ÁñÁ«ÊÓƵ

Today most people think of Lamarckism as the antithesis of Darwinism. This is because Lamarckism has come to mean a belief in the inheritance of acquired characteristics. This in its turn implies that heredity exists to reproduce the organism, to some extent as a photocopier reproduces documents. The "neo-Darwinian modern synthesis" that, according to the authors of this book, has "imposed an ideological strait-jacket on evolutionary theory", puts things exactly the opposite way round. This "DNA-centric" view, as they call it, sees the organism as evolving to reproduce its DNA. According to this way of looking at things, the gene and not the organism is the sole agent of evolutionary change and the ultimate unit of selection.

Eva Jablonka and Marion Lamb claim "there are mechanisms that enable the genome to sense an environmental change, respond to it, and transmit the response to descendants". They criticise Richard Dawkins for comparing genes to a recipe and suggest instead that there are "multiple inheritance systems" for which a musical score and a recording of a performance are a better analogy. Jablonka and Lamb "do not doubt that the basic mechanisms underlying the inheritance of acquired variations evolved in a Darwinian fashion", but ask "why should the evolutionary origins of Lamarckian inheritance systems be of any relevance when considering the effects these systems have once they are in place"?

The kind of thing they have in mind is imprinting, "the only biochemically well-characterised mechanism of epigenetic inheritance". Imprinted genes effectively carry biochemical gift tags that say "From Father", or "From Mother". A maternally imprinted gene is expressed only if inherited from the father, when it is said to be paternally active, while paternally imprinted genes are correspondingly maternally active. The biochemical labels are not DNA, so imprinting is what Jablonka and Lamb define as "epigenetic inheritance". It is Lamarckian to the extent that a characteristic of the parent is transmitted to the offspring.

Igf2 is an imprinted mammalian gene so-called because it codes for insulin-like growth factor 2. Its effect is exactly like the raising-agent that Dawkins mentions in his recipe analogy: it makes the fetus larger, just as more raising agent makes a cake larger. Applying Jablonka and Lamb's analogy, the DNA sequence of Igf2 is more like a score, and imprinting like a recording of the score. The organism is constructed according to both because a double dose of the father's Igf2 gene makes mice larger than normal, whereas a double dose of the mother's makes them smaller. Since normal development requires suitably imprinted genes from both parents, Jablonka and Lamb imply that a DNA-centric view is inadequate and requires us to take the epigenetic inheritance of imprinting into account.

David Haig and others have interpreted exactly the same facts in terms of intragenomic conflict: the realisation that, if selection acts on individual genes, it may do so even within the same genome, and probably will do so if genes are imprinted, thereby revealing their relatedness to a parent-of-origin. Haig argues that (except under strict monogamy) a paternally active gene like Igf2 has no necessary reason for ever finding itself in offspring of the same mother more than once. A maternally active one, however, will always have a 50:50 chance of being in any offspring a mother may have (since parents pass half their genes to offspring). Consequently, paternally active genes for growth like Igf2 are selected to exploit the mother much more than maternally active ones because they have no vested interest in her reproductive future beyond the immediate pregnancy.

If Jablonka and Lamb are to be believed, this should be no problem to a mammalian mother for three reasons. First, their claim that evolution is "environmentally directed" to the extent that the genome can allegedly sense an environment and respond to it predicts that the fetus should be more adapted to the mother than the mother should be to it simply because the mother is the environment of the mammalian fetus. Second, whereas fathers pass on only nuclear DNA, mothers bequeath an entire cell to the offspring (the egg), along with the epigenetic inheritance systems that go with it and are so exhaustively described in the bulk of the book. If epigenetic inheritance is as important as the authors claim, this should decisively add to the mother's advantage. Third, the mother and fetus constitute a unified group on which the authors claim selection can act. To revert to their analogy, the father may have written half the score, but the mother, along with composing the other half, is the sole performer with a claque for an audience. If the next generation are what they are because of both score and performance, mothers should have the upper hand.

But research by Haig and others into conflicts between mothers and fetuses during human pregnancy gives no grounds for such a conclusion. On the contrary, the evidence suggests an evolutionary arms race culminating in a stable if dangerous balance of forces. Even in the case of Igf2, mouse mothers respond, not by relying on epigenetic inheritance systems and group selection, but with a counter gene - Igf2r - that acts as a sink for the growth factor concerned, mopping it up to reduce its effect. Conflict between the individual genes that Jablonka and Lamb dismiss so disparagingly as part of the "ideology" of modern Darwinism seems a much better explanation of their best case of epigenetic inheritance than does Lamarckian evolution. Far from imposing a "strait-jacket" on evolutionary thinking, the DNA-centric view reveals conflicts Lamarckians would never have suspected and, despite Jablonka and Lamb's best efforts, still cannot explain.

Christopher Badcock is reader in sociology, University of London, and author of PsychoDarwinism.