Bully for Brontosaurus

by Stephen Jay Gould

  An amazing and famous photo of a female kiwi one day before laying its enormous egg. COURTESY OF THE OTOROHANGA ZOOLOGICAL SOCIETY, NEW ZEALAND.

  But the greatest of kiwi oddities centers upon reproduction. Females are larger than males. They lay one to three eggs and may incubate them for a while, but they leave the nest soon thereafter, relegating to males the primary task of incubation, a long seventy to eighty-four days. Males sit athwart the egg, body at a slight angle and bill stretched out along the ground. Females may return occasionally with food, but males must usually fend for themselves, covering both eggs and nest entrance with debris and going forth to forage once or twice on most nights.

  The kiwi egg is a wonder to behold, and the subject of this essay. It is, by far, the largest of all bird eggs relative to body size. The three species of kiwis just about span the range of domestic poultry: the largest about the size of Rhode Island Reds; the smallest similar to bantams—say five pounds as a rough average (pretty meaningless, given the diversity of species, but setting the general domain). The eggs range to 25 percent of the female’s body weight—quite a feat when you consider that she often lays two, and sometimes three, in a clutch, spacing them about thirty-three days apart. A famous X-ray photo of kiwi and egg taken at the kiwi sanctuary of Otorohanga, New Zealand, tells the tale more dramatically than any words I could produce. The egg is so large that females must waddle, legs spread far apart, for several days before laying, as the egg passes down the oviduct toward the cloaca. The incubation patch of male kiwis extends from the top of the chest all the way down to the cloaca—in other words, they need almost all their body to cover the egg.

  A study of the general relationship between egg size and body size among birds shows that average birds of kiwi dimensions lay eggs weighing from 55 to 100 grams (as do domestic hens). Eggs of the brown kiwi weigh between 400 and 435 grams (about a pound). Put another way, an egg of this size would be expected from a twenty-eight-pound bird, but brown kiwis are about six times as small.

  The obvious question, of course, is why? Evolutionary biologists have a traditional approach to riddles of this sort. They seek some benefit for the feature in question, then argue that natural selection has worked to build these advantages into the animal’s way of life. The greatest triumphs of this method center upon odd structures that seem to make no sense or (like the kiwi egg) appear, prima facie, to be out of proportion and probably harmful. After all, anyone can see that a bird’s wing (although not a kiwi’s) is well designed for flight, so reference to natural selection teaches you little about adaptation that you didn’t already know. Thus, the test cases of textbooks are apparently harmful structures that, on closer examination, confer crucial benefits upon organisms in their Darwinian struggle for reproductive success.

  This general strategy of research suggests that if you can find out what a structure is good for, you will possess the major ingredient for understanding why it is so big, so colorful, so peculiarly shaped. Kiwi eggs should illustrate this basic method. They seem to be too big, but if we can discover how their large size benefits kiwis, we shall understand why natural selection favored large eggs. Readers who have followed my essays for some time will realize that I wouldn’t be writing about this subject if I didn’t think that this style of Darwinian reasoning embodied a crucial flaw.

  The flaw lies not with the claim of utility. I regard it as proved that kiwis benefit from the unusually large size of their eggs—and for the most obvious reason. Large eggs yield large and well-developed chicks that can fend for themselves with a minimum of parental care after hatching. Kiwi eggs are not only large; they are also the most nutritious of all bird eggs for a reason beyond their maximal bulk: they contain a higher percentage of yolk than any-other egg. Brian Reid and G. R. Williams report that kiwi eggs may contain 61 percent yolk and 39 percent albumin (or white). By comparison, eggs of other so-called precocial species (with downy young hatching in an active, advanced, and open-eyed state) contain 35 to 45 percent yolk, while eggs of altricial species (with helpless, blind, and naked hatchlings) carry only 13 to 28 percent yolk.

  The lifestyle of kiwi hatchlings demonstrates the benefits of their large, yolky eggs. Kiwis are born fully feathered and usually receive no food from their parents. Before hatching, they consume the unused portion of their massive yolk reserve and do not feed (but live off these egg-based supplies) for their first seventy-two to eighty-four hours alfresco. Newly hatched brown kiwi chicks are often unable to stand because their abdomens are so distended with this reserve of yolk. They rest on the ground, legs splayed out to the side, and only take a first few clumsy steps when they are some sixty hours old. A chick does not leave its burrow until the fifth to ninth day when, accompanied by father, it sallies forth to feed sparingly.

  Kiwis thus spend their first two weeks largely living off the yolk supply that their immense egg has provided. After ten to fourteen days, the kiwi chick may weigh one-third less than at hatching—a fasting marked by absorption of ingested yolk from the egg. Brian Reid studied a chick that died a few hours after hatching. Almost half its weight consisted of food reserves—112 grams of yolk and 43 grams of body fat in a 319-gram hatchling. Another chick, killed outside its burrow five to six days after hatching, weighed 281 grams and still held almost 54 grams of enclosed yolk.

  I am satisfied that kiwis do very well by and with their large eggs. But can we conclude that the outsized egg was built by natural selection in the light of these benefits? This assumption—the easy slide from current function to reason for origin—is, to my mind, the most serious and widespread fallacy of my profession, for this false inference supports hundreds of conventional tales about pathways of evolution. I like to identify this error of reasoning with a phrase that ought to become a motto: Current utility may not be equated with historical origin, or, when you demonstrate that something works well, you have not solved the problem of how, when, or why it arose.

  I propose a simple reason for labeling an automatic inference from current utility to historical origin as fallacious: Good function has an alternative interpretation. A structure now useful may have been built by natural selection for its current purpose (I do not deny that the inference often holds), but the structure may also have developed for another reason (or for no particular functional reason at all) and then been co-opted for its present use. The giraffe’s neck either got long in order to feed on succulent leaves atop acacia trees or it elongated for a different reason (perhaps unrelated to any adaptation of feeding), and giraffes then discovered that, by virtue of their new height, they could reach some delicious morsels. The simple good fit of form to function—long neck to top leaves—permits, in itself, no conclusion about why giraffes developed long necks. Since Voltaire understood the foibles of human reason so well, he allowed the venerable Dr. Pangloss to illustrate this fallacy in a solemn pronouncement:

  Things cannot be other than they are…. Everything is made for the best purpose. Our noses were made to carry spectacles, so we have spectacles. Legs were clearly intended for breeches, and we wear them.

  This error of sliding too easily between current use and historical origin is by no means a problem for Darwinian biologists alone, although our faults have been most prominent and unexamined. This procedure of false inference pervades all fields that try to infer history from our present world. My favorite current example is a particularly ludicrous interpretation of the so-called anthropic principle in cosmology. Many physicists have pointed out—and I fully accept their analysis—that life on earth fits intricately with physical laws regulating the universe, in the sense that were various laws even slightly different, molecules of the proper composition and planets with the right properties could never have arisen—and we would not be here. From this analysis, a few thinkers have drawn the wildly invalid inference that human evolution is therefore prefigured in the ancient design of the cosmos—that the universe, in Freeman Dyson’s words, must have known we were coming. But the
current fit of human life to physical laws permits no conclusion about the reasons and mechanisms of our origin. Since we are here, we have to fit; we wouldn’t be here if we didn’t—though something else would, probably proclaiming, with all the hubris that a diproton might muster, that the cosmos must have been created with its later appearance in mind. (Diprotons are a prominent candidate for the highest bit of chemistry in another conceivable universe.)

  But back to kiwi eggs. Most literature has fallen into the fallacy of equating current use with historical origin, and has defined the problem as explaining why the kiwi’s egg should have been actively enlarged from an ancestor with an egg more suited to the expectations of its body size. Yet University of Arizona biologist William A. Calder III, author of several excellent studies on kiwi energetics (see 1978, 1979, and 1984 in the bibliography), has proposed an opposite interpretation that strikes me as much more likely (though I think he has missed two or three good arguments for its support, and I shall try to supply them here).

  The alternative interpretation holds that kiwis are phyletic dwarfs, evolved from a lineage of much larger birds. Since these large ancestors laid big eggs appropriate to their body size, kiwis just never (or only slightly) reduced the size of their eggs as their bodies decreased greatly in bulk. In other words, kiwi eggs never became unusually large; kiwi bodies got small—and these statements are not equivalent, just as we know that an obese man is not short for his weight, despite the old jest.

  (Such a hypothesis is not anti-adaptationist in the sense that maintenance of a large egg as size decreases—and in the face of energetic and biochemical costs imposed by such a whopping contribution to the next generation—may well require a direct boost from natural selection to prevent an otherwise advantageous decrease more in keeping with life at Colonel Sanders’s favorite size. Still, there is a world of difference between retaining something you already have, and first developed for other reasons [in this case simple appropriateness for large body size], and actively evolving such a unique and cumbersome structure for some special benefit.)

  Calder’s interpretation might seem forced or farfetched but for the outstanding fact of taxonomy and biogeography cited as the introduction to this essay. Moas are the closest cousins of kiwis, and most moas were very large birds. “Is the kiwi perhaps a shrunken moa?” Calder asks. Unfortunately, all moa fossils lie in rocks of a geological yesterday, and kiwi fossils are entirely unknown—so we have no direct evidence about the size of ancestral kiwis. Still, I believe that all the inferential data support Calder’s alternative hypothesis for the great size of kiwi eggs—a “structural” or “historical” explanation if you will, not a conventional account based on natural selection for immediate advantages.

  Although the best argument for viewing kiwis as much smaller than their ancestors must be the large size of their closest moa cousins, Calder has also developed a quirky and intriguing speculation to support the dwarfed status of kiwis. (I hasten to point out that neither of these arguments amounts to more than a reasonable conjecture. All evidence can be interpreted in other ways. Both moas and kiwis, for example, might have evolved from a kiwi-sized common ancestor, with moas enlarging later. Still, since the kiwi is the smallest of all ratites—a runt among ostriches, rheas, emus, and cassowaries—its decrease seems more probable than moa increase. But we will not know until we have direct evidence of fossil ancestry.)

  Calder notes that in many respects, some rather curious, kiwis have adopted forms and lifestyles generally associated with mammals, not birds. Kiwis, for example, are unique among birds in retaining ovaries on both sides (the right ovary degenerates in all other birds)—and eggs alternate between sides, as in mammals. The seventy-to eighty-four-day incubation period matches the eighty-day pregnancy expected for a mammal of kiwi body size, not the forty-four days predicted for birds of this weight. Calder continues: “When one adds to this list, the kiwi’s burrow habit, its furlike body feathers, and its nocturnal foraging highly dependent on its sense of smell, the evidence for convergence seems overpowering.” Of course, this conjunction of traits could be fortuitous and each might mean something quite unmammalian to a kiwi, but the argument does gain strength when we remember that no terrestrial mammals reached New Zealand, and that the success of many introduced species indicates a hospitable environment for any creature that could exploit a mammalian way of life.

  You will be wondering what these similarities with mammals could possibly mean for my key claim that kiwis are probably descendants of much larger birds. After all, mammals are superior, noble, and large. But they aren’t. The original and quintessential mammalian way of life (still exploited by a majority of species) is secretive, furtive, nocturnal, smell-oriented in a non-visual world—and, above all, small. Remember that for two-thirds of their geological history, all mammals were little creatures living in the interstices of a world ruled by dinosaurs. If a large bird converged upon a basically mammalian lifestyle in the absence of “proper” inhabitants as a result of geographic isolation, decrease in size would probably be a first and best step.

  Perhaps I have convinced you that kiwis probably decreased in size during their evolution. But why should this dwarfing help to explain their large eggs? Why didn’t egg size just keep pace with body size as kiwis scaled down? We now come to the strong evidence of the case.

  The study of changes in form and proportion as organisms increase or decrease in size is called allometry. It has been a popular and fruitful subject in evolutionary research since Julian Huxley’s pioneering work of the 1920s. One of Huxley’s own classic studies (Journal of the Linnaean Society of London, 1927) bore the title: “On the Relation between Egg-weight and Body-weight in Birds.” Huxley found that if you plot one point for each species on the hummingbird-to-moa curve for egg weight versus body weight, relative egg size decreases in an even and predictable way. The eggs of large birds, he found, are absolutely larger, but relatively smaller in proportion to body weight, than those of small birds.

  Huxley’s work has since been extended several times with more voluminous and consistent data. In the two best studies that I know, Samuel Brody (in his masterful compendium, Bioenergetics and Growth, 1945) calculated a slope of 0.73, while H. Rahn, C. V. Paganelli, and A. Ar (1975), with even more data from some 800 species, derived a similar value of 0.67. This means that as birds increase in body weight, egg weight enlarges only about two-thirds as fast. Conversely, as birds decrease in size, egg weight diminishes more slowly—so little birds have relatively heavy eggs.

  This promising datum will not, however, explain the kiwi’s outsized egg, for the two-thirds slope represents the general standard for all birds. Kiwi eggs are huge compared with the expected egg weight for a bird of kiwi body weight along this standard curve.

  But the literature of allometry has also yielded a generality that will, I think, explain the kiwi’s massive egg. The two-thirds slope of the egg weight/body weight curve represents a type of allometry technically called interspecific scaling—that is, you plot one point for each species in a related group of organisms and attempt to establish the characteristic change of proportion along a gradient of increasing size. (These curves are popularly called mouse-to-elephant for relationships among mammals—hence my designation hummingbird-to-moa for birds.) Allometricians have established hundreds of interspecific curves for birds and mammals.

  Another kind of allometry is called intraspecific scaling. Here you plot one point for each individual among adults of varying body weights within a single species—the Tom Thumb-to-Manute Bol curve for human males, if you will. Since the similarity of these technical terms—interspecific and intraspecific—is so confusing, I shall call them, instead, among-species (for mouse-to-elephant) and within-species (for Thumb-to-Bol).

  As an important generality in allometric studies, within-species curves usually have a substantially lower slope than among-species curves for the same property. For example (and in our best-studied case), the
mouse-to-elephant curve for brain weight versus body weight in mammals has a slope of about two-thirds (as does the egg weight/body weight curve for birds). But the within-species curve from small to large adults of a single species, while varying from one group to another, almost always has a much lower slope in the range of 0.2 to 0.4. In other words, while brains increase about two-thirds as fast as bodies among species (implying that large mammals have relatively small brains), brains only increase about one-fifth to two-fifths as fast as bodies when we move from small to large adults within a single mammalian species.

  Such a regularity, if it applied to egg weight as well, could resolve the kiwi paradox—if kiwis evolved from larger ancestors. Suppose that kiwi forebears start at moa size. By the hummingbird-to-moa among-species standard, egg size should decrease along the two-thirds slope. But suppose that natural selection is operating to favor small adults within a population. If the within-species curve for egg weight had a slope much lower than two-thirds, then size decrease by continued selection of small adults might produce a new species with outsized eggs well above the two-thirds slope, and therefore well above the expected weight for a bird of this reduced size. (Quantitative arguments like this are always easier to grasp by picture than by words—and a glance at the accompanying graph should resolve any confusion.)

  But what is the expected within-species relationship for egg weight? Is the shape of the curve low, as for brain weight, thus affirming my conjecture? I reached for my well-worn copy of Brody’s unparalleled compendium and found that for adults of domestic fowl, egg weight increases not two-thirds as fast, but only 15 percent as fast as body weight! (Brody uses this fact to argue that small hens are usually better than large, so long as egg production remains the same—for egg size diminishes very little with a large decrease in body mass, and the small loss in egg volume is more than compensated by large decreases in feeding costs.)