Evolution in Four Dimensions: Genetic, Epigenetic, Behavioral, and Symbolic Variation in the History of Life. Eva Jablonka and Marion J. Lamb. MIT Press, Cambridge, MA, 2005. 472 pp. $34.95 (ISBN 0262101076 cloth).
In the preface to the second edition of Descent of Man in 1874, Darwin re-marked,
My critics frequently assume that I attribute all changes of corporeal structure and mental power exclusively to the natural selection of such variations as are often called spontaneous; whereas, even in the first edition of the ‘Origin of Species,’ I distinctly stated that great weight must be attributed to the inherited effects of use and disuse, with respect both to the body and mind.
The authors argue that four inheritance systems characterize biological systems. The first is the DNA-based genetic system. The second is the epigenetic inheritance system, whose role in the development of multicellular organisms has led to such wonderful molecular work in recent decades. The third is what they call the behavioral inheritance system, encompassing most forms of social learning in animals. Their fourth form is symbolic inheritance, exemplified by the human use of language to spread ideas. Each of these systems has a considerable number of subsystems. Epigenetic inheritance includes self-sustaining loops of gene activity, the transmission of variable cytoarchitectures, chromatin marking, and RNA silencing of genes. One could quibble with the four-part taxonomy, but it serves the purpose of the book well enough.
The existence of such a wide variety of inheritance systems will be an eye-opener for many biologists who have not had the opportunity to peruse work outside their own fields. I found the chapters on the genetic and epigenetic systems full of new and interesting information. Molecular biologists are likely to find the chapters on the behavioral and symbolic systems equally enlightening. Clear writing, illustrated with stories and quirky cartoons, makes the book a good read for any biologist.
But this book is not a simple recitation of interesting bits of recent biology. It is a hard-argued polemic for the importance of the inheritance of acquired variation across the whole spectrum of inheritance systems. If Jablonka and Lamb are correct, some form of inheritance of acquired variation is important in every organism. Some of these cases are uncontroversial. The behavioral and symbolic systems probably owe their adaptive properties to the fact that what one individual learns can be acquired by others via social learning. In one case that Jablonka and Lamb describe, a population of black rats exploits an Israeli pine plantation as if the rats were squirrels. Some ancestral animal or animals discovered a technique for efficiently extracting seeds from pinecones. Now, in a considerable population of rats, the behavior is sustained by behavioral inheritance. Rat pups learn the technique by handling the partly opened cones their mother discards. Individual rats without such exposure have never learned the technique on their own in the laboratory. In the epigenetic inheritance system, cell lineages up-regulate some genes and silence others in response to signals received during embryogenesis. The chromosome marks that transmit this information do not alter DNA sequences, only gene activity. Epigenetically acquired variation is the fundamental mechanism by which the development of multicellular organisms takes place.
More controversially, does DNA inheritance exhibit the inheritance of acquired variation from generation to generation? Jablonka and Lamb are inclined to think that it commonly does. They argue that abundant evidence exists for contingent adaptive modification of the DNA inheritance system. Many species undergo facultative sexual reproduction in stressful environments. The mutation rates of bacteria, at least, seem to increase with stress. Pathogenic bacteria are hypermutable at loci that deal with the host's rapidly evolving immune system, even while other loci have normal rates of mutation. These features are not quite “instructed” evolution, but they verge on it. Mutation rates increase adaptively, but the direction of mutation is still blind with respect to adaptation.
In any single-celled creature, epigenetic changes will be transmitted to offspring, and some evidence suggests that the epigenetic inheritance systems that underlie the development of multicellular animals are derived from epigenetic systems that single-celled organisms used for the adaptive inheritance of acquired variation between generations. Multicellular organisms without the separation of the germ line from the soma early in development can likewise transmit acquired variations across the generations. Plants do not exhibit early germ-line segregation and hence can transmit epigenetic variation from generation to generation. E. J. Steele's highly controversial idea that somatically selected immune system variants might somehow be spliced by viruses into the germ line is mentioned in passing, though Jablonka and Lamb do not defend its plausibility. Perhaps the Weismannian animals can afford to segregate the germ line because they mainly use behavior to deal with variable environments and can often use behavioral transmission in lieu of other Lamarckian systems.
The interaction of inheritance systems is another way in which “instructed” Lamarckian effects can influence the DNA inheritance system indirectly. This idea goes back to Weismann's original idea of germ-line segregation. Conwy Lloyd Morgan, James Mark Baldwin, and Henry Fairfield Osborn all independently discovered what has come to be known as the Baldwin effect in the mid-1890s, after Weismann's influence was felt but before the rediscovery of Mendel's laws launched genetics. The Baldwin effect reconciled the apparent inheritance of acquired variation with Weismann's doctrine of segregation of the germ line. In the mid-20th century, C. H. Waddington explicitly linked instructed effects to genes. Animals learn, and what they learn alters the selection pressures that bear on their genes. Any mechanism of phenotypic flexibility will do. For example, an animal introduced into a new habitat may use some form of phenotypic flexibility to survive and reproduce. In so doing, it exposes its genes to selection, which will tend to make traits originally acquired innate.
Organisms also modify their niches in many ways, leading to “niche construction,” as John Odling-Smee and his colleagues call this effect. Selection will then adapt a species to the niche that it has constructed. Beaver dams and the aquatic adaptations of the beaver are an example. Social learning can do the same thing. Human symbolic culture leads to massive changes in our physical and social environment, and surely selection has favored genes adapted to such environments. Genes and culture can be said to coevolve. Jablonka and Lamb illustrate the concept of gene–culture coevolution with the example of adult lactose absorption. In all other mammals and in most human populations, lactase synthesis in the gut ends after weaning. In European and African populations with a long history of dairying, lactase synthesis continues in most adults, allowing these populations to make efficient use of fluid milk.
Jablonka and Lamb go some way beyond current evidence in envisioning major roles for Lamarckian processes in evolution. For example, the hard evidence from experimental studies of non-human social learning indicates that behavioral transmission is present in many social species, but that most such systems support only the transmission of a few simple variants. To be sure, observers of animal behavior in the field typically report more, and more complex, social learning than experimentalists can replicate in the lab. Some time will pass before this gap is closed. Much of the evidence in support of Lamarckian processes is still fragmentary, and the final weighing of the evidence might find them to be of substantial importance only in special cases like human culture, and to be curiosities elsewhere.
Darwin's acceptance of the inheritance of acquired variation turned substantially on his acceptance of evidence from poor experiments. But he also had an adaptive intuition. Given that organisms have sophisticated systems for acquiring adaptive variation, why would selection favor writing off each generation's investment in adaptive acquired variation and force its offspring to repeat a costly course of phenotypic adaptation? Theoretical models of cultural evolution show that this intuition is cogent, given spatial and temporal variation that is autocorrelated on the generation-to-generation time scale. Skepticism about Lamarckian processes is warranted, but Jablonka and Lamb marshal enough evidence to make dogmatic claims of the absence of such processes equally deserving of skepticism. Where the weight of the evidence eventually comes down will be of great interest.
