Like books along a bookshelf

Mark Ridley

  • The Wisdom of the Genes by Christopher Wills
    Oxford, 351 pp, £6.99, January 1991, ISBN 0 19 286113 1

The inherited instructions by which a body is built are carried from parent to offspring in molecular form, in the DNA. The instructions come in units, called genes, and for most purposes the DNA can he imagined as a long row of genes, arranged like books along a bookshelf. It is not known exactly how many genes are used to build a human body, but the number is between 100,000 and a million, and is probably nearer the former. Evolution (on this simple conception of the DNA) happens when the environment changes and one form of a gene rather than another builds a better body in the new conditions, or when a gene mutates to an improved form. Either way, natural selection favours the superior form of the gene, which becomes more numerous. Evolution in this sense means a change in gene frequency. It is a powerful way of thinking about evolution, and evolutionary biologists use it all the time. Taking one gene at a time, they study (or, at a more rarefied theoretical level, imagine) different forms of the gene, and see how natural selection works on them.

Like all good things, however, it has its critics, Ernst Mayr, for instance, strongly objects to it. He says it makes evolution seem like changing the bean composition in a beanbag, and refers to it derisively as ‘beanbag genetics’. One of Mayr’s ‘beanbag’ geneticists was J.B.S. Haldane – and one year before his death Haldane still knew how to conduct a controversy. He ironically adopted Mayr’s term, and showed how the beanbag model had led to almost all our understanding of evolution. Thirty years later, much the same still applies. Mayr’s mistake was to confuse a cartoon with a portrait. Scientists make simplifying assumptions on purpose, in order to identify the essence of a problem and eliminate the unnecessary detail; their models do not attempt to describe nature accurately, in all its complexity. When they learn how to handle sonic previously ignored feature of nature, it can be built into a model soon enough. Meanwhile, not much is to be gained by talk, particularly if it is stuffed with slogans about reductionism and holism.

Christopher Wills is completely secure with genetic theory, and makes no attempt to belittle the beanbag model. The Wisdom of the Genes, a generally admirable book, suggests how modern genetics is leading evolutionary biologists away from the beanbag model, to think instead about how genes are organised in our DNA. Evolution has not scattered the genes in any old order, but like a wise librarian has ordered the genes in subtle schemes which enable them to operate more efficiently. Genes (unlike the contents of books) evolve through time, and Wills suggests that some genes are arranged in such a way that they can change to form new types of body more rapidly. Such is the ‘wisdom of the genes’. Wills, by the way, can be rather erratic on history: in the case of beanbag genetics, he credits the term to Haldane and ignores the ironic borrowing from Mayr. Wills thus effectively distances himself from the earlier critics of beanbag genetics – and that is undoubtedly the best relationship to have with them.

Wills’s book is not a closely, or forcefully, reasoned argument. What he does is to build up an impression with a series of suggestive examples. His main skill – and the reason I strongly recommend the book – is in explaining individual genetic case-studies. It is a rare skill. We have several excellent writers on theoretical genetics; and every genetic laboratory is full of people with an alarming but seemingly incommunicable knowledge of the most recent discoveries. Wills both under stands the latest molecular genetics and writes clear and attractive English. He also knows when to give the reader a rest with an entertaining anecdote, and he comes across as a civilised and liberal observer of the world. Nonspecialists should find it quite easy to understand his account of the extraordinary behaviour of jumping genes, of the way the structure of real proteins controls their function, and how DNA is sequenced. There are at least twenty pieces of genetic research that I did not understand before reading Wills’s book – but that does not mean I found his main theme compelling.

Let us look at one of his examples. Papilio and Heliconius are two beautiful kinds of butterflies. The papilios are the swallowtails; they are typically coloured in yellow and black and have a tail to their hindwings. They are difficult to find in the UK, being confined to a small spot in the Norfolk Broads, but they are much commoner elsewhere and I often see them flapping around gardens in Atlanta. The interesting species, in Wills’s story, are the ones that have evolved complex kinds of mimicry. Swallowtails are not poisonous, and are readily eaten by birds; but some kinds mimic other butterflies that are poisonous. Two well-studied species even produce multiple mimics. Thus, in the Indonesian species Papilio memnon the males all look the same and are nonmimetic. The females mimic other species, and are variously coloured in black, white, and shades of red, orange, pink or yellow, many in decidedly non-swallowtail-like patterns. Some kinds of female have tails, others do not. Many of the different female forms live in different places, where they mimic whatever other poisonous butterfly is locally abundant; but in any one place there is usually more than one mimetic female form. In Borneo, for instance, there is both the form called venusia, which has all orange hindwings and abdomen, and the form laomedon, which has mainly black hindwings, with a region in yellow, and a black abdomen.

How do the genes control the mimetic colour patterns? An individual’s full pattern is made up of at least five components: the general colour, the forewing colour pattern, the hind wing colour pattern, the presence or absence of a tail, and the colour of a conspicuous spot on the forewing called the epaulette; each of the five components, it turns out, is probably controlled by one gene. In this multiple genetic control lies a great danger. For a butterfly to mimic its model successfully, it should match it in all five respects. If, in a region where there is more than one kind of mimic, a female of one mimetic form were to mate with a male containing genes for another form, their progeny might be a collage of colours from the two different forms. The offspring might have the forewing to mimic one model together with the hindwings of another, or have no tail together with the colour patterns of a tailed model; and they would then mimic neither. In fact, this does not seem to happen in nature. All the swallowtails belong to the forms that are good mimics; no collages are found. The reason is that the five genes are inherited as a single unit, called a ‘supergene’. The five genes in the supergene are physically very close together in the DNA, and crosses between two forms generate offspring that have one of the complete sets of mimetic genes.

This cannot have happened by chance. It must mean that the five genes have moved together during evolution, to form the supergene. Indeed, in the comparable African species Papilio dardanus, the gene controlling presence or absence of a tail is separate from the genes for colour patterns, unlike in the Indonesian P. memnon in which the tail gene is part of the supergene, having presumably joined the colour genes during the evolution of P. memnon.

If we go back to the library analogy, it is as if a five-volume set existed in two editions, and in both cases constant cross-reference between volumes was required if they were to be used efficiently. If both sets were scattered randomly through the library, an unlucky reader might take out a mismatched set – only to get lost in blind cross-references, contradictions and incompatible terminology. A wise librarian would ensure that the compatible five volumes were shelved next to each other, and in no danger of being used with the other edition. He or she might even glue the five volumes together. In the swallowtail, the five co-operating genes encoding the mimetic pattern are physically so close they behave as it they were glued together. It was a major exercise to show they were five genes, and not just one.

Wills contrasts the mimetic swallowtails with the South American butterflies called Heliconius. Heliconius also has multiple mimetic forms. And they are also controlled by many different genes. In Heliconius, however, only one mimetic form lives in any one place and there is little danger two forms would interbreed. Unlike the swallowtails, their mimicry genes are scattered at random through their DNA. If you take two forms and cross them experimentally, you generate a mimetic disaster. The offspring contain an array of non-mimetic forms, with parental colour patterns all jumbled up. This is presumably so unlikely to happen in nature that the genes have not been assembled into units. (Actually, there are some small regions where more than one kind lives, but I am not sure what happens there.)

So in swallowtails it is wise to have the mimetic genes in a unit, whereas in Heliconius it does not matter how they are arranged. The reasoning in the case of these butterflies is relatively clear-cut. If we look into the DNA of any organism, including humans, we find groups of genes arranged in apparently non-random fashion: real genes are often not scattered haphazardly through the DNA. Some process analogous to the selection that favours complete mimicry in swallowtails is probably the explanation, but it is difficult to say exactly what. Wills uses the example only to illustrate a possibility – that genes can be wisely arranged. There is a lot of work left to do before we know what the analogous force would be in other species and other genes.

Jumping genes are another of Wills’s examples. They are particular genetic sequences which have the property, under some conditions, of copying themselves and reinserting at various places in the DNA. They may also pick up adjacent genes and carry them about. Jumping genes may have only trivial effects on evolution; but Wills argues the opposite. Some environmental changes, or shocks, may stimulate the jumping genes into action: under normal conditions they may sit still but when things change they may move. When jumping genes move, they cause mutations. As the sequence inserts itself at a new site, the gene on the receiving end necessarily mutates. An environmental shock could therefore lead to a general increase in the mutation rate via its effect on jumping genes. The higher mutation rate might in turn increase the rate of evolution. It is precisely in times of environmental change that rapid evolution may be needed, and the possession of jumping genes would then be an evolutionary advantage– another instance of Wills’s genetic wisdom.

He describes an experiment to test the idea. Some classes of jumping genes can be made to move or to keep still, by appropriate treatment. Trudy McKay, for example, was able to create two lines of fruitfly: one with still and the other with mobile jumping-gene sequences. She then artificially selected both lines for high, or for low, bristle number (the bristles on a fruitfly, being easy to see, are often used in artificial selection experiments). That is, she formed the next generation of fruitflies by picking out the individuals that possessed high (or low) numbers of bristles, and breeding only from them. She then measured the ‘response’ to selection: i.e. any change in average bristle number in the fly population between generations. In the first experiment, the line with mobile jumping genes increased (or decreased) in bristle number at twice the rate of the line in which the jumping genes were kept still Unfortunately, later experiments were less clear: but the first one shows the sort of thing that might occur.

Once a species has obtained jumping-gene sequences, it might become (as Wills puts it) better at evolving than it was before. Wills even proposes that species, over evolutionary time, become ever better at achieving future evolutionary change. A species, like the mimetic swallowtail butterfly, with its genes arranged in units can better track any future need for changed mimetic patterns; by possessing jumping genes, a species can adjust its mutation rate at times of environmental change ... Wills has other ideas along the same lines. The wisdom of the genes, he believes, has increased through time.

How convincing is his argument? His intention is, in the main, only to be suggestive, and in that he is successful. But a sterner scepticism is also in order, and it should be said that Wills is supported neither by evidence nor by probability. We have no evidence that recent species evolve any more (or any less) readily in relation to environmental change than their ancestors did. A genetic mechanism which facilitates evolutionary change can only be favoured when that change is needed – and these times will be rare. A mechanism for rapid adaptation to the circumstances of Ice Ages can only have been favoured once every 100,000 years or so, and it will be a weak force if there is any counter-selection. As Sydney Brenner once remarked, natural selection could not cause a Cambrian species to retain a gene ‘because it might come in useful in the Cretaceous’. Wills is well aware of this problem, and has sensible things to say about it, but he does not conclusively answer it. For much of the book, however, all this remains behind the scenes. The stage is filled with amazing genetic phenomena, and Wills is at his best when he explains them. I am quite willing to put up with his dubious (as I believe) main argument, for the sake of his highly readable accounts of the mutation rate of the Japanese survivors of atomic bombs (almost unaffected, it seems), the workings of the masseter muscle, and the elevated activity of DNA repair mechanisms in the workers at Europe’s still fashionable radioactive health spas.