Darwin’s argument is theoretically sound as an abstract resolution of a conceptual puzzle—one of the logical frustrations of my introduction. But have we any evidence that nature actually bows to the reason of our arguments? This issue is particularly important in evolutionary biology because we so often make the mistake of assuming that we understand the origin of a feature just because it now works so well. Consider, for example, the large set of “showy” male organs—from peacocks’ tails to deer antlers to elaborate behavioral displays in birds-of-paradise—that presumably evolved in the process identified by Darwin as “sexual selection.” In one category of sexual selection, called “female choice” by Darwin, these elaborate structures (encumbrances in any other context) develop and enlarge because females prefer the bigger or more decorated males. Female choice may explain the extensive and gaudy patterning, but why tail feathers in the first place? Why not one of a hundred alternatives—head plumes, elaborate calls, or the more common mammalian analog of old-fashioned male bullying?
Several evolutionists, in the past few years, have thought more deeply about the difference between origins and later pathways and have taken Darwin’s problem and solution more seriously. They have realized that the pathways set by female choice must often involve an important initiating component of preexisting bias in sensory and cognitive systems. Females, after all, perceive and process information in a limited number of ways based on broad features of brain and sense organs that obviously did not evolve in order to prefer showy tail feathers in some unspecified future. Relative to tail feathers, or antlers, or complex behavioral displays, these biases are components of good fortune that permit the initiation of a particular trend. Two recent studies have provided excellent evidence for the great seal principle by combining experimental data on female choice, with a documentation of preexisting sensory biases in a genealogical context that validates an evolutionary argument.
Fish tails. When we think of the conjunction of weaponry and fishes, we usually picture a large and graceful marine species with a sword for a snout and a lovely name of classical redundancy—Xiphias gladius, the swordfish (xiphias is a Greek sword; gladius, a Latin counterpart, as in the gladiator who wields it in combat). But a much smaller, Central American, fresh-water relative of the guppy, also bears a sword—this time at the rear end, formed after sexual maturity, and only in males, by an elongation of rays at the base of the caudal (tail) fin.
Alexandra Basolo of the University of California at Santa Barbara performed behavioral experiments on the swordtail, Xiphophorus helleri, and proved that females do prefer males with longer swords, thus establishing the efficacy of female choice in maintaining and enlarging the male sword. But such information tells us very little about the origin of swords. These projections do males a world of Darwinian good, but why swords, rather than big eyeballs, funny fins, or elaborate swimming displays? Fortunately, we have enough information about the genealogy of swordtails to reconstruct a historical sequence, and to recognize an important component of preexisting female sensory bias in the evolution of swords.
The close relatives of X. helleri are all swordless, and we may conclude that ordinary swordless tails represent the original state of this lineage. In particular, Xiphophorus maculatus, the closest living relative of the swordtail, lacks a rear projection (despite its taxonomic residence in the same genus, with its etymology of “sword bearer”). Basolo therefore performed a series of ingenious and elegant experiments on the swordless X. maculatus.
In her basic procedure, she placed a female in the center section of an aquarium constructed with two side compartments of equal volume. She then put a single male into each of the side compartments and noted female preference by time spent in the vicinity of males, and by performance of courtship behaviors. In these particular experiments, she surgically implanted, into the tail of swordless X. maculatus males, swords of the same relative length and form as in X. helleri. In some males, the swords had the same distinctive yellow color and bold black border as in X. helleri; in others, the swords were transparent (and shown in behavioral experiments to be invisible to females).
Basolo placed a male with a colored tail into one side compartment and a male with an invisible tail into the other chamber. (She followed this elaborate procedure of implanting invisible tails, rather than simply using ordinary tailless males, because she needed to control for the results of surgery and the effect of a tail upon swimming and other male behaviors. If females just reacted differently to an ordinary, unoperated male, than to a male with an experimentally implanted tail, we would not know whether this disparity recorded the tail’s presence or the results of surgical intervention.)
Basolo used six pairs of males, each containing one fish with a visible and the other with a transparent sword. She tested each pair with nine to sixteen female fishes. Invariably, females preferred males with visible swords—even though males of this species have no swords at all in nature. As in all good experiments, Basolo then performed a variety of additional tests to eliminate other interpretations. She changed sides for males with visible and invisible swords—just in case females were choosing left or right sides of the aquarium, rather than visible or invisible swords. The females preferred the visible sword, regardless of position in the aquarium. She even performed a second operation and switched swords—placing the transparent sword into the fish that previously carried the colored version, and implanting the colored sword into the fish that had borne the invisible addition. The fish that had previously been shunned (presumably for its invisible sword) was now favored when bearing the sword with the prominent black border. Basolo concludes: “These data suggest that the females were basing their choice on sword preference and not other traits.”
The title of Basolo’s article says it all—“Female preference predates the evolution of the sword in swordtail fish” (see bibliography). Something in the sensory system of ancestral fishes evidently predisposed females of the X. helleri line to prefer males with swords. Since no previous member of this large and successful group of fishes possessed swords (so far as we know), this sensory and cognitive bias exists for other reasons, and must be regarded as fortuitous with respect to the evolution of swords. Male Xiphophorus helleri must, in this sense, thank Lady Luck for their graceful extensions.
Frog calls. So much of what we view as most aesthetic and charming in nature—the singing and plumage of birds, as a prime example—actually functions as part of the great Darwinian struggle for reproductive success. Chorusing of males in crickets, frogs, or birds, is no paean of praise to the night, no hosanna to the joys of life, but a complex tapestry of challenge (to other males) or advertisement (to females).
In many frogs, the female choice model of sexual selection seems to apply, and males call to win the sexual attention of females—all in the service of the great Darwinian attempt to avoid croaking (vernacular sense) of family lines. Michael J. Ryan and colleagues at the University of Texas in Austin have applied the preexisting bias model (“sexual selection for sensory exploitation” in their terms) to the complex call of the Tungara frog, Physalaemus pustulosus (see bibliography).
This Panamanian frog has an unusually complex call, consisting of two sequential components with expressive names: a whine and a chuck. The call begins with the longer whine, about 350 milliseconds in duration, that gradually decreases in fundamental frequency from 900 to 400 Hertz. Although the whine contains up to three harmonics, most energy resides in the fundamental. (Harmonics are overtones generated from the fundamental and having higher frequencies at integral multiples of the fundamental. If that sounds like a mouthful, the first harmonic of a 220 Hz fundamental frequency is 440 Hz, the second 660 Hz, the third 880 Hz, etc.). The whine is followed by a series of one to six chucks. These chucks are much shorter in duration (about 40 milliseconds) and have a lower fundamental frequency of about 220 Hz. But, unlike the whine, chucks have much higher energy in the fifteen harmonics above the fundamental.
In fact, some 90 percent of the energy resides in harmonics above 1500 Hz, with the peak frequency above 2000 Hz.
This complexity becomes important in Ryan’s argument for an interesting reason based on the physiology of amphibian hearing. Uniquely among terrestrial vertebrates, amphibians possess two inner-ear organs that pick up airborne vibrations—the amphibian papilla and the basilar papilla. The amphibian papilla is most sensitive to frequencies below 1200 Hz, while the basilar papilla responds best to higher frequencies above 1500 Hz.
Direct study of the inner ear in Physalaemus pustulosus shows that the most sensitive fibers of the amphibian papilla are tuned to about 500 Hz, while all fibers in the basilar papilla are most sensitive to about 2100 Hz. These facts suggest an obvious hypothesis for evolution of the complex call in P. pustulosus, particularly for the addition of chucks to the presumed ancestral call of whine alone—namely, that the whine only stimulates the amphibian papilla, while addition of the chuck takes advantage of a latent capacity already present but unutilized in ancestral calls: the acoustical properties of a basilar papilla tuned to higher frequencies concentrated in the chuck. The basilar papilla provides the preexisting sensory bias (sensitivity to higher frequencies), and the chuck finally contacts this everpresent, but initially unexploited, capacity.
Since the calls elicit female attention (with approach and eventual mating), Ryan and colleagues performed an interesting and successful experiment. They synthesized a variety of calls and broadcast two different versions from opposite ends of an indoor arena measuring 3 square meters. They put a female in the center of the arena and covered her with an opaque cone. They gave her five minutes to acclimate as they played the calls. A remote device then lifted the cone, and the female was free to approach a speaker. If she consistently preferred one to another, then the relative evolutionary value of whines and chucks might be assessed.
Females consistently favored the complex call of whine plus chuck over the whine alone. This preference is not a simple result of adding more total energy by including the chucks, but seems to be set by distinctive characters of the chuck. If females are given a choice between whine plus chuck and an enhanced whine alone (with 50 percent more total energy than whine plus chuck), they still prefer the complex call of whine plus chuck, despite its lower energy. Finally, Ryan and colleagues determined that females respond equally well to both the low and the high harmonics of the chuck. In other words, they are equally positive towards components of the chuck that stimulate either the amphibian or the basilar papilla.
So far so good. The added chuck does elicit female preference, and the component of the chuck that stimulates the previously unexploited basilar papilla also appeals to females. But to argue for the preexisting sensory bias model, we need to know that ancestral species (or relatives maintaining the ancestral state) are also inclined to react favorably to the chuck, even though their distinctive call contains no such component—just as Basolo showed that females in unarmored species prefer males with surgically implanted swords. Ryan was able to supply this final piece of evidence by measuring the tuning of basilar papillas in seven individuals of the closely related species, Physalaemus coloradorum. Ryan writes: “This species does not produce chucks, and the ability to produce chucks was derived in P. pustulosus after these species diverged.” Ryan found no statistically significant difference between the most sensitive frequencies for P. pustulosus (2130 Hz) and for the chuckless P. coloradorum (2230 Hz). The basilar papilla of P. coloradorum is therefore tuned to high frequencies not found in their actual call. A preexisting sensory bias for potentially advantageous evolutionary change can therefore be specified—a pathway actually followed by P. pustulosus.
These conclusions about preexisting sensory bias, while satisfying, also present another paradox. How can something so specific as the preference for an extension to a tail or a funny sound be encoded by accident into ancestral behaviors? Can we seriously believe that an animal might be adapted to favor swords never seen or chucks never heard?
The probable resolution of this paradox (another logical frustration in the terms of my introduction) may be illustrated by a famous experiment on quail, done ten years ago by the British ethologist Patrick Bateson (see bibliography). Avoidance of incest is very common in vertebrates with complex behaviors and high cognitive capacity. The evolutionary rationale is easy to express: Mating with closest kin produces a high frequency of genetically compromised offspring with disadvantageous traits in double recessive doses (a phenomenon called inbreeding depression). But quail don’t know Mendelian genetics (and neither did people before this century). So what can be leading animals to this evolutionarily advantageous behavior?
Bateson built an ingenious device that exposed individual quail to five birds of the opposite sex, but of different degrees of relationship: a sibling nestmate, a sibling never seen before, a first cousin, a third cousin, and an unrelated bird. Both males and females generally preferred first cousins over all alternatives.
One popular hypothesis (applicable to humans in some interpretations) holds that we avoid closest kin by a simple learning rule that derails later sexual feelings towards individuals reared with us from our earliest days (as one wag said, if we share potties early, we don’t party later—and remember, where I live in Boston, the two words are pronounced nearly alike). On this argument, since rearingmates are usually sibs, the simple learning rule turns the proper evolutionary trick. But this explanation will not suffice for Bateson’s data, for quail prefer first cousins over true siblings never seen before.
Bateson therefore concludes, from this and other arguments, that quail may be following a highly abstract aesthetic rule—prefer intermediary degrees of familiarity: not so close as to be cloying, not so distant as to be overly strange. If he is right, an elegant solution to the problem of avoiding incest suggests itself. Quail are not Mendelian calculators. They are, rather, following a deeper, and more abstract, rule of aesthetic preference that may be common to a wide range of animals and neurologies. Maximal attraction to intermediate familiarity will automatically exclude disadvantageous closest kin as potential mates. Natural selection need not work for the specific goal of avoiding incest. By good fortune, a deeper cognitive principle engenders this result as a consequence. (Of course, one might turn the argument around and claim that the aesthetic principle arose because incest avoidance is so important, and animals could only achieve this result by such an indirect route. But I prefer to view the specific as a manifestation of the general, rather than the rule as a surrogate for the example.)
This same style of argument makes the preexisting bias hypothesis more sensible. We need not postulate a preexisting bias for seeing and favoring swords on tails. We only require a general behavioral rule (like intermediate familiarity), that might render the specific result (avoiding incest) as a reasonable manifestation. In fact, Basolo suggests that swords may be preferred by females in swordless species because the implanted weapon makes the male look larger in general, and bigger size is a potent spur to female choice in many regimes of sexual selection. Thus, the general cognitive rule would proclaim: Prefer larger males. The specific solution in this case would be: Extended swords give an impression of larger size with little addition of actual bulk. A thousand other pathways might have satisfied the same broad rule, but Xiphophorus evolved a sword. Similarly, the preexisting bias in frogs is a basilar papilla tuned to high frequencies, not an irresistible urge to hear a chuck. Again, this bias might have been exploited in many other ways, but P. pustulosus evolved a chuck.
The solution is elegant (and probably even true in these cases—what a rare and lovely combination). Evolution is always a wondrous mixture of the quirky and unpredictable (usually expressed as historical legacies brought to different modern contexts) with the sensible adaptive improvements wrought by natural selection. The quirky component of historical legacy constrains the predictable force of immediate selection, so we usually think of history as res
trictive and selection as flexible. But the stories of this essay reverse the usual perspective. Here, historical legacy is a broad cognitive rule bursting with potential along a thousand possible pathways—prefer larger males, or prefer individuals of intermediate familiarity. And adaptation then clamps the limit by choosing one manifestation—a sword, a chuck, or a first cousin. If Lady Luck smiles on the beginning of such an evolutionary trend from one side of our great seal, I am tempted to quote the more familiar motto from the other side to describe the final choice of a singular solution: E pluribus unum, one from many.
27 | A Dog’s Life in Galton’s Polyhedron
IN THE OPENING sentence of Hereditary Genius (1869), the founding document of eugenics, Francis Galton (Charles Darwin’s brilliant and eccentric cousin—see Essay 31 for another tale of this remarkable man) proclaimed that “a man’s natural abilities are derived by inheritance.” He then added, making an appeal by analogy to changes induced by domestication:
Consequently, as it is easy…to obtain by careful selection a permanent breed of dogs or horses gifted with peculiar powers of running, or of doing anything else, so it would be quite practicable to produce a highly-gifted race of men by judicious marriages during several consecutive generations.
Darwin had also invoked domestication as his first argument in the Origin of Species. Darwin began his great treatise, not with fanfare or general proclamation, but with a discussion of breeding in domestic pigeons (see Essay 25).
Darwin attributed the wondrous variety among pigeons, dogs, and other domesticated animals to the nearly limitless power of selection: “Breeders habitually speak of an animal’s organization as something plastic, which they can model almost as they please.” He quotes one authority on the “great power of this principle of selection”: “It is the magician’s wand, by means of which he may summon into life whatever form and mold he pleases.”