind the skull there are two singular post-occipital plates, one on each side of the anterior section of the vertebral column. The tail was apparently furnished with a fringing caudal fin, supported dorsally by the long forked, neural spines, and below by the much shorter haemal spines. There is no trace {522}of limbs, limb-girdles, jaws, or branchial arches, nor any evidence of the existence of scales, denticles, or other exoskeletal structures.

Palaeospondylus has been regarded as a Cyclostome, a view which derives its principal support from the resemblance of the cirri-encircled orifice at the anterior end of the skull to an unpaired nasal or naso-pituitary aperture, and perhaps some measure of credence from such purely negative evidence as the apparent want of limbs, and of any structures comparable to jaws. But even if it be admitted that there is some reason for this view, it is certain that Palaeospondylus obtained a far higher grade of specialisation in certain respects than any of the existing Cyclostomata; the presence of calcified vertebral centra and neural arches is conclusive on this point.[616] Palaeospondylus has also been compared with a larval Arthrodiran and with a larval Dipnoid.[617]

II. OSTRACODERMI.

The Palaeozoic fish-like forms, which, more as a matter of convenience than as an expression of real kinship, it has been customary to include in this group, are amongst the earliest Craniates of which we have any precise knowledge. {523}Of the three subordinate groups or "Orders" into which they have usually been divided hitherto, two, the Heterostraci and the Osteostraci, may, with some show of reason, be considered as related forms, and although they are characterised by much specialisation on independent lines, there is yet some evidence of connecting links between the two. The organisms comprising the third group, the Antiarchi, stand upon a very different footing, and at present it cannot be said that they are in any way related to either the Heterostraci or the Osteostraci, or indeed to any other Craniates whatsoever. The association of the Ostracodermi with the Cyclostomata, a view which has received more influential support than it deserves, is based on the presumed absence of jaws and paired fins. The absence of jaws, which, if present, were almost certainly cartilaginous, has yet to be proved, and even in the latter group it is by no means certain that they do not possess structures which, morphologically if not functionally, are veritable jaws. Nor is it quite certain that the lateral lobes of some Ostracodermi are neither pectoral flaps nor lateral fin-folds, to say nothing of the lateral appendages of the Antiarchi. And to these objections there is the further difficulty that there is absolutely no evidence that the Ostracodermi are monorhinal in the sense in which this term is applied to the Cyclostomata.[618] On these grounds it would seem more in accordance with our present knowledge to regard the Ostracodermi as an independent group whose exact position in the system has yet to be determined, including, however, besides the generally accepted orders Heterostraci and Osteostraci, the recently founded provisional order Anaspida, but excluding the Antiarchi as a separate and distinct section; rather than to crystallise in a definite system of classification views which are either purely conjectural or wholly unjustifiable. Even with this limitation the Ostracodermi are by no means easy to define, especially if we include those remarkable shark-like forms from the Upper Silurian rocks of the south of Scotland which have been so admirably described in the recent classical memoirs of Dr. Traquair. As a rule, the head and the anterior part of the body are laterally expanded, and more or less sharply defined from the rest of the body by prominent postero-lateral angles. The exoskeleton, which exhibits an extraordinary variety of {524}structure in the different families, ranges from a uniform covering of dermal denticles to a condition in which the denticles fuse to form anteriorly a highly characteristic tessellated or continuous dorsal shield, while posteriorly they become replaced by a nearly typical rhombic squamation. The tail is heterocercal. Paired fins of the ordinary piscine type are absent. In some Ostracodermi it seems probable that the gill-clefts opened into a common branchial chamber on each side, with a single external aperture, but in others they may have been ventral. The endoskeleton, jaws, dentition, and the nostrils are unknown.

ORDER I. HETEROSTRACI.

The exoskeletal structures consist of dentine, or of a tissue resembling it, never of true bone. The orbits are marginal or lateral in position. With the exception of a caudal fin there are no median fins.

FAM. 1. COELOLEPIDAE.[619]—Head and anterior portion of the body flattened and expanded, with prominent lappet-like postero-lateral lobes, which may represent continuous lateral fin-folds or a very primitive type of pectoral fin. Nothing is known of the mouth, but it must have been ventral, nor of the position of the orbits. Branchial apertures unknown, but transverse markings on each side of the anterior part of Thelodus pagei may be indications of a branchial apparatus. The exoskeleton consists of a uniform covering of hollow pointed spines, devoid of a basal plate and open below (Lanarkia); or of minute shagreen-like tubercles (Thelodus). The tubercles or spines consist of dentine coated by ganoin. Of the only two known genera, Thelodus is a characteristic Upper Silurian genus {525}from the Ludlow and Downtonian Beds of Lanarkshire. Detached scales are also known in the Upper Silurian of England. One species (Th. pagei) occurs in the Lower Old Red Sandstone of Forfarshire, and another (Th. tulensis) in the Upper Devonian of Russia. Lanarkia has only been found in the Downtonian Beds. None of the Coelolepidae exceed fourteen to fifteen inches in length.

FAM. 2. DREPANASPIDAE.—This family[620] affords an interesting transition to the more highly specialised and carapaced Pteraspidae. The head and anterior part of the trunk now form a broad oblong shield, rounded in front and abruptly marked off from the tail by conspicuous rounded angles. The exoskeleton is no longer uniform. In the caudal region the scattered spines or shagreen tubercles of the Coelolepids have become transformed into tuberculated quadrangular scales, which are further differentiated along the dorsal and ventral margins into ridge scales or fulcra; and from a similar source by a process of basal fusion a series of larger or smaller dermal plates are formed as components of large dorsal or ventral shields. The dorsal shield (Fig. 314) is formed by a large central plate; the postero-lateral portions by two narrow falciform plates; and the anterior margin by a series of smaller rostral plates. Between the larger plates the shield is completed by numerous {526}small polygonal plates. All the plates are superficially ornamented by small stellate tubercles. The ventral armature (Fig. 315) is similar to the dorsal. A large mental plate forms the hinder margin of the transverse slit-like mouth, the anterior limit of which is defined by the rostral plates already mentioned. Laterally may be seen a pair of small plates (x), each perforated by a small aperture, and probably indicating the position of some kind of sense-organ. Posteriorly there is a large median ventral plate, in relation with a pair of anterior and a pair of posterior ventral plates. The areas between the larger plates are filled in by numerous small polygonal plates. It is possible that there is a single external branchial aperture on each side, near the postero-lateral angle of the shield and behind the posterior ventro-lateral plate. The sole representative of the family is Drepanaspis gemündenensis, from the Lower Devonian of Gemünden in Rhenish Prussia. Large examples of this fossil must have exceeded two feet in length.

FAM. 3. PSAMMOSTEIDAE.—To this family are referred certain dermal plates occurring, in a more or less fragmentary condition, in the Old Red Sandstone and Devonian formations of Great Britain and Russia. In their size and shape, and in their stellate tubercles, these have been compared to the dorsal, postero-lateral, and ventral plates of {527}Drepanaspis. That Psammosteus is closely allied to Drepanaspis seems certain, but for the present the two genera may be retained in separate families.

FAM. 4. PTERASPIDAE.[621]—Until the recent inclusion of the three preceding families, the Pteraspidae were the only representatives of the Heterostraci. In the best known genus, Pteraspis, there is a marked reduction in the number of the component plates of the carapace, and only seven can now be distinguished (Fig. 316): (a) a large posterior dorsal plate, supporting behind a stout spine; (b) a conical rostral plate, covering the preorbital part of the head; (c) a pair of small marginal orbital plates, each with a small aperture, probably for the eye; (d) a pair of posterior lateral or cornual plates, each of which is perforated by a large oblique foramen, conjecturally an external branchial aperture; and (e) a large ventral plate. There is probably, also, a small median "parietal," or "pineal," plate, with a pit on its inner surface, situated between the rostral and posterior dorsal plates. Externally the plates are sculptured into fine ridges, which in their minute structure and their crenated free margins are suggestive of linear series of fused denticles. The tail appears to have been invested by imbricated rhombic scales. Pteraspis (Lower Old Red Sandstone of Scotland and England, and the Lower Devonian of Galicia); Cyathaspis (Upper Silurian and Lower Old Red Sandstone), known only by its dorsal and ventral shields; and Holaspis (Lower Old Red Sandstone of Monmouthshire, and the Upper Silurian of Pennsylvania), are the only genera.

ORDER II. OSTEOSTRACI.

While agreeing with the more specialised Heterostraci in the division of the body into an anterior carapaced portion and a free {528}hinder part invested by a rhombic squamation, the Osteostraci are distinguished by the presence of bone as a histological component of the dermal hard parts; by the position of the orbits, which, instead of being marginal in position, are close together on the dorsal aspect of the carapace; and by the possession of a median dorsal fin.

FAM. 1. ATELEASPIDAE.[622]—The general shape of the body is much the same as in the Coelolepidae, but the exoskeleton consists of numerous polygonal tuberculated plates in front of the postero-lateral lobes, and of sculptured rhombic scales behind. A pair of crescentic markings, placed close together about the middle of the dorsal surface of the head, probably indicate the outer margins of orbital recesses (Fig. 317). The only species at present known (Ateleaspis tessellata) occurs in the Downtonian beds.

FAM. 2. CEPHALASPIDAE.[623]—In this family the dorsal shield is rounded in front, strongly arched above, with its postero-lateral angles produced into highly characteristic cornua (Fig. 318). The shield consists of a single piece, but as the outer surface is ornamented by small tubercles arranged in polygonal areas, it is probable that it has been formed by the basal fusion of {529}numerous primitively distinct polygonal plates (Fig. 319, A). Between the orbits there is a separately calcified but fixed plate, which bears a hollow prominence, probably for the reception of a parietal organ. In some genera certain of the anterior dorsal and ventral scales of the trunk fuse into a continuous plate. Internally to the postero-lateral cornua the middle layer of the shield is prolonged backwards into a pair of singular flap-like lobes, which have been variously interpreted as corresponding to the lateral lobes of the Coelolepidae, to pectoral fins, or to opercula. The scales of the trunk and tail are rhombic and imbricated; on the sides of the body they are remarkably high and narrow.

The best known genus is Cephalaspis. The earliest remains are found in the Ludlow Tilestones. The genus is also represented in the Ledbury Passage Beds, the Lower Old Red Sandstone of Scotland, and the Upper and Lower Devonian of Canada. Most of the species are of small size, but C. magnifica,[624] from the Caithness Flagstones, the largest of all the Cephalaspids, has a shield 8½ inches long, and 12 inches across the widest part. Auchenaspis occurs in the Ludlow Tilestones and the Ledbury Passage Beds, and also in the Upper Silurian of the Isle of Oesel in the {530}Baltic. Another genus, Didymaspis, has been found in the Lower Old Red Sandstone of Ledbury.

FAM. 3. TREMATASPIDAE.—The interorbital plate is free, and hence it is often lost in the fossils. Several species of Tremataspis occur in the Upper Silurian of the Isle of Oesel.

As regards the origin and mutual relationships of the different families comprising the Heterostraci, it has been urged with great force by Dr. Traquair[625] that they constitute a natural sequence of forms, beginning with organisms whose Elasmobranch ancestry is extremely probable, and leading to highly-specialised types, which, considered by themselves, possess little to justify any conclusions whatever as to their origin or kinship. The Coelolepidae form the starting-point, and in the light of their exoskeleton of dermal denticles, their derivation from some primitive Elasmobranch prototype seems a reasonable inference.[626] From the Coelolepids the path of specialisation through the Drepanaspidae and Psammosteidae to the Pteraspidae is marked (i.) by the basal concrescence of isolated denticles to form, first, numerous small polygonal plates, and then larger and less numerous plates, as the constituent elements of a characteristic dorsal shield, leaving, however, the denticles of the rest of the body to become converted into a rhombic squamation; (ii.) by modifications in the "lateral fin-lobes," which may become enclosed in the developing dermal armour (e.g. Drepanaspis), or cease to be recognisable (e.g. Pteraspis). The affinities of the Osteostraci are very obscure, and their inclusion with the Heterostraci in the same group (Ostracodermi) has hitherto rested mainly on such negative evidence as the supposed absence of paired limbs, jaws, and teeth; in fact, it has been affirmed that "there is absolutely no reason for regarding Cephalaspis as allied to Pteraspis beyond that the two genera occur in the same rocks."[627] It is possible, however, that in Ateleaspis we have an annectent form, which in some measure combines the structural peculiarities of the two groups. That this singular genus belongs to the Osteostraci is proved by the presence of bone lacunae in its dermal hard parts, a conclusion which is strengthened by the apparently dorsal position of the orbits and the presence of a dorsal fin. On the {531}other hand, its close resemblance to the Coelolepids in the general contour of its laterally-lobed body, and the probability that its mosaic and tuberculated head-shield has been formed by the concrescence of Coelolepid denticles, is at least significant of a relationship to the more primitive Heterostraci. Little can be conjectured as to the habits of these ancient "Fishes." The form and regional proportions of the body, which in some respects often remind one of organisms so diverse as a King Crab, or a Loricaroid Teleost (such as Liposarcus), are strongly suggestive of a grovelling, bottom-feeding, sluggish habit of life, in sharp contrast to the more active and predaceous Fishes whose appearance is coincident with the extinction of the Ostracodermi at the close of the Devonian period. Habits such as these may well be associated with much structural degeneration, even, it may be, with the loss of paired fins, and hence it is not altogether improbable that the Ostracodermi are outcasts from the Elasmobranchs, a degenerate race which has sought safety in a sequestered life and a coat of mail.

ORDER III. ANASPIDA.

This group has been instituted by Traquair[628] for the provisional reception of two remarkable genera, which, owing to the absence of precise knowledge of the histology of their exoskeletal structures, cannot at present be referred either to the Heterostraci or the Osteostraci, and for which, as their discoverer remarks, no place can be found in the system unless they are admitted to the Ostracodermi.

FAM. 1. BIRKENIIDAE.—Body fusiform and fish-like. Head bluntly rounded, without a cranial shield. Caudal fin bilobate and heterocercal A median row of scales with recurved spines arranged along the ventral surface. Orbits, jaws, teeth, paired fins, and endoskeleton unknown.

In Birkenia (Fig. 320) the body is invested by longitudinal rows of narrow scales arranged in oblique transverse rows, which are replaced on the head by much smaller, peculiarly disposed, spindle-shaped scutes. On the side of the hinder part of the head there is an oblique row of small apertures, possibly branchial. A small remote dorsal fin, invested by the trunk scales, is present. {532}Birkenia elegans, the only species known, does not exceed 3½ inches in length. Less is known about the second genus, Lasanius, of which there are two species. Except for the mid-ventral series of spiny scutes, and a row of slender, parallel, rod-like structures, the body appears to have been naked (Fig. 321). The two genera belong to the remarkable series of fossil Fishes from the Silurian rocks of Lanarkshire. Rare in the Ludlow series, Birkenia is by far the most common of the Fishes of the over-lying Downtonian Beds. Lasanius is confined to the latter horizon. Euphanerops, from the Upper Devonian of Canada, is probably related to this family, but lateral branchial apertures are not known.[629]

III. ANTIARCHI.

The organisms comprising this group[630] resemble the Ostracodermi in possessing a well-developed carapace of bony plates and a heterocercal tail, as well as in many of the purely negative features which are characteristic of the latter group.

{533}[Illustration: FIG. 322.—Restored outline of Pterichthys milleri. The upper figure represents a dorsal view, and the lower a lateral view. The dotted lines indicate the course of the lateral line system. a.d.l, Antero-dorso-lateral; ag, angular; a.m.d, anterior median dorsal; a.v.l, anterior ventro-lateral; e.l, extra-lateral or operculum; l, lateral; l.occ, lateral occipital; m, median or interorbital plate; m.occ, median occipital; o, orbit; p.d.l, posterior dorso-lateral; p.m, pre-median; p.m.d, posterior median dorsal; pt.m, post-median; p.v.l, posterior ventro-lateral. ——, Plates investing the limbs: c, central; d.a, dorsal anconeal; d.ar, dorsal articular; e.m, external marginal; i.m, internal marginal; m.m, marginals; t, terminal. (From Traquair.)]

The remarkable dorsal shield is divided into a small cephalic portion and a much larger hinder part investing the greater part of the trunk, both of which are strongly arched above and flattened ventrally, with a movable articulation between the two. The cephalic shield is formed by numerous symmetrically-disposed tuberculated plates, suturally connected with one another, and, like the other exoskeletal structures, containing bone lacunae (Fig. 322).[631] The orbits are close together, near the middle of the dorsal surface, and between them there is a small median interorbital plate, with a deep pit on its inner surface, possibly for a parietal organ. A small lateral plate (e.l.), evidently free behind, suggests the presence of an operculum. Nothing is certainly known about the jaws or the nostrils. The mouth is situated just behind the anterior margin of the cephalic shield on the ventral surface, and in front of it there are two plates, {534}which in Bothriolepis canadensis have their oral margins fringed by small "denticles"; it is possible that these plates represent the components of a secondary upper jaw. The dorsal armature of the trunk is shown in Fig. 322. Ventrally it is completed by a pair of anterior ventro-lateral plates and a pair of posterior ventro-lateral plates with a small median plate between the two pairs. Articulating with the anterior ventro-lateral plates by means of a complex hinge joint there is a pair of pectoral appendages of a kind entirely without parallel in any other vertebrated animals. Each appendage is completely encased by numerous suturally connected plates, and about the middle of its length there is a second movable joint. The appendages are hollow, and their cavities probably contained the muscles by which the limbs were moved, and the blood-vessels and nerves for their nutrition and innervation. A lateral line system of the normal type is present in Pterichthys, consisting of a lateral groove along the side of the trunk, and of supra-orbital and infra-orbital grooves, and post-temporal and infra-orbital commissures, on the head. The free portion of the body and the tail are invested by imbricated and finely tuberculated scales, which form fulcra in front of and behind the small dorsal fin. There are no pelvic fins. The caudal fin is heterocercal.

FAM. 1. ASTEROLEPIDAE.—The best known genera are Pterichthys from the Lower Old Red Sandstone of Scotland and the Devonian of Eifel, and Bothriolepis, a more widely distributed genus which occurs in the Upper Old Red of Scotland and Shropshire, and in the Upper Devonian of Russia and Canada. Two other genera, Asterolepis and Microbrachius, are also found in the Old Red Sandstone of Scotland.[632]

Beyond an uncertain and shadowy relationship to the Ostracodermi, and perhaps some points of resemblance to the Arthrodira, the Antiarchi stand alone among Craniates. Nothing is known of their origin; no intermediate forms link them to any other groups, and the high specialisation they have attained is sufficient to negative any idea that they can "be credited with any share in the evolution of the Fishes of more recent periods."

{535}IV. ARTHRODIRA

This group has been instituted for the reception of a number of remarkable armoured Fishes of uncertain relationships which flourished in Europe during the Devonian and Old Red Sandstone periods, and in North America from the Devonian to the Lower Carboniferous. The head (e.g. in Coccosteus)[633] is invested dorsally by a series of median and lateral symmetrically-disposed tuberculated plates (Fig. 323). Two of the lateral plates are notched for the orbits, and between them there is an interorbital plate which either has a pit on its inner surface or is perforated by an open funnel, as in Dinichthys, possibly for a parietal or a pineal organ. Some of the bones present some analogy, to say the least, to certain of the dermal bones of a typical Teleostome, apparently representing such elements as paired parietals and frontals, a dermal mesethmoid, and toothless premaxillae and maxillae (Fig. 324, A). As in the Antiarchi, the anterior portion of the trunk is also armoured, above by a dorsal shield, formed by median and lateral plates, and below by a similarly constructed ventral shield (Fig. 324, B). A huge joint connects the head and trunk shields: hence the term Arthrodira or "joint-neck." The rest of the body is naked.

Pectoral fins are unknown, but pelvic fins, each supported by a stout basal plate or basipterygium, and with traces of radials, are present. There is a small dorsal fin. Little is known of the primary cranium, but in the trunk and tail it is evident {536}that there are well-developed and partially calcified neural and haemal arches associated with a persistent notochord. It is possible that the skull is autostylic. Gill-arches are not known. A pair of plates (Fig. 324, A, j) at the postero-lateral angles of the cephalic shield may perhaps be opercula. The teeth are conical. Those in the upper jaw are supported by two pairs of plates, probably vomers and palatines. In the lower jaw there are two series of teeth, one in front near the symphysis, and the other behind, supported by a single bone in each ramus. There is a well-developed lateral line system, indicated by surface markings on the head and trunk shields.

FAM. 1. COCCOSTEIDAE.—Coccosteus occurs in the Devonian of Europe and North America, and includes species of relatively small size, not exceeding half a metre in length. C. decipiens, the best known species, is a characteristic fossil in the Old Red Sandstone of Scotland. Phlyctaenaspis[634] is found in the Lower Devonian of Canada, England, and Poland. A larger Arthrodiran, with slender toothless jaws, Homosteus,[635] is met with in the Lower Old Red Sandstone of the North of Scotland, and in the Devonian of Germany and Russia. The Old World Arthrodira must yield, however, to those of the New World for variety in size and shape, and in the character of their dentition.[636] {537}Some of the North American genera (e.g. Dinichthys) probably attained a length of ten feet, or even, as in Titanichthys, a much greater size. Some are fusiform in shape, but Mylostoma is flattened and Ray-like, and, judging from the dentition, their food and habits must have been equally varied. Mylostoma has tritoral plates not unlike those of Neoceratodus or Chimaera. In others the teeth are single, and conical or pointed; in Titanichthys the front teeth in both jaws are beak-shaped structures. It is highly probable that Titanichthys, Mylostoma, and perhaps other genera, are types of distinct families.

The Arthrodira have been regarded as armoured Dipneusti, a view which is mainly based on their supposed autostylism and the nature of the dentition. But this autostylism has yet to be verified, and, if proved, the possibility that it may be a secondary feature, associated with the evolution of a peculiar dentition, must not be forgotten. Much more may be said for their claim to be regarded as a highly specialised race of primitive Teleostomi. Besides a well-developed lower jaw, bones comparable to the elements of a secondary upper jaw are known, and in a general way the disposition of the cranial roofing bones, and the arrangement of the endoskeletal elements of the pelvic fins, tend to conform to the normal Teleostome type. In fact, Dr. Traquair has expressed the opinion that the Arthrodira are Teleostomi and Actinopterygii.[637]

TELEOSTEI (SYSTEMATIC PART)

BY

G. A. BOULENGER, F.R.S., V.P.Z.S.

Of the British Museum (Natural History)

{541}CHAPTER XXI

TELEOSTEI: GENERAL CHARACTERS—MALACOPTERYGII—OSTARIOPHYSI

ORDER IV. TELEOSTEI.[638]

As stated above (p. 495), the Holostean Ganoids pass very gradually into the Teleosteans, the lower groups of which appear to have been directly derived from them. The precise definition of the Order Teleostei, as compared with the Ganoid Order Holostei, is a matter of some difficulty. The most important character appears to be the presence of an ossified supraoccipital bone.[639] Remnants of primitive characters, such as ganoid scales, fulcra, rudiments of a splenial bone, spiral valve to the intestine, multivalvular conus arteriosus, are still found in some lower Teleosteans, but no longer in that combination which serves to define the preceding order. Although Albula is exceptional among all Teleosteans in having two transverse series of valves to the bulbus arteriosus instead of one, no Ganoid has fewer than three.

The first remains of Teleosteans appear scantily in the Upper Trias, and it is not before we reach the Upper Cretaceous that they assume preponderance over other Teleostomes; whilst in the Upper Eocene they have already attained a development and variety of types comparable to their present condition. Out of some 12,000 well-established species of Fishes known to exist at {542}the present day, about 11,500 belong to this order. The classification of such an array of forms is, of course, a matter of great difficulty, and gives scope for much difference of opinion among those who have attempted to grapple with the subject. It is now recognised that the study of the skeleton affords the safest guide to a natural arrangement of the families and higher divisions. Much has been done in this line by Cope, Gill, Sagemehl, A. S. Woodward, and Jordan and his pupils; but the osteology of many important types still remains unknown. For some years a large number of skeletons have been prepared in the British Museum with the object of settling open questions, and this material has enabled me to draw up a scheme of classification which, whatever its defects, and however provisional, I feel sure is on the whole an improvement on those hitherto proposed, and especially on that generally in use in this country. The latter was, to a great extent, based on physiological principles; the present aims at being phylogenetic. In its preparation I have derived great benefit from the labours of the authors quoted above, but have endeavoured in every instance to verify their statements on a larger osteological material than appears to have been available to them. I have also had the advantage of the criticism, on many points, of my young colleague, Mr. C. Tate Regan, who has himself endeavoured to settle some important questions of classification.[640]

The Order Teleostei is divided into thirteen sub-orders, the probable relations of which are expressed in the following diagram:—

11. Opisthomi. 13. Plectognathi. 12. Pediculati. –––+ | | | | +––––––––––––––––––+––––––––––––––––––––+ | | | 9. Anacanthini. 10. Acanthopterygii. 8. Percesoces. | | | | | +––––––––––––––––––+––––––––––––––––––––+ |––Teleostei. | | 7. Catosteomi. 5. Haplomi. 6. Heteromi. | | | | | +––––––––––––––––––+––––––––––––––––––––+ | | | | | 4. Apodes. | | 3. Symbranchii. | | | 1. Malacopterygii. 2. Ostariophysi. | | | –––+ +––––––––––––––––––––––+ | Ganoidei Holostei.

{543}In the classification of Günther, which has been generally in use in this country for the last thirty years, the Teleosts were divided into six principal groups, of ordinal rank: I. Acanthopterygii; II. Acanthopterygii Pharyngognathi; III. Anacanthini; IV. Physostomi; V. Lophobranchii; VI. Plectognathi. Group I. corresponds to Sub-Order 6 (part), 7 (part), 8 (part), 10 (part), 11 and 12 of the present work; Group II. to Sub-Order 10 (part); Group III. to Sub-Order 9 and 10 (part); Group IV. to Sub-Order 1, 2, 3, 4, 5, 6 (part), and 8 (part); Group V. to Sub-Order 7 (part); and Group VI. to Sub-Order 13.

SUB-ORDER 1. MALACOPTERYGII.

Air-bladder, if present, communicating with the digestive tract by a duct. Opercle well developed. Pectoral arch suspended from the skull; mesocoracoid arch present.[641] Fins without spines, the ventrals abdominal, if present. Anterior vertebrae distinct, without Weberian ossicles.

This sub-order, which corresponds to the Isospondyli and Scyphophori of Cope and to a part of the Isospondyli of A. S. Woodward, embraces the most generalised of the Teleosts, and is intimately connected with the Ganoids by the fossil forms which are placed at the base of the series of families. The physostomous condition of the air-bladder, the connexion of the pectoral arch with the skull, the presence of the mesocoracoid arch, the backward position of the many-rayed ventral fins, the normal condition of the anterior vertebrae, the absence of true spines to the fins, and the separation of the supraoccipital bone from the frontals by the parietals, are primitive characters which among the Teleosts occur combined in some families of this suborder only. The mesocoracoid arch is retained by the Ostariophysi, which differ in the remarkably modified condition of the anterior vertebrae, but it disappears in all other Teleosts, which gradually acquire a more forward position of the ventral fins and a reduction in the number of their rays, develop spines in the vertical fins, and lose the communication of the air-bladder with the outside.

The Malacopterygii may be divided into twenty-one families, the characters of which are contrasted in the following synopsis:—

I. Fins fringed with fulcra, or scales coated with ganoin; {544} notochord usually continuous through the vertebrae.

Vertebral centra not more than rings; fins with fulcra; scales rhombic, united by peg-and-socket joints 1. Pholidophoridae.†

Vertebral centra not more than rings; fins with fulcra; scales cycloid 2. Archaeomaenidae.†

Vertebral centra complete or with minute perforation; fins with fulcra; scales cycloid 3. Oligopleuridae.†

Vertebral centra nearly complete, but with perforation; no fulcra; scales cycloid 4. Leptolepididae.†

II. Fins without fulcra; scales without ganoin; vertebral centra complete.

A. Supraoccipital separated from the frontals by the parietals or upper surface of skull; no adipose fin.

1. Ventral fins with 10 to 16 rays.

An intergular bone; parasphenoid narrow 5. Elopidae.

No intergular bone; parasphenoid very broad 6. Albulidae.

2. Ventrals with not more than 7 rays.

a. Supratemporal very large, plate-like.

Praemaxillary single, its posterior extremity free from the maxillary; symplectic absent; basis cranii simple 7. Mormyridae.

Praemaxillary paired, its posterior extremity firmly attached to the maxillary; symplectic present; basis cranii double 8. Hyodontidae.

b. Supratemporal small; maxillary firmly attached to posterior extremity of praemaxillary.

Praemaxillary paired; a large hole on each side of the skull, between the post-frontal and the squamosal; basis cranii double; suboperculum absent; ribs sessile 9. Notopteridae.

Praemaxillary paired; basis cranii simple; suboperculum reduced; ribs inserted on parapophyses 10. Osteoglossidae.

Praemaxillary single; basis cranii simple; suboperculum and interoperculum absent; ribs inserted on parapophyses 11. Pantodontidae.

c. Supratemporal small; maxillary movable; ribs sessile; ventral fins below the pectorals 12. Ctenothrissidae.†

B. Supraoccipital separating parietals, or adipose fin present.

1. Interoperculum enormous; symplectic absent; basis cranii simple 13. Phractolaemidae.

2. Interoperculum normal; symplectic present; basis cranii double.

a. Teeth in sockets; maxillary firmly attached to praemaxillary. Symplectic exposed 14. Saurodontidae.†

b. Teeth not in sockets.

Symplectic concealed between the quadrate and the hyomandibular 15. Chirocentridae.

Postclavicle on outer side of clavicle; no adipose fin 16. Clupeidae.

Postclavicle on inner side of clavicle; an adipose dorsal fin 17. Salmonidae.

Postclavicle absent; no adipose fin 18. Alepocephalidae.

3. Interoperculum normal; basis cranii simple. {545}

Maxillary large, toothed; praecaudal vertebrae without well-marked parapophyses; scales cycloid or absent; adipose dorsal fin present or absent 19. Stomiatidae.

Mouth small, toothless; vertebrae with strong parapophyses; head and body covered with spiny scales 20. Gonorhynchidae.

Mouth small, toothless; no symplectic; head and body naked 21. Cromeriidae.

† This sign indicates that the group is represented by fossil forms only.

FAM. 1. PHOLIDOPHORIDAE.—Margin of the upper jaw formed by the praemaxillaries and the maxillaries, the latter large and loosely attached; teeth small and conical. Parietal bones separating the supraoccipital from the frontals; opercular bones well developed. Vertebral centra never advanced beyond the annular stage; ribs delicate; no fused or expanded haemal arches at the base of the tail; no epipleurals or epineurals. Fin-fulcra present, but usually small; dorsal and anal fins small, the former above or behind the ventral fins, which are small or may be absent. Scales ganoid, rhombic, those on the flanks united by peg-and-socket joints.

This family, which appears to me to be related to the Dapediidae, is provisionally placed here by A. S. Woodward on account of its resemblance to the Leptolepididae, but it is not yet quite clear that the mandible was destitute of splenial and coronoid elements, while the bones at the base of the pectoral fin have not hitherto been observed. The principal genera are Pholidophorus, ranging from the Upper Trias to the Purbeck; Thoracopterus, from the Upper Trias; and Pleuropholis, from the Upper Jurassic. The species of Pholidophorus are very numerous in the Jurassic period, and Woodward has observed that the scales of the later species are more elaborately ornamented than those of earlier date.

FAM. 2. ARCHAEOMAENIDAE.—Distinguished from the preceding by the thin, cycloid scales. Conspicuous obtuse ridge-scales are present along the dorsal and ventral lines. Archaeomenes, from the Jurassic (?) of New South Wales.

FAM. 3. OLIGOPLEURIDAE.—Characters as in Pholidophoridae, but scales cycloid and vertebrae completely or nearly completely ossified.

Oligopleurus, from the Jurassic of England and France; Oenoscopus, from the Jurassic and Cretaceous of France, Germany, and Italy; and Spathiurus, from the Cretaceous of Mount Lebanon and Dalmatia.

{546}FAM. 4. LEPTOLEPIDIDAE.—Praemaxillaries very small; maxillaries large, loosely attached; teeth small and conical. Parietal bones separating the supraoccipital from the frontals; opercular bones well developed. Vertebral centra well ossified, but always pierced by the notochord; ribs delicate; epipleurals present; no fused or expanded haemal arches at the base of the caudal fin. Dorsal and anal fins small, the former above or behind the ventrals. Ventrals with 5 to 10 rays. Scales thin, cycloid and deeply imbricate, usually coated with ganoin in their exposed portion.

Leptolepis, with numerous species, from the Jurassic and Cretaceous of Europe and New South Wales; Vidalia, Jurassic of France; Aethalion, Jurassic of Bavaria; Thrissops, Jurassic and Cretaceous of Europe; and Lycoptera, Jurassic of Asia.

FAM. 5. ELOPIDAE.—Margin of the upper jaw formed by the praemaxillaries and the maxillaries, the latter the more developed, and movably articulated above the former to the ethmoid. Parietal bones in contact behind the frontals; opercular bones well developed. Basis cranii double. A bony intergular or sublingual plate. Jaws, palatines, pterygoids, vomer, parasphenoid, glossohyal, and pharyngeals toothed. Ribs mostly sessile, inserted very low down, behind parapophyses; epineurals similar to the ribs, but directed upwards. Pectorals low down, folding like the ventrals. Post-temporal forked, the upper branch attached to the epiotic, the lower to the opisthotic; post-clavicle small; scapular foramen in the scapula; pterygials well developed, three in contact with coracoid. Ventrals with 10 to 16 rays. Branchiostegal rays very numerous (over 20). Air-bladder large.

{547}This family is abundantly represented in Cretaceous times by the genera Osmeroides and Elopopsis, and from the Lower Eocene to the present day by Elops and Megalops. Elops saurus is a handsome elongate silvery Fish, found in all the warm and tropical seas; the young are ribbon-shaped like those of Albula. A second species, E. lacerta, is from the West Coast of Africa, entering rivers. Megalops, distinguished by larger scales, the absence of pseudobranchiae, and the curious prolongation of the last ray of the dorsal fin, includes the well-known Tarpon M. atlanticus, and the Indian M. cyprinoides. The Tarpon occurs from the south-eastern coasts of North America and the West Indies to Brazil, and reaches a length of 6 feet and a weight of 110 lbs. It often leaps out of the water, after the manner of Grey Mullets, and its chase when hooked affords good sport, the landing of so active a giant being attended with great difficulties. Its remarkably large scales, over two inches in diameter, are much prized for fancy work in the Florida curiosity shops.

FAM. 6. ALBULIDAE.—Margin of the upper jaw formed by the praemaxillaries and the maxillaries, the latter movably articulated above the former to the ethmoid. Parietal bones separating the supraoccipital from the frontals; suboperculum large; interoperculum small and entirely or nearly entirely hidden below the praeoperculum. Basis cranii double. Praemaxillaries, mandible, vomer, palatines, pterygoid, parasphenoid, glossohyal, and pharyngeals toothed. Ribs sessile, inserted behind and somewhat below small parapophyses, which are absent or merely indicated on the anterior vertebrae, and gradually increase in size towards the caudal region; these parapophyses, as well as the neural and haemal arches, are autogenous bones; epineurals, no epipleurals. {548}Pectorals low down, folding like the ventrals. Post-temporal trifid, the upper branch attached to the epiotic, the median to the squamosal, the lower to the opisthotic; post-clavicle large (formed of three pieces); scapular foramen between scapula and clavicle; pterygials well developed, two in contact with coracoid. Ventrals with 10 to 14 rays. Branchiostegal rays 6 to 14. Air-bladder large, not communicating with the ear.

Elongate fusiform Fishes, covered with large silvery scales forming regular longitudinal series; head naked; mouth rather small, with thick lips; gill-openings wide. Dorsal fin originating in front of ventrals; anal short; caudal well developed, forked.

The type of this family, Albula or Butirinus, is remarkable among Teleosts in possessing a rudiment of a conus arteriosus to the heart, provided with two rows of valvules.[642] Its single species inhabits all the warm and tropical seas. Prof. Gilbert has ascertained that the young pass through a metamorphosis, analogous to that of the Eels. The "Leptocephalid" described as Esunculus by Kaup is probably a larval Albula.

The deep-sea Japanese Pterothrissus (Bathythrissa) must be placed in this family; its skeleton is very similar to that of Albula. It differs in the elongate dorsal fin, in the presence of small teeth on the maxillary bone, and in the small number of branchiostegal rays (6 instead of 12 to 14).

{549}Albula is represented in the Eocene (London Clay and Bruxellian); and the Cretaceous Istieus and Anogmius are believed to be possibly related to Pterothrissus. Four Cretaceous types (Plethodus, Thryptodus, Syntegmodus, and Ancylostylus) are referred with doubt to the Albulidae.

FAM. 7. MORMYRIDAE.—Margin of the upper jaw formed by the single praemaxillary and the maxillaries, the latter articulated above the former to the ethmoid. Parietal bones separating the supraoccipital from the frontals; a large hole on each side of the skull, between the squamosal, the epiotic, and the opisthotic, covered by a large, thin, bony plate (the supratemporal), which may extend over a part of the parietal; symplectic absent; suboperculum small and hidden under the operculum, or absent; interoperculum well developed. Basis cranii simple. No pharyngeal teeth. Opercular bones hidden under the skin; gill-clefts narrow. Anterior ribs sessile; epineurals, no epipleurals. Pectorals directed upwards. Ventrals with 6 or 7 rays. Air-bladder communicating with the ear.

Curious-looking Fishes, very variable in the form of the head and body and in the extent of the fins. Mouth often very small; teeth in jaws usually few; teeth usually present on the parasphenoid, working against a similar patch on the glossohyal; eye covered over by skin, sometimes very indistinct; scales small, cycloid; branchiostegal rays 4 to 8. The dorsal and anal fins may be nearly equally developed (Genyomyrus, Gnathonemus); or the former (Mormyrus) or the latter (Hyperopisus) are several times the longer. Gymnarchus, Eel-shaped, apodal, and deprived of the caudal fin, lacks the anal fin, the dorsal extending over the whole extent of the body. Some species of Mormyrops show how a form {550}like Gymnarchus may have been evolved out of a more typically-formed Fish. Nothing is more striking than the variation in shape of the snout within one and the same genus, and the names given to some of the species (ovis, caballus, elephas, tamandua, numenius, ibis) are suggestive of resemblances with the heads of various animals.

The Mormyrids are highly remarkable for the enormous development of the brain, the weight of which equals 1/52 to 1/82 of the total, a thing unparalleled among lower Vertebrates; and for the problematic organ which surmounts it; also as being among the few Fishes in which an electric organ has been discovered. The organ, situated on each side of the caudal region, is derived from the muscular system and is of feeble power, as ascertained by Babuchin and by Fritsch; it was long considered as "pseudo-electric." The natural affinities of this family appear to be with the Albulidae, and there is nothing to justify the term "Nilhechte" (Nile-pike) which has been bestowed on them by German {551}authors. Ninety-three species are known from the fresh waters of Africa north of the Tropic of Capricorn, and may be referred to two sub-families and ten genera[643]:—

(i.) MORMYRINAE, with teeth on the parasphenoid and tongue, with ventral, anal, and caudal fins, and a simple air-bladder; vertebrae 37 to 64; peculiar (Gemmingerian) linear bones, without known homologues, along each side of the tail, above and beneath the electric organ; scapular foramen in the scapula, or between the scapula and the coracoid. Mormyrops, Petrocephalus, Isichthys, Marcusenius, Stomatorhinus, Myomyrus, Gnathonemus, Genyomyrus, Mormyrus.

(ii.) GYMNARCHINAE, without teeth on the parasphenoid and tongue, without ventral, anal, or caudal fins, and with a cellular air-bladder; vertebrae about 120; Gemmingerian bones absent; scapular foramen in the coracoid. Gymnarchus.

Fossil Mormyrids are unknown.

Venerated by the ancient Egyptians, the Mormyrs of the Nile are frequently represented on hieroglyphics and mural paintings as well as in bronze models. Very little is known of the habits of these Fishes. Prof. G. Fritsch, of Berlin, during his stay in Egypt for the purpose of experimenting on electric Fishes, observed that they perish very rapidly when removed from the river, and he had the greatest difficulty in keeping some alive in an aquarium for two or three days. The species with comparatively large mouths (Mormyrops, Gymnarchus) feed principally on fishes and crustaceans, the others on tiny animals and vegetable and more or less decomposed matter. Delhez, on the Congo, found that many are attracted to the borders of the river in the neighbourhood of human dwellings, where they feed on the refuse thrown into the water. It is probable that the species with a rostrum use it to procure small prey hidden between stones or buried in the mud, and that the fleshy mental appendage with which many are provided is a tactile organ compensating the imperfection of the vision in the search for food. A small Mormyrid from the Congo (Stomatorhinus microps) has the eyes so much reduced and the skin so feebly pigmented as to convey the impression of a cave Fish. Until quite recently, absolutely nothing was known {552}of the breeding habits and development in this important family. To the late J. S. Budgett we owe some very interesting observations made in the Gambia on Gymnarchus niloticus.[644] The Fish makes a floating nest, emerging on three sides, over which the male keeps a fierce watch; the recently-hatched larvae are remarkable for the enormous size of the yolk-sac, which hangs down, acting as a sort of anchor, and for the presence of long external branchial filaments, as in Selachian embryos. The Fish propels itself through the water entirely by the action of its dorsal fin, forwards and backwards with equal facility; when swimming rapidly backwards, it may be seen to use the end of its tail as a feeler to guide the way. Budgett has also identified, with some doubt, the eggs of Hyperopisus bebe, out of which emerged embryos not unlike those of some tailless Batrachians, which hung suspended to rootlets of grass in swamps by means of threads of viscid mucus secreted from glands on the top of the head.

FAM. 8. HYODONTIDAE.—Margin of the upper jaw formed by the praemaxillaries and the maxillaries, the latter the more developed and firmly united to the end of the former. Parietal bones separating the supraoccipital from the frontals; a large hole on each side of the skull, between the parietal, the squamosal, and the epiotic (paroccipital), closed by a large, thin, bony plate (the supratemporal), which extends over the greater part of the parietal; suboperculum and interoperculum small, the latter partly hidden below the praeoperculum. Basis cranii double. Jaws, palatines, pterygoids, vomer, parasphenoid, and glossohyal toothed; no pharyngeal teeth. Ribs sessile, inserted above and behind well-developed parapophyses; epineurals, no epipleurals. {553}Pectorals low down, folding like the ventrals. Post-temporal forked; the upper branch attached to the epiotic, the lower to the squamosal; no post-clavicle; coracoids forming together a ventral keel; scapular foramen between scapula and clavicle; pterygials well developed, three in contact with coracoid. Ventrals with 7 rays. Branchiostegal rays in moderate number (8 to 10). Air-bladder communicating with the ear. No oviducts, the eggs falling into the abdominal cavity before exclusion.

Elongate, compressed, silvery Fishes, covered with moderate-sized cycloid scales; head naked; mouth large, with strong dentition; gill-openings wide; dorsal fin short, posterior to the ventrals; anal rather elongate; caudal well developed, forked.

A single genus (Hyodon) with three species (Moon-Eyes of the Americans), all reaching the length of about a foot, inhabiting the fresh waters of North America east of the Rocky Mountains.

{554}FAM. 9. NOTOPTERIDAE. The Fishes which form this family may be regarded as an eccentric modification of a type very similar to the preceding, with which they agree in most osteological features as well as in the dentition, in the connexion between the air-bladder and the ear, and in the absence of oviducts. They differ in the absence of the lid-like supratemporal, the hole which it covers in Hyodon being here bordered by the post-frontal and the squamosal (fused with the opisthotic), sometimes also by the epiotic, in the absence of the suboperculum, in the absence or incomplete ossification of the upper branch of the post-temporal (the lower articulating with the opisthotic), and in the presence of accessory bones (named adpleurals) attached to or fused with the distal extremity of the anterior ribs, which they prolong to the mid-ventral line, where they are embraced by dermal ossifications forming a doubly serrated crest.

The bones of the head are cavernous, the mouth is large; the anterior nostril is produced into a tentacle. The body is very strongly compressed, with very short precaudal region, attenuate behind; the ventral fins are much reduced or absent; the dorsal is short or absent, {555}and the anal is much elongate and confluent with the caudal, which may be regarded as aborted. The scapular foramen is entirely in the scapula. The air-bladder is very large, with several divisions, forked in front and behind, and prolonged along each side of the caudal region; its extraordinary condition has been described by Bridge.[645]

These Fishes live in marshes and lakes, fresh-water or brackish, and feed on worms and insects. Nothing is known of their breeding habits and development.

Two genera: Notopterus, with a dorsal fin and 6 to 9 branchiostegal rays—three species from India, Burma, and the Malay region, and one from West Africa; Xenomystus, without dorsal fin and with only 3 branchiostegal rays, the unique species inhabiting the White Nile and West Africa. Remains of Notopterus have been found in the marl slates (Tertiary) of Padang, Sumatra. The largest species, the Indian N. chitala, attains 4 feet in length; its flesh is said to be uncommonly rich and well flavoured, but a strong prejudice exists against it, owing to the Fish being supposed to live on human carcases.

FAM. 10. OSTEOGLOSSIDAE.—Margin of the upper jaw formed by the praemaxillaries and the maxillaries, the latter the more developed and firmly attached to the end of the former. Parietal bones separating the supraoccipital from the frontals; suboperculum much reduced, and entirely or partially concealed under the praeoperculum. Basis cranii simple. Teeth in jaws, and on pterygoid and hyoid bones; no pharyngeal teeth. Head scaleless, the thin skin confluent with the bones; body covered with large bony scales, composed of pieces like mosaic. Ribs inserted on the strong parapophyses; epineurals, no epipleurals. Pectoral fins low down. Post-temporal forked, the upper branch attached to the epiotic, the lower to the squamosal; post-clavicle present; scapular foramen in scapula; pterygials well developed, only one in contact with coracoid. Dorsal and anal fins originating behind the ventrals; latter with 5 or 6 rays. No oviducts, the eggs falling into the abdominal cavity before exclusion (at least in Heterotis, as observed by Budgett).

This family is represented at the present day by five species, referred to four genera; thus characterised:—

{556}[Illustration: FIG. 335.—Principal forms of Osteoglossids. A, Dapedoglossus testis (restoration); B, Scleropages leichardti; C, Osteoglossum bicirrhosum; D, Arapaima gigas; E, Heterotis niloticus. All much reduced.]

{557}Scleropages.—Mouth large; vomer, palatines, pterygoids, and glossohyal toothed; mandibular barbels; branchiostegal rays 15 to 17; body compressed, with trenchant abdomen; coracoids forming a ventral keel; dorsal fin short; ventral fins nearly equally distant from end of snout and caudal fin; vertebrae 29 to 31 + 30; air-bladder not cellular. One species from the northern parts of Australia, and one from Sumatra, Banka, and Borneo.

Osteoglossum.—Mouth large; vomer, palatines, pterygoids, and glossohyal toothed; mandibular barbels; branchiostegal rays 10; body compressed, with trenchant abdomen; coracoids forming a ventral keel; dorsal fin long; ventral fins nearly twice as far from the caudal as from the end of the snout; vertebrae 28 + 59; air-bladder not cellular.—South America (Guianas, Brazil).

Arapaima.—Mouth rather large; vomer, palatines, pterygoids, and glossohyal toothed; branchiostegal rays 16; belly rounded; dorsal fin rather long; ventral fins equidistant from head and caudal fin; vertebrae 36 to 38 + 41 to 42; air-bladder cellular.—South America (Guianas, Brazil).

Heterotis.—Mouth moderate; branchiostegal rays 7; belly rounded; dorsal fin rather long; ventral fins nearer end of snout than caudal fin; vertebrae 27 + 42 to 43; air-bladder cellular; fourth branchial arch with an accessory breathing-organ. Africa (Nile, Senegal, Gambia, Niger).

Dapedoglossus, from the Eocene of Wyoming, appears to be nearest to Scleropages, and Brychaetus, from the Eocene (London Clay) of Sheppey, Kent, to Arapaima, so far as the state of preservation of these fossils enables us to form an opinion.

Dr. Günther has directed attention to the remarkable {558}coincidence of the geographical distribution of this family and the Dipneusti, although, however, the latter are not known to be represented in the Malay Archipelago. "Not only," he adds, "are the corresponding species found within the same region, but also in the same river systems; and although such a connexion may and must be partly due to a similarity of habit, yet the identity of this singular distribution is so striking that it can only be accounted for by assuming that the Osteoglossidae are one of the earliest Teleosteous types which have been contemporaries of and have accompanied the present Dipnoi since or even before the beginning of the Tertiary epoch."

The Queensland species of Scleropages (S. leichardti) is known to the settlers by the name of Barramunda, which has also been applied to Neoceratodus. Arapaima gigas is one of the largest fresh-water Fishes known, exceeding a length of 15 feet and a weight of 400 pounds. Its flesh is much valued. Sir R. Schomburgh has observed that the mother protects the young, who, for some time after their birth, always swim in front of her. A similar observation has been made in the Gambia on Heterotis niloticus by the late J. S. Budgett, who states that the Fish builds enormous nests in swamps, in about two feet of water; the walls of the nest are made of the stems of the grasses removed by the Fish from the centre; the floor is the swamp-bottom, and is made perfectly smooth and bare. The nest appears to be used for at most four or five days; the newly-hatched larvae are provided with long external gill-filaments of a blood-red colour.[646]

FAM. 11. PANTODONTIDAE.—The little West African Fish described by Peters as Pantodon buchholzi is the unique representative of a family closely related to the Osteoglossidae, but distinguished by the very small, single praemaxillary and the absence of suboperculum and interoperculum. The pectoral fins are very large and are remarkable for the fleshy process to which the inner ray is adnate; the ventrals, formed of 7 rays, some of which are simple and prolonged into filaments, are placed more forward than in any other type of this sub-order, the Ctenothrissidae excepted, viz. immediately behind the pectorals. Teeth in the jaws and on the vomer, palatines, pterygoids, parasphenoid, {559}glossohyal, and pharyngeal bones. Mesocoracoid arch slender, strongly curved, and meeting its fellow on the median line; coracoids forming a ventral keel. Vertebrae few (16 + 14).

Observed by M. de Brazza to be a freshwater Flying-Fish.

FAM. 12. CTENOTHRISSIDAE.—A curious type characterised by small praemaxillaries, large maxillaries, with feeble dentition, {560}the parietals in contact on the median line, vertebral centra without transverse processes, a moderately large dorsal with simple anterior rays, and large ventrals advanced far forwards and formed of 8 rays. Its affinities are still obscure, but the condition of the jaws decides its allocation to the suborder Malacopterygii, whilst in the position of the ventrals it is most nearly approached by the Pantodontidae. Small Fishes known only by two genera, of the Cretaceous period (England and Mount Lebanon), one with ctenoid scales (Ctenothrissa), the other with cycloid scales (Aulolepis).

FAM. 13. PHRACTOLAEMIDAE.—Mouth edentulous, projectile, bordered by the very slender praemaxillaries and maxillaries. Supraoccipital in contact with the frontals, widely separating the small parietals; operculum and suboperculum well developed; praeoperculum small; interoperculum enormous, covering the gular region and overlapping its fellow; symplectic absent. Basis cranii single. No pharyngeal teeth. Only 3 slender branchiostegal rays. Ribs stout, sessile, nearly completely encircling the body; slender epineurals; no epipleurals; caudal region very short. Supratemporal small, simple, fixed to the parietal and squamosal; no postclavicle; scapular foramen in the scapula. Pectoral fin inserted low down, folding like the ventrals; latter with 6 rays.

The remarkable little Fish, Phractolaemus ansorgii, discovered by Dr. W. J. Ansorge in the Niger Delta in 1900, and which has since also been found in the Congo, stands quite apart among the Malacopterygians, its nearest allies being apparently the Osteoglossidae. The body is elongate and subcylindrical, covered with large striated scales; the head is small, the skull strongly ossified, covered with thin skin; the mouth small, proboscidiform, {561}capable of being thrust forwards, when at rest folded over and received into a depression on the upper surface of the head; the narial orifice is single, and preceded by a barbel; the gill-openings are narrow, restricted to the sides. The ventral fins are inserted far back, the dorsal and anal are short. The air-bladder is very large, and the intestine extremely long and much convoluted. Vertebrae 26 + 8.

FAM. 14. SAURODONTIDAE.—Margin of the upper jaw formed by the praemaxillaries and the maxillaries, the latter the more developed and firmly united to the former; these bones, as well as the mandible, with teeth implanted in deep sockets; palate toothless. Supraoccipital separating the small parietals; opercular bones well developed; symplectic present, exposed. Basis cranii double. Ribs sessile, very low down on the centra; no parapophyses; neural arches not fused with the centra. Pectorals inserted very low down; postclavicle apparently absent. Caudal fin deeply forked, without fused hypurals.

This family, comprising several Cretaceous genera, may be regarded as ancestral to the Chirocentridae, with or near which it has been placed by Cope and various later authors. The normal position of the symplectic, however, entitles its members to rank as a separate family, and the autogenous neural arch, as well as the distinctness of the bones supporting the caudal fin, are also indicative of a greater generalisation. The restoration of Ichthyodectes as given by Loomis, shows a general form similar to an ordinary Herring, but it does not appear to be reliable.

The members of the Saurodontidae have been referred to two groups: (a) with praedentary (praesymphysial) bone, Saurocephalus, Saurodon; (b) without praedentary, Chirocentrites, Portheus, Ichthyodectes, Spathodactylus, Cladocyclus. These Fishes are from the Chalk of Europe and North America, and some among them attain a very large size, perhaps not less than two metres in length.

FAM. 15. CHIROCENTRIDAE.—Margin of the upper jaw formed by the praemaxillaries and the maxillaries, the latter the more developed and firmly united to the former; these bones, as well as the mandible, with large teeth not implanted in true sockets; minute teeth on the palatines, pterygoids, and hyoid bones, Supraoccipital in contact with the frontals, separating the small parietals; opercular bones well developed; symplectic hidden {562}between the inner surface of the quadrate and a descending process of the hyomandibular. Basis cranii double. Ribs very slender, sessile, very low down on the centra; no parapophyses; epipleurals and epineurals. Pectorals inserted very low down. Post-temporal forked; postclavicle absent; a thin bony lamina, similar to the postclavicle, above the pectoral fin, attached to the scapula; scapular foramen in scapula; coracoids in contact with each other, forming a keel. Ventrals very small, with 7 rays. Brachiostegal rays 8. Air-bladder large, not communicating with the ear, incompletely divided into cells. Mucous membrane of the intestine forming a spiral fold.

The body is very elongate and strongly compressed, covered with thin, deciduous scales; the vertebrae number 75. The dorsal fin is short and opposite to the anal, which is long.

Chirocentrus dorab, the only representative of this family, inhabits the Indian Ocean and the seas of China and Japan.

FAM. 16. CLUPEIDAE.—Margin of the upper jaw formed by the praemaxillaries and the maxillaries. Supraoccipital separating the small parietals; opercular bones well developed. Basis cranii double. Ribs mostly sessile, inserted behind parapophyses; intermuscular bones (epineurals, epipleurals, adpleurals) usually numerous. Post-temporal forked, the upper branch attached to the epiotic, the lower to the opisthotic; post-clavicle applied to outer side of clavicle. Ventrals with 6 to 11 rays. Air-bladder large, communicating with the ear.

Four sub-families:—

(i.) THRISSOPATRINAE.—Mouth large; praemaxillaries very small; maxillaries large, with rather narrow supplemental bone, firmly attached to praemaxillaries; branchiostegals about 30; abdomen compressed to an edge, without serration; no lateral line. Thrissopater, from the Gault of Folkestone.

{563}(ii.) ENGRAULINAE.—Mouth moderate or large; praemaxillaries very small; maxillaries large, with narrow supplemental bones, more or less firmly attached to praemaxillaries; branchiostegals 6 to 19; abdomen rounded or more or less compressed, with or without serration; no lateral line. Recent genera: Dussumieria, Etrumeus, Engraulis, Cetengraulis, Heterothrissa, Coilia. Fossil: Spaniodon, Upper Cretaceous.

(iii.) CLUPEINAE.—Mouth small or moderate; maxillaries freely movable behind the praemaxillaries, usually with large supplemental bones; branchiostegals 5 to 10; abdomen usually serrated; lateral line usually absent. Recent genera: Clupea, Hyperlophus (Diplomystus), Opisthonema, Brevoortia, Pellonula, Clupeichthys, Odaxothrissa, Pellona, Chirocentrodon, Pristigaster, Raconda, Chatoessus. Fossil: Pseudoberyx, Histiothrissa, Scombroclupea, Leptichthys, Upper Cretaceous.

(iv.) CHANINAE.—Mouth small, toothless; maxillaries firmly attached to praemaxillaries; branchiostegals 4, very broad; abdomen rounded or flattened; lateral line distinct. Chanos, recent; Chanoides, Upper Eocene; Prochanos, Cretaceous.

{564}Heralded by the genus Thrissopater,[647] which may be regarded as a connecting type between the Elopidae and the Clupeidae, this family is largely represented in Cretaceous times, more abundantly still in the Eocene and Miocene, where Clupea and Engraulis occur in numerous species; Hyperlophus, distinguished from Clupea by the presence of a dorsal serrated ridge similar to the ventral, occurs in the Upper Cretaceous of Syria, Southern Europe, and South America, in the Eocene of North America and Europe, and is represented at the present day on the West Coast of South America and on the coast and in the rivers of New South Wales. About 200 Clupeids are known to live at the present day, mostly marine species, but a few are confined to fresh-waters; none may be termed deep-sea forms; some, like the Allis Shad (Clupea alosa) and Twait Shad (C. finta), are anadromous, ascending rivers to spawn. The range of the family is almost cosmopolitan. Several species are remarkable for the extreme abundance of individuals, as for example the Herring (Clupea harengus), the Pilchard or Sardine (C. pilchardus), and the Anchovy (Engraulis encrasicholus). The Herring inhabits the northern parts of the Atlantic and the seas north of Asia. As Dr. Günther first showed, the so-called "Whitebait" consists chiefly of the fry of Herrings, which, like those of the Sprat (C. sprattus), have a predilection for brackish water. The Anchovy and the Pilchard, on the other hand, seldom if ever enter estuaries. The eggs of the Herring, contrary to those of most British marine food-fishes, are heavy and adhesive, sticking firmly to stones or fixed objects on the sea bottom, whilst those of the Sprat and Pilchard float on the surface. The larvae are long, slender, and transparent. The Sardine, which affords so valuable a fishery on the West Coast of France, is the immature state of the Pilchard, which grows to a length of 10 to 14 inches. Its movements are not yet well understood, and its scarcity during certain years in the waters where it usually swarms has caused periodical crises in an important industry. Ripe Pilchards are mostly found at a considerable distance from the coasts. The Anchovy is especially abundant in the {565}Mediterranean, but it is also regularly fished in Holland, especially in the Zuydersee, where it breeds, as well as in the Mediterranean; it makes only temporary appearances, and has not been observed to spawn, in the English Channel, although eggs have recently been obtained off the coast of North Lancashire.[648]

The imperfectly known Cretaceous Crossognathidae (Crossognathus and Scyllaemus), referred by some authors to the Percesoces, should probably be placed with or near the Clupeidae.

FAM. 17. SALMONIDAE.—Margin of the upper jaw formed by the praemaxillaries and the maxillaries. Supraoccipital in contact with the frontals, but frequently overlapped by the parietals, which may meet in a sagittal suture; opercular bones all well developed. Basis cranii double. Ribs sessile, parapophyses very short or absent; epineurals, sometimes also epipleurals, present. Post-temporal forked, the upper branch attached to the epiotic, the lower to the opisthotic; postclavicle, as usual, applied to inner side of clavicle. A small adipose dorsal fin. Air-bladder usually present, large. Oviducts rudimentary or absent, the ova falling into the cavity of the abdomen before exclusion.

Marine and fresh-water Fishes, mostly from the temperate and Arctic zones of the northern hemisphere: one genus (Retropinna) on the coasts and in the rivers of New Zealand; a few deep-sea forms (Argentina, Microstoma, Nansenia, Bathylagus) occur in the Arctic Ocean, the North Atlantic Ocean, the Mediterranean, and the Antarctic Ocean, down to 2000 fathoms. Apparently of comparatively recent age, no remains older than Miocene (Osmerus, Thaumaturus, Prothymallus) being certainly referable to this family. The recent genera may be grouped as follows:—

A. Air-bladder present.

a. Branchiostegal rays 8 to 20; ventral rays 9 to 13; stomach siphonal; pyloric appendages more or less numerous (17 to 200). Breed in fresh water. Salmo, Brachymystax, Stenodus, Coregonus, Phylogephyra, Thymallus.

b. Branchiostegal rays 6; ventral rays 11 to 14; stomach caecal; pyloric appendages in moderate numbers (12 to 20). Argentina.

{566}c. Branchiostegal rays 6 to 10; ventral rays 6 to 8; stomach caecal; pyloric appendages few (2 to 11) or rather numerous. Osmerus, Thaleichthys, Mallotus, Plecoglossus, Hypomesus.

d. Branchiostegal rays 3 or 4; ventral rays 8 to 10; stomach caecal; pyloric appendages absent. Microstoma, Nansenia, Bathylagus.

B. Air-bladder absent; branchiostegal rays 3 to 6; ventral rays 6 or 7; stomach siphonal; pyloric appendages absent. Retropinna, Salanx.

Only about 80 species can, at present, be regarded as valid.

The beauty, gameness, and great economical value of the Salmonids have caused more attention to be bestowed on them than probably upon any other group of fishes. As Professor Smitt tells us, a Swedish proverb says "A dear child has many names," and this applies well to our Salmon and Trout, the species of which have been unduly multiplied by some writers. The genus Salmo, characterised by a large mouth and powerful dentition, is divided into three sections: Oncorhynchus, Quinnat Salmon, of the North Pacific, ascending rivers in North America and Asia, with 12 to 17 developed rays in the anal; Salmo, Salmon and Trout, with 8 to 12 rays in the anal, and teeth not only on the head of the vomer but also along its shaft, at least in the young, {567}represented in the seas and freshwaters of Europe, Asia, and North America, extending southwards to North-West Africa, Asia Minor, Northern Persia, the Hindu Kush, the head of the Gulf of California, and the Rio Grande; Salvelinus, Charr, with 8 to 10 rays in the anal, and teeth on the raised head of the vomer only, of the lakes of Northern and Central Europe and the rivers of the northern parts of Asia and North America as far north as 82° 34´, sometimes descending to the sea.

The changes in form and colour which these fishes undergo when passing from fresh water into the sea or when artificially transported from one place to another are very great, and this plasticity, together with the connecting links which render the naming of not a few specimens impossible, have caused most recent students of the genus Salmo, in Europe at least, to reduce many of the so-called species to the rank of local varieties, and even our common Brown Trout or Brook Trout (S. fario) is now generally regarded as not specifically separable from the anadromous Sea Trout (S. trutta). The anadromous true Salmon (S. salar) may be distinguished by its somewhat larger scales, there being only 11 or 12 in a transverse series running from the posterior border of the adipose fin forwards to the lateral line, Trout having 13 to 16. The Charr of the lakes of Wales, the North of England, Scotland, and Ireland are also regarded as mere varieties of the common Northern migratory Charr (S. alpinus), of which the "Omble Chevalier" of the Swiss lakes and the "Saeblings" of the Alpine lakes of Germany and Austria are likewise varieties. An allied species (S. fontinalis) has been introduced into England from North America, as well as a true Trout (S. irideus). The large size of the eggs, their lack of {568}adhesiveness, and the fact that the ova fall into the abdominal cavity, out of which they may easily be squeezed, renders artificial impregnation particularly easy, and the species of Salmo have always occupied the first place in the annals of fish-culture. Fertilised eggs are transported in ice, the development being simply suspended for several weeks, and several forms of British and American Salmonidae have thus been introduced into New Zealand and Tasmania, where some have thoroughly established themselves.

The White-Fish, Coregonus, are more numerous in species than Salmo, and as a rule more readily defined. They are easily recognised by their large silvery scales and their smaller mouth without or with minute teeth. Some, like the Houting (C. oxyrhynchus) of Northern Europe, occur in the sea, entering rivers to spawn, whilst others, like the Sik, Weiss, Felchen, or Lavaret (C. lavaretus), are confined to lakes. British species are the Gwyniad (C. clupeoides), of Loch Lomond, Haweswater, Ullswater, and Bala, the Vendace (C. vandesius), of Loch Maben, and the Pollan (C. pollan) of Lough Neagh in Ireland.

The Grayling (Thymallus vulgaris or vexillifer), with its high dorsal fin formed of about 20 rays, one of the handsomest British fishes, inhabits the rivers and lakes of Northern and Central Europe, and is represented by a few allied species in Asia and North America. It derives its name from having the odour of thyme.

The Smelt (Osmerus eperlanus) breeds in salt water, and although it often enters rivers, it does not ascend beyond tidal influence. The Capelin (Mallotus villosus), of the coasts of Arctic America and North-eastern Asia, deposits its eggs in the sand along the shores in incredible numbers, the beach becoming a {569}quivering mass of eggs and sand. Plecoglossus, from Japan and Formosa, is highly remarkable for its lamellar, comb-like, lateral teeth. The Siel-Smelts (Argentina) are deep-sea Salmonids of which examples have occasionally been taken off the coasts of Scotland and Ireland; large numbers have been brought from Norway to English markets. Bathylagus is still better adapted for life at great depths (down to 1700 fathoms), the eyes being of enormous size. As Dr. Günther has observed, "these fishes must be entirely dependent for vision on the phosphorescent light which is produced by other abyssal creatures. Not being fish of prey themselves, or only to a slight degree, they would be attracted by the light issuing from the Pediculates and Stomiatids of the deep, and thus form an easy prey to these fishes."

Secondary sexual characters are very strongly developed in many Salmonids. In adult males of Salmon, Trout, and Quinnat the snout becomes greatly distorted, both jaws being hooked and the base of the teeth more or less enlarged; in the latter species a fleshy hump is developed before the dorsal fin, and the scales of the back become embedded in the flesh. Pearl-like excrescences appear on the scales of many of the White-Fish during the breeding season, being more prominent in males than in females, and Mallotus villosus is so called from the villous bands formed by the scales of mature males, the scales above the lateral line and along each side of the belly becoming elongate-lanceolate, densely imbricated and produced into free, projecting points.[649]

The Pachyrhizodontidae, with the Cretaceous genus Pachyrhizodus, are placed by some authors with the Salmonidae, but the remains at present known are too fragmentary to afford a correct idea of their exact systematic position. There seems to be less justification for placing them among the Elopidae.

FAM. 18. ALEPOCEPHALIDAE.—Deep-sea Fishes similar in general structure to the Clupeidae and Salmonidae, but destitute of a postclavicle and of an adipose dorsal fin,[650] the rayed fin being situated far back on the body, in the caudal region, and opposed {570}or slightly anterior to the anal. The skeleton of Alepocephalus[651] is remarkable for its feeble ossification. Epipleurals and epineurals are present, and the bilateral division of the neural arch remains perfectly distinct throughout the praecaudal region, both halves being very loosely apposed. The air-bladder is absent. Ventrals are absent in Platytroctes, and the snout is much produced in Aulostomatomorpha.

Eleven genera are distinguished:—A, with scales: Alepocephalus, Conocara, Bathytroctes, Leptochilichthys, Narcetes, Platytroctes, Aulostomatomorpha. B, without scales:—Xenodermichthys, Aleposomus, Leptoderma, Anomalopterus.

Represented by about 35 species in nearly all the seas; as usual with deep-sea forms, individuals of the same species have been obtained from stations very remote from one another.

FAM. 19. STOMIATIDAE.—I would unite under this name the Stomiatidae and Sternoptychidae of Günther, an assemblage of aberrant deep-sea Fishes which agree in having the maxillary bone more developed than the praemaxillary, and beset with teeth, a character which differentiates them at once from all other deep-sea forms of this sub-order, as well as from the Scopelidae among the Haplomi. The ventral fins are usually inserted very far back, and the number of their rays varies from 5 to 8. Contrary to what occurs in other groups of fishes, the pectoral fins have a tendency to reduction, and actually disappear in some genera, whilst the ventrals remain well developed; whenever the pectoral fins are fully developed, as in Maurolicus, Chauliodus, Astronesthes, and Photichthys, the mesocoracoid arch is present.[652] The form of the body varies exceedingly, even within the smaller groups into which this family has been divided; it may be excessively short and compressed, or excessively elongate, {571}but the mouth and eyes are always large, these fish being essentially predatory; the dentition is often very powerful, and may extend to the palate or be confined to the jaws. The body is naked or scaly; luminous spots (photophores) are more or less developed.[653] The development and position of the vertical fin is highly variable within this group, and the several families which have been founded upon this character have no more taxonomic importance than in the better-understood groups Characinidae and Siluridae. All authors, besides, have been compelled to admit that the presence or absence of an adipose dorsal fin has no high significance in this case, a view which is further strengthened by Dr. Gilchrist's discovery, off the Cape of Good Hope, of a deep-sea Fish agreeing in every respect with Astronesthes, but for the presence of a small adipose fin, absolutely similar to the dorsal, but situated on the ventral side, immediately in front of the anus. Two species with similar ventral adipose fins have just been discovered by Dr. Brauer and referred to Astronesthes. I am therefore unable to adopt the elaborate arrangement in favour with the modern American school.

The genera may be arranged in five sub-families:—

I. Anal not exactly opposed to the rayed dorsal, or much longer than the latter; no hyoid barbel.

A. Rayed dorsal far forward, between pectorals and ventrals; pectorals well developed (CHAULIODONTINAE). Chauliodus.

B. Rayed dorsal above or behind the ventrals; pectorals well developed.

1. Body more or less elongate; ventrals well developed (GONOSTOMATINAE).

a. A hyoid barbel. Astronesthes.

b. No barbel. Bathylychnus, Gonostoma, Cyclothone, Triplophos, Photichthys, Bathylaco, Diplophos, Maurolicus, Ichthyococcus.

2. Body short and deep; ventrals rudimentary or absent (STENOPTYCHINAE). Argyropelecus, Sternoptyx, Polyipnus.

II. Dorsal and anal opposed to each other and very far back on the caudal region; pectorals often reduced or absent; hyoid barbel often present. {572}(STOMIATINAE). Stomias, Macrostomias, Echiostoma, Opostomias, Pachystomias, Photonectes, Malacosteus, Thaumatostomias, Photostomias.

This family, comprising about 55 species, has a world-wide distribution, but most of the known forms have been obtained from the Atlantic; some of the species occur both in the Atlantic and the Indo-Pacific. Chauliodus, Astronesthes, and Stomias are among the fishes with the most formidable dentition.

FAM. 20. GONORHYNCHIDAE.—Margin of the upper jaw formed by the praemaxillaries and the maxillaries, the latter articulated above the former to the ethmoid. Supraoccipital in contact with the frontals, widely separating the small parietals; opercular bones well developed; symplectic present. Basis cranii simple. Mouth small and toothless, inferior, surrounded by thick, fringed lips. Four branchiostegal rays. Head and body entirely covered with small spiny scales. Praecaudal vertebrae with strong parapophyses, to the extremity of which slender ribs and epipleurals are attached. No postclavicle. Pectoral fins inserted low down, folding like the ventrals; latter with 10 rays.

The single existing species, Gonorhynchus greyi, is characterised by an elongate, cylindrical body, a pointed projecting snout bearing a single barbel, short dorsal and anal fins, the former opposed to the ventrals, and the gill-membranes broadly attached to the isthmus. Teeth are present on the pterygoid and hyoid bones. No suborbital arch. Vertebrae, 45 + 20. Air-bladder absent. Its distribution is a very wide one, the species being on record from the coasts of the Cape of Good Hope, Australia, New Zealand, and Japan.

The genus Notogoneus, from the freshwater Eocene beds of France and North America, has been referred to this family by Cope, and has been shown by A. S. Woodward to be closely related to Gonorhynchus, differing only in the absence of teeth on the palate and tongue, and in the more forward position of the dorsal fin. The genus Charitosomus, with several species {573}from the Upper Cretaceous of Westphalia and Mount Lebanon, has also been included in this family, but the precise shape and character of the scales have not yet been ascertained.

FAM. 21. CROMERIIDAE.—Margin of the upper jaw formed by the praemaxillaries and the maxillaries. Supraoccipital large and widely separating the very small parietals; opercular bones well developed; symplectic absent. Basis cranii simple. Mouth small and toothless, inferior; gill-opening narrow. Three branchiostegal rays. Body naked. Praecaudal vertebrae with parapophyses; ribs and epipleurals slender. No postclavicle. Pectoral fin inserted low down, folding like the ventrals.

A single genus, Cromeria, recently discovered in the White Nile. In its elongate, naked body and the posterior position of the dorsal fin, it resembles the Galaxiidae, to which it was at first referred. But this allocation has proved to be incorrect, now that the osteological structure of the minute Fish (only about 30 mm. long) has been worked out by Swinnerton.[654] The vertebrae number 42 to 45 (28-30 + 14-15). A long, slender air-bladder is present.

SUB-ORDER 2. OSTARIOPHYSI.

Air-bladder, if well developed, communicating with the digestive tract by a duct. Pectoral arch suspended from the skull; mesocoracoid arch present. Fins without spines, or dorsal and pectoral with a single spine formed by the co-ossification of the segments of an articulated ray. The anterior four vertebrae strongly modified, often co-ossified and bearing a chain of small bones (so-called Weberian ossicles) connecting the air-bladder with the ear.

This is one of the most natural groups of the Class Pisces, although its members are so diversified in outward appearance as to have been widely separated in the systems of older authors. It is to Sagemehl[655] that is due the credit of having first grouped, under the above name, the Characines, the Carps, the Cat-Fishes, and the Gymnotids, the relations of which had been realised, to a certain extent, by Cope. But it was not until the homology throughout the group of the ossicula auditus, first described by E. H. Weber in 1820, had been demonstrated by Sagemehl that the justification for the course here followed appeared in its full {574}strength, as such an agreement in the structure of so complicated and specialised an apparatus can only be the result of a community of descent of the families which are possessed of it. It is invariably the anterior four vertebrae that take part in the support of the Weberian apparatus. The first vertebra is much reduced; its upper arch is absent and replaced by the ossicles termed claustrum and scaphium[656] (the former being perhaps nothing but the modified neural arch), which fill in the space between the exoccipital and the neural arch of the second vertebra; the principal piece of the apparatus, the tripus, variable in form, is related to the third vertebra, of which it is regarded as a modified rib; a fibrous ligament extends from the anterior extremity of the tripus to the scaphium, and in this ligament is inserted the fourth piece, the intercalarium. The various forms of this sub-order also show a complete agreement in the spinal nerves which pass through these ossicles. The parietal bones either separate the frontals from the supraoccipital or are fused with the latter.

This sub-order is divided into six families. The Characinids are the most generalised, and the others are probably derived from them in the manner expressed by the following diagram:—

Loricariidae Aspredinidae | | +––––––––––––+––––––––––+ | Cyprinidae Siluridae Gymnotidae | | | +–––––––––––+–––––––––––+ | | | +–––––––––––––––––––+–––––––––––––––––+ | Characinidae.

SYNOPSIS OF THE FAMILIES

I. Parietal bones distinct from the supraoccipital; symplectic present; ribs mostly sessile, all or the greater number of the praecaudal vertebrae without parapophyses.

Mouth not protractile, usually toothed; pharyngeal bones normal; body scaly; an adipose dorsal fin often present 1. Characinidae.

Mouth not protractile, usually toothed; pharyngeal bones normal; body Eel-shaped, naked or scaly; vent under the head or on the throat 2. Gymnotidae.

Mouth usually more or less protractile, toothless; lower {575} pharyngeal bones large, falciform; body naked or scaly; no adipose dorsal fin 3. Cyprinidae.

II. Parietal bones usually fused with the supraoccipital; symplectic absent; body naked or with bony scutes; mouth usually toothed, with barbels; adipose fin often present.

Ribs attached to strong parapophyses; operculum well developed 4. Siluridae.

Ribs sessile; parapophyses absent; operculum more or less developed; mouth inferior 5. Loricariidae.

Ribs sessile; strong parapophyses to the vertebrae; operculum absent 6. Aspredinidae.

FAM. 1. CHARACINIDAE.—Mouth non-protractile, usually bordered by the praemaxillaries and the maxillaries, rarely by the praemaxillaries only; jaws usually toothed. Parietal bones united in a sagittal suture or separated by a fontanelle; opercular bones well developed; symplectic present. Pharyngeal bones normal, with small teeth. Ribs mostly sessile; no parapophyses in the thoracic region; epipleurals and epineurals, mostly free floating. Pectoral fins inserted very low down, folding like the ventrals. Body covered with scales. An adipose dorsal fin often present.

This is a very generalised type, although perhaps not directly derived from the bony Ganoids, as believed by Sagemehl. The species number about 500, and are confined to the freshwaters of Africa and Central and South America. The classification of the family is still in an unsatisfactory state, but the division into the following groups (hardly deserving the rank of sub-families), although quite provisional, appears preferable to the highly artificial arrangement hitherto adopted:—

I. No adipose fin.

A. ERYTHRININAE.—Carnivorous; teeth strong; maxillary large; gill-openings wide; scales cycloid. American: Macrodon, Erythrinus, Lebiasina, Pyrrhulina, Corynopoma.

II. Adipose fin usually present.

B. HYDROCYONINAE.—Entirely or partially carnivorous; teeth strong; maxillary well developed; scales cycloid; lateral line usually nearer ventral than dorsal outline (sometimes only on the tail). African: Sarcodaces, Hydrocyon, Bryconaethiops, Alestes, Micralestes, Petersius. American: Acestrorhynchus, Boulengerella, Acestrorhamphus, Crenuchus, Chalceus, Brycon, Bryconops, Bryconodon, Creagrutus, Chalcinus, Brachychalcinus, Pseudocorynopoma, Stichonodon, Gastropelecus, Tetragonopterus, Scissor, Chirodon, Piabucina, Iguanodectes, Aphiocharax, Salminus, Oligosarcus, Agoniates, Paragoniates, Leptagoniates, Anacyrtus.

{576}C. SERRASALMONINAE.—Carnivorous; teeth strong; belly serrated; scales cycloid. American: Serrasalmo, Myletes, Myleus, Metynnis, Catoprion.

D. ICHTHYOBORINAE.—Carnivorous; teeth strong; maxillary very small; upper jaw movable; scales ciliated. African: Eugnathichthys, Paraphago, Mesoborus, Phago, Ichthyoborus, Neoborus.

E. XIPHOSTOMINAE.—Carnivorous; teeth very small; maxillary rather small; scales ciliated. American: Xiphostoma.

F. ANOSTOMINAE.—Herbivorous, entirely or partially; teeth well developed in both jaws; maxillary very small; gill-openings narrow; scales cycloid. American: Anostomus, Leporinus, Characidium, Chorimycterus, Nanostomus, Nanognathus.

G. HEMIODONTINAE.—Partially herbivorous; dentition imperfect; maxillary well developed; scales cycloid. American: Hemiodus, Caenotropis, Saccodon, Parodon.

H. DISTICHODONTINAE.—Entirely or partially herbivorous; teeth small but well developed; maxillary well developed; scales ciliated. African: Nannaethiops, Neolebias, Distichodus, Nannocharax, Xenocharax.

I. CITHARININAE.—Herbivorous; teeth minute or absent; maxillary small; scales cycloid or ciliated. African: Citharinus, Citharidium. American: Prochilodus, Curimatus.

The genera in the above sub-families are mostly founded on the dentition and the extent of the praemaxillary and maxillary bones, which are astonishingly varied, as may be seen from the annexed figures showing the open mouths of a few of the most remarkable types. As I have already pointed out, the character often given as diagnostic of this family, viz. the maxillary forming part of the oral border, is not absolutely constant; this bone is often much reduced, and it is entirely excluded from the mouth in Ichthyoborus and Neoborus. The branchiostegal rays {577}number 3 to 5 only. The fins never bear pungent spines, and the ventrals have 6 to 13 rays. Barbels are absent. In most of the herbivorous forms the brain-case is produced forward to the nasal capsule, whilst in most of the carnivorous forms they are separated by an interorbital septum; but there are exceptions to this correlation, and as otherwise closely related genera may differ in this respect, I have not been able to make use of the character in defining sub-families.

{578}The air-bladder is divided into an anterior and a posterior part by a constriction; the posterior part is the longer, and its anterior portion is cellular in Erythrinus and Lebiasina. Pyloric appendages to the stomach, which are constantly absent in the Cyprinids, are more or less numerous. An accessory respiratory organ in a diverticulum above the fourth branchial arch has been observed in Xenocharax and Citharinus.

The appearance and habits of the genera which compose this family vary greatly. Some resemble the Cyprinids and are mainly vegetarians, whilst others recall Salmonids and Pike. Among the most formidable are Hydrocyon, the Dogs of the Water, or Kelb-el-Bahr of the Arabs, with their powerful jaws with shark-like teeth, visible when the mouth is closed, and which grow to the size of the Salmon. The five known species inhabit the Nile and the rivers and lakes of tropical Africa. No less ferocious are the "Piranha" or "Cariba" (Serrasalmo) of South America, whose bite has been compared to the cut of a razor. They abound in some rivers and are much dreaded by people having to enter the water, as they fiercely bite off big pieces of flesh as with a pair of scissors, and the smell of blood is said to attract them by thousands; they show a great tenacity of life and can remain for hours out of the water. Serrasalmo niger has been observed by Schomburgk to produce a grunting noise in the water. Salminus orbignianus, of the Plate River, "Dorado" of the Spaniards, which reaches a length of 3 feet, has the predacious habits of the Pike, and follows other fishes moving in shoals; its flesh is much valued, although very full of bones, like that of all Characinids.

{579}As an example of phytophagous types may be mentioned the Moon-Fish of the Nile (Citharinus geoffroyi), with its feeble dentition, deep compressed body, and falciform dorsal fin; it is often represented on the monuments of the ancient Egyptians.[657]

FAM. 2. GYMNOTIDAE.—Mouth non-protractile, bordered by the praemaxillaries and the maxillaries, the latter sometimes much reduced; jaws usually toothed. Parietal bones united in a sagittal suture, or separated by a fontanelle; opercular bones well developed; symplectic present. Pharyngeal bones normal, with small teeth. Anterior ribs sessile, the posterior inserted on transverse processes; epipleurals and epineurals. Body much elongate, Eel-like, naked or scaly; dorsal fin absent or reduced to an adipose strip; anal very long; caudal rudimentary or absent;[658] ventrals absent. Vent under the head or at a very short distance behind the throat. Gill-openings narrow.

In spite of their external appearance, these fishes have nothing to do with the Eels; they are strongly modified, degraded Characinids, as first pointed out by Reinhardt. The few genera and species (about 30) are confined to the fresh waters of Central and South America. No fossils are known. Eight genera may be distinguished:—

A. A cranial fontanelle; maxillary bone larger than the praemaxillary; anterior nostril on the upper surface of the head; vent below the head; body scaly: Sternopygus, Eigenmannia, Sternarchus, Rhamphosternarchus, Rhamphichthys, Steatogenys.

B. No cranial fontanelle; maxillary bone very small; anterior nostril on the upper lip; vent on the throat.

a. Body scaly: Carapus.

b. Body naked; an electric organ: Gymnotus.

The mouth is small or very small, and the modifications of the snout in the genera Sternarchus and Rhamphichthys recall those noticed among the Mormyridae. The air-bladder is divided into an anterior and a posterior part, united by a slender duct.[659] The vertebrae vary in number from 70 (Sternopygus) to 240 (Gymnotus). Gymnotus is unique in this sub-order in having as many as 8 pterygials (actinosts) to the pectoral fin, as in Anguilla.

{580}[Illustration: FIG. 351.—Outlines of heads, showing shape of snout and position of vent (v). A, Sternarchus albifrons; B, Sternarchus macrostoma; C, Rhamphosternarchus curvirostris; D, Rhamphosternarchus tamandua.]

The best known member of this family is the so-called Electric Eel (Gymnotus electricus), of the Orinoco, Amazons, and intermediate river-systems. It grows to a length of 8 feet and the thickness of a man's thigh, and is much feared for the electric shocks it is able to discharge. The "Tremblador," as it is called by the Spanish-speaking inhabitants of the Orinoco district, is found only in marshes and in comparatively shallow parts of rivers, to the great annoyance of travellers who have to ford at such points, beasts of burden being frequently knocked down by the electric shock. Specimens have often been exhibited alive in this country; two brought to London in the year 1842, neither of them weighing more than one pound, had by 1848 reached the weights of 40 and 50 pounds respectively. About four-fifths of the length of the fish is occupied by the tail, which contains the electric organ; this is formed by modified muscular tissue, and consists of two huge masses, longitudinal bands or columns, of cells filled with a jelly-like substance, occupying the whole of the caudal region below the vertebral column and separated by a narrow median septum; a smaller body, of similar structure, extends along each side at the base of the anal fin. The whole apparatus is supplied with a great number of nerves branching from the spinal nerves. The electrical apparatus is exercised by the will of the fish, even to {581}a distance, but this faculty is exhausted by continuous employment, and is recovered during repose. Although apparently not exempt from exaggeration and fable, Humboldt's account in Observations de Zoologie, p. 497, is recommended for further information on the habits and modes of capture of Gymnotus.[660]

FAM. 3. CYPRINIDAE.—Mouth usually more or less protractile, toothless, bordered by the praemaxillaries and the maxillaries, or, more frequently by the praemaxillaries only. Parietal bones united in a sagittal suture, or separated by a fontanelle; opercular bones well developed; symplectic present. Lower pharyngeal bones falciform, subparallel to the branchial arches, provided with teeth arranged in one, two, or three series, and often remarkably specialised. Ribs mostly sessile; no parapophyses in the thoracic region; epipleurals and epineurals, mostly free, floating. Pectoral fins inserted very low down, folding like the ventrals. Body naked or scaly. No adipose dorsal fin.

The brain-case is produced forward to the nasal capsule. The branchiostegal rays are reduced to 3; the branchiostegal membrane is usually more or less extensively grown to the isthmus. The suborbital branch of the sensory canals is usually produced on the operculum, as in the Characinidae. The ventral rays number 7 to 12, rarely 5 or 6. Pyloric appendages to the stomach are absent.

Freshwater fishes feeding on vegetable substances or small animals, and dispersed over the whole world with the exception of South America, Madagascar, Papuasia, and Australasia. The species are exceedingly numerous, about 1300 being known, referable to four sub-families, as proposed by Sagemehl.

(i.) CATOSTOMINAE.—Margin of upper jaw formed in the middle by the small praemaxillaries and on the sides by the maxillaries, which are hidden in thick fleshy lips; no barbels; pharyngeal teeth in a single row, very numerous, comb-like; air-bladder large, divided into two or three parts by transverse constrictions, not surrounded by a bony capsule. Mostly from North America; two species from China and one from Eastern Siberia. Fossil in the Lower Tertiary of North America.

Principal genera:—Sclerognathus, Carpiodes, Catostomus, Moxostoma.

{582}(ii.) CYPRININAE.—Maxillaries not bordering the mouth; barbels absent, or one or two pairs; pharyngeal teeth in one to three rows, in small number, often very large, and working against a sclerous plate attached to a ventral process of the basi-occipital, which extends under the anterior vertebrae. Air-bladder usually large and divided into an anterior and a posterior part, rarely tripartite, not surrounded by a bony capsule. The great bulk of the family, represented in every part of its range. Remains of several of the existing genera have been found in Oligocene and later beds of Europe, Sumatra, and North America.

Principal genera:—Cyprinus, Catla, Catlocarpio, Osteochilus, Labeo, Discognathus, Psilorhynchus, Capoëta, Barbus, Gobio, Pseudogobio, Saurogobio, Rhinogobio, Oreinus, Schizothorax, Ptychobarbus, Gymnocypris, Diptychus, Aulopyge, Ceratichthys, Pimephales, Campostoma, Cochlognathus, Exoglossum, Meda, Lepidomeda, Rhinichthys, Rohteichthys, Leptobarbus, Rasbora, Luciosoma, Nuria, Amblypharyngodon, Cyprinion, Semiplotus, Xenocypris, Leuciscus, Tinca, Leucosomus, Chondrostoma, Achilognathus, Rhodeus, Danio, Pteropsarion, Hypophthalmichthys, Abramis, Nematabramis, Aspius, Leucaspius, Alburnus, Barilius, Bola, Neobola, Chelaethiops, Chela, Culter, Pelecus, Parapelecus, Cachius, Opsariichthys, Scombrocypris, Squaliobarbus, Luciobrama.

(iii.) COBITIDINAE.—Maxillaries not bordering the mouth; barbels three to six pairs; pharyngeal teeth in one row, in moderate number. Anterior part of the air-bladder divided into a right and left chamber separated by a constriction, and enclosed in a bony capsule, the posterior part free, or absent. Loaches, characterised externally by a low, elongate body, without or with minute scales. Europe, Asia, Abyssinia. Miocene of Oeningen.

Principal genera:—Botia, Lefua, Diplophysa, Nemachilus, Misgurnus, Cobitis, Lepidocephalichthys, Acanthophthalmus, Eucirrhichthys, Apua.

(iv.) HOMALOPTERINAE.—Maxillaries not bordering the mouth, which is inferior; barbels three or four pairs; pharyngeal teeth in one row, in moderate number. Air-bladder rudimentary, divided into two lateral halves, encased in a bony capsule. Mountain forms with depressed head and horizontally expanded paired fins. China, India, Further India, Malay Peninsula and Archipelago.

Genera:—Homaloptera, Helgia, Glaniopsis, Gastromyzon.

The recently described Gyrinochilus, from Borneo, resembling {583}Homaloptera in habit, with two gill-clefts on each side, an upper and a lower, a tadpole-like mouth without barbels, and a small, free air-bladder, should probably be regarded as the type of a fifth sub-family.

Many of the genera of the Cyprininae are partly founded on the shape and the disposition of the pharyngeal teeth, which, adapted to various requirements, may be conical, hooked, spoon-shaped, molariform, etc., etc. The importance attached to the disposition of these teeth in one, two, or three series for the definition of genera, has been rather exaggerated.[661]

The Cyprinids constitute the majority of the freshwater fishes in Europe, Asia, and North America; they are comparatively few (about 100 species) in Africa, where they coexist with the Characinids. Some, like the Carp (Cyprinus carpio) and the Tench (Tinca vulgaris), are sluggish, except during the breeding season, when they show great excitement and indulge in leaps out of the water; others, like the Bleak (Alburnus lucidus) are constantly on the move in large shoals near the surface; whilst others again, like the M‘Biriki of Lake Tanganyika (Barbus tropidolepis), behave after the manner of Salmon and Trout, {584}travelling long distances, against rapids and over waterfalls, to reach their breeding places at the heads of rivers. During the breeding season, the males of many species assume a more brilliant livery, or develop pearl-like or spiny excrescences on various parts of the head, or also on the body and fins.[662] Cyprinids are oviparous, with the exception of a small Barbel from Natal, discovered and described by Prof. Max Weber as Barbus viviparus.

A most striking instance of symbiosis is offered by a little Carp-like fish of Central Europe, the "Bitterling" (Rhodeus amarus). The genital papilla of the female acquires a great development during the breeding season, becoming produced into a tube nearly as long as the fish itself; by means of this ovipositor the comparatively few and remarkably large eggs, measuring 3 millimetres in diameter—the fish being only 60 to 80 millimetres long—are introduced through the gaping valves, between the branchial laminae of pond mussels (Unio and Anodonta) where, after being inseminated, they undergo their development, the fry leaving their host about a month later, having attained a length of 10 or 11 millimetres.[663] The mollusc reciprocates by throwing off its embryos on the parent fish, in the skin of which they remain encysted for some time, the period of reproduction of the fish and mussel coinciding.

Some members of this family grow to a very large size,—4 to 6 feet; such is the case with the Carp, a native of Asia, introduced into England towards the beginning of the seventeenth century; the Catla (Catla buchanani) of India, Burma, and Siam; the Mahaseer (Barbus mosal) of the mountain streams of Asia, the scales of which may be as large as the palm of a hand; and Hypophthalmichthys molitrix of China and Manchuria, remarkable for the low position of the eyes, the fusion of the gill-rakers into thin plates of spongious appearance, which must act as a most efficient sifting apparatus, and the presence of an involuted problematic superbranchial organ to each branchial arch.[664]

Among well-known aberrations produced by artificial selection may be mentioned the "Leather Carp," a race in which the scales are either lost or much reduced in number, and enlarged {585}along the lateral line and the back, and the Gold-Fish, a variety of Cyprinus carassius, remarkable for its golden or bright red colour, or its perfect albinism, as well as its monstrous form the Telescope Fish, with enormously projecting eyes, and enlarged, horizontally spread caudal fin.[665] This family has also yielded numerous more or less well-established examples of hybridism, congeneric and digeneric, originally described as distinct species, the produce of which is believed to be in some cases fertile for at least one generation.

The crystalline silvery colouring matter of various Cyprinids is said to have been employed from time immemorial for ornamental purposes by the Chinese. The well-known and important industry of "Essence Orientale" and artificial pearls, carried on in France and Germany with the scales of the Bleak, was not introduced before the middle of the seventeenth century.

The Loaches, Cobitidinae, which form a very natural sub-family, are small fishes, few species growing to a foot in length, mostly living in small streams and ponds. Many delight in the mud at the bottom, in which they move like Eels. In some cases, the branchial respiration appears to be insufficient, and the intestinal tract acts as an accessory breathing organ. The air-bladder, which is partially encased in a bony capsule, may be so reduced as to lose its hydrostatic functions and becomes transformed into a sensory organ, its outer exposed surface being connected with the skin by a meatus between the bands of muscle, and conveying the thermo-barometrical impressions to the auditory nerves; hence the name of "Wetterfisch," by which Loaches are known in some parts of Germany.

The Homalopterinae are more or less perfectly adapted to life in rapid streams, the most remarkable in this respect being {586}Gastromyzon of North Borneo, in which the pectoral and ventral fins are much expanded to form, with the belly, a sucker by which the fish adhere to the stones of mountain torrents, showing a remarkable analogy to Exostoma among the Silurids.[666]

FAM. 4. SILURIDAE.—Mouth non-protractile, bordered by the praemaxillaries and the maxillaries, or by the praemaxillaries only, the maxillaries being often rudimentary and supporting the base of a barbel; jaws usually toothed. Parietal bones usually confluent with the supraoccipital, forming a single large plate (parieto-occipital); symplectic and suboperculum absent. Pharyngeal bones normal, with small teeth. Ribs attached to the lower surface of long parapophyses; epipleurals absent. Pectoral fins inserted very low down, folding like the ventrals, often armed, like the dorsal, with a strong bony spine. Body naked or with bony plates. An adipose dorsal fin often present. One to four pairs of barbels.

The skull and the opercular apparatus show a reduction in the number of elements as compared with the Characinids and Cyprinids, such as the absence of the metapterygoid, the often rudimentary, rod-like condition of the palatine, and the fusion of the parietals with the supraoccipital.[667] The scapular arch is solidly united to the skull and is often very massive, and the occiput may be connected with the base of the dorsal fin by a buckler formed by the expansion of the first and second inter-neural bones. The pterygials or supports of the pectoral rays are large and reduced to two or three.[668] Teeth are rarely present {587}on the maxillary bones (Diplomystes, Eutropiichthys), being usually confined to the praemaxillaries and dentaries; they often occur on the palate. The branchiostegal rays vary from 4 to 17. The lips are sometimes much developed, and may form a sucking disk, as in Euchilichthys and Exostoma. As in the Cyprinids, the pungent spines which may arm the fins have nothing in common with the true spines of Acanthopterygians; they result from the co-ossification, with age, of successive articles; but, contrary to the condition in the Cyprinids, the axis of the spine is single, not double. The ventral rays vary from 6 to 16, 6 being the most frequent number. Some of the exterior vertebrae may be solidly fused together, and also with the occipito-nuchal buckler. Prof. Ramsay Wright[669] has shown, by a study of the development, that the complex which follows the first vertebra, which is more or less rudimentary, if distinct, represents the fusion of the 2nd, 3rd, and 4th vertebrae, without even nerve-foramina denoting its compound origin; the first, strongly developed, transverse process represents that of the 4th vertebra. The air-bladder is usually large and trilocular, but additional septa may greatly complicate its structure, and external diverticula may also exist;[670] it may be more or less reduced and entirely or partially enclosed in a bilateral bony capsule formed by the transverse processes of the vertebrae, and sometimes (Nematogenys) ankylosed to the skull. In a few genera, like Cetopsis, the air-bladder seems to be altogether absent: it is reduced to two small oval sacs encased in the large compound anterior vertebra. As in Loaches, the air-bladder is often in immediate contact with the skin behind the shoulder-girdle. The intestinal tract may be simple and short (carnivorous forms) or extremely long and convoluted (Callichthys); as in Cyprinids, pyloric appendages are absent.

Cat-Fishes, as Silurids are usually called, are a large family embracing some 1000 species, spread over the fresh waters of all parts of the world, but mostly from between the tropics.[671] Only a few are marine (Plotosus, Arius, Galeichthys).

This family may be divided into eight subfamilies

{588}(i.) CLARIINAE.—Dorsal and anal fins very long, extending to the caudal; gill-membranes free, or narrowly united to the isthmus. Asiatic-African genera: Clarias, Heterobranchus, Plotosus. Asiatic-Australian: Copidoglanis. Asiatic: Cranoglanis. African: Clariallabes, Allabenchelys, Gymnallabes, Channalabes. Australian: Cnidoglanis.

(ii.) SILURINAE.—Dorsal fin very short or absent, anal very long; gill-membranes free. Europaeo-Asiatic: Silurus. Asiatic: Silurodon, Silurichthys, Saccobranchus, Wallago, Belodontichthys, Cryptopterus, Callichrous, Hemisilurus, Ailiichthys, Ailia, Schilbichthys, Laïs, Pseudeutropius, Pangasius, Osteogeniosus, Helicophagus, Silondia. African: Eutropius, Schilbe, Siluranodon, Physailia, Parailia. Australian: Eumeda, Neosilurus.

(iii.) BAGRINAE.—Dorsal fin short, followed by a more or less elongate adipose fin; anal short or moderate; gill-membranes free. Asia, Africa, America, Australia: Arius. Asia and America: Amiurus. Asiatic: Macrones, Pseudobagrus, Liocassis, Bagroides, Bagrichthys, Rita, Acrochordonichthys, Acysis, Olyra, Hemipimelodus. African: Bagrus, Clarotes, Chrysichthys, Gephyroglanis, Auchenoglanis, Notoglanidium, Anoplopterus, Galeichthys. American: Diplomystes, Paradiplomystes, Aelurichthys, Genidens, Noturus, Callophysus, Pimelodus, Pimelodina, Nanoglanis, Heptapterus, Nematogenys, Pariolius, Pirinampus, Conorhynchus, Notoglanis, Callophysus, Sorubim, Piramutana, Bagropsis, Sciades. Australian: Nedystoma, Pachyula.

(iv.) DORADINAE.—A short-rayed dorsal fin and an adipose, the latter sometimes replaced by a second rayed dorsal; anal short or moderate; gill-clefts more or less widely interrupted below. African: Synodontis, Chiloglanis, Atopochilus, Euchilichthys, Mochocus, Doumea, Phractura, Paraphractura, Andersonia, Trachyglanis, Belonoglanis. Asiatic: Bagarius, Glyptosternum, Gagata, Pseudecheneis, Exostoma, Sisor, Breitensteinia, Sosia, Chaca. South American: Doras, Oxydoras, Leptodoras, Physopyxis, Glanidium, Centromochlus, Wertheimeria, Cetopsis.

(v.) MALOPTERURINAE.—No rayed dorsal fin, an adipose; anal short; gill-clefts interrupted below. African: Malopterurus.

(vi.) CALLICHTHYINAE.—Dorsal, anal, and adipose fins short; body completely cuirassed; praemaxillaries much reduced, the border of the upper jaw formed mainly by the maxillaries. South American: Callichthys, Corydoras.

{589}(vii.) HYPOPHTHALMINAE.—Dorsal fin short, behind the ventrals, anal long; gill-clefts wide or interrupted below. South American: Ageniosus, Trachelyopterus, Auchenipterus, Epapterus, Tetranematichthys, Hypophthalmus, Helogenes.

(viii.) TRICHOMYCTERINAE.—Dorsal fin short, far back, behind the ventrals; no adipose fin; anal short; operculum and interoperculum armed with erectile spines. South American: Trichomycterus, Eremophilus, Stegophilus, Vandellia, Acanthopoma.

Our knowledge of the distribution in time of the Silurids is still very scanty, and throws no light on the derivation of the group. Arius, and two genera apparently related to it, Rhineaster and Bucklandium, have left remains in the Eocene of Europe and North America, and traces of various recent genera have been found in later Tertiary deposits in Europe, Asia, and North and South America.

The habits of the Silurids are extremely diversified, and the shape of the body varies accordingly. The body may be very short and the head enormous and excessively depressed, for instance in the Indo-Burmese Chaca lophioides, which, as its name implies, resembles the Fishing-Frog or Angler; stout and Cottus-like in some South American Pimelodus; Loach-like in Trichomycterus and Stegophilus; more or less Eel-shaped in Clarias and its allies, etc.; the extreme of slenderness obtains in the African Channalabes, the body being excessively elongate (over 100 vertebrae), the ventral fins absent, and the pectorals rudimentary or absent. Among other remarkable forms may be mentioned the Indian Sisor, which resembles Aspredo, and in which the upper caudal ray is much thickened and greatly prolonged; Pseudecheneis, living in rapids of the Himalayas and Khasia hills, provided with a transversely plaited ventral disk between the pectoral fins; the African Phractura and Andersonia, resembling Loricaria; and the likewise African Belonoglanis, comparable to a Needle-Fish. The spines which so frequently arm the dorsal and pectoral fins may be barbed or serrated, and constitute formidable defensive weapons; in the South American Ageniosus valenciennesi, the maxillary bone is transformed into a strong, barbed, erectile spine, replacing the barbel. Stings of even the smaller Cat-Fish are at least as painful as that of a bee, and this is probably due to some poisonous property of the dermal secretion of the Fish.

{590}[Illustration: FIG. 356.—Harmout, Clarias anguillaris (after Valenciennes). ¼ nat. size.]

Cope believed an orifice at or above the axil of the pectoral fin in Noturus to be the opening of the duct of a poison-gland; "from it may frequently be drawn a solid gelatinous style ending in a tripod, each limb of which is dichotomously divided into short branches of regular length." I think this condition of things has nothing to do with a poison-organ, and is merely a repetition of what is observed in Loaches and in the Characinid Xenocharax, where I have found a gelatinous substance filling the short duct by which the membrane of the air-bladder is placed in communication with the skin and the sensory organ of the lateral line. Most Silurids can live in very foul water, taking in air from the surface, and spend a comparatively long time out of the water, without being possessed of any special apparatus for atmospheric respiration. A few genera, however, are provided with an accessory breathing organ: in Clarias, Heterobranchus, and allies, there is a dendritic superbranchial organ, in Saccobranchus a long air-sac, extending from the first branchial cleft along the side of the body, as described above, p. 295; and these Fish can live for days on land. Clarias lazera has been observed, in Senegambia, to spend several months of the dry season in burrows, from which it emerges at night to crawl about in search of food. Many Silurids, but especially Doras and Synodontis, are known to produce sounds in and out of the water by means of a special mechanism of the air-bladder and the processes of the vertebrae above it, combined with the movements of the pectoral spine grinding in the glenoid cavity.[672] In South America, Doras has been observed to move rapidly on land, projecting itself forward on the pectoral spines by the elastic spring of the tail, travelling long journeys over land, from one drying pond to another, spending whole nights on the way; these migrations sometimes take place {591}in numerous bands, baskets of the small Fish being filled by the Indians who come across them.[673] The African Synodontis are much in the habit of floating or swimming leisurely on the surface with the belly in the air, as was well known to the ancient Egyptians, who have frequently depicted the Fish in this anomalous position. A curious fact in connexion with this habit is that S. membranaceus and S. batensoda, in which it has most frequently been observed, show an inversion of the ordinary mode of coloration, the lower parts being dark brown or black and the upper pale silvery grey. The electric Cat-Fish (Malopterurus electricus), is also a native of Africa, occurring all over the tropical parts of that continent and also in the Lower Nile, growing to a length of three feet. Its flesh is more esteemed than that of other Silurids. It avoids light and is slow in its movements. The electrical apparatus differs absolutely from that of all other Fishes, being derived from the integument, belonging to the glandular system, and surrounding the whole body with a thick coat of grease or gelatinous substance; the apparatus is governed by a single nerve on each side proceeding from a huge ganglionic cell at the anterior extremity of the spinal cord.[674] The shocks given by Madopterurus are very powerful, and the Fish is called "Raad" by the Arabs, a name which means "thunder." Kept in an aquarium with other Fishes, even of the same species, the "Raad" soon kills its companions.

{592}[Illustration: FIG. 358.—Callichthys littoralis, from South America. ⅔ nat. size.]

In this family the eggs and young are usually looked after by the parents. Aristotle observed that the male of the European Silurus glanis watches over and defends the eggs. In one of the commonest North American Cat-Fishes, Amiurus nebulosus, a species which has been largely introduced into some parts of Europe of late, now thriving in many ponds and more or less polluted streams of the Continent, the eggs are deposited near the banks of weedy ponds and rivers without currents, in concealed places beneath logs, stumps, or even in pails or other receptacles, failing which both parents join in excavating a sort of nest in the mud, a work often requiring two or three days of incessant labour. The male watches over the eggs, and later leads the young in great schools near the shore, seemingly caring for them as the hen for her chickens.[675] The Doras and the Callichthys of South America, according to Hancock[676] and Vipan,[677] build regular nests of grass or leaves, sometimes placed in a hole scooped out in the bank, in which they cover their eggs and defend them, male and female sharing in this parental duty. In the likewise South American Corydoras (Callichthys paleatus), as observed by Carbonnier,[678] a lengthy courtship takes place, followed by an embrace, during which the female receives the seminal fluid in a sort of pouch formed by the folded membranes of her ventral fins; immediately after, five or six eggs are produced and received in the pouch, to be afterwards carefully placed in a secluded spot. This operation is repeated many times, until the total number of eggs, about 250, have been deposited. In {593}accordance with these pairing habits, the pectoral spines of the male, which are used in amplexation, are longer and stronger than those of the female. These Fish are monogamous, and both parents remain by the side of the nest, furiously attacking any assailant. Dr. R. Semon[679] has made observations in Queensland on the habits of Arius australis, which builds nests in the sandy bed of the Burnett River. These nests consist of circular basin-like excavations, about 20 inches in diameter, at the bottom of which the eggs are laid, and covered over by several layers of large stones. A still more efficient protection is afforded their progeny by the marine and estuarine species of Arius,[680] Galeichthys,[681] and Osteogeniosus,[682] the male, more rarely the female, carrying the eggs in the mouth and pharynx; these eggs, few in number, are remarkably large, measuring as much as 17 or 18 millimetres in diameter in Arius commersonii, a Fish of three or four feet in length. According to Babuchin, Malopterurus also is said by the Nile fishermen to shelter its fry in the mouth.

Some of the Silurids attain to a very large size. Among these is the type of the family, Silurus glanis, the "Wels" of the Germans, its only European representative, which occurs over a great part of Europe, but is absent from the British Isles, France, the Spanish Peninsula, and Italy. It is most abundant in the Danube basin, where it sometimes reaches a length of 10 feet or more and a weight of 400 lbs. It is the largest strictly fresh-water Fish of Europe. Among the smallest species, we have to mention the "Candiru" of Brazil, Vandellia cirrhosa, 60 millimetres in length and 3 or 4 in diameter, which is believed to enter and ascend the urethra of people bathing, being attracted by the urine; the Fish, having once made its way into the urethra, cannot be pulled out again, owing to the erectile spines which arm its gill-covers. The natives of some parts of the Amazons are in great dread of this Fish, and protect themselves when entering the water by wearing a sheath formed of a small, minutely-perforated cocoanut-shell suspended from a belt of palm-fibres.[683] According {594}to Reinhardt[684] the allied Stegophilus insidiosus, a small colourless Fish, 30 to 40 mm. long, from Brazil and Argentine, lives parasitically in the gill-cavity of large Cat-Fishes (Platystoma). Dr. F. Silvestri has noticed that it sucks the blood in the gills of Platystoma coruscans, a Silurid growing to a length of 6 feet.

FAM. 5. LORICARIIDAE.—Distinguished from the preceding by the sessile ribs and the absence of the transverse processes in the praecaudal vertebrae, which have bifid neural spines. The air-bladder is always much reduced, and enclosed in a right and a left bony capsule formed by the skull and the anterior vertebrae. Gill-openings narrow clefts. The mouth is inferior, with more or less developed circular lips and feeble dentition; it is used as a sucker, by which the Fish fixes itself to any hard object with such strength that it cannot be pulled off without great difficulty. The teeth are usually slender and bicuspid. The food consists of very small prey and more or less putrefied organic substances, the intestine being usually extremely elongate and much convoluted. The habits of these Fish are very little known, but the fact that the males of many species have the pectoral fins much stronger than the females renders it probable that they pair like {595}Callichthys. There are other sexual differences in many species of Plecostomus, Chaetostomus, and Loricaria, as the presence of dermal tentacles on the snout (see Fig. 359), or of hair-like bristles on various parts of the head and fins in the males, which are usually of larger size.

About 200 species are known, all from the tropical and subtropical parts of Central and South America. The largest species (Chaetostomus gigas) measures 2½ feet; many are of very small size.[685] The genera may be referred to two sub-families:—

(i.) ARGINAE.—Body naked; ribs strong. Arges, Stygogenes, Astroblepus.

(ii.) LORICARIINAE.—Body cuirassed by bony plates; ribs very slender. Plecostomus, Liposarcus, Chaetostomus, Cochliodon, Pterygoplichthys, Rhinelepis, Acanthicus, Otocinclus, Hypoptopoma, Loricaria, Acestra.

The "Prenadillas" of the Andes, Arges and Stygogenes, were believed to live in subterranean waters within the bowels of active volcanoes, and to be ejected with streams of mud and water during eruptions, a story that has been repeated by Humboldt. The fact is that they live in small torrents at great altitudes (up to 10,700 feet), and are swept down during periods of disturbance caused by the eruption of the volcano.[686] The members of the sub-family Loricariinae vary much in the shape of the body, which may be short and stout, or more or less slender, the extreme in the latter respect being attained by the species of the genus Acestra.

{596}FAM. 6. ASPREDINIDAE.—This family is also closely related to the Siluridae. The ribs are sessile as in the Loricariidae, but inserted very low down on the centra, which higher up bear strong transverse processes. The opercular bone is entirely absent. The gill-opening is reduced to a foramen in front of the pectoral fin. The head is extremely depressed and the mouth terminal; the tail is very slender; the body is naked. The air-bladder is large and free, the intestinal canal short. Four genera from South America: Aspredo, Bunocephalus, Bunocephalichthys, Dysichthys. Species 18. Aspredo, of the Guianas, the largest form, reaching to about a foot in length, is remarkable for the manner in which the female carries her eggs. The skin of the lower parts assumes a spongy condition about the breeding season, and the eggs, after being deposited, become attached to the lower surface of the head, belly, and paired fins, forming a single layer; each egg becomes connected with the skin of the mother by a cup-shaped, pedunculate base, supplied with blood-vessels and coated with a layer of epithelium, the formation of which is still unexplained.[687]

{597}CHAPTER XXII

TELEOSTEI (CONTINUED): SYMBRANCHII—APODES—HAPLOMI—HETEROMI—CATOSTEOMI— PERCESOCES—ANACANTHINI

SUB-ORDER 3. SYMBRANCHII

Eel-shaped Fishes without paired fins, with the pectoral arch free or suspended from the skull, and with the anterior vertebrae distinct, without Weberian ossicles. Gill-openings confluent into a single, ventral slit. Air-bladder absent.

The structure of the skull conforms to that of typical Malacopterygians. The praemaxillary and maxillary are well developed, the latter placed behind the former, and forming but a very small part of the oral border; the symplectic is present; the parietals form a long sagittal suture, and separate the frontals from the supraoccipital. The vertebrae are very numerous, the praecaudal bearing very strong parapophyses, to which short, slender ribs are attached. The skin is naked (Symbranchidae) or covered with minute scales (Amphipnoidae), and the vertical fins are rudimentary, reduced to mere dermal folds.

Like the Apodes, which they resemble in general appearance, these Fishes are no doubt derived from some low type with abdominal ventral fins, but whether from the Malacopterygii or the Haplomi we have as yet no data from which to conclude. Only two families are known.

FAM. 1. SYMBRANCHIDAE.—Post-temporal well developed, forked, attached to the skull. Inhabitants of the fresh or brackish waters of South-Eastern Asia, Tropical America, Australia, and Tasmania. Three genera are known: Symbranchus, with two species from India and the Malay region, and one from Central and South America; Monopterus, with a single species {598}from China, Japan, and the Malay region; and Chilobranchus, with two species from Australia and Tasmania. Although the South American Symbranchus has been observed to live in marshes which periodically dry up, the Fish burying itself in the mud like a Lepidosiren, the branchiae are fully developed on the four branchial arches. In Monopterus, of similar habits, the branchial laminae are rudimentary, and on three arches only. No accessory breathing organ is known to exist.

FAM. 2. AMPHIPNOIDAE.—Post-temporal absent, the shoulder-girdle free from the skull. The Cuchia, Amphipnous cuchia, the sole representative of this family, an inhabitant of the fresh and brackish waters of India and Burma, growing to two feet in length, is remarkable for the presence of a respiratory air-sac on each side of the neck behind the head, communicating with the gill-cavity. Of the three branchial arches the second alone possesses gill-filaments; the third supports, in their place, a thick and semi-transparent tissue; the principal organs of respiration are two small bladders, resembling the posterior portions of the lungs of snakes, which the animal has the power {599}of filling with air immediately derived from the atmosphere. Although covered over by the common integuments, these bladders present externally, when inflated, two protuberances of a round shape. Of the whole volume of blood contained in the branchial artery, one-third passes through the gills and respiratory bladders, whilst the other two-thirds are conveyed directly from the heart to the aorta without being exposed to the action of the air.[688] This amphibious Fish, when in the water, constantly rises to the surface for the purpose of respiration, and it is often found lying in the grassy sides of ponds after the manner of Snakes.

SUB-ORDER 4. APODES.

Air-bladder, if present, communicating with the digestive tract by a duct. Praemaxillaries absent; the maxillaries, if present, separated on the median line by the coalesced ethmoid and vomer. Pectoral arch, if present, not connected with and remote from the skull; mesocoracoid arch absent. Fins without spines, the ventrals absent. Anterior vertebrae distinct, without Weberian ossicles.

The Apodes or Eels are elongate, serpentiform Fishes with naked skin, or with minute scales imbedded in the skin, the opercular bones small and completely hidden under the integument; narrow or minute gill-openings; the vertical fins, if present, confluent behind or separated by the projecting tip of the tail. The pterygo-palatine arch is often reduced or absent, and there is no distinct symplectic; the supraoccipital bone is small, separated from the frontals by the parietals, which meet on the middle line. The vertebrae are very numerous (up to 225), and the praecaudals bear strong parapophyses, to which short, slender ribs may be attached; epineurals are sometimes present. The five families into which this sub-order is divided show remarkable degrees of simplification of the skull, through reduction or loss of either the maxillary or the pterygo-palatine arches.

There has been much difference of opinion in the determination of the bones of the upper jaw in these Fishes. Cuvier regarded the lateral bones of the upper jaw as praemaxillaries, Owen and Richardson as palatines (at least in Muraenas), whilst Peters {600}and most recent authors have identified them throughout the order as maxillaries.[689] The conclusion I have come to from the examination of numerous skulls belonging to various genera, is that the praemaxillaries have disappeared in all, whilst the maxillaries have persisted in the true Eels (Anguillidae) and disappeared in the Muraenidae, their place being taken by the fused palato-ectopterygoids, which may even join the mandibular suspensorium. The vestigial bone, regarded by Jacoby as the pterygoid in Muraena helena, may be identified as the meta-pterygoid, and therefore does not disprove the homology, here suggested, of the other elements of the palate.

FAM. 1. ANGUILLIDAE.—Maxillaries present, separated on the median line by the ethmo-vomer; palato-pterygoid bone present, connected with the hyomandibular and quadrate; gill-clefts separate, opening into the pharynx by wide slits; tongue present; vent far removed from the head.

Spread over all the seas of the temperate and tropical zones, often descending to the greatest depths, a few entering fresh waters. Many are known to undergo very striking metamorphoses, the pellucid, strongly compressed larvae (Leptocephalus) having long been a puzzle to naturalists.

{601}Nearly 150 recent species are known, of which some 50 are deep-sea forms, occurring down to 2500 fathoms. Scanty fossil remains, referable to recent genera or scarcely different from them, are known from the Eocene of Europe. The Cretaceous genus Urenchelys, from England and the Lebanon, is interesting as representing a more generalised type, the hindmost vertebrae bearing a pair of expanded hypural bones, showing the diphycercal Eels to have been derived from Fishes with a normal caudal fin.

The genera are numerous. The following are the principal:—Anguilla, Simenchelys, Ilyophis, Conger, Coloconger, Congromuraena, Uroconger, Heteroconger, Muraenesox, Nettastoma, Nettophichthys, Saurenchelys, Nettenchelys, Myrus, Myrophis, Derichthys, Chilorhinus, Muraenichthys, Liuranus, Ophichthys, Moringua.

In the first four genera, small, more or less lineal rudimentary scales are embedded in the skin, arranged in small groups, which are placed obliquely at right angles to one another, forming a curious pattern; but these scales are so small that they escape the notice of the superficial observer, hence Eels have been improperly included among the Fishes forbidden as food by the Mosaic prescriptions. In the other genera, including the exclusively marine Conger of our coasts, scales are really absent.

The Common Eel (Anguilla vulgaris) has a very wide distribution, being found over the greater part of Europe, North Africa, Temperate Asia, and perhaps also North America east of the Rocky Mountains, Mexico, and the West Indies. Its record from Australia and New Zealand is probably due to the imperfection of our knowledge of the specific characters. It is not found in the Black Sea nor in the rivers flowing into it, owing, no doubt, to the sulphurous nature of the bottom of the sea, to which, as we now know, these Fish would have to resort for breeding.

The mode of propagation of the Eel long remained a mystery, from the fact that individuals found in fresh water never show ripe genital glands. The idea had been entertained of their being hermaphrodite, and internal parasites had also given rise to the belief in their viviparous nature. The genital glands of the female were first investigated by Rathke in 1838, but it was not until 1874 that those of the male were discovered by Syrski, and shortly after fully described by L. Jacoby, who, in his final contribution to the subject, concluded that Eels need salt water for the development of their organs of generation, and that this development takes place, not near the coast, but further out in deep water.

{602}[Illustration: FIG. 363.—Larva of Common Eel, Leptocephalus brevirostris of Kaup. (After Kaup.)]

As a rule it is not until the fifth or sixth year that the Eels go to the sea for the purpose of propagation, which takes place at great depths—at least 200 fathoms. Males have been observed to precede the females. The breeding season over, the Eels do not return to fresh waters, but are believed to die soon after. The eggs were discovered by Raffaele in 1888 in the Gulf of Naples, and shortly after Grassi and Calandruccio finally settled the question of the breeding and development of the Fish from observations made in the Mediterranean. Their conclusions are thus summed up:—"The Common Eel matures in the depths of the sea, where it acquires larger eyes than are ever observed in individuals which have not yet migrated to deep water. The abysses of the sea are its spawning places; its eggs float in the sea water. In developing from the egg, it undergoes a metamorphosis, it passes through a larval form denominated Leptocephalus brevirostris." What length of time the development requires is not yet fully established, since the Leptocephali are rarely found at the surface, most of the specimens studied by Grassi and Calandruccio having been obtained from the stomach of the Sun-Fish (Orthagoriscus mola) in the Straits of Messina; but it is believed that the young Eels or "elvers," which ascend our rivers in such prodigious numbers in spring and summer ("Eel-Fares") are already one year old. Some individuals apparently spend their whole life in fresh waters, but they are barren.[690] A specimen was kept in confinement in the family of the French naturalist Desmarest for upwards of 40 years, growing to a length of 4½ feet, being already of large size at the time of {603}its capture. Eels are extremely voracious, and endowed with an extraordinary tenacity of life; they can live for many hours out of the water, and are often met with at night creeping through the grass of meadows from one pond or stream to another.

One of the most remarkable among the deep-sea Eels is the Snub-nosed Eel (Simenchelys parasiticus), which has been found in great numbers off Newfoundland and the Azores, at depths of 200 to 900 fathoms. The maxillary and mandibular bones are very short and massive, provided with large obtuse teeth; the head is short and bulldog-like in aspect, the mouth small and bordered by a thick circular lip. Some specimens have been observed to burrow in the muscles of living Halibut and other large Fishes, after the manner of Myxine.

FAM. 2. NEMICHTHYIDAE.—Distinguished from the preceding by the position of the vent, which is close to, or at no great distance from, the gill-openings. The rays of the vertical fins are connected by thin membrane instead of being imbedded in thick skin, as in most Eels; in some of the genera the jaws are excessively prolonged, needle-like, sometimes recurved. Deep-sea Eels of small size, represented in the Atlantic, Pacific, and Indian Oceans by about 10 species, referred to 6 genera: Dysomma, Dysommatopsis, Nemichthys, Spinivomer, Serrivomer, Gavialiceps.

FAM. 3. SYNAPHOBRANCHIDAE.—Maxillaries narrowly separated on the median line, their extremity strongly attached by ligament to the mandible; pterygo-palatine arch absent. Gill-openings externally confluent into a single ventral slit. Deep-sea Fishes, resembling the true Eels in the general form and in the presence of linear scales placed at right angles, but differing in the absence of the pterygo-palatine arch, as in the Saccopharyngidae. Eight species of Synaphobranchus are known, from the Atlantic, Pacific, and Indian Oceans, at depths of 200 to 2000 fathoms.

FAM. 4. SACCOPHARYNGIDAE.—Maxillaries narrowly separated on the median line, extremely elongate; mouth enormous; pterygo-palatine arch absent; hyomandibular arch slender and movably articulated to the cranium, the two bones (hyomandibular and quadrate) of which it is composed being capable of being swung in all directions; branchial arches far behind the skull; no branchiostegal rays or pharyngeal bones.

Extraordinary-looking deep-sea Fishes allied to the Eels, of which they appear to be a further degraded type, the muscles {604}being feebly developed and the skeleton imperfectly calcified. The mouth, furnished with rather long but feeble, or even minute teeth, and the pharynx and stomach are capable of great distension, these Fish being able to get outside a prey very much larger than themselves; the eyes are situated far forward on the head; the tail is extremely slender and elongate. Four genera are known, each with a single species, from the Atlantic: Saccopharynx, Eurypharynx, Macropharynx, and Gastrostomus. The depths at which they have been obtained vary between 389 and 1467 fathoms, but three out of the four known specimens of Saccopharynx were brought to the surface by having swallowed a Fish too large for the capacity of the stomach. The length of the largest specimen is about 6 feet, of which the tail constitutes nearly three-fourths.

FAM. 5. MURAENIDAE.—Maxillaries absent, replaced by the palato-pterygoid, the mouth bordered by the latter and the {605}ethmo-vomer; palato-pterygoid bone separated from hyomandibular arch; branchial openings into the pharynx narrow slits; no tongue.

The body is naked, pectoral fins are usually absent, and the gill-cleft is a small round opening. The opercular bones are much reduced in size, and the pectoral arch may be totally absent.

Voracious marine Fishes, inhabiting tropical and subtropical waters, being especially abundant about coral reefs. Some 120 species are known, many reaching a very large size, and being also remarkable for their variegated coloration. The genera are mostly founded on the dentition, which shows much diversity; the following are the principal:—Myroconger, Enchelycore, Muraena, Thyrsoidea, Lycodontis, Pythonichthys, Echidna, Channomuraena. The Muraena of the ancient Romans, Muraena helena, of the Mediterranean, Eastern Atlantic, and neighbouring parts of the Indian Ocean, occurring exceptionally as far north as the English coast, grows to 4 feet, and its flesh was more esteemed than that of any other Fish, large numbers being reared in specially constructed reservoirs near the sea, and fed on the corpses of slaves. Channomuraena vittata, from the coast of Cuba, is known to attain a length of 8 feet, and Thyrsoidea macrura, from the Indian Ocean and the Malay Archipelago, to upwards of 10 feet.

SUB-ORDER 5. HAPLOMI.

Air-bladder, if present, communicating with the digestive tract by a duct. Opercle well developed. Pectoral arch suspended from the skull; no mesocoracoid arch. Fins usually without, rarely with a few spines; ventrals abdominal, if present. Anterior vertebrae distinct, without Weberian ossicles.

The absence of the mesocoracoid arch distinguishes the Haplomi from the Malacopterygii, with which they are united by various authors. They lead to the Percesoces through the Cyprinodontids, and to the Lower Acanthopterygians, such as the Berycidae, through the Scopelids, Stephanoberycids, and Percopsids, as is evidenced by the structure of the mouth and the forward position, in some of the genera, of the ventral fins, which, however, are never attached to the pectoral girdle. Most of the forms which are here included inhabit either fresh water or the deep sea.

{606}SYNOPSIS OF THE FAMILIES.

I. Parietals separating the frontals from the supraoccipital; post- temporal simple; praecaudal vertebrae with autogenous parapophyses.

Margin of the upper jaw formed by the praemaxillaries and the maxillaries; basis cranii simple; no adipose dorsal fin 1. Galaxiidae.

Margin of the upper jaw formed by the praemaxillaries only; basis cranii double; adipose dorsal fin present 2. Haplochitonidae.

II. Frontals in contact with the supraoccipital.

A. Praecaudal vertebrae without parapophyses.

1. Margin of the upper jaw formed by the praemaxillaries and the maxillaries.

Body without or with minute scales, usually with rows of scutes; adipose dorsal fin usually present 3. Enchodontidae.†

Body scaly; post-temporal forked; no adipose dorsal fin; ventrals with 6 to 11 rays 4. Esocidae.

Body scaly; post-temporal incompletely ossified; pectoral fin without pterygials; no adipose dorsal fin; ventrals with 3 rays only 5. Dalliidae.

2. Maxillaries excluded from the oral border.

a. Adipose dorsal fin usually present; ventral fin with 7 to 10 rays.

Post-temporal forked; dorsal fin formed of articulated rays 6. Scopelidae.

Post-temporal simple; dorsal fin very long, formed of slender, non-articulated, simple or bifid rays 7. Alepidosauridae.

b. No adipose dorsal fin; head and mouth enormous, dentition feeble; body naked; ventral fins, if present, with 5 rays 8. Cetomimidae.

B. Praecaudal vertebrae with well-developed parapophyses; maxillaries excluded from the oral border.

1. Dorsal and anal fins without spines; scales cycloid, or with erect spines; no adipose fin.

Mouth not protractile; ventral fins far forward, with 7 to 17 rays 9. Chirothricidae.†

Mouth not protractile; ventral fins remote from the pectorals, with 9 rays 10. Kneriidae.

Mouth protractile; ventral fins, if present, with 5 to 7 rays 11. Cyprinodontidae.

Mouth scarcely protractile; ventral fins rudimentary or absent; vent close to the gills 12. Amblyopsidae.

Mouth slightly protractile; ventral fins with 5 or 6 rays 13. Stephanoberycidae.

2. Dorsal and anal fins with true spines; scales ctenoid; an adipose dorsal; ventral fins with 9 rays 14. Percopsidae.

† Fossil only.

{607}FAM. 1. GALAXIIDAE.—Margin of the upper jaw formed by the praemaxillaries and the maxillaries, the latter behind the former, and toothless. Parietals in contact with each other, and separating the frontals from the supraoccipital; opercular bones all well developed. Basis cranii simple. Ribs inserted on strong, autogenous parapophyses; epipleurals and epineurals. Post-temporal simple, attached to the epiotic; post-clavicle present. Body naked. Vertical fins far back; no adipose dorsal fin. Pectoral fins inserted very low down. Ventrals, if present, with seven rays. Air-bladder present. Ova falling into the cavity of the abdomen before exclusion.

The genus Galaxias has an interesting distribution, the species of which it is made up occurring in the fresh waters of the southern hemisphere, viz. 8 in New Zealand and neighbouring islands, 7 in New South Wales, 3 or 4 in South Australia, 1 in West Australia, 2 in Tasmania, 7 in South America, from Chili southwards, and 1 at the Cape of Good Hope. One species (G. attenuatus) is even believed to be identical in New Zealand, Tasmania, South Australia, the Falkland Islands, and South America. This conclusion is probably correct from the fact, which may account for the distribution of the whole genus, {608}that it is not confined to fresh waters, but occurs also in the sea. Specimens were observed by Mr. Rupert Vallentin in the Falkland Islands, where the Fish is known to the inhabitants as "Smelts," in shoals in the shallow water along the shore; and, according to Mr. F. E. Clarke, the same species, in New Zealand, periodically descends to the sea, where it spawns, from January to March, and returns from March to May. A marine species has recently been discovered at the Chatham Islands. In New Zealand, the Galaxias were called "Trout" by the settlers before the introduction of Salmonids, whilst the fry of G. attenuatus are eaten as "Whitebait." The largest species reach the length of a foot. Neochanna, from New Zealand, differs from Galaxias in the absence of ventral fins; it has been found in burrows, which it excavates at a distance from water.

FAM. 2. HAPLOCHITONIDAE.—Small fresh-water Trout-like Fishes, agreeing in most respects with the Galaxiidae, to which they are unquestionably closely related, differing only in the greater extent of the praemaxillaries, which exclude the maxillary from the oral border, in the double basis cranii (the prootics uniting under the brain, leaving a canal between them and the parasphenoid), in the shorter parapophyses, which, like the neural arches of the praecaudal vertebrae, are autogenous, and in the presence of a small adipose dorsal fin, opposed to the anal.

Two genera: Haplochiton, naked, with a single species from Chili, the southern extremity of South America, and the Falkland Islands, and Prototroctes, covered with small scales, of which one species inhabits Queensland, another South Australia, and a third New Zealand. In Haplochiton, the urogenital orifice of both sexes is produced into a cylindrical tube, which lies concealed in a groove in front of the anal fin.

FAM. 3. ENCHODONTIDAE.—Margin of the upper jaw formed by the praemaxillaries and the maxillaries, the latter sometimes toothed like the former. Frontals in contact with the supraoccipital; basis cranii simple. Ribs sessile; praecaudal vertebrae without transverse processes. Rayed dorsal fin never much extended; sometimes an adipose fin behind it. Scales delicate or absent, but occasional longitudinal series of scutes occur, the dorsal series, when present, being unpaired.

Cretaceous Fishes allied to, and apparently more generalised than, the Esocidae and Scopelidae. Numerous remains from {609}Europe and North America, referred to 7 genera: Enchodus, Eurypholis, Palaeolycus, Halec, Cimolichthys, Prionolepis.

FAM. 4. ESOCIDAE.—Margin of the upper jaw formed by the praemaxillaries and maxillaries, the latter behind the former, and toothless. Supraoccipital in contact with the frontals, separating the small parietals; opercular bones all well developed; basis cranii simple. No parapophyses, except to the hindermost praecaudal vertebrae; epipleurals and epineurals. Post-temporal forked, the upper branch attached to the epiotic, the lower to the exoccipital; post-clavicle present. Vertical fins far back; no adipose dorsal fin. Pectoral fins inserted very low down; ventrals with 6 to 11 rays. Air-bladder present.

As in the Haplochitonidae, the neural and haemal arches are bones distinct from the centra, and although parapophyses are not developed, the ribs are not inserted on the centra, but on distinct bases wedged into the latter.[691] Teeth are present on the vomer, palatine, and pharyngeal bones.

A small family of carnivorous freshwater Fishes, including the Pike (Esox), of predaceous habits, unsurpassed in greediness and voracity, and the small and insignificant-looking Umbra, distinguished by the more anterior position of the dorsal fin, the larger scales, and the moderately large gape, with feeble villiform teeth. The range of the Esocidae is restricted to the cold and temperate parts of the northern hemisphere. Besides the well-known Esox lucius of Europe, Northern Asia, and the northern parts of North America, growing to a length of 4 feet, and the Maskinongy (E. nobilior) of north-eastern North America, reaching twice that length, the first genus comprises three smaller species from the Eastern United States. Remains of Esox have been found in {610}various freshwater deposits in Europe as far back as the Oligocene. E. lepidotus, of which very perfect specimens have been found in the Upper Miocene beds of Oeningen in Baden, differs from the living species in its much larger scales and in the greater approximation of the ventral and anal fins, two characters in which it approaches Umbra. Only two species of the latter are known: U. crameri ("Hundsfisch"), from the stagnant waters of Austria-Hungary, and U. limi ("Mud-Fish"), living in swamps and ditches in Canada and the north-eastern United States, often remaining imbedded in the mud of prairie sloughs and bog-holes.

FAM. 5. DALLIIDAE.—The genus Dallia, with a single species inhabiting the streams and ponds of Alaska and Siberia, is related to Umbra, but differs in the very thin and papery skeleton, with the post-temporal imperfectly ossified and the pectoral fin without pterygials or actinosts. The dorsal fin is far back and opposite to the anal, as in the Pike. The ventral fins are composed of three rays only, and the pectorals, which have a somewhat fleshy base, have as many as 36. The scales are extremely small, and partly imbedded in the skin. The Black-Fish, D. pectoralis, abounds in Sphagnum ponds, feeding on plants and worms, and forming the chief food of the natives of some parts of Northern Alaska, where, with the exception of the Salmonids, it is the only freshwater Fish. Turner, its discoverer, says its vitality is {611}extraordinary: Black-Fishes will remain frozen in baskets for weeks, and when thawed are as lively as ever, one having been swallowed in a congealed condition by a dog, thawed out by the heat of the stomach, and vomited up alive.

FAM. 6. SCOPELIDAE.—Praemaxillaries much elongate, and completely excluding the maxillaries from the oral border. Supraoccipital in contact with the frontals, sometimes partly covered by the parietals; opercular bones all well developed. Basis cranii simple. Ribs sessile; no parapophyses on the praecaudal vertebrae; epipleurals and epineurals. Post-temporal forked, the upper branch in contact with the epiotic or the supraoccipital, the lower with the opisthotic; post-clavicle present. An adipose dorsal fin often occurs; luminous spots often present on head and body. Ventral fins with 7 to 10 rays. Air-bladder sometimes absent.

A large family (over 100 known living species), mostly of pelagic and deep-sea Fishes. A great number of fossil types have been described.

Recent genera:

A. Without photophores: Saurus, Saurida, Bathysaurus, Harpodon, Scopelarchus, Aulopus, Odontostomus, Omosudis, Sudis, Paralepis, Bathypterois, Benthosaurus, Chlorophthalmus, Ipnops. B. With photophores: Scopelus, Dasyscopelus, Neoscopelus, Scopelengys, Nannobrachium, Scopelosaurus.

Fossil genera:

A. Cretaceous: Sardinioides, Acrognathus, Leptosomus, Sardinius, Dactylopogon, Nematonotus, Microcoelia, Opisthopteryx, Apateodus, Rhinellus. B. Eocene, Oligocene, and Miocene: Omiodon, Scopeloides, Parascopelus, Anapterus.

{612}[Illustration: FIG. 371.—A, Scopelus crocodilus (after Goode and Bean). B, Bathypterois dubius (after Collett). C, Ipnops murrayi, with dorsal view of head (after Goode and Bean).]

{613}The members of this Family vary much in form, and among them are to be found some of the most curious adaptations to bathybial existence. One of the best known is Harpodon nehereus, which, when newly taken, is brilliantly phosphorescent all over the body; in a salted and dry condition it is the "Bombay-duck," a delicacy eaten with curries, and exported in large quantities from the west coast of India. It is not known to occur at any great depth, and is not even restricted to the sea, being very abundant in the rivers and estuaries of Bengal and Burma; whilst an allied species, H. squamosus, is found in the Indian Ocean at depths of 120 to 300 fathoms. In Bathypterois, the eyes are very small; some of the rays of the paired fins being excessively prolonged, acting as tactile organs, and compensating the reduction in the eyes. Sir John Murray has observed about B. longipes: "When taken from the trawl [from 2650 fathoms] they were always dead, and the long pectoral rays were erected like an arch over the head, requiring considerable pressure to make them lie along the side of the body; when erected they resembled Pennatulids like Umbellula." In Ipnops, which resembles in general form the large-eyed Chlorophthalmus gracilis, the upper surface of the broad spatulate snout is occupied by a luminous organ longitudinally divided into two symmetrical halves, and the eyes are absent, unless, as first supposed, this extraordinary organ be a modification of them; but Professor Moseley's examination seems to have proved beyond doubt that it is a special form of phosphorescent organ, the object of which would be to attract other creatures to the wide gape of a Fish which, living in the abysses of the sea and deprived of organs of sight and touch, would have great difficulty in procuring its food. Odontostomus, with a very large eye which can be turned upwards and sidewards, and enormous compressed curved teeth, barbed at the tip and depressible backwards, is one of the few Scopelids in which scales are completely absent.

The numerous species (about 50) of Scopelus and their allies are moderate-sized or small pelagic and deep-sea forms found in nearly all the seas, some coming to the surface at night, whilst others are confined to great depths; they are remarkable for the series of phosphorescent spots (photophores) on the body, and in some species also on the head, where they may form large patches on the snout. The arrangement of these photophores is a very {614}definite one, and it has been used for the division of these Fishes into genera or sub-genera.[692] The ventral fins have a more forward position than in most other members of the Family.

FAM. 7. ALEPIDOSAURIDAE.—Characters as in the preceding, but supratemporal simple, attached to the opisthotic, and dorsal fin very long, formed of slender, non-articulated, simple or bifid rays, extending along nearly the whole length of the back, followed by a small adipose fin. The air-bladder is absent and the body scaleless. The skeleton is feebly ossified; the dentition is very powerful, some of the teeth on the palate and mandible being very strongly enlarged. 4 or 5 species are known, from considerable depths in the Atlantic and Pacific Oceans, referable to one genus, Alepidosaurus or Plagyodus. A. ferox, from the Atlantic, reaches a length of 4 feet.

FAM. 8. CETOMIMIDAE.—The affinities of the recently discovered genera Rondeletia and Cetomimus, deep-sea Fishes from the North Atlantic, at depths of 1000 to 1600 fathoms, are still uncertain, as the skeleton could not be examined; they are probably most nearly related to the Scopelidae. The head is enormous, with very wide gape, that of Cetomimus being suggestive of that of a Right Whale; the teeth are small and coarsely granular; the gill-openings are very wide; the body is more or less compressed and scaleless; the dorsal and anal fins are opposed to each other; no adipose dorsal fin. In Rondeletia, the eyes are moderately large, and ventral fins, with 5 rays, are present; in Cetomimus, the eyes are very small, and ventral fins are absent.

{615}FAM. 9. CHIROTHRICIDAE.—Praemaxillaries delicate and styliform, completely excluding the maxillaries from the upper border of the mouth; jaws with feeble dentition or toothless; opercular apparatus complete. Praecaudal vertebrae with robust parapophyses, to which ribs are attached. Ventral fins far forwards.

These Fishes, of which three fossil genera are known from the Cretaceous of Germany and Syria, appear to be related to the Scopelidae, from which the strong parapophyses distinguish them. Chirothrix is remarkable for its excessively enlarged ventral fins with about 17 rays; these fins were taken for the pectorals by the early describers. In Telepholis and Exocoetoides, the ventral fins are smaller than the pectorals, and formed of 7 or 8 rays only; the dorsal region, in the former, is protected by a covering of small, thin, rounded or polygonal dermal scutes, each bearing a median tubercle.

FAM. 10. KNERIIDAE.—Margin of the upper jaw formed by the praemaxillaries; mouth toothless, not protractile. Parietals separated by the supraoccipital. Pharyngeal bones toothless. Praecaudal vertebrae with parapophyses. Body covered with small scales. Ventrals with 9 rays. No adipose dorsal fin. Air-bladder present.

{616}The genus Kneria comprises two species from the fresh waters of tropical Africa, one from Angola, the other from East Africa.[693] Small Loach-like Fishes, two to four inches long, with the upper jaw projecting beyond the mouth, which is inferior and transverse; no barbels; gill-membranes entirely grown to the isthmus, the gill-opening being a rather narrow vertical slit; dorsal and anal fins short, the former opposite, or nearly opposite, to the ventrals; the snout of the male(?) of K. angolensis is described as beset with small spine-like excrescences; the intestinal tract makes several convolutions.

FAM. 11. CYPRINODONTIDAE.—Mouth protractile, the maxillaries excluded from the oral border; teeth in the jaws and on the pharyngeal bones; pterygo-palatine arch weak or rudimentary; opercular bones all well developed. Basis cranii simple. Praecaudal vertebrae with strong parapophyses, bearing the ribs; epipleurals inserted on the ribs. Post-temporal forked. Ventrals, if present, with 5 to 7 rays. No adipose dorsal fin. Air-bladder sometimes absent.

From a physiological point of view, this Family may be divided into carnivorous forms, with short digestive tract, and phytophagous or limnophagous ones, in which the intestine forms numerous coils. To the first division belong the living genera Cyprinodon, Characodon, Tellia, Haplochilus, Fundulus, Rivulus, Cynolebias, Orestias, Empetrichthys, Jenynsia, Pseudoxiphophorus, Belonesox, Gambusia, Anableps, among existing forms, and the fossil genera Prolebias (Oligocene and Miocene) and Pachylebias (Miocene); to the second, the living genera Poecilia, Mollienesia, Platypoecilus, and Girardinus. Procatopus, a near ally of Haplochilus, recently discovered in South Cameroon, is remarkable for having the ventral fins inserted far forward, below the pectoral fins.

These are small or very small Fishes,[694] only a few reaching a length of about a foot, confined to fresh or brackish waters, recognisable externally by the flat head with protractile mouth, the usually large scales, and the absence of a well-developed lateral {617}line. The teeth vary much in shape: cardiform, villiform, or compressed, and bi- or tri-cuspid; the palate is either toothless, or teeth are present on the vomer. About 200 species are known, mostly from the American continent, only about 30 being known from other parts of the world, viz. Southern Europe, Southern Asia and Japan, and Africa. In many species the sexes are dissimilar, the female being larger and less brilliantly coloured, with smaller fins; the anal fin of the male may be modified into an intromittent organ by means of which internal fertilisation takes place, the ova developing in a sort of uterus, which the young leave in a more or less advanced stage of growth. The most curious of the Cyprinodontids is the genus Anableps, of Central and South America, surface-swimming Fishes, the strongly projecting eyes of which are divided by a horizontal band of the conjunctiva into an upper part adapted for vision in the air, and a lower for vision in the water, and the pupil is divided into two parts by a constriction; the larger species grows to the length of a foot.

{618}FAM. 12. AMBLYOPSIDAE.—Mouth scarcely protractile, the maxillaries excluded from the oral border; teeth small, in jaws and palate, and on the pharyngeal bones. Praecaudal vertebrae with very strong parapophyses, bearing the ribs on their upper surface; epipleurals inserted on the ribs. Ventral fins rudimentary or absent. Vent jugular, close to the gill-clefts. Air-bladder present.

Small ovoviviparous Fishes, closely related to, and evidently derived from, the Cyprinodontids, measuring from 1 to 5 inches, inhabiting ditches and small streams, or confined to subterranean waters of limestone caves, in the United States east of the Rocky Mountains. Six species, referable to three genera, are known. In Chologaster, the eyes are well developed and the body is coloured. C. cornutus inhabits the lowland streams and swamps of the South Atlantic States, from Virginia to Florida; C. agassizii is found in the underground streams of Kentucky and Tennessee; and C. papilliferus occurs under stones in the springs of south-western Illinois. Amblyopsis and Typhlichthys, which are evidently derived from the former, or from forms closely related to it, have the eyes rudimentary and more or less concealed under the skin, and the body is colourless. Amblyopsis spelaea is widely distributed in the caves east of the Mississippi, {619}both north and south of the Ohio River; it is common in the River Styx of the Mammoth Cave. Typhlichthys subterraneus is found with the latter species in the caves east of the Mississippi, but is confined to the south side of the Ohio River, whilst T. (Troglichthys) rosae is found in the caves west of the Mississippi River. Of Amblyopsis spelaea, the late Professor Cope has observed: "If these Amblyopses be not alarmed, they come to the surface to feed, and swim in full sight, like white aquatic ghosts. They are then easily taken by the hand or net, if perfect silence is preserved, for they are unconscious of the presence of an enemy except through the medium of hearing; this sense, however, is evidently very acute, for at any noise they turn suddenly downwards, and hide beneath stones, etc., on the bottom." Dr. Garman thinks, on the contrary, that such a sense can hardly be developed in recesses where we are accustomed to think any sounds other than those made by the rippling or dripping water are almost unknown, and that it is through the sense of touch, and not through hearing, that the Fish is disturbed. In fact, the head is provided with a great number of tactile papillae, arranged in transverse ridges, provided with nervous filaments, which evidently compensate the loss of the visual organ.[695]

FAM. 13. STEPHANOBERYCIDAE.—This Family has hitherto been placed near the Berycidae, among the Acanthopterygii, but there are no spinous rays in the dorsal and anal fins; and the ventrals, formed of one simple and four or five branched rays, are abdominal. The genus Stephanoberyx, with two species from the Atlantic, at depths of 535 to 2949 fathoms, is characterised by a large, thick, cavernous head, with thin bony spine-bearing ridges, a large mouth bordered by the protractile praemaxillaries, behind which are the large maxillaries, a short dorsal and a short anal, opposed to each other behind the ventrals, and the body covered with feebly imbricated scales, each bearing in the centre one or several erect spines. The largest specimen measures 6 inches. Malacosarcus, a small Fish from the Pacific, at depths of 2350 and 2425 fathoms, is very closely allied to Stephanoberyx, but its scales are very thin and cycloid. The striking resemblance which the head {620}bears to that of the Berycid Melamphaes may be merely a case of convergence, and it must be borne in mind that this appearance is approached by some species of Scopelus, with which both Malacosarcus and Melamphaes were originally confounded. The praecaudal vertebrae are provided with parapophyses. I have ascertained on a specimen of Stephanoberyx monae that the air-bladder is connected with the dorsal side of the stomach by a short and comparatively wide duct.

FAM. 14. PERCOPSIDAE.—Margin of the upper jaw formed by the praemaxillaries; mouth small, not protractile, toothed; palate toothless. Supraoccipital in contact with the frontals, separating the small parietals. Basis cranii simple. Most of the praecaudal vertebrae with parapophyses, on the upper surface of which the ribs are inserted; no epipleurals. Post-temporal forked; post-clavicle present; scapular foramen in the scapula, on which three hour-glass-shaped pterygials are inserted, a fourth being inserted on the coracoid. Dorsal fin with two true spines; anal with one or two; ventrals far forward, with 9 rays; pectorals inserted rather high. A small adipose dorsal fin. Body covered with strongly ctenoid scales. Air-bladder present (with open duct).

This is a most interesting group of Fishes, from the resemblance which they bear to the Perches, and they have therefore been raised to the rank of a sub-order, Salmopercae, by Jordan and Evermann, who regard them as "archaic fishes, relics of some earlier fauna, and apparently derived directly from the extinct transitional forms through which the Haplomi and Acanthopteri have descended from allies of the Isospondyli [Malacopterygii]." On the other hand, an analysis of their characters shows them to belong to the Haplomi, of which they {621}may be regarded as highly specialised members, having evolved in the direction of the Acanthopterygii.

Only two genera are known, each with a single species: Percopsis, from the rivers and streams of Canada and the north-eastern United States, and Columbia, more recently discovered in the sandy or weedy lagoons along the Columbia River. These Fishes are of small size, not exceeding 6 inches in length. Their eggs are unusually large.

SUB-ORDER 6. HETEROMI.

Air-bladder without open duct. Opercle well developed; parietal bones separating the frontals from the supraoccipital. Pectoral arch suspended from the supraoccipital or the epiotic, the post-temporal small and simple or replaced by a ligament; no mesocoracoid. Ventral fins abdominal, if present.

The Halosauridae and Notacanthidae are deep-sea Fishes of obscure affinities. In the abdominal position of the many-rayed ventral fins and in the absence of the mesocoracoid arch they agree with the Haplomi; but if, as the investigations of Günther lead us to believe,[696] there is really no open communication between the air-bladder and the digestive tract, they {622}should be removed from this physostomous sub-order. The two families have many characters in common, such as the attachment and structure of the pectoral arch, which is devoid of a post-clavicle, the position of the pectoral fins high up the sides, the strong parapophyses inserted very low down on the centra of the vertebrae, the extent of the parietal bones, which meet in a sagittal suture and separate the frontals from the supraoccipital. The recent discovery of a third family, the Lipogenyidae, which, in the structure of the dorsal fin, is exactly intermediate between the two others, has lessened the gap between the Lyomeri (Halosauridae) and Heteromi (Notacanthidae) of Gill, which I have proposed to unite in a suborder under the latter name.

These Fishes are no doubt derived from forms in which a separate caudal fin existed; such a type must have been near the Dercetidae, as defined by A. S. Woodward, which may provisionally be placed here.

An imperfectly known Fish from the Chalk of Mount Lebanon, Pronotacanthus sahelalmae, appears to bear some affinity to Notacanthus, and has been placed in the same family; but its characters are not sufficiently defined to refer it without doubt to this division.

There is a fifth family which may enter this sub-order: the Fierasferidae, the structure of which has been exquisitely described and figured by Emery.[697] Hitherto placed with or near the Ophidiidae, they differ widely from them, as well as from all Acanthopterygians, in the conformation of the skull, the supraoccipital being separated from the frontals by the parietals, which form a long median suture. This is a feature which has only been observed in Fishes with abdominal ventral fins, and although the total absence of those fins in Fierasfer deprive us of an important criterion in deciding on its affinities, I am inclined to regard this family as derived from an "abdominal" type. The conformation of the pectoral arch has much in common with that of the Halosaurs, and, notwithstanding the interpretation that has been given to the bones at the back of {623}the cranium in the latter type, the same may be said, in a general way, of the skull.

As pointed out by Emery, the very anterior position of the vent in the Fierasferidae is directly related to the curious mode of life of these Fishes, and the analogous condition obtained in various families, such as the Gymnotidae, Nemichthyidae, Amblyopsidae, shows it to be a character of relatively small systematic importance.

SYNOPSIS OF THE FAMILIES.

A. Vent posterior.

a. A distinct caudal fin; ordinary scales small or wanting, but enlarged scutes along the side 1. Dercetidae.†

b. Tail tapering to a point; scales cycloid.

No spines; dorsal fin short, anal very long 2. Halosauridae.

Fins with spines, dorsal short, anal long 3. Lipogenyidae.

Dorsal fin formed of a series of spines, anal long, formed partly of spines and partly of soft rays 4. Notacanthidae.

B. Vent immediately behind the gill-opening; no caudal fin; scales absent 5. Fierasferidae.

FAM. 1. DERCETIDAE.—Body much elongate; ordinary scales small or wanting, but two or more continuous series of enlarged scutes along each side; mouth large, praemaxillaries apparently forming the greater part of the upper border of the mouth, which is toothed; opercular apparatus complete. Dorsal fin more or less extended, without spines; anal short, caudal separate; ventrals with not less than 7 or 8 rays.

Dercetis, Leptotrachelus, Leptecodon, Pelargorhynchus, and Stratodus, from the Upper Cretaceous of Europe, Syria, and North America.

FAM. 2. HALOSAURIDAE.—Body elongate, covered with cycloid scales, the tail tapering to a point, without caudal fin; head with scales; mouth moderate, bordered by the praemaxillaries and the maxillaries, both toothed; suborbitals large; praeopercle rudimentary. Dorsal fin short, formed of soft rays, above or a little behind the ventrals, which are rather far back, and formed of 9 or 10 rays; anal very long, without spines, extending to the end of the tail. Ovaries transversely laminated, the ova falling into the abdominal cavity. Some 10 living species are known, referred to three genera, inhabiting the Atlantic, Pacific, and Indian Oceans, at depths of 500 to 1400 fathoms.

{624}In Halosaurus the scales of the lateral line, which runs near the lower profile, are scarcely enlarged, and are destitute of luminous organs. Halosaurichthys differs in the union of the ventral fins with each other, as in Notacanthus. In Halosauropsis the scales of the lateral line are strongly enlarged and pouch-like, and bear photophores.

This family is one of great antiquity, being represented in the Upper Cretaceous of Westphalia by Echidnocephalus, which, as shown by A. S. Woodward, appears to have been closely related to Halosaurus.

FAM. 3. LIPOGENYIDAE.—Similar to the preceding in shape and in the position of the dorsal fin, but with a toothless, roundish, inferior, suctorial mouth, and with the short dorsal and the long anal formed partly of spines and partly of soft rays. Head and body covered with minute scales; lateral line nearer the dorsal than the ventral profile. Ventrals with 3 spines and 7 soft rays. A single species, Lipogenys gillii, from the North Atlantic, 865 fathoms.

FAM. 4. NOTACANTHIDAE.—Body elongate, covered with very small cycloid scales, the tail tapering to a point, without caudal fin; head scaly; mouth small, inferior, bordered by the praemaxillaries only; jaws toothed; no suborbitals; praeoperculum small; post-temporal replaced by ligament. Dorsal fin formed of a series of short disconnected spines; anal very long, formed partly of spines and partly of soft rays, extending to the end of the tail. Ventrals with 1 to 5 spines and 7 to 10 soft rays.

{625}Two genera: Notacanthus, with the ventrals connate or confluent and with 6 to 12 dorsal spines; and Polyacanthonotus, with the ventrals separated and 27 to 38 dorsal spines. Nine species, from the Mediterranean, the Atlantic, and the Pacific, at depths of 400 to 1875 fathoms.

FAM. 5. FIERASFERIDAE.—Body elongate or extremely attenuate, naked, the tail tapering to a point or truncate, without distinct caudal fin; mouth small, inferior, bordered by the praemaxillaries; jaws toothed; no suborbitals; praeoperculum well developed. Dorsal and anal fins very long, extending to the end of the tail, and formed entirely of soft rays. Ventral fins absent. Vent situated immediately behind the gill-opening. Air-bladder with a muscular apparatus for dilatation of its anterior part.

A single genus, Fierasfer, with about 10 species, distributed over nearly all the warm and tropical seas, rarely found as far north as the west coast of Ireland. Encheliophis, without pectoral fins, is the larval form of Fierasfer.

Fierasfer spends the greater part of its existence in the interior of Holothurians and other Echinoderms as well as in bivalve Mollusca. It has been observed to enter Holothurians by the posterior or anal aperture, either head first or tail foremost, in the latter case availing itself of the suction which takes place alternately with the expulsion of water by that orifice; it remains near the anus, from which it projects its head in search of food outside its host. It is neither a true parasite nor a commensal or mutualist, in the sense given to these terms by Van Beneden, but simply a lodger, "inquilino," as Emery puts it. Semper, however, regards Encheliophis vermicularis as a true parasite, feeding on the viscera of the Holothurian in which it lives. Putnam has examined eight specimens of a Fierasfer from the Bay of Panama, which were obtained alive from pearl oysters, and also one beautifully {626}enclosed in a pearly covering deposited upon it by the oyster; a similar specimen is preserved in the British Museum.

SUB-ORDER 7. CATOSTEOMI.

Air-bladder, if present, without open duct. Parietal bones, if present, separated by the supraoccipital. Pectoral arch suspended from the skull; no mesocoracoid arch; coracoid usually very large. Ventral fins, if present, abdominal, or pelvis attached to the coracoid bones.

The mouth is small and bordered by the praemaxillaries or by the praemaxillaries and a small portion of the maxillaries. The air-bladder is present, except in the Solenostomidae and Pegasidae.

Following the suggestions of Kner and Steindachner and Cope to their logical conclusion, A. S. Woodward, in his valuable catalogue of the Fossil Fishes in the British Museum, has united the Lophobranchs of Cuvier with the Hemibranchs of Cope, a course which seems fully justified, and has received {627}further support from the recent investigations of Swinnerton,[698] who has proposed to unite the two groups under the new name of Thoracostei. The name Phthinobranchii has also been suggested by O. P. Hay for the same association. The structure of the Lophobranchs (Solenostomidae and Syngnathidae) shows that these fishes are only extremely specialised forms of the group of which the Sticklebacks are the well-known type, and the character of the "tufted" gills alone is surely not of sufficiently great importance to warrant the retention of the Lophobranchii as a division equivalent to the sub-orders adopted in the present classification. Besides, as recently pointed out by A. Huot,[699] there is no fundamental difference, but only one of degree, between the so-called tufted gill and the normal type; each "tuft" corresponds to one branchial lamella, and at a certain stage of development the disposition of the branchial lamella is the same in a Syngnathus and in an ordinary Teleostean. I have recently attempted to show[700] that the Lamprididae are related to the Hemibranchii, although sufficiently distinct to warrant the establishment of a division, named Selenichthyes.[701]

SYNOPSIS OF THE FAMILIES.

I. Praeoperculum and symplectic distinct; branchial apparatus fully developed; gills pectinated; mouth terminal, toothless; post-temporal forked, free; pelvic bones connected with scapular arch; ventral fins with 15 to 17 rays; ribs long, sessile; fins without spines (SELENICHTHYES) 1. Lamprididae.

II. Praeoperculum and symplectic distinct, latter much elongate; branchial apparatus more or less reduced; gills pectinate; post-temporal simple, immovable; mouth terminal (HEMIBRANCHII).

A. Mouth toothed.

1. Pelvic bones usually connected with scapular arch; spinous dorsal represented by isolated spines.

Snout conical or but slightly tubiform; ventral fins with 1 spine and 1 or 2 soft rays; ribs slender, free; anterior vertebrae not enlarged 2. Gastrosteidae.

Snout tubiform; ventral fins with 1 spine and 4 soft rays; {628} ribs flattened, fused with the lateral bony shields; anterior vertebrae not enlarged 3. Aulorhynchidae.

Snout tubiform; ribs slender, free; first vertebra enlarged 4. Protosyngnathidae.

2. Pelvic bones not connected with scapular arch; ventrals without spine, with 5 or 6 rays; snout tubiform; first vertebra very elongate, formed by the fusion of several.

Isolated dorsal spines; body scaly 5. Aulostomatidae.

No dorsal spines; body naked 6. Fistulariidae.

B. Mouth toothless; snout tubiform; two short dorsal fins, the first with a few spines; ventral fins with 3 to 5 rays; anterior vertebrae elongate.

Body covered with bony shields and small rough scales 7. Centriscidae.

Body completely cuirassed by bony shields which are fused with the endoskeleton 8. Amphisilidae.

III. Praeoperculum absent; symplectic much elongate; branchial apparatus more or less reduced; gill-lamellae reduced in number and enlarged, forming rounded lobes; post-temporal simple, immovably attached to the skull; mouth toothless, at the end of a tubiform snout; body covered with bony plates (LOPHOBRANCHII).

Two dorsal fins; ventral fins present, with 7 rays; gill-openings wide; exoskeleton of large star-like plates 9. Solenostomidae.

A single dorsal fin; no ventral fins; gill-openings very small; exoskeleton in the form of rings 10. Syngnathidae.

IV. Praeoperculum and symplectic absent; gills pectinated; mouth inferior, toothless; body entirely covered with bony plates; ventral fin with 2 or 3 rays (HYPOSTOMIDES) 11. Pegasidae.

FAM. 1. LAMPRIDIDAE.—Body short and deep, with minute scales. Snout short; mouth toothless, bordered by the praemaxillaries and, to a small extent, by the maxillaries; opercular bones well developed. Gills four, pectinated; branchial apparatus fully developed. Post-temporal bone forked. Vertebrae very numerous (21 + 25), without transverse processes; ribs strong, long. Fins without spines; dorsal and anal elongate. Pectoral fins with very short pterygials folding downwards against the body. Pelvic bones connected with the coracoids, which are very large, and do not form a suture at their ventral extremity. Ventral fins with 15 to 17 rays.

The Opah or King-Fish (Lampris luna), the sole representative of this family, is remarkable for its large size (growing to a length of four feet) and its vivid colours. Its flesh is rich, and intermediate between that of the Salmon and that of the Tunny. It is a pelagic fish of wide distribution, known from the North {629}Atlantic and Mediterranean and from distant points in the Pacific; specimens are occasionally captured on our coasts. It feeds on other fish, but little is known of its habits and nothing of its development.

The affinities of the Lamprididae are very doubtful. Lampris has usually been placed with the Acanthopterygians, a view which is still upheld by Gill.[702] I now agree with this high authority in regarding the bone which I took for an infraclavicle as a much developed coracoid, and the bone termed by me the coracoid as a pterygial. But it has also been shown, by Starks, that such a thing as an infraclavicle does not exist in the Stickleback, the bone so-called being only a part of the coracoid; and as in most of the Sticklebacks the pelvic bones join the latter, the resemblance between them and Lampris remains. As I have previously pointed out, the absence of spines in the fins, and the position of the ventral fins, together with the great number of rays in the latter, which is only met with in the lower Teleosteans, are characters which necessitate the removal of Lampris from the Acanthopterygians, and I cannot find a better place for them than near the Gastrosteidae.

The whole question of the arrangement of the Physoclists with abdominal ventrals (Catosteomi and Percesoces) is, I feel, much in need of revision, and it may be found advisable to break up this group into a greater number of sub-orders, in which case the Selenichthyes would stand by themselves; the Hemibranchii and Lophobranchii would be united under the former name, as proposed by Woodward, or under that of Thoracostei (Swinnerton) or Phthinobranchii (Hay). The position in the system of the Pegasidae is still somewhat doubtful. This family is regarded by some authors as related to the mail-cheeked Acanthopterygians.

FAM. 2. GASTROSTEIDAE.—Body more or less elongate, naked or protected by bony shields, tapering to a slender caudal peduncle. Head moderate, with short or elongate and tubiform snout; mouth small, terminal, toothed; opercular bones well developed; suborbitals in contact with praeoperculum, protecting the cheek. Gills four, pectinated. Praecaudal vertebrae with strong transverse processes and slender, free ribs. Spinous dorsal represented by isolated spines. Pectoral fins with short {630}pterygials. Pelvic bones usually connected with scapular arch. Ventral fins with one spine and one or two soft rays.

Four genera: Gastrosteus, Apeltes, Eucalia, Spinachia.

The little Three-spined and Two-spined Sticklebacks (Gastrosteus aculeatus and G. pungitius), which include many varieties that have been regarded as distinct species, are among the best known of our British Fishes. They are remarkable for the perfect indifference with which they can be transported from fresh into salt water, and vice versa, and for the elaborate nests which the males build in fresh or brackish water, and over which they watch with the greatest vigilance after the female has deposited her relatively large eggs.[703] These nests are made of weeds and twigs fastened together by threads secreted by the kidneys of the male. The {631}larger fifteen-spined Stickleback (Spinachia vulgaris) is entirely marine; its nests are to be found on our coasts in sheltered rock-pools, and they are made chiefly of sea-weeds and Hydrozoa. Sticklebacks are short-lived, and are believed to breed only once.

The Gastrosteidae are restricted to the northern hemisphere, being more abundant in the higher latitudes, extending to Iceland, Greenland, and Bering Straits; the southernmost points of their distribution are Algeria in the Old World, and Lower California in the New.[704]

A very large number of species have been described, but probably only about a dozen deserve to stand.

FAM. 3. AULORHYNCHIDAE.—The genera Aulorhynchus and Auliscus, each with one species from the Northern Pacific, much resemble Spinachia in outward form and in the equal size of the anterior vertebrae, but the snout is still more produced, tubiform, and the ventral fins are formed of one spine and four soft rays. The difference which justifies their separation as a distinct family resides in the disposition of the ribs, which are flattened and ankylosed to the lateral bony shields.

FAM. 4. PROTOSYNGNATHIDAE.—This family appears to be intermediate between the Gastrosteidae and the Aulostomatidae, agreeing with the former in possessing slender, free ribs, with the latter in having the first vertebrae elongate, though to a {632}less degree than in Aulostoma. Its only representative is Protosyngnathus sumatrensis, from a Tertiary freshwater formation in Sumatra, which has been referred, without adequate grounds, to Aulorhynchus or Auliscops.[705]

FAM. 5. AULOSTOMATIDAE.—Allied to the Aulorhynchidae, differing in the ventral fins devoid of spines, formed of 5 or 6 rays, widely removed from the pectoral arch, the very elongate, saddle-shaped anterior vertebra formed by the fusion of several, the large supratemporals produced backward over the anterior vertebra, the very elongate pterygials of the pectoral fin, and the compressed body covered with small ctenoid scales. Ribs are rudimentary or absent. No suborbitals. The snout is long, tubiform; the small terminal mouth bears bands of minute teeth, and the lower jaw has a small barbel at the symphysis. A single genus, Aulostoma, with two species from the Atlantic coasts of tropical America, and two from the Eocene and Miocene of Europe. A. coloratum grows to a length of 26 inches.

FAM. 6. FISTULARIIDAE.—Body greatly elongate, naked. First vertebra much elongate, formed by the fusion of several; strong transverse processes to the ribs in front and behind, those of two vertebrae suturally united; ribs rudimentary or absent. Supratemporal much produced posteriorly, extending over the anterior vertebrae; suborbitals absent; snout forming a long tube, which terminates in a narrow mouth with minute teeth. Spinous dorsal entirely absent. Pterygials of pectoral fin very elongate. Ventral fins very small, with 6 soft rays, inserted far behind the pectoral girdle.

The Flute-mouths, Fistularia, which Dr. Günther describes as "gigantic marine Sticklebacks living near the shore, from which they are frequently driven into the open sea," are represented by three species, from the tropical and subtropical parts of the Atlantic and Indo-Pacific. The middle rays of the forked caudal fin are produced into a long filament. The largest species, F. tabaccaria, reaches a length of 6 feet. The same genus is represented by two species in the Upper Eocene and Oligocene of Europe, and Urosphen, from the Upper Eocene, is regarded as allied to it.

{633}FAM. 7. CENTRISCIDAE.—Body moderately elongate, partially enclosed in a bony armour, which is distinct from the endoskeleton. Anterior vertebrae elongate, with strong parapophyses ankylosed to the exoskeleton; no ribs. Suborbitals absent; snout forming a long tube, with small, terminal, toothless mouth. Two dorsal fins, the anterior with a very strong spine. Pterygials of pectoral fin very small. Ventral fins small, with 4 or 5 rays, the pelvic bones in contact with the postclavicles.

Centriscus, with five species in the Atlantic and Pacific Oceans, represents this family at the present day. C. scolopax has occasionally been found on the English coast. Isolated spines from the Pliocene of Tuscany have been referred to the same genus. Rhamphosus, from the Eocene of Monte Bolca, is believed to have been allied to Centriscus.

FAM. 8. AMPHISILIDAE.—Near the preceding, but body extremely compressed and completely enclosed in a thin bony armour which is fused with the endoskeleton; the caudal region, much abbreviated, is free and relegated to the ventral surface, the body terminating in the two dorsals, of which the first bears a strong spine. The ventral fins are far back, very small, formed of 3 or 4 rays.

Amphisile is represented by three or four recent species in the Indian and Pacific Oceans, and two are known from Upper Eocene and Oligocene beds in Europe. Dr. Arthur Willey has observed these fishes in the Southern Pacific. A. strigata "lives in small shoals of about half-a-dozen individuals, and swims about with rapidity in a vertical position, cleaving the water with its razor-shaped body."

FAM. 9. SOLENOSTOMIDAE.—Body moderately elongate, with large star-like ossifications. Anterior vertebrae elongate, without transverse processes; no ribs. Snout much produced, tubiform; mouth, small, terminal, toothless; no praeoperculum; symplectic elongate; gill-opening wide; gill-lamellae small rounded lobes. Two short dorsal fins, the rays of the anterior not articulated, flexible spines. Pterygials of pectoral fin very small. Ventral fins large, with 7 rays, behind the pectoral arch. No air-bladder.

The unique genus, Solenostomus, with three or four species from the Indian and Pacific Oceans, may be regarded as in many respects intermediate between the Centriscidae and the Syngnathidae. In the female the inner side of the ventral fins coalesces {634}with the integuments of the body, forming a large pouch for the reception of the eggs.

Solenorhynchus, from the Upper Eocene of North Italy, probably belongs to this family, but its form is much more elongate, and the exoskeleton is in regular rings.

FAM. 10. SYNGNATHIDAE.—Body more or less elongate, protected by an exoskeleton forming rings. Anterior vertebrae not elongate; parapophyses strong, ankylosed to the exoskeleton; no ribs. Snout much produced, tubiform; mouth small, terminal, toothless; no praeoperculum; symplectic elongate; gill-cleft reduced to a very small opening near the upper posterior angle of the gill-cover; gill-lamellae small rounded lobes. A single dorsal fin. Pectoral fins, if present, with very small pterygials; ventrals absent. Caudal fin often absent; tail sometimes prehensile.

This family embraces about 175 marine species, and is represented over the greater part of the world. Principal genera: Siphonostoma, Syngnathus, Penetopteryx, Ichthyocampus, Nannocampus, Osphyolax, Urocampus, Doryichthys, Coelonotus, Stigmatophorus, Nerophis, Protocampus, Gastrotoceus, Solenognathus, Hippocampus, Acentronura, Phyllopteryx.

Remains are found in the Upper Eocene and Miocene beds of Europe, and have been referred to Siphonostoma and Syngnathus, and to the extinct genus Calamostoma. It is probable that Pseudosyngnathus, from the Upper Eocene of Monte Bolca, is the type of a distinct family.

The best known members of this family are the Needle-Fish or Pipe-Fish (Siphonostoma and Syngnathus) and Sea-Horse (Hippocampus) of our coasts. The latter, like Amphisile, swims with the body in a vertical position. In most species the male takes charge of the eggs, in a pouch under the tail (Siphonostoma, {635}Syngnathus, Penetopteryx, Nannocampus, Stigmatophorus, Hippocampus), in a groove under the tail (Phyllopteryx), or in a groove on the abdomen (Doryichthys, Coelonotus, Nerophis, Gastrotoceus).

An Australian species of Syngnathus has been described by E. P. Ramsay under the name of S. intestinalis, from its living inside Holothurians, in the manner of Fierasfer, and G. Lunel has observed a Doryichthys to offer a similar instance of inquilinism.

One of the most remarkable types of Syngnathids is Phyllopteryx, from Australia. The spines and knobs of the head and body are furnished with dermal appendages, which closely imitate the fucus among which they live.

FAM. 11. PEGASIDAE.—Body short or moderately elongate, encased in an exoskeleton forming rings. Anterior vertebrae not elongate; no ribs. Snout produced beyond the mouth, which is small, inferior, and toothless; no praeoperculum, no symplectic; gill-opening very small; gills pectinated. A single dorsal fin. Pectoral fins large, horizontal; ventrals reduced to one or two filamentous rays, behind the scapular arch. Air-bladder absent.

{636}Five or six species, referable to two genera, Pegasus and Parapegasus, make up this family. They are very small fishes, inhabiting the coasts of China, Japan, Arabia, the Malay Archipelago, and Australia. Pegasus is remarkable among all fishes in having the five anterior rays of the pectoral fin transformed into strong spines.

SUB-ORDER 8. PERCESOCES.

Air-bladder, if present, without open duct. Parietal bones separated by the supraoccipitaL Pectoral arch suspended from the skull; no mesocoracoid arch. Ventral fins, if present, abdominal, or at least with the pelvic bones not solidly attached to the clavicular arch.

This group connects the Haplomi with the Acanthopterygii, the Scombresocidae being somewhat related to the Cyprinodonts,[706] whilst the Anabantidae show distinct affinity to the Osphromenidae in the following sub-order. Other families, previously included among the Scombriform Acanthopterygians, are placed here on the assumption that the loose attachment of the pelvic bones to the clavicles is a primitive character, and not the result of degeneration, such as occurs in some cases among true Acanthopterygians. Although this sub-order is perhaps only an artificial association, it must be borne in mind that, notwithstanding the very wide divergence which exists between the first and last families, and however dissimilar their members may appear to be at first sight, a gradual passage may be traced connecting the most aberrant types.

SYNOPSIS OF THE FAMILIES.

I. Ventral fins, if present, inserted far behind the pectorals; no spines to the fins.

Ribs attached to the extremity of much-developed parapophyses; lower pharyngeal bones completely united; pectoral fins inserted very high up 1. Scombresocidae.

Ribs mostly sessile; lower pharyngeal bones distinct; pectoral {637} fins nearer the ventral than the dorsal line 2. Ammodytidae.

II. Ventral fins, if present, more or less approximated to the pectorals.

A. Two well-developed dorsal fins, the anterior small and formed, at least in part, of spinous rays.

1. Ribs attached to strong parapophyses.

Pelvic bones free or connected with the clavicles by ligament; pectoral fins inserted high up 3. Atherinidae.

Pelvic bones suspended from the postclavicles; pectoral fins inserted very high up; teeth very feeble or absent 4. Mugilidae.

Pelvic bones suspended from the postclavicles; pectoral fins low down, with detached lower rays 5. Polynemidae.

Pelvic bones connected with the clavicles by ligament; pectoral fins nearer the ventral than the dorsal line; dentition powerful, cardiform; scales minute or absent 6. Chiasmodontidae.

2. Anterior ribs sessile; pelvic bones not connected with the scapular arch; pectoral fins nearer the ventral than the dorsal line 7. Sphyraenidae.

B. Spinous dorsal, if present, connected with the soft.

1. Anterior vertebrae without parapophyses; scales on head, if present, small.

Oesophagus with lateral sacs which are beset with papillae internally; spinous dorsal long; scales rhomboidal, in oblique transverse series; pelvic bones free 8. Tetragonuridae.

Oesophagus with lateral sacs which are beset with toothed papillae internally; spinous dorsal, if distinct, shorter than the soft dorsal; scales moderate or small, cycloid, often deciduous 9. Stromateidae.

No sacs in the oesophagus; fins without spines; scales very small or absent 10. Icosteidae.

2. All, or all but the anterior two vertebrae with parapophyses; scales on head large; a superbranchial cavity.

No spines to the fins 11. Ophiocephalidae.

Strong spines to the dorsal, anal, and ventral fins 12. Anabantidae.

FAM. 1. SCOMBRESOCIDAE.—Maxillary entering the border of the upper jaw; dentition moderately strong or feeble. Lower pharyngeal bones united. Praecaudal vertebrae with strong parapophyses supporting the ribs. Body covered with cycloid scales. Pectoral fins inserted very high up; ventral fins widely separated from the pectorals, without spines, with 6 rays. Dorsal fin opposed to the anal, and likewise formed entirely of soft rays. Air-bladder generally present, sometimes cellular.

The shape of the head and body vary greatly, and the pectoral fin may reach an extraordinary wing-like development. The dorsal fin may be followed by a series of finlets, as in many of the Scombridae. Most of the Scombresocidae, of which about 200 species are known, are marine; some are carnivorous, {638}others (Hemirhamphus) mainly herbivorous, feeding on green algae. Nearly all are in the habit of making great leaps out of the water, this tendency culminating in the Flying-Fish (Exocoetus), which skip or sail through the air in a manner the explanation of which has given rise to much controversy. According to the latest evidence[707] the sole source of motive power is the action of the strong tail while in the water; no force is acquired while the fish is in the air. The pectorals are not used as wings but as parachutes. There is every passage between the small pectoral fin of a Saurie (Scombresox) or a Hemirhamphus and the swallow-like wings of the most developed Exocoetus. The genus Hemiexocoetus is a very remarkable connecting form. The Gar-Pike (Belone), of which one species is common on our coasts, have both jaws produced into a long slender beak; the bones are green. In Hemirhamphus the lower jaw only is prolonged; some of the species, living in fresh water, are viviparous, the anal fin being modified into a copulatory organ, as in many Cyprinodonts.

Scombresocidae occur in all the tropical and temperate seas. Belone, Scombresox, and Hemirhamphus are found in Upper Eocene and Miocene beds of Europe, and, as stated above, Protaulopsis should perhaps be referred to this family.

{639}FAM. 2. AMMODYTIDAE.—Maxillary excluded from the border of the upper jaw; mouth protractile; dentition feeble or absent. Lower pharyngeal bones separate. Praecaudal vertebrae without parapophyses. Body covered with very small cycloid scales. Pectoral fins nearer the ventral than the dorsal line; ventral fins, if present, widely separated from the pectorals, without spine, with 6 rays. Dorsal and anal fins more or less elongate, formed of soft rays. Air-bladder absent.

The existing genera, Ammodytes, with 8 species, from the temperate coasts of the northern hemisphere, and Hypoptychus, from northern Japan, with a single species, are deprived of ventral fins, and their exact relations remained obscure until the structure of the Oligocene Cobitopsis revealed their affinity to the Scombresocidae, or at least their pertinence to the present suborder. The Greater Sand-Eel or Launce (Ammodytes lanceolatus) and the Lesser Sand-Eel (A. tobianus) are common on our coasts, and are remarkable for the manner in which, by means of their sharp-pointed snout, they bury themselves with great rapidity in the sand, darting in and out like arrows.

FAM. 3. ATHERINIDAE.—Maxillary excluded from the border of the upper jaw; dentition more or less developed. Body covered with cycloid or ctenoid scales. Ribs attached to strong parapophyses. Pectoral fins inserted high up; ventral fins more or less approximated to the pectorals, with one spine and five soft rays; pelvic bones connected with the clavicular symphysis by a ligament. Two well-separated dorsal fins, the anterior small and formed, at least in part, of spinous rays. Air-bladder present.

Carnivorous Fishes, mostly marine and of small size, much valued as food, and distributed along the coasts of most tropical and temperate seas; some inhabit fresh waters. A silvery lateral band, or "stole," is usually present. About 65 species are known, referred to 14 genera: Atherina, Iso, Chirostoma, Thyrina, Atherinella, Labidesthes, Atherinopsis, Atherinops, Telmatherina, Neatherina, Pseudomugil, Rhombatractus, Aida, Melanotaenia.

{640}Represented in the Upper Eocene of Europe by several species of Atherina and by the extinct genus Rhamphognathus.

FAM. 4. MUGILIDAE.—Maxillary excluded from the border of the upper jaw; dentition feeble or absent. Body covered with cycloid scales. Ribs attached to the extremity of strong parapophyses. Pectoral fins inserted high up; ventral fins more or less approximated to the pectorals, with one spine and five soft rays; pelvic bones suspended from the post-clavicles. Two well-separated dorsal fins, the anterior formed of a small number of spines. Air-bladder present.

These Fishes are closely related to the preceding, of which they are a further specialisation, the pharyngeal bones having a complicated structure, much reducing the oesophageal opening, and the vertebrae being reduced in number (24 to 26 instead of 32 to 60). They feed on organic matter contained in mud, and inhabit the fresh waters and coasts of the temperate and tropical regions. The species number about 100. Principal genera: Mugil, Myxus, Anostomus, Joturus. Grey Mullets (Mugil) are represented on our coasts by three species, valued as food, one of which (M. capito) has a remarkably wide range, occurring from Scandinavia to the Cape of Good Hope. Remains referred to the same genus occur in the Miocene and Oligocene.

FAM. 5. POLYNEMIDAE.—Maxillary excluded from the border of the upper jaw; dentition feeble. Body covered with ctenoid {641}scales. Ribs attached to the extremity of strong parapophyses. Pectoral fin inserted low down, with a lower portion consisting of free rays; the upper portion, or fin proper, attached to the scapula, the lower to a fenestrate bone which appears to be formed by coalesced pterygials (Fig. 391, pt.). Ventral fin more or less approximated to the pectoral, with one spine and five soft rays; pelvic bones suspended from the post-clavicles. Two well-separated dorsal fins, the anterior formed of a small number of spines. Air-bladder, if present, very large.

The vertebrae number 24 (10 + 14).

Three closely allied genera: Polynemus, Pentanemus, and Galeoides, with about 25 species, from the shores of tropical seas, often entering rivers. Some attain a length of 4 feet, and are valued as food or for the isinglass yielded by their air-bladder. The free pectoral filaments are organs of touch, and can be moved independently of the fins.

FAM. 6. CHIASMODONTIDAE.—The deep-sea genera, Chiasmodon, Pseudoscopelus, and Champsodon, which have been placed either with the Gadidae, the Trachinidae, or the Berycidae, may be referred to the Percesoces, as the pelvic fins have only a ligamentous connexion with the pectoral arch. Unfortunately, the skeleton has only been examined in Champsodon; it is remarkably similar to that of the Atherinidae. As in Atherinichthys, the posterior extremity of the air-bladder is protected by a bony sheath formed by the expanded ring-like haemal processes of the anterior caudal vertebrae. Vertebrae 32 (16 + 16). The {642}scales are absent or very small and spinulose, the mouth large, with cardiform teeth; spinous dorsal short, soft dorsal and anal elongate. Chiasmodon and Pseudoscopelus have a complicated system of sensory organs on the body, which in the latter suggest the photophores of Scopelids. Champsodon vorax is a fish of extreme voracity, swallowing prey much larger than itself. Only four species of this family are known.

FAM. 7. SPHYRAENIDAE.—Maxillary excluded from the border of the upper jaw; dentition very strong. Body covered with cycloid scales. Anterior ribs sessile, the rest inserted on parapophyses. Pectoral fin nearer the ventral than the dorsal outline; ventral fin more or less approximated to the pectoral, with 1 spine and 5 soft rays; pelvis not connected with the pectoral arch. Two well-separated dorsal fins, the anterior formed of a small number of spines. Air-bladder large. Vertebrae 24.

Carnivorous Pike-like Marine Fishes from the tropical and sub-tropical seas, often found at the mouths of rivers. The "Barracudas" form a single genus, Sphyraena, with about 20 species, the largest of which grow to 8 feet and are dangerous to people bathing; many are valued as food, but some are reported to be poisonous, at least at certain seasons. Remains of several species are known from the Eocene and later periods in Europe and North America.

Spotted a typo or a problem in this chapter? Tell us