The Flamingo’s Smile

by Stephen Jay Gould

  Once we abandon the alluring, but fallacious, image of Darwin winning his intellectual battle utterly alone at sea, we can ask the really interesting question that begins to probe Darwin’s particular genius. Gould was the expert. Gould resolved the details correctly. Gould, a staunch creationist in taxonomy, nonetheless recognized that he had to abandon beaks as key characters. Darwin could accomplish none of this. But Darwin, not Gould, recognized that all the pieces required a stunningly new explanation—evolution—to make a coherent story. The amateur triumphed when the stakes were highest, while the professional got the details right and missed the organizing theme.

  Darwin continued to work in this way throughout his career. Somehow, as an amateur, he could cut through older patterns of thought to glimpse new modes of explanation that might better fit an emerging, detailed story constructed by experts who, somehow, could not take the big and final step. But Darwin worked with his culture and with his colleagues. Science is a collective endeavor, but some individuals operate with an enlarged vision—and we would like to know how and why. We can ask no harder question, and I propose no general solution. But we do need to clear away heroic legends before we can begin.

  24 | A Short Way to Corn

  SINCE IT WAS only a few miles to Tipperary, not the long way of song and legend, I took a detour to visit the town. Soon I felt like the city slicker of that old New England joke. Looking for a small town, he stops before a general store and asks an old-timer, “Where is Pleasantville?” “Don’t you move a god-damned inch,” comes the reply.

  Tipperary, made large by its fame and my imagination, is but one main street with a few stores and houses. This eerie scene repeated itself again and again during my visit to this most beautiful of European lands. For Ireland, contrary to the trend of most other countries, is a depopulated nation. Its current count of some 3 million inhabitants includes but half of the 1840 total. Abandoned homes, farms, and even towns lie strewn about the countryside.

  The beginning of the great emigration that so enriched my native city of New York and my current Boston home dates to the great potato famines of 1845 and 1846, when half a million people starved to death and another million left. The potato is a remarkable food. It contains so well balanced an array of nutrients that people can live on virtually nothing else for years on end. Monotonous perhaps, spuds being spuds, but quite viable. Irish peasants often ate nothing but potatoes through the long winter months. But disease attacked the crop in 1845 and virtually destroyed it, producing unprecedented starvation and the great exodus to Liverpool and beyond.

  The Irish potato blight illustrates a classic dilemma in agriculture. To produce the “best” plant for maximal yields, farmers and scientists will hone and select for many generations until they obtain just the right combination of features. They will then propagate their entire crop from this improved form. These plants, as offspring of a single parental type, are genetically uniform and depleted in variability. In other words, we trade genetic diversity for an unvarying optimum.

  All may be well for a while, but uniform stocks are exquisitely susceptible to the ravages of disease. If some virus, bacterium, or fungus successfully attacks the plants, it can destroy every one, thus devastating the crop. In natural populations, on the other hand, genetic variation among individuals insures that some will enjoy protection against the agent of disease and part of the crop will survive. Since next year’s plants are offspring of these immune survivors, populations with abundant variability maintain a natural mechanism to purge themselves of disease.

  The Irish, growing their potatoes from a uniform stock, lost their entire crop in 1845. The same story can be told for most agricultural mainstays. Some scholars believe that the mysterious collapse of classic Maya civilization was precipitated by a virus, borne by leaf hoppers dispersed on high-altitude air currents, that wiped out their corn crop virtually overnight. Corn continues to plague us with similar problems. During the summer of 1970, a new mutant strain of Southern Leaf Blight Fungus swept across American cornfields at rates of fifty miles or more a day, devastating all plants bred to contain a genetic element called Texas cytoplasmic male sterility factor.

  To avoid this dilemma, breeders try to beef up genetic variability by hybridizing their successful but uniform stocks with different strains. For corn, a major source of potential hybridization lies in a plant of markedly different appearance, the New World grass known as teosinte. For example, Zea diploperennis, a recently discovered species of teosinte, is the only known source of immunity to three of the major viruses that afflict domestic corn. (This species is also a perennial rather than an annual like corn, thus giving potential substance to an old dream that, by hybridization, breeders might produce a perennial corn that survives from season to season and need not be replanted from seed each year.)

  It may seem strange at first that a plant so different in appearance from corn should be sufficiently similar in genetic structure to permit hybridization. True, young plants of corn and teosinte are indistinguishable, but after they flower, the differences in adult structures could hardly be more profound. The business end of corn is a large cob bearing numerous rows of kernels (the technical term, polystichous—simply meaning many rowed—has a lovely ring). The cob and kernels are female, and they reside at the terminal end of stout branches lateral to the main stem (mark this well, for these positions become crucial in my developing argument). Many people don’t recognize this position because corn ears just seem to be stuck to the sides of the main stem. But the husks that so completely enclose the ear are actually remnants of leaves that originally formed on a longer lateral branch. They cover the cob, which is, indeed, terminal on a drastically shortened lateral branch. The central stem bears a male tassel, the source of pollen, at its terminal end. Thus, corn grows separate male and female structures: the tassel, terminal on the main stem, is male; the ears, terminal on lateral branches, are female.

  Teosinte, on the other hand, grows a central stem and many long lateral branches of comparable length and strength. Each branch ends in a male tassel. The female ears, quite unlike corn, grow laterally, not terminally, from the lateral branches. The teosinte “ear” is also a miserable analog or runt compared with the majestic ear of corn. It contains (depending on the race of teosinte) six to twelve triangular kernels in two rows (technically distichous) telescoped into one because the triangular ends of the opposing kernels interdigitate. The kernels are surrounded by a stony outer covering and are quite useless as human food unless popped (as in popcorn) or laboriously ground and separated from their inedible covering. (Corn kernels are soft and naked, immediately available for food because their covering structures are not only pliant, but so reduced in size that they surround only the base of the kernel.)

  In modern corn (left), female ears are terminal on lateral branches; in teosinte (right), male tassels are terminal on lateral branches, while female ears are lateral on the lateral branches. Thus, the modern corn ear is the homolog of a teosinte tassel spike. See text for explanation. REPRINTED FROM NATURAL HISTORY.

  Yet, despite these differences, corn and teosinte hybridize without any impediment, producing cobs of intermediate size. This paradoxical compatibility exists for two basic reasons that reflect the subject of this essay—a disquisition on the ancestry and origin of corn. First, teosinte is probably the direct ancestor of domestic corn (some experts disagree, although no one denies the close relationship). Second, no chromosomal disparities or even simple and consistent differences in single genes have been found between teosinte and corn. (Of course, the two forms could not be so different in appearance without some genetic divergence, but ease of hybridization and our failure to find differences indicate that genetic distinction between the two forms must be minuscule. Indeed, botanists place corn and the annual teosintes in the same species, Zea mays.)

  The teosinte theory for the origin of corn has always suffered from one major dilemma: How could it happen? How could the teo
sinte ear, so different from corn, become the modern cob? Corn, like all our major domestic cereals, is a grass. The evolutionary origin of other major grains, wheat for example, presents fewer problems. Wheat ears differ only quantitatively from their wild grassy ancestor—they are essentially just more of the same, but bigger. We can easily understand how agricultural selection could translate a wild ancestor into domestic wheat. But how can you make a corncob from a teosinte ear or from any part of teosinte? They are constructed so differently.

  In the standard version of the teosinte hypothesis—which I will reject here in favor of a radical alternative—the teosinte ear is, nonetheless, gradually transformed into a modern corn ear. It adds rows gradually, while the hard outer covering softens and retracts from the kernels. This scenario seems so obvious and so consistent with our usual view of evolutionary transformation. The tradition of gradual change from teosinte ear to corn ear dates at least to Luther Burbank, the great “wizard” of early twentieth-century plant breeding, who claimed that he had transformed teosinte to corn in eighteen generations of selection. He was wrong. He had started, not with teosinte, as he thought, but with a corn-teosinte hybrid—and his selection had merely segregated and accumulated the genetic factors for corn. But his general argument for a gradual transformation of the teosinte ear into a corncob persisted. In a Scientific American article of January 1980, George Beadle, one of the great corn scientists of our age, proclaimed that “the cobs can be placed in an evolutionary continuum from teosinte to modern corn on the basis of progressive modifications.”

  But this theory of gradual derivation from the teosinte ear encounters three great problems, perhaps fatal. First, corn appears suddenly in the archeological record about 7,000 years ago. The earliest ears, to be sure, are not as fat or as many-rowed as a modern cob, but they clearly represent corn, not something in between corn and teosinte. Second, as stated before, breeders have found no consistent genetic difference between corn and teosinte. If corn were the product of long and slow selection from teosinte, a considerable number of genetic changes should have accumulated. Both these arguments are negative and therefore not conclusive. Perhaps sudden appearance merely records our failure to find intermediates; perhaps the absence of genetic difference only means that we haven’t looked at the right parts of the right chromosomes.

  The third argument is positive and more troubling for the hypothesis that corn ears arose from teosinte ears. Remember the point I asked you to flag some paragraphs back: the positions of teosinte and corn ears are not equivalent. The teosinte ear sprouts laterally from lateral branches; the corn ear grows terminally on lateral branches. In teosinte, the terminal structure on the main lateral branches is a male tassel, not a female ear. Therefore, by position—and I shall say in a moment why position is so important a criterion—the modern female ear of corn is equivalent to (or, as we say in technical parlance, is the homolog of) a male tassel spike.

  This homology of male tassel spike to female ear has long been recognized (and stated) by many corn experts, but no one has previously exploited this fact to develop a hypothesis for the origin of corn. The obvious theory suggested by this homology may, at first, sound absurd, but it solves plausibly and with elegance all the classical problems of the teosinte hypothesis. In short, this new theory proposes that corn ears evolved rapidly from male tassel spikes by shortening of the lateral branches and suppression of teosinte ears below. Instead of a slow and continuous enlargement of female teosinte ears, we envision an abrupt transformation of male tassel spikes to small and primitive versions of a modern female corn ear.

  Hugh H. Iltis, professor of botany and director of the Herbarium at the University of Wisconsin in Madison, developed this heterodox theory and recently published it in America’s leading professional journal (see bibliography).* I have no corn credentials and cannot make any proclamation about the truth or falsity of this intriguing idea. But I do want to illustrate its status as a plausible, potential example of an evolutionary process often dismissed with ridicule for want of understanding—the so-called hopeful monster.

  We call parts of two organisms “homologous” when they represent the same structure by a criterion of evolutionary descent from a common ancestor. No concept is more important in unraveling the pathways of evolution, for homologies record genealogy, and false conclusions about homology invariably lead to incorrect evolutionary trees.

  Homologous structures need not look alike. Indeed, the standard examples invoke organs quite dissimilar in form and function, for these “classics” are chosen to illustrate the idea that mere resemblance does not qualify as a criterion. Examples include the homology of the hammer and anvil bones of the mammalian middle ear with the jaw articulation bones of reptiles, and the lung of land vertebrates with the air bladder of bony fishes.

  How then can we recognize homology and thereby reconstruct the pathways of evolution? This most difficult question in evolutionary theory has no definite answer. No single criterion works in every case; all rules have well-known exceptions. We must evaluate proposed homologies by all available standards and accept or reject a hypothesis by the joint and independent affirmation of several criteria. Similarity in early embryology often works well for structures that become very different in adults: early mammalian embryos first develop their ear bones at the ends of their jaws—and this fact harmonizes with a well-established fossil sequence showing continual decrease of these two jaw bones and their eventual movement to the middle ear. But truly homologous organs may be modified by evolutionary changes in embryos that mask the pathways of descent.

  A seemingly superficial detail—simple spatial relation with other parts—often serves well as a criterion of homology. As the old song goes, the foot bone truly is connected to the ankle bone, and such fundamental relationships are not easily altered in evolution. Thus, the so-called “positional criterion” of homology is probably the most respected and most often utilized of all standards. And by this criterion, modern female corn ears must have descended from teosinte male tassel spikes (for both features are alike in position at the terminal ends of lateral branches), and not from female teosinte ears.

  Lest it seem absurd that male structures be transformed into female organs of such different appearance, I remind readers that male and female parts often have a common basis in embryology, one developing directly from the other under the influence of different hormones. The external genitalia of all mammals, for example, begin as female structures: the clitoris enlarges, folds over and fuses to form a cylinder with a central tube, the male penis; the labia majora expand and fuse at the midline to form a scrotal sac. In essay 11 of Hen’s Teeth and Horse’s Toes, I used these equivalences to argue that the remarkable male-mimicking genitalia of female spotted hyenas arise automatically from these common pathways of sexual development, because females of this species secrete unusually high levels of testosterone during growth and become both larger than males and dominant over them.

  Tassel spike and corn ear are also equivalent structures and the transformation of one to the other is equally plausible. Indeed, such interchanges often occur as teratologies, or abnormalities of development, in modern corn. Male tassel spikes may grow as female ears or partial ears with male ends, for several reasons: genetic mutations and diseases that drastically shorten the central branch, for example. Iltis sent me the accompanying photograph of such a feminized tassel, sold as an ordinary ear of corn at Kohl’s supermarket in Madison, Wisconsin, for thirty-nine cents. The central axis has grown as a complete female ear. The three lateral branches are female at the bottom, grading to male at the top, with the male parts “arranged,” Iltis tells me, “exactly as in any maize or teosinte tassel branch.” Of course, such teratologies only show the interchangeability of corn tassels and corn ears, not the evolutionary derivation of corn ears from teosinte tassels. But they surely illustrate, by strong analogy, why the genealogical path from teosinte tassel to corn ear remains so reasonable, however
peculiar at first hearing.

  In calling his theory the “catastrophic sexual transmutation,” Iltis forcefully identifies its two outstanding and unconventional properties. First, using the positional criterion of homology as a guide, female corn ears arose by transmutation of a male teosinte tassel spike, not by gradual enlargement of a female teosinte ear. Second, the transformation occurred rapidly under the guidance of little (or even no) genetic change, despite the sudden and striking alteration of form. I shall try to epitomize Iltis’s argument in the following basic steps:

  1. In both corn and teosinte, hormones are distributed along simple gradients in long stems, with male zones at the tops passing through a threshold to female zones below.

  2. A gradient in time of differentiation during growth also accompanies this hormonal distribution. Structures at the tops of stems develop earlier than those lower down. On a lateral teosinte branch, the terminal male tassel differentiates before the female ears below.

  A corn “monster” purchased for thirty-nine cents at a Madison, Wisconsin, supermarket by Hugh Iltis. PHOTO COURTESY OF HUGH ILTIS.

  3. The nutritional needs of a male tassel are small, those of a female ear (particularly a large and polystichous ear of corn) much larger. The differentiation of a tassel at the terminal end of a branch still leaves most nutrients available for development of subsequent female structures below (see point 2).