Chapter 5 Chapter 3 The Immortal Spiral Circle

We are survival machines, but here we don't just refer to people, it includes all animals, plants, bacteria and viruses.The total number of survival machines on Earth is difficult to calculate, and even the total number of species is unknown.In the case of insects alone, it is estimated that there are about three million extant species, and that individual insects may number in the tens of billions. Different kinds of survival machines have an ever-changing variety of external shapes and internal organs.An octopus has nothing in common with a mouse.And both of these are very different from oak trees.But their basic chemical structures are pretty much the same, not least because they have replicating genes that are basically the same type of molecules that we find in everything from elephants to bacteria.We are all the same survival machine that replicates our genes, the molecule that we call DNA.But the way of living in the world is very different, so the replicators have created a large number of various survival machines for their use.Monkeys are machines that preserve genes to live in trees, fish are machines that preserve genes to live in water, and there is even a kind of worm that is machines to preserve genes to live in German beer mug mats. The way DNA works is truly mysterious.

For the sake of brevity, I have described the modern gene, made of DNA, as almost the same as the first replicators in the primordial soup.This doesn't matter much to the argument, but it probably doesn't.The original replicator may be a molecule similar to DNA, or it may be quite different. In the latter case, we may say that the replicator's survival machinery was captured by DNA at a later stage.If the above is true, then the original replicators have been completely wiped out, since there is no trace of them in modern survival machines.On the basis of such inferences, A.G. Cairns-Smith (A.G. Cairns-Smith) made an interesting suggestion that our ancestors, the first replicators, may not be organic molecules at all, but inorganic crystals. Some minerals and small pieces of clay etc.Regardless of whether DNA is predators or not, it is indisputable that it is the juggernaut today, unless, as I tried to suggest in the last chapter, a new predatory force is now on the rise.

A DNA molecule is a long chain of building blocks, small molecules called nucleotides.Just as protein molecules are chains of amino acids, DNA molecules are chains of nucleotides. The size of the DNA molecule cannot be seen by the naked eye, but its exact shape has been subtly revealed by indirect methods.It consists of a pair of nucleotide strands intertwined in an elegant helix; this is the double helix or immortal coil.There are only four kinds of nucleotide building blocks, which can be referred to as A, T, C and G for short.These four are the same in all animals and plants, the only difference being the order in which they are intertwined.The human G-block is exactly the same as the snail's G-block.But the sequences of human building blocks differ not only from those of snails, but also differ, albeit to a lesser extent, between individuals (except in the special case of identical twins).

Our DNA resides in our bodies.It is not concentrated in a specific part of the body, but distributed among all cells.The average human body is made up of approximately one quadrillion (one thousand and fifteen) cells.Except in some special cases we can ignore, each cell contains a complete set of DNA of the human body.This DNA can be thought of as a set of instructions on how to make a human body.Represented by the A, T, C, G alphabet of nucleotides.It's like having a bookcase in every room in a huge building, and in the bookcases are the architect's plans for building the whole building.This bookcase in each cell is called the nucleus.There are forty-six volumes of this blueprint of the architect, and we call them chromosomes.In different species, its number is also different.Chromosomes are visible under a microscope and are shaped like long lines.Genes are arranged sequentially along these chromosomes.But figuring out where genes meet end to end is difficult, and in fact may even be meaningless.Fortunately, this chapter will show that this is of little relevance to our thesis.

I will use the analogy of an architect's blueprint, mixing figurative language with professional language at will.The terms volume isochromosomal will be used interchangeably.Pages are temporarily used interchangeably with genes, although the boundaries between genes are not as sharp as the pages of books.We will use this metaphor at length.When this metaphor does not solve the problem, I will refer to other metaphors.By the way, of course there is no such thing as an architect, DNA instructions are arranged by natural selection. There are two important things that DNA molecules do: First, they replicate, that is, replicate themselves.Since life, such replication activities have never been interrupted.Now the DNA molecule is really skilled at replicating itself, and it is easy to handle.An adult has a thousand and fifteen cells in his body, but in an embryo starts out as a single cell with a master copy of the architect's blueprint.This single cell was divided into two, and each of the two cells received its own copy of the blueprint.Cells divide in turn by four, eight, sixteen, thirty-two, etc. multiples until billions.Every time it splits, the blueprint of DNA is copied without any distortion, and errors are rare.

Speaking of DNA replication is only one aspect.But if DNA is really a set of blueprints for building a human body, how does that work?How will they be transformed into human tissue?This is the second important thing that DNA does that I want to talk about.It indirectly oversees the manufacture of different kinds of molecular proteins.Hemoglobin, mentioned in the previous chapter, is an example of an extremely wide variety of protein molecules.The DNA code information represented by the four-letter nucleotide alphabet is mechanically translated into another alphabet.This is what spells out the amino acid alphabet for protein molecules.

Making proteins may seem like a long way from making a human body, but it is the first small step in that direction.Proteins are not only the main components of human tissue, but they also sensitively control all chemical processes in cells, and selectively make this chemical process continue or stop at accurate time and exact place.Exactly how this process develops into a baby is a long story, and embryologists will take decades, perhaps centuries, to figure it out.But it is an indisputable fact that the final result of this process is a baby.Genes do indirectly control the making of the human body, and their influence is entirely one-way: acquired traits are not inherited.No matter how much intelligence you acquire in life, none of it will be passed on to your children through genetic pathways.The new generation starts from scratch.The human body is nothing more than a means for the genes to keep themselves unchanged.

The evolutionary significance of the fact that genes control embryonic development is this: It means that genes are at least partially responsible for their own future survival, since their survival depends on the efficiency of the human body they inhabit and help build .Long ago, natural selection consisted of the differential survival of replicators floating freely in the primordial soup.Today, natural selection favors replicators that are adept at building survival machines—genes that are adept at controlling embryonic development.In this respect, replicators are as sensual and purposeful as ever.The automatic selection among competing molecules for their respective longevity, fecundity, and capacity for exact reproduction continues as blindly and inevitably as in distant epochs.Genes don't have foresight, they don't plan ahead.Genes are just that, some more than others.This is the case.But the traits that determine genetic longevity and fecundity are not as simple as they once were, far from being so.

In recent years (meaning the past 600 million years or so), replicators have made remarkable achievements in the craft of building survival machines, such as muscles, hearts, and eyes (through several separate evolutionary processes).By that time, as replicators, the basic features of their way of life had changed radically.We need to understand this if we are to proceed with our argument. The first thing to know about modern replicators is that they are highly social.A survival machine is a vehicle that contains not just one gene but thousands.Making the human body is a coordinated, intricate venture, in which the contribution of one gene is almost inseparable from the contribution of another to the common cause.A single gene can have many different effects on different parts of the body.A part of the human body is influenced by many genes, and the effect of any one gene depends on the interaction with many other genes.Certain genes act as master genes, controlling the activity of a group of other genes.In analogy, any one page of the blueprint provides references to many different parts of the building, and each page only makes sense as a cross-reference with many other pages.

This intricate interdependence of genes might confuse you, why do we use the word gene?Why not use a collective noun like gene complex?We think it's actually a pretty good idea in many ways.But if we look at things from another angle, it also makes sense to think of the gene complex as dividing into separate replicators.The problem arises due to the presence of sexual phenomena.Sexual reproduction has the effect of mixing genes, which means that any individual is just a temporary vehicle for a short-lived gene combination.The combination of genes in any one individual may be short-lived, but the genes themselves can live for a long time.Their paths intersect and intersect each other, and in successive generations a gene can be regarded as a unit, which survives through the continuation of a series of individuals.This is the central theme of this chapter.Some of my highly respected colleagues stubbornly refuse to accept this argument.So please forgive me if I seem to be a bit tedious in my arguments!First I must briefly state some facts concerning sex.

I have said that the blueprint for building a human body was written in forty-six volumes.In fact, this is an oversimplified approach.The reality is quite bizarre.Forty-six chromosomes consist of twenty-three pairs of chromosomes.We might say that each cell nucleus stores two sets of twenty-three interchangeable volumes of blueprints.We may call them Volume 1a, Volume 1b, Volume 2a, Volume 2b, and Volume 23a, Volume 23b.Of course the numbers I use to identify each volume and each page thereafter are chosen arbitrarily. We receive each complete chromosome from either father or mother, where they are assembled in the testes and ovaries, respectively.For example, volume 1a, volume 2a, volume 3a comes from father, volume 1b, volume 2b, volume 3b comes from mother.Although practically impossible, you could theoretically use a microscope to look at the forty-six chromosomes in any one of your cells and tell which twenty-three came from your father and which twenty-three came from your mother. In fact, pairs of chromosomes do not stick together for life, and they are not even close to each other.So in what sense are they paired?To say that they are in pairs means that each volume originally from the father can be considered to directly replace, page by page, the corresponding volume originally from the mother.For example, page 6 of volume 13a and page 6 of volume 13b may both be concerned with eye color, perhaps the upper page says blue and the other page says brown.Sometimes the two alternate pages are identical, but in other cases, as in our eye color example, they are different from each other.What happens to the human body if they make conflicting recommendations?There are different outcomes for each bet.Sometimes one page has more impact than another.In the eye color example just given, the person might actually have brown eyes, since the instructions to make blue eyes might have been ignored during the construction of the human body.Still, that wouldn't stop the order to make blue eyes from being passed on to future generations.A gene that is ignored in this way is called a recessive gene.The opposite of recessive genes are dominant genes.The gene for brown eyes has an advantage over the gene for blue eyes.A person gets a pair of blue eyes only if both copies of the relevant page unanimously recommend blue eyes.More often, the two alternative genes are not exactly alike, and the result is some type of compromise that builds the body into an intermediate shape, or a completely different shape. When two genes, such as brown eyes and blue eyes, compete for the same position on a chromosome, we call one an allele of the other.For our purposes, allele is synonymous with competitor.Think of the architect's volumes of blueprints as binders with pages that can be pulled out and exchanged.Every volume thirteen must have a sixth page, but several sixths could go in the binder, sandwiched between pages five and seven.One version says blue eyes; another may say brown eyes: there may be other versions throughout the population that say other colors like green.Perhaps six alternative alleles occupy the sixth position on the thirteenth chromosome scattered throughout the population.Each person has only two rolls of thirteen chromosomes.Therefore, there can only be a maximum of two alleles at the position on page six.For example, a person with blue eyes may have two copies of the same allele, or he may choose two of the six alternative alleles in the population. Of course, you can't really go to the gene pool of the entire population to choose your own genes.At any one time, all genes are tightly knit together within an individual survival machine.We each receive all our genes as embryos, and there is nothing we can do about it.In the long run, however, it makes sense to call the genes of an entire population collectively the gene pool.It's actually a term used by geneticists.The gene pool is a rather useful abstraction because sexual activity mixes genes, albeit a carefully orchestrated process.In particular, something like pulling out pages, stacks, and swapping them from a binder does happen, as we'll see shortly.I have described the normal division of a cell into two new cells.Each cell that divides receives a complete copy of all forty-six chromosomes.This normal cell division is called mitosis.But there is another type of cell division called meiosis.Meiosis occurs only during the production of sex cells, namely sperm and eggs.Sperm and eggs have a unique aspect in our cells that they only have twenty-three chromosomes instead of forty-six.This number, of course, happens to be half of forty-six.How convenient it is for them to fuse together after fertilization or insemination to make a new individual!Meiosis is a specialized type of cell division that occurs only in the testes and ovaries.In this process, a cell with a complete double of a total of 46 chromosomes divides into a sex cell with only a single copy of a total of 23 chromosomes (all take the number of chromosomes in the human body as an example). A sperm with 23 chromosomes is produced by meiosis of an ordinary cell with 46 chromosomes in the testis.Which twenty-three chromosomes go into a sperm cell?It is obviously important that a sperm should not acquire any twenty-three chromosomes, i.e. it should not have two copies of Volume XIII and none of Volume XVII.It is theoretically possible that an individual can endow one of his sperm with all the chromosomes from his mother (ie, volume 1b, volume 2b, volume 3b, volume 23b).In this unlikely scenario, a child conceived with this type of sperm would inherit half of her genes from her grandmother but not from her grandfather.In practice, however, this total chromosome-wide distribution does not occur.The actual situation is much more complicated.Please don't forget that the volumes of blueprints (chromosomes) are treated as binders.During the production of spermatozoa, many single pages or stacks of single pages of a certain volume of blueprints are extracted and exchanged with alternative corresponding single pages of another volume.Thus, volume one of a particular sperm cell may be organized in such a way that the first sixty-five pages are taken from volume one a, and pages sixty-six through the last are taken from volume one b.The other twenty-two volumes of this spermatid are organized in a similar manner.Thus, even though all the twenty-three chromosomes of a man's sperm are composed of segments of the same set of forty-six chromosomes, each sperm cell he makes is unique.Eggs are made in the ovary in a similar way, and they are all unique and different. This mixed composition method in real life is well known.During the making of a sperm (or egg), pieces of each paternal chromosome separate and exchange places with the exact corresponding pieces of the maternal chromosome (remember, we are talking about the chromosomes originally from the making of this sperm). Chromosomes from the parents of an individual, i.e., the grandparents of the child resulting from fertilization by this sperm).This process of exchanging chromosome segments is called crossover.This is a point that is crucial to the entire argument of this book.That said, it would be futile if you looked at the chromosomes of one of your own sperm (or egg in the case of a female) with a microscope and tried to identify which chromosomes were originally the father's and which were originally the mother's (This is in stark contrast to somatic cells in general, see p. 33).Any one chromosome in the sperm is a kind of improvisation, that is, a mosaic of the mother's gene and the father's gene. The page-to-gene analogy can no longer be used from here on.In binders, complete pages can be inserted, removed, or exchanged, but not fragments.But the gene complex is just a long string of nucleotide letters, not clearly divided into separate pages.Of course, there are special symbols for the head of the protein chain information and the tail of the protein chain information, that is, the same four-letter alphabet as the protein information itself.Between these two punctuation marks are the coded instructions for making a protein.If we like, we can understand a gene as the sequence of nucleotide letters between the head and tail symbols and the code for a chain of proteins.We use the word cistron to denote a unit with such a definition.Some people use gene and cistron as two words that can be used interchangeably.But exchange does not respect the boundaries between cistrons.Splitting can occur not only between cistrons, but also within cistrons.It was as if the architect's blueprints were drawn on forty-six ticker rolls rather than separate pages.A cistron has no fixed length.Only by looking at the symbols on the paper strips and looking for the symbols of the head and tail of the information can we find where the previous cistron ends and the next cistron starts.The exchange is represented by the process of taking matching paternal and maternal strips, cutting out and exchanging their matching parts, regardless of what is drawn on them. The word gene, as used in the title of this book, does not refer to a single cistron but to something more nuanced and complex.My definition won't be to everyone's taste, but again there is no universally accepted definition of genes.Even if there were, definitions are not sacrosanct.If our definitions are clear and unambiguous, it is all right to define a word as we like, for our own purposes.The definition I'm going to use comes from Williams.Gene is defined as any part of chromosomal material which is capable of functioning as a unit of natural selection over successive generations.In the terms of the previous chapter, a gene is a replicating gene that replicates with great precision.The ability to reproduce exactly is another way of saying longevity through copying, which I will simply call longevity.The correctness of this definition needs further proof. By any definition, a gene must be part of a chromosome.The question is how big is this part, ie how many long ticker strips?Let us imagine any sequence of adjacent code letters on the slip of paper; call this sequence the genetic unit.It may be a sequence of only ten letters within a cistron; it may be a sequence of eight cistrons; it may have its head and tail in the middle of the cistron.It must overlap with other genetic units.It will include smaller genetic units, and it will also form part of a larger genetic unit.Regardless of its length, for the convenience of the present argument we shall call it a hereditary unit.It is nothing but a segment of a chromosome, without any substantial difference from the rest of the chromosome. Now, this is important: the shorter the genetic unit, the longer it is likely to live in generations.Especially the less likely it is to be split by a swap.Assuming that on average, every time meiosis produces a sperm or egg, the entire chromosome may undergo an exchange, and this exchange may occur on any segment of the chromosome.If we imagine that this is a large genetic unit, say half the length of a chromosome, then every time meiosis occurs, the chance of this genetic unit dividing is 50%.If the genetic unit we envision is only 1% as long as a chromosome, we can assume that it will divide only 1% of the time in any one meiosis.That is to say, this genetic unit is able to survive many generations in the offspring of that individual.A cistron is probably much shorter than 1% of a chromosome.Even a group of several adjacent cistrons can live for many generations before being broken up by exchange. The average life expectancy of a genetic unit can be conveniently expressed in terms of generations, which can also be converted to years.If we take the whole chromosome as the assumed genetic unit, its life history will only last one generation.Now assuming that eight a is your chromosome, inherited from your father, it was made in one of your father's testes shortly before you were conceived.Before that, the world had never existed before.This genetic unit is the product of the mixing process of meiosis, which brings together segments of chromosomes from your grandfather and grandmother.This genetic unit is placed within a particular sperm and is therefore unique.This sperm is one of millions sailing into your mother's womb in this vast fleet of tiny ships.This particular sperm (unless you're a non-identical twin) is the only one in the fleet to find a port of call in one of your mother's eggs.This is why you exist.The genetic unit we envision, your eight a chromosomes, begins to replicate itself along with the rest of your genetic material.Now it exists in duplicate all over your body.But when it's your turn to have a baby, that chromosome is destroyed when you make an egg (or sperm).Segments of this chromosome will exchange with segments of your mother's eight b chromosome.In any one sex cell there will be a new chromosome eight, which may be better or worse than the old one.But unless it's a very rare coincidence, it's sure to be different, unique.The lifespan of a chromosome is one generation. How long does a small genetic unit, say one percent as long as your chromosome 8a, live?This genetic unit also came from your father, but probably wasn't originally assembled in him.According to the previous reasoning, there is a ninety-nine percent chance that he received it intact from his father or mother.For now let's just assume he got it from his mother, your grandmother.It is also 99% likely that she has received it intact from her father or mother as well.If we trace the ancestry of a small genetic unit, we eventually find its original creator.At some stage, this genetic unit must have been first created in the testes or ovaries of one of your ancestors. Let me repeat once again the rather peculiar sense I have used of the word creation.The smaller subunits that we envision making up the genetic unit may have existed long ago.When we say that a genetic unit was created at a particular moment, we mean only that the particular arrangement of subunits that make up the genetic unit did not exist before that moment.Perhaps the creation was fairly recent, such as within your grandfather or grandmother.But if we're thinking of a very small genetic unit, it might have been assembled for the first time by a very distant ancestor, perhaps a prehuman ape.And the small genetic unit in your body can also last for a long time in the future, passing on from generation to generation intact. Also don't forget that the offspring of an individual are not single-line, but branched.Whichever of your ancestors created your particular short chromosome 8a, he or she likely had many other descendants besides you.One of your genetic units may also exist in your second cousin.It could be in me, it could be in the Prime Minister, it could be in your dog.Because if we go back far enough, we all share a common ancestor.Even this small genetic unit may have happened to undergo several independent assemblies: if the genetic unit is small, such a coincidence is not very improbable.But even a close relative is unlikely to share an entire chromosome with you.The smaller the genetic unit, the greater the possibility of sharing a whole chromosome with another individual, that is, the greater the possibility of manifesting in the world many times in the form of copies. The chance aggregation of preexisting subunits by crossover is the usual way to form a new genetic unit.Another way is called point mutation.This approach, although rare, is evolutionarily significant.A single genetic point mutation is equivalent to a typographical error of a single letter in a book.Although this is rare, it is clear that the longer the genetic unit, the more likely it is to be altered by a mutation at some point. Another error or mutation that is less common but has important long-term consequences is called an inversion.The chromosome separates a section of itself at both ends, reverses the head and tail, and rejoins in this reversed position.Following the previous analogy, it was necessary to renumber some of the page numbers.Sometimes parts of a chromosome are not simply inverted, but join to a completely different part of the chromosome, or even join with a completely different chromosome.This situation is like changing a stack of loose-leaf pages from one binder to another.While this type of error is usually catastrophic, it can sometimes make tightly linked segments of genetic material that happen to work well together, which is why it matters.Perhaps an inversion could bring the two cistrons closer together.They are beneficial only when they are together, that is, in some way complement or reinforce each other.Natural selection then tends to favor new genetic units constituted in this way, so that such genetic units will spread through future populations.Gene complexes may have been comprehensively rearranged or edited in this way over the distant past. One of the best examples of this involves a phenomenon called mimicry.Certain butterflies have a repulsive odor, and their colors are usually brilliant and gorgeous.Birds learn to avoid them by virtue of their warning signs.So some other species of butterflies that don't have this disgusting smell take advantage of the phenomenon.They mimic those weird-smelling butterflies.So they are born with the same color and shape as those weird-smelling butterflies, but with a different taste.They have often fooled human naturalists, and they have often fooled birds.A bird that has eaten butterflies that actually smell weird will usually avoid all butterflies that look the same, including the mimics.Natural selection thus favors genes that promote mimicry.This is how mimicry evolved. There are many different species of smelly butterflies, and they don't all look the same.It's impossible for a Simulator to be like all the weird-smelling butterflies.So, they had to mimic a particular species of butterfly that tastes weird.Any specific imitator species is generally specialized in imitating a specific weird-smelling species.But some kinds of mocks have a very peculiar behavior.Some individuals of these species imitated one odd-smelling species, while others imitated another species.Any individual that is intermediate or tries to imitate both species will be quickly eaten.But butterflies are not born that way.Just as an individual is definitely either male or female, so an individual butterfly either imitates one odd-smelling species or the other.A butterfly may imitate species A, while its siblings may imitate species B. Whether an individual mimics species A or species B appears to depend on just one gene.But how can a single gene determine every aspect of a simulation—color, shape, pattern of patterns, rhythm of flight?The answer is that a gene understood as a cistron is probably not possible, but a large group of old genes that were separated in the past can be found in the Combined into a tight linkage group on one chromosome.The entire linkage group behaves like a single gene (by our definition, it is indeed now a single gene).It also has an allele, which is actually another linkage group.One linkage group contains cistrons mimicking species A, while the other contains cistrons mimicking species B.Each linkage group is rarely split by crossovers, so intermediate butterflies have never been seen in nature.But if the butterflies are fed in large numbers in the laboratory, this intermediate form occasionally appears. I use the word gene to mean a unit of heredity that is small enough to last many generations and to be spread around in many copies.It's not a dead-on definition of all-or-nothing, but like the big or old definition, it's a definition that's gradually blurring.The more easily a segment of chromosome is split by crossovers, or altered by various types of mutations, the less it corresponds in the sense I would call a gene.A cistron can probably be called a gene, but a unit larger than a cistron should also be a gene.Twelve cistrons may be so closely associated with each other on a chromosome that they appear to us as a long-lived genetic unit.The mimic group in butterflies is a good example.As cistrons leave an individual into the next generation, as they ride sperm or eggs into the next generation, they may find that the boat also carries their neighbors from the previous voyage.These neighbors are the companions they have sailed with on this long voyage that began in the bodies of distant ancestors.Adjacent cistrons on the same chromosome form a group of closely linked travel partners. When the time for meiosis comes, they are often able to board the same boat and rarely separate. Strictly speaking, this book should be called neither the selfish cistron nor the selfish chromosome, but the slightly selfish large segment and the even more selfish small segment.But it should be said that such a title is at least not so attractive.Since I pictured the gene as a small stretch of chromosome that lasts many generations, I titled this book The Selfish Gene. 現在我們又回到了第一章結尾的地方。在那裡我們已經看到，在任何稱得上是自然選擇的基本單位的實體中，都會發現自私性。我們也已看到，有人認為物種是自然選擇單位，而另有些人則認為物種中的種群或群體是自然選擇單位，還有人認為個體是自然選擇單位。我曾講過，我寧可把基因看作是自然選擇的基本單位，因而也是自我利益的基本單位。我剛才所做的就是要給基因下這樣的定義，以便令人信服地證明我的論點的正確性。 自然選擇的最普通形式是指實體的差別性生存。某些實體存在下去，而另一些則死亡。但為了使這種選擇性死亡能夠對世界產生影響，一個附加條件必須得到滿足。每個實體必須以許多拷貝的形式存在，而且至少某些實體必須有潛在的能力以拷貝的形式生存一段相當長的進化時間。小的遺傳單位有這種特性，而個體、群體和物種卻沒有。孟德爾（Gregor Mendel）證明，遺傳單位實際上可以認為是一種不可分割和獨立的微粒。這是他的一項偉大的成就。現在我們知道，這種講法未免有點過分簡單。甚至順反子偶然也是可分的，而且在同一條染色體上的任何兩個基因都不是完全獨立的。我剛才所做的就是要把基因描繪為一個這樣的遺傳單位，它在相當大的程度上接近不可分的顆粒性這一典型。基因並不是不可分的，但很少分開。基因在任何具體個體中要麼肯定存在要麼肯定不存在。一個基因完整無損地從祖父母傳到孫子女，徑直通過中間世代而不同其他基因混合。如果基因不斷地相互混和，我們現在所理解的自然選擇就是不可能的了。順便提一句，這一點還在達爾文在世時就已被證實，而且使達爾文感到莫大的憂慮。因為那時人們認為遺傳是一個混和過程。孟德爾的發現那時已經發表，這本來是可以解除達爾文的焦慮的，但天啊，他卻一直不知道這件事。達爾文和孟德爾都去世之後許多年，似乎才有人讀到這篇文章。孟德爾也許沒有認識到他的發現的重要意義，否則他可能會寫信告訴達爾文的。 基因的顆粒性的另一個方面是，它不會衰老，即使是活了一百萬年的基因也不會比它僅活了一百年更有可能死去。它一代一代地從一個個體轉到另一個個體，用它自己的方式和為了它自己的目的，操縱著一個又一個的個體；它在一代接一代的個體陷入衰老死亡之前拋棄這些將要死亡的個體。 基因是不朽的，或者更確切地說，它們被描繪為接近於值得賦予不朽稱號的遺傳實體。我們作為在這個世界上的個體生存機器，期望能夠多活幾十年，但世界上的基因可望生存的時間，不是幾十年，而是以千百萬年計算。 在有性生殖的物種中，作為遺傳單位的個體因為體積太大而且壽命也太短，而不能成為有意義的自然選擇單位。由個體組成的群體甚至是更大的單位。在遺傳學的意義上，個體和群體像天空中的雲彩，或者像沙漠中的塵暴。它們是些臨時的聚合體或聯合體，在進化的過程中是不穩定的。種群可以延續一個長時期，但因為它們不斷地同其他種群混合，從而失去它們的特性。它們也受到內部演化的影響。一個種群還不足以成為一個自然選擇的單位，因為它不是一個有足夠獨立性的實體。它的穩定性和一致性也不足，不能優先於其他種群而被選擇。 一個個體在其持續存在時看起來相當獨立，但很可惜，這種狀態能維持多久呢？每一個個體都是獨特的。在每個實體僅有一個拷貝的情況下，在實體之間進行選擇是不可能實現進化的！有性生殖不等於複製。就像一個種群被其他種群所玷污的情況一樣，一個個體的後代也會被其配偶的後代所玷污，你的子女只一半是你，而你的孫子孫女只是你的四分之一。經過幾代之後，你所能指望的，最多是一大批後代，他們之中每個人只具有你的極小一部分幾個基因而已，即使他們有些還姓你的姓，情況也是如此。 個體是不穩定的，它們在不停地消失。染色體也像打出去不久的一副牌一樣，混和以致被湮沒。但牌本身雖經洗牌而仍存在。這裡，牌就是基因。基因不會被交換所破壞，只是調換夥伴再繼續前進。它們繼續前進是理所當然的，這是它們的本性。它們是複製基因，而我們則是它們的生存機器。我們完成我們的職責後就被棄之一旁，但基因卻是地質時代的居民：基因是永存的。 基因像鑽石一樣長存，但同鑽石長存的方式又不盡相同。長存的一塊塊的鑽石水晶體，它們以不變的原子模型存在。但ＤＮＡ分子不具備這種永恆性。任何一個具體的ＤＮＡ分子的生命都相當短促，也許只有幾個月時間，但肯定不會超過一個人的一生時間。但一個ＤＮＡ分子在理論上能夠以自己的拷貝形式生存一億年。此外，一個具體基因的拷貝就像原始湯中的古代複製基因一樣，可以分佈到整個世界。所不同的是，這些基因拷貝的現代版本都有條不紊地裝入了生存機器的體內。 我所說的一切都是為了要強調，基因通過拷貝形式的存在幾乎是永恆的，這種永恆性表明了基因的特性。將基因解釋為一個順反子適用於某些論題，但運用於進化論，定義就需要擴大。擴大的程度則取決於定義的用途。我們需要找到自然選擇的一個切合實際的單位。要做到這點，首先要鑒別出一個成功的自然選擇單位必須具備哪些特性。用前面一章的話說，這些特性是：長壽，生殖力以及精確複製。那麼我們只要直截了當地把基因解釋為了個至少有可能擁有上述三種特性的最大的實體。基因是一個長久生存的複製基因，它以許多重複拷貝的形式存在著。它並非無限期地生存下去。嚴格地說，甚至鑽石也不是永恆的，順反子甚至也能被交換一分為二。按照定義，基因是染色體的一個片段，它要短得使其能夠延續足夠長的時間，以便它作為一個有意義的自然選擇單位而發生作用。 確切地說，到底多長才算足夠長的時間呢？這並沒有嚴格的規定。問題取決於自然選擇的壓力達到多大的嚴峻程度。就是說，要取決於一個壞的遺傳單位死亡的可能性比它的好的等位基因大到多大程度。這個問題牽涉到因具體情況不同而各異的定量方面的細節。自然選擇最大的切合實際的單位基因，一般界於順反子同染色體之間。 基因之成為合適的自然選擇基本單位，其原因在於它的潛在的永恆性。現在是強調一下潛在的這個詞的時候了。一個基因能生存一百萬年，但許多新的基因甚至連第一代也熬不過。少數新基因成功地生存了一代，部分原因是它們運氣好，但主要是由於它們具有一套看家本領，就是說它們善於製造生存機器。這些基因對其寄居其中的一個個連續不斷的個體的胚胎發育都產生一定的影響。這樣就使得這個個體生存和繁殖的可能性要比其處在競爭基因或等位基因影響下的可能性稍大一些。舉例說，一個好的基因往往賦予它所寄居其中的連續不斷的個體以長腿，從而保證自己的生存。因為長腿有助於這些個體逃避捕食者。這只是一個特殊的例子，不具普遍意義。因為長腿畢竟不是對誰都有好處的。對鼴鼠來說，長腿反而是一種累贅。我們能不能在所有好的（即生存時間長的）基因中找出一些共同的特性，而不要使我們自己糾纏在煩瑣的細節中呢？相反，什麼是能夠立即顯示出壞的即生存短暫的基因的特性呢？這樣的共同特性也許有一些，但有一種特性卻同本書特別有關，即在基因的水平上講，利他行為必然是壞的，而自私行為必定是好的。這是從我們對利他行為和自私行為的定義中得出的無情結論。基因為爭取生存，直接同它們的等位基因競爭，因為在基因庫中，它們看等位基因是爭奪它們在後代染色體上的位置的對手。這種在基因庫中犧牲其等位基因而增加自己生存機會的任何基因，我再嚕囌一句，按照我們的定義，往往都會生存下去。因此基因是自私行為的基本單位。 本章的主要內容已敘述完畢，但我一筆帶過了一些複雜的問題以及一些潛在的假設。第一個複雜的問題我已扼要地提到過。不論基因在世世代代的旅程中多麼獨立和自由，但它們在控制胚胎發育方面並不是那麼非常自由和獨立的行為者。它們以極其錯綜複雜的方式相互配合和相互作用，同時又和外部環境相互配合和相互作用。諸如長腿基因或者利他行為基因這類表達方式是一種簡便的形象化講法，但理解它們的含義是重要的。一個基因，不可能單槍匹馬地建造一條腿，不論是長腿或是短腿。構造一條腿是多基因的一種聯合行動。外部環境的影響也是不可或缺的。因為實際上腿畢竟是由食物鑄造出來的！但很可能有這樣的一個基因，它在其他條件不變的情況下，往往使腿生得比在它的等位基因的影響下生長的腿長一些。 作為一種類比，請想像一下如硝酸鹽那一種肥料對小麥生長的影響。小麥這種植物施以硝酸鹽要比不施硝酸鹽長得大，這是盡人皆知的事實。但恐怕沒有這樣的傻瓜會宣稱，單靠硝酸鹽能生長小麥。種子、土壤、陽光、水分以及各種礦物質顯然同樣不可缺少。但如果上述的其他幾種因素都是穩定不變的，或者甚至在一定範圍內有某些變化，硝酸鹽這一附加因素能使小麥長得更大一些。單個基因在胚胎發育中的作用也是如此。控制胚胎發育的各種關係像蜘蛛網一樣交織連鎖在一起，非常錯綜複雜，我們最好不要去問津。任何一個因素，不論是遺傳上的或環境上的，都不能認為是導致嬰兒某部分形成的唯一原因。嬰兒的所有部分都具有幾乎是無窮數量的先前因素（antecedent causes）。但這一嬰兒同另一嬰兒之間的差別，如腿的長短差別，可以很容易地在環境或基因方面追溯到一個或幾個先前差別（antecedent differences）。就是這些差別才真正關係到生存競爭和鬥爭；對進化而言，起作用的是受遺傳控制的差別。 就一個基因而言，它的許多等位基因是它的不共戴天的競爭者，但其餘的基因只是它的環境的一個組成部分，就如溫度、食物、捕食者或夥伴是它的環境一樣。基因發揮的作用取決於它的環境，而這一所謂環境也包括其餘的基因。有時，一個基因在另一個特定基因在場的情況下所發揮的是一種作用，而在另一組夥伴基因在場的情況下所發揮的又是一種截然不同的作用。一個個體的全部基因構成一種遺傳氣候或背景，它調整和影響任何一個具體基因的作用。 但現在我們似乎有一種佯謬現象。如果建造一個嬰兒是這樣的一種複雜的相互配合的冒險事業，如果每一個基因都需要幾千個夥伴基因配合共同完成它的任務，那麼我們又怎麼能把這種情況同我剛才對不可分的基因的描述一致起來呢？我曾說，這些不可分的基因像永生的小羚羊一樣年復一年、代復一代地從一個個體跳躍到另一個個體：它們是自由自在，不受約束地追求生命的自私行為者，難道這都是一派胡言嗎？絲毫也不是。也許我為了追求詞藻絢麗的章句而有點神魂顛倒，但我絕不是在胡言亂語，事實上也不存在真正的佯謬。我可以用另外一個類比的方法來加以說明。 單靠一個划槳能手在牛津和劍橋的划船競賽中是贏不了的。他需要有八個夥伴。每個划手都是一個專門家，他們總是分別在特定的位置上就坐前槳手或尾槳手或艇長等。划船是一項相互配合的冒險行動，然而有些人划船比另一些人划得好。假使有一位教練需要從一夥候選人中挑選他理想的船員，這些船員中有的人必須是優秀的前槳手，其他一些人要善於執行艇長的職務等等。現在我們假設這位教練是這樣挑選的：他把應試的船員集合在一起，隨意分成三隊，每一隊的成員也是隨意地安排到各個位置上，然後讓這三條船展開對抗賽。每天都是如此，每天都有新的陣容。幾周之後將會出現這樣的情況：贏得勝利的賽艇，往往載有相同的那幾個人。他們被認為是划槳能手。其他一些人似乎總是在划得較慢的船隊裡，他們最終被淘汰。但即使是一個出色的划槳手有時也可能落入划得慢的船隊中。這種情況不是由於其他成員技術差，就是由於運氣不好，比如說逆風很大。所謂最好的划槳手往往出現在得勝的船上，不過是一種平均的說法。 划槳手是基因。爭奪賽艇上每一位置的對手是等位基因，它們有可能佔據染色體上同一個位置。划得快相當於建造一個能成功地生存的個體。風則相當於外部環境。候選人這個整體是基因庫。就任何個體的生存而言，該個體的全部基因都同舟共濟。許多好的基因發現自己與一群壞的基因為伍，它正在同一個致死基因共一個個體。這一致死的基因把這一尚在幼年時期的個體扼殺。這樣，好的基因也就和其餘基因同歸於盡。但這僅是一個個體，而這個好的基因的許多副本卻在其他沒有致死基因的個體中生存了下來。許多好基因的拷貝由於碰巧與壞基因共一個個體而受累；還有許多由於其他形式的厄運而消亡，如它們所寄居的個體被雷電所擊。但按照我們的定義，運氣不論好壞並無規律可循，一個一貫敗陣的基因不能怪它的運氣，它本來就是個壞的基因。 好槳手的特點之一是相互配合好，即具有同其餘槳手默契配合的能力。對於划船來講，這種相互配合的重要性，不下於強有力的肌肉。我們在有關蝴蝶的例子中已經看到，自然選擇可能以倒位的方式、或染色體片段的其他活動方式無意識地對一個基因複合體進行編輯。這樣就把配合得很好的一些基因組成緊密地連接在一起的群體。但在另外一個意義上說，一些實際上並不相互接觸的基因也能夠通過選擇的過程來發揮其相容性（mutual compatibility）。一個基因在以後歷代的個體中將會與其他的基因，即基因庫裡的其他基因相遇，如果它能和這些基因中的大多數配合得很好，它往往從中得到好處。 舉例說，有效的肉食動物個體要具備幾個特徵，其中包括鋒利的切嚼牙齒，適合消化肉類的腸胃，以及其他許多特徵。但另一方面，一個有效的草食動物卻需要扁平的磨嚼牙齒，以及一副長得多的腸子，其消化的化學過程也不同。在草食動物的基因庫中，任何基因，如果它賦於其主人以鋒利的食肉牙齒是不大可能取得成功的。這倒不是因為食肉對誰來說都是一種壞習慣，而是因為除非你有合適的腸子，以及一切食肉生活方式的其他特徵，否則，你就無法有效地吃肉類。因此，影響鋒利的食肉牙齒形成的基因並非本來就是壞基因。只有在草食動物種種特徵形成的基因所主宰的基因庫中，它們才算是壞基因。 這是個複雜的微妙的概念。它之所以複雜，是因為一個基因的環境主要由其他基因組成。而每一個這樣的基因本身又依它和它的環境中的其他基因配合的能力而被選擇。適合於說明這種微妙概念的類比是存在的，但它並非來自日常生活的經驗。它同人類的競賽理論相類似，這種類比法將在第五章講到個體動物間進行的進犯性對抗時加以介紹。因此，我把這點放到第五章的結尾處再進一步討論。現在我回過頭來繼續探討本章的中心要義。這就是，最好不要把自然選擇的基本單位看作是物種，或者是種群，甚至是個體；最好把它看作是遺傳物質的某種小單位。為方便起見，我們把它簡稱為基因。前面已經講過，這個論點的基礎是這樣一種假設：基因能夠永存不朽，而個體以及其他更高級的單位的壽命都是短暫的。這一假設以下面兩個事實為依據：有性生殖和交換；個體的消亡。這是兩個不容否認的事實。但這不能阻止我們去追問一下：為什麼它們是事實。我們以及大多數的其他生存機器為什麼要進行有性生殖？為什麼我們的染色體要進行交換？而我們又為什麼不能永生？ 我們為什麼要老死是一個複雜的問題，其具體細節不在本書的探討範圍。除各種特殊原因之外，有人提出了一些比較普遍的原因。例如有一種理論認為，衰老標誌著一個個體一生中發生的有害的複製錯誤以及其他種類的基因損傷的積累。另外一種理論為梅達沃（Peter Medawar）爵士所首創，它是按照基因選擇的概念思考進化問題的典範。他首先擯棄了此類傳統的論點：老的個體之死亡屬於對物種其他成員的一種利他主義行為。因為假如他們衰老得不能再生殖還留戀塵世，他們就會充塞世界對大家都無好處。梅達沃指出，這是一種以假定為論據的狡辯。因為這種論點，以它必須證實的情況作為假定，即年老的動物衰老得不能再生殖。這也是一種類似群體選擇或物種選擇的天真的解釋方法，儘管我們可以把有關部分重新講得更好聽一些。梅達沃自己的理論具有極好的邏輯性，我們可以將其大意綜述如下： 我們已經提出了這樣的問題，即哪些是好的基因的最普遍的特性。我們認為自私是其中之一。但成功的基因所具有的另一個普遍特性是，它們通常把它們的生存機器的死亡至少推遲至生殖之後。毫無疑問，你有些堂兄弟或伯祖父是早年夭折的，但你的祖先中一個也沒有是幼年夭折的。祖先是不會年幼喪生的。 促使其個體死亡的基因稱為致死基因。半致死基因具有某種使個體衰弱的作用，這種作用增加了由於其他因素而死亡的可能性。任何基因都在生命的某一特定階段對個體施加其最大的影響，致死和半致死基因也不例外。大部分基因是在生命的胚胎期間發生作用的，有些是在童年，有些是在青年，有些是在中年，而還有一些則是在老年。請思考一下這樣一個事實：一條毛蟲和由它變成的蝴蝶具有完全相同的一組基因。很明顯，致死基因往往被從基因庫中清除掉。但同樣明顯的是，基因庫中的晚期活動的致死基因要比早期活動的致死基因穩定得多。假如一個年紀較大的個體有足夠的時間至少進行過若干次生殖之後，致死基因的作用才表現出來，那麼這一致死基因在基因庫中將仍舊是成功的。例如，使老年個體致癌的基因可以遺傳給無數的後代，因為這些個體在生癌之前就已生殖。而另一方面，使青年個體致癌的基因就不會遺傳給很多的後代；使幼兒得致死癌症的基因就不會遺傳給任何後代。根據這一理論，年老體衰只是基因庫中晚期活動致死基因同半致死基因的一種積累的副產品。這些晚期活動的致死和半致死基因之所以有機會穿過了自然選擇的網，僅僅是因為它們是在晚期活動。 梅達沃本人著重指出的一點是：自然選擇有利於這樣一些基因，它們具有推遲其他的致死基因的活動的作用；也有利於這樣一些基因，它們能夠促進好的基因發揮其影響。情況可能是，基因活動開始時受遺傳控制的種種變化構成了進化內容的許多方面。 值得重視的是，這一理論不必作出任何事先的假設：即個體必須到達一定的年齡才能生殖。如果我們以假設一切個體都同樣能夠在任何年齡生一個小孩作為出發點，那麼梅達沃的理論立刻就能預測推斷出晚期活動的有害基因在基因庫中的積累，以及由此而導致的老年生殖活動的減少的傾向。 這裡就此說幾句離題的話。這一理論有一個很好的特點，它啟發我們去作某些相當有趣的推測。譬如根據這一理論，如果我們想要延長人類的壽命，一般可以通過兩種方式來實現這個目的。第一，我們可以禁止在一定的年齡之前生殖，如四十歲之前。經過幾世紀之後，最低年齡限制可提高到五十歲，以後照此辦理。可以想見，用這樣的方法，人類的長壽可提高到幾個世紀。但我很難想像會有任何人去認真嚴肅地制定這樣一種政策。 第二，我們可以想辦法去愚弄基因，讓它認為它所寄居的個體比實際要年青。如果付諸實踐，這意味著需要驗明隨著年紀的增大，發生在個體內部化學環境裡的種種變化。任何這種變化都可能是促使晚期活動的致死基因開始活動的提示（cues）。以倣傚青年個體的表面化學特性的方法，有可能防止晚期活動的有害基因接受開始活動的提示。有趣的是，老年的化學信號本身，在任何正常意義上講，不一定是有害的。比如，我們假設偶然出現了這種情況：一種S物質在老年個體中的濃度比在青年個體中來得高，這種S物質本身可能完全無害，也許是長期以來體內積累起來的食物中的某種物質。如果有這樣一個基因，它在S物質存在的情況下碰巧產生了有害的影響，而在沒有S物質存在的情況下卻是一個好基因，這樣的基因肯定在基因庫中自動地被選擇，而且實際上它成為一種導致年老死亡的基因。補救的辦法是，只要把S物質從體內清除掉就行了。 這種觀點的重大變革性在於，S物質本身僅是一種老年的死亡，可能認為S物質是一種有毒物質，他會絞盡腦汁去尋找S物質同人體機能失常之間的直接的、偶然的關係。但按照我們假定的例子來講，他可能是在浪費時間！ 也可能存在一種y物質，這種物質在青年個體中要比在老年個體中更集中。從這一意義上講，y物質是青春的一種標誌。同樣，那些在有y物質存在的情況下產生好的效果，而在沒有y物質存在的情況下卻是有害的基因會被選擇。由於還沒有辦法知道S物質或y物質是什麼東西可能存在許多這樣的物質我們只能作這樣的一般性的推測：你在一個老年個體中越能模仿或模擬青年個體的特點，不論這些看來是多麼表面化的特點，那個老年個體就應該生存得越久。 我必須強調指出，這些只是基於梅達沃理論的一些推測。儘管在某種意義上說，梅達沃理論在邏輯上是有些道理的，但並無必要把它說成是對任何年老體衰實例的正確解釋。對於我們現在的論題密切有關的是，基因選擇的進化觀點對於個體年老時要死亡這種趨勢，能毫無困難地加以解釋。對於個體必然要死亡的假設是本章論證的核心，它是可以在這一理論的範圍內得到圓滿解釋的。 我一筆帶過的另一個假設，即存在有性生殖和交換，更加難以解釋清楚。交換並不總是一定要發生。雄果蠅就不會發生交換。雌果蠅體內也有一種具有壓抑交換作用的基因。假定我們要飼養一個果蠅種群，而這類基因在該種群中普遍存在的話，染色體庫中的染色體就會成為不可分割的自然選擇基本單位。其實，如果我們遵循我們的定義直到得出其邏輯結論的話，就不得不把整條染色體作為一個基因。 還有，性的替換方式是存在的。雌蚜蟲能產無父的、活的雌性後代。每個這樣的後代具有它母親的全部基因（順便提一下，一個在母親子宮內的胎兒甚至可能有一個更小的胎兒在它自己的子宮內。因此，一個雌蚜蟲可以同時生一個女兒和一個外孫女，它們相當於這個雌蚜蟲自己的雙胞胎）。許多植物的繁殖是以營養體繁殖的方式進行，形成吸根。這種情況我們寧可稱其為生長而不叫它生殖。然而你如果仔細考慮一下，生長同無性生殖之間幾乎無任何區別，因為二者是細胞簡單的有絲分裂。有時以營養體方式生長出來的植物同母體分離開來。在其他情況下，如以榆樹為例，連接根出條保持完整無損。事實上，整片榆樹林可以認為是一個單一的個體。 因此，現在的問題是：如果蚜蟲和榆樹不進行有性生殖，為什麼我們要費這樣大的周折把我們的基因同其他人的基因混合起來才能生育一個嬰兒呢？看上去這樣做的確有點古怪。性活動，這種把簡單的複製變得反常的行為，當初為什麼要出現呢？性到底有什麼益處？ 這是進化論者極難回答的一個問題。為了認真地回答這一問題，大多數的嘗試都要涉及到複雜的數學上的推理。除一點外，我將很坦率地避開這個問題。我要說的一點是，理論家們在解釋性的進化方面所遇到的困難，至少在某些方面是由於他們習慣於認為個體總是想最大限度地增加其生存下來的基因的數目。根據這樣的講法，性活動似乎是一種自相矛盾的現象，因為個體要繁殖自己的基因，性是一種效率低的方式：每個胎兒只有這個個體的基因的百分之五十，另外百分之五十由配偶提供。要是他能夠像蚜蟲那樣，直接芽出（bud off）孩子，這些孩子是他自己絲毫不差的複製品，他就會將自己百分之百的基因傳給下一代的每一個小孩。這一明顯的佯謬促使某些理論家接受群體選擇論，因為他們比較容易在群體水平上解釋性活動的好處。用博德默（w．F．Bodmer）簡單明瞭的話來說，性促進了在單個個體內積累那些以往分別出現於不同個體內的有利突變。 但如果我們遵循本書的論證，並把個體看作是由長壽基因組成的臨時同盟所造成的生存機器，這一佯謬看起來就不那麼自相矛盾了。從整個個體的觀點來看，有效性就無關緊要了。有性生殖對無性生殖就被認為是在單基因控制下的一種特性，就同藍眼睛對棕色眼睛一樣。一個負責有性生殖的基因為了它自私的目的而操縱其他全部基因。負責交換的基因也是如此。甚至有一種叫作突變子的基因，它們操縱其他基因中的拷貝錯誤率。按照定義，拷貝錯誤對錯誤地拷貝出來的基因是不利的。但如果這種拷貝錯誤對誘致這種錯誤的自私的突變基因有利的話，那麼這種突變基因就會在基因庫裡擴散開。同樣，如果交換對負責交換的基因有好處，這就是存在交換現象的充分理由；如果同無性生殖相對的有性生殖有利於負責有性生殖的基因，這也就是存在有性生殖現象的充分理由。有性生殖對個體的其餘基因是否有好處，比較而言也就無關緊要了。從自私基因的觀點來看，性活動畢竟也就不那麼異乎尋常了。 這種情況非常接近於一種以假定為論據的狡辯，因為性別的存在是整個一系列推論的先決條件。而這一系列推論的最後結果認為基因是自然選擇單位。我認為是有辦法擺脫這一困境的。但本書宗旨不在於探索這一問題。性毫無疑問是存在的。這一點是真實的。我們之所以能將這種小的遺傳單位，或基因，看作是最接近於基本的和獨立的進化因素，正是性和交換的結果。 只要學會按照自私基因的理論去思考問題，性這一個明顯的佯謬就變得不那麼令人迷惑不解了。例如有機體內的ＤＮＡ數量似乎比建造這些有機體所必需的數量來得大，因為相當一部分ＤＮＡ從未轉化為蛋白質。從個體有機體的觀點來看，這似乎又是一個自相矛盾的問題。如果ＤＮＡ的目的是監造有機體，那麼，一大批ＤＮＡ並不這樣做，這是令人奇怪的。生物學家在苦思冥想地考慮，這些顯然是多餘的ＤＮＡ正在幹些什麼有益的工作呢？但從自私的基因本身的角度上看，並不存在自相矛盾之處。ＤＮＡ的真正目的僅僅是為了生存。解釋多餘的ＤＮＡ最簡單的方法是，把它看作是一個寄生蟲，或者最多是一個無害但也無用的乘客，在其他ＤＮＡ所創造的生存機器中搭便車而已。 有些人反對這種他們認為是過分以基因為中心的進化觀點。他們爭辯說，實際上生存或死亡的畢竟是包括其全部基因在內的完整個體，我希望我在本章所講的足以表明：在這一點上其實並不存在分歧。就像划船比賽中整條船贏或輸一樣，生存或死亡的確實是個體，自然選擇的直接形式幾乎總是在個體水平上表現出來。但非隨機的個體死亡以及成功生殖的遠期後果，表現為基因庫中變化著的基因頻率。對於現代複製基因，基因庫有保留地起著原始湯對於原始複製基因所起的同樣作用。性活動和染色體交換起著保持原始湯的現代相等物的那種流動性的作用。由於性活動和交換，基因庫始終不停地被攪混，使其中基因部分地混和。所謂進化就是指基因庫中的某些基因變得多了，而另一些則變得少了的過程。每當我們想要解釋某種特性，如利他性行為的演化現象時，最好養成這樣一種習慣只要問問自己：這種特性對基因庫裡的基因頻率有什麼影響？有時基因語言有點乏味，為簡潔和生動起見，我們不免要借助於比喻。不過我們要以懷疑的目光注視著我們的比喻，以便在必要時能把它們還原為基因語言。 就基因而言，基因庫只是基因生活於其中的一種新湯。所不同的是，現在基因賴以生存的方式是，在不斷地製造必將消亡的生存機器的過程中，同來自基因庫的一批批絡繹不絕的夥伴進行合作。下面一章我們要論述生存機器本身以及在某一個意義上我們可以說基因控制其生存機器的行為。