The Talk.Origins Archive: Exploring the Creation/Evolution Controversy

The Evolution of Improved Fitness
Correspondence with Critic Lee Spetner
Copyright © 2001 by Edward E. Max, M.D., Ph.D.
[Last Update: January 10, 2001]



1. Introduction

In June of 2000, a reader of my Talk.Origins essay The Evolution of Improved Fitness suggested that I solicit a response from Dr. Lee M. Spetner. Spetner's book "Not by Chance" attacks the theory of evolution using arguments from information theory, in an effort to support a poorly articulated "creation" alternative. I Emailed Spetner, and an extensive (and interesting, in my view) correspondence ensued, which we agreed to link to the original Talk.Origins essay. The unedited correspondence is difficult to follow. There were no length limitations placed on either of us, and many points made by each of us were addressed by the other in comments spread over several successive Emails. To make the correspondence readable I have assembled together the various arguments relevant to a particular issue in the debate. This version contains most of the arguments central to the correspondence, but due to time constraints, some have been omitted. This edited versiion has not been approved by Spetner, and he will likely have further rebuttals. The correspondence will be updated periodically as time permits.

In his first posting of our correspondence, Spetner set the stage, expressing his view that if modern complex species occurred through the "grand sweep of evolution" as envisioned by neo-Darwinian theory (NDT), mutations that added novel information to the "biocosm" must have occurred. According to Spetner, however, none of the mutations observed in experimental models of evolution actually add information as he analyzes it; therefore, Spetner concludes that no matter how many mutations of the observed sort occurred, they could not explain the increase in genomic information postulated by evolution. Here is the text of the beginning of Spetner's posting, and my responses to a few minor points. Spetner's comments are in smaller type, and successive indentations indicate successive responses on individual points. Editorial comments made for this collated assembly of our correspondence are indicated [in brackets].

I am writing this essay in response to a request from Edward E. Max to comment on his posting The Evolution of Improved Fitness (updated July 12 1999). His essay is an attempt to defend evolutionary theory against attacks by creationists. Although Max scored some points against some alleged creationist arguments, he failed to defend Darwinian evolution against my attack on it in my book Not By Chance. He did not mention my book in his posting, but he referred to my book in his request for my comments. I shall also take this opportunity to clarify some issues in my book about which some readers have written me.

The principle message of evolution is that all life descended with modification from a putative single primitive source. I call this the grand sweep of evolution. The mechanism offered for the process of modification is basically the Darwinian one of a long series of steps of random variation, each followed by natural selection. The variation is generally understood today to be random mutations in the DNA.

That primitive source of life is assumed to be sufficiently simple that it could have arisen from nonliving material by chance. There is no theory today that can account for such an event, but I shall not address that issue here. That is for another place and another time. What is relevant to this discussion is that the requirement that life arose spontaneously sets, at the very least, a stringent upper limit on the complexity and information content of the putative first organism that could reproduce itself, and thus serve as a vehicle from which to launch Darwinian evolution. The issue I address here is the alleged development of all life by the neo-Darwinian process of random mutation and natural selection, starting from a sufficiently simple beginning.

Despite the insistence of evolutionists that evolution is a fact, it is really no more than an improbable story. No one has ever shown that macroevolution can work. Most evolutionists assume that macroevolution is just a long sequence of microevolutionary events, but no one has ever shown it to be so. (Those few evolutionists who hold that macroevolution is really different from microevolution have changed their story several times since they first came out with it, and their mechanism is so fuzzy that I cannot tell what it is. John Maynard Smith seems to be of a similar opinion.)

For the grand process of evolution to work, long sequences of "beneficial" mutations must be possible, each building on the previous one and conferring a selective advantage on the organism. The process must be able to lead not only from one species to another, but to the entire advance of life from a simple beginning to the full complexity of life today. There must be a long series of possible mutations, each of which conferring a selective advantage on the organism so that natural selection can make it take over the population. Moreover, there must be not just one, but a great many such series.

The chain must be continuous in that at each stage a change of a single base pair somewhere in the genome can lead to a more adaptive organism in some environmental context. That is, it should be possible to continue to climb an "adaptive" hill, one base change after another, without getting hung up on a local adaptive maximum. No one has ever shown this to be possible.

Now one might say that if evolution were hung up on a local maximum, a large genetic change like a recombination or a transposition could bring it to another higher peak. Large adaptive changes are, however, highly improbable. They are orders of magnitude less probable than getting an adaptive change with a single nucleotide substitution, which is itself improbable. No one has shown this to be possible either.

Moreover, as I have noted in my book, the large mutations such as recombinations and transpositions are mediated by special enzymes and are executed with precision - not the sort of doings one would expect of events that were supposed to be the products of chance. Evolutionists chose the mechanism of randomness, by the way, because we can't think of any other way beneficial mutations might occur in the absence of a law that might govern them. Genetic rearrangements may not be really random at all. They do not seem to qualify as the random mutations neo-Darwinists can invoke whenever needed to escape from a local adaptive maximum.

Evolutionists can argue, and rightly so, that we have no way of observing long series of mutations, since our observation time is limited to a relatively short interval. Our genetic observations over the past 100 years are more like a snapshot of evolution rather than a representative interval in which we can search for the required long series of changes. But our inability to observe such series cannot be used as a justification for the assumption that the series Darwinian theory requires indeed exist.

I agree that there are no definitive examples where a macroevolutionary change (such as the development of cetaceans from terrestrial mammals) has been shown to result from a specific chain of mutations. And I agree with your further comment that "we have no way of observing a long series of mutations." But you go on to say that "our inability to observe such series cannot be used as a justification for the assumption that the series Darwinian theory requires indeed exist." An equally reasonable conclusion, in my view, would be that our inability to observe such series cannot be used as a justification for the assumption that such a series of mutations did NOT occur. In the absence of conclusive data defining such a series, if we want to distinguish between various hypotheses to explain the origin of species we must rely on other data, such as from various laboratory model systems that show adaptations in short enough timeframes that we can observe them. Then we must extrapolate as best we can the information learned from these model systems to the questions of species origins. This extrapolation from laboratory model systems to systems unobservable in the laboratory is the method of science common to medicine, astronomy, chemistry, meteorology, physics, etc.

Now Ed, that's ridiculous! Those two statements are not symmetrical. I don't have to assume the series did not occur to make a case for the inadequacy of NDT. You, who are basing your theory of evolution on the occurrence of such a series, are required to show that it exists, or at least that it is likely to exist. You are obliged to show an existence. I am not obliged to prove a non-existence.

[I think there is some semantic confusion here about the word "justification" in Spetner's sentence "But our inability to observe such series cannot be used as a justification for the assumption that the series Darwinian theory requires indeed exist." He is correct that acceptance of the NDT implies the belief that a series of successive mutations (including duplications and translocations) occurred in the evolution of an ancient primitive genome into the complex genome of a modern species. Because we can access only genomes of modern (or very recent) species, we can never obtain the direct evidence - i.e. a complete list of those mutations - that some anti-evolutionists (e.g. Behe) seem to think would be necessary to support NDT. In the absence of such direct evidence, it seems pointless to argue which side is "obliged" to provide what indirect evidence; certainly neither side can hope for anything close to "proof." Although Spetner denies that he is "obliged to prove a non-existence" of such a chain of mutations, his whole effort in the correspondence seems to be directed to just that aim. Evolutionists have the job of defending the reasonableness of such a series of mutations. I believe that Spetner would agree with this.]

But the argument against Darwinian theory is considerably stronger than that. The theory requires there be a vast number of possible point mutations which, coupled with natural selection, can produce the evolutionary advances that could produce the grand sweep of evolution. Because there must be a large number of qualifying mutations, at least a few of them should have been observed in some of the many genetics laboratories around the world. All the mutations in these long series must not only confer selective advantage on the organism but they must, on the average, also contribute to the information, or complexity, increase that surely distinguishes present-day life from the putative primitive organism.

These mutations must have whatever characteristics are necessary for them to serve as elements of the grand sweep of evolution. Thus, for a mutation to qualify as a representative member of the required multitude of long series that are supposed to produce evolution, it must bring new information not just to the genome of the organism, but the information must be new to the entire biocosm . The horizontal transfer of a gene from one species to another is not information new to the biocosm. To show evolution in action, one must at least demonstrate examples of a mutation that can serve as a prototype of those required by the theory. Such a mutation must be one that could be a contributing member of a series of mutations that could lead to the vast increase in information required by the theory. Thus, for example, a mutation that disables a repressor gene causing a constitutive synthesis of an enzyme might be advantageous to an organism under special circumstances, but the disabling of a gene does not represent the mutations required by the theory.

Max devotes a good portion of his essay to refuting what he calls the "creationist" argument against evolution. Although some opponents of evolutionary theory may have advanced the arguments he attacks, those arguments are in large measure straw men that Max busies himself with refuting. If some creationists have claimed that all mutations are harmful, they would be wrong, but Max's observation that there are mutations that are beneficial, while true, is hardly a telling argument for evolution.


2. The B Cell Hypermutation Model

The next major point of discussion in the correspondence has been about how well the model of immunoglobulin gene somatic hypermutation in B cells serves as an analog to genomic mutation in evolution. The following section contains the salient points of our exchange about this question, beginning with Spetner's initial response to my essay on Talk.Origins.

Max's pièce de résistance was the somatic mutations in B lymphocytes (B cells) of the vertebrate immune system as examples of random mutations that add information. He implied that Evolution could follow this method to achieve baboons from bacteria. I agree with him that these mutations add information to the B-cell genome. I also agree that they are random, but they are random only in the base changes they make; they are not random in where in the genome they can occur. More important, I do not agree that the grand sweep of evolution could be achieved through such mutations.

Although the somatic mutations to which Max referred are point mutations that do indeed add information to the genome of the B cells, they cannot be applied to Darwinian evolution. These are not the kind of mutations that can operate as the random mutations required by NDT that can, through chance errors, build information one base change at a time.

For one thing, the rate of the somatic mutations in the immune system is extremely high - more than a million times normal mutation rates. For this reason they are called hypermutations. If an organism had a mutation rate that was even a small fraction of this rate it could not survive. For a second thing, the hypermutations in the B cells are restricted to a specific tiny portion of the genome, where they can do no harm but only good. The entire genome of the B cell could not mutate at this rate; the hypermutation must be restricted only to the portion that encodes selected portions of the variable part of the antibody.

The mutation rate of the hypermutating part of the B cell's genome is about 10-3 per base pair per replication (Darnell et al., 1986, Molecular Cell Biology, Scientific American Books, p. 1116.), and it can be as high as one in 500 base pairs per replication (Shen, 1998 Science 280: 1750). These rates are incompatible with Darwinian evolution. If an organism's genome were to mutate at this rate, there would be, on the average, about one mutation in every gene, with a high probability that many of them would be fatal for the organism. No, Darwinian evolution could not occur with such rates.

These high rates are essential for the working of the immune system. In each replication of a B cell, about 30 of the 300 or so gene regions encoding the CDR's will have a mutation. A lower mutation rate would make for a less efficient immune system. The high mutation rates, so necessary for the immune system, if applied to an entire organism for evolutionary purposes, would be fatal many times over.

Note that these hypermutations are limited to a restricted portion of the genome. Moreover, the hypermutations are mediated by special enzymes. Thus, although the hypermutations are random in the changes they make in the bases of the genome, they are not random in the positions in which they occur. They occur only in the small region in which they are needed, and occur there through enzymes that apparently play only that role. Furthermore, they occur only when they are switched on by the controlling mechanism of B-cell maturation. Thus it is clear that the hypermutations in B cells cannot serve as a prototype for the random mutations required for NDT.

You agreed with me that the model system of random somatic mutations and selection that occurs in immunoglobulin genes in B lymphocytes can "add information to the B-cell genome." I am glad that you accept the idea that random mutation and selection can lead to an increase in information, since this idea directly refutes the notion of Dembski and others who believe that there is some theoretical bar that would prevent achieving what they call "complex specified information" through random mutation and selection. (Incidentally, I don't think they would appreciate your characterization of them as "straw men.") However, you then go on to declare that the B cell example is a poor model for what happens in "Darwinian" evolution, and you cite two reasons: (1) the mutation rate in this model is much higher than what is seen in non-immunoglobulin genes and in non-B-cells; and (2) these "hypermutations" are mediated by "special enzymes." With regard to your first point, I agree that the mutation rate is higher in the B cell example than in evolution, but I fail to see why that fact weakens the usefulness of the example as a model for evolution. If adaptive mutations that increase information in the genome of a B lymphocyte population can occur over one week given a high mutation rate, what theoretical argument would lead you to reject the idea that adaptive mutations that increase information in the genome of a germ cell population could occur over many millions of years given a much lower mutation rate?

Your second objection to the somatic mutation model in B-cells, that "special enzymes" are involved, is unsupportable. As far as I can tell from my reading of the literature, the mechanism of somatic hypermutation in B-cells is not currently known. The mechanism could perhaps involve "special" enzymes that create mutations, but an alternative possibility is that the high rate of accumulation of mutations simply reflects selective inhibition of normal proof-reading mechanisms. But again, I fail to see why the source of the random mutations should influence the general validity of the conclusion that random mutations and selection can increase genomic information, or why you feel that these mutations cannot serve as a model for evolutionary adaptations.

Indeed, both the rate and predominant mechanism of mutation may be different in different species of organisms, depending on whether they have more or less exposure to cosmic rays and other environmental mutagens, and depending on the nature and robustness of their genomic error-correction mechanisms. Therefore, if we accept your argument against extrapolation from B cell adaptation to species adaptation, should we reject the extrapolation of any information learned from studying one organism to understand adaptations in a second organism, unless it is shown that both the rate and mechanism of mutation are the same in both organisms? In my view this would be like refusing to use the gravitational constant determined in laboratories on earth to analyze stellar physics. Such a reluctance to extrapolate would certainly prevent the use of modern organisms as a basis for understanding evolutionary events that occurred millions of years ago (which may be precisely your intent). I sometimes hear arguments like yours from creationists who are demanding rigorous "proof" of evolution. These creationists do not seem to understand the distinction between mathematics, where a rigorous proof is expected, versus most experimental and observational science, where all we are seeking is the best theory that explains observed data. Of course it is possible to extrapolate unreasonably, but I do not see that you have shown how evolutionary theory (or my essay) does this.

Yes, Ed, the hypermutation in the B cells cannot be a prototype of the kind of mutation required by NDT for Evolution A for the two reasons I gave. You question both those reasons, so I shall elaborate to explain to you why they are valid reasons for rejecting your example of B-cell hypermutation as support for NDT.

One of my arguments to invalidate hypermutation as a model for NDT is that this kind of mutation requires "special enzymes", and is not the kind of mutations held to be responsible for the variation required in NDT. You rejected that argument as unsupportable, but that rejection is unjustified. These mutations, unlike ordinary errors in DNA replication in the germline, are turned on precisely when they are needed and turned off when they have done their job. They are accurately targeted to the very small regions of the genome where they can provide variability to the CDR's, which form the antibody binding site. Although the mechanism of this precisely targeted phenomenon is not yet known in complete detail, enough is known to say that there has to be a "mechanism" - it doesn't just happen by chance.

[At this point Spetner quotes a number of speculative statements in the scientific literature, to the effect that B cell somatic hypermutation involves a "special mechanism." The enzymes involved in somatic hypermutation in B cells remain unknown, so if Spetner is correct that "special enzymes" is supportable, he is correct only in the sense that the idea of "special enzymes" is supported by speculation in the literature. It is unsupported by any evidence, which is what I meant by "unsupportable." To be fair I should note that an enzyme known as Activation Induced Deaminase (AID), reported after my initial comments to Spetner, has been shown necessary for somatic hypermutation to occur, but it is not clear whether this enzyme participates directly in the introduction of mutations. Indeed, since absence of AID also blocks isotype switch recombination, a phenomenon not obviously related to hypermutation, and also leads to enlarged germinal centers, it is possible that this enzyme is required for a step in B cell developmental maturation that triggers both hypermutation and switch recombination, and that the enzyme plays no direct role in mutating DNA. In any case, I never have questioned the idea that somatic hypermutation in B cells involves a "special mechanism"; the question of whether unique enzymes are directly involved in creating the mutations seems to me rather tangential to the present discussion, but it is accurate to say that this question has not been settled as of yet.]

It thus seems quite clear to me that informed opinion in this field supports my contention and rejects your suggestion that "an alternative possibility is that the high rate of accumulation of mutations simply reflects selective inhibition of normal proof-reading mechanisms". Please let me know if you agree or disagree.

[As indicated above, I disagree.]

You ask, why does the existence of a special mechanism for the hypermutation in B cells preclude the example from being a model of mutations for NDT? The simple answer is that if you really want to suggest that mutations for NDT are capable of hypermutations as are the B cells, you have to show two things. First you have to show that such a highly complex system with its requisite enzymes actually exists in germ cells, where they can play a role in evolution. As far as I know, there is no such mechanism in germ cells. If you know of anything like this please let me know.

[I have pointed out that the enzymatic mechanisms creating somatic mutations in B cells is not known. I have never claimed that it is the same mechanism as causes mutations in germ cells where they play a role in phylogenetic evolution, so I am not obligated to show what Spetner says I am.]

Furthermore, according to the evolutionary paradigm, you must account for the origin and development of such a mechanism in the germline, or at the very least, you must suggest how such a development could reasonably occur. You are obligated to do this because you hold that all characteristics of life have evolved through random variation and natural selection.

[Here Spetner, like Behe, seems to demand that I provide an "origin and development" scenario that he knows he will be able to disparage as another "just-so story." Furthermore, the question of how somatic hypermutation evolved is totally irrelevant to the question of whether it is a good model for the efficacy of random mutation and selection in promoting "increased fitness," which is the subject of my Talk.Origins essay.]

You are not entitled to postulate a mechanism that could not have evolved. Such a mechanism of germline mutation would have to produce accurately targeted mutations that could play a role in evolution. For such a system to develop by neo-Darwinian evolution, a long series of evolutionary steps in ordinary evolution would have to play the role of a single step in the evolution of this mechanism, because selection here is based on successful evolution of the ordinary kind. Thus if a million generations are necessary for the evolution and perfection of a new phenotypic character, then a million times that, or a trillion generations, would be required for the evolution and perfection of this mechanism.

[These are totally unsupported quantitative speculations.]

It seems to me that any selective pressure to raise the spontaneous mutation to benefit evolution would be overwhelmed by the selective pressure to keep the mutation rate low. Perhaps I'll look into the mathematics of such a phenomenon and prepare a paper on it.

My second argument invalidating hypermutations as a model of evolution is that the high mutation rate enabling the B cells to play their role in the immune system would be a disaster in ordinary evolution. You also questioned that argument. You asked, "If adaptive mutations that increase information in the genome of a B lymphocyte population can occur over one week given a high mutation rate, what theoretical argument would lead you to reject the idea that adaptive mutations that increase information in the genome of a germ cell population could occur over many millions of years given a much lower mutation rate?"

The theoretical argument is the following. Evolution requires a long series of steps each consisting of an adaptive mutation followed by natural selection. In this series, each mutation must have a higher selective value than the previous. Thus, the evolving population moves across the adaptive landscape always rising toward higher adaptivity. It is generally accepted that the adaptive landscape is not just one big smooth hill with a single maximum, but it is many hills of many different heights. Most likely, the population is on a hill that is not the highest in the landscape. It will then get stuck on a local maximum of adaptivity and will not be able to move from it. This is particularly likely because the steps it takes are very small - only one nucleotide change at a time. The problem is compounded by the lack of freedom of a single nucleotide substitution to cause a change in the encoded amino acid. A single nucleotide substitution does not have the potential to change an amino acid to any one of the other 19. In general, its potential for change is limited to only 5 or 6 others. To evolve off the "dead point" of adaptivity, a larger step, such as the simultaneous change of more than one nucleotide, is required. Moreover, the probability is close to 1 that a single mutation in a population, even though it is adaptive, will disappear without taking over the population (see my book, Chapter 3). Therefore, many adaptive mutations must occur at each step.

The hypermutation in the B cells does this. It achieves all possible single, double, and triple mutations for the immune system, which allows them to obtain the information necessary to match a new antigen. Ordinary mutations, at the normal low rate, cannot add this information - even over long times. I shall explain why. Consider a population of antigen-activated B cells of, say, a billion individuals. In two weeks, there will be about 30 generations. Let's say the population size will remain stable, so in two weeks there will be a total of 30 billion replications. With a mutation rate of 1 per 1000 nucleotides per replication, there will be an average of 30 million changes in any particular nucleotide during a two-week period. The probability of getting two particular nucleotides to change is one per million replications. Thus in two weeks, there will be an average of 30 thousand changes in any two particular nucleotides. There will be an average of 30 changes in any three particular nucleotides.

How many generations, and how long, would it take to get a particular multiple nucleotide change in a germ cell to have an effect on neo-Darwinian evolution? Here, the mutation rate is about one per billion nucleotides per replication. Let's suppose we're doing this experiment with a population of a billion bacteria. Then, in one generation, there will be an average of one change in a particular base. A particular double base change has a probability of one per quintillion, or 10-18. To get one of these would take a billion generations, or about 100,000 years. To get a triple change would take 1014, or a hundred trillion, years. That is why a long waiting time cannot compensate for a low mutation rate. I've given numbers here for a laboratory experiment with bacteria. Many more mutations would be expected world-wide. But the same kind of thing has to happen under NDT with multicelled animals as well. With vertebrates, for example, the breeding populations seldom exceed a few thousand. Multicelled animals would have many fewer mutations than those cited above for bacteria.

Extrapolations made in astrophysics and cosmology may not be entirely valid, but at least they are reasonable based on everything we know. The extrapolation you propose from B-cell hypermutation to neo-Darwinian evolution is unreasonable based on present knowledge, and it is therefore unjustified.

I still fail to see how the particular enzymes involved have any bearing on the applicability of the model to Darwinian species evolution. Indeed, a variety of laboratory mechanisms for generating mutations in antibody genes (chemical mutagens, randomized oligonucleotides, etc.) all lead to pools of mutated antibody genes from which higher affinity proteins can be obtained, so the principle that random mutation and selection can lead to improved function appears to be independent of the mechanism of generating the mutations. There is no logical reason why mutation and selection in species adaptation should be strictly dependent on the mechanism of mutation either; indeed, a variety of different mechanisms are known to contribute to varying extents under different conditions, including copying errors, radiation, chemical mutagens, slipped mispairing, deamination, etc.

You are missing my point. I am not focusing on the source of mutation as the distinguishing factor between the somatic mutations and germline mutations, but I am noting that the hypermutations do have a special mechanism that controls them whereas the germline mutations have no such mechanism available. The important features of the somatic mutations that are unavailable to germline mutations include some (unknown) trigger that turns them on at the right time and directs them to the right place on the genome. Without these controls, hypermutations would destroy the B cells.

[Clearly, the lower rate of mutation that occurs in the germline does not destroy genomes for the next generation. Is this rate high enough to generate enough mutations to account for adaptive phylogenetic evolution? This is a critical issue, and while I don't have a quantitative answer, I don't find Spetner's negative answer to this question supported by convincing logic. He makes reference to his book as offering some arguments, but has not discussed this evidence in our correspondence.]

On the question of the frequency of mutation, in your last posting you included numerical models for B cell hypermutation and for species mutation, and arrived at conclusions by reasoning that I find illogical. You calculate the time required for one, two or three particular nucleotide changes to occur as though these calculations would be relevant to the times required for changes to occur in either B cell or species adaptations. Your reasoning seems to be predicated on the following logic. (1) By single nucleotide mutations, most triplet codons of amino acids can be mutated to code for only "5 or 6" different altered amino acids out of the 20 amino acid constituents of proteins. (2) This limitation would restrict the changes available by single nucleotide changes, such that certain adaptive changes would require two or three particular nucleotide mutations in order avoid getting "stuck on a low local maximum of activity" in the "adaptive landscape."

Your reasoning seems flawed to me in part because you are considering the time to achieve a particular change rather than the time necessary to achieve an improvement in function. The illogic of this is similar to that of equating the odds of being dealt any hand that beats a "bust" hand with the odds of being dealt a particular poker hand that beats a "bust" hand.

I am considering in my analysis what would be analogous to being dealt any hand that beats (i.e. that has a higher selective value than) a given hand.

[This seems to be contradicted by the following sentences Spetner wrote (quoted a few paragraphs back:

With a mutation rate of 1 per 1000 nucleotides per replication, there will be an average of 30 million changes in any particular nucleotide during a two-week period. The probability of getting two particular nucleotides to change is one per million replications. Thus in two weeks, there will be an average of 30 thousand changes in any two particular nucleotides. There will be an average of 30 changes in any three particular nucleotides.

"Particular nucleotides" will be generated by mutation at a far lower frequency than adaptive mutations, as discussed below.]

The B cell system selects for improvements in function and for not particular sequences. Furthermore, there is no reason to assume that the highest theoretical peaks on the adaptive landscape are ever achieved -- either by the B cell system or Darwinian species evolution. Finally, there is no reason to assume that functional improvements cannot arise from the small subset of amino acid replacements accessible from single nucleotide changes.

On the contrary, there is no justification in assuming that one can always obtain a selective advantage with one nucleotide substitution.

[I am not assuming a priori that one can ALWAYS obtain a selective advantage with one nucleotide substitution. But where this has been investigated by experimentally mutating immunoglobulin genes to introduce single mutations corresponding to changes observed in natural somatic hypermutation, it has been found that improvements in antibody affinity can be attributed to specific single nucleotide changes; so it is reasonable to assume that this potential is not so rare as to require the assumption that most increases in antibody affinity require multiple simultaneous mutations. ]

Indeed, if one looks at actual sequences of somatically mutated antibody genes (e.g. Cumano and Rajewsky EMBO J 5:2459, 1986, Figure 2), one finds plenty of single nucleotide mutations that change amino acids, and their presence at a frequency higher than would be predicted by random mutations suggests that most have been selected for on the basis of improved antigen binding. (In this study of nine somatically mutated antibodies, there were 23 amino acid changes caused by single mutations within a codon, and only 4 caused by double mutations; in addition there were only 6 silent single mutations, i.e. not causing amino acid changes.)

This is all consistent with what I have written. I do not say that an adaptive mutation cannot be achieved by a single mutation. I only say that target set of single mutations is smaller than that for single plus double mutations, which in turn is smaller than that for single plus double plus triple, etc. I do not claim that functional improvements cannot be achieved by single mutations, only that the choice is much wider for multiple mutations. For example, three single mutations are not the equivalent of a triple mutation. The first mutation will not remain in the population unless it has a selective advantage. I am sure you will agree that it is possible for a triple nucleotide substitution to have a selective advantage without any single one of them having an advantage. In fact, I would say that is the most likely case.

If single nucleotide changes can lead to selection for improved function, then if one wants to calculate the time necessary to achieve the even greater improvement that might be achieved by three mutations, this time would be found not by a calculation like yours, based on the product of the odds for a single mutation, but rather by multiplying the time for a single adaptive nucleotide mutation by three. Suppose it would take one week for the first adaptive change. By the selection mechanism in the germinal center, the population of B cells would soon be overtaken by B cells with this first change, so that the time required for the second adaptive change would again be one week; and similarly the third change would require a third week to yield a protein of even greater affinity than could be achieved by one or two amino acid replacements.

You are assuming that the single mutation will be selected. I say that is unlikely. The time to achieve a triple change would be equal to the sum of the times necessary to get a single one only if all those single changes had selective value. and that is too unlikely to bank on.

What is the basis for your judgment that such selectable single nucleotide mutations are "too unlikely"?

A similar argument would apply to estimates for adaptive mutations in bacteria. Because of the flawed assumptions built into your approach it seems to me that your calculations grossly overstate the time required for evolving adaptive changes by random mutation and selection.

My assumptions are not flawed. You just don't understand them. Read again what I wrote above and see if you don't agree with me.

[I have reread what Spetner wrote. He agrees that single nucleotide changes can lead to selectable advantages (even if the available codons are restricted so the scope of amino acid changes is less than would be possible with multiple simultaneous mutations). Such selectable single mutations could spread throughout the population, to be followed by successive additional selectable point mutations. By this model, three selectable point mutations could occur in a time frame measured by three times the time for a single mutation, rather than taking the time necessary for three simultaneous mutations as Spetner has calculated. We both agree that simultaneous double and triple mutations have greater scope for amino acid replacement, but are very rare. I honestly do not understand the basis for our disagreement here, but perhaps Spetner will clarify why he considers evolution by successive single selectable point mutations "too unlikely" even though he agrees that single point mutations leading to selectable advantage can occur.]


3. The Role of Gene Duplication

In his first response to my essay, Spetner was critical of the role he thought I claimed for gene duplication in evolution. When he understood that he had originally misread the essay, he had no quarrel with this aspect. Here is the short discussion of this point.

Max cited gene duplication as an example of a mutation that increases information. A favorite scenario for molecular evolution is that a gene gets duplicated and then gradually mutates to become something useful that did not exist before. Such a proposed scenario does not constitute evidence for evolution, it proves nothing, and indeed such a scenario itself requires proof. I do not, of course, mean to say that one has to prove that genes can be duplicated. That is well known. But gene duplication alone does not constitute an increase of information in the biocosm or even in the genome of the organism itself. Two copies of today's newspaper contain no more information than one copy. Gene duplication, in any case, cannot play the role of the mutations that could produce the grand sweep of evolution.

Gene duplication alone cannot add information to the genome. The purpose of the gene duplication in the above scenario is simply to provide raw material from which a new gene could evolve without having to give up any functions the organism already had. New information would then supposedly be built up by point mutations and natural selection. And this is precisely the process I discussed in my book and about which I said that all known examples of these mutations lose information rather than gain it. Note that I did not say that it is impossible in principle for random mutations to add information to the genome. But it just turns out that that is what has been found.

You state: "Max cited gene duplication as an example of a mutation that increases information." On the contrary, I believe that I was careful to avoid saying that gene duplication alone increases information. I do not believe such a statement is correct and agree fully with your statement that "Two copies of today's newspaper contain no more information than one copy.". Please let me know exactly what words in my essay (or in my letter to you) suggested that I believed duplication by itself increases information, and I will try to change the phraseology so as to reduce the likelihood that other readers will misconstrue my meaning.

On the other hand - and this is the major point of all that follows - I do believe that gene duplication is a critical component of what I will call the evolutionary triad: namely gene duplication, random mutation and selection. To illustrate the role of gene duplication in this triad, let's extend your own newspaper analogy. Suppose we have a copy of the early edition of today's newspaper and a copy of the final edition. In the final edition several paragraphs of certain articles have been altered to include late breaking events. Each article has remained the same length in the two editions because certain less important information in each article was deleted to make room for the late breaking news. Now it is clear that having these two copies of today's newspaper does give us more information than either copy alone, since the early edition lacks the late breaking events and the late edition lacks the information that was deleted to make room for the late breaking news.

You seem to allude to this possibility in evolution when you suggest that in the evolutionary model, after gene duplication "[n]ew information would then supposedly be built up by point mutations and natural selection."

You deny suggesting that gene duplication alone adds information. I accept your denial and I apologize for incorrectly attributing that view to you. What led me to believe that you did suggest this is the statement in point 1 of your letter to me, saying. "Gene duplications occur, and there is no reason to postulate supernatural processes to account for them. Does the ID argument about impossibility of naturalistic information increase include an assumption that naturalistic gene duplications cannot occur?" This is what led me to think that you were suggesting gene duplications as a method of adding information.


4. Interpretations of the Word "Evolution"

Spetner tried to clarify different interpretations of "evolution" that frequently cause people confusion if one meaning is intended but another is meant. (For the text of Spetner's comments on this issue, I have taken his True.Origins posting, which begins with this discussion.) I countered that there were several more identifiable meanings of evolution, and that Spetner seemed to be avoiding the burden of having to defend his position by being intentionally vague about where he stood. My response to this point has not been answered.

At the outset, I shall establish an important and necessary guideline in this discussion of evolution. The word evolution is generally used in at least two different senses, and the distinction between them is important. On the one hand, the word evolution is used to denote the descent of all life from a putative single primitive source. It is the grand sweep of evolution that is supposed to have led from a simple beginning, something perhaps simpler than a bacterium, to all organisms living today, including humans. This descent is supposed to have occurred through purely natural means. Neo-Darwinian theory (NDT), which is the prevailing theory of evolution, teaches that this development occurred through random heritable variations in the organisms followed by natural selection. I shall denote the word evolution used in this sense as Evolution A. When evolution is discussed for popular consumption, it is most often Evolution A.

The second sense in which the word evolution is used is to denote any kind of change of a population. The change can sometimes occur in response to environmental pressure (artificial or natural selection), and sometimes it can just be random (genetic drift). I shall denote the word used in this second sense as Evolution B. Evolution B has been observed. Evolution A is an inference, but is not observable. The distinction between these two meanings of evolution parallels the distinction between macroevolution and microevolution, but the two pairs of terms are not identical. Evolution A is certainly what is called macroevolution, but what is called macroevolution is not identical with Evolution A. In any case, I prefer to use the A and B to avoid having to carry whatever baggage might go with the macro/micro distinction.

The distinction between these two meanings of evolution is often ignored by the defenders of neo-Darwinian evolution. But the distinction is critical. The claim is made for Evolution A, but the proof offered is often limited to Evolution B. The implication is that the observation of Evolution B is a substantiation of Evolution A. But this is not so. Since Evolution A is not an observable, it can only be substantiated by circumstantial evidence. This circumstantial evidence is principally the fossil record, amino-acid-sequence comparisons, and comparative anatomy. Circumstantial evidence must be accompanied by a theory of how it relates to what is to be proved. NDT is generally accepted to be that theory. The strength of the circumstantial evidence for Evolution A can therefore be no better than the strength of NDT.

I can't tell exactly what you accept in your distinction between Evolution A and Evolution B. I actually think that there are finer distinctions between the various meanings of evolution than encompassed by your A vs B. I would distinguish several more possible meanings:

#1. Living forms are different now from what they were in the past. This seems to be well documented by fossil evidence. This slow change is sometimes referred to as evolution.

#2. Random mutation and selection can lead to "microevolution," i.e., small changes in gene frequencies that follow an environmental shift and leave a population on average more fit to cope with the new environment. I think you accept this, since I think it corresponds to what you mean by Evolution B. I certainly accept it.

#3. Various different modern species share a common ancestry. Since the time of the common ancestor, the divergence into the various modern species has involved changes much greater than microevolution. This is the idea of "common descent." I am really not sure whether you accept this notion. I think there is excellent evidence for common descent of some groups of species, as outlined in my essay. If you do not accept common descent, at least for the cases I cite in my essay, I would be interested in hearing what alternative interpretations you can offer for the observations I cite in that essay.

#4. All of the nucleotide discrepancies between modern species, or between a modern species and its ancestral species, arose as a result of random mutation (including gene duplications, insertions and deletions caused by naturalistic processes) and natural selection, without the intervention of an "intelligent designer." I do not believe that there is any evidence for the preceding statement, and indicate as much in my essay. Nor do I believe that an "intelligent designer" can be ruled out as an explanation for hurricanes, disease, or stock market fluctuations. However, I have never seen a convincing argument that an intelligent designer must be hypothesized in order to explain any of these kinds of events, or to explain species change through time.

#5. The origin of life came about through exclusively naturalistic processes operating on prebiotic chemicals, which evolved into replicating life forms. We have almost no scientific evidence about the origin of life and so there is no scientific evidence to support a purely naturalistic origin of life. I feel the same way about this meaning of "evolution" as I do about #4.

In my judgment, there is good scientific evidence for #1, #2 and #3. From your dismissal of evidence for what you call Evolution A, I can't tell what you believe about #3. On #4 and #5 I assume we are in agreement on the insufficiency of scientific evidence to support a purely naturalistic mechanism, but we obviously differ on whether arguments such as yours are sufficient to rule out a purely naturalistic mechanism. I think that it would be an improvement in the dialogue/ document to clarify both of our opinions on these finer distinctions. Incidentally, I am not clear exactly on the difference you see between EvolutionA and macroevolution.

I don't know what version of creation you accept, but it seems to me that even if the supernatural played a role in past events, those past events leave traces. By refusing to specify an alternative scenario that you consider more believable than evolution, you hide behind vagueness in order to avoid having to defend potential contradictions between your scenario and the traces from the past that point in a different direction.


5. Information Content of Proteins

The central theme of Spetner's position, and the focus of his book, is that information theory can shed light on the likelihood of the evolutionary scenario envisioned by the NDT. In particular, he believes that observed mutations do not provide increases in information that would be required by the NDT to produce what he calls Evolution A. Spetner included several graphic Figures in his discussion of ribitol dehydrogenase (section 6.1 below) which I have not been able to reproduce in the text below. I feel that the essence of his arguments is comprehensible even without the Figures, but I will attempt to insert them in the future.

Mutations have indeed been observed that confer an adaptive advantage, but that alone does not qualify them to serve as components of a series of neo-Darwinian steps.

In my critique, I included for pedagogical purposes the following short explanation of information and its measurement:

I shall emphasize again: There is no theorem requiring mutations to lose information. I can easily imagine mutations that gain information. The simplest example is what is known as a back mutation. A back mutation undoes the effect of a previous mutation. If the change of a single base pair in the genome were to change to another and lose information, then a subsequent mutation back to the previous condition would regain the lost information. Since these mutations are known to occur, they form a counterexample to any conjecture that random mutations must lose information. An important point I make in my book, and which I emphasize here, is that no mutations observed so far qualify as examples of the kind of mutations required for Evolution A.

In discussing mutations in my book I noted in each case in which the molecular change was known, that it could not serve as a prototype for the mutations required by NDT. In all the cases I discussed, it was the loss of information that prevented the mutation from serving as a prototype of those required by NDT. The back mutation likewise cannot serve as a prototype of the NDT-required mutations. Here, the reason is not that it loses information - it actually gains information. But the information it gains is already in the biocosm and the mutation contributes nothing new. Evolution cannot be accounted for if the only information gain was by back mutations.

In my book, I did not quantify the information gain or loss in a mutation. I didn't do it mainly because I was reluctant to introduce equations and scare off the average reader. And anyway, I thought it rather obvious that a mutation that destroys the functionality of a gene (such as a repressor gene) is a loss of information. I also thought it rather obvious that a mutation that reduces the specificity of an enzyme is also a loss of information. But I shall take this opportunity to quantify the information difference before and after mutation in an important special case, which I described in my book.

The information content of the genome is difficult to evaluate with any precision. Fortunately, for my purposes, I need only consider the change in the information in an enzyme caused by a mutation. The information content of an enzyme is the sum of many parts, among which are:


  • Level of catalytic activity
  • Specificity with respect to the substrate
  • Strength of binding to cell structure
  • Specificity of binding to cell structure
  • Specificity of the amino-acid sequence devoted to specifying the enzyme for degradation

These are all difficult to evaluate, but the easiest to get a handle on is the information in the substrate specificity.

To estimate the information in an enzyme I shall assume that the information content of the enzyme itself is at least the maximum information gained in transforming the substrate distribution into the product distribution. (I think this assumption is reasonable, but to be rigorous it should really be proved.)

We can think of the substrate specificity of the enzyme as a kind of filter. The entropy of the substances separated after filtration is less than the entropy of the original mixture. We can therefore say that the filtration process results in an information gain equal to the decrease in entropy. Let's imagine a uniform distribution of substrates presented to many copies of an enzyme. I choose a uniform distribution of substrates because that will permit the enzyme to express its maximum information gain. The substrates considered here are restricted to a set of similar molecules on which the enzyme has the same metabolic effect. This restriction not only simplifies our exercise but it applies to the case I discussed in my book.

The products of a substrate on which the enzyme has a higher activity will be more numerous than those of a substrate on which the enzyme has a lower activity. Because of the filtering, the distribution of concentrations of products will have a lower entropy than that of substrates. Note that we are neglecting whatever entropy change stems from the chemical changes of the substrates into products, and we are focusing on the entropy change reflected in the distributions of the products of the substrates acted upon by the enzyme.

The entropy of a distribution of n elements with fractional concentrations f1,...,fn is given by

H = Sum(i=1..N) fi log fi

and if the base of the logarithm is 2, the units of entropy are bits.

As a first illustration of this formula let us take the extreme case where there are n possible substrates, and the enzyme has a nonzero activity on only one of them. This is perfect filtering. The input entropy for a uniform distribution of n elements is, from (1), given by

HI = log n since the fi's are each 1/n. The entropy of the output is zero,

H0 = 0,

because all the concentrations except one are zero, and the concentration of that one is 1. Then the decrease in entropy brought about by the selectivity of the enzyme is then the difference between (2) and (3), or

H = HI - H0 = log n

Another example is the other extreme case in which the enzyme does not discriminate at all among the n substrates. In this case the input and output entropies are the same, namely

HI = H0 = log n

Therefore, the information gain, which is the difference between HO and HI, in this case is zero,

H = 0

We normalize the activities of the enzyme on the various substrates and these normalized activities will then be the fractional concentrations of the products. This normalization will eliminate from our consideration the effect of the absolute activity level on the information content, leaving us with only the effect of the selectivity.

Although these simplifications prevent us from calculating the total entropy decrease achieved by action of the enzyme, we are able to calculate the entropy change due to enzyme specificity alone.

5.1 The dangers of conclusion jumping

As a final example let me take part of a series of experiments I discussed in my book, which demonstrate the dangers of conclusion jumping. This subject bears emphasis because evolutionists from Darwin on have been guilty of conclusion jumping. I shall here take only a portion of the discussion in my book, namely, what I took from a paper by Burleigh et al. (Biochem. J., 1974, 143: 341), to illustrate my point.

Ribitol is a naturally occurring sugar that some soil bacteria can normally metabolize, and ribitol dehydrogenase is the enzyme that catalyzes the first step in its metabolism. Xylitol is a sugar very similar in structure to ribitol, but does not occur in nature. Bacteria cannot normally live on xylitol, but when a large population of them were cultured on only xylitol, mutants appeared that were able to metabolize it. The wild-type enzyme was found to have a small activity on xylitol.

Fig. 1 shows the activity of the wild-type enzyme and the mutant enzyme on both ribitol and xylitol. Note that the mutant enzyme has a lower activity on ribitol and a higher activity on xylitol than does the wild-type enzyme. An evolutionist would be tempted to see here the beginning of a trend. He might be inclined to jump to the conclusion that with a series of many mutations of this kind, one after another, evolution could produce an enzyme that would have a high activity on xylitol and a low, or zero, activity on ribitol. Now wouldn't that be a useful thing for a bacterium that had only xylitol available and no ribitol? Such a series would produce the kind of evolutionary change NDT calls for. It would be an example of the kind of series that would support NDT. The series would have to consist of mutations that would, step by step, lower the activity of the enzyme on the first substrate while increasing it on the second.

But Fig. 1 is misleading in this regard because it provides only a restricted view of the story. Burleigh and his colleagues also measured the activities of the two enzymes on another similar sugar, L-arabitol, and the results of these measurements are shown in Fig. 2. With the additional data on L-arabitol, a different picture emerges. No longer do we see the mutation just swinging the activity away from ribitol and toward xylitol. We see instead a general lowering of the selectivity of the enzyme over the set of substrates. The activity profile of the wild-type enzyme is more selective on ribitol than is the mutant enzyme.

In Fig. 1 alone, there appears to be a trend evolving an enzyme with a high activity on xylitol and a low activity on ribitol. But Fig. 2 shows that such an extrapolation is unwarranted. It shows instead a much different trend. An extrapolation of the trend that appears in Fig. 2 would indicate that a series of such mutations would result in an enzyme that had no selectivity at all, but exhibited the same low activity on a wide set of substrates. Both these extreme extrapolations are, of course, unwarranted from the limited amount of data available. But, Fig. 2 does show that the mutation decreased the information in the enzyme.

The point to be made from this example is that conclusion jumping from the observation of an apparent trend is a risky business. From a little data, the mutation appears to add information to the enzyme. From a little more data the mutation appears to be degrading the enzyme's specificity and losing information.

Just as we calculated information above, we can calculate the information in the enzyme acting on a uniform mixture of the three substrates for both the wild type and the mutant enzyme. These values turn out to be 0.74 and 0.38 bits respectively. The information in the wild-type enzyme then turns out to be about twice that of the mutant. The evolutionist community, from Darwin to today, has based its major claims on unwarranted conclusion jumping. Darwin saw that pigeon breeders could achieve a wide variety of forms in their pigeons by selection, and he assumed that the reach of selection was unlimited. Evolutionists, who have seen crops and farm animals bred to have many commercially desirable features, have jumped to the conclusion that natural selection, in the course of millions of years, could achieve many fold greater adaptive changes than artificial selection has achieved in only tens of years. I have shown in my book that such extrapolations are ill founded because breeding experiments, such as those giving wheat greater protein content or vegetables greater size, result from mutations that disable repressor genes. The conclusions jumped to were false because they were based on data that could not be extrapolated to long sequences. One cannot gain information from a long sequence of steps that all lose information. As I noted in my book, that would be like the merchant who lost a little money on each sale, but thought he could make it up on volume.

5.2 Antibiotic Resistance as an Example of Evolution

Continuing his effort to show the evolutionary efficacy of beneficial mutations, Max presented in his essay the acquisition of antibiotic resistance by microorganisms as an example of evolution. He said one can "demonstrate a beneficial mutation with laboratory organisms that multiply rapidly, and indeed such experiments have shown that rare beneficial mutations can occur. For instance, from a single bacterium one can grow a population in the presence of an antibiotic, and demonstrate that organisms surviving this culture have mutations in genes that confer antibiotic resistance." Such an experiment shows that "de novo beneficial mutations" can arise.

My response to this is that I have shown in my book that mutations leading to antibiotic resistance fail the test of representing the mutations necessary for evolution. I summarize that argument here.

All antibiotics are derived from microorganisms. Recall the story of the serendipitous discovery of penicillin by Alexander Fleming in 1928, when he noticed that his plate of Staphylococcus bacteria was clear in the vicinity of a bread-mold contaminant. The mold was found to produce something that could lyse and kill the bacteria. That something was a molecule later named penicillin. Afterwards, other antibiotics were found to be produced by other microorganisms, such as soil bacteria. Soil has long been recognized in folk medicine as a cure for infections.

The antibiotics produced by these microorganisms serve them as a defense against attack by other microorganisms. Some microorganisms are endowed with genes that grant resistance to these antibiotics. This resistance can take the form of degrading the antibiotic molecule or of ejecting it from the cell. Unfortunately for human health care, the organisms having these genes can transfer them to other bacteria making them resistant as well. Although the resistance mechanisms are specific to a particular antibiotic, most pathogenic bacteria have, to our misfortune, succeeded in accumulating several sets of genes granting them resistance to a variety of antibiotics.

The acquisition of antibiotic resistance in this manner qualifies as evolution only in the sense that it is an adaptive hereditary change. It is an example only of Evolution B. It is not the type of evolution that can make a baboon out of a bacterium. The genetic change is not the kind that can serve as a prototype for the mutations needed to account for Evolution A. The genetic changes that could illustrate the theory must not only add information to the bacterium's genome, they must add new information to the biocosm. The horizontal transfer of genes only spreads around genes that are already in some species.

It turns out, however, that a microorganism can sometimes acquire resistance to an antibiotic through a random substitution of a single nucleotide, and this is the kind of example Max presented. Streptomycin, which was discovered by Selman Waksman and Albert Schatz and first reported in 1944, is an antibiotic against which bacteria can acquire resistance in this way. But although the mutation they undergo in the process is beneficial to the microorganism in the presence of streptomycin, it cannot serve as a prototype for the kind of mutations needed by NDT. The type of mutation that grants resistance to streptomycin is manifest in the ribosome and degrades its molecular match with the antibiotic molecule. This change in the surface of the microorganism's ribosome prevents the streptomycin molecule from attaching and carrying out its antibiotic function. It turns out that this degradation is a loss of specificity and therefore a loss of information. The main point is that Evolution A cannot be achieved by mutations of this sort, no matter how many of them there are. Evolution cannot be built by accumulating mutations that only degrade specificity.

In the final paragraph of my original critique, I said the following:

The mutations needed for macroevolution have never been observed. No random mutations that could represent the mutations required by NDT that have been examined on the molecular level have added any information. The question I address is: Are the mutations that have been observed the kind the theory needs for support? The answer turns out to be NO! Many have lost information. To support NDT one would have to show many examples of random mutations that add information. Unless the aggregate results of the genetic experiments performed until now is a grossly biased sample, we can safely dismiss neo-Darwinian theory as an explanation of how life developed from a single simple source.

You cite the fact that some bacteria grown under selective pressure of this antibiotic become resistant through a mutation that "degrades the molecular match with the antibiotic molecule" representing "a loss of specificity and therefore a loss of information." Some streptomycin resistance mutations do, as you point out, reflect mutations of the ribosomal protein S12 which cause loss of binding of this antibiotic, which you interpret as "loss of information." However, you ignore other mutations of this protein that do not lead to loss of antibiotic binding (e.g. Timms et al., Mol Gen Genet 232:89, 1992). According to your formulation, these mutations would not represent a loss of information, yet they are represent natural mutations that are adaptive under conditions of exposure to streptomycin. Would you accept that this kind of mutation is a good model for an adaptive evolutionary change consistent with neo-Darwinian Theory?

You misunderstood the paper by Timms et al., which you cited. All of the adaptive mutations reported in that paper show reduced binding of the streptomycin molecule. The 12 adaptive mutations reported in the S12 protein fall into two categories. There was no example of what you claimed I ignored. Five of those mutants are designated as streptomycin resistant (Smr), and seven are designated as streptomycin dependent (Smd). All 12 of them, in the words of the authors "reduce the affinity of the ribosome for streptomycin." Perhaps you would like to point out to me where in that paper they mention mutations in S12 do not lead to reduced binding, and which you claim I have ignored.

My citation of this paper was based on its description of the streptomycin-dependent mutants, which require streptomycin for growth as a result of mutations in the S12 protein. Clearly such mutants have not lost streptomycin binding completely; however it is possible that they have reduced binding affinity, so that according to your criteria -- which I do not accept as valid -- they might have "lost information." However, your whole argument about streptomycin seems to be based on the misconception that streptomycin works by binding to the S12 protein. In fact, as mentioned in the Timms paper, the binding is primarily to the 16S ribosomal RNA, not to S12, and the mutations in the S12 protein function to decrease streptomycin by stabilizing a specific conformation of the 16S rRNA that does not bind streptomycin well (Carter et al., Nature 407: 340, 2000; Moazed & Noller, Nature. 327:389, 1987; Gravel et al., Biochemistry. 26:6227, 1987; Montandon et al, EMBO J. 5:3705, 1986; Pinard et al, FASEB J. 7:173, 1993; Melancon et al., Nucleic Acids Res. 16:9631, 1988). A mutation that causes a specific conformational change in another molecule that in turn prevents efficient binding of a third molecule does not necessarily suggest a "loss of information" to me, even if your protein information metric were valid.

There are several other ways of considering how mutations affect information. In my view, even if all S12 mutations that caused streptomycin resistance abolished antibiotic binding, a reasonable argument could still be made that such mutations represent a gain of information rather than a loss. In the universe of all the possible S12 amino acid sequences that can function in the ribosome, essentially all S12 proteins found in "wild-type" bacteria (i.e., those grown in the absence of streptomycin) bind to this antibiotic. The S12 sequences that allow bacterial growth in the presence of streptomycin respresent a small subset of the universe of functional S12 sequences. Therefore by growing bacteria in streptomycin we select for a specific and small subset of possible S12 sequences; thus it might be argued that we have forced a small increase the information content of the genome by narrowing the choice of S12 sequences.

The set of S12 proteins that allow bacterial growth in streptomycin (i.e. that do not bind to the antibiotic) form a disparate subset of the universe of S12 proteins. My intuition tells me that the set that binds (the susceptible set) is smaller, and therefore has a smaller entropy, than the set that does not bind (the resistant set). Mutations that appear in the presence of the antibiotic convert one subset to the other. A mutation that transfers the enzyme from a low-entropy set to a higher-entropy set loses information; it does not gain it.

There are many sequences of S12 proteins in a variety of "wild type" bacteria. Different species of Gram negative bacteria are commonly sensitive to streptomycin despite variations in S12 sequence; organisms with S12 mutations are very rarely found except under streptomycin selection. Therefore, MY intuition tells me that most S12 sequences bind streptomycin and that the set of S12 sequences conferring streptomycin resistance is smaller than the set conferring sensitivity. What supports your "intuition" that the susceptible set is smaller and therefore has smaller entropy?

However, I want to make it clear that I don't buy your interpretation of certain specific mutations as reflecting a "loss of information." You state that the "information content of an enzyme is the sum of many parts, among which are: level of catalytic activity, specificity with respect to the substrate, strength [and specificity] of binding to cell structure, [and] specificity of the amino-acid sequence devoted to specifying the enzyme for degradation." This formulation is vague, non-quantitative, not supported by clear logic, not accepted in the scientific literature (to the best of my knowledge; please educate me if I am wrong), and in my view not useful.

Ed, the level of your argument here is quite low. You have seen this entire section (above), and you took from the introduction my list of what characteristics can contribute to the information content of an enzyme and criticized it for being non-quantitative (followed by other pejorative epithets). Is that supposed to be some sort of debating tactic? In any case, the tactic is out of place in this discussion. From the context of what I wrote, it should have been clear to you that this partial list of characteristics that can contribute to the information in an enzyme was an introduction to my quantitative estimate of one of the characteristics of specificity of an enzyme. After I showed how one might calculate the information related to a type of specificity, I showed how a mutation that appeared to enhance activity on a new substrate actually reduced the information by about 50%.

It is elementary that specificity translates into information and vice versa. Have you ever played 20 questions? With the YES/NO answers to 20 judicious questions, one can discover a previously-chosen number between 1 and a million. If the questions are well chosen, those YES/NO answers can be worth one bit of information each, and 20 bits can specify one object out of a million. Twenty bit of information translates to specificity of one part in a million. Ten bits - to one part in a thousand.

The Zip codes in the US also demonstrate that specificity and information are two sides of the same coin and go hand in hand. An address in the United States can be completely specified by the nine-digit zip code. One digit of information will narrow down the address from being anywhere in the United States to being in just a few states. Thus if the first digit is a 6, the address is located somewhere in Illinois, Missouri, Kansas, or Nebraska.

A second digit of information will add specificity by narrowing down the address further. A 3, 4, or 5 in the second digit puts the address in Missouri. A 3 in the second digit puts it in the eastern portion of the state. Two digits of information are more specific than one.

A third digit of information is still more specific, narrowing down the address even more, making it still more specific. If the third digit is a 1, the address is specific to St. Louis and its suburbs. The next two digits of information pin down the address to within a few blocks. The remaining 4 digits of information can locate a specific building. Thus, it is clear that the information contained in the digits of the zip code translate into specificity.

There is no question about it: SPECIFICITY = INFORMATION.

Not only have I made it clear above that my criterion for gain/loss of information is quantitative, and supported by logic and the conventional understanding of these notions in information theory, I included that section in my first critique of your posting. You chose not to relate to it at all, and instead you made up the above criticism out of thin air.

In my previous comments about your calculation of the "information gain or loss in a mutation" I made some criticisms which you called "pejorative epithets" and which you suggested were "some sort of debating tactic" or "made out of thin air"; but you did not address any of the criticisms substantively, so I will repeat them with more detail in hopes that you will address them.

1. I suggested that your formulation is vague and non-quantitative and not supported by clear logic.

You have stated:

The information content of an enzyme is the sum of many parts, among which are:
  • Level of catalytic activity
  • Specificity with respect to the substrate
  • Strength of binding to cell structure
  • Specificity of binding to cell structure
  • Specificity of the amino-acid sequence devoted to specifying the enzyme for degradation

First of all, I note that of these five components, you have suggested for only one -- specificity with respect to the substrate -- how you would quantitate its contribution to the information content of the protein. In discussing this component you state:

To estimate the information in an enzyme I shall assume that the information content of the enzyme itself is at least the maximum information gained in transforming the substrate distribution into the product distribution. (I think this assumption is reasonable, but to be rigorous it should really be proved.)

You may think that this assumption is reasonable, but I think that it is totally unreasonable. This assusmption forms the basis for almost your entire argument, yet even you admit that it has not been "proven," which would be necessary for your analysis to be rigorous, as you state. You therefore agree with me that your analysis is not rigorous, but based an unproven assumption.

Secondly, you omitted any description of how the other components you listed would be used to assess information. Yet, you have claimed that because the mutations in the ribitol dehydrogenase system suggest a decrease in the substrate specificity component of information, the mutation represents a loss of information. But how can you claim this when you have not evaluated quantitatively all the other components that you say contribute to information? To me, for you to make a judgment about the quantitative information change due to the mutation when you have left out an evaluation of four of the five components of your proposed information metric is a rather serious lapse, especially for one who accuses others of "conclusion jumping."]

Thirdly, you have not specified whether all the five components in your list should be given equal weight. If you do not give them equal weight, please explain your weighting system and justify it.

Fourthly, you imply ("sum of many parts, among which are") that there are additional "parts" that might contribute to the information content; but you never specify what these are.

[Fifthly, you have not justified why any of these parameters should be considered in a metric quantitating the information of a protein. One might argue that the information content of the wild type and mutated ribitol dehydrogenase proteins were the same because - regardless of the substrate specificities -- the amount of information necessary to define their amino acid sequence has not changed.

Your analogies (the 20 questions game, or zip codes) that encourage you to proclaim that "Specificity = Information" don't clarify anything about the "information of a protein" in that a 200 amino acid protein A that has high levels of all of the components of your information metric can be specified by exactly as much information as a 200 amino acid protein B that is low in all your components. Indeed, ] I believe most scientists who have considered the information represented by genes or enzymes would conclude that a large complex protein involves much more information than a short polypeptide. Certainly it requires more information to specify the sequence of a large protein. Yet in your list of five components of information you have completely omitted that one parameter that most scientists would consider most important in comparing information content.

In summary, you have evaluated only one of your five components of protein information quantitatively, and that analysis you admit is not rigorous; you have not yet defined how four of your five parameters would be quantified; you have not yet described how the parameters would be weighted in combining them into a metric of information; you have not presented a justification of why each parameter should be included in the metric; you have not specified what other parameters need to be included; and you have not justified the exclusion of sequence length, the parameter most scientists would include in an information estimate. These are the reasons I considered your formulation vague and non-quantitative and not supported by clear logic.

2. Your formulation is not accepted in the scientific literature.

This is obvious to you and to me, but I wanted to make it clear to other potential readers of this correspondence. Although components of your analysis may include elements of accepted information theory analysis, your inclusion of the 5 items above as the elements contributing to a quantifiable information metric is original with you and has never (to my knowledge) been published in the peer-reviewed scientific literature; please correct me if I am wrong. Some readers might conclude from your pejorative (and unnecessarily personal and condescending) comments (e.g. "I would recommend that you not refer to my criteria of information loss as "questionable" until you understand them" ) that I am a loose cannon who blasts accepted theories without clearly understanding them. These readers should be aware that your theories have not met the normal criterion for a scientific idea to be worthy of acceptance or even serious consideration, namely publication in the peer-reviewed professional scientific literature. Although a computer search of the literature showed me that you wrote exactly two papers on information and protein sequences that were published in peer-reviewed journals, both more than 30 years ago (J THEOR BIOL 7 : 412, 1964; and NATURE 226: 48, 1970), neither of these papers contains discussion of your estimate of information content of a protein as measured by the parameters listed above. As far as I have been able to determine (and please correct me if I am wrong) the latter ideas were published only in your book, a non-peer reviewed publication; and the ideas from the book have been mentioned in the peer-reviewed professional literature only once, in a recent paper (Schneider Nucl Ac Res 28:2794, 2000) that disputes the validity of your analysis. The fact that your information metric has not been published in the peer-reviewed professional literature does not in itself make the analysis wrong, any more than the absence of flat-earth papers in the professional planetary astronomy journals or the absence of Holocaust denial papers in the professional history literature makes those two theories wrong. Each theory stands or falls on its merits (or lack thereof). But readers should know that you have not undertaken a novel application of a generally accepted metric to draw novel conclusions that confound evolutionists; rather, you have applied an eccentric metric never accepted by the science community, and not surprisingly have drawn eccentric (and in my view invalid) conclusions.


6. Gene Families as Examples of Duplication, Mutation and Selection

A commonly cited observation consistent with the neo-Darwinian model of evolution is that in the DNA of humans there are many genes with similar sequences that have similar function, yet play distinct physiological roles. The multiple globin genes are an example I cite in my essay. Frequently genes with similar sequences are found in more primitive organisms, but in these species the number of related genes is much smaller. The evolutionary interpretation is that the last common ancestor of humans and the primitive modern species had a smaller genome than modern humans and that as the human lineage evolved, there were multiple duplications which generated extra gene copies; these mutated independently and evolved to take on slightly different functions. Spetner, of course, does not accept this scenario. I begin this part of the exchange with a description of such a gene system.

Let's consider a gene locus that I have studied in my lab: the human immunoglobulin heavy chain (or IgH) locus. In the human locus one sees evidence of a large DNA duplication that created two copies that are highly similar in both coding and non-coding flanking regions. One duplicate includes constant region sequences known as gamma3, gamma1, pseudo-epsilon and alpha1, while the second copy contains gamma2, gamma4, epsilon and alpha2. More primitive primates like the New World monkeys appear to have a single copy of this locus and a single gamma gene. The four human gamma chain genes are thus thought to have derived from a single ancestral gamma chain gene in a primate ancestor by a series of duplications and mutations. The four kinds of antibody proteins encoded by the human genes serve very similar functions, but they are not identical. They differ from one another in their "effector functions" such as their ability to activate serum complement proteins or to bind the various Fc receptors on cells of the immune system. For example, antibodies with gamma2 protein work best for recognizing polysaccharide antigens found on certain bacteria, while gamma4 antibodies work best for fighting parasites. Presumably the single ancestral gamma gene was not specialized and had to serve as a "jack-of-all-trades." If you were to consider the mutations of that gene that led to the specialized function of the polysaccharide-binding gamma2 protein you could probably argue for "loss of information" in that, by mutating from primordial gamma, the protein may have "lost specificity" for battling parasite infestation; and if you looked at the mutations that led to the "parasite specialist" gamma4 protein, you could argue for "loss of information" in that the protein may have "lost specificity" for binding to polysaccharides. If you put on blinders and looked at one gene at a time you could make your argument that both genes "lost information," but if you look at the whole picture you see that there is a gain in information for the whole system. In the ancestral primate we had one non-specialized gene whereas in modern humans we have four specialized genes. This is exactly the sort of genetic change that would be consistent with neo-Darwinian evolution leading to an increase in complexity. In your newspaper example it corresponds to having both the early and the final edition of today's paper. A merchant who makes a little money on each transaction can certainly make a bundle if he works long enough at it.

Yes, information would have been increased if what you speculate had indeed happened. The proof would only lie in showing that it has indeed happened. Let us not lose sight of the requirement of neo-Darwinian evolution for long series of single-nucleotide substitutions, where each mutation makes the phenotype sufficiently more adaptive than it was to permit the mutated phenotype to take over the population through natural selection with a high probability. It is far from clear that the individual mutations you suggest will each be adaptive and selected at each step. You cannot show this - you merely assume it. You are postulating an historical event that cannot possibly be verified. It seems that all of your arguments are based on postulating events that are inherently not observable. That should make one a little suspicious of the theory, shouldn't it?

I realize that the above model for the human IgH locus is hypothetical and assumes that the evolutionary triad of duplication, random mutation and selection is a reasonable naturalistic explanation for the four human gamma genes. We cannot verify this explanation since we can never know the properties of the primordial ancestral gamma immunoglobulin, or know the series of mutations that occurred in the various duplicate gamma genes during our evolution from that primordial ancestor. What I am asking is: is there anything so implausible in this model to justify your suggestion that we should "dismiss neo-Darwinian theory" as an explanation for this example?

Yes, it is implausible because you are postulating a series of events of a type for which there is evidence that they have not occurred. If they had occurred to produce Evolution A, there should have been a vast number of them, and there aren't. Had there been the required large number of them, we should have seen some of them in all the genetic experiments performed in all the laboratories of the world. And we haven't, to my knowledge, seen a single one.

Or more to the point, exactly what alternative explanation for the origin of the four human gamma genes do you propose that is more plausible than the one I offered? This is important, because considering the weaknesses I have pointed out in your arguments, you are far from having definitively ruled out the neo-Darwinian evolutionary triad as the correct explanation for what you call the "grand sweep of evolution." A mathematical proof that a conjecture is false, or a proof that a proposed invention is impossible because it violates the Second Law of Thermodynamics, may checkmate further discussion; but the points that you make (at least the ones that are sound) simply highlight gaps in our knowledge and force one to evaluate the validity of extrapolations we make from observable data. These points are useful contributions, but they would seriously damage the credibility of evolution only if there were an alternative explanation that did not suffer from similar gaps and challengable extrapolations. Since you haven't ruled out evolution, the best you can do to "unseat" this theory from its present acceptance by scientists is to show that it is inferior to some other theory of species origins, but you have not described any alternative theory, so it is not clear that evolution needs to be "unseated."

How does creation grab you? You probably don't want to admit that possibility, but you can think of it as a default position. It cannot be demonstrated scientifically, not because of any philosophical defect in the proposition, but because of the limitations of Science. Because Science is incapable of dealing with it does not mean it hasn't happened. There are, after all, some truths in the physical world that cannot be reached by Science, just as there are mathematical truths that cannot be reached by mathematical proof. If we don't have a scientifically viable theory to account for the origin of the four human gamma genes, or for the origin of life itself, we needn't despair. Not every mystery necessarily has a scientific solution. I do not mean to say that one should not look for a scientific solution. One should. But not having such a solution is not a license to make up stories and pass them off to a gullible public as Science. Because I don't have a (scientifically) 'plausible' explanation of the origin of life, does not mean that your improbable stories are correct and should be foisted on the public under the guise of scientific truth.


7. Summaries by the Correspondents

Spetner's Summary

I have shown here, with references to my book, that the examples most often cited by evolutionists as evidence for evolution occurring now are not evidence at all for the grand sweep of evolution, which I have called here Evolution A. For an example of evolution happening now to have any relevance to Evolution A, it must be based on a mutation that could be typical of those alleged to be in the long series of steps that lead from a bacterium to a baboon. The mutation must at least be one that when repeated again and again will build up enough information to turn a bacterium into a baboon. The favorite example cited for evolution is antibiotic resistance. I have shown that the mutations leading to antibiotic resistance do not add any information to the biocosm. In some cases, they actually lose information. I have shown an example of a mutation that can easily be misconstrued to demonstrate the addition of information to the genome. Upon the gathering of further data, this example turned out to be a demonstration of information loss and not gain. Conclusion jumping is always risky, because we seldom have enough data. Yet, the evolutionist community has persisted in making the shakiest of extrapolations. Max has tried to argue that his triad of gene duplication, random mutation, and natural selection, can add information to the collective genome of the biocosm. I have exposed his argument as being nothing more that offering possible scenarios - it is argument by just-so-stories. But the argument against NDT does not stop with the failure of its supporters to show proper theoretical or empirical evidence for it. The telling blow against NDT is that examples of information addition have never been exhibited. The absence of such examples is more than just the absence of evidence for evolution. It is actually evidence against evolution because if NDT were correct, there should be millions of such examples and in all the genetic experiments performed until now we should have seen many. Finally, the example of mutations in the B cells of the immune system carries no weight as an example of a mutation that adds information. Although these mutations do add information to the B-cell genome, they cannot be applied to evolution for the reasons I laid out above. Dr. Edward Max made a valiant attempt to present a solid case for evolution in his posting on the URL cited above. That he failed is not because of any defect in the author. Dr. Max is an intelligent, competent, and articulate scientist. He has a PhD and an MD, and for many years has done research and published on the genetics of the immune system, and he has added to our knowledge in this field. If he could not make a good case for evolution, there must be something woefully wrong with evolution.


Max's summary

Although Spetner claims that mutations observed in experimental models of evolution uniformly lose information, I have tried to show that his metric for evaluating the information content of proteins has not been rigorously validated, and that his whole argument is therefore based on an untenable foundation.

He has also argued that immunoglobulin affinity maturation, which depends on mutation and selection of randomly mutated immunoglobulin genes, is not a useful model for phylogenetic evolution, but none of his objections convince me. The mechanism generating mutations may be different in the two cases, but since many experimental methods for generating mutations yield pools of mutants from which individuals with improved function can be selected, the specifics of the mechanism seem irrelevant to the idea that mutation and selection can lead to increased fitness. Spetner's argument about the differences in the rates of mutation in B cells versus germline cells also seems irrelevant, since we both seem to agree on these essential points: that single mutations can provide selectable advantages that could spread through the population after multiple cycles of reproduction; and that phylogenetic evolution is much slower than the B cell example because the mutation rate in germ cells must be much lower than what is feasible in the immunoglobulin genes of B cells.

Spetner has avoided specifying precisely what he means by his preferred model of "creation," so he avoids having to defend his model against scrutiny similar to what he has applied to evolutionary theory. Even a supernatural "creation" should leave traces that might be different from those expected from evolutionary theory. If his "creation" alternative does not make specific predictions that might distinguish it from evolution, it is not a useful scientific model. This may not bother Spetner, who has said that science is not the only source of knowledge; but as discussed below, it suggests that Spetner's views do not deserve consideration in science classrooms or textbooks.

Spetner's idea that evolution is being "foisted on the public under the guise of scientific truth" reveals a blurring of the distinction between scientific knowledge and religious dogma. Religious dogma based on an unchanging holy text may provide a "truth" that Spetner can accept without feeling any need to explain or justify it; such dogma, being immune from scrutiny, may be immune from revision and therefore represents an immutable "truth." In contrast, no responsible scientist suggests that our current scientific theories are immune from revision based on future evidence. We simply claim that, even despite areas of controversy and perplexing gaps in our current knowledge, evolution is the scientific theory most compatible with existing scientific evidence. When we discuss the origin of species in our science classes, there is no alternative theory in the scientific literature that we can teach. We therefore teach ("foist on the public"?) the only theory about this question found in that literature: the theory of evolution. (It is unfortunate that some science teachers go beyond the scientific evidence to claim that the theory of evolution rules out the existence of God; this is not a valid extrapolation, as I mentioned in section 4 above.) If Spetner feels he has evidence that the scientific literature on which classroom instruction is based is in error, he should argue his case in the professional scientific literature, not in an unrefereed book. Controversial views expressed in books but not in the professional literature cannot justifiably be foisted on students in science classrooms (at least not in public elementary/secondary classrooms) because there are no consistent standards of scholarship for book publication (as demonstrated by books on psychic powers, extrasensory perception, astrology and Holocaust denial).

Despite our opposing viewpoints, the correspondence has been interesting (to me at least) because Spetner is an intelligent and articulate scientist, who seems genuinely interested in a dialogue that tries to analyze where the differences in our positions lie. I hope that the correspondence will continue. If it does, I will update this summary. Readers who would like to make additional points relevant to this correspondence can Email me at


Home Page | Browse | Search | Feedback | Links
The FAQ | Must-Read Files | Index | Creationism | Evolution | Age of the Earth | Flood Geology | Catastrophism | Debates