Missing Link, or Just Jawboning About Ear Evolution?
Tetrapod vertebrates (four-footed animals with backbones) comprise a dizzying array of species, both living and extinct. When is it justifiable to arrange different forms into an ancestral evolutionary sequence, especially when some members are extinct and others are still alive today? On what basis can scientists claim that a discovery demonstrates evolution? Some Chinese scientists entered their latest attempt to exhibit a missing link. They discovered a fossilized “primitive” mammal they claim fills a gap in the evolution of the mammalian middle ear. Their paper was published last week in Nature.1 A press release from the National Science Foundation reproduced on EurekAlert emphasized the evolutionary message: “This early mammalian ear from China is a rosetta-stone type of discovery which reinforces the idea that development of complex body parts can be explained by evolution, using exquisitely preserved fossils,” said Richard Lane of the NSF. The lead author said, “This new fossil offers a rare insight in the evolutionary origin of the mammalian ear structure.” He also said, “Yanoconodon clearly shows an intermediate condition in the evolutionary process of how modern mammals acquired their middle ear structure.” Nothing in the press release indicated anything short of complete vindication for evolutionary theory with this find.
Zhe-Xi Luo et al described a new eutriconodont mammal species, Yanoconodon, they found in the fossil-rich province of Liaoning, China. After providing the customary description and classification sections, they attempted to explain two observations. First and most notable was the structure of the middle ear. They claimed it represents a clear transitional form between the attached bones of mammaliaformes (“mammal-like reptiles”) and detached middle ear bones of mammals. Second, they noted the extra vertebra (26 instead of 19 for most mammals, and 22 for the nearest relative) and the presence of lumbar ribs, unusual for mammals but present in some widely-separated groups. This they explained by convergent evolutionary manipulation of Hox genes that govern the divisions of the vertebral column (sacrum, ilium, lumbar) and the presence or absence of lumbar ribs. Experiments on lab rats show that these traits can be manipulated by knockout or overexpression of these master-switch regulatory genes.
There is no clear evolutionary transition in the vertebral characteristics. The authors noted that Yanoconodon’s nearest alleged relative, Jeholodens, lacks lumbar ribs. Moreover, they found another pair of relatives on a different branch, one that has lumbar ribs and one that doesn’t. For an animal to have 26 vertebra is “exceptional,” they said; the only other one is Repenomamus, a dog-size Cretaceous mammal that preyed on dinosaurs.
The focus of the paper, though, was on the middle ear bones. The authors went into great detail to try to establish Yanoconodon as a transitional form. In the evolutionary scenario, primitive mammals emerged from reptiles with the bones at the rear of the jaw still attached to the jaw. Over evolutionary time, the rear bones began to gradually separate from the back of the jaw. Presumably, this gave some advantage to hearing because the incipient middle ear bones (ossicles), the hammer, anvil and stirrup (or malleus, incus and stapes) were more free to vibrate. Eventually, the ossicles became completely separated from the jaw as in modern mammals and were devoted solely to hearing. The authors identified parts they called malleus and incus, but did not find a stapes.
In support of the story, the authors found that in Yanoconodon the ossicles had partially detached, remaining connected to the jaw only by an ossified Meckel’s cartilage. Further, the malleus and incus bore a resemblance to the completely-detached ossicles of the platypus (Ornithorhynchus paradoxus), a fur-bearing monotreme. They also pointed out that platypus ossicles emerge during development with an attachment to the jaw via Meckel’s cartilage, then become detached later. The arrangement in Yanoconodon, they said, may be pedomorphic – a case of “arrested development” in which the embryonic attachment was maintained into adulthood.
So that’s the story. How good a transition is it? The well-known skeptic and pseudoscience fighter James Randi thought this was ”very cool” as a demonstration of evolution. He gave it a big write-up at his James Randi Educational Foundation where he reproduced the figures from the original paper. The figure he left out, however, is the cladogram (phylogenetic tree), which, surprisingly, shows “homoplasies [convergent evolution] of DMME [definite mammalian middle ear] in basal mammals” in six places on the tree. One of them is Hadrocodium, a lower Jurassic mammal lacking some of the typical features of mammals but having a complex hearing system (see Reference.com). Does this mean that an even less derived species had mammal-like middle ear bones, separated from the jaw? If so, Yanoconodon is too late to be considered a transitional form. Randi displayed some ignorance of modern evolutionary theory by using the pedomorphy hypothesis to conjure up the ghost of Haeckel: “It’s one module in development that flaunts a lovely example of embryonic recapitulation of evolutionary history,” his article boasts. (Recapitulation is dismissed by most Darwinists these days.)
The authors of the original paper did not even claim that Yanoconodon represents a straight-line evolution from attached middle ear to separate middle ear. It represents, rather, a possible pedomorphic characteristic capable of two different evolutionary explanations. The placement of Yanoconodon in an evolutionary sequence also required a tool that maximizes “parsimony” (simplicity) between competing possibilities of phylogenetic trees when all the characteristics are analyzed. In the technical explanation below, keep in mind that homoplasy refers to convergent evolution – similar traits appearing in unrelated groups.
Yanoconodon and its eutriconodontan kin are nested within the crown Mammalia (Fig. 2) by the parsimony of all characters. The absence of DMME [definite mammalian middle ear] in eutriconodonts, an in-group of crown Mammalia, is in sharp contrast to modern monotremes [like Platypus] and therians [more derived mammals, including marsupials and placentals] that have DMME. This phylogeny requires one of the following two evolutionary scenarios: either (1) DMME was present in the common ancestor of monotremes, eutriconodonts and therians; but eutriconodonts re-evolved the middle ear attachment to mandible, or (2) DMME was absent in the common ancestor of monotremes, eutriconodonts and therians, and this is retained as a paedomorphosis in eutriconodonts; but DMME evolved in extant monotremes, and separately in therians. Paedomorphosis, or retention of fetal or juvenile characteristics of ancestors and relatives through developmental heterochrony [differences in developmental rates], is a common phenomenon in vertebrate evolution. The heterochronic (‘premature’) ossification of Meckel’s cartilage in eutriconodonts is the immediate cause for this paedomorphic connection of middle ear and mandible, and is why there is an overall homoplastic distribution among therians (with DMME), eutriconodonts (without DMME), monotremes (with DMME) and pre-mammalian relatives (without DMME) (triangles in Fig. 2). The paedomorphic connection of the middle ear to mandible of eutriconodonts and mammaliaforms is consistent with their lack of the long-bone epiphyses for terminating skeletal growth, as seen in modern mammals.
Clearly, they are favoring scenario 2. The cost, though, is to believe that definite mammalian middle ear (DMME) evolved twice – once in monotremes (platypus), and separately in therian mammals. Maybe that’s why the editors of Nature hedged a little in their praise of this paper:2
The formation of the three tiny bones of the middle ear from components of the reptilian lower jaw was a key event in mammalian evolution. Never before has this transition been seen so clearly as in a primitive fossil mammal found recently in a new locality of the Yixian Formation in China, 300 km west of the classic sites in Liaoning. In this specimen the middle-ear bones remain connected to the lower jaw by Meckel’s cartilage – a transition associated with a corresponding remodelling of the lower back. But the situation is not as clear-cut as it seems. The evolutionary relationships of the fossil suggest that either the ‘modern’ middle ear evolved twice, independently or that it evolved and was then lost in at least one ancient lineage.
The complexity of the situation did not stop the authors from ending their paper on a triumphant Darwinian note. Speaking only about the vertebral column in their fossil, they felt that juggling Hox genes provides “a plausible mechanism for the evolutionary patterns” in lumbar ribs and numbers of vertebra. Again, homoplasy (convergent evolution) comes to the rescue: either the uneven patterns of lumbar ribs in the phylogenetic tree arose because they had similar functions, or were lost in others for the same reason. In conclusion, they felt they had illustrated “two cases for extrapolating the Hox gene patterning of laboratory mice to early mammal phylogeny on a grand evolutionary scale.”
1Zhe-Xi Luo, Peiji Chen, Gang Li, and Meng Chen, “A new eutriconodont mammal and evolutionary development in early mammals,” Nature 446, 288-293 (15 March 2007) | doi:10.1038/nature05627.
2Editor’s summary, “An early look at mammals,” Nature 446:7133.
Evolutionary papers are like contracts: the bold print giveth, and the fine print taketh away. Darwin just handed us a reverse mortgage without telling us the benefits are all coming from a reduction of equity [equity n.: fairness, impartiality, justice]. James Randi performed his glitzy commercial like James Garner, touting all the wonderful benefits of evolutionary reverse mortgages, but we just read the fine print and found the alleged benefits overpowered by serious problems.
Anyone who thinks explanations just “jump right out of the data” should listen to the Philosophy of Science lecture series from the Teaching Company. Jeffrey Kasser shows that it is devilishly hard to prove the simplest scientific statement, such as “all copper conducts electricity” by either deduction, induction, empiricism or anything else. In a related course from the Teaching Co., Steven Goldman shows how philosophers have struggled for centuries with the question of whether the observations of our senses actually tell us anything about the real world. If such basic and simple explanations about things right under our noses that we take for granted cannot be established with certainty, how much less inferences about the unseen past? The issues of proof and explanation are far more difficult and complex than most people realize.
Did you know, for instance, that David Hume in the 18th century argued that induction provides no justification for explanation or prediction, that we do not “see” causes, and that philosophers of science have never answered his challenge successfully? Did you know that Karl Popper essentially dismissed induction as having anything to do with scientific explanation? Did you know that induction is not necessary to discover anything in science, and that deduction has about as many problems as induction? How about the facts that philosophy of science since the heady days of logical positivism in the 1930s has become a welter of conflicting opinions between realists and anti-realists and every position in between, with no one being able to define with any justification what constitutes an observation, let alone a scientific explanation or theory?
It would do you good to struggle with some of these issues for awhile before evaluating a paper claiming to have found a transitional form in Darwin’s storyland.
Surprisingly, the usual Darwin propaganda outlets who usually go ape over every missing link story didn’t seem to pay this claim much attention. Maybe they realized critics could easily shoot it down. We’re going to display this as an example, though, to educate our readers on how to evaluate such claims in general. OK, so a four-legged fossil animal was found. There are lots of four-legged fossil animals. What right does anyone have to arrange them into an ancestral sequence? What justification is there for the implication that complex hearing bones were evolving by a naturalistic, aimless, pointless, purposeless process, with advanced mammalian hearing as the product? If you take away the evolutionary assumption, no such inference jumps out of the observations.
To see why, visualize a 3-dimensional plot with data points scattered throughout like dust particles in the air. Here and there, some seem to cluster together, but no obvious trend reveals itself. How can one justify drawing lines through the dots that show an ancestral tree pattern? An almost infinite number of patterns could connect the dots. What right does one scientist have to claim his pattern is the “true” pattern that tells what actually happened before there were observers?
One needs to understand the Darwinist process of drawing phylogenetic trees, because the story really breaks down right there. These scientists used a “maximum parsimony” computer program to produce their tree. Why did they select that? They could have chosen instead to use “maximum likelihood” or Bayesian analysis, or some yet-to-be-discovered new method that will enjoy fad status for awhile. Even so, the method they used required choosing between 218 equally parsimonious trees. For those who need proof how arbitrary this process is, read the following quote from the paper:
This is based on the strict consensus of 218 equally parsimonious trees of PAUP (Phylogenetic Analysis Using Parsimony and Other Methods, version 4.0b) analysis of 436 characters (1,000 heuristic runs with unordered multi-state characters) that can be scored for 102 comparative taxa (97 mammaliaforms including 25 extant mammals, plus three cynodonts as outgroups). For each equally parsimonious tree, tree length = 2,188, consistency index = 0.375, retention index = 0.803.
What would happen if they included different organisms as outgroups, or lightened the consistency requirements, or used some other criterion for establishing the consensus? How much was the result dependent on the way the traits of the various organisms were described by others? To what extent were evolutionary assumptions embedded in all the data to start with? Anyone thinking that The One and Only True Tree came out of this exercise is a good customer for the next beachfront property sale on the Isle of DeBris.
We have reported previously (07/25/2002, 10/01/2005) that all such trees are compromises between contradictory data points, and therefore represent human assumptions imposed on the data, not independently-verifiable patterns that exist “out there” in some value-free, assumption-free Logicland. This is an important fulcrum of this discussion; all the team’s inferences about evolutionary transitions hinge on the validity of their phylogenetic tree. But their tree is a product of human imagination, not a fact of history. Understand that and the whole paper falls apart. The only thing that is left is an unusual fossil that fits who-knows-where into some scheme by who-knows-who about who-knows-what that happened who-knows-when. Insert your guess here. It can be considered just as valid as anyone else’s, including the one in this paper.
Suppose you sent four different teams into a grocery market on a mission to read the ingredients on every box of processed food and arrange them all into a phylogenetic tree. The hapless victims would soon find themselves making arbitrary decisions among a mess of complications. One group might try to locate all the products containing lecithin, while another would use statistical analysis based on percentages of recommended daily allowances of vitamins, carbohydrates, fats and proteins. A third team might approach the problem by sorting everything according to texture, color, odor or some other physical characteristic, while the last team would feel confident they had the final answer, because a clear phylogenetic pattern emerged when they arranged the products by geometrical shape. Cheerios evolved from granola, they say.
Each team has its own concept of a phylogenetic tree, and has plenty of evidence to support it. Which one is right? Clearly, none is right. In this contrived case, all the products were created by intelligent design. Arranging designed items into an evolutionary pattern is an exercise in futility. General Mills can produce both cheerios and granola. But even if nobody knew the products were designed, arranging them into an ancestral tree pattern would be hopelessly messy. Group one might get a good-looking pattern arranged by lecithin content, only to find that separate branches have members with and without coconut. Scratching their heads, they might wonder if the coconut evolved by convergent evolution (homoplasy), or whether the coconut was present in the common ancestor but was lost in various lineages and retained in others. No amount of energy, enthusiasm, diligence and detail is going to justify what inferences these researchers are making. They are just playing games. Data notwithstanding, the patterns they are producing are nothing but mythical constructions of their own imaginations.
This illustration helps answer a counter-argument from some Darwinists. They complain that every time they produce a good transitional form, creationists allege that two more are produced: one on each side of the transitional form. Such a complaint is bogus, because it embeds a prior evolutionary assumption. It presupposes a linear sequence. The “missing link” terminology misleads the reader into visualizing a straight line, with an early “primitive” animal, a later “derived” animal with complex features (like detached middle-ear bones), and a gap in between that the evolutionist triumphantly fills in with his new transitional form. Toss out that picture, because it is nothing like what is found in either the fossil record or the zoo. Recall the supermarket or the 3-dimensional plot with scattered data points. No amount of new data points or processed food products added to the mix will justify inferring an evolutionary pattern. It cannot be done. Unless you already assume that the data points or the supermarket products evolved, you cannot convince a critic that the data imply an ancestral relationship, and you could never convince a well-trained philosopher of science that you are justified in making such an inference. Conversely, there is no way an evolutionist can deny the creationists’ inference that God created a wide variety of creatures, and gave each the kind of bones and ears and ribs and toes that it needed for its ecological niche. Bring up plesiomorphies (close similarities), and he could answer that a certain amount of sorting out of characters occurred since the creation, but no new information was added. Prove him wrong.
So is Yanoconodon a transitional form on the “grand evolutionary scale” of life? On what basis could anyone make such a claim? There are plenty of contradictions in the story. For one thing, if Hadrocodium already had a DMME, then Yanoconodon surely wasn’t inventing it for the first time. Considering this skeleton to be a new species is fraught with difficulties in the first place. You can’t tell whether a fossil is reproductively isolated from other fossils. How do they know it was not some other animal whose skeleton got deformed in the fossilization process? What is a “species” anyway, if not a human concept imposed on a bewildering array of complicated data? Why do some of its features show up in some distant branches, but some not show up in its alleged closest relative? How do they know the developmental stage of this fossil? Maybe it was a juvenile and its ossified Meckel’s cartilage had not yet dissolved. How much can you tell about the morphology, lifestyle and complexity of an animal from its bones? Imagine the surprise if these scientists had classified a platypus knowing only the skeleton, and then were shown a living one, with its duck bill, poison spur, electrical sensors, specialized fur and numerous other advanced characteristics.
The paper is filled with jargon (plesiomorphy, apomorphy, pedomorphy, heterochrony, homoplasy), but no amount of hand-waving justifies their interpretation. Jargon is, at best, a convenience for those playing the game; at worst, it is bluffing and obfuscation. Most important, their story is based on the flimsiest piece of ossified cartilage and says nothing about how a complex system of hearing, employing finely-tuned ossicles, evolved. The complexity of mammal hearing is staggering. The ability of tympanic membranes and ossicles and cochleae to amplify the minutest pressure waves in air is breathtaking in its efficiency. But then, the body has to be able to transmit that information to the brain, process it and respond to it. Hearing cannot be evaluated on the basis of an ossified bit of cartilage. It must be understood in context of the whole process of hearing, and how it fits into the lifestyle and survival of the entire organism. Darwinian theory disallows teleology or goal-seeking behavior. A consistent Darwinist must erase from his theory any suggestion that these tetrapods somehow “wanted” to evolve advanced sound systems by modifying their jaws. Each tiny alteration, wrought by mistake, would have had to confer a selective advantage with sufficient benefit to outcompete all the population lacking it. Remove the assumption of evolution and the story becomes implausible to the point of absurdity.
There is a somewhat trusted method of scientific reasoning called inference to the best explanation. It is not infallible, but applied to the complex design of the mammalian ear, would certainly prefer the design explanation over the chance explanation. A corollary would be that simpler animals would be designed with simpler equipment. This would provide a “top-down” approach, similar to explaining supermarket products as all designed, with some complex products designed for formal dinners and simpler ones for snacks. Scientifically and logically, the “top-down” explanation is just as valid as the “bottom-up” explanation. For those not already blinded by Darwin dogma, most people, we expect, would feel the design inference much better supported by the evidence.