Climbing Mount Improbable

by Richard Dawkins

  These astonishing orchids epitomize an important aspect of pollination strategy. Many flowers seem to take great pains to get pollinated {262} by one particular kind of animal but not any other. In the New World tropics, red tubular flowers are diagnostic of humming-bird pollination. Red is a bright and attractive colour to bird eyes (insects can't see red as a colour at all). Long, narrow tubes exclude all but specialist pollinators with long narrow beaks — humming-birds. Other flowers go out of their way to be pollinated only by bees, and we've already noted that their flowers are often coloured and patterned in the invisible (to humans) ultraviolet part of the spectrum. Yet others are pollinated only by night-flying moths. They are often white and they make use of scents in preference to visible advertisements. Perhaps the climactic stage in the progression towards an exclusive pollination partnership is the close hand-in-glove duo of fig trees with their own particular fig wasps, the example with which our book begins and ends. But why should plants be so fussy about who pollinates them?

  Presumably the advantage of cultivating specialist pollinators is a more extreme version of the advantage of having animal pollinators at all, rather than wind. It narrows the target. Wind pollination is supremely extravagant, wastefuUy bathing the entire countryside in a rain of pollen. Pollination by jack-of-all-trades flying animals is better, but still pretty wasteful. The bee who visits your flower may fly on to a flower of a quite different species and your pollen will be wasted. Pollen borne by ordinary bees is not exactly rained over the countryside like that of a wind-pollinated grass, but it is still relatively indiscriminately splashed about. Contrast this with the bucket orchid's private species of bee, or a fig tree's private fig wasp. The insect flies unerringly, like a tiny guided missile, or like what medical journalists call a ‘magic bullet’, to exactly the right target from the point of view of the plant whose pollen it bears. In the case of the fig wasp, this means, as we shall see, not just another fig tree but another fig tree of precisely the right species out of the 900 fig species available. Employing specialist pollinators must permit huge savings in pollen production. On the other hand, as we shall also see, it raises other costs of its own, and it is not surprising that some plants are led by their way of life to stay with the wasteful wind as their pollinator. Other plant species are best suited by an intermediate technique along the spectrum from scattergun to magic bullet. Figs are perhaps the ultimate in dependency on the magic bullet {263} of a particular species of pollinator and we reserve them for our climax, in the final chapter.

  Returning to bees, the pollination services that they offer are truly massive. It has been calculated that, in Germany alone, honey-bees pollinate about ten trillion flowers in the course of a single summer day. It has also been calculated that 30 per cent of all human foods are derived from bee-pollinated plants, and that the economy of New Zealand would collapse if bees were wiped out. Bees, flowers might say, are put into the world to carry our pollen around for us.

  The coloured and fragrant flowers of the world, then, although they may seem to be placed there for our benefit, are definitely not so. Flowers live in an insect garden, a mysterious ultraviolet garden in which, for all our vanities, we are irrelevant. Flowers have always been cultivated and domesticated but, until very recent times, the gardeners were bees and butterflies, not us. Flowers use bees, and bees use flowers. Both sides in the partnership have been shaped by the other. Both sides, in a way, have been domesticated, cultivated, by the other. The ultraviolet garden is a two-way garden. The bees cultivate the flowers for their purposes. And the flowers domesticate the bees for theirs.

  Partnerships like this are quite common in evolution. There are so-called ant gardens consisting of epiphytes (plants that grow on the surface of other plants), which ants sow by bringing seeds of the right type and burying them in the soil of their nests. The plants grow out of the surface of the nest and their leaves provide food for the ants. It has been shown that some plants grow better if their roots are in an ants’ nest. Other ant and termite species are specialized to cultivate fungi underground, planting the spores, weeding the gardens to rid them of competing fungus species, and fertilizing them with compost mulched from chewed-up leaves. In the case of the famous leafcutter ants of the New World tropics, all the foraging efforts of their eight-million-strong colonies are directed towards harvesting fresh-cut leaves. They can devastate an area with a ruthless efficiency reminiscent of a locust plague. Yet the leaves that they take are not to be eaten by the ants or their larvae, but are gathered purely to fertilize the fungus gardens. The ants themselves eat only the fungi, which arc of a species that grows nowhere else than in the nests of this kind of {264} ant. These fungi might say that ants are there purely to cultivate fungi, and the ants might say that the fungi exist purely to feed ants.

  Perhaps the most remarkable of all the ant-loving plants are the South-East Asian epiphytes which grow a large bulbous swelling in the stem called a pseudo-bulb. The pseudo-bulb is hollowed with a labyrinth of cavities. These cavities are so like the ones ants commonly dig for themselves in soil that one would naturally suspect ants of fashioning them. This is not the case, however. The cavities are made by the plant and ants live in them (Figure 8.4).

  Figure 8.4 A plant that provides custom-made accommodation for ants in return for protection. Cross-section of a pseudo-bulb of Myrmecodia pentasperma. {265}

  Figure 8.5 Acacia thorn. Another example of cooperation between ants and plants. These bulbous thorns are hollowed out conveniently for ants.

  Better known are the species of ants that live only in special hollow thorns of acacia trees (Figure 8.5). The thorns are thick and bulbous and the plant makes them already hollowed out, apparently for no other purpose than to house ants. What the plants gain from the arrangement is protection, provided by the ants’ vicious stings. This has been shown by elegantly simple experiments. Acacias whose ants have been killed by insecticide soon suffer marked increases in depredation from herbivores. Ants, if they think at all, think that acacia thorns are for the benefit of ants. Acacias think that ants are for protecting them from browsers. Should we, then, think of each member of such partnerships as working for the good of the other? It is better to think of each as using the other for its own good. It is a kind of mutual exploitation in which each benefits from the other enough to make the costs of helping it worth paying.

  There is a temptation, for which ecologists have been known to fall, to see all of life as a sort of mutual-support encounter-group. Plants are the community's primary energy harvesters. They trap the sun's rays and make its energy available to the whole community. {266} They contribute to the community by being eaten. Herbivores, including the very abundant herbivorous insects, are the conduit by which the suns energy is channelled from the primary producers, the plants, to higher stages in the food chain, the insectivores, small carnivores and large carnivores. When animals defecate or die, their vital chemicals are recycled by the scavengers such as dung beetles and burying beetles who hand the precious burden over to soil bacteria who eventually make it available to plants again.

  There would not be too much wrong with this cosily benign picture of the circulation of energy and other resources, if only it were clearly understood that the participants are not doing it for the good of the circle. They are in the circle for the good of themselves. A dung beetle scavenges dung and buries it for food. The fact that she and her kind thereby perform a cleaning-up and recycling service which is valuable to the other inhabitants of the area is strictly incidental.

  Grass provides the staple diet for a whole community of grazers, and the grazers manure the grass. It is even true that, if you removed the grazers, many of the grasses would die. But this does not mean that a grass plant exists to be eaten, or in any sense benefits by being eaten. A grass plant, if it could express its wishes, would much rather not be eaten. How, then, do we resolve the paradox that if the grazers were removed the grasses would die? The answer is that, although
no plant wants to be eaten, grasses can tolerate it better than many other plants can (which is why they are used in lawns that are designed to be mown). As long as an area is heavily grazed or mown, plants that would compete with grasses cannot establish themselves. Trees cannot get a foothold because their seedlings are destroyed. Grazers, therefore, are indirectly good for grasses as a class. But this still does not mean that an individual grass plant benefits by being grazed. It may benefit from other grasses being grazed, including other plants of its own species, since this will have dividends in manure and in helping to remove competitor plants. But if the individual grass plant can get away with not being grazed itself, so much the better.

  We began by lampooning the common fallacy that flowers and animals are placed in the world for the benefit of humans, cattle are {267} docilely eager to be eaten, and so on. Marginally more defensible was the idea that they are placed in the world for the benefit of others with whom they have a naturally evolved mutualism: flowers for the benefit of bees, bees for the benefit of flowers, acacia bullhorns for the benefit of ants and their ants for the benefit of acacias. But this notion of creatures being ‘for the good’ of other creatures is in peril of reductio ad absurdum. We must have no truck with the pop ecologist s fallacy, the holisty grail of all individuals striving for the good of the community, the ecosystem, ‘Gaia’. It is time to get fussy and sharpen up what we mean whenever we talk of a living creature being there ‘for the benefit of anything. What does ‘for the good of really mean? What are flowers and bees, wasps and figs, elephants and bristlecone pines — what are all living things really for? What kind of an entity is it whose ‘benefit’ will be served by a living body or a part of a living body?

  The answer is DNA. It is a profound and precise answer and the argument for it is watertight, but it needs some explanation. It is this explanation that I want to come on to now and in the next chapter. I'll begin by returning to my daughter.

  She was once suffering from a high fever and I suffered vicariously with her as I took my turns sitting by her bedside, sponging her down with cool water. Modern doctors could assure me that she was not in serious danger but the sleep-deprived mind of a loving father could not help recalling the countless childhood deaths of earlier centuries and the agony of each individual loss. Charles Darwin himself never recovered from the uncomprehended death of his beloved daughter Annie. The apparent injustice of her illness was said to have contributed to his loss of religious faith. If Juliet had turned to me and asked, in a piteous echo of our earlier and happier conversation, ‘What are viruses for?’, how should I have answered?

  What are viruses for? To make us better and stronger through triumphing over adversity? (Like the ‘benefits’ of Auschwitz as was suggested by a professor of theology with whom I shared a debating platform on British television.) To kill enough of us to prevent the overpopulation of the world? (An especial boon in countries where effective contraception has been prohibited by theological authority.) {268} To punish us for our sins? (In the case of the AIDS virus, you will find plenty of enthusiasts to agree. One feels almost sorry for medieval theologians that this admirably moralistic pathogen was not around in their time.) Once again, these replies are too humancentred, albeit in a negative way. Viruses, like everything else in nature, have no interest in humans, positive or negative. Viruses are coded program instructions written in DNA language, and they are for the good of the instructions themselves. The instructions say ‘Copy Me and Spread Me Around’ and the ones that are obeyed are the ones that we encounter. That is all. That is the nearest you will come to an answer to the question ‘What is the point of viruses?’ It seems a pointless point, and that is precisely what I now wish to emphasize. I shall do so using the parallel case of computer viruses. The analogy between true viruses and computer viruses is extremely strong and it is also illuminating.

  A computer virus is just a computer program, written in the same sort of language as any other computer program and travelling via the same range of media, for instance floppy discs, or the network of computers, telephone wires, modems and software that is called the Internet. Any computer program is just a set of instructions. Instructions to do what? It could be essentially anything. Some programs are sets of instructions to reckon accounts. Word processors are sets of instructions to accept typed words, move them around the screen and eventually print them. Yet other programs, like Genius 2 which recently defeated Kasparov, the Grand Master, are instructions to play chess very well. A computer virus is a program consisting of instructions that say something like this: ‘Every time you come across a new computer disc, make a copy of me and put it on to that disc’ It is a ‘Duplicate Me’ program. It may incidentally say something more, for instance, ‘Erase the entire hard disc’ Or it may cause the computer to speak, in tinny robotic tones, the words ‘Don't panic’. But that is by the way. The hallmark of a computer virus, its identifying feature, is that it contains the instructions ‘Duplicate me’, written in a language that computers will obey.

  Humans may see no reason to obey such starkly peremptory commands, but computers slavishly obey anything so long as it is written {269} in their own particular language. ‘Duplicate me’ will be obeyed just as readily as ‘Invert this matrix’ or ‘Italicize this paragraph’ or ‘Advance this pawn two squares’. Moreover, there is plenty of opportunity for cross-infection. Computer-users profligately exchange floppy discs, passing game programs around to friends, and useful programs too. You can easily see that, when there are lots of discs being promiscuously shared around, a program that said ‘Copy me on to every disc you encounter’ would spread around the world like chicken-pox. There would soon be hundreds of copies about, and the number would tend to increase. Nowadays, with information highways crisscrossing cyberspace, the opportunities for high-speed cross-infection by computer viruses are even better.

  It is tempting to expostulate about the pointlessness of such parasitic programs, as I did when talking about disease viruses. What on earth is the use of a program that says nothing but ‘Duplicate this program’? Admittedly it will be duplicated but isn't there something ridiculously otiose about such purely self-referential efforts? Of course there is! It is viciously futile. But it doesn't matter that it is futile and pointless in that sense. It can be utterly pointless and still spread. It spreads because it spreads because it spreads. The fact that it does nothing useful on the way — may even do something harmful on the way — is neither here nor there. In the world of computers and disc-swapping, it survives simply because it survives.

  Biological viruses are just the same. Fundamentally a virus is just a program, written in DNA language, which is very much like a computer language even to the point of being written in a digital code. Like a computer virus, the biological virus simply says ‘Copy me and spread me around’. As in the case of computer viruses we aren't suggesting that the DNA in a virus wants to get itself copied. It is just that, of all ways in which DNA could be arranged, only the arrangements that spell out the instructions ‘Spread me’ spread. The world willy-nilly becomes full of such programs. Once again, like the computer viruses, they're here because they're here because they're here. If they didn't embody instructions to ensure that they exist, they would not exist.

  The only important difference between the two kinds of virus is that computer viruses are designed by the creative efforts of mischievous {270} or evil humans, while biological viruses evolve by mutation and natural selection. If a biological virus has bad effects like sneezing or death, these are by-products or symptoms of its methods of spreading. The bad effects of computer viruses are sometimes of this type. The famous Internet Worm, which raced around the networks of the United States on 2 November 1988, had bad effects that were all non-deliberate by-products (a computer worm is technically distinct from a computer virus but the difference need not trouble us here). Copies of the program expropriated memory space and processor time, and brought around 6,000 computers to a standstil
l. Computer viruses, as we have seen, sometimes have bad effects which are not byproducts or necessary symptoms but gratuitous manifestations of pure malice. Far from assisting the spread of the parasite these malicious effects, if anything, slow it down. Real viruses would do nothing so human-centred unless they were designed in a biological-warfare laboratory. Naturally evolved viruses don't go out of their way to kill us or make us suffer. They have no interest in whether we suffer or not. If we suffer, it is a by-product of their self-spreading activities.

  'Duplicate me’ instructions, like any instructions, are no use unless there is machinery set up to obey them. The world of computers is a fine and friendly place for a Duplicate Me program. Computers, linked by the Internet, abetted by people borrowing and lending discs, constitute a kind of paradise for a self-copying computer program. There is ready-made instruction-copying and instruction-obeying machinery humming and whirring and, in a sense, begging to be exploited by any program that says ‘Duplicate me’. In the case of DNA viruses, the ready-made copying and obeying machinery is the machinery of cells, the whole elaborate paraphernalia of Messenger RNA, of Ribosomal RNA and of the various Transfer RNAs, each one hooking on to its own, key-coded amino acid. Never mind the details, or look them up in J. D. Watson's superbly clear Molecular Biology of the Gene. For our purposes it is enough to understand, first, that every cell contains a miniature analogue of a computer's instruction-obeying machinery and, second, that the machine code of all cells, in all creatures on Earth, is identical. (Computer viruses don't have that {271} luxury, by the way: DOS viruses cannot infect Macs, and vice versa.) Computer virus instructions and DNA virus instructions are obeyed because they are written in a code that is slavishly obeyed in the environments in which they respectively find themselves.