Turtles feed and breed on separate grounds for good reasons. They feed on sea grasses in protected, shallow-water pastures, but breed on exposed shores where sandy beaches develop—preferably, on islands where predators are rare. But why travel 2,000 miles to the middle of an ocean when other, apparently appropriate breeding grounds are so much nearer? (Another large population of the same species breeds on the Caribbean coast of Costa Rica.) As Carr writes: “The difficulties facing such a voyage would seem insurmountable if it were not so clear that the turtles are somehow surmounting them.”
Perhaps, Carr reasoned, this odyssey is a peculiar extension of something much more sensible, a journey to an island in the middle of the Atlantic, when the Atlantic was little more than a puddle between two continents recently separated. South America and Africa parted company some 80 million years ago, when ancestors of the genus Chelonia were already present in the area. Ascension is an island associated with the Mid-Atlantic Ridge, a linear belt where new sea floor wells up from the earth’s interior. This upwelling material often piles itself high enough to form islands.
Iceland is the largest modern island formed by the Mid-Atlantic Ridge; Ascension is a smaller version of the same process. After islands form on one side of a ridge, they are pushed away by new material welling up and spreading out. Thus, islands tend to be older as we move farther and farther from a ridge. But they also tend to get smaller and finally to erode away into underwater seamounts, for their supply of new material dries up once they drift away from an active ridge. Unless preserved and built up by a shield of coral or other organisms, islands will eventually be eroded below sea level by waves. (They may also sink gradually from sight as they move downslope from an elevated ridge into the oceanic depths.)
Carr therefore proposed that the ancestors of Ascension green turtles swam a short distance from Brazil to a “proto-Ascension” on the late Cretaceous Mid-Atlantic Ridge. As this island moved out and sank, a new one formed at the ridge and the turtles ventured a bit farther. This process continued until, like the jogger who does a bit more each day and ends up a marathoner, turtles found themselves locked into a 2,000-mile journey. (This historical hypothesis does not deal with the other fascinating question of how the turtles can find this dot in a sea of blue. The hatchlings float to Brazil on the Equatorial Current, but how do they get back? Carr supposes that they begin their journey by celestial cues and finally home in by remembering the character [taste? smell?] of Ascension water when they detect the island’s wake.)
Carr’s hypothesis is an excellent example of using the peculiar to reconstruct history. I wish I could believe it. I am not troubled by the empirical difficulties, for these do not render the theory implausible. Can we be confident, for example, that a new island always arose in time to replace an old one—for the absence of an island for even one generation would disrupt the system. And would the new islands always arise sufficiently “on course” to be found? Ascension itself is less than seven million years old.
I am more bothered by a theoretical difficulty. If the entire species Chelonia mydas migrated to Ascension or, even better, if a group of related species made the journey, I would have no objection, for behavior can be as ancient and as heritable as form. But C. mydas lives and breeds throughout the world. The Ascension turtles represent only one among many breeding populations. Although its ancient ancestors may have lived in the Atlantic puddle 200 million years ago, our record of the genus Chelonia does not extend back beyond fifteen million years, while the species C. mydas is probably a good deal younger. (The fossil record, for all its faults, indicates that few vertebrate species survive for as many as ten million years.) In Carr’s scheme, the turtles that made the first trips to proto-Ascension were rather distant ancestors of C. mydas (in a different genus at least). Several events of speciation separate this Cretaceous ancestor from the modern green turtle. Now consider what must have happened if Carr is right. The ancestral species must have been divided into several breeding populations, only one of which went to proto-Ascension. This species then evolved to another and another through however many evolutionary steps separated it from C. mydas. At each step, the Ascension population kept its integrity, changing in lock step with other separate populations from species to species.
But evolution, so far as we know, doesn’t work this way. New species arise in small, isolated populations and spread out later. Separate subpopulations of a widely dispersed species do not evolve in parallel from one species to the next. If the subpopulations are separate breeding stocks, what is the chance that all would evolve in the same way and still be able to interbreed when they had changed enough to be called a new species? I assume that C. mydas, like most species, arose in a small area sometime within the last ten million years, when Africa and South America were not much closer together than they are today.
In 1965, before continental drift became fashionable, Carr proposed a different explanation that makes more sense to me because it derives the Ascension population after C. mydas evolved. He argued that ancestors of the Ascension population accidentally drifted on the Equatorial Current from west Africa to Ascension. (Carr points out that another turtle, the west African ridley, Lepidochelys olivacea, colonized the South American coast by this route.) The hatchlings then drifted to Brazil in the same east-to-west current. Of course, getting back to Ascension is the problem, but the mechanism of turtle migration is so mysterious that I see no barrier to supposing that turtles can be imprinted to remember the place of their birth without prior genetic information transmitted from previous generations.
I don’t think that the validation of continental drift is the only factor that caused Carr to change his mind. He implies that he favors his new theory because it preserves some basic styles of explanation generally preferred by scientists (incorrectly, in my iconoclastic opinion). By Carr’s new theory, the peculiar Ascension route evolved gradually, in a sensible and predictable fashion, step by step. In his former view, it is a sudden event, an accidental, unpredictable vagary of history. Evolutionists tend to be more comfortable with nonrandom, gradualistic theories. I think that this is a deep prejudice of Western philosophical traditions, not a reflection of nature’s ways (see essays of section 5). I regard Carr’s new theory as a daring hypothesis in support of a conventional philosophy. I suspect that it is wrong, but I applaud his ingenuity, his effort, and his method, for he follows the great historical principle of using the peculiar as a sign of change.
I am afraid that the turtles illustrate another aspect of historical science—this time a frustration, rather than a principle of explanation. Results rarely specify their causes unambiguously. If we have no direct evidence of fossils or human chronicles, if we are forced to infer a process only from its modern results, then we are usually stymied or reduced to speculation about probabilities. For many roads lead to almost any Rome.
This round goes to the turtles—and why not? While Portuguese sailors hugged the coast of Africa, Chelonia mydas swam straight for a dot in the ocean. While the world’s best scientists struggled for centuries to invent the tools of navigation, Chelonia looked at the skies and proceeded on course.
3 | Double Trouble
NATURE MARKS Izaak Walton as a rank amateur more often than I had imagined. In 1654, the world’s most famous fisherman before Ted Williams wrote of his favorite lure: “I have an artificial minnow…so curiously wrought, and so exactly dissembled that it would beguile any sharpsighted trout in a swift stream.”
An essay in my previous book, Ever Since Darwin, told the tale of Lampsilis, a freshwater clam with a decoy “fish” mounted on its rear end. This remarkable lure has a streamlined “body,” side flaps simulating fins and tail, and an eyespot for added effect; the flaps even undulate with a rhythmic motion that imitates swimming. This “fish,” constructed from a brood pouch (the body) and the clam’s outer skin (fin and tails), attracts the real item and permits a mother clam to shoot her larvae from the brood pouch towar
d an unsuspecting fish. Since the larvae of Lampsilis can only grow as parasites on a fish’s gill, this decoy is a useful device indeed.
I was astounded recently to learn that Lampsilis is not alone. Ichthyologists Ted Pietsch and David Grobecker recovered a single specimen of an amazing Philippine anglerfish, not as a reward for intrepid adventures in the wilds, but from that source of so much scientific novelty—the local aquarium retailer. (Recognition, rather than machismo, is often the basis of exotic discovery.) Anglerfish lure their dinner, rather than a free ride for their larvae. They carry a highly modified dorsal fin spine affixed to the tips of their snouts. At the end of this spine, they mount an appropriate lure. Some deep-sea species, living in a dark world untouched by light from the surface, fish with their own source of illumination: they gather phosphorescent bacteria in then lures. Shallow-water species tend to have colorful, bumpy bodies, and look remarkably like rocks encrusted with sponges and algae. They rest inert on the bottom and wave or wiggle their conspicuous lures near their mouths. “Baits” differ among species, but most resemble—often imperfectly—a variety of invertebrates, including worms and crustaceans.
Anglerfish
DAVID B. GROBECKER
Pietsch and Grobecker’s anglerfish, however, has evolved a fish lure every bit as impressive as the decoy mounted on Lampsilis’s rear—a first for anglerfish. (Their report bears as its appropriate title “The Compleat Angler” and cites as an epigraph the passage from Walton quoted above.) This exquisite fake also sports eyelike spots of pigment in the right place. In addition, it bears compressed filaments representing pectoral and pelvic fins along the bottom of the body, extensions from the back resembling dorsal and anal fins, and even an expanded rear projection looking for all the world like a tail. Pietsch and Grobecker conclude: “The bait is nearly an exact replica of a small fish that could easily belong to any of a number of percoid families common to the Philippine region.” The angler even ripples its bait through the water, “simulating the lateral undulations of a swimming fish.”
These nearly identical artifices of fish and clam might seem, at first glance, to seal the case for Darwinian evolution. If natural selection can do this twice, surely it can do anything. Yet—continuing the theme of the last two essays and bringing this trilogy to a close—perfection works as well for the creationist as the evolutionist. Did not the psalmist proclaim: “The heavens declare the glory of God; and the firmament showeth his handiwork.” The last two essays argued that imperfection carries the day for evolution. This one discusses the Darwinian response to perfection.
The only thing more difficult to explain than perfection is repeated perfection by very different animals. A fish on a clam’s rear end and another in front of an anglerfish’s nose—the first evolved from a brood pouch and outer skin, the second from a fin spine—more than doubles the trouble. I have no difficulty defending the origin of both “fishes” by evolution. A plausible series of intermediate stages can be identified for Lampsilis. The fact that anglerfish press a fin spine into service as a lure reflects the jury-rigged, parts-available principle that made the panda’s thumb and the orchid’s labellum speak so strongly for evolution (see the first essay of this trilogy). But Darwinians must do more than demonstrate evolution; they must defend the basic mechanism of random variation and natural selection as the primary cause of evolutionary change.
Anti-Darwinian evolutionists have always favored the repeated development of very similar adaptations in different lineages as an argument against the central Darwinian notion that evolution is unplanned and undirected. If different organisms converge upon the same solutions again and again, does this not indicate that certain directions of change are preset, not established by natural selection working on random variation? Should we not look upon the repeated form itself as a cause of the numerous evolutionary events leading toward it?
Throughout his last half-dozen books, for example, Arthur Koestler has been conducting a campaign against his own misunderstanding of Darwinism. He hopes to find some ordering force, constraining evolution to certain directions and overriding the influence of natural selection. Repeated evolution of excellent design in separate lineages is his bulwark. Again and again, he cites the “nearly identical skulls” of wolves and the “Tasmanian wolf.” (This marsupial carnivore looks like a wolf but is, by genealogy, more closely related to wombats, kangaroos, and koalas.) In Janus, his latest book, Koestler writes: “Even the evolution of a single species of wolf by random mutation plus selection presents, as we have seen, insurmountable difficulties. To duplicate this process independently on island and mainland would mean squaring a miracle.”
The Darwinian response involves both a denial and an explanation. First, the denial: it is emphatically not true that highly convergent forms are effectively identical. Louis Dollo, the great Belgian paleontologist who died in 1931, established a much misunderstood principle—“the irreversibility of evolution” (also known as Dollo’s law). Some ill-informed scientists think that Dollo advocated a mysterious directing force, driving evolution forward, never permitting a backward peek. And they rank him among the non-Darwinians who feel that natural selection cannot be the cause of nature’s order.
In fact, Dollo was a Darwinian interested in the subject of convergent evolution—the repeated development of similar adaptations in different lineages. Elementary probability theory, he argued, virtually guarantees that convergence can never yield anything close to perfect resemblance. Organisms cannot erase their past. Two lineages may develop remarkable, superficial similarities as adaptations to a common mode of life. But organisms contain so many complex and independent parts that the chance of all evolving twice toward exactly the same result is effectively nil. Evolution is irreversible; signs of ancestry are always preserved; convergence, however impressive, is always superficial.
Consider my candidate for the most astounding convergence of all: the ichthyosaur. This sea-going reptile with terrestrial ancestors converged so strongly on fishes that it actually evolved a dorsal fin and tail in just the right place and with just the right hydrological design. These structures are all the more remarkable because they evolved from nothing—the ancestral terrestrial reptile had no hump on its back or blade on its tail to serve as a precursor. Nonetheless, the ichthyosaur is no fish, either in general design or in intricate detail. (In ichthyosaurs, for example, the vertebral column runs through the lower tail blade; in fish with tail vertebrae, the column runs into the upper blade.) The ichthyosaur remains a reptile, from its lungs and surface breathing to its flippers made of modified leg bones, not fin rays.
Ichthyosaur
COURTESY OF THE AMERICAN MUSEUM OF NATURAL HISTORY
Koestler’s carnivores tell the same tale. Both placental wolf and marsupial “wolf” are well designed to hunt, but no expert would ever mistake their skulls. The numerous, small marks of marsupiality are not obliterated by convergence in outward form and function.
Second, the explanation: Darwinism is not the theory of capricious change that Koestler imagines. Random variation may be the raw material of change, but natural selection builds good design by rejecting most variants while accepting and accumulating the few that improve adaptation to local environments.
The basic reason for strong convergence, prosaic though it may seem, is simply that some ways of making a living impose exacting criteria of form and function upon any organism playing the role. Mammalian carnivores must run and stab; they do not need grinding molar teeth since they tear and swallow their food. Both placental and marsupial wolves are built for sustained running, have long, sharp, pointed canine teeth and reduced molars. Terrestrial vertebrates propel themselves with their limbs and may use their tails for balance. Swimming fish balance with their fins and propel from the rear with their tails. Ichthyosaurs, living like fish, evolved a broad propulsive tail (as whales did later—although the horizontal flukes of a whale’s tail beat up and down, while the vertical flukes offish and ic
hthyosaurs beat from side to side).
No one has treated this biological theme of repeated, exquisite design more eloquently than D’Arcy Wentworth Thompson in his 1942 treatise, On Growth and Form, still in print and still as relevant as ever. Sir Peter Medawar, a man who eschews hype and exaggeration, describes it as “beyond comparison the finest work of literature in all the annals of science that have been recorded in the English tongue.” Thompson, zoologist, mathematician, classical scholar, and prose stylist, won accolades as an old man but spent his entire professional life in a small Scottish university because his views were too unorthodox to win prestigious London and Oxbridge jobs.
Thompson was more a brilliant reactionary than a visionary. He took Pythagoras seriously and worked as a Greek geometrician. He took special delight in finding the abstract forms of an idealized world embodied again and again in the products of nature. Why do repeated hexagons appear in the cells of a honeycomb and in the interlocking plates of some turtle shells? Why do the spirals in a pine cone and a sunflower (and often of leaves on a stem) follow the Fibonacci series? (A system of spirals radiating from a common point can be viewed either as a set of left- or right-handed spirals. Left and right spirals are not equal in number, but represent two consecutive figures of the Fibonacci series. The Fibonacci series is constructed by adding the previous two numbers to form the next: 1, 1, 2, 3, 5, 8, 13, 21, etc. The pine cone may, for example, have 13 left spirals and 21 right spirals.) Why do so many snail shells, ram’s horns, and even the path of a moth to light follow a curve called the logarithmic spiral?
The Panda’s Thumb Page 3