How to Build a Dinosaur (18 page)

This made the steplike sequence of development clear, but more evidence was needed to link the developmental sequence to an evolutionary sequence. Some of this was readily available in the great variety of feathers in modern birds. Each evolutionary stage of feather development could be seen on some living bird. So, Prum was not inventing any structures that were unknown. All these feather types had appeared on birds at one stage or another.

Nothing they had learned had falsified their hypothesis. Nothing had proved it either. Of course, in historical sciences, like paleontology or evolutionary molecular biology, proof is not possible in the way that it can be obtained in a physics experiment. But predictions can be made and evidence produced that supports or refutes the validity of the predictions. Prum and his colleagues, in describing the sequence of evolution, were, in effect, predicting that extinct organisms existed that had primitive feathers, mere tubes, and downy feathers, and that these should have existed before
Archaeopteryx
.

Paleontology came to the rescue with the discoveries of feathered dinosaurs, which I described in the last chapter, in the 1990s in China. These were just what had been predicted. As Prum writes, “The first feathered dinosaur found there, in 1997, was a chicken-size coelurosaur
(Sinosauropteryx);
it had small tubular and perhaps branched structure emerging from its skin.” Later, other dinosaurs were found with pennaceous feathers. The variety of feathers, including the simple tufted sort that would correspond to the second stage of feather evolution in the Prum plan, all of them on dinosaurs, gave further support to this idea of feather evolution.

Prum’s exhilaration in the
Scientific American
article produced one of the great scientific sentences: “These fossils open a new chapter in the history of vertebrate skin.” Indeed.

Birds became a subset of theropod dinosaurs. Dinosaurs acquired feathers.
T. rex
may even have had them. The idea of feathers evolving from scales was undermined. Scales don’t grow as cylinders, but with a distinct top and bottom. And it became clear that feathers did not evolve for the purposes of flight. Why they evolved we don’t know. Nor can we say when they evolved. And we have found that we will probably never be able to say when birds evolved. All evolution in reality is a continuum, with no sharp distinctions. And nowhere is this clearer than in the transition from theropod dinosaur to avian dinosaur. Arguments now exist over whether some of the Chinese dinosaurs are birds.

In the work on feathers Prum demonstrated and articulated the direction that paleontology and evolutionary biology must take: the same direction that others have favored. As he concluded, “Feathers offer a sterling example of how we can best study the origin of an evolutionary novelty: Focus on understanding those features that are truly new and examine how they form during development in modern organisms.” In fact, he refers to it as a “new paradigm in evolutionary biology” and one that is likely to be very productive. In a forgivable pun he ends by saying, “Let our minds take wing.”

Another example of how developmental evidence can be used to infer what happened in evolution has to do with the bird hand. Hans Larsson at the Redpath Museum, McGill University, and Günter Wagner at Yale, along with others, have been occupied with a problem that is obvious on the surface, but leads to murky twists and turns when you start to look at it more closely.

Certainly development of an embryo has some parallels to the evolutionary history of organisms. And it may be tempting, as has happened in the past, to come up with a just-so story of an evolutionary process that would follow the developmental process that we can see. But how does one justify the conclusion? What counts as evidence? What are the rules of logic and experiment that constrain scientists who want to point to the ways the evolution of the feather or bird hand occurred?

For laboratory sciences the problem is simple. It’s the old scientific method. You come up with a hypothesis and then use experiments to test the hypothesis. It has to be falsifiable so that it can be proved wrong. This works with microevolution, changes in specific genes that we can see. We could hypothesize that if we put bacteria in an environment laced with amoxicillin, the amoxicillin-resistant ones will live and prosper. The bacterial population will evolve to become untouchable by that antibiotic. In fact, this is an experiment being conducted right now in the ears of American toddlers. A Mississippi of pink liquid amoxicillin flows through the nation’s pharmacies and the bacteria that cause ear infections are becoming tougher and more resistant.

We could no doubt find the genes responsible for bacterial resistance and demonstrate evolution in action. But macroevolution occurs over time. The study of how birds evolved, of where mammals came from, of how primates appeared—these issues have to be studied historically. And here, the logic of science becomes a bit different. There is, of course, no proof in science as there is in mathematics. You can prove something wrong. And you can accumulate evidence in support of a theory until it becomes strong and well-founded. But any theory is always susceptible to new evidence, new theoretical approaches.

Hans turns to two ideas as the philosophical basis for his use of developmental stages in attempting to understand evolutionary events.

One is the idea of forensic evidence. Just as coroners determine the manner of death by looking at a corpse, he writes, so scientists can reason the course of evolution by looking at the fossil record.

Another important notion is at the heart of the reasoning that ties changes in the development of the embryo to changes in the shape of animals in the course of evolution. And that is that for a developmental event, a change in how an embryo grows, to be linked to an evolutionary event, a change in the form of adult animals over the course of evolutionary time, the two events have to be of comparable complexity. There has to be a kind of symmetry.

We can see, in the fossil record, how nonavian dinosaurs gave rise to avian dinosaurs and how those avian dinosaurs, the birds, themselves evolved. Along the way a five-fingered hand changed to a three-fingered hand, changed to three fingers stretched into a wing. And we can see the development of the wing as a chicken embryo grows. If we want to draw conclusions to connect the laboratory and the fossil evidence, we need scientific rules of engagement, a clear understanding of what constitutes scientific proof in linking development and evolution.

The symmetry that Hans has argued must exist between the two events is not the supersymmetry of theoretical physics that holds that for each subatomic particle there is a supersymmetrical “swarticle”—requiring squarks, selectrons, and sprotons , all of which may have something to do with the dark matter that seems to make up most of the universe. No. Evolutionary theory may get complicated, but it is not yet ready to match theoretical physics in its complexity.

The symmetry that Hans is talking about is between cause and effect. In this case the principle is that the cause must be as complex as the effect. In practice, what this means is that if you are looking at a change in embryonic development and believe that this developmental event is what caused an evolutionary event, the developmental event must be at least as complex as the evolutionary event.

Keeping this principle in mind, you can propose an idea, a hypothesis for how a cause led to an effect. And you can test it in the laboratory, by making a prediction. For instance, in the case of the bird hand, there has been a debate about how the five digits of early dinosaurs led to three digits in later dinosaurs and finally to what have appeared to be three different digits in birds.

When four-limbed animals first appeared, the evolution of the hand (and foot) was still in flux. An early tetrapod,
Acanthostega,
had seven digits on its hind limbs and eight on the “hand” of the forelimb. The number of fingers and toes was reduced, until the standard body plan of tetrapods specified five digits. Over the course of time some of these digits have been lost or become vestigial in different animals, but in embryonic development, as the hand grows, the beginning of the five digits can be seen. Changes in the course of development result in three-fingered hands, in some dinosaurs, and in birds, although in birds those fingers have elongated and changed shape to form wings.

In observing the development of embryos, the limb buds can be observed. You can watch how, in certain animals, they appear and then are lost during development. In birds, until recently, only four buds had been seen. And the identity of these digits, as established by the conventions of embryology, was a puzzle. The digits are numbered I-V in Roman numerals, going from thumb to pinkie.

The small theropod dinosaurs, like coelurosaurs, that gave rise to the birds had three-fingered hands, and the fingers have been numbered as digits I-III. Birds also have three-fingered hands, of a sort, although the bones in the digits are part of their wings. But they appear to have digits II-IV, according to the observations of embryologists. If birds descended from dinosaurs, this arrangement would not make sense. And some critics of the idea that birds are dinosaurs argued that despite the overwhelming evidence, the digit discrepancy showed that birds could not have descended from dinosaurs.

One piece of contrary evidence does not demolish a larger idea supported by a varied body of evidence from the fossil record. So even if the puzzle of the digits remained unsolved, the descent of birds from dinosaurs would still be the most convincing account of bird evolution. But the puzzle did call out for a solution. Hans and Günter Wagner, a colleague at Yale, worked on the digit problem both together and independently, coming up with an answer that not only solved the problem but demonstrated the way developmental and evolutionary events could be linked.

The essence of their approach was that there are separate stages in development of the chicken hand for which symmetrical events exist in evolution. In development the first is the appearance of the autopodial field, an area or zone of cells that are organizing themselves to create the beginnings of a hand. The second is the growth of digits, and the third the differentiation of the digits into distinct sizes and shapes.

In evolution Hans mapped comparable stages, tracing a symmetry between developmental and evolutionary events. The evolutionary event that he points to as parallel to the stage in development when cells organize to become a hand, is the appearance of fish called tetrapodomorphs. These were the fish that preceded the move to land. They had four fins that look like they were thinking of becoming limbs, so to speak.

The next developmental step is the growth of digits, and the parallel evolutionary event is the appearance among these sorts of fish of digitlike structures in a somewhat jumbled handlike paddle at the end of the fin.

But there is a third stage in development, when these growing digits acquire an identity, a characteristic structure. In evolution that stage occurred with the appearance of four-limbed creatures, intermediate between fish and amphibians. These are tetrapods, like
Acanthostega,
with seven digits on the hind limbs and eight on the front limbs. These digits were different in structure, so that you could distinguish one from the other.

Acanthostega
may not have been able to walk well on land. Its limbs probably helped it to navigate shallows near the water’s edge.
Ichthyostega
was another tetrapod, also a shallow-water creature that may have been able to walk on land. It had seven digits on its hind limbs, of identifiably different shapes. A recent discovery,
Tiktaalik,
a four-limbed fish that has many of the characteristics of later tetrapods, is sometimes called a “fishapod.” Eventually, four-limbed creatures colonized the land and settled on five digits, and all the shapes we see today, including wings, hooves, and the hands of concert pianists evolved from the five-digit hands and feet of our lumbering ancestors.

In both the developmental and evolutionary stages, each step is built on the preceding one, just as with the growth and evolution of the feathers that Richard Prum worked on. Digits begin to grow before they take different shapes and sizes. Just because a digit starts growing in the spot where we might expect the first digit to be doesn’t mean that it necessarily has to become the first digit. That may be the normal course of development, but it can be altered experimentally, and it could have been changed in the course of evolution.

Suppose the bud (anlage) of the second digit appears, but the sequence of HOX genes and sonic hedgehog and bone morphogenetic proteins that would normally turn it into the second digit are altered. Then that bud could turn into the first digit.

That would suggest that the developing bird embryo could start on the path to developing digits II, III, and IV, but end up with digits I, II, and III. Arguments over development can become so elaborate that they are hard to follow, but what this would mean, in brief, is that the evolutionary road from dinosaurs to birds would be cleared up. If, however, this change in development occurred, then a fifth bud ought to show up in development and ought to be in the right place. And a fifth bud had not been discovered until recently, when it was observed and reported by Hans and two other groups as well.

The other groups used different techniques, which were suggestive, but not as definitive as the work by Hans, which tracked the condensation of cells as they developed into buds, and then into digits. Hans and Günter Wagner then joined together to work on interpreting the evidence Hans had developed. They have argued cogently that in the developing chick embryo the growth that begins as anlagen II, III, and IV develop into digits I, II, and III.

This work, however, has importance far beyond the specific case of the bird’s digits. The application of the experimental process to issues of macroevolution, the same process Richard Prum and others have used, marks a new and more rigorous way to understand the past. Paleontology has given us wonderful creatures, dug up from the past. It has provided the raw material for analysis and tracking of evolutionary change on a small and grand scale. It has not provided the mechanism, however. The mechanism of evolution, molecular-level changes in DNA and gene regulation, has been studied in the laboratory but has been restricted to small changes. Evolutionary developmental biology puts the two together, and the result for all of us is a more coherent and detailed understanding of how evolution proceeds.