The most important and useful of these predictions involves a paradox under older Darwinian views. If selection controls evolutionary rate, one might think that the fastest tempos of alteration would be associated with the strongest selective pressures for change. Speed of change should vary directly with intensity of selection. Neutral theory predicts precisely the opposite—for an obvious reason once you start thinking about it. The most rapid change should be associated with unconstrained randomness—following the old thermodynamic imperative that things will invariably go to hell unless you struggle actively to maintain them as they are. After all, stability is far more common than change at any moment in the history of life. In its ordinary everyday mode, natural selection must struggle to preserve working combinations against a constant input of deleterious mutations. In other words, natural selection, in our technical parlance, must usually be “purifying” or “stabilizing.” Positive selection for change must be a much rarer event than watchdog selection for tossing out harmful variants and preserving what works.
Now, if mutations are neutral, then the watchdog sees nothing and evolutionary change can proceed at its maximal tempo—the neutral rate of substitution. But if a molecule is being preserved by selection, then the watchdog inhibits evolutionary change. This originally counterintuitive proposal may be regarded as the key statement of neutral theory. Kimura emphasizes the point with italics in most of his general papers, writing for example (in 1982): “Those molecular changes that are less likely to be subjected to natural selection occur more rapidly in evolution.”
Both the greatest success, and the greatest modification, of Kimura’s original theory have occurred by applying this principle that selection slows the maximal rate of neutral molecular change. For modification of the original theory, thousands of empirical studies have now shown that watchdog selection, measured by diminished tempo of change relative to predictions of randomness, operates at a far higher relative frequency than Kimura’s initial version of neutralist theory had anticipated. For success, the firm establishment of the principle itself must rank as the greatest triumph of neutralism—for the tie of maximal rate to randomness (rather than to the opposite expectation of intense selection) does show that neutralism exerts a kind of base-level control over evolution as a whole.
The most impressive evidence for neutralism as a maximal rate has been provided by forms of DNA that make nothing of potential selective value (or detriment) to an organism. In all these cases, measured tempos of molecular change are maximal, thus affirming the major prediction of neutralism.
1. Synonymous substitutions. The genetic code is redundant in the third position. A sequence of three nucleotides in DNA codes for an amino acid. Change in either of the first two nucleotides alters the amino acid produced, but most changes in the third nucleotide—so-called synonymous substitutions—do not alter the resulting amino acid. Since natural selection works on features of organisms, in this case proteins built by DNA and not directly on the DNA itself, synonymous substitutions should be invisible to selection, and therefore neutral. Rates of change at the third position are usually five or more times as rapid as changes at the functional first and second positions—a striking confirmation of neutralism.
2. Introns. Genes come in pieces, with functional regions (called exons) interrupted by DNA sequences (called introns) that are snipped out and not translated into proteins. Introns change at a much higher evolutionary rate than exons.
3. Pseudogenes. Certain kinds of mutations can extinguish the function of a gene—for example, by preventing its eventual translation into protein. These so-called pseudogenes begin with nearly the same DNA sequence as the functional version of the gene in closely related species. Yet, being entirely free from function, these pseudogenes should exert no resistance against the maximal accumulation of changes by random drift. Pseudogenes become a kind of ultimate test for the proposition that absence of selection promotes maximal change at the neutral rate—and the test has, so far, been passed with distinction. In pseudogenes, rates of change are equal, and maximal, at all three positions of the triplet code, not only at the third site, as in functional genes.
I was inspired to write about neutral theory by a fascinating example of the value of this framework in assessing the causes of evolutionary rates. This example neither supports nor denies neutralism but forms a case in the middle, enlightened by the more important principle that random models provide simple and explicit criteria for judgment.
While supposedly more intelligent mammals are screwing up royally above ground, Near Eastern mole rats of the species Spalax ehrenbergi are prospering underneath. Subterranean mammals usually evolve reduced or weakened eyes, but Spalax has reached an extreme state of true blindness. Rudimentary eyes are still generated in embryology, but they are covered by thick skins and hair. When exposed to powerful flashes of light, Spalax shows no neurological response at all, as measured by electrodes implanted in the brain. The animal is completely blind.
What then shall we make of the invisible and rudimentary eye? Is this buried eye now completely without function, a true vestige on a path of further reduction to final disappearance? Or does the eye perform some other service not related to vision? Or perhaps the eye has no direct use, but must still be generated as a prerequisite in an embryological pathway leading to other functional features. How can we decide among these and other alternatives? The random models of neutral theory provide our most powerful method. If the rudimentary eye is a true vestige, then its proteins should be changing at the maximal neutral rate. If selection has not been relaxed, and the eye still functions in full force (though not for vision), then rates of change should be comparable to those for other rodents with conventional eyes. If selection has been relaxed due to blindness, but the eye still functions in some less constrained way, then an intermediate rate of change might be observed.
The eye of S. ehrenbergi still builds a lens (though the shape is irregular and cannot focus an image), and the lens includes a protein, called A-crystallin. The gene for this protein has recently been sequenced and compared with the corresponding gene in nine other rodents with normal vision (see article by W. Hendriks, J. Leunissen, E. Nevo, H. Bloemendal, and W. W. de Jong in bibliography).
Hendriks and colleagues obtained the most interesting of possible results from their study. The A-crystallin gene is changing much faster in blind Spalax than in other rodents with vision, as relaxation of selection due to loss of primary function would suggest. The protein coded by the Spalax gene, for example, has undergone nine amino acid replacements (of 173 possible changes), compared with the ancestral state for its group (the murine rodents, including rats, mice, and hamsters). All other murines in the study (rat, mouse, hamster, and gerbil) have identical sequences with no change at all from the ancestral state. The average tempo of change in A-crystallin among vertebrates as a whole has been measured at about 3 amino acid replacements per 100 positions per 100 million years. Spalax is changing more than four times as fast, at about 13 percent per 100 million years. (Nine changes in 173 positions is 5.2 percent; but the Spalax lineage is only 40 million years old—and 5.2 percent in 40 million corresponds to 13 percent in 100 million years.) Moreover, Spalax has changed four amino acids at positions that are absolutely constant in all other vertebrates studied—seventy-two species ranging from dogfish sharks to humans.
“These findings,” Hendriks and colleagues conclude, “all clearly indicate an increased tolerance for change in the primary structure of A-crystallin in this blind animal.” So far so good. But the increased tempo of change in Spalax, though marked, still reaches only about 20 percent of the characteristic rate for pseudogenes, our best standard for the maximal, truly neutral pace of evolution. Thus, Spalax must still be doing enough with its eyes to damp the rate of change below the maximum for neutrality. Simple models of randomness have taught us something interesting and important by setting a testable standard, approached but not met in this cas
e, and acting as a primary criterion for judgment.
What then is A-crystallin doing for Spalax? What can a rudimentary and irregular lens, buried under skin and hair, accomplish? We do not know, but the established intermediate rate of change leads us to ask the right questions in our search for resolution.
Spalax is blind, but this rodent still responds to changes in photoperiod (differing lengths of daylight and darkness)—and apparently through direct influence of light regimes themselves, not by an indirect consequence that a blind animal might easily recognize (increase in temperature due to more daylight hours, for example). A. Haim, G. Heth, H. Pratt, and E. Nevo (see bibliography) showed that Spalax would increase its tolerance for cold weather when exposed to a winterlike light regime of eight light followed by sixteen dark hours. These mole rats were kept at the relatively warm temperature of 22°C, and were therefore not adjusting to winter based on clues provided by temperature. Animals exposed to twelve light and twelve dark hours at the same temperature did not improve their thermoregulation as well. Interestingly, animals exposed to summerlike light regimes (sixteen light and eight dark), but at colder temperatures of 17°C, actually decreased their cold-weather tolerance. Thus, even though blind, Spalax is apparently using light, not temperature, as a guide for adjusting physiology to the cycle of seasons.
Hendriks and colleagues suggest a possible explanation, not yet tested. We know that many vertebrates respond to changes in photoperiod by secreting a hormone called melatonin in the pineal gland. The pineal responds to light on the basis of photic information transmitted via the retina. Spalax forms a retina in its rudimentary eye, yet how can the retina, which perceives no light in this blind mammal, act in concert with the pineal gland? But Hendriks and colleagues note that the retina can also secrete melatonin itself—and that the retina of Spalax includes the secreting layer. Perhaps the retina of Spalax is still functional as a source of melatonin or as a trigger of the pineal by some mechanism still unknown. (I leave aside the fascinating, and completely unresolved, issue of how a blind animal can respond, as Spalax clearly does, to seasonal changes in photoperiod.)
If we accept the possibility that Spalax may need and use its retina (in some nonvisual way) for adaptation to changing seasons, then a potential function for the lens, and for the A-crystallin protein, may be sought in developmental pathways, not in direct utility. The lens cannot work in vision, and A-crystallin focuses no image, but the retina does not form in isolation and can only be generated as part of a normal embryological pathway that includes the prior differentiation of other structures. The formation of a lens vesicle may be a prerequisite to the construction of a retina—and a functioning retina may therefore require a lens, even if the lens will be used for nothing on its own.
Evolution is strongly constrained by the conservative nature of embryological programs. Nothing in biology is more wondrously complex than the production of an adult vertebrate from a single fertilized ovum. Nothing much can be changed very radically without discombobulating the embryo. The intermediate rate of change in lens proteins of a blind rodent—a tempo so neatly between the maximal pace for neutral change and the much slower alteration of functioning parts—may point to a feature that has lost its own direct utility but must still form as a prerequisite to later, and functional, features in embryology.
Our world works on different levels, but we are conceptually chained to our own surroundings, however parochial the view. We are organisms and tend to see the world of selection and adaptation as expressed in the good design of wings, legs, and brains. But randomness may predominate in the world of genes—and we might interpret the universe very differently if our primary vantage point resided at this lower level. We might then see a world of largely independent items, drifting in and out by the luck of the draw—but with little islands dotted about here and there, where selection reins in tempo and embryology ties things together. What, then, is the different order of a world still larger than ourselves? If we missed the different world of genic neutrality because we are too big, then what are we not seeing because we are too small? We are like genes in some larger world of change among species in the vastness of geological time. What are we missing in trying to read this world by the inappropriate scale of our small bodies and minuscule lifetimes?
8 | Reversals—Fragments of a Book Not Written
29 | Shields of Expectation—and Actuality
DISCOVERY, like its soul mate love, is a many-splendored thing. Stumbling serendipity surrounds some great finds—like Archaeopteryx, the first bird, unearthed by a quarryman at Solnhofen. Others are the product of dogged purpose. Consider Eugène Dubois who, as a Dutch army surgeon, posted himself to Indonesia because he felt sure that human ancestors must have inhabited East Asia (see Essay 8). There he found, in 1893, the first human fossils of a species older than our own—the Trinil femur and skull cap of Homo erectus (“Java man” of the old texts).
The most beautiful specimens in my office, which I happily share with about 50,000 fossil arthropods, rest in the last cabinet of the farthest corner. They are head shields of Eurypterus fischeri, a large, extinct freshwater arthropod related to horseshoe crabs. These exquisite fossils are preserved as brown films of chitin, set off like an old rotogravure against a surrounding sediment so fine in grain that the background becomes a uniform sheet of gray (see figure). They were collected in Estonia by William Patten, a professor of biology at Dartmouth.
When I first came to Harvard twenty years ago, I made a reconnaissance of all our 15,000 drawers of fossils—an adventure surely surpassing anything ever achieved by the smallest boy in the largest candy store. I found some of the great specimens of my profession—Agassiz’s echinoderms, Raymond’s collection from the Burgess Shale. But I got a particular thrill from Patten’s eurypterids because I knew exactly why he had gathered them. Patten, like Dubois, had collected with a singular purpose. I had read his 1912 book—The Evolution of the Vertebrates and Their Kin—one of the curiosities of my profession. Patten’s book represents the last serious defense of the classic, though incorrect, theory for vertebrate origins—the attempt to link the two great phyla of complex animals by arguing that vertebrates arose from arthropods.
A head shield of Eurypterus fischeri collected by William Patten in Estonia. Photograph by Rosamond W. Purcell.
Patten identified eurypterids as the arthropod ancestors of vertebrates—hence his strong desire to collect them. But Patten was even more interested in a group that occurred with the eurypterids in some localities—jawless fishes of the genus Cephalaspis (meaning head shield). We now recognize these jawless fishes (class Agnatha) as the oldest vertebrates and precursors of all later forms, ourselves included. The Agnatha survive today as a small remnant of naked eel-shaped forms—the lampreys (genus Petromyzon) and the distantly related hagfishes (genus Myxine). But the original armored agnathans, popularly called ostracoderms (shell skinned), dominated vertebrate life for its first hundred million years and included a large array of diverse forms. Patten’s fascination with ostracoderms arose from his misinterpretation of their anatomy. Patten viewed Cephalaspis and its relatives as intermediary forms between arthropods and true fishes.
We usually tell the history of a profession as a pageant of changing ideas and their proponents. But we can also render a different and equally interesting account from the standpoint of objects studied. One could provide a fascinating history of astronomy from the moon’s point of view, and genetics receives a different, multifaceted account through the eyes of a fruit fly. Cephalaspis may be our best standard bearer for evolution.
The history of ideas about Cephalaspis—from its original misinterpretation as the head of a trilobite in the early 1800s to its present status as the archetypal ostracoderm for all aficionados of the group—provides more than a synopsis of evolutionary thinking. It also illustrates, in an unusually forceful way, the fundamental process of scientific discovery itself.
Popular misunderstandin
g of science and its history centers upon the vexatious notion of scientific progress—a concept embraced by all practitioners and boosters, but assailed, or at least mistrusted, by those suspicious of science and its power to improve our lives. The enemy of resolution, here as nearly always, is that old devil Dichotomy. We take a subtle and interesting issue, with a real resolution embracing aspects of all basic positions, and we divide ourselves into two holy armies, each with a brightly colored cardboard mythology as a flag of struggle.
The cardboard banner of scientific boosterism is an extreme form of realism, the notion that science progresses because it discovers more and more about an objective, material reality out there in the universe. The extreme version holds that science is an utterly objective enterprise (and therefore superior to other human activities); that scientists read reality directly by invoking the scientific method to free their minds of cultural superstition; and that the history of science is a march toward Truth, mediated by increasing knowledge of the external world.
The cardboard banner of the opposition is an equally extreme form of relativism, the idea that truth has no objective meaning and can only be assessed by the variable standards of different communities and cultures. The extreme version holds that scientific consensus is no different from any other arbitrary set of social conventions, say the rules for Chinese handball set by my old crowd on 63d Avenue. Science is ideology, and scientific “progress” is no improving map of external reality, but only a derivative expression of cultural change.
These positions are so sharply defined that they can only elicit howls of disbelief from the opposition. How can relativists deny that science discovers external truth? say the realists. Cro-Magnon people could draw a horse as beautifully as any artist now alive, but they could not resolve the structure of DNA or photograph the moons of Uranus. How, reply the relativists, can anyone deny the social character of science when Darwin needed Adam Smith more than Galápagos tortoises and when Linnaeus matched his taxonomy to prevailing views of divine order?