No More Learning

This will amount to equal numbers of males and females, except in those cases where offspring of one sex cost more to rear than offspring of the other. If, for example, it costs twice as much food to rear a son than a daughter (perhaps to make sons big enough to compete effectively with rival males) the stable sex ratio will be twice as many females as males. This is because the strategic alternative to one son is not one daughter but two. Fisher's powerful logic has been extended and refined in various ways, for
44 45 example by W. D. Hamilton and E. L. Charnov .
Once again, and notwithstanding the quotation above from the Second Edition of The Descent of Man, Darwin himself, in the First Edition, came remarkably close to anticipating Fisher, although without the economic language of parental expenditure:
Let us now take the case of a species producing, from the unknown causes just alluded to, an excess of one sex - we will say of males - these being superfluous and useless, or nearly useless. Could the sexes be equalized through natural selection? We may feel sure, from all characters being variable, that certain pairs would produce a somewhat less excess of males over females than other pairs. The former, supposing the actual number of the off- spring to remain constant, would necessarily produce more females, and would therefore be more productive. On the doctrine of chances a greater number of the offspring of the more productive pairs would survive; and these would inherit a tendency to procreate fewer males and more females. Thus a tendency toward equalization of the sexes would be brought about.
Sadly, Darwin deleted this remarkable passage when he came to prepare the Second Edition, preferring the more cautious paragraph later to be quoted by Fisher. Darwin's partial anticipation of Fisher in the First Edition of Descent is all the more impressive because, as Alan Grafen points out to me, Fisher's argument depends crucially on a fact which was not available to Darwin, namely that the two parents make an equal genetic contribution to every offspring. Indeed, in historical times, different schools of thought (the spermists and the ovists respectively)
70
? had held that the male, or the female, sex had a monopoly on heredity. The whole question of Fisher's sources for the sex ratio theory has
been meticulously sleuthed by Professor A. W. F. Edwards of Cambridge
46
University, himself one of Fisher's most distinguished pupils. Edwards
not only notes Darwin's priority over the essential argument and the odd fact that he deleted it from the Second Edition. He also shows how Darwin's argument was taken up and developed by a series of other workers whose writings were probably known to Fisher. First Carl Diising of Jena, in 1884, reiterated and clarified Darwin's argument. Next, in 1908 the Italian statistician Corrado Gini discussed the argument more critically. Finally in 1914, the eugenicist J. A. Cobb gave a form of the argument which seems to have all the refinements of Fisher's own of 1930, including the economic idea of parental expenditure. Cobb seems to have been unaware of Darwin's priority, but Edwards is persuasive that Fisher was aware of Cobb's. Edwards remarks that:
commentators have assumed, and most have firmly stated, that the argument was original to Fisher, though he did not claim it to be, nor did he refer to it either before or after 1930 in any of his other publications. Indeed, there is no evidence that he saw it as particularly novel, remarkable, or likely to lead to major developments in evolutionary biology . . . he may well have regarded it as public property by 1930.
Edwards himself is one of those (I am another) who once overlooked the crucial difference between the First and Second Editions of The Descent.
Fisher's economic view of sex was developed further by Robert L. Trivers, writing in a volume published to commemorate the centenary of The Descent of Man. " Trivers's subtle application of the theory of parental investment (his name for what Fisher had called parental expenditure) to male and female roles in sexual selection greatly illuminates the facts collected by Darwin in the middle chapters of Descent. Trivers defines parental investment (PI) as (what economists would call) an opportunity cost. The cost to a parent of investing in a particular child is measured in correspondingly lost opportunity to invest in others, present or future. Sexual inequality is fundamentally economic. The mother typically invests more in any individual offspring than the father does, and this inequality has far-reaching consequences, which reach even further in a kind of self-feeding process. A member of the low-investing sex (usually male) who persuades a member of the high-investing sex (usually female) to mate with him has gained an economic prize worth fighting (or otherwise competing) for. This is why males typically devote more effort to competing with other males, while females typically shunt their
LIGHT WILL BE THROWN
71
? LIGHT WILL BE THROWN
effort away from competing with other females and into investing in offspring. It is why, when one sex is more brightly coloured than the other, it is typically the male. It is why, when one sex is more choosy in selecting a mate, it is typically the female. And it is why variance in reproductive success is typically higher among males than among females: the most successful male may have many times more descendants than the least successful male, where the most successful female is only somewhat more successful than the least successful female. The Fisher/ Trivers economic inequalities between the sexes should be kept in mind while reading Darwin's enthralling review of sexual selection through the animal kingdom. It is a most striking example of a single idea uniting and explaining, at one blow, a multitude of seemingly disparate facts.
Now, to the descent of man itself. Darwin's guess that our species arose in Africa was typically ahead of its time, amply confirmed today by numerous fossils, none of which was available to him. We are African apes, closer cousins to chimpanzees and gorillas than they are to orang utans and gibbons, let alone monkeys. Darwin's 'quadrumana' were denned so as to exclude humans: they were all the apes and monkeys, with a hand bearing an opposable digit on the hindlegs as well as the forelegs. The early chapters of his book are concerned to narrow the perceived gap between ourselves and the quadrumana, a gap which Darwin's target audience would have seen as yawning between the top rung of a ladder and the next rung down. Today we would not (or should not) see a ladder at all. Instead, we should hold in our minds the branching tree diagram which is the only illustration in The Origin ofSpecies. Humanity is just one little twig, nestling among many others somewhere in the middle of a thicket of African apes.
Two vital techniques which were unavailable to Darwin are radio- active dating of rocks, and molecular evidence including the 'molecular clock'. Where Darwin, in his quest to demonstrate the similarity between ourselves and the quadrumana, could point to comparative anatomy supplemented by charming anecdotes of psychological and emotional resemblance (arguments extended in The Expression of the Emotions), we are privileged to know the exact letter-by-letter sequence of massive DNA texts. It is claimed that more than 98 per cent of the human genome, when measured in this way, is identical with chimpanzees'. Darwin would have been spellbound. Such closeness of resemblance, and such precision in measuring it, would have delighted him beyond dreams.
Nevertheless, we must beware of being carried away by the euphoria of it all. That 98 per cent doesn't mean we are 98 per cent chimpanzees. And
72
? it really matters which unit you choose to make your comparison. If you count the number of whole genes that are identical, the figure for humans and chimpanzees would be close to zero. This is not a paradox. Think of the human genome and the chimpanzee genome as two editions of the same book, say the first and second editions of The Descent ofMan. If you count the number of letters that are identical to their opposite numbers in the other edition, it is probably well over 90 per cent. But if you count the number of chapters that are identical, it may well be zero. This is because it takes only one letter to be different, anywhere in a chapter, for the whole chapter to be judged different between the two editions. When you are measuring the percentage similarity between two texts, whether two editions of a book or two editions of an African ape, the unit of comparison you choose (letter or chapter, DNA base pair or gene) makes
a huge difference to the final percentage similarity.
The point is that we should use such percentages not for their
absolute value but in comparisons between animals. The 98 per cent figure for humans and chimpanzees starts to make sense when we compare it with the 96 per cent resemblance between humans and orang utans (it is the same 96 per cent between chimpanzees and orang utans, and the same between gorillas and orang utans, because all the African apes are connected to the Asian orang utans via a shared African ancestor). For the same kind of reason, all the great apes share 95 per cent of their genomes with the gibbons and siamangs. And all the apes share 92 per cent of their genomes with all Old World monkeys.
The hypothesis of a molecular clock allows us to use such percentage figures to put a date on each of the splits in our family tree. It assumes that evolutionary change, at the molecular genetic level, proceeds at an approximately fixed rate for each gene. This is in accordance with the widely accepted neutral theory of the Japanese geneticist Motoo Kimura. Kimura's neutral theory is sometimes seen as anti-Darwinian but it is not. It is neutral with respect to Darwinian selection. A neutral mutation is one that makes no difference to the functioning of the protein produced. The post-mutation version is no better and no worse than the pre-mutation version, where both may be vital to the life of the organism.
From a Darwinian point of view, neutral mutations are not mutations at all. But from a molecular point of view they are extremely useful mutations because their fixed rate makes the clock reliable. The only point of controversy introduced by Kimura is how many mutations are neutral. Kimura thought it was the great majority which, if true, is very nice for the molecular clock. Darwinian selection remains the only explanation for adaptive evolution and it is arguable (I would argue)
LIGHT WILL BE THROWN
73
? LIGHT WILL BE THROWN
that most if not all of the evolutionary changes we actually see in the macroscopic world (as opposed to those concealed among the molecules) are adaptive and Darwinian.
As so far described, the molecular clock gives relative timings but not absolute ones. We can read off timings of evolutionary splits, but only in arbitrary units. Fortunately, in another great advance that would have entranced Darwin, various absolute clocks are available for dating fossils. These include the known rates of radioactive decay of isotopes in volcanic rocks sandwiching the sedimentary strata in which fossils are found. By taking a group of animals with a rich fossil record and dating the splits in their family tree two ways - by the molecular genetic clock and by radioactive clocks - the arbitrary units of the genetic clock can be validated, and simultaneously calibrated in real millions of years. This is how we can estimate that the split between humans and chimpanzees occurred between 5 and 8 million years ago, the split between African apes and orang utans about 14 million years ago, and the split between apes and Old World monkeys about 25 million years ago.
Fossils, all discovered after Descent was published, provide us with a sporadic picture of some possible intermediates connecting us to our common ancestor with chimpanzees. Unfortunately, there are no fossils connecting modern chimpanzees to that shared ancestor, but on our side of the split reports of new fossil finds are coming in at a rate which
I find exciting and surely Darwin would have too. Going back in steps
of roughly one million years we find: Homo erectus, Homo habilis, Australopithecus afarensis, Australopithecus anamensis, Ardipithecus, Orrorin and, a recent discovery which may date from as long ago as 7 million years, Sahelanthropus. That last find is from Chad, far to the west of the great Rift Valley which had hitherto been thought to constitute
a geographic barrier dividing our lineage from that of the chimpanzees. It is good for our orthodoxies to be upset from time to time.
We must beware of assuming that this temporal series of fossils represents an ancestor/descendant series. It is always safer to assume that fossils are cousins rather than ancestors, but we need not be shy of guessing that earlier cousins may tell us at least something about the true ancestors among their contemporaries.
What are the main changes that occurred since our split from the chimpanzees? Some, such as our loss of body hair, are interesting, but fossils can tell us nothing directly about them. The two main changes that fossils can help us with, and where we therefore have a big advantage over Darwin, are that we rose up on our hind legs, and that our brains got rather dramatically bigger. Which of these changes came
74
? first, or did they happen together? All three views have been supported,
and controversy has gone back and forth over the decades. Darwin
thought the two big changes happened in concert, and he makes out a
plausible case. But this is a rare instance where Darwin's tentative guess
has turned out wrong. The fossils give a satisfyingly decisive and clear
48
answer. Bipedality came first, and its evolution was more or less
complete before the brain started to swell. Three million years ago, Australopithecus was bipedal and had feet like ours, although it probably still retreated up trees. But its brain, relative to its body size, was the same size as a chimpanzee's, and presumably the same as the shared ancestor with chimpanzees. Nobody knows whether the bipedal gait set up new selection pressures that encouraged the brain to grow, but Darwin's original arguments for simultaneous evolution can be adapted to make that plausible. Perhaps the enlargement of the brain had some- thing to do with language, but here nobody knows and disagreements abound. There is evidence that particular parts of the human brain are uniquely pre-wired to handle specific universals of language, although
49 the particular language spoken is, of course, locally learned.
Another twentieth-century idea which is probably important in human evolution, and which again would have intrigued Darwin, is neoteny: evo- lutionary infantilization. The axolotl, an amphibian living in a Mexican lake, looks just like the larva of a salamander, but it can reproduce, and has chopped off the adult, salamander stage of the life history. It is a sexually mature tadpole. Such neoteny has been suggested as a way in which a lineage can suddenly initiate an entirely new direction of evolution, at a stroke. Apes don't have a discrete larval stage like a tadpole or a caterpillar, but a more gradualistic version of neoteny can be discerned in human evolution. Juvenile chimpanzees resemble humans far more than adult chimpanzees do. Human evolution can be seen as infantilism. We are apes
50 that became sexually mature while still morphologically juvenile. If
humans could live for 200 years, would we finally 'grow up', drop on all fours and develop huge prognathous chimpanzee-like jaws? The possibility has not been lost on writers of ironic fiction, notably Aldous Huxley in After Many a Summer. He presumably learned about neoteny from his elder brother Julian, who was one of the pioneers of the idea and did amazing research on axolotls, injecting hormones to make them turn into salamanders never before seen.
Let me end by bringing together once again the two halves of Darwin's book. He went to town on sexual selection in The Descent of Man because he thought it was important in human evolution, and especially because he thought it was the key to understanding the
LIGHT WILL BE THROWN
75
? LIGHT WILL BE THROWN
differences among human races. Race, in Victorian times, was not the political and emotional minefield it is today, when one can give offence j by so much as mentioning the word. I shall tread carefully, but I cannot I ignore the topic because it is prominent in Darwin's book and especially germane to the unification of its two parts.
Darwin, like all Victorians, was intensely aware of the differences among humans but he also, more than most of his contemporaries, emphasized the fundamental unity of our species. In Descent he carefully considered, and decisively rejected, the idea, rather favoured in his own time, that different human races should be regarded as separate species. Today we know that, at the genetic level, our species is more than usually uniform. It has been said that there is more genetic variation among the chimpanzees of a small region of Africa than among the entire world population of humans (suggesting that we have been through a bottleneck in the past hundred thousand years or so). Moreover, the great majority of human genetic variation is to be found within races, not between them. This means that if you were to wipe out all human races except one, the I great majority of human genetic variance would be preserved. The variance between races is just a bit extra, stuck on the top of the greater quantity of variation within all races. It is for this reason that many geneticists advocate the complete abandonment of the concept of race.
At the same time - the paradox is similar to one recognized by Darwin - the superficially conspicuous features characteristic of local populations around the world seem very different. A Martian taxonomist who didn't know that all human races happily interbreed with one another, and didn't know that most of the underlying genetic variance in our species is shared by all races, might be tempted by our regional differences in skin colour, facial features, hair, body size and proportions to split us into more than one species. What is the resolution of the paradox? And why did such pronounced superficial differences evolve in different geographical areas, while most of the less conspicuous variation is dotted around across all geographical areas? Could Darwin have been right all along? Is sexual selection the answer to the paradox? The distinguished
51
biologist Jared Diamond thinks so, and I am inclined to agree.
Utilitarian answers have been suggested to the question of the evolution of racial differences, and there may well be some truth in them. Dark skin may protect against skin cancer in the tropics, light skin admit beneficial rays in sun-starved latitudes where there is a danger of Vitamin D deficiency. Small stature probably is of benefit to hunters in dense forest, such as the pygmies of central Africa, and various independently evolved hunter gatherers of Amazon and South East Asian
76
1
? forests. The ability to digest milk when adult seems to have evolved in peoples who, for cultural reasons, prolong the use of this primitively juvenile food. But I am impressed by the diversity of features that are superficial and conspicuous, while deeper differences are so slight.
What sexual selection explains, better than natural selection, is diversity that seems arbitrary, even driven by aesthetic whim. Especially if the variation concerned is geographical. And also especially if some of the features concerned, for example beards and the distribution of body hair and subcutaneous fat deposits, differ between the sexes. Most people have no problem in accepting an analogue of sexual selection for culturally mediated fashions like headdresses, body paint, penis sheaths, ritual mutilations or ornamental clothes. Given that cultural differences such as those of language, religion, manners and customs certainly provide resistance to interbreeding and gene flow, I think it is entirely plausible that genetic differences between peoples of different regions, at least where superficial, externally prominent features are concerned, have evolved through sexual selection. Our species really does seem to have unusually conspicuous, even ostentatious, superficial differences between local populations, coupled with unusually low levels of overall genetic variation. This double circumstance carries, to my mind, the stamp of sexual selection.
52
In this respect, human races seem a lot like breeds of dog, another
favourite topic of Darwin. Superficially, the domestic breeds of dogs are astonishingly varied, even more so than human races, yet the under- lying genetic differences are slight, and they are all clearly descended
53
from wolves within the past few thousand years. Reproductive
isolation is today maintained by disciplined pedigree breeders, and the shapes and colours of the dogs themselves are steered through their rapid evolution by the whim of the human eye rather than the whim of female dogs. But the essential features of the situation, as Darwin realized, are similar to those of sexual selection.
In this, as in so much else, I suspect that Darwin was right. Sexual
selection really is a good candidate for explaining a great deal about the
unique evolution of our species. It may also be responsible for some
unique features of our species which are shared equally by all races, for
54
example our enormous brain. Geoffrey Miller, in The Mating Mind, strongly developed precisely this case, and Darwin would have loved it
no less because Miller takes a Wallacean view of sexual selection. It is starting to look as though, despite initial appearances, Darwin really was right to bring together, in one volume, Selection in Relation to Sex and The Descent ofMan.
LIGHT WILL BE THROWN
77
has
? ? En ? AMI
Darwin Triumphant55 Darwinism as a Universal Truth
If we are visited by superior creatures from another star system - they will have to be superior if they are to get here at all - what common ground shall we find for discussion with them? Shall we overcome the barriers simply by learning one another's language, or will the subjects that interest our two cultures be so divergent as to preclude serious conversation? It seems unlikely that the star travellers will want to talk about many of our intellectual stocks-in-trade: about literary criticism or music, religion or politics. Shakespeare may mean nothing to those without human experiences and human emotions, and if they have a literature or an art these will probably be too alien to excite our sensi- bilities. To name two thinkers who have more than once been promoted as Darwin's equals, I rather doubt whether our visitors will have much interest in talking about Marx or Freud, other than perhaps as anthropological curiosities. We have no reason to suppose that these men's works are of more than local, parochial, human, earthly, post- Pleistocene (some would add European and male) significance.
Mathematics and physics are another matter. Our guests may find our level of sophistication quaintly low, but there will be common ground. We shall agree that certain questions about the universe are important, and we shall almost certainly agree on the answers to many of these questions. Conversation will flourish, even if most of the questions flow one way and most of the answers the other. If we discuss the histories of our respective cultures, our visitors will surely point with pride, however far back in time, to their equivalents of Einstein and Newton, of Planck and Heisenberg. But they won't point to an equivalent of Freud or Marx any more than we, visiting a hitherto undiscovered tribe in a remote forest clearing, would nominate our civilization's equiva- lent of the local rainmaker or gully-gully man. One does not have to disparage the local achievements of Freud and Marx on this planet to agree that their findings have no universality.
78
? What about Darwin? Will our guests revere another Darwin as one of their greatest thinkers of all time? Shall we be able to have a serious conversation with them about evolution? I suggest that the answer is yes (unless, as a colleague suggests to me, their Darwin is on the expedi- tion and we are her Galapagos*). Darwin's achievement, like Einstein's, is universal and timeless, whereas that of Marx is parochial and ephemeral. That Darwin's question is universal, wherever there is life, is surely undeniable. The feature of living matter that most demands explanation is that it is almost unimaginably complicated in directions that convey a powerful illusion of deliberate design. Darwin's question, or rather the most fundamental and important of Darwin's many questions, is the question of how such complicated 'design' could come into being. All living creatures, everywhere in the universe and at any time in history, provoke this question. It is less obvious that Darwin's answer to the riddle - cumulative evolution by nonrandom survival of random hereditary changes - is universal.
It is at first sight conceivable that Darwin's answer might be valid only parochially, only for the kind of life that happens to exist in our own little clearing in the universal
56
forest. I have previously made the case that this is not so, that the
general form of Darwin's answer is not merely incidentally true of our kind of life but almost certainly true of all life, everywhere in the universe. Here, let me for the moment make the more modest claim that, at the very least, Darwin's bid for immortality is closer to the Einstein end of the spectrum than to the Marx end. Darwinism really matters in the universe.
When I was an undergraduate in the early nineteen sixties, we
were taught that although Darwin was an important figure in his own
time, modern neo-Darwinism was so much further advanced that it
hardly deserved the name Darwinism at all. My father's generation of
biologist undergraduates read, in an authoritative Short History of 57
. . . the struggle of living forms leading to natural selection by the survival of the fittest, is certainly far less emphasized by naturalists now than in the years that immediately followed the appearance of Darwin's book. At the time, however, it was an extremely stimulating suggestion.
And the generation of biologists before that could read, in the words of William Bateson, perhaps the dominant British geneticist of the time,
This is how my friend worded her suggestion. The joke was rather ruined by the political scruples of the original article's copy-editor, who changed 'her Galapagos' to 'his or her Galapagos'.
Biology , that
DARWIN TRIUMPHANT
79
? LIGHT WILL BE THROWN
We go to Darwin for his incomparable collection of facts [but] . . . for us he speaks no more with philosophical authority. We read his scheme of Evolution as we would those of Lucretius or Lamarck . . . The transformation of masses of populations by imperceptible steps guided by selection is, as most of us now see, so inapplicable to the fact that we can only marvel . . . at the want of
58 penetration displayed by the advocates of such a proposition.
And yet the editors of this volume can commission an article with the title 'Darwin Triumphant'. I do not normally like writing to titles that others have proposed, but I can accept this one without reservation. In the last quarter of the twentieth century, it seems to me that Darwin's standing among serious biologists (as opposed to nonbiologists influenced by religious preconceptions) is rightly as high as it has been at any time since his death. A similar story, of even more extreme eclipse in earlier years followed by triumphant recent rehabilitation, can be told of Darwin's 'other theory', that of sexual selection. *
It is only to be expected that, a century and a quarter on, the version
of his theory that we now have should be different from the original. Modern Darwinism is Darwinism plus Weismannism plus Fisherism plus Hamiltonism (arguably plus Kimuraism and a few other isms). But when I read Darwin himself, I am continually astonished at how modern
he sounds. Considering how utterly wrong he was on the all-important topic of genetics, he showed an uncanny gift for getting almost every- thing else right. Maybe we are neo-Darwinists today, but let us spell the neo with a very small n\ Our neo-Darwinism is very much in the spirit of Darwin himself. The changes that Darwin would see if he came back today are in most cases changes that, I venture to suggest, he would instantly approve and welcome as the elegant and obviously correct answers to riddles that troubled him in his own time. Upon learning that evolution is change in frequencies within a pool of particulate hereditary elements, he might even quote T. H. Huxley's alleged remark upon reading the Origin itself: 'How extremely stupid not to have thought of thatl't
*See 'Light Will Be Thrown' (pp. 63-77).
tOf the two stories about Huxley that have become chestnuts, I greatly prefer this to the one about his so-called 'debate' with the Bishop of Oxford, Sam Wilberforce. There is something admirably honest about Huxley's exasperation at not having thought of such a simple idea. I have long found it a complete mystery why it had to wait until the nineteenth century before anyone thought of it. Archimedes' and Newton's achievements seem, on the face of it, far more difficult. But the fact that nobody did think of natural selection before the nineteenth century clearly shows that I am wrong. As does the fact that so many people, even today, don't get it.
80
? I referred to Darwin's gift for getting things right, but surely this can only mean right as we see it today. Shouldn't we be humble enough to admit that our right may be utterly wrong in the sight of future scientific generations? No, there are occasions when a generation's humility can be misplaced, not to say pedantic. We can now assert with confidence that the theory that the Earth moves round the Sun not only is right in our time but will be right in all future times even if flat- Earthism happens to become revived and universally accepted in some new dark age of human history. We cannot quite say that Darwinism is in the same unassailable class. Respectable opposition to it can still be mounted, and it can be seriously argued that the current high standing of Darwinism in educated minds may not last through all future generations. Darwin may be triumphant at the end of the twentieth century, but we must acknowledge the possibility that new facts may come to light which will force our successors of the twenty-first century to abandon Darwinism or modify it beyond recognition. But is there, perhaps, an essential core of Darwinism, a core that Darwin himself might have nominated as the irreducible heart of his theory, which we might set up as a candidate for discussion as potentially beyond the reach of factual refutation?
Core Darwinism, I shall suggest, is the minimal theory that evolution is guided in adaptively nonrandom directions by the nonrandom survival of small random hereditary changes. Note especially the words small and adaptively. Small implies that adaptive evolution is gradualistic, and we shall see why this must be so in a moment. Adaptive does not imply that all evolution is adaptive, only that core Darwinism's concern is limited to the part of evolution that is. There is no reason to assume
59
that all evolutionary change is adaptive.
change is not adaptive, what is undeniable is that enough of evolution- ary change is adaptive to demand some kind of special explanation. It is the part of evolutionary change that is adaptive that Darwin so neatly explained. There could be any number of theories to explain non- adaptive evolution. Nonadaptive evolution may or may not be a real phenomenon on any particular planet (it probably is on ours, in the form of the large-scale incorporation of neutral mutations), but it is not a phenomenon that awakes in us an avid hunger for an explanation. Adaptations, especially complex adaptations, awake such a powerful hunger that they have traditionally provided one of the main motiva- tions for belief in a supernatural Creator. The problem of adaptation, therefore, really was a big problem, a problem worthy of the big solution that Darwin provided.
DARWIN TRIUMPHANT
But even if most evolutionary
81
? LIGHT WILL BE THROWN
60
R. A. Fisher developed a case, which did not make any appeal to
particular facts, for the armchair deducibility of Mendelism.
It is a remarkable fact that had any thinker in the middle of the nineteenth century undertaken, as a piece of abstract and theoretical analysis, the task of constructing a particulate theory of inheritance, he would have been led, on the basis of a few very simple assumptions, to produce a system identical with the modern scheme of Mendelian or factorial inheritance.
Is there a similar statement that could be made about the inevitability of the core of Darwin's scheme of evolution by natural selection? Although Darwin and Wallace themselves were field naturalists who made extensive use of factual information to support their theory, can we now, with hindsight, argue that there should have been no need for the Beagle, no need for the Galapagos and Malay Archipelagos? Should any thinker, faced with the problem formulated in the right way, have been able to arrive at the solution - core Darwinism - without stirring from an armchair?
Part of core Darwinism arises almost automatically from the problem that it solves, if we express that problem in a particular way, as one of mathematical search. The problem is that of finding, in a gigantic mathe- matical space of all possible organisms, that tiny minority of organisms that is adapted to survive and reproduce in available environments. Again, Fisher put it with characteristically powerful clarity.
An organism is regarded as adapted to a particular situation, or to the totality of situations which constitute its environment, only in so far as we can imagine an assemblage of slightly different situations, or environments, to which the animal would on the whole be less well adapted; and equally only in so far as we can imagine an assemblage of slightly different organic forms, which would be less well adapted to that environment.
Imagine some nightmarish mathematical menagerie in which is found the all but infinitely large set of conceivable animal forms that could be cobbled together by randomly varying all the genes in all genomes in all possible combinations. For brevity, although it is not as precise a phrase as its mathematical tone leads one to think, I shall refer to this as the set of all possible animals (fortunately the argument I am developing is an order-of-magnitude argument which does not depend on numerical precision). Most of the members of this ill-favoured bestiary will never develop beyond the single-cell stage. Of the very few that manage to be born (or hatch, etc. ), most will be hideously mis- shapen monstrosities who will die early. The animals that actually exist,
82
? or have ever existed, will be a tiny subset of the set of all possible animals. Incidentally, I use animal purely for convenience. By all means substitute plant or organism.
It is convenient to imagine the set of all possible animals as arrayed in a multidimensional genetic landscape. * Distance in this landscape means genetic distance, the number of genetic changes that would have to be made in order to transform one animal into another. It is not obvious how one would actually compute the genetic distance between any two animals (because not all animals have the same number of genetic loci); but again the argument does not rely upon precision, and it is intuitively obvious what it means, for instance, to say that the genetic distance between a rat and a hedgehog is larger than the genetic distance between a rat and a mouse. All that we are doing here is to place as well, in the same multidimensional system of axes, the very much larger set of animals that have never existed. We are including those that could never have survived even if they had come into existence, as well as those that might have survived if they had existed but as a matter of fact never came into existence.
Movement from one point in the landscape to another is mutation, interpreted in its broadest sense to include large-scale changes in the genetic system as well as point mutations at loci within existing genetic systems. In principle, by a sufficiently contrived piece of genetic engineering - artificial mutation - it is possible to move from any point in the landscape to any other. There exists a recipe for transforming the genome of a human into the genome of a hippo or into the genome of any other animal, actual or conceivable. It would normally be a very large recipe, involving changes to many of the genes, deletion of many genes, duplication of many genes, and radical reorganizations of the genetic system. Nevertheless, the recipe is in principle discoverable, and obeying it can be represented as equivalent to taking a single giant leap from one point to another in our mathematical space. In practice, viable mutations are normally relatively small steps in the landscape: children are only slightly different from their parents even if, in principle, they could be as different as a hippo is from a human. Evolution consists of step-by-step trajectories through the genetic space, not large leaps.
*I find this image, which is modified from the venerable American population geneticist Sewall Wright, a helpful way to think about evolution. I first made use of it in The Blind Watchmaker and gave it two chapters in Climbing Mount Improbable, where I called it a 'museum' of all possible animals. Museum is superficially better than landscape because it is three- dimensional, although actually, of course, we are usually dealing with many more than three dimensions. Daniel Dennett's version, in Darwin's Dangerous Idea, is a library, the vividly named 'Library of Mendel'.
DARWIN TRIUMPHANT
83
? LIGHT WILL BE THROWN
Evolution, in other words, is gradualistic. There is a general reason why this has to be so, a reason that I shall now develop.
Even without formal mathematical treatment, we can make some statistical statements about our landscape. First, in the landscape of all possible genetic combinations and the 'organisms' that they might generate, the proportion of viable organisms to nonviable organisms is very small. 'However many ways there may be of being alive, it is certain
61
that there are vastly more ways of being dead. ' Second, taking any
given starting point in the landscape, however many ways there may be of being slightly different, it is obvious that there are vastly more ways of being very different. The number of near neighbours in the landscape may be large, but it is dwarfed by the number of distant neighbours. As we consider hyperspheres of ever increasing size, the number of progres- sively more distant genetic neighbours that the spheres envelop mounts as a power function and rapidly becomes for practical purposes infinite.
The statistical nature of this argument points up an irony in the claim, frequently made by lay opponents of evolution, that the theory of evolution violates the Second Law of thermodynamics, the law of increasing entropy or chaos* within any closed system. The truth is opposite. If anything appeared to violate the law (nothing really does), it would be the factst, not any particular explanation of those facts! The Darwinian explanation, indeed, is the only viable explanation we have for those facts that shows us how they could have come into being without violating the laws of physics. The law of increasing entropy is, in any case, subject to an interesting misunderstanding, which is worthy of a brief digression because it has helped to foster the mistaken claim that the idea of evolution violates the law.
62
The Second Law originated in the theory of heat engines, but the
form of it that is relevant to the evolutionary argument can be stated in more general statistical terms. Entropy was characterized by the physicist Willard Gibbs as the 'mixed-upness' of a system. The law states that the total entropy of a system and its surroundings will not decrease. Left to itself, without work being contributed from outside, any closed system (life is not a closed system) will tend to become more mixed-up, less orderly. Homely analogies - or they may be more than analogies - abound. If there is not constant work being put in by a librarian, the orderly shelving of books in a library will suffer relentless degradation due to the inevitable if low probability that borrowers will return them
"Chaos here has its original and still colloquial meaning, not the technical meaning which it has recently acquired.
tAbout life's functional complexity or high 'information content'.
84
? to the wrong shelf. We have to import a hard-working librarian into the system from outside, who, Maxwell's-Demon-like, methodically and energetically restores order to the shelves.
The common error to which I referred is to personify the Second Law: to invest the universe with an inner urge or drive towards chaos; a positive striving towards an ultimate nirvana of perfect disorder. It is partly this error that has led people to accept the foolish notion that evolution is a mysterious exception to the law. The error can most simply be exposed by reference to the library analogy. When we say that an unattended library tends to approach chaos as time proceeds, we do not mean that any particular state of the shelves is being approached, as though the library were striving towards a goal from afar. Quite the contrary. The number of possible ways of shelving the N books in a library can be calculated, and for any nontrivial library it is a very, very large number indeed. Of these ways, only one, or a very few, would be recognized by us as a state of order. That is all there is to it. Far from there being any mystical urge towards disorder, it is just that there are vastly more ways of being recognized as disorderly than of being recognized as orderly. So, if a system wanders anywhere in the space of all possible arrangements, it is almost certain - unless special, librarian-like steps are taken - that we shall perceive the change as an increase in disorder. In the present context of evolutionary biology, the particular kind of order that is relevant is adaptation, the state of being equipped to survive and reproduce.
Returning to the general argument in favour of gradualism, to find viable life forms in the space of all possible forms is like searching for a modest number of needles in an extremely large haystack. The chance of happening to land on one of the needles if we take a large random mutational leap to another place in our multidimensional haystack is very small indeed. But one thing we can say is that the starting point of any mutational leap has to be a viable organism - one of the rare and precious needles in the haystack. This is because only organisms good enough to survive to reproductive age can have offspring of any kind, including mutant offspring. Finding a viable body-form by random mutation may be like finding a needle in a haystack, but given that you have already found one viable body-form, it is certain that you can hugely increase your chances of finding another viable one if you search in the immediate neighbourhood rather than more distantly.
The same goes for finding an improved body-form. As we consider mutational leaps of decreasing magnitude, the absolute number of destinations decreases but the proportion of destinations that are
DARWIN TRIUMPHANT
85
? LIGHT WILL BE THROWN
improvements increases. Fisher gave an elegantly simple argument to ] show that this increase tends towards 50 per cent for mutational changes of very small magnitude. * His argument seems inescapable for any single dimension of variation considered on its own. Whether his precise conclusion (50 per cent) generalizes to the multidimensional case I shall not discuss, but the direction of the argument is surely indisputable. The larger the leap through genetic space, the lower is the probability that the resulting change will be viable, let alone an improve- ment. Gradualistic, step-by-step walking in the immediate vicinity of already discovered needles in the haystack seems to be the only way to find other and better needles. Adaptive evolution must in general be a crawl through genetic space, not a series of leaps.
But are there any special occasions when macromutations areI incorporated into evolution? Macromutations certainly occur in the laboratory,t Our theoretical considerations say only that viable macromutations should be exceedingly rare in comparison with viable micromutations. But even if the occasions when major saltations are viable and incorporated into evolution are exceedingly rare, even if they have occurred only once or twice in the whole history of a lineage from Precambrian to present, that is enough to transform the entire course of evolution. I find it plausible, for instance, that the invention of segmentation occurred in a single macromutational leap, once during the history of our own vertebrate ancestors and again once in the ancestry of arthropods and annelids. Once this had happened, in each of these two lineages, it changed the entire climate in which ordinary cumulative selection of micromutations went on. It must have resembled, indeed, a sudden catastrophic change in the external climate. Just as a lineage can, after appalling loss of life, recover and adapt to a catastro- phic change in the external climate, so a lineage might, by subsequent micromutational selection, adapt to the catastrophe of a macromutation as large as the first segmentation.
In the landscape of all possible animals, our segmentation example might look like this. A wild macromutational leap from a perfectly viable parent lands in a remote part of the haystack, far from any needle of viability. The first segmented animal is born: a freak; a monster none of whose detailed bodily features equip it to survive its new, segmented
*He used the analogy of perfecting the focus of a microscope. A very small movement of the objective lens has a 50 per cent chance of being in the right direction (which will improve the focus). A large movement is bound to make things worse (even if it was in the right direction, it will overshoot).
tMacromutations, or saltations, are mutations of large magnitude. A famous example in fruit flies is antennapedia. Mutant flies grow a leg where an antenna should be.
86
;
? architecture. It should die. But by chance the leap in genetic space has coincided with a leap in geographical space. The segmented monster finds itself in a virgin part of the world where the living is easy and competition is light. What can happen when any ordinary animal finds itself in a strange place, a new continent, say, is that, although ill- adapted to the new conditions, it survives by the skin of its teeth. In the competition vacuum, its descendants survive for enough generations to adapt, by normal, cumulative natural selection of micromutations, to the alien conditions. So it might have been with our segmented monster. It survived by the skin of its teeth, and its descendants adapted, by ordinary micromutational cumulative selection, to the radically new conditions imposed by the macromutation. Though the macromutational leap landed far from any needle in the haystack, the competition vacuum enabled the monster's descendants subsequently to inch their way towards the nearest needle. As it turned out, when all the compensating evolution at other genetic loci had been completed, the body plan represented by that nearest needle eventually emerged as superior to the ancestral unsegmented body plan. The new local optimum, into whose vicinity the lineage wildly leapt, eventually turned out superior to the local optimum on which it had previously been trapped.
This is the kind of speculation in which we should indulge only as a last resort. The argument stands that only gradualistic, inch-by-inch walking through the genetic landscape is compatible with the sort of cumulative evolution that can build up complex and detailed adaptation. Even if segmentation, in our example, ended up as a superior body form, it began as a catastrophe that had to be weathered, just like a climatic or volcanic catastrophe in the external environment. It was gradualistic, cumulative selection that engineered the step-by- step recovery from the segmentation catastrophe, just as it engineers recoveries from external climatic catastrophes. Segmentation, according to the speculation I have just given, survived not because natural selection favoured it but because natural selection found compensatory ways of survival in spite of it. The fact that advantages in the segmented body plan eventually emerged is an irrelevant bonus. The segmented body plan was incorporated into evolution, but it may never have been favoured by natural selection.
But in any case gradualism is only a part of core Darwinism. A belief in the ubiquity of gradualistic evolution does not necessarily commit us to Darwinian natural selection as the steering mechanism guiding the search through genetic space. It is highly probable that Motoo Kimura
DARWIN TRIUMPHANT
87
? LIGHT WILL BE THROWN
is right to insist that most of the evolutionary steps taken through genetic space are unsteered steps. To a large extent the trajectory of small, gradualistic steps actually taken may constitute a random walk rather than a walk guided by selection. But this is irrelevant if - for the reasons given above - our concern is with adaptive evolution as opposed to evolutionary change per se. Kimura himself rightly insists* that his 'neutral theory is not antagonistic to the cherished view that evolution of form and function is guided by Darwinian selection'. Further,
the theory does not deny the role of natural selection in determining the course of adaptive evolution, but it assumes that only a minute fraction of DNA changes in evolution are adaptive in nature, while the great majority of phenotypically silent molecular substitutions exert no significant influence on survival and reproduction and drift randomly through the species.
The facts of adaptation compel us to the conclusion that evolutionary trajectories are not all random. There has to be some nonrandom guidance towards adaptive solutions because nonrandom is what adap- tive solutions precisely are. Neither random walk nor random saltation can do the trick on its own. But does the guiding mechanism necessarily have to be the Darwinian one of nonrandom survival of random spontaneous variation? The obvious alternative class of theory postulates some form of nonrandom, i. e. directed, variation.
Nonrandom, in this context, means directed towards adaptation. It
does not mean causeless.