Home | Intro | About | Feedback | Prev | Next

When The Tiger Comes, Freeze

What the Chinese whispers game tell us about evolution of complexity

MENDEL'S DEMON reviewed by Gert Korthof
24 Aug 2001 (update: 3 Feb 2002)


contents EVOLUTION
Chinese whispers. Asha Kaji Thaku
  DNA copying accuracy limits the increase of complexity of life on Earth. High error rates limit the number of genes that can be accurately copied. Furthermore and unexpectedly the Mendelian type of inheritance and sexual reproduction is a requirement for complex life. On the basis of this Mark Ridley predicts that complex life on other planets also must have the Mendelian type of sex.

Mendels Demon. 2000
Mark Ridley (2000) Mendel's Demon. Gene Justice and the complexity of life.


A year ago an alarming article on error rates in hard disks appeared in the Scientific American [2]. It stated that if disks grow in size the total number of errors grows too. Assuming a constant relative error rate, we don't notice the errors on small disks, but there is a real risk of encountering errors on disks with Giga and Terabyte capacity. This means that without increasing the error-correcting rate big hard disks are just unreliable. Ridley did not use this example, but it would be a perfect introduction to the main problem of his book: keeping the error rate down by all kinds of error correcting mechanisms based upon redundant information in DNA. (See for more metaphors the box Technological metaphors).

    Ridley uses the 'Chinese whispers' (or 'telephones' in North America) game as a metaphor. This metaphor does a marvellous job. It is the game in which a message is repeated from one person to the next, down a line. Ridley uses the message: WHEN THE TIGER COMES, FREEZE. (This one is easy to remember, but make it longer or try a string of 100 random letters). By the end the original message is laughably corrupted (see illustration). Ridley uses the following mutations of the original message:


These messages become a matter of life or death when real tigers are around! Short DNA messages can be copied reliably. The longer the message the more mistakes. RNA viruses have an error rate of 1 per 104 bases. Bacteria perform better. An error rate of 1 per 109 - 1 per 1010 bases and a message length of 1 million (106) bases produces 1 error per 1000 to 1 per 10,000 copies (offspring). There is a 100,000 reduction in error rates between RNA viruses and bacteria.

organismrelative errorabsolute error
RNA-viruses1 error per 104 bases <= 1 per offspring
bacteria 5 per 1010 bases 2 per 1000 offspring
worm, fruitfly1 per 108 bases 2 - 4 per offspring
humans 3 per 108 bases 200 per offspring

"Copying accuracy per letter is the same in all life except viruses" (p. 93). So for the rest of life the absolute number of errors simply depends on the length of the DNA message (compare with the errors on hard disks). Worm and fruitfly are just below 1 harmful error per offspring. Humans have a DNA length 2000 times larger than bacteria (about 33,000 genes) and produce 200 copying mistakes per offspring of which 2-20 are harmful. Stunning and alarming information. It's clear that it makes no sense to have 1 million genes. It would imply 6000 mutations of which 60-600 per offspring are harmful. No surviving offspring means extinction of the species. So there is an upper limit of complexity (number of genes). For other reasons Ridley states that "a life form cannot exist if it makes more than ONE mistake per offspring" (p. 78). So here is a paradox. We should all be dead. The human species should have been extinct for long. How we escape the paradox is the story of the book.

Unexpected predictions and explanations

Ridley's analysis leads him to predict that if complex life exists on other planets, then it must use Mendelian inheritance. To me this was a really unexpected prediction. Ridley's prediction is based on the insight that the details of Mendelian inheritance are crucial for complex life. For example the following puzzling details of the design of the process that produces haploid gametes from diploid cells (meiosis):

  diploid (2n) >> tetraploid (4n) >> diploid (2n) >> haploid (n)  
in stead of one step:
  diploid (2n) >> haploid (n)  

Ridley's analysis leads him to predict that the simple one step meiosis should not exist. This is an eminently refutable prediction. I never realised the details of meiosis were unexplained. Ridley has an evolutionary explanation for it. In my view this proves the power and profoundness of his evolutionary analysis. No other alternative theory ever gave such detailed explanations. No other alternative theory pointed out the problem in the first place. If the success of a scientific explanation is measured by its ability to elucidate brute facts, then surely we have here a successful explanation at hand.
book "Mendel's Demon. Gene Justice and the Complexity of Life"
by Mark Ridley

Weidenfeld & Nicolson, Great Britain
2000, hardback 337 pages.
ISBN: 0 297 6 4634 6
(USA title: "The Cooperative Gene: How Mendel's Demon Explains the Evolution of Complex Beings").

 1Keeping living things simple1
 2The gene number of the beast28
 3The mutational meltdown56
 4The history of error80
 5The ultimate existential absurdity109
 6Darwinian mergers and acquisitions134
 7The justice of the peas167
 8The long reach of the lawbreaker202
 9The human condition233
10A complex future257
 Glossary, notes, bibliography, index281-337
     Another example of an evolutionary prediction is the origin of gender and the unexpected relation with the origin of mitochondria (the famous 'Symbiosis theory' of Lynn Margulis). Ridley predicts that complex life elsewhere in the universe will usually lack gender. To predict these kind of things one really needs a deep understanding of life. Religions say that God created man and woman. Yes, we know that there are two sexes. We can see that. But why? And why gender? Why 2 different sized gametes? Why one gamete with and the other gamete without mitochondria? Does religion give the answers? According to Ridley gender completely depends on a historical accident: the Margulian merger. Without that merger gender would not exist. This is the most profound explanation of gender I encountered so far.

     There are many more rewarding insights in Ridley's book such as the relation between the existence of sex and the existence of mutations (sex is pointless without mutations); Kondrashov's prediction about the effect of mutations in sexual reproducing organisms; why female chromosomes are expelled in meiosis at fertilisation; the reason for gestational diabetes in humans; why genes inherited from the father stimulate fetal growth and inherited from the mother limit fetal growth; why color blindness has increased in rich countries and why death rates of single versus married people differ. Ridley even proposes a new cure for AIDS based upon the principle of mutational meltdown. In short: Mark Ridley shows that evolutionary theory makes sense of puzzling details of cloning, sex, meiosis, Mendelian inheritance, gender, the Margulian merger, pregnancy and mutations rates (cloning of humans is in the news. He tells a few useful things about the strategy of cloning from a genetic-evolutionary point of view). I found it extra rewarding that Ridley applies all his knowledge to our own species (chapter 9 and 10). In fact the book is more about the humans species, than complex life in general. Many of the theories Ridley explained in his book turn out to be of great help to understand new findings in biology, which otherwise would be incomprehensible.

Is error elimination the core of neo-Darwinism?

The Chinese whispers metaphor dominates the book and his thinking. "The Chinese whispers is actually a powerful way to think about mutation." (p. 63). "The analogy points to a property not just of children's games, but of life too." (p. 63). "There is only one way of being right (the original message) and so many ways of being wrong (anything other than the original message)." (p57). I have two problems with the Chinese whispers metaphor and the underlying idea:
(1) Typically the Chinese whispers metaphor starts with an error-free, uncorrupted, meaningful message: THE ORIGINAL. However evolution is not a game starting with a correct and meaningful message. (Do 'correct' messages exist anyway?). The CW metaphor obscures the question of the origin of correct and meaningful messages, because it simply assumes the existence of a correct and meaningful message.
(2) If the original is by definition error-free, then all DNA mutations must be corruptions of the original. There is no place for improvement. The notion of a 'creative mutation' seems absurd and a 'creative error' becomes self-contradictory. Ridley strengthens this idea by using a speech from Hamlet as the original message (3). No wonder Ridley is inclined to think that zero errors is the ideal (4). It looks as if life is just a Xerox machine. The goal of a Xerox machine to make 100% identical copies and nothing more. That's exactly what copying machines are for. We are not interested in a copying machine that makes all kinds of creative variations.
    Please note that the Chinese whisper metaphor is compatible with the created and fixed species model of the Bible. Species contain the original God-given DNA message and that message must be safeguarded. There is no need to evolve new genes, because they are there from the Beginning. Surely Ridley does not support such a view!
    Of course Ridley has the right to choose his subject: "our topic here is not evolutionary change. Our topic is mutational decay and how life preserves itself against it." (p. 62). In that context the metaphor is helpful indeed. But clearly Ridley wants more than explaining mutational decay alone. He wants to understand "the Evolution of Complex Beings". In that wider context the metaphor does more damage than good, because improvement and conservatism are antagonistic. For innovation the more mutations the better; for conservation : the less mutations the better. I think that finding a balance between too much and too few mutations, is what makes evolution an exciting subject. It's just too easy to focus on one aspect and forget the other. Proofreading enzymes cannot selectively correct harmful mutations. So if proofreading enzymes lower the total number of mutations, the number of useful mutations is lowered too.

Let's have a look at 8 examples of how the Chinese Whispers metaphor controls Ridley's evolutionary thinking:

  1. "Most of the work of natural selection would be concerned with preserving messages" (p. 106). "Natural selection is described as a theory of evolution, and indeed it is one, but it is a theory of non-evolution too" (!) (p. 106). Do we really need a theory of 'non-evolution'? Yes, we need a theory of non-evolution for everything that is conserved: Histone-H4, 'living fossils', and every successful invention. However we certainly need a theory of evolutionary innovation to explain complex life!
  2. For Ridley 'enhancing the power of selection' does not mean to enhance natural selection as a creative power, but as an eliminative force (p75).
  3. "Errors can be good, bad or indifferent. We can assume that the good ones are a negligible fraction of the total." (p82). But then evolutionary innovations are based on a negligible fraction! Whether or not Ridley is interested only in harmful mutations in the context of his book, any evolutionary theory designed to explain increase in complexity would have difficult times. If good mutations are negligible, it will be harder to explain complex life, certainly not easier. If somebody tells you that a lottery is a waste of money, because the chance to win is negligible, would he understand the game?
  4. In discussing the high mutation rates of HIV Ridley states: "On the argument of this book, however, mutations are a net disadvantage: more means worse" (!) (p89), (12). If mutations are a net disadvantage, can neo-Darwinism really explain innovations? Are we better off with no mutations? But then evolution would come to a halt (11).
  5. Ridley states that 2-20 of the 200 copying mistakes per offspring are harmful. But how much are beneficial? The question is not even asked. Evolution needs those mutations! Are evolutionary biologists unable to recognise beneficial mutations?
  6. Sex helps to remove bad genes. There is no mention of the possibility of concentrating good genes (p. 112). Kondrashov's theory of sex is about bad genes. There is one exception in Ridley's book: courtship reveals good genes (p126).
  7. The broken car analogy is beautiful, but it's again about combining errors. There is no improvement of the original car design. There is no innovation in this analogy.
  8. When discussing the future of the human species Ridley writes: "Evolution can be propelled forwards by the new technologies of error prevention and error therapy" (p. 265), (5). Forwards? Go back to the ORIGINAL and stay there!
    Summing up: Is error elimination really the core of neo-Darwinism? (6). Or is Ridley's focus on disadvantageous mutations simply a consequence of the topic he chose? Or is it a reflection of the current research in neo-Darwinism? Or are we witnessing here "the neglect of the creative role of mutation" by mainstream neo-Darwinism? (9) Or is it a reflection of the fact that evolutionary innovations are rarer than harmful mutations and therefore harder to study? (10) (certainly no reason to ignore innovations!).
Error reduction and innovations are intimately connected. Error reduction itself depends on the evolution of 50 new proofreading genes. Where did they come from? The secret of writing good books is not good spellcheckers, proofreaders, or editors, but a good author. Although Ridley discusses increase in complexity in chapter 2, that (complex!) story lacks a powerful metaphor and is contradicted by the Chinese whispers metaphor.
    Generalising beyond Ridley's book: is the Evolutionary Synthesis (7) really completed as long as we don't have a theory that explains the origin of evolutionary innovations: new genes, new organs, new organisms? Wouldn't it be time to sort out if enough resources (time, reproductive capacity, number of beneficial mutations) are available to create complex adaptations? I hope Ridley's next book (Evolutionary Innovations (15) or Creative Evolution) contains a chapter discussing how life produced those 50 proofreading and repair enzymes. And a chapter discussing how all this can be done with a 'negligible fraction of mutations'. In short: we need a theory of evolution that does not take the core question for granted, but systematically explains how life got from 0 to 35,000 genes. And how animal body plans got from 1 to 35. I think it can be done, but somebody has to write it down and do the necessary research. Yes, I have somebody in mind! (14), (15).
    Is there an alternative for the Chinese whisper metaphor? I think the 'trial and error' does a better job in the context of evolutionary innovations necessary for complex life. Mutations are trials. Trials can be successful or unsuccessful, but they always are trials. We can eliminate error completely by eliminating trial at the cost of never finding anything new, and be in a frozen state forever. This is incompatible with evolution. Stuart Kauffman uses the trial and error metaphor extensively, almost exclusively (8). In this metaphor evolution is a search in sequence space, protein space or shape space. A search is not assuming a correct message. Indeed we need to conserve our successes, but we have to discover them in the first place. A complete theory of evolution needs to integrate the conflicting metaphors of the Chinese whispers and the trial and error metaphor.


Is the book difficult? Mendel's Demon is about concepts, not about animal and plants. Knowledge about fundamental biological concepts (mitosis, meiosis, Mendelian inheritance, mitochondria, Margulis symbiosis theory, haploid, diploid) is helpful, but these things are explained and illustrated in the book (glossary). Apart from that, one does not need to study many other books to understand Ridley (13). The success of Ridley's book lies in his creative use of well known metaphors: Chinese whispers, encyclopaedia, broken cars, meltdown, the holy writ, business merger, etc.
As an exposition of the limiting factors in evolution of complex life, Ridley's book is splendid. The explanation of many odd details of meiosis, Mendelian inheritance, sexual reproduction and gender demonstrates the fruitfulness of current neo-Darwinist research. Far from being side issues, these issues are fundamental to understand multicellular life. There is simply no competing theory in the whole field of biology that equals the explanatory power of neo-Darwinism. At the same time Ridley's book shows the limitation of current neo-Darwinism: the avoidance of the issue of innovation. I really hope that Ridley's next book is about evolutionary novelty. If anyone can do it, he is the man.


  1. I wish to thank Asha Kaji Thaku for his beautiful Chinese whisper drawing. Thaku is a design artist and cartographer with the International Center for Integrated Mountain Development in Nepal.
  2. Paul Wallich To Err Is Mechanical, Scientific American, 1999.
  3. (p57). Compare also with the metaphor 'genomes-are-compositions-of-J.S. Bach'. Please don't change them, they are error-free and cannot be improved, let alone that random errors could improve a Bach composition.
  4. in the New Scientist 21 July 2001 p49 Mark Ridley states that the idea in a novel - a biologist develops mutation-free DNA to do away with sex - "exactly matches my current thinking".
  5. I would like to add two useful features to evolution: (1) a mechanism that keeps successful genes frozen, (2) a mechanism that exchanges successful innovations between (multicellular) species so they profit from each others inventions.
  6. Probably I wouldn't have asked these questions if I had not read Lee Spetner, Fred Hoyle and Michael Behe (see reviews on this site). Remarkably the critics of evolution ask the core questions! Furthermore, good questions are asked by the new evo-devo discipline 'Evolutionary Developmental Biology'.
  7. See review of Smocovitis' Unifying Biology. The Evolutionary Synthesis and Evolutionary Biology.
  8. Stuart Kauffman At home in the universe, p161.
  9. Wallace Arthur(2000): The Origin of Animal Body Plans. A study in Evolutionary Developmental Biology, p244. He called the emphasis on the creative versus the destructive forces of mutation and selection the 'Fourth Eternal Metaphor' in the history of evolutionary biology.
  10. Indeed during the last 500 million years no new animal body plan emerged.
  11. However on the argument of Stephen Palumbi's book, The Evolution Explosion, p 121, a decrease of the error rate of HIV will slow down the development of drug-resistance, so will be disadvantageous for the virus (and advantageous for the patient). So in a stable environment a low mutation rate is advantageous, in a challenging environment (drugs) a higher mutation rate is advantageous for the virus.
  12. This is confirmed by evolutionary biologist Michael Majerus (2003) "Sex Wars", p.205: "it is fair to say that, on the whole, mutation tends to destroy adaptation".
  13. A good one would be Matt Ridley (1994) The Red Queen. Sex and the Evolution of Human Nature. (Penguin paperback). This is a very good introduction to evolutionary thinking of the last 25 years. [ 11 Dec 2004 ]
  14. A few years later, I think Sean B. Carroll's The New Science of Evo Devo (2005) is the best book on evolutionary innovations, because evo devo has the tools and the evidence to detect evolutionary innovations. [ 20 May 2005 ]
  15. The book has been written by Andreas Wagner (2011) The Origins of Evolutionary Innovations: A Theory of Transformative Change in Living Systems. [ 11 Oct 2011 ]

Further Reading:

  • email Email from Mark Ridley on the Feedback page of this site.
  • A review of Mendel's Demon by Sean Nee: "We do need sex, really". New Scientist 9 Sep 2000. (Apparently the page has been removed: a bad idea).
  • A review of Mendel's Demon by Andrew Berry: "Mitigating mutations", Nature 412, 379-380 (2001). (Subscription required).
  • Mutated into Oblivion is a review of Mendel's Demon by Lynda F. Delph in Science, Vol 292, Issue 5526, 2437-2438 , 29 June 2001. (Subscription required).
  • A short review of Mendel's Demon by William C. Summers: "Genes by Context" (together with 2 other books). American Scientist, Nov-Dec 2001, Bookshelf.
  • A review by Mark Ridley of Ernst Mayr's What Evolution is in Nature 417, 223-224 (2002).
  • A review by Mark Ridley of Tom Wakeford's "Liaisons of Life. From Hornworts to Hippos, How the Unassuming Microbe Has Driven Evolution" in The New York Times, April 15, 2001.
  • A review by Mark Ridley of Simon Conway Morris' "The Crucible of Creation. The Burgess Shale and the Rise of Animals" in The New York Times, May 10, 1998.
  • A review by Mark Ridley of Darwin's "The Expression of the Emotions in Man and Animals" in the Scientific American.
  • A review by Mark Ridley of two books of Richard Lewontin (It Ain't Necessarily So: The Dream of the Human Genome and Other Confusions; The Triple Helix: Gene, Organism, and Environment) in Nature 406, 347-348 (2000).
  • A review by Mark Ridley of Michael Ruse(2003) Darwin and Design: Does Evolution Have a Purpose? in Nature 423, 686-687 (2003) 12 June 2003.
  • John Waller shows why Ridley is probably wrong when stating: "Mendel's insight was that the attributes of an individual are controlled by pairs of hereditary factors..." (p.170). See the review of Fabulous Science. Fact and Fiction in the history of scientific discovery (John Waller).
  • Miguel Constância et al (2004) "Resourceful imprinting", Nature, 432, 53-57 (4 Nov 2004). "A child's genes are not all equal: in some cases, the copy from either the mother or the father is tunrned off. This affects the child's ability to acquire resources in the whomb, after birth, and perhaps throughout life." Keywords: 'kinship theory' or 'genetic conflict theory'.

Technological metaphors

    Everybody familiar with Shannon's theory of information transmission and Hubert Yockey's work will immediately recognise the theme of Ridley's book: Data transmission, noise and noise reduction. See: review of Hubert Yockey's 'Information theory and molecular biology'.
    Another example from technology could introduce Ridley's theme: digital audio-CD's. The CD's contain errors caused during production or damage thereafter. However the information can be reconstructed by the extra redundant data that are added at production time. There is a limit to the damage that the errorcorrecting mechanism can cope with. Especially when we are dealing with copies of copies of copies (...) there is no way to reconstruct the original or anything close to it, analogous to mutational meltdown.
    In the area of webgraphics there is still another metaphor: lossy and lossless compression. GIF images produce a lossless compression. The decompressed image is exactly the same as the original. The JPEG process is always lossy compression (even with a minimal compression factor), so there is always loss of information.The Chinese whispers game can be simulated by subsequently compressing and decompressing a jpg image and repeating this several times. The result will be a steady degradation of the image until only noise rests. This simulates a mutational meltdown. The speed of degradation is determined by the compression factor and is analogous to the mutation rate.
    Richard Dawkins(1995) in his River out of Eden (p22) referred to Chinese Whispers when explaining the difference in reliability between an analog genetic system and a digital genetic system. Analog systems degrade when copied, while the DNA code is a digital code and is the only code that can be reliably copied over eons of geological time.

    Mark Ridley is the author of the following textbook of evolution.
[ in 2004 the third edition was published ]

book Evolution
by Mark Ridley
Second Edition
Blackwell Science
719 pages.
ISBN 0-86542-495-0

Part 1 Introduction.
 Ch  1 The Rise of Evolutionary Biology
 Ch  2 Molecular and Mendelian Genetics
 Ch  3 The Evidence for Evolution
 Ch  4 Natural Selection and Variation
Part 2 Evolutionary Genetics
 Ch  5 The Theory of Natural Selection
 Ch  6 Random Events in Population Genetics
 Ch  7 Molecular Evolution and the Neutral Theory
 Ch  8 Two-Locus and Multi-Locus Population Genetics
 Ch  9 Quantitative Genetics
 Ch 10 Genome Evolution
Part 3 Adaptation and Natural Selection
 Ch 11 The Analysis of Adaptation
 Ch 12 The Units of Selection
 Ch 13 Adapative Explanation
Part 4 Evolution and Diversity
 Ch 14 Evolution and Classification
 Ch 15 The Idea of a Species
 Ch 16 Speciation
 Ch 17 The Reconstruction of Phylogeny
 Ch 18 Evolutionary Biogeography
Part 5 Paleobiology and Macroevolution
 Ch 19 The Fossil Record
 Ch 20 Rates of Evolution
 Ch 21 Macroevolutionary Change
 Ch 22 Coevolution
 Ch 23 Extiction
Answers to study and review questions

Valid HTML 4.01 Transitional
guestbook (moderated) homepage: wasdarwinwrong.com wasdarwinwrong.com/kortho27.htm
Copyright ©G.Korthof 2001 First published: Aug 24 2001 Update 3 Feb 2002. Links/Notes: 11 Oct 11