Biology_of_Turtles

3.1Origins

The body form of a turtle, with the shoulder girdles enclosed between the ribs and the plastron, is so peculiar that it could easily be used to justify a fundamental dichotomous division of vertebrates into “turtles” (including tortoises, softshells, and so on) and “non-turtles” (giraffes, lampreys, salamanders, vultures, and so on). Agassiz (1857) was one of the earliest to contemplate this strange body plan, but as a pre-evolutionist his explanation bordered upon the mystical:

There is hardly any other type in the whole animal kingdom, in which the direct intervention of thought, as the first cause of its characteristic features, can be so fully and so easily illustrated as in the order of Testudinata. In the first place, these animals are so peculiar in their form and in their structure, that they strike, at first sight, every observer as belonging almost to another creation. They have been represented as inverted Vertebrata; and the peculiarity in the position and connection of their limbs has been so magnified, even to the rank of a class character, that very special conditions would seem necessary for their existence; and yet they are so extensively scattered upon the whole surface of the globe, among other animals of entirely different form and structure, upon land, in the fresh waters, and in the ocean, that, unless it can be shown that, besides its known properties, matter possesses also a turtle-making property, it must be granted that there are special thoughts expressed both in their structure and in their forms, and that the plan to which they belong, notwithstanding their striking differences, must have been designed and executed by a thinking being.

Ironically, in addition to his profound interest in turtles Agassiz was also an exceptional embryologist. His engravings of turtle embryos at all stages of development are among the best ever, and his zeal in acquiring turtle eggs at the earliest stages of development is the subject of a brilliant essay, “Turtle Eggs for Agassiz” (Sharp, 1916). Ontogeny, we hope, repeats phylogeny. Twentiethcentury authors (Ruckes, 1929; Walker, 1947) have, like Agassiz, looked to ontogeny as revealed by embryology to solve the riddles of turtle evolution. As Zangerl (1969) wrote:

Ruckes … showed that a great proliferation develops in the mid-dorsal region of the embryonic body, and that this Anlage of the dorsal disc spreads rapidly in a peripheral direction. The cartilaginous ribs do not remain beneath the dermis, but enter and grow longitudinally within the dorsal disc. The rate of growth of this system notably exceeds the rate of growth of the rest of the embryo with the result that the girdle elements (which remain relatively stationary—an observation confirmed by Walker, 1947)—become dorsally and laterally overgrown.

It would be satisfying if a range of really early chelonian fossils were available to confirm and document this transition but, alas, missing links remain missing. The fossil record of turtles is astonishing in its complexity and in the frequency with which genuinely novel offshoots are discovered, even within the relatively conservative pleurodire suborder (Gaffney et al., 2006). But the earliest turtles—the Proganochelyidae of the Triassic—were fully formed, differing from certain extant types mostly in technical osteologic details, and their distribution, which included what are now Germany, Greenland, and Thailand, suggests that turtles attained a worldwide distribution very early in their evolution. As to how they got there, both morphologically and geographically, Zangerl once again sums up the situation succinctly: “The fossil record, at present, provides no clues.”

If a living Proganochelys were to cross the path of the average Homo sapiens, the man would dismiss the animal as a slightly odd-looking snapping turtle; the giants, the truly weird species, and other instantly noteworthy chelonians of the past—Stupendemys, Drazinderetes, Meiolania,

Archelon, Sinemys gamera, Kallokibotion, Nahnsiungchelys—came much later. The heavily sculptured and tuberculate dorsal scutes, the long tail, and the serrate carapace margin of Proganochelys would not look unfamiliar to those on the Suwannee or the Apalachicola who had seen Macrochelys. Already in this earliest known turtle, the carapace scutes included a vertebral series, a set

articulation, and lack of bony walls to the middle ear (Gaffney, 1986). The slightly later Kayentachelys (early Jurassic of Arizona, U.S.A.) had lost the basipterygoid articulation but retained the palatal teeth typical of Proganochelys. Kayentachelys is identifiable as a cryptodire by the typically cryptodiran trochlear device allowing the jaw musculature to slide over the top of the otic chamber (Gaffney et al., 1987). The early existence of the pleurodire lineage is exemplified by the very distinctive Araripemys (Early Cretaceous of the Santana formation, Ceará, Brazil), in which the long neck and flattened head suggest adaptation for piscivory. Gaffney et al. (2006) discuss anatomical details.

Once the early turtle ancestors had developed the shell, with both carapace and plastron primarily for purposes of passive defense, this elaborate, literally all-embracing structure turned out to be invaluable for a great many other functions ranging from crypsis (many turtles having highly disruptive carapace patterns and others having uniform earth tones) to thermoregulation, the favorable surface/volume ratio favoring thermal stability, especially in the larger, terrestrial species. The shell also provided an encasement for the voluminous viscera, and especially the bulky digestive tract characteristic of grazing species, notably tortoises. Furthermore, the presence of the shell provided some surprising incidental biochemical advantages. These include the function of providing a buffer for lactic acid during periods of anoxia, permitting protracted survival without access to oxygen, as may occur during hibernation underwater for several months at temperature around 3°C. The bone in the shell may constitute as much as 32% of the body mass of a small turtle such as Chrysemys picta, with the rest of the skeleton being an addition 5.5%—together comprising more than three times the percentage found in a small mammal or non-chelonian reptile. Over 99% of the total body calcium, magnesium, and phosphate, over 95% of the carbon dioxide, and over 60% of the body’s sodium may reside in the shell and bones (Jackson, 2000). The buffering occurs in two ways: by releasing calcium and magnesium carbonates and by storing and buffering lactic acid within the bone of the shell.

3.2Standard Configuration and Variant Forms

3.2.1Bones

Among extant turtle species, certain aspects of the bony shell mosaic are very conservative whereas others may show extreme variation. The plastron shows less variation in its components than the carapace, although just as much or more variation in shape, and most turtles have four pairs of plastral bones, evolutionarily derived from the clavicles, interclavicle, and the abdominal ribs or gastralia, and perhaps some new bones. They are the epiplastra, hyoplastra, hypoplastra, and xiphiplastra, together with an anterior entoplastron. One extant family (the Pelomedusidae) includes an additional pair of mid-plastral bones, the mesoplastra. These are wide and together span the whole width of the plastron in Pelusios (Figure 3.2), whereas in Podocnemis they are laterally placed and widely separated (Figure 3.2). Such mesoplastra are also often present in the testudinid cryptodire Malacochersus (Figure 3.3). In Pelomedusa, they are separated by the large mid-plastral fontanel. Irregular bones present between the hyoplastra and the hypoplastra of many specimens of Manouria may also be called mesoplastra, and Pritchard (1966) records their anomalous presence in a specimen of Lepidochelys olivacea. Other variant plastral conditions are described later, under Kinosternidae, Trionychidae, Chelydridae, and so on.

The carapace includes a large anterior element, the nuchal, present in all turtles, even in Dermochelys. This important bone is the origin of numerous muscles and usually has a sutural connection with the first pair of marginals, the first costals, and the first neural bone. Posterior to the nuchal is a series of neural bones. In the typical condition, each neural bone alternates with the underlying, attached vertebral centrum, and the rib heads lie approximately halfway along the length of the corresponding neural bone. However, throughout the Testudines there is an astonishing amount of variation:

Figure 3.3  Plastron of Malacochersus tornieri showing mesoplastra. Right: Plastron of Podocnemis erythrocephala with abnormal absence of mesoplastra.

The presumed ancestral condition of about eight predominantly hexagonal neurals, each of which has the broad end anteriorly directed, may be modified in many ways—by proliferation of elements (to as many as 15), as in Lepidochelys; by reduction (usually at the ends of the series) as in many kinosternids and chelids; by total loss of exposed neurals, as in most Chelodina and Emydura; and by changing in shape—by broadening, as in Deirochelys, Macrochelys, etc.; narrowing as in Carettochelys; reduction to isolated kite-shaped elements as in many Phrynops gibbus and Cycloderma senegalensis; reversal (to hexagons with the broad end posteriorly directed as in Rhinoclemmys and various other smaller batagurids, or to an alternating series of octagonal and quadrilateral elements in most testudinids (Pritchard, 1988).

The pygal bone is usually single but the suprapygals are very variable, the principal variants being complete absence (Podocnemis lewyana, most trionychids); single, as in the other Podocnemis species; or paired, as in the majority of turtles, with an anterior elongate element and a posterior wider one. The suture between the two is often strongly curved, as in many testudinids and cheloniids. In extreme cases, the anterior element develops posterolateral arms that embrace the posterior element (Figure 3.4), although this condition is not fully developed until maturity is approached (Auffenberg, 1976).

The costal bones number eight pairs in almost all turtles, each costal bearing a rib and corresponding to vertebral II to vertebral IX. Dorsal vertebra 1 thus usually lacks both ribs and costal bones, but Kordikova (2000) reported a specimen of Pelodiscus sinensis (PCHP 2771) in which these elements were present, giving a total of nine pairs of costals, and Pritchard (1993) reported a Dogania (PCHP 3368) with the same condition on one side. A similar abnormal individual of P. sinensis with nine pairs of costals was the type specimen of Heude’s Tortisternum novemcostatum, and one assumes that his Tortisternum novemcostatum also should have had nine costals, although his illustration showed only eight (Heude, 1880). In Rafetus, the eighth costals are strongly reduced (Figure 3.5), and this reduction is further advanced in American trionychids, in which the eighth costals are either reduced (Figure 3.6) or completely missing (or sometimes present on one side only). The peripheral bones are also relatively stable, generally numbering 11 pairs, although they

Figure 3.7  Carapace of Callagur borneoensis (right) and Emydura australis (left) showing extensive inguinal buttressing.

are reduced to 10 pairs in kinosternids and carettochelyids and they disappear completely in dermochelyids and trionychids.

Buttressing between carapace and plastron, as already evident to White in 1774, is very variable among turtles. Extensive shell buttressing is shown by Callagur, Batagur, Kachuga, and also certain sidenecks (e.g., Emydura australis), probably associated with defense against predators with powerful jaws, especially crocodilians (Figure 3.7). At the other extreme, buttressing is eliminated in species with hinged plastra, although a remnant of the axillary buttresses not contacting the inner surface of the carapace may be present in Pelusios, where they serve as a pair of fixed levers for the muscles that elevate the kinetic anterior plastral lobe (Figure 3.8). In some batagurids in which the posterior plastral lobe shows minor kinesis, the inguinal buttress is not completely eliminated but is shortened and connects to the inner surface of the carapace by means of a pad of fibrous tissue rather than an immovable suture. Buttresses are absent in chelydrids, cheloniids, trionychids, and so on.

3.2.2Scutes

Whereas some turtle families (Dermochelyidae, Trionychidae, Carettochelyidae) have dispensed entirely with scutes (large external scales that cover the carapace and plastron), they are retained by the vast majority of turtles. Scutes may be remarkably thick (Eretmochelys, Astrochelys yniphora) or may be paper-thin (Dermatemys, Callagur). Usually, they are juxtaposed but they are imbricate (overlapping) in most Eretmochelys.

Scutes grow by proliferation of keratin around the periphery (as well as the entire undersurface of the scute) as the turtle itself grows, and it is common for this growth to be seasonal, or at least episodic, so that rings or bands may be formed on each scute throughout the area of post-hatchling growth. Typically, the outline of each scute of the hatchling is preserved as an “areola,” a small, stippled or shagreened area somewhere near the center of the scute. The growth rings or annuli around the aureolae may often be counted to give an index of the age of the animal (Figure 3.9), although this is reliable only in younger specimens and those that live in temperate latitudes, where winter dormancy occurs. Also, some tropical species (e.g., in the llanos of Venezuela) develop annuli that