For the first dozen years or so the annual rings can be easily followed, but when the creature approaches maturity each shield adds very little to its growth, and the rings become very fine, crowded and irregular. Only by careful counting and comparison of the rings on the costals, marginals, and plastrals, can a reliable average be arrived at. In some tortoises, e.g. Chrysemys, the whole outer layer of the shields peels off periodically; only a thin smooth layer like mica or tracing-paper remains, of course without any indication of rings. The pigment is formed in the Malpighian layer, but it frequently diffuses into the horny shields themselves, notably in Chelone imbricata, which yields the beautiful "tortoise-shell." The colour of the pigment is either black, yellow, or red, with resulting combinations. The green colour, often so beautiful in baby-specimens of Chrysemys, is optical, produced, according to Agassiz, by a network of black pigment, spread over a layer of yellow oil.

Horny scales, sometimes forming spines, and covering a nodule of dermal ossification, are also common on other parts of the skin, especially on the limbs of land-tortoises, and also on the tail of Chelydra. Sometimes the end of the tail is protected by a claw-like nail, for instance in Pyxis. In some of the gigantic land-tortoises, and in Chelone mydas, this nail assumes large dimensions, and several of the terminal caudal vertebrae are fused together into a regular urostyle. In some subfossil specimens of Mauritian tortoises, these ankylosed complexes are 12 cm. long and more than 5 cm. broad!

Before leaving the description of the shell, it is worth while to draw attention to the enormous correlative changes in other organs produced by this case. Nearly the whole organism has been altered. The hard, firm carapace has partly rendered the supporting functions of the vertebral column unnecessary or impossible. In many tortoises, especially in the large land-tortoises, the vertebrae and the capitular portions of the ribs are reduced to mere bony outlines; the reduction to thin paper-like bony lamellae proceeds with age. The iliac bones find a better support in the costal plates; the contact with the sacral ribs is given up, and these ribs fuse partly with the costal plates, or they are absorbed. The whole mass of muscles of the trunk is completely lost in the region of the shell, but traces of them exist in young specimens. Neck, limbs, and tail can in most cases be withdrawn and hidden in the shell. When this is not possible it is due to secondary changes. The neck is withdrawn either by being tucked away sideways (Pleurodira^[129]), or by being bent in an S-shaped curve in a vertical plane. In a left-sided profile-view of the animal, the head represents the tail of the S. The neck is withdrawn by long muscles, which are inserted into the ventral side of the middle of the neck, and extend in the shape of vertical ribbons far back into the shell, arising from the centra of some of the middle or even more posterior thoracic vertebrae.

Lastly, a few remarks on the partial regeneration, or the mending of injuries to the shell. If part of the horny covering is badly bruised, torn off, or rubbed through, or if part of the shell is crushed, the underlying portion of the bony plate becomes necrotic, and the horny covering also dies so far as its Malpighian layer is destroyed. Soon, however, the uninjured Malpighian cells, around the margin of the wound, multiply, grow into and beneath the injured portion of the bone, and form a new horny layer, casting off the necrotic portion. After several months the deficiency is patched up; new bone has grown in the deeper remaining strata of the cutis, and the outside is covered by a continuous horny layer, without, however, reproducing the original concentric moulding of the shields. In badly crushed shells sometimes almost one-third of the whole shell is thus cast off and mended within one or two years. The regeneration of the forcibly stripped-off shields of Chelone imbricata is described on p. 386. Bitten-off tails and limbs, rather frequent occurrences in water-tortoises, are of course not reproduced, but the wounds are healed and covered again with scaly skin.

Sense-organs.–The EYE is by far the best developed sense-organ. It is comparatively small. The pupil is round. The iris is mostly dark in terrestrial forms, while in water-tortoises it is often brightly coloured, for instance pale yellow in Chelodina, greenish and mottled with black, pale grey, brown, etc., in various species of Chrysemys. Cistudo presents a curious sexual dimorphism; the males have red, the females brown, eyes. The sclerotic wall contains a ring of numerous small ossified plates. There is no trace of a pecten. The eye is protected externally by the two lids and the nictitating membrane. In some water-tortoises, notably in Chelodina, the lower lid is transparent. Lacrymal and Harderian glands are present.

The SENSE OF HEARING is apparently not very acute, although tortoises and turtles are frightened by noise, and can distinguish sounds; otherwise they would have no voice, which is very tiny and piping in most tortoises during the pairing season. In most water-tortoises the tympanic membrane is thin and quite exposed; in land-tortoises it is often thick and covered by the ordinary skin; lastly, in Chelone the tympanic cavity is filled with a plug of the much-thickened skin, possibly in adaptation to the water-pressure when these creatures dive to considerable depths. The ossicular chain is mostly reduced to a long, bony, columellar rod.

The SENSE OF SMELL is well developed. All Chelonians carefully smell their food, in the air as well as under water. The individual predilection shown by many species for different kinds of animal and vegetable food,–since they are, for instance, able to distinguish between the various sorts of cabbage, cauliflower, sprouts, etc.,–proves that they possess a considerable amount of smell and taste.

Tortoises have a fine sense of touch; even the slightest tap on the shell is noticed, and the skin of the soft parts is extremely sensitive. Tickling of the sides of the tail, or of the hinder surface of a thigh, produces ridiculous scratching actions of the same or of the opposite foot.

The digestive apparatus is simple. Only a few peculiarities need be mentioned. The tongue is mostly broad and soft; it cannot be protruded. The oesophagus of the Chelonidae is covered with many conical projections pointing towards the stomach. The latter is simple, except in Sphargis. The intestine is devoid of a caecum, but the difference between the small intestine and the rectum is very marked and often abrupt. The cloaca is very roomy. It contains the large copulatory organ, which is unpaired, grooved on its dorsal side, and is altogether constructed like that of the Crocodilia. The large bladder opens ventrally into the urodaeum, a recess of the cloaca; near its base open the urinary and genital ducts. Many water-tortoises possess also a pair of lateral thin-walled sacs, the so-called anal sacs, dorso-lateral diverticula of the walls of the urodaeum. These sacs, which have highly vascularised walls, are incessantly filled and emptied with water through the vent, and act as important respiratory organs. When such a water-tortoise, for instance an Emys or a Clemmys, is suddenly taken out of the water, it squirts out a stream of this water, which is not, as is generally supposed, the urine from the bladder.

The mode of respiration is interesting. The lungs are very complicated, highly-developed, spongy structures. They are attached by their whole dorsal surface to the inner lining of the shell. As they cannot expand through their own initiative, and since the shell has made costal and abdominal expansion impossible, the tortoise has to resort to other means of producing the necessary vacuum. This is done partly by the neck and the limbs, which act like pistons in being drawn in and out; partly by the greatly developed hyoidean apparatus, by which, when the neck is stretched out, the throat is alternately inflated and emptied, the air being swallowed, or pumped into the lungs. Additional respiration, besides that of the anal sacs mentioned above, is effected in various aquatic tortoises by slightly vascularised recesses of the pharyngeal region. Most Chelonians can exist for a very long time without breathing; sulky individuals remain for hours or days under water. Cistudo can shut itself up for an equally long time. Nevertheless this and other land-tortoises easily get drowned.

All Chelonians lay white eggs, round or oval, according to their kind, but the shape of the eggs of one set sometimes varies within the greatest limits. The shell varies from a parchment-like, flexible, scarcely calcareous cover to a hard, well-polished case. As a rule the eggs, imbedded in the ground, are hatched after a few months, but in some of the northern kinds, e.g. Emys orbicularis, the hatching is deferred until the next spring, the embryo's development being arrested during the winter. How such eggs, buried a few inches only below the surface, withstand the often very severe North German and Russian winter is a mystery. Whilst the plastron is generally flat, it is more or less concave in the males of many species, notably in Testudo, Cistudo, and Emys.

The general conclusions which can be drawn from the present geographical distribution of the Chelonia are as few and unsatisfactory as those applying to the Crocodilia, since all the main groups of Chelonians, and many more extinct families, occurred together in bygone ages in the same countries, for instance in Europe.

The marine forms are naturally cosmopolitan, but the Testudinidae are likewise cosmopolitan, except in the Australian region. The Chelydridae, now restricted to North and Central America, occurred formerly also in Europe. The Pleurodira, in Mesozoic times plentiful in Europe, India, and North America, are now restricted to South America, Australia, and Africa; the Pelomedusidae to Africa, Madagascar, and South America; the Chelydidae to South America and Australia. In the latter country all the Chelonians belong to the Chelydidae. The Trionychoidea, occurring since the Cretaceous epoch in North America, in Early and Mid-Tertiary times in Europe, are now restricted to North America, Asia, and Africa. The country richest in Chelonians is America; North and Central America together possessing representatives of all the families except the Pleurodira, and these we know to have died out there. The Dermatemydidae, Cinosternidae, and Chelydridae are now restricted to the Nearctic sub-region (including Central America). Poorest in genera and species, all of them Chelydidae, is the Australian region, where no fossils of other families have yet been discovered. Europe, with its few Testudinidae, does not come into consideration; Asia has at least Testudinidae and Trionychidae, and in addition the solitary Platysternum in Indo-China, representative of a family whose affinities with the Chelydridae again proclaim the validity of the Periarctic region.

Order I. ATHECAE.

The vertebrae and ribs are not fused with, but are free from, the carapace, which consists of numerous small polygonal plates and is covered with leathery skin without any epidermal shields. The limbs are transformed into paddles. The neck is not retractile. Marine.

Fam. Sphargidae.–Sphargis s. Dermatochelys coriacea, the Leathery Turtle or Luth, is the only recent species and is the largest of all recent Chelonians. The biggest specimen in the national collection is about six feet and a half long, from the nose to the end of the shell, which latter is about four feet long; such a specimen may weigh half a ton. Agassiz, however, says that he has seen some "weighing over a ton." The general colour is dark brown, either uniform or with yellow spots. The Leathery Turtle has a wide distribution, ranging over all the intertropical seas, but it is rare everywhere; least so perhaps in the Western Atlantic from Florida to Brazil and in the Indian Ocean.

According to Agassiz it breeds regularly every year in the spring on the Bahamas, on the Tortugas, and on the coast of Brazil, depositing its many eggs on the sandy shore like other turtles. Accidentally it visits the northern coast up to Long Island, and specimens, perhaps carried with the Gulf Stream, have been caught on the coasts of Europe, for instance off Dorsetshire. One was caught near Nantes in 1729, and is said to have made a terrible noise when being killed. This is perhaps the reason why Merrem in 1820 invented the generic name Sphargis, supposed to be derived from σφαραγέω (I make a noise). It has also been recorded from the Mediterranean. It seems to be entirely carnivorous, living upon Molluscs, Crustacea, and fish. The flesh is supposed to be unwholesome. It is a very curious fact that of this rare species only large specimens, besides a very few baby-turtles, are known or preserved in collections, while individuals of intermediate size, say from four inches to three feet in length, have never been recorded. If it were not for the fact that they are still known to breed, it would look as if the species were dying out. Perhaps they are very shy, leading a pelagic life, diving at the least sign of danger, and coming near the land only for the sake of breeding.

The structure of Sphargis is so peculiar in many respects that it deserves a somewhat full account. The neuro-central sutures persist on all the vertebrae. The eight cervicals are short. All the ten trunk-vertebrae carry ribs, and these, with the exception of the last, articulate between the centra and with the neural arches; the first and tenth ribs are short, the others are long and flattened, but not broad, with wide spaces between them. The tail is short, although it consists of about twenty vertebrae; these are devoid of chevrons.

The skull superficially resembles that of Chelone, chiefly owing to the completely roofed-in temporal region. The supraoccipital crest is rather short, covered completely by the parietals, the posterior margin of which is rounded off instead of forming, as in the Chelonidae, a long projecting triangular crest with the supra-occipital. The parietals are in broad contact with the postfrontals, posteriorly they are just reached by the squamosals. The quadrato-jugal is small, separated from the postfrontal by the meeting of the squamosal with the jugal. The quadrate is notched behind, and it separates the opisthotic from the squamosal. The basisphenoid is large and broad, extending far forwards so as to separate the pterygoids widely from each other except in their anterior portions, which, instead of sending a lateral arm to the jugal and maxillary, as in Chelone, are widely separated from these bones by the palatines. The choanae lie on either side of the anterior half of the vomer, and are not roofed over by ventral vomero-palatine wings.

The limbs and their girdles are essentially like those of the Chelonidae, but are not derivable from them. The most remarkable feature is the shell. The dorsal and ventral halves are directly continuous, forming one unbroken case all round, which is composed of many hundreds of little bony plates, irregularly polygonal, fitting closely into each other with their sutural edges, and giving the shell a beautiful mosaic appearance. On the dorsal side are a median row and three pairs of lateral rows of larger plates, and these form seven longitudinal blunt ridges which all converge towards the triangularly pointed tail-end of the shell. The ridges are not so much produced by thickened or spine-like edges of the plates, but by the right and left halves of the plates being actually bent at an angle. This is most conspicuous at the sides of the shell where it passes into the ventral portion. The latter has two pairs of lateral and one median ridge. The whole shell has consequently twelve ridges. The mosaic plates are deeply imbedded in the cutis, being externally as well as internally covered or lined with dense leathery skin. The epiderm is thin, and shows no indications of horny scales. In young specimens the whole shell is soft and very imperfectly ossified, later on it is quite rigid, although comparatively thin. It is nowhere in contact with the internal skeleton, except by a nuchal bone, which by a descending process articulates with the neural arch of the eighth cervical vertebra.

The affinities of the Sphargidae and their position in the system are still debatable. Whilst some authorities, e.g. Cope, Dollo, and Boulenger look upon Sphargis as the sole remnant of a primitive group in opposition to all the other recent Chelonia, Baur considered it the most specialised descendant of the Chelonidae. Dames agreed with him. Van Bemmelen has modified this view in so far as he regards Sphargis as the most specialised Chelonian, but considers the differences between it and the Chelonidae great enough to conclude that both Sphargidae and Chelonidae represent two independent, partly parallel, branches which have arisen from two different groups of terrestrial tortoises. Case,^[130] from the study of Protostega and other fossil forms, tends towards Baur's view. He believes that Sphargis is the culminating form of a branch which through Psephophorus and with Eosphargis has sprung from some creature like Lytoloma, which at the same time is the starting-point of another branch which culminates in the genera Thalassochelys and Chelone, while lastly a third branch contains Protostega, Protosphargis, and Pseudosphargis. In other words, he considers them all Chelonidae. If he is right we have of course no business to separate Sphargis with its fossil allies from the rest of the Chelonia as "Athecae."

However, Case has not proved his point. It is easy enough to understand that the characters of the cranium and plastron of Sphargis are in a condition which by partial reduction can be derived from that of typical Chelonidae. The structure of the cervical vertebrae, the absence of the marginal plates and the peculiar articulation of the nuchal with the last cervical vertebra can be explained as convergent analogies, just like the paddles of Carettochelys. But the shell of Sphargis is fundamentally different from and not homologous with that of the others. Cope was therefore quite justified in distinguishing the Sphargidae as "Athecae" in opposition to the others which Dollo later on, by contrast, named "Thecophora." Unfortunate names, since both groups are undeniably in possession of a θήκη or shell. Both authors meant, however, by Theca the epidermal shields, but even this distinction is rendered invalid by Carettochelys.

The most reasonable explanation has been suggested by Hay.^[131] The mosaic polygonal components of the shell of Sphargis are, so to speak, an earlier generation of osteodermal plates than the later generation of longer and broader bony plates which in the Thecophora come into contact, and fuse with, the neural arches and ribs. The osteoderms of Sphargis belong to the same category as the dermal ossifications in the scutes of Crocodilia, whilst the plates of the carapace and plastron of the Thecophora belong to the category of the abdominal ribs. Sphargis has the first kind in its peculiar shell, the second kind in the deeper lying plastron and in its nuchal plate. But it has lost, or perhaps had never developed, the horny shields. The only difficulty is, however, the presence of a plastron and of a typical neural plate in Sphargis. This difficulty is not very serious. The plastron is a very old institution. It occurs together with the more superficial osteoderms in Caiman, and the nuchal plate may be the oldest of all dorsals. We can scarcely imagine that the direct ancestors of Sphargis had developed both kinds of shells, and that comparatively recently the inner shell of the carapace was lost, leaving only the nuchal plate. Fossils do not support such an assumption. Undoubted ancestral forms of Sphargis are very rare. Psephophorus of the Oligocene and Miocene of Europe had a continuous mosaic shell much resembling that of Sphargis; Eosphargis is represented by a well-preserved skull from the London clay. Then follows a wide gap until we come to Psephoderma of the Rhaetic, or Upper Trias of Bavaria; the large fragment of whose dorsal shell is composed of about 200 mosaic pieces. If this fragment really formed part of the shell of a Chelonian, its age would speak greatly in favour of the Athecae being a very primitive and independent group.

Order II. THECOPHORA.

Thoracic vertebrae and ribs united with a series of median or neural and a paired series of lateral or costal plates. Parietals prolonged downwards, meeting the pterygoids directly or by interposition of an epipterygoid.

Sub-Order 1. Cryptodira.–The carapace is covered with horny shields. The neck, if retractile, bends in an S-shaped curve in a vertical plane. The pelvis is not fused with the shell.

Fam. 1. Chelydridae.–The plastron is small and cross-shaped (Fig. 61, 2, p. 315); the bridge is very narrow, and the displaced abdominal shields are widely separated from the marginals by a few irregularly shaped inframarginals. The tail is long. The limbs, neck, and head are so stout that they cannot be completely withdrawn into the shell. Snout with a powerful hooked beak. American; only two genera, each with one species.

The temporal region is roofed very incompletely and only anteriorly by the expanded parietals and postfrontals, which form a long suture. The plastron consists of nine bony plates, a small entoplastron being present; there are lacunae in the middle line, the plates meeting imperfectly, and the horny abdominal shields are likewise separated by soft skin. The carapace has a nuchal with long rib-like processes which underlie the marginals; the neural plates form a continuous series. There are twenty-three marginal plates. The pubic and ischiadic symphyses remain separate, enclosing one large heart-shaped foramen. The five fingers and toes are webbed and are protected by claws except the outer toe, the nail of which is usually suppressed.

Chelydra serpentina, the Snapping Turtle, attains a large size, namely, a shell-length of more than one foot, and a total length from the nose to the tip of the tail of more than three feet. Its range extends from the Canadian lakes east of the Rocky Mountains, through the United States and Central America. The carapace of young specimens has three very marked series of keels, which gradually disappear with age, until in very old individuals the shell becomes quite smooth. The skin is very warty, especially on the neck, and there is a pair of minute barbels on the chin. The tail carries three series of originally triangular horny crests, which with age are transformed into blunt knobs. The general colour of this rather ugly creature is olive, mottled with dark brown above and with yellowish below.

According to Holbrook the Snapping Turtle is found in stagnant pools, or in streams where the waters are of sluggish motion. Generally they prefer deep water, and live at the bottom of rivers; at times, however, they approach the surface, above which they elevate the tip of their pointed snout, all other parts being concealed; and in this way they float slowly with the current, but if disturbed they descend speedily to the bottom. They are extremely voracious, feeding on fish, reptiles, or any animal substance that falls in their way. They take the hook readily, whatever may be the bait, though most attracted by pieces of fish; in this way many are caught for the market. It is, however, necessary to have strong hooks and tackle, otherwise they would be broken, for the animal puts forth great strength in his struggles to escape, both with his firm jaws and by bringing his anterior extremities across the line. When caught they always give out an odour of musk, which in very old animals is sometimes disagreeably strong.

Occasionally the Snapping Turtle leaves the water, and is seen on the banks of rivers or in meadows, even at a distance from its accustomed element. On land his motions are awkward; he walks slowly, with his head, neck, and long tail extended, elevating himself on his legs like the Alligator, which at that time he greatly resembles in his motions; like the Alligator also, after having walked a short distance, he falls down to rest for a few moments, and then proceeds on his journey. In captivity they prefer dark places, and are exceedingly ferocious; they will seize upon and bite severely anything that is offered them, and their grasp upon the object with their strong jaws is most tenacious.

The Snapping Turtles, or "Snappers," are feared on account of the ferocious bites which they inflict, and they are hated because of the destruction of valuable fish and water-fowl. They in turn atone for this damage by being eaten, especially the younger half-grown individuals, the flesh of the older ones being too much tainted with the odour of musk. The round eggs, which are laid to the number of twenty to thirty in the summer (in the Northern States about June), are likewise good to eat. The first act of the young creature on leaving the shell is said to be snapping and biting. In captivity they are often very sulky, and refuse food stubbornly for many months, perhaps for a whole year, and apparently without much harm to themselves, since they lie quietly in the distant corner of the tank, now and then slowly rising to the surface to breathe. Fresh-water algae grow on the shell and in the mud which settles on it, and since this happens also in the wild state, they are rendered as inconspicuous as old rotten logs. In order to attract fishes they protrude a pair of worm-like, pale pink filaments from the tip of the tongue.

Macroclemmys temmincki, the "Alligator Turtle."–In size and general appearance much like the other Snapping Turtle, but the dorsal shields have each a strong and prominent keel, and these three series increase in size with age. The costal shields are separated from the marginals by an additional series of about four supramarginals, well shown in the illustration. The shields of the cross-shaped plastron are subject to much individual variation, small shields being frequently intercalated, or rather retained, between the usual ones, especially between the pectorals and abdominals, in the gular region, and on the narrow bridge, where the inframarginals number one to three or even more. This species inhabits, broadly speaking, the whole basin of the Mississippi and Missouri rivers.

This beast is as vicious as the other Snapping Turtle. According to Agassiz it does not withdraw its head and limbs on the approach of danger, but resorts to more active defence. It raises itself upon the legs and tail, highest behind, opens the mouth widely, and throwing out the head quickly as far as the long neck will allow, snaps the jaws forcibly upon the assailant, at the same time throwing the body forward so powerfully as often to come down to the ground when it has missed its object.

It lives mostly in the water, but makes considerable journeys overland. Both in the water and on dry land the limbs move nearly perpendicularly, and the body is raised high. On dry land a considerable part of the weight of the body is borne by the long, strong tail.

"They are as ferocious as the wildest beast of prey, but the slowness of their motions, their inability to repeat the attack immediately, their awkwardness in attempting to recover their balance when they have missed their object, their haggard look, and the hideous appearance of their gaping mouth, constitute at such times a picture as ludicrous as it is fearful and revolting. Their strength is truly wonderful. I have seen a large specimen bite off a piece of a plank more than an inch thick. They take hold of a stick with such tenacity that they may be carried for a considerable distance suspended to it free above the ground. Fishes and young ducks are their ordinary prey. They lay from twenty to forty or more round eggs only about the size of a small walnut in holes which they dig in sloping banks not far from the water" (Agassiz).

Fam. 2. Dermatemydidae.–The pectoral shields are widely separated from the marginals by inframarginals, the gular shields are very small or absent, and the tail is extremely short. Only two or three genera, with three or four species in Central America.

The plastron is composed of nine plates. In Dermatemys mawi it is large, firmly joined to the carapace, covered with eleven or more shields, and there are four inframarginals; in Staurotypus salvini of Mexico the plastron is cruciform, with the anterior lobe movable, covered with seven or more shields, according to the fusion of the anal shields and the presence or absence of the gulars; there are only two inframarginals. The pubic and ischiadic symphyses remain separate; the temporal fossa remains widely open, the postfrontals scarcely touching the parietals. There are 23 marginal shields in Staurotypus, 25 in Dermatemys, including the unpaired nuchal. The nuchal plate has a pair of rib-like processes like those of the Chelydridae, but some of the posterior costal plates, sometimes only one pair, meet in the middle line, overlying or suppressing the corresponding neural plates. The shell of these aquatic tortoises is rather flat, more or less keeled, especially in young specimens, and in the fully adult condition is about one foot in length.

Fam. 3. Cinosternidae, represented by the single genus Cinosternum, with about ten species in North and Central America, and one in Guiana. Closely allied to the two previous families, with which it agrees by the separation of the pubic and ischiadic symphyses, the presence of an ento-plastral plate, the possession of inframarginal shields (Fig. 61, 3, p. 315), the widely open temporal fossae, and the rib-like pair of processes to the nuchal plate. It agrees with the Dermatemydidae in the interruption of the neural plates by the meeting of several pairs of the costal plates. There are 23 marginal shields; five or four shields, according to the presence or absence of the gular on the plastron, and in some species these plastral shields become, with age, more and more separated from each other by soft skin (see Fig. 75). The shape and size of the plastron differ considerably in the various species; in most of them, e.g. in C. pennsylvanicum and C. leucostomum, but not in C. odoratum, the anterior and posterior lobes are movable, with transverse soft hinges, so that the animal can completely close its shell. The skin of the legs and neck is so baggy and loose that these parts slip in, the skin rolling off, when the creature withdraws into its shell. They lay only a few–from three to five–elliptical eggs, which have a shining, glazed, and thick, but very brittle shell.

Cinosternum odoratum, the Mud-Turtle, or Stinkpot Terrapin, so called on account of the disagreeable smell which exudes from the inguinal glands. The head is disproportionately large, with the snout rather compressed laterally, and pointed underneath, with several short barbels. The neck is long and slender. The carapace of the young is keeled, each of the neural shields being raised in the middle line; but in full-grown specimens the shell becomes quite smooth and rounded. The horny shields of the plastron are relatively largest in the young, but they soon leave ever-increasing spaces between them, which are then filled with soft skin only, which thinly covers the underlying bone. The fore- and hind-limbs, especially the latter, are extensively webbed, and are provided with five short claws. The general colour of the shell is horny brown, either uniform or with darker spots or streaks. The neck and limbs are mottled brown. The only ornamental colouring is a pair of clear yellow broad lines on each side of the head, and a similar streak on each side of the lower jaw. On the chin and upper throat are two pairs of small tentacles. The tail of the male is of about the length of the hind-limbs, while that of the female is so short that its tip scarcely reaches beyond the hinder margin of the carapace. Length of the shell of full-grown specimens between four and five inches. Very young specimens have a rather droll appearance, owing to the long and slender neck with the large head, and the humpy back.

This species is common in the eastern half of North America, from Canada to Texas. It is mainly aquatic, and is one of the dullest and shyest species. My own specimens spend most of their time in the water, invariably in the darkest corners, preferably under a stone or a log, and they do not leave their hiding places until dark, in search of worms, meat, and all sorts of animal food. For months I could never induce them to take food from a stick, or even to eat in my presence, and it was not until after many weeks that one of them at last protruded its head far enough to exhibit the yellow stripes. When taken out of the water they draw in their heads, just allowing the vicious little eyes to be visible, and opening the sharp-edged mouth widely to bite deliberately and furiously at the unwary finger. Some spent the winter in the water, in the greenhouses, feeding as usual, others crept on land, hiding under moss, half buried in the soil, where they slept for several months, but with interruptions in order to soak and to drink. When spring is well advanced they prefer the water for their regular sojourn. Some which had been sent over from New York arrived in a deplorably dried-up condition, the skin being quite flabby and shrivelled, but after a few hours' soaking they came round, and increased considerably in weight, the limbs and neck becoming turgid.

C. pennsylvanicum of Eastern North America has a larger, more oval plastron. The head is not so strikingly large as in the other species and, like the neck, is brown with yellowish spots, and often has streaks on the sides. The tail of the male ends in a nail-like horny point. The lobes of the plastron are well hinged in the adult.

C. leucostomum of Central America is larger, with a shell-length of six inches. The plastron is not at all cruciform, but has a broad bridge, and fills the box, moreover it has an anterior and a posterior hinge, so that the box can be completely closed. Hence the vernacular name of the Box-Terrapin.

Fam. 4. Platysternidae, represented by the single species Platysternum megacephalum in Burma, Siam, and Southern China.

The pectoral shields are widely separated from the marginals by inframarginals, the plastron is large, oblong, not cruciform, and the tail is long.

The plastron consists of nine plates, and is covered with six pairs of shields, the most anterior of which are the broad gulars. The nuchal plate has no rib-like processes. The neurals form a continuous series, and there are twenty-three marginal scutes. The temporal fossae are completely roofed over, owing to the long sutures formed by the parietals with the postfrontals, moreover the postfrontals expand laterally so much that they posteriorly come into broad contact with the quadrato-jugals and squamosals, anteriorly with the maxillaries, so that the jugals are completely surrounded by bones, and are shut off from the orbits and from the temporal fossae. This is a unique arrangement, found nowhere else in Tortoises. The pubic and ischiadic symphyses are connected with each other by ligaments only.

The general appearance of this water-tortoise is rather curious, since the carapace is much depressed, looking, especially in younger specimens, as if it had been crushed in. The head, provided with very strong hooked jaws, is strikingly heavy and large, and is covered above with one single large shield. The tail is longer than the shell, which, in full-grown specimens, reaches about six inches in length; it is, throughout its length, covered with rings of squarish shields. A large specimen measures 14 inches in total length, of which only five fall to the shell.

Fam. 5. Testudinidae.–The shell is always covered with well-developed horny shields. Those which form the plastral bridge are in direct contact with the marginals. The plastron is composed of nine bones. The digits have four or five claws. The neck is completely retractile. The skull is devoid of parieto-squamosal arches.

This large family is cosmopolitan, with the exception of the Australian and the adjoining Austro-Malayan countries. It contains genera which form a continuous gradation between absolutely terrestrial and thoroughly aquatic tortoises; and many are truly amphibious. As a general rule the typically terrestrial kinds have a more curved or arched shell, the digits are short, the eggs are more oval or round, and they are chiefly herbivorous; the essentially aquatic kinds have a flatter or depressed shell, webbed feet, with longer, often slender claws, the eggs are more cylindrical, and they live on animal diet. About 20 genera, with more than 110 species, are recognised by Boulenger, but their essential characters are nearly all internal, and therefore of no avail for the determination of live or entire specimens.

Chrysemys.–One of the most typical and widely distributed genera of American Terrapins or water-tortoises. The carapace is flat; the plastron is quite immovable, with a strongly developed bridge. Feet well webbed. Tail short. Skull with a broad, complete, lateral, temporal arch. About one dozen species, mostly in the eastern half of the United States, but the whole genus ranges from Canada to Argentina.

Most of the young Chrysemys are very pretty, the ground-colour of the upper shields being green, variegated with yellowish-brown or blackish markings, which often form exquisitely delicate patterns, either concentrical (Ch. concinna, Ch. rubriventris), or more longitudinal (Ch. elegans), or apparently quite irregular. The ground-colour of the plastron is yellow, but the various species are best distinguished, at least in very young individuals, by the arrangement of the dark brown spots and patches. There are, for instance, several pairs of bold lateral and several median patches in Ch. rubriventris; five pairs of ocellated spots in Ch. elegans; only small median patches, where four plastral shields meet, in Ch. concinna; while the plastron of Ch. picta is uniformly yellow.

These water-tortoises are very lively and shy, most so perhaps Ch. picta, which is very quick and active. The food varies, often according to individual fancy. Most of them eat fish. Ch. picta is partial to insects, but it also takes worms. Some of my specimens refused meat for a long time, but ultimately they became so fond of it and of worms, that they came out of the pond to take the food from the fingers; those in the Zoological Gardens of London have developed a taste for biscuits. One of my largest Ch. concinna fasted deliberately for eight months, refusing worms, insects, meat, and frogs, only occasionally sniffing at the food, until it was tempted with whitebait, which it took greedily. It refused, however, smelts and pieces of soles, but after another month it condescended to take meat regularly. Very young individuals live chiefly on flies, which they watch for near the surface of the water; and they are fond of smooth caterpillars, maggots, the larvae of humble-bees, and similar soft creatures. They all spend most of their time in the water, preferably floating near the surface, hidden between weeds; and they are fond of basking. Some of them spend the night in the water, lying motionless on the bottom, with heads and limbs turned in. Others prefer hiding under moss. Those species, which, like Ch. concinna and Ch. picta, are common in the North, are of course perfectly hardy. For the winter they dig themselves holes in the banks near the water, and they do not come out again until the spring is well advanced. The eggs are hard-shelled, mostly long and oval, and they are hatched before the end of the summer. The larger species of Terrapin are eaten.

Ch. picta (Fig. 76), the "Painted Terrapin," of the Eastern United States, e.g. of New York and Long Island, is easily recognised by the much depressed shell, which is absolutely smooth, and without a trace of a keel. The colour above is dark olive-brown or blackish, with broad yellow bands across the anterior ends of the neural and costal shields. Three or four of these transverse bands are very conspicuous. The marginals are red, with more or less concentric black and yellow markings. The pretty red colour, with some black stripes, extends over the bridge, but the plastron itself is uniformly yellow. The soft parts are likewise prettily marked, the ground-colour is black-brown, with delicate bright yellow and red stripes on the sides of the neck, limbs, and tail. The stripes are originally yellow, but they develop an orange or red line in the middle, so that each red stripe is ultimately narrowly edged with yellow; or the yellow and red stripes alternate, for instance on the tail, which is short, narrow, and pointed. The head is further adorned with a pair of conspicuous bright yellow patches behind the eyes, and a smaller pair on the occiput. The black and yellow stripes run across the gape of the mouth, some of the lines even looking as if they had been painted across. The nuchal shield is elongated and very narrow, its anterior edge and that of the neighbouring marginals are finely serrated. Very young individuals are at once recognised by the prominent longitudinal median stripe of bright orange extending over the nuchal and neural shields; the yellow transverse bands are still absent; they appear when the longitudinal line vanishes.