This argument doesn't, by itself, constitute an explanation for why we have sexual reproduction in the first place, although it might be made the basis for such an explanation. To say, as I have, that good genes can benefit from the existence of sex whereas bad genes can benefit from its absence, is not the same thing as explaining why sex is there at all. There are many theories of why sex exists, and none of them is knock-down convincing. One of the earliest to be suggested, ‘Muller's Ratchet’, is a more disciplined version of the theory that I've informally expressed in the form of ‘wishes’ of good genes and bad genes. My discussion of mutator genes can be seen as adding an extra fillip to the Muller's Ratchet theory. Asexual reproduction doesn't just allow bad genes to accumulate in the population. It gives active encouragement to mutator genes. This is likely to hasten the extinction of asexual clones or, in other words, speed up the operation of Muller's Ratchet. But the whole question of sex and why it is there, Muller's Ratchet and all, is another story and a difficult one to tell. Maybe one day I'll summon up the courage to tackle it in full and write a whole book about the evolution of sex.
But that was a digression. The upshot is that, where there is sexual reproduction, the phenomenon of mutation is penalized by natural selection, even though individual mutations (a minority of them) may occasionally be favoured by natural selection. This is true even in a tone of stress where you can make a superficially plausible case for in-Creased mutation rates. The predilection to mutate is always bad, even though individual mutations occasionally turn out to be good. It is best, if more than a little paradoxical, to think of natural selection as ftrouring a mutation rate of zero. Fortunately for us, and for the continuance of evolution, this genetic nirvana is never quite attained. Natural selection, the second stage in the Darwinian process, is a ion-random force, pushing towards improvement. Mutation, the first ftage in the process, is random in the sense of not pushing towards {85} mprovement. All improvement is therefore, in the first place, lucky, which is why people mistakenly think of Darwinism as a theory of chance. But mistaken they are.
The belief that natural selection favours a mutation rate of zero and that mutation is undirected does not preclude an intriguing possibility, which I have called ‘the evolution of evolvability’, and advocated in an essay of that title. I'll explain a new version of the idea — kaleidoscopic embryology — in Chapter 7. Meanwhile, let us return to natural selection itself, the other half of the Darwinian partnership. Though mutation is allowed to be random and in one important sense almost certainly is random, the whole essence of natural selection is that it is not random. Of all the wolves that might survive, a non-random sample — the fleetest of foot, the canniest of wit, the sharpest of sense and tooth — are the ones that do survive and pass on their genes. Consequently, the genes that we see in the present are copies of a non-random sample of the genes that have existed in the past Every generation is a gene sieve. The genes that still exist after a million generations of sieving have what it takes to get through sieves. They have participated in the embryonic constructing of a million bodies without a single failure. Every one of those million bodies has survived to adulthood. Not one of them was too unattractive to find a mate — unattractive meaning whatever seems unattractive to would-be mates of the species concerned. Every single one of them proved capable of bearing or begetting at least one child. The sieve is an exacting one. The genes that penetrate to the future are not a random sample, they are an elite. They have survived ice ages and droughts, plagues and predators, busts and booms of population. They have survived shifting climates not just in the conventional sense of rain, ice and drought. They have survived shifting climates of companion genes, for the lot of a gene where there is sexual reproduction is to change partners in every generation; surviving genes are those that flourish when rubbing shoulders with successive samplings from the genes of the whole species, and this means other genes that are good at cooperating with the other genes of the species. The dominant part of the climate in which a gene has to survive is the other genes of the species: its companions in the ‘River out of Eden that flows through successive bodies down the generations. We {86} can think of different species, as they split apart in branchings of the river, as separate micro-climates in which different sets of genes have to survive.
For simplicity we speak of mutation as the first stage in the Darwinian process, natural selection as the second stage. But this is misleading if it suggests that natural selection hangs about waiting for a mutation which is then either rejected or snapped up and the waiting begins again. It could have been like that: natural selection of that kind would probably work, and maybe does work somewhere in the universe. But as a matter of fact on this planet it usually isn't like that. There is actually a large pool of variation, originally fed by a trickle of mutations but importantly stirred and agitated into greater variation by sexual reproduction. Variation comes originally from mutation but the mutations may be quite old by the time natural selection gets around to working on them.
For instance, my Oxford colleague the late Bernard Kettlewell famously studied the evolution of dark, almost black, moths in species that had hitherto been light-coloured. In the species he especially studied, Biston betularia, dark individuals tend to be slightly hardier than light ones, but in unpolluted country districts they are rare because they are conspicuous to birds and are promptly picked off. In industrial areas where tree trunks have been blackened by pollution, they are less conspicuous than the light-coloured forms and consequently less likely to be eaten. This also allows them to enjoy the additional advantage of their natural hardiness. The consequent increase in numbers of dark forms, to overwhelming numerical dominance in industrial areas since the mid-nineteenth century, has been dramatic and is one of the best attested examples of natural selection in action. And now we come to the reason for introducing the case here. It is often wrongly thought that after the Industrial Revolution natural selection worked on a single brand-new mutation. On the contrary, we can be sure that there have always been dark individuals — they just haven't lasted very long. Like most mutations, this one will have been recurrent but the dark moths were always rapidly snapped up by birds. When conditions changed after the Industrial Revolution, natural selection found a {87} ready-made minority of dark genes in the gene pool to work upon.
We have identified the ingredients that must be present before evolution can occur as being mutation and natural selection. These two will follow automatically on any planet given a more fundamental ingredient, one that is difficult, but obviously not impossible, to procure. This difficult basic ingredient is heredity. In order for natural selection to occur, anywhere in the universe, there must be lineages of things that resemble their immediate ancestors more than they resemble members of the population at large. Heredity is not the same thing as reproduction. You can have reproduction without heredity. Bush fires reproduce but without heredity.
Imagine a dry, parched grassland, stretching for mile after mile in every direction. Now, in a particular place, a careless smoker drops a lighted match and in no time at all the grass has flared up into a racing fire. Our smoker runs away from it as hard as his coughing lungs will let him, but we are more concerned with the way the fire spreads. It doesn't just swell steadily outwards from the original starting point. It also sends sparks up into the air. These sparks, or burning wisps of dry grass, are carried by the wind far away from the original fire. When a spark eventually comes down, it starts a new fire somewhere else on the tinder-dry prairie. And later, the new fire sends off sparks which kindle yet more new fires somewhere else. We could say that the fires are indulging in a form of reproduction. Each new fire has one parent fire. This is the fire that spat out the spark that started it. And it has one grandparent fire, one great-grandparent fire, and so on back until the ancestral fire started by the wayward match. A new fire can have only one mother but it can have more than one daughter, because it can send out more than one spark in
different directions. If you watched the whole process from above and were able to record the history of each flare-up, you could draw out a complete family tree of the fires on the prairie.
Now the point of the story is that although there is reproduction among the fires there is no true heredity. For there to be true heredity, each fire would have to resemble its parent more than it resembled the other fires in general. There is nothing wrong with the idea of a fire resembling its parent. It could happen. Fires do vary, do have {88} individual qualities, just as people do. A fire may have its own characteristic flame colour, its own smoke colour, flame size, noise level and so on. It could resemble its parent fire in any of these characteristics. If, in general, fires did resemble their parents in these ways, we could say that we had true heredity. But in fact fires don't resemble their parents any more than they resemble the general run of fires dotted around the prairie. An individual flare-up gets its characteristic qualities, its flame size, smoke colour, crackle volume and so on, from its surroundings; from the kind of grass that happens to be growing where the spark lands; from the dryness of the grass; from the speed and direction of the wind. These are all qualities of the local area where the spark lands. They are not qualities of the parent fire from which the spark came.
In order for there to be true heredity, each spark would have to carry with it some quality, some characteristic essence, of its parent fire. For example, suppose that some fires have yellow flames, others red flames, others blue. Now, if yellow-flamed fires give off sparks that start new yellow-flamed fires, whereas red-flamed fires give off sparks that start new red fires, and so on, we'd have true heredity. But that isn't what happens. If we see a blue flame we say, ‘There must be some copper salts in this area.’ We don't say, ‘This fire must have been started by a spark from another blue-flamed fire somewhere else.’
And this, of course, is where rabbits and humans and dandelions differ from fires. Don't be misled, incidentally, by the fact that rabbits have two parents and four grandparents whereas fires have only one parent and one grandparent. That is an important difference, but it is not the one we are talking about at the moment. If it helps, think not of rabbits but of stick insects or aphids, where females can have daughters, granddaughters and great-granddaughters without males ever being involved. The shape, colour, size and temperament of a itick insect is influenced, no doubt, by the place and climate of its upbringing. But it is influenced, too, by the spark that flies only from parent to child.
So what is it, this mysterious spark that flies from parent to off-ipring but not from fire to fire? On this planet it is DNA. The most tmazing molecule in the world. It is easy to think of DNA as the {89} information by which a body makes another body like itself. It would be more correct to see a body as the vehicle used by DNA to make more DNA like itself. All the DNA in the world at a given time, such as now, has come down through an unbroken chain of successful ancestors. No two individuals (except identical twins) have exactly the same DNA. Differences between DNA in individuals really contribute to their survival and chances of reproducing that same DNA. To repeat because it is so important, the DNA that has made it down the river of time is DNA that has, for hundreds of millions of years, inhabited the bodies of successful ancestors. Lots of would-be ancestors have died young, or failed to find a mate. But none of their DNA is still with us in the world.
It would be easy, at this point, to make the mistake of thinking that something, some elixir of success, some odour of sanctity from the good, successful, ancestral bodies ‘rubs off on the DNA as it passes through them. Nothing of the kind occurs. The river of DNA that flows through us into the future is a pure river that (mutations apart) leaves us exactly as it finds us. To be sure, it is continually mixed in sexual recombination. Half the DNA in you is from your father and the other half is from your mother. Each one of your sperms or eggs will contain a different mixture assembled from the genetic streamlet that came from your father and the genetic streamlet that came from your mother. But the point I was making remains true. Nothing about successful ancestors ‘rubs off on the genes as they pass through on their way to the distant future.
The Darwinian explanation for why living things are so good at doing what they do is very simple. They are good because of the accumulated wisdom of their ancestors. But it is not wisdom that they have learned or acquired. It is wisdom that they chanced upon by lucky random mutations, wisdom that was then selectively, non-ran-domly, recorded in the genetic database of the species. In each generation the amount of luck was not very great: small enough to be believable even by the sceptical physicists whom I quoted earlier. But, because the luck has been accumulated over so many generations, we are eventually very impressed by the apparent improbability of the end product. The whole Darwinian circus depends upon — follows {90} from — the existence of heredity. When I called heredity the basic ingredient, I meant that Darwinism, and hence life, will follow, more or less inevitably, on any planet in the universe where something equivalent to heredity arises.
We have arrived back at Mount Improbable, back to ‘smearing out’ the luck: to taking what looks like an immense amount of luck — the luck needed to make an eye where previously there was no eye, say — and explaining it by splitting it up into lots of little pieces of luck, each one added cumulatively to what has gone before. We have now seen how this actually works, by means of the accumulation of lots of little pieces of ancestral luck in the DNA that survives. Alongside die minority of genetically well-endowed individuals who survived, (here were large numbers of less favoured individuals who perished. Every generation has its Darwinian failures but every individual is de-icended only from previous generations’ successful minorities.
The message from the mountain is threefold. First is the message we have already introduced: there can be no sudden leaps upward — no precipitous increases in ordered complexity. Second, there can be no going downhill — species can't get worse as a prelude to getting better. Third, there may be more than one peak — more than one way of solving the same problem, all flourishing in the world.
Take any part of any animal or plant, and it is a sensible question to ask how that part has been formed by gradual transformation from lome other part of an earlier ancestor. Occasionally we can follow the process through successively younger fossils. A famous example is the gradual derivation of our mammalian ear bones — the three bones that relay sound (with exquisite impedance-matching, if you should happen to know the technical jargon) from the ear-drum to the inner ear. Fossil evidence clearly shows that these three bones, called the hammer, the anvil and the stirrup, are lineally descended from three Corresponding bones that, in our reptile ancestors, formed the jaw joint.
Often the fossil record is less kind to us. We have to guess at possible intermediates, sometimes with a bit of inspiration from other modern animals which may or may not be related. Elephant trunks contain no bones and do not fossilize, but we don't need fossils to {91} realize that the elephant's trunk started out as just a nose. It is now ... well, let me quote from a book that has me abashedly forcing back the tears whenever I read it: Battle for the Elephants, by a couple of heroes, Iain and Oria Douglas-Hamilton. They wrote alternate chapters and here is Oria's horrified description, on page 220, of an elephant ‘cull’ that she witnessed in Zimbabwe:
I looked at one of the discarded trunks and wondered how many millions of years it must have taken to create such a miracle of evolution. Equipped with fifty thousand muscles and controlled by a brain to match such complexity, it can wrench and push with tonnes of force. Yet, at the same time, it is capable of performing the most delicate operations such as plucking a small seed-pod to pop in the mouth. This versatile organ is a siphon capable of holding four litres of water to be drunk or sprayed over the body, as an extended finger and as a trumpet or loud speaker.
The trunk has social functions, too; caresses, sexual advances, reassurances, greetings and mutually intertwining hugs; and
among males it can become a weapon for beating and grappling like wrestlers when tusks clash and each bull seeks to dominate in play or in earnest. And yet there it lay, amputated like so many elephant trunks I had seen all over Africa.
The paragraph has had the usual effect on me ...
Here, the message from the mountain is that the ancestors of elephants must have included a continuous series of intermediates with more or less longish noses like tapirs, or elephant shrews, or proboscis monkeys, or elephant seals. None of these creatures is closely related to elephants (or to each other). All have evolved their long noses independently of each other and probably for different reasons (Figure 3.1).
In the evolution of the elephant from its short-nosed ancestors, there must have been a smooth, gradual succession of steadily longer noses, a sliding gradient of thickening muscles and more intricately dissected nerves. It must have been the case that, as each extra inch was added to the length of the average trunk, the trunk became better at its job. It must never be possible to say anything like: {92}