The strength of any theory comes not from a single measurement or a single confirmed prediction, but from the theory’s many predictions being confirmed by many independent tests, samples, and methods. Often, attacks on evolutionary theory take the form of showing that some measurement technique is not infallible, or that some measurement technique depends on assumptions that could be wrong. Examples, of course, include creationist criticisms of radiometric dating, which is a common technique for dating rocks.
When the creationist adds to these types of criticisms the charge that there are some things evolutionary theory cannot explain, it can seem to some that evolutionary theory is a weak, speculative hypothesis. But as we will show below, scientific theories aren’t strong because the measurement techniques that confirm them are perfect, or because they have no open questions. If this were the case, all science would be weak and speculative.
The best way to illustrate the fundamental problem with the creationist critique is by way of an analogy. Let’s say I’m the prosecuting attorney in a felony case, and my case depends on placing the suspect in a certain location between 4:55 PM and 5:05 PM, during which the suspect shot someone. Now let’s say twenty of my witnesses happened to look at their watches (a combination of analog and digital watches) when they saw the suspect shoot. Seventeen other witnesses separately recall seeing the suspect shoot just as the 5 O’clock news was coming on in each of their apartments in the neighboring building. Twenty six other witnesses saw the suspect shoot just as the 5 O’clock whistle blew at the nearby quarry.
Now, imagine the defense carefully pointing out that watches have been known to be wrong. Watches, he stresses to the jury, are not infallible. He proudly brings in people to testify that they have made mistakes at some point in their lives about the time of day because their watch’s battery had run down, or their watch had gotten wet, or damaged by various means. As for the 5 O’clock news, the defense points out that scheduling errors in TV programming have been known to cause shows to come on more than five minutes after or five minutes before their scheduled time, and he even finds a TV producer to testify about a time that this did, in fact, happen. And as for the quarry, the defense was able to find several expert witnesses (people who blow whistles at quarries) to testify that there had been times when they blew the whistle at more than five minutes before or after the scheduled time, at each of their respective quarries.
As is obvious in this example, it is not enough for the defense to show that the methods used by each of the prosecution’s witnesses are fallible ways of telling the time. Not only is it extremely unlikely that something was wrong with everyone’s measurement of the same event, but even if everyone were in error, we would hardly expect them to be in agreement with each other—especially when the measurements involve unrelated technologies and methods. Creationist criticisms of the evidence that support evolutionary theory or modern geology (e.g., dating rock samples) are often analogous to this defense attorney’s criticism of the evidence supporting the prosecution’s case. To take the defense’s (and creationists’) criticisms seriously would literally mean that we could not tell time at all simply because watches are individually imperfect. But in science that fallibility has been specifically taken into account through the use of multiple samples and multiple independent measurement methods. Just like our imaginary prosecutor, scientists look for independent corroboration before they consider a theory to be well supported.
In the case of rock dating, and in all significant aspects of evolutionary theory, we have just this type of wide agreement spanning many different methods and many different samples (far more samples and methods, in fact, than described in the courtroom analogy). This first article will focus on just five of the many independent, corroborating lines of evidence that confirm evolutionary theory, with a focus on one of its key hypotheses: Descent with Modification. The definition of this hypothesis (taken from the subject refresher article on Evolution in Freethought Debater) follows:
The descent with modification component of evolutionary theory is that all life forms can trace their lineages back to earlier classes of life forms in a branching, “nested” hierarchy (forming what looks like a bush or tree, as in the “Tree of Life”), which can ultimately be traced back to the beginning of life on earth…During the long period since that time, changes in body plans have accumulated in diverse directions to make all the differences we now see between all life forms. To continue the tree analogy, you can start from the tip of any arbitrarily chosen twig and follow it back to the point where it joins another twig, the point of origin of the two (or more) twigs. That now thicker twig or branch can then be traced back to where it joins another branch. This process can be repeated until you ultimately reach the trunk. Applied to evolution, each twig and branch represents a particular lineage; the points where there is a joining of those twigs and branches represents the common ancestor of all the lineages that can be traced back to that point.
The importance of recognizing the cross checking nature of science cannot be overstated. When one forgets that this is the basis of good theory, one can quickly lose sight of the “forest for the trees,” and literally end up haggling over irrelevant individual cases in which a result was unexpected or inconsistent with the theory. However, some exceptions are, in fact, expected so long as they can be shown to be “outlier” results (random or statistically rare errors) that do not significantly affect the conclusion drawn from all of the data taken as a whole. This is exactly why statistics plays such a large role in analyzing scientific data, and why scientific results are usually stated along with a statistical level of certainty.
It is perhaps understandable why creationists make these sort of irrelevant criticisms of outlier results. Unlike actual science, creationism does “not seek organizational relationships or look for relationships in terms of universal physical laws.”1 Instead, by definition, they exist to defend a point of view—a view rooted in absolute certainty in an unchanging Truth (with a capital T).
The lines of evidence we will briefly look at are the fossil record in the geologic column; the classification of living forms based on comparative anatomy; biochemistry; embryology; and finally, biogeography. As you review the evidence notice how the independently corroborating nature of the different lines of evidence precludes any appeal to “there was this one fossil they couldn’t explain” or “someone once made a mistake in classifying an animal,” or even, “there was once this hoax. . .” As in the court case example, the proof is in the wide statistical agreement of many measurements spanning all of these methods, not in any single data point.
The geologic column, which is the identification and classification of the different rock layers (strata), was essentially completed by 1815, almost 50 years before Darwin’s theory. Importantly, this work was done by the creationists of the time. They noticed that each rock strata contained a distinct collection of animals and plants. So unique and consistent were these fossil collections in each of the layers, that certain fossils in them could then be matched to fossils in other continents to locate that layer within the geologic column (index fossils). It is important to notice that this all predates Darwin and is completely independent of any assumptions about evolution. Inexplicably, a common creationist complaint is that the geologic column presupposes the truth of evolution, which in turn presupposes the truth of the geologic column, resulting in a circular argument for evolution. But as this very short history should already make clear, the geologic column makes no assumptions about evolution since it was established by before there even was a theory of evolution.
By comparing the fossil contents from the lower layers (or strata) to the higher ones, these pre-Darwin creationists (such as Cuvier, the father of paleontology and a deeply religious creationist) could see a pattern in which an individual life form would appear in one layer and then be replaced in the next layer up by multiple variations on that original, single form. These layers were often themselves followed by layers in which all or most of those variations (and variations of variations) suddenly and completely disappeared. These creationists also noticed that the deeper the fossil, the less recognizable it was. As one moved higher, the forms became increasingly recognizable. As these scientists examined more and more strata, this cycle of emergence followed by branching variations continued.
Again, this pattern was recognized before Darwin’s theory, though the pattern was assumed not to reflect any kind of relationship through descent. Based on their observations, these creationists (and basically all geologists were creationists at this time) concluded that God must have inflicted not one, but a series of cataclysms, each followed by a new creation. However, as they collected ever more data, they realized that there had to have been ever more of these cataclysm and creation events.
To recap, at this point in pre-Darwinian history the geologic column had been established with the lower layers understood to be older than the higher layers. The actual ages of the layers remained unknown, however, but were assumed to be consistent with a literal reading of the Bible. Also, at this point in history, it was apparent that a slice through the earth—like cutting through a layer cake to reveal the layers—revealed not a hodgepodge mix of living and extinct forms (as one might expect from a world-wide deluge), but an extremely ordered and consistent pattern of fossils throughout the world. That ordered and consistent pattern was that a form would appear in one layer, and in higher layers that same form would be replaced by similar though different forms, which became progressively more different in even higher strata. This typically ended in the wide-spread disappearance of many of the forms, and then this cycle of appearance followed by a kind of “radiation” would start over. One thing that was also apparent even then was that the higher the layer in which a fossil was found, the more recognizable the fossil usually was, while at lower layers the forms were less recognizable, and harder to tell apart and categorize. For example, it is easy to tell mammals from reptiles today, but if one goes deep enough, mammals and reptiles become essentially indistinguishable (e.g., you find groups like the Therapsids, which were one of the mammal-like reptiles that blended features of both groups). Creationists, like Cuvier, argued that the data could be explained by a series of divine cataclysms and creations, the last of which was the Biblical Flood.
Classification of Living Forms
Setting aside fossils altogether for a moment, but keeping the observed pattern we saw in mind, let us separately look at what happens when we classify today’s living forms based on their physical forms and structures (called “morphology”). A fair question to ask before classifying living things is, “What characteristics should we use: size, weight, what?” Well, the pre-Darwinian, Linnaeus, who came up with the animal naming system we still use today, grouped animals by overall large-scale anatomical similarity, though this left some room for arbitrary decisions about what should be considered “similar.” Later, the anti-Darwinian, Richard Owen, argued that if a feature could be shown to be the same structure modified for different purposes—as revealed by comparative anatomy and embryological development–then the animals should be classified closer together. The more such structures were shared, the closer would be the classification. He called such structures “homologous,” while structures that looked superficially the same, but were based on completely different structures and embryological development, he called “analogous.”
This homology / analogy distinction works very well because by using it one can predict other things that the animals would also have in common. This approach allows one to gain new insights. If you had classified by size alone, say, you wouldn’t gain any additional insights, and would know little more than the just size of the animal. It’s interesting to ask why this approach makes for such effective predictions. For Owen, and the creationist establishment of the time, the answer was that homologous structures revealed part of God’s plan: He used relatively few basic templates that He modified to create all the species. Different species that shared homologous structures were based on the same template, but each had customized modifications to meet the functional needs of the individual “kinds.”
Owen’s approach, however, also produces observations that conflict with expectations based on God’s having created each kind directly. If each kind were separately created, then there would be no restriction preventing God from mixing and matching useful structures. If God designed a structure to serve a purpose, then all species could benefit from that originally perfect design. There would be no constraint on God that says once He designs a useful structure, He can only give it to other species that happen to share with that first species a set of completely unrelated characteristics. For example, there would not be a rule that says God can only give three middle ear bones to species that have milk-producing (mammary) glands.
An all-powerful Designer should have at least as much flexibility as human designers. Things designed and manufactured by humans show no such restrictions. Design ideas are shared across widely different “kinds” of human creations. This is quickly apparent when you try to create a classification of modern computers or aircraft based on their shared components. For example, Global Positioning Satellite (GPS) technology can now be found on helicopters, biplane crop dusters, high performance fighter aircraft, and even rental cars and fishing boats. If you had earlier created a hierarchical classification that had winged aircraft as the basis for one branch and wingless (e.g., helicopter) aircraft as the basis for another, you would now have to add GPS to both branches. In other words, GPS technology would not be nested within just one branch, but would cut across the branches. As a result, one cannot create a stable, hierarchical classification for human-created things in which sets of characteristics are nested one within the other.
On the other hand, if each species is related through common descent, and not individually created, then we would expect that a classification of them based on shared structures would reveal just such a nested hierarchy of structures. We would expect this since the structural inheritance of each branch is different and any new structures have to be built from the materials at hand; that is, they have to derive from these inherited “components.” As a result, any new structures would be confined to the one lineage in which it first appeared, and to its descendent branches. Consequently, we wouldn’t expect to see structures cutting across the branches, as we did in the GPS example.
When we classify living life forms we see the nested hierarchy predicted by Descent with Modification. For example, consider the following classification (cladogram)2 of a few representative, but disparate forms:
Descent with Modification also makes some additional, more specific predictions about the morphology of living forms. First, since species can only work within their structural inheritance to solve adaptive problems, we would expect that as descendent species in one lineage branch out into different niches (some that involve flying and others swimming, for example), that their adaptations will involve different modifications of the same underlying structures. These are the homologies we discussed earlier, and they are a specific prediction of the Descent with Modification hypothesis. A corollary prediction of this view is that we would never actually find the “winged horse” seen in mythology, since that would involve entirely new appendages appearing “out of thin air,” as it were, rather than coming from existing structures. This is why, for example, flying birds and mammals had to modify two of their four inherited appendages to get wings; that is, they had to modify their “arms” to make wings.
Second, where species from different lineages find themselves filling the same environmental niche (for example, living in different parts of the world and filling the bamboo-eating niche) we would expect to see cases of different structures being used for similar functions (as opposed to similar structures being used for different functions in the homology example); that is, we would expect to see different lineages solve the same adaptive problems with different structures—structures which reflect each lineage’s unique inheritance. Such cases are examples of “analogies,” (or “convergent evolution” as it’s often called today) and are also predicted by Descent with Modification. But a further interesting prediction of this theory is that some analogous structures will be inferior to others in fulfilling similar functions. This is expected under this theory, since the inherited components of some lineages will make better designs easier to “get to” than the components inherited in other lineages.
In the case of homologies, we see exactly what is predicted. For example, the forelimbs of all mammals are composed of the same bones arranged in similar ways. It’s only the proportions of these bones that differ. For example, compare the bones of bat, human, and dolphin. While each fills a very different niche, and uses these bones for very different functions, they all have the scapula, humerus (upper arm), radius (forearm bone 1), ulna (forearm bone 2), carpals (wrist), metacarpals (hand), and phalanges (fingers). Again, only their proportions differ. Creationists often say that there are similar structures for similar functions, but if that’s true, then why does the bat’s wing have far more structural similarity with the human hand than it does with a bird’s wing?
In the case of analogies we see again exactly what is predicted. For example, the shark (a fish), Ichthyosaur (an extinct swimming reptile), penguin (a bird), and dolphin (a mammal) all have forelimbs adapted for swimming. Outwardly, these forelimbs look very, very similar; however, the internal structures are radically different: the dolphin’s fin has more in common with the bat’s wing and the human hand (including the same five finger bones) than it does with the Shark’s purely cartilaginous fin (no bones at all). This makes sense in the evolutionary context given that the genetic inheritance in the fish lineage didn’t have bone to work with when the sharks first appeared. So, again, the Creationist claim of “similar structures for similar functions” is falsified. What we typically see is different structures for similar functions and similar structures for different functions.
The Australian marsupials are a particularly startling example of this kind of convergent evolution (different structures adapted to solve the same functional needs). There are marsupial versions of mice, flying squirrels, moles, ground hogs, rabbits and wolves, just to name a few. As different as they are from each other, they still have more in common with each other than to any placental mammal. For example, the marsupial mouse has more in common with the marsupial wolf than it does with the placental mouse with which we are all familiar. (This Australian example is discussed in more detail below.)
Owen’s supernatural “cost-control” explanation for these patterns is an ad hoc theory tacked on after the fact to explain data that contradicts the predictions of a theory based on an all-powerful Designer creating each “kind” separately. Worse, his explanation requires one to contradict God’s omnipotence and omniscience. People such as Owen apparently felt that God used relatively few templates for reasons of efficiency in the design process, in the same way that GM, for example, would not want to design each car from scratch due its being an inefficient use of limited and expensive resources. Of course, it is entirely unclear how one is to reconcile this explanation with the notion of an all-powerful, all-knowing, and perfect Creator, who presumably isn’t operating under such resource constraints. Unless there was a second act of creation after the “Fall” (one in which inferior designs were introduced) then one can only conclude that many “kinds” were rendered imperfect from the very beginning (i.e., before the Fall), since many of the analogous structures we see in the different “kinds” are not equally effective in serving similar functions.
An all-powerful and perfect creator who created each kind directly would be expected to have originally used the optimal design when the same function was needed. If any inefficient functionality that we see today was due to “degeneration” following the “Fall,” then this would appear as resulting from defects appearing in a common structure, not from the use entirely different structures, which would reflect a choice that that had been made from the very beginning.
Relationship to Fossil Record and Descent with Modification
Now, let’s go back to the fossil record. If the nested hierarchy we see in living forms reveals the particular design approach of a Creator who made all “kinds” at the same time, then there should be no connection between the characteristics of life forms (like having 3 ear bones) and the depth of the rock layers in which these characteristics first appear. In other words, if all species were created at the same time, then there should be no correlation between elapsed time and the appearance of characteristics that define each species.
On the other hand, if all life is related through the process of descent with modification, then there should be a very specific correlation between time and characteristics; which is to say, there should be a correlation between the depth of rock strata, and the first appearance of structural characteristics in the fossils of those layers. How should they correlate? Remember that a structural characteristic that is shared by more branches, such as jaws or a vertebral column, should first appear in layers that are deeper than those containing the first appearance of any nested characteristic. For example, the earliest appearance of animals possessing 3-ear bones (ossicles) should appear higher in the rock layers (i.e., in younger rock formations) , than the earliest appearance of animals with vertebral columns.
This pattern is exactly what is found in the fossil record. Referring back to our earlier tree with the lamprey, perch, lizard, mouse, and cat (which was made independently of fossils) what we find in the fossil record is that the jawless fishes (agnaths) first appeared in the deepest layers, followed in higher layers by the first appearance of bony fishes, followed in yet higher layers by the first appearance of reptiles, which in still higher layers were followed by the first appearance of mammals. The theory of Descent with Modification not only predicts this, but is alone in being able to explain it in a productive way (that is, based on a theory with testable predictions and not by invoking Divine, mystical, or magical causes and purposes, which can be made to fit any data—which, of course, means it can explain no data).
Note that I said “first” appearance. Obviously jawless fishes are still with us, as are the reptiles, and even the much earlier forms of life like blue-green algae. Too often one hears comments like, “…but if we evolved from them, then why are they still here.” It should be clear from the discussion so far that such a view represents a serious misunderstanding of descent with modification and evolution. A parent species does not have to go extinct before a child species can emerge. To take just one simple example, a child species can emerge when the parent species is split into two by the disappearance of, say, a land bridge, creating two islands where there had been just one. The two groups might now diverge into two species, though one of the species may show almost no change from the original population, while the other shows dramatic change. This could be the result of very different environmental pressures on each of the two groups.
To recap so far, we see that a tree of life made from comparative anatomy of living forms, which has nothing to do with timing or fossils, independently correlates to a tree of life based only on the appearance of characteristics as you go from deeper to shallower rock strata.
Now if Descent with Modification is the right explanation of the these two independent phenomena, (i.e., the pattern in fossil record and the pattern in living organisms) then any newly discovered features should continue to corroborate these same patterns, while perhaps adding even more detail and filling in open questions.
Since Darwin’s time, the life and geological sciences have exploded, complete with the emergence of whole new sciences that Darwin could never have imagined, such as molecular genetics, plate tectonics, and geochronology (rock dating). Any good falsifiable theory should not only hold up under the scrutiny of these new sciences, but thrive and contribute to this progress of knowledge. Looking at the new field of molecular genetics, what happens if we try to make yet another “tree of life” diagram, but based only on DNA, independent of anatomy and independent of the fossil record? If we end up with a tree of life that is completely different from the diagrams we got based on fossils and comparative anatomy, then Descent with Modification will have a big problem.
Briefly, DNA is a molecule on which is strung a code made up of just a four-letter alphabet. For example, part of the whole string might read, “…GCCTTACGGA…,” where the whole string is actually about 3 billion letters long in human DNA. When that code is read across the entire DNA, one ends up with a recipe for a growing a life form. This four-letter alphabet really represents the four chemical “bases,” each of whose initials are A, T,G, and C (adenine, thymine, guanine, and cytosine).
Until the DNA copies itself (a process called “replication”), this string of letters exists as one side of a two-sided zipper-like structure called a double helix. The other side of the zipper is the mirror image of the left, where T is the mirror of A, and C is the mirror of G. When a cell divides, this zipper unzips into two halves. Each half then builds a new sister half, letter by letter: Ts attract As, As attract Ts, etc. When all is done and said, we end up with two complete “zippers,” each identical to the original single zipper before it split.
Now, each group of three letters makes a “word,” (e.g., CCA is a word, or UCU, etc.). These three-letter words are called “codons” and come in two basic types: instructions for building things, or exons; and regulatory instructions, or introns, which can sometimes be thought of as punctuation marks or “stop reading,” “start reading” instructions. In the case of exons, each three-letter word corresponds to a particular building block of a protein molecule. These building blocks are called amino acids. A protein is essentially a string of amino acids in a particular sequence all hooked up end-to-end. Proteins in living things are made up of very long sequences of just 20 different types of amino acids. The DNA code for making the whole protein is an example of a “gene,” each of which can be thousands of letters long.
A genetic mutation amounts to A, U, G, or C (actually the U stands for urasil, which is used instead of T in RNA, where RNA is an intermediate step before the protein-building process) being accidentally “flipped” to something else, or having a one or more letters deleted, added, reversed, or put somewhere else in the sequence. For example, if the word UCA becomes UCG, then there is no effect since they both code for the same amino acid: serine. However, if the first position is flipped, turning UCA into GCA, then the amino acid in this part of the completed protein will be alanine instead of serine. This change to the protein may or may not affect the way it functions or the final “fitness” of the completely grown life form.
Some mutations may prevent the protein from even forming at all, such as a mutation affecting one of the “punctuation marks.” If this happens, the entire gene of some thousands of those letters (A,C, T, U) are rendered useless, or “dead.” (Of course, from the standpoint of affecting offspring, and therefore evolution, the only mutations that matter are those that affect the DNA copies that go into the sex cells, since only they are going to be used to grow another organism.)
In the interest of space, I’ve tried to limit this introduction to genetics to only those points bearing directly upon our discussion. Hopefully I’ve described enough of it to make the significance of the next point clear: only around 5% of DNA actually codes for building anything, the rest are the introns along with a vast amount of dead or “pseudo” genes. This means that most random mutations will have no impact because they will land in this vast sea of ignored dead genes. (All introns, including these dead genes, are actually cut out to leave only the exons just before the instructions are read prior to building proteins.)
Keeping in mind that some 95% of the genetic code is nonfunctional, and also keeping in mind that random mutations appear on a fairly regular basis, we can expect that mutations affecting these dead areas of the DNA will tend to spread throughout the breeding population. The reason they will tend to spread is that specific mutations are very rare, and they have no effect on survivability (since the dead genes don’t do anything). Consequently, as the creature possessing this mutation reproduces, it will likely pass it on and it will multiply over the generations. If one could find very old DNA of a particular species, and compare it to new DNA from that same species, one would expect the newer DNA to show an accumulation of these types of mutations. Even more importantly, if that earlier population had split into two, say by a land bridge being cut creating two islands from one, then each of the two descendant populations should be accumulating mutations independently of each other—getting more and more different from each other over time—since mutations in one group cannot spread to the other group. The bigger the accumulated difference, the longer the time they have been separated.
Earlier we created a tree based on comparative anatomy, and saw that it corresponded closely to a tree independently built from the fossil record. In that example, we looked at representatives from disparate branches, but similar anatomical and fossil classifications of just the primates invariably show that the great apes are very close to each other and to humans, with monkeys being more distantly related, and lemurs further still (i.e., “more distantly related” means it represents an earlier split, in the same way that the perch, in our earlier diagram, represents a lineage that split off prior to the lizard’s lineage).
With our understanding of genetic mutation and of the large amount of non-functional DNA we can make yet another “tree of life.” One way to build a tree based on genetic differences is to “unzip” the chromosomes of different species, and take the left half of one species and see how well it bonds to the right half of the other species. Naturally the bonding will be 100% for the same species. The greater the genetic similarity, the tighter the bond (measured by heat required to re-separate them). Here are the results from this DNA-DNA binding technique 3:
|Species||Percent DNA Binding|
This certainly isn’t expected if each kind were separately created, especially when you consider that the vast majority of the sequences have nothing to do with function. The dead genes, however, provide much more powerful corroboration of evolution than does this impressive result.
Shared Typographical Errors
The discovery of the high proportion of dead or “pseudo” genes provides corroboration of the Descent with Modification hypothesis in a startlingly different way. To see how, let’s start with an analogy.
Imagine you are a teacher grading essays. The essays are each a response to the same question you posed to the whole class. You, of course, made clear to the students that they are to work separately, and that they are not to copy each other’s work. So how do you tell whether matching passages in some of the returned essays were plagiarized or were just the result of a coincidentally similar choice of words in response to the same question? What if 30 consecutive words were identical? Well, that certainly is not likely, but what if in addition all the punctuation marks also matched exactly? Well, now things are looking even worse for student honesty; but what if in addition to all of that, grammatical, spelling, and punctuation errors match exactly between the two passages and in exactly the same locations? Well now the odds of a coincidence are so small as to be considered zero. An extension of this kind of analysis can also reveal whether one person copied from another, or if two students each separately copied from the same third source, and whether each of those “child” copies were themselves copied by others, etc. The “junk” genes, of which so much of DNA is made, allow just this type of analysis on gene “copying” between species.
As mentioned earlier, mutations can prevent a gene from making anything at all. A gene is just a long run of code (often made up of thousands of the code letters A, T, C, and G) that when read produces, for example, an entire protein. If mutations occur at critical locations , then that protein may not be made at all. Such a mutation would render the whole string of code, of some thousands of letters, “dead.” Since only a few errors in a string of thousands can make a gene non-functional, then an analysis of the string can often reveal what the gene originally coded for.
For example, humans and the other primates require vitamin C in their diets. In humans and primates, vitamin C (ascorbic acid) deprivation will lead to debilitating diseases such as scurvy. “So what?” you ask? Well, this dependence on vitamin C is the exception among mammals, since, besides primates and the guinea pig, other mammalian species produce an enzyme protein that allows their bodies to synthesize their own vitamin C. This enzyme protein is called LGGLO (which, if you must know, stands for L-gulono-gamma-lactone-oxidase). Now, the LGGLO gene that codes for this protein has been identified.
Under the evolutionary hypothesis all mammals inherited the LGGLO gene from a common ancestor. Any mutation that would render this gene nonfunctional in any of this common ancestor’s descendent species would be pretty rare, and when it did occur it would be fatal unless the gene were no longer needed, such as if the diets of those affected species just happened to be rich in vitamin C. Now what if one of these descendent species develops a mutation in the LGGLO gene that “kills” the gene—that is, makes it a dead gene? Well, as I mentioned above, this would only be a problem if the diet of this species is not rich in vitamin C. Bu what if this species with a dead LGGLO does have a diet rich in vitamin C, and it branches into multiple descendent species of its own over the course of evolution? Well, now we have a testable prediction. If Descent with Modification is true, we would expect that this same dead gene would appear in each of those descendent species. In the case of the primates those descendent species are alive today, and include us. The prediction is confirmed: that broken relic of a once working LGGLO gene has indeed been found in humans and in the other primates. 4
Now humans and the other primates are believed to share a recent common ancestor based on an enormous array of other converging lines of evidence that have nothing to do with the LGGLO gene—such as morphological, other genetic, molecular, and fossil lines of evidence. So the additional fact that humans and primates as a group need vitamin C in their diets, and the fact that this condition is extremely rare among the mammals, suggests yet another very specific testable prediction—namely, that the particular genetic “typo” that makes this LGGLO gene “dead” would be the exact same typo in humans as it is in the other primates. The reason we would expect this is that any number of potential defects can “kill” a gene, so if the defect occurred independently in each primate, then it would be extremely unlikely to be the result of the same typo each time—and it would certainly be completely mysterious why such a cluster of independent events would target the primates as a group when it is extremely rare among all mammals. This “identical typo” prediction for the primates/human relationship has now been confirmed: “A small section of the GLO pseudogene sequence was recently compared from human, chimpanzee, macaque and orangutan; all four pseudogenes were found to share a common crippling single nucleotide deletion that would cause the remainder of the protein to be translated in the wrong triplet reading frame.” 5
It is important to see that the strength of this prediction in no way depends on primates being the only group of species with a dead LGGLO gene. The strength of the prediction comes from two powerful facts: first, that the mutation is extremely rare among mammals but found in all primates as a group (why would this be if the primates are not related?); and second, that for primates as a group the gene is not just dead, but dead from precisely the same genetic typographical error. Now it shouldn’t be very surprising if we find a dead LGGLO gene in some other mammalian species that does not share a common ancestor with the primates, so long as it is a rare find. This is so because it is certainly conceivable for a mutation that kills the LGGLO gene to appear independently in another lineage; however, the strong expectation is that if it did occur independently, then the genetic typo involved would be a different typo from the one found in the primates. This is the expectation, for example, in the case of the guinea pig, which is an example of another rare occurrence of a mammalian species with a dead LGGLO gene.
So the LGGLO case is yet another powerful independent line of evidence that converges with so many others. The creationist has to explain not just how each of these lines of evidence might be wrong, but how, if they are wrong, they all point to the same answer. Convergence makes appeals to potential errors extremely implausible since it relies on either a fantastic coincidence (that all these errors are just coincidentally consistent with a single theory), or a fantastic conspiracy (that all international universities have been secretly colluding to hide the truth for over a century). If the LGGLO case is an example of degeneration after the Fall, then why did it just happen to primates and humans as a group in such a way as to independently corroborate that identical grouping constructed from all the other independent sources already mentioned (fossil layering, morphology, etc.) —was it to “test our faith”?
A functional gene that is rendered useless and then persists as a dead gene in the population is rare because most such errors are fatal or horribly debilitating (In the case of LGGLO, the primate diet was rich in vitamin C, so its loss was not a disadvantage). A more common error found is that of gene duplication, which occurs during the DNA replication process. It is as if in copying a book someone copies the same paragraph or page twice. Since one gene is functional, and the copy usually defective and nonfunctional (i.e., dead), mutations can accumulate in the defective copy without affecting the animal carrying it, since they have a second working copy. Consequently, more of this type of dead gene is found than is the other type.
One example (of many) of duplicated dead genes is the gene that codes for an enzyme that is involved in the metabolism of steroid hormones. Right next to this gene in humans is a defective copy of the same gene; that is, it’s a non-functional, “dead” copy of the still functional gene right next to it. Many mutations can render the copy dead; in this case, the particular typo that ruins this gene copy is the deletion of a particular string of just 8 letters (out of a much, much larger number of letters). Now, Chimpanzees have the same dead gene and here is the “smoking gun”: Chimpanzees have the exact same 8 letter deletion.6 This is important: they don’t just have the same dead gene, they have the same typographical error that we do in a very large “book” of letters.
Any appeal to “similar design for similar function” is irrelevant here since these are non-functional errors. Any appeal to “degeneration since the Fall” is also irrelevant, unless one is to assume that God intervenes directly to cause this degeneration by first creating redundant copies of a gene and then targeting particular letters out of many thousands in the genetic code, and only in those animals that are grouped closely together based on comparative anatomy and fossil layering. Keep in mind that this is only one of many examples of specific typos in the same locations of dead genes that are shared between humans and primates.
Of course, God can do anything (which is why saying “Because God did it that way” explains nothing), but if He did, it only serves to create false corroboration of evolution. If we allow for that type of explanation, then we can just as easily accept that the universe was created ten minutes ago with our memories in tact, and that all evidence to the contrary is either a test, or just “because God did it that way for His reasons, which we are too lowly to comprehend”). Again, not only can anything be believed under such an approach, but it also denies us any insights in to the workings of nature since it undermines all of science.
What else did all those dead genes code for?
Those dead and dormant genes code for some strange things, indeed—but unsurprising from an evolutionary framework. For example, embryonic tissue from the jaw of a chicken was induced to grow teeth.7 Please pause and consider the significance of this: genes for teeth are in the chicken’s DNA, but ignored because the chemical signal that activates them no longer occurs. Of course, this experiment proves that the genes for teeth are in there. Why a creator would give chickens the genes for teeth, but keep them turned off, is something difficult to image without the help of a qualified creationist.
To recap up to this point, we’ve seen that a conclusion of descent with modification is corroborated by a tree of life based only comparative anatomy; and again independently corroborated by a tree based only on the position of extinct species in the fossil layers; and again independently corroborated by a tree based only on DNA, which is made up largely of code that doesn’t do anything; and again independently corroborated by the pattern of shared identical typographical errors in the dead genes found in DNA; and again independently corroborated by the kinds of things those dead genes used to code for—things like teeth in chickens, which fits with where chickens are in a tree of life with respect to their toothed, reptilian ancestors.
Earlier I described DNA as a kind of recipe. This analogy is apt because the form of the final organism depends on the precise timing of various other genes during embryological development. Indeed, embryological structures are an important factor in identifying homologous structures (as they were for the earlier taxonomists such as the anti-Darwinian, Richard Owen, mentioned above).
A gene typically doesn’t create an anatomical feature all by itself; instead, the feature arises from the action of the gene working in concert with many other genes, and with each operating under a complex schedule of timing. Modifications of genes that control timing can cause some features to be suppressed, others to be dramatically modified, and still others that are partially developed at one stage of embryological developed to be completely erased at a later stage. Because genes work in this manner, we would expect that the developing embryo will sometimes reveal certain aspects of its evolutionary past.
Examples of this include the whalebone whale and the anteater, both of whom develop teeth in an early embryological stage only to reabsorb them in a later embryological stage before birth.8 (Actually, this example also demonstrates the presence of teeth genes in yet more toothless animals, like the chicken.) Terrestrial salamanders at one stage develop both fins and gills, but then lose them before hatching.
Not only do examples like this show the presence of silent genes for characteristics alien to the definition of the species in which we find them, but these dormant traits are consistent with the placement of these animals in the tree of life as having descended from species that did express these traits: birds descended from toothed reptiles, amphibians descending from an ancestral fish, etc.
Of course, this embryological process may not cause complete disappearance of ancestral traits as it did in the above examples. When they don’t completely disappear, but are still nonfunctional, the structures are described as “vestigial.” Flightless beetles, for example, have wings that remain forever sealed under permanently fused wing covers. Even Darwin commented on many such examples, including that of the rudimentary hind legs found in Boa Constrictors. Seeming to anticipate that some might claim these to have an as yet unknown function, Darwin asked, “why…have they not been retained by other snakes, which do not possess even a vestige of these same bones?”9
This final section is perhaps the most straightforward and certainly one of the most persuasive stand-alone bits of evidence that support Descent with Modification (in other words, even if it weren’t corroborated by all the independent lines of evidence we’ve discussed so far). Based on Descent with Modification, if one species is the descendent of another, then there had to be some geographical continuity from where the parent species is found to where the child species is found—they had to be able to get there.
Of course, if this geographical continuity were broken at some point in the past, then there are predictable consequences—but only if Descent with Modification is true. Without going into the many examples of biodiversity that support Descent with Modification, I will focus only on the Australia example, since it alone is such an overwhelmingly persuasive example—particularly against any notion that all of today’s air-breathing species came from one point on the globe, such as from an “Ark.”
Deeper layers of the fossil record show that marsupial mammals (pouched mammals like the kangaroo) were more common than placental mammals (mammals like us that gestate their young inside their bodies with the use of a placenta). During this time (i.e., in these layers) some parts of the world were populated only by marsupial mammals, including the land mass that would eventually become Australia. Shallower (more recent) layers of the fossil record show that placental mammals had displaced the marsupials over much of the earth.
But what if a barrier appeared before the expanding placentals could invade a particular area that had been occupied only by marsupials? For example, what if a peninsula that had been occupied only by marsupials, became an island before the new placentals migrated there? Well, Descent with Modification would predict that the isolated marsupials might not only survive, but fill many, if not all, of the same ecological niches that placental mammals occupy elsewhere in the world. In other words, they would evolve many analogies to placentals, and only in one place: their isolated island.
Of course, this is exactly what we see in Australia. In the table below10, keep in mind that all of the animals in the Marsupial column are more closely related to each other than they are to their counterparts in the other column. This is an extremely telling observation; it really should make you say, “Wow!”
Consider that the Tiger cat is more closely related to the marsupial mouse than it is to the Bob Cat, which looks superficially almost the same. The same can be said about the Tasmanian Wolf, which looks almost identical to a “regular” wolf, but is also in fact a closer relative to the Marsupial mouse, who for all the world looks like a “regular” mouse.
|Flying Squirrel||Flying Phalanger|
|Bob Cat||Tasmanian Tiger Cat|
Keep in mind that all these marsupial species exist in only one part of the world. Fascinating to be sure, though this is not only explained by Descent with Modification, it is practically expected. Moreover, it adds yet another independent cross-check of the tree you get based only on the comparative anatomy of marsupials and placentals, which, in turn, is independently cross-checked by the tree drawn only from the layer positions of fossils, which is cross-checked by the tree based on biochemistry, etc., etc.
On the other hand, this is not only completely inexplicable under the creationist “model,” but it actually falsifies that “model.” What can the creationist say about such a pattern in biogeography? All they can say is that God created parallel versions of each of these animals (which alone contradicts “similar structures for similar functions”), that they left the Ark at the same time from Mt. Ararat and that somehow the marsupial mouse, Tasmanian wolf, Tiger Cat and the many, many other marsupial species (not shown in the table) that exist only in Australia all cooperated as a group to get to get to Australia ahead of all placental mammals. As Philip Kitcher puts it,
If they all started at the same time in the same place, as the creationists claim, what was it about their lack of a placenta that made them move as a group, predator and prey, large and small ahead of very fast placental predators to just this one part of the globe? Without a direct Divine assist, it’s hard to imagine a coherent explanation.
Let’s think back to the original example of the court case that we discussed at the very beginning. Can analog watches be wrong? Of course. Can certain fossils be misidentified, or identified as coming from the wrong layers? Of course. Can digital watches make mistakes? Definitely. Are animals sometimes misclassified based on their anatomies? Definitely. What about the timing of the 5 O’clock news—is it infallible? Definitely not. What about the reading of DNA sequences—is it infallible? Definitely not.
But just as in the court case, such criticisms miss the whole point. In the court case example, we don’t believe the suspect is guilty just because of what someone’s watch said, or just because someone heard the 5 O’clock news coming on at 5 O’clock, or just because someone heard the 5 O’clock whistle blow. We believe it, because what someone’s watch said was the same thing as what the timing of the 5 O’clock news was telling us, which was the same thing as what the timing of the 5 O’clock whistle was telling us. In other words, we believe because of the agreement between multiple independent sources (not to mention the agreement between multiple samples from the same source—e.g., many watches of different types).
When you find yourself talking about highly technical minutia regarding some particular measurement or method, remember that one particular measurement or one particular method is not why scientists believe that evolution is true. The critic of Evolution has to show not how a measurement or method may be wrong (we all know that), but how all of the thousands of different measurements using many different independent methods come up with the same wrong answer. In the court case, ask yourself what the odds are that all those analog and digital watches were all broken in different ways, but still all said 5 O’clock at the same time; and further that this matched the mis-scheduling of the 5 O’clock news, which in turn, coincided with the mis-scheduling of the 5 O’clock whistle. This conspiracy of errors would have to ensure that as a group they all agreed it was 5:00 PM when it was really, say, 2:02 PM.
It is in fact likely that errors will be made, precisely for the reasons creationists give: these techniques are not perfect. However, if the prosecution is to have a convincing case, errors should appear as a couple of watches that said it was 4:35 or 5:20, with one perhaps saying it was 11:00 AM, but with the overwhelming majority of independent measurements and methods showing tight agreement around 5 O’clock, plus or minus a minute or two. Naturally that would be extremely convincing, and the errors would be recognized as statistical outliers—due precisely to the known fallibility of individual measurements. This is precisely why science doesn’t consider any theory strong on a few data points, but only when there are a many data points and a good deal of independent corroboration. Keep in mind that errors and unexpected results are reported with the rest of the data. This is how science accounts for the fallibility of measurements, and the imperfections of individual scientists.
So it is with evolution. To the creationist one has to ask: How did all the possible errors that could happen in any separate case not only did happen, but conspired together so that as a group they would have tight agreement around the same wrong answer? That is what we mean by independent corroboration; that is what we mean when we say that a theory is well supported by the evidence; and that is what the critic needs to explain. Indeed, his or her alternative theory must not only explain the same phenomena, but must account for that agreement, and not simply point to the obvious fact that mistakes can be made, or that some questions remain, as they do in every field of science.
1 Arthur N. Strahler, Science and Earth History: The Evolution / Creation Controversy (Buffalo: Prometheus Books, 1987), p. 108.
2 Adapted from Joel Cracraft, “Systematics, Comparative Biology, and the Case Against Creationism,” in Scientists Confront Creationism, ed. Laurie R. Godfrey (New York: W. W. Norton & Company, 1983), p. 171.
3 Adapted from Strahler, p. 352
4 Edward E. Max, “Plagiarized Errors and Molecular Genetics,” Creation/Evolution XIX (1986) , p. 34. Reprinted and updated 7/12/99 in TalkOrigins.
6 Kawaguchi, American Journal of Human Genetics 50:766-80 (1992), cited in Max.
7 E.J. Ckollar and C. Fisher, Science 207:993 (1980) cited cited in Douglas J. Futuyma, Science on Trial: The Case for Evolution (Sunderland: Sinaur Associates, 1995), p. 48.
8 Stephen Jay Gould, “Evolution as Fact and Theory,” in Science and Creationism, ed. Ashley Montagu (New York: Oxford University Press, 1984), p. 122.; see also, Futuyma, p. 189.
9 Quoted in Futuyma, p. 49.
10 Adapted from Tim M. Berra, Evolution and the Myth of Creationism (Stanford: Stanford University Press, 1999), Fig. 16.
11 Philip Kitcher, Abusing Science: The Case Against Creationism (Cambridge: The MIT Press, 1993), pp. 141.