Thinking about Turtles and Evolution

Part One: How the Turtle Gets its Shell

This introductory section consists of 5 topics. (1) An initial discussion of “fine tuning” explains that natural historians have long accepted that natural selection could cause superficial changes in lineages of organisms. (2) The question of species’ origins, however, is more profound. Here we explain that, anatomically, turtles are fundamentally different from other vertebrates; the evolution of turtles is not merely a matter of superficial tinkering with pre-existing body-plans. (3) In some cases, profound structural changes in plant or animal types are documented by an extensive fossil record. Such is not the case with turtles, which appear suddenly in the early Triassic Period. The more we learn about the origin of turtles, the better we will understand the important concept of evolution and punctuated equilibria. (4) Keys to the mysteries of turtle evolution should be sought in turtle embryogenesis. The basic question is, how do the shoulder and hip girdles end up inside of the turtle’s rib-cage? (5) Recent advances in genetics and developmental biology suggest how such a major anatomical re-arrangement (in the relative position of ribs and girdles) might result from the redeployment of ancient genetic “construction programs” to a new purpose.

1. Darwin’s easy sell: fine-tuning by natural selection. General. In the late 1850’s, when Charles Darwin began to get serious about presenting his views on the origin of species, he did not have to start from absolute scratch. Most Victorian intellectuals were well aware that selective breeding could, across multiple generations, lead to discernable differences in separate lineages of plants or animals. Therefore, Darwin could begin a public lecture by reminding his audience of how British dairy farmers had developed cows that gave more milk or how pigeon fanciers had established strains of birds that looked, sounded, and even behaved differently. Furthermore, many 19th century naturalists were willing to accept that populations of wild animals could, over generations, also adapt to their particular surroundings. Few thoughtful biologists were surprised, for example, that European vipers had darker colors in the colder portions of their range. And these scientists did not in general attribute such coloration to the direct, causal activity of a divine creator with a particular concern for heat absorption by small, venomous snakes.

Thinking turtles. Thus biologists operating in a pre-Darwinian paradigm would not have been surprised to learn that, within a given species of turtle, populations sharing habit with crocodilians (important turtle predators) have thicker, higher-domed shells than populations living in croc-free waters. This correlation has been formally demonstrated in North America and Africa. I suspect it is also true in Asia, Australia, and South America. Evolving a thicker shell hardly seems surprising. Evolving a shell at all—well, that’s a different matter.

2. The harder question: the origin of species. General. Darwin, of course, was not primarily interested in showing one more time that selection, natural or otherwise, could fine-tune the fit of a biotype with its environment. Rather, he wanted to argue that, through descent with heritable modification, selection could actually result in the origin of radically different biotypes. This radical idea was hardly even considered by most biologists. Instead, species were thought to be fundamental elements of the natural world. Sure, biotypes could be manicured a little, by nature or by human breeders, and perhaps the resulting individuals would fit in a bit better with their environment. But the basic types, the species, were immutable categories. A cow that gave more milk was still a cow; a pigeon that rolled was still a pigeon, and a river cooter with a thicker shell is still a river cooter.

Thinking turtles. Of all the vertebrate animals, none appears so fundamentally different, so morphologically isolated, as the familiar turtle. Every turtle, of course, sports a suit of protective armor. But armor in itself does not make the turtle so very different. After all, the number of armored fishes was once legion. The armored skeletons of dinosaurs populate dioramas in many of our better museums. Many lizards—and all crocodilians—have their scales underlain by an armor of dermal bone. The mammalian glyptodonts once carried shells as big as Volkswagens across the Lowcountry of Pleistocene South Carolina. And living mammals include more than a dozen armored species of armadillos and pangolins.

However, in two fundamental respects, turtles are different from these other armored creatures. (1) First, we should recognize that tissues we call “bone” can have two different embryological origins. The first sort of bone to evolve (in very ancient fish, more than 300 million years ago) formed by the organization of calcium and phosphorus within the deep layer of the skin called the dermis. This dermal bone (which persists in human beings as most of the facial bones in our skulls) entirely comprises the bony shell of every armored vertebrate—except turtles. Turtles, too, incorporate dermal bone within their shells, but the basic architecture of the carapace (upper shell) is made of something entirely different: endochondral bone. In vertebrates, endochondral bone is typically deep bone. It is bone that forms embryologically (or shortly after birth) within or around cartilage. It is the bone of the vertebrate skeleton—our backbones, our limb bones, our braincase, our ribs. And how it structures the outside of the turtle is something rather remarkable. (2) Second, although the turtle’s shell is entirely exterior, it is constructed of deep-body elements. This is obvious when one observes the interior of an empty turtle-carapace: ten vertebrae are incorporated into this upper shell, and, extending laterally from these vertebrae, also as parts of the shell, are pairs of ribs (see figure below). In other words, if one looks at an empty carapace and contemplates the entire turtle, one immediately realizes the deepest mystery of the turtle’s anatomical evolution: somehow the animal’s limb-girdles are inside the animal’s rib-cage!

This is no small thing. It is not like a cow that gives more milk or a pigeon that flies in peculiar patterns. It is a radical variation from the fundamental vertebrate body-plan. Indeed I think that the evolution of turtles from their shell-less, proto-reptilian ancestors is a critical test-case for Darwin’s theory on the origin of species. If evolution can explain the origin of turtles, then everything else should be pretty darn easy.

3. Fossils and the origin of species. General. In some cases an extensive fossil record precisely chronicles a series of changes leading from one biotype to another. This, however, is not the general case; indeed, transitions between fossil-bearing geological strata often reflect the “sudden” appearance of noticeably new organisms as well as the disappearance of older forms. (Biologists Stephen J. Gould and NilesEthridge philosophize extensively about this phenomenon, which they named “punctuated equilibrium.” This topic lies beyond the current scope of Biology 150, but Gould’s writings are quite interesting and certainly fall within the intellectual range of a serious college student.)

Thinking turtles. The absence of intermediate forms is particularly evident in the case of turtle evolution. Approximately 225 million years ago, well before the rise of the dinosaurs, an animal now called Proganochelys entered the fossil record. We need not concern ourselves with this varmint’s exact appearance or its poorly understood ecology, but you should understand two things about beast. First, if you saw the fossils, you would notice the shell and immediately recognize Proganochelys as a turtle. And, second, this early turtle has no obvious relationship to any previously living creature. As herpetologist Wayne VanDevender puts it, “Turtles enter the fossil record with great suddenness, like Athena springing full-blown from the head of Zeus.”

Turtles, then, are not just a quintessential test-case for the explanatory power of Darwinian theory. Their abrupt appearance, with complete shells, also makes them poster-children forevolution through punctuated equilibria. For more than a century post-Darwin, the evolutionary biology of turtles remained essentially unknown, and indeed many mysteries persist today. On the other hand, new understandings of genetics and embryology are shedding additional light on the biology of turtles. In fact, turtle-shell formation is a fundamental question of “Evo-Devo,” a new scientific discipline that is uniting genetics, evolution, and developmental biology.

4. The embryological development of the skeleton. General. In a typical terrestrial-vertebrate embryo, the pectoral girdle develops as a pair of cartilaginous half-rings that are superficial to (outside of) the transverse processes of the vertebrae. And, inside the developing girdle, the ribs extend deep and downward from the vertebrae to form a sort of bone-basket that will support the viscera. All these deep skeletal elements, initially laid down as cartilage, will, as the embryo develops, be replaced by bone.

Thinking turtles. The development of the limb-girdles in turtles basically follows the typical amniote pattern. But the ribs are captured by the spreading pre-skin cells and carried laterally (not downward), to the outside of the pectoral girdle. In this superficial location they will establish the skeletal structure of the carapace. Here’s roughly how that works.

All turtles lay eggs, and depending on the kind of turtle, these eggs develop at very different rates; overall, however, the pattern is pretty much the same across species. At laying, the embryo within a turtle egg is a hollow mass of cells that has begun to turn inward upon itself to form what will become three general layers of tissue. Within a few days after laying, the embryo will have a front end and a back end as well as a top side and a bottom side. This “geography” of the tiny organism will be mapped by sets of genes that will eventually instruct the embryo to build particular features in particular places. Among the first recognizable structures to appear are four limb-buds. These aggregations of cells are located at the front and rear quarters of the embryo; these will develop according to general patterns recognizable in other tetrapods such as chickens, mice, and people.

Shortly after the formation of limb-buds, the turtle embryo develops a structure that distinguishes it from all other vertebrates. This is called the carapacial ridge, and as the name would indicate, it is associated with the eventual formation of the upper shell. The carapacial ridge first appears as a slightly thickened line of cells running along each side of the embryo, just above the limb buds. Tissue-wise, it consists of ectoderm from the embryo’s dorsal side folded atop dermal mesoderm. In cross-section, the embryo looks like this (From Burke, Amer. Zool., 1991):

In early cross-sections (see above) the carapacial ridge looks like a structure that might “flow” down the sides of the turtle embryo, but its actual growth is more to the side, as the entire embryo flattens. And as this lateral growth accelerates, the cartilaginous rib-precursors (marked above by yellow arrows) grow laterally along with it. Meanwhile the limb-buds continue to grow, eventually producing rather conventional front and rear legs for our developing turtle. And as this occurs, the carapacial ridge, now in apparent control of ribs’ directional growth, spreads above those legs (and above the shoulder and hip girdles), eventually to form the turtle’s upper shell.

So that’s how the turtle-shell is formed outside the limb-girdles. The cartilaginous rib-precursors continue to expand laterally, all the way to what will become the edge of the shell. Then new cells called osteoblasts migrate from the embryo’s spinal region out and along the rib-precursors, eventually replacing the softer cartilage with harder bone. Typically this process is more or less complete by the time the turtle-egg hatches, and visualized by X-ray (to emphasize the bony structures) the hatchling turtle looks a bit like an elliptical wagon-wheel, with the spine as an elongated hub and the hardened ribs as the spokes (photo is of hatchling Trachemysscripta, by Scott Gilbert’s “Team Turtle” at Swarthmore University):

The anatomical emphasis of this essay is the relative position of the turtle’s rib-cage and limb-girdles. You have now learned the basic embryology of this strange contortion, and we shall presently speculate on how this alteration of vertebrate body-plan might have evolved. First, however, for completeness’ sake, we’ll briefly consider three other facets of shell-development. (1) The vertebrae, the ribs, and the margin of the carapace are endochondral bone, “deep bone,” formed by the replacement of initially cartilaginous tissues (see the photograph immediately above). The remainder of the carapace—the bone in between the ribs—is dermal bone, laid down by embryonic skin tissues, typically after the hatching of the little turtle. Herpetologists do not yet understand exactly how the skin-cells between the ribs “know” that they are supposed to build bones (while the skin-cells on the tail, legs, and feet, etc., do not); currently most researchers suspect that the cells within the recently-hardened ribs somehow signal the skin cells so that they express genes that lead to the construction of the bone. (2) In a healthy, living turtle the bones of the carapace are covered with other material. In the vast majority of species this covering consists of laminated plates calledscutes. Technically, scutes are scales, and like other reptile scales they are formed in the epidermis, the outer layer of the skin. Like our fingernails, turtle scutes are composed of a tough, waterproof, wear-resistant material called keratin. (In case you’re ever on Jeopardy, you might need to know that human fingernails are made of α-keratin while most turtle scutes are made of β-keratin, but for our purposes that’s a pretty trivial difference.) Scutes add strength to a turtle’s shell; because they are pigmented, they also comprise the patterns that help us distinguish between turtle species. (3) The formation of a turtle’s plastron, or under-shell, is a topic about which biologists disagree. One theory is that the bony plastron is formed by ossification of cells whose embryological origin is in the neural crest. This theory has aroused controversy because cells in the post-cranial neural crest are not otherwise known to form bone, and cells in the cranial neural crest (which, for example, form human facial bones) are not otherwise known to migrate posterior to the shoulders. Eventually we might suggest that a re-mapping of the turtle embryo could cause neural-crest cells to “think” they should move to the bottom side of the turtle embryo and produce bone. As with the carapace, bones of the plastron are covered in most turtle species by keratinaceousscutes.

5. Macro-evolutionary changes and tool-kit genes. As explained above, thoughtful pre-Darwinian biologists were willing to entertain the idea that modest changes in plant or animal biotypes could result from “selection” (either artificial or natural). Darwin’s principal difficulties therefore lay in persuading the scientific community that evolution could cause fundamental alterations of the biotypes themselves. As you probably know, Mr. Darwin’s recourse was to suggest that small changes, accumulated over vast amounts of time, could result in radical differences. This remains the basic, generally accepted rationale for macro-evolution. However, many evolutionists have not felt entirely comfortable with this explanation. Part of the problem lies in the Darwinian necessity that all forms intermediate between Biotype Alpha and Biotype Omega should work well—indeed, should work very well, should in general be “fitter” than their relatives not lying along the trajectory of evolutionary change. In this age of word processors, we can restate the old monkey-and-typewriter metaphor inflicted upon too many generations of high school students. The monkey starts withHamlet, and if she strikes enough computer keys, then as time stretches towards infinity, her random edits should at some point produceTo Kill a Mockingbird. That sounds believable, given enough random typing. On the other hand, to assume that an unbroken succession of intermediate drafts would be worth reading (would exhibit “literary fitness,” to mix academic languages)—well, that would seem to require a mega-eon of time and a blue bazillion computer-monkeys! But what if a random monkey-keystroke produced not a letter but an entire word? And what if, by some underlying rule of grammar, the random change of a word called forth a cascade of other words that would complete a coherent sentence, and what if other rules of grammar assembled sentences into paragraphs? Even then a population of computer-monkeys, starting with Hamlet, might notreproduce Harper Lee’s inspiring novel, but maybe they would eventually write something interesting.

NOTE: Biology 150 is not by any means a genetics course, but questions aboutHow the Turtle Gets its Shell are best addressed through the incorporation of evolutionary theory, developmental biology, and genetics. As background to this discussion, we need to understand, at least on a metaphorical level, the “grammar” of how the random mutation of one gene can elicit a cascade of effects that have coherent consequences for a developing organism. In the long boxed topic presented below, we focus on genetic “switches” to offer a simplistic explication of that grammar. We hope that you are already familiar with these matters—or that you can speed-read through our box-topic and move forward to our thinking-turtles section on toolkit genes.