Chapter 8 — Heredity

A mango seed grows into a mango tree and a hen's egg into a chicken, never the other way about, and a child has the eyes of one parent and the chin of the other. The transmission of characters from parents to their offspring is heredity, and the branch of biology that studies it is genetics. The characters that are inherited, the shape of the nose, the colour of the eyes, blood group, the ability to roll the tongue, the colour of a flower, the length of a pea stem, are called traits, and each trait is controlled by units of inheritance carried in the gametes, the genes.

Offspring resemble their parents because they receive their genes from them, but they are not identical to them or to one another. The differences among the individuals of a species, and between offspring and parents, are variations. Look round a classroom: every student is human, yet each differs in height, complexion, face, voice and a thousand other ways. Some variations are small and continuous, such as height, which ranges smoothly from short to tall; others are sharp and discontinuous, such as blood group, where a person is A, B, AB or O and nothing in between. Variations that are inherited arise in two ways. In sexual reproduction the genes of two parents are shuffled and combined in new ways at every generation, so that each offspring receives a different half of each parent's genes; this recombination is the main source of variation among brothers and sisters. And the copying of DNA, though remarkably accurate, occasionally makes an error, a mutation, which produces a gene that did not exist before; mutations are rare and usually harmful, but they are the ultimate source of all new variation. Asexually reproducing organisms show very little variation, only that due to mutation, which is why a field of cloned bananas is uniform and a garden of pea seedlings is not.

Not every difference is inherited. A boy who lifts weights develops big muscles, a labourer's skin darkens in the sun, a tree bent by the wind grows crooked, a mouse that loses its tail in a trap has no tail, and a person who learns Telugu speaks it. These are acquired characters, produced during the life of the individual by use, environment or accident, and they affect only the body cells, not the genes in the gametes. They are therefore not passed to the offspring: the weightlifter's children are born with ordinary muscles, and the mouse's children have tails. The failure of acquired characters to be inherited was one of the decisive facts in the history of the theory of evolution, as we shall see. Only variations that are written into the DNA of the gametes, whether from recombination or mutation, are inherited, and only they matter for heredity and evolution.

Variation is not a defect; it is the reason a species survives. In a population where every individual is alike, a new disease or a change of climate that kills one kills all; where individuals vary, some will resist, and the species goes on. This chapter first explains how characters are inherited, and then how inherited variation, acted on by the environment, leads to evolution.

Gregor Johann Mendel (1822-1884) was an Austrian monk who taught science at a school in Brünn (now Brno) and grew peas in the monastery garden. Between 1856 and 1863 he crossed thousands of pea plants, counted their offspring, and worked out the rules of inheritance with a clarity that no one before him had approached. He read his results to the local natural history society in 1865 and published them in 1866, but the paper lay unnoticed for thirty-four years until three botanists rediscovered the same laws in 1900 and found that Mendel had been there first. He is called the father of genetics.

Mendel succeeded where others had failed for four reasons. First, he chose his organism well. The garden pea, Pisum sativum, is easy to grow, completes its life in one season, and produces many seeds. Its flowers are bisexual and their stamens and stigma are enclosed within the keel petals, so that each flower normally pollinates itself; Mendel could be sure that his pure lines stayed pure, and when he wanted a cross he could remove the anthers of a flower before they ripened (emasculation), dust its stigma with pollen from another plant with a brush, and bag the flower to keep out insects. Second, he chose his characters well. He studied seven contrasting pairs of traits, each of which occurred in two clearly different forms with nothing in between: tall or dwarf stem (about 2 m or 0.5 m), round or wrinkled seed, yellow or green seed (cotyledon), purple or white flower, inflated or constricted pod, green or yellow pod, and flowers along the stem (axial) or at the tip (terminal). Third, he began with pure lines: plants which, self-pollinated for several generations, always produced offspring like themselves, so that a tall pure line gave only tall plants. Fourth, and most important, he counted. Earlier breeders had described their hybrids in words; Mendel recorded the number of each kind of offspring in each generation and looked for ratios, and he studied one character at a time before combining them. The ratios he found, 3 : 1 and 9 : 3 : 3 : 1, were the key to everything.

Mendel's method was that of an experimental scientist: a testable question, a suitable organism, controlled crosses, large numbers, and mathematical analysis. He had studied physics and mathematics at the University of Vienna, and it was his mathematical habit of mind, unusual in a biologist of his day, that let him see the pattern in his counts. He also had luck: the seven characters he chose happened to lie on different chromosomes or far apart on the same one, so that they were inherited independently; had he chosen linked characters his second law would not have appeared.

The terms used in describing his crosses are these. The plants first crossed are the parental generation, P. Their offspring are the first filial generation, F₁. The offspring of the F₁ plants, when self-pollinated or crossed among themselves, are the second filial generation, F₂. A cross involving one pair of contrasting traits is a monohybrid cross; one involving two pairs is a dihybrid cross.

Mendel's first experiments each followed a single pair of contrasting traits. Take his cross of a pure tall pea plant with a pure dwarf one. He expected, as most people would, that the offspring might be of medium height, a blend of the parents. They were not. Every plant of the F₁ generation was tall, as tall as the tall parent, and it made no difference which parent supplied the pollen; the dwarf character had apparently vanished. Mendel then let the F₁ plants self-pollinate and grew the F₂ generation. The dwarf character reappeared: of 1064 F₂ plants, 787 were tall and 277 dwarf, a ratio of 2.84 : 1, which Mendel recognised as 3 : 1. The dwarf plants of F₂ bred true when selfed, giving only dwarfs; of the tall F₂ plants, one-third bred true and gave only tall, while two-thirds behaved like the F₁ and gave tall and dwarf in the ratio 3 : 1 again. He obtained the same pattern for all seven traits: in each pair one form appeared alone in F₁ and both appeared in F₂ in a 3 : 1 ratio; for seed shape, for example, 5474 round to 1850 wrinkled, 2.96 : 1.

From these results Mendel drew his first conclusions. Each trait is controlled by a pair of factors (we now call them genes, and the two alternative forms of a gene alleles), one inherited from each parent. When the two factors of a pair are different, one of them, the dominant factor, expresses itself and masks the other, the recessive factor, which is present but hidden. Tallness is dominant over dwarfness, round over wrinkled, yellow seed over green, purple flower over white, inflated pod over constricted, green pod over yellow, and axial flowers over terminal. This is the law of dominance: when two contrasting factors are present together, only the dominant one is expressed.

The convention for writing the cross uses a capital letter for the dominant allele and the same small letter for the recessive. The pure tall parent is TT, the pure dwarf tt. Each parent's gametes carry only one factor of the pair, T or t. Every F₁ plant receives a T from one parent and a t from the other and is Tt; since T is dominant, all are tall. The F₁ plants form two kinds of gametes, T and t, in equal numbers; when they self-pollinate, the gametes combine at random and the F₂ consists of TT, Tt, tT and tt in the proportion 1 : 2 : 1. The TT and Tt plants, three-quarters, are tall; the tt plants, one-quarter, are dwarf. The recessive factor was not lost in F₁; it was carried unexpressed in every Tt plant and reappeared whenever two t gametes met.

The set of factors an organism carries for a trait is its genotype (TT, Tt or tt); its outward appearance is its phenotype (tall or dwarf). An organism with two identical alleles (TT or tt) is homozygous or pure, and breeds true; one with two different alleles (Tt) is heterozygous or hybrid, and does not. The F₂ of a monohybrid cross therefore has a phenotypic ratio of 3 : 1 and a genotypic ratio of 1 : 2 : 1. Two plants of the same phenotype, tall, may differ in genotype, TT or Tt, and the only way to tell them apart is to breed from them.

The reappearance of the recessive trait in F₂, in exactly one-quarter of the plants, led Mendel to his second and most important law. The two factors of a pair, present together in the plant, do not blend or contaminate each other; they remain distinct, and when the plant forms its gametes they separate, so that each gamete receives only one factor of each pair, and receives either one with equal probability. This is the law of segregation, also called the law of purity of gametes: the gametes are pure for one factor or the other, never mixed. At fertilisation the factors pair up again at random. We now know the physical basis: the two alleles of a gene lie at the same position on the two chromosomes of a homologous pair, and these chromosomes separate into different gametes at meiosis.

Punnett square. The English geneticist R. C. Punnett devised a simple grid for working out the results of a cross. The gametes of one parent are written along the top and those of the other down the side, and each cell of the grid is filled with the combination of the row and column gametes; the cells show all possible offspring genotypes in their expected proportions. For the F₁ selfing Tt × Tt, the gametes are T and t on each side, and the four cells read TT, Tt, Tt, tt. The square makes it easy to read off the genotypic ratio 1 : 2 : 1 and the phenotypic ratio 3 : 1, and it works for any cross once the gametes are known. It gives probabilities, not certainties: among four seeds from such a cross one expects one dwarf, but any particular four may contain none or four, just as four tosses of a coin need not give two heads; only with large numbers, as in Mendel's hundreds of plants, does the ratio appear clearly.

Test cross. A tall plant may be TT or Tt, and its appearance does not tell which. Mendel's method of finding out is the test cross: the plant of unknown genotype is crossed with the homozygous recessive, tt. If the unknown is TT, all its gametes carry T, every offspring is Tt, and all are tall. If it is Tt, half its gametes carry T and half t, and the offspring are half Tt (tall) and half tt (dwarf), a 1 : 1 ratio. The appearance of even one dwarf offspring proves the unknown plant is heterozygous. Breeders use the test cross constantly to find out whether an animal or plant carrying a dominant trait will breed true. A cross of an F₁ hybrid with either parent is a back cross; the test cross is the back cross with the recessive parent.

Human examples. Mendel's rules apply to us. The ability to roll the tongue into a U is dominant over the inability; free ear lobes are dominant over attached; a widow's peak over a straight hairline; brown eyes over blue; and the ability to taste the bitter chemical PTC over the inability. Two tongue-rolling parents, each Rr, can have a child who cannot roll the tongue (rr), with a probability of one in four at each birth, and puzzled parents sometimes wonder how. Many inherited diseases are recessive: thalassaemia, sickle-cell anaemia and albinism appear in a child only when both parents are carriers (heterozygous), which is why they are commoner in communities where cousins marry, since relatives are more likely to carry the same hidden allele. Some conditions are dominant, such as polydactyly (extra fingers) and Huntington's disease, and appear whenever one allele is present.

Having settled how one trait is inherited, Mendel asked whether two traits are inherited together or separately. He crossed a pure pea plant with round yellow seeds (both dominant) with a pure plant with wrinkled green seeds (both recessive). All the F₁ seeds were round and yellow, as the law of dominance predicts. He then selfed the F₁ plants and examined 556 F₂ seeds. He found four kinds: 315 round yellow, 108 round green, 101 wrinkled yellow and 32 wrinkled green, a ratio very close to 9 : 3 : 3 : 1. Two of the four kinds, round green and wrinkled yellow, were new combinations not seen in either parent.

The ratio told Mendel two things. First, if he looked at each trait alone, the old 3 : 1 ratio was still there: 315 + 108 = 423 round to 101 + 32 = 133 wrinkled, about 3 : 1; and 315 + 101 = 416 yellow to 108 + 32 = 140 green, again 3 : 1. Each pair of factors was segregating exactly as in a monohybrid cross. Second, the two pairs were being inherited independently of each other: the factor for seed shape went into a gamete without regard to which factor for seed colour went with it, so that all four combinations of gametes, RY, Ry, rY and ry, were formed in equal numbers. If the two pairs are independent, three-quarters of the seeds should be round and, of those, three-quarters yellow, giving 3/4 × 3/4 = 9/16 round yellow; 3/4 × 1/4 = 3/16 round green; 1/4 × 3/4 = 3/16 wrinkled yellow; and 1/4 × 1/4 = 1/16 wrinkled green. That is the 9 : 3 : 3 : 1 ratio. This is Mendel's third law, the law of independent assortment: when two or more pairs of contrasting factors are followed together, each pair segregates independently of the others, and the factors combine at random in the gametes.

The cross is written thus. Parents RRYY (round yellow) × rryy (wrinkled green). The gametes are RY and ry respectively; the F₁ is RrYy, round yellow. The F₁ forms four kinds of gametes, RY, Ry, rY, ry, each with probability one-quarter. A 4 × 4 Punnett square with these gametes along the top and side has sixteen cells, and counting the phenotypes gives 9 round yellow (containing at least one R and one Y), 3 round green (R with yy), 3 wrinkled yellow (rr with Y), 1 wrinkled green (rryy). There are nine different genotypes among the sixteen cells: RRYY, RRYy, RrYY, RrYy, RRyy, Rryy, rrYY, rrYy, rryy, in the proportion 1 : 2 : 2 : 4 : 1 : 2 : 1 : 2 : 1. Only one cell in sixteen, RRYY, breeds true for both dominant traits, and only one, rryy, for both recessive.

Independent assortment is why sexual reproduction produces so much variation: with just two pairs of genes a hybrid makes four kinds of gametes, with three pairs eight, and with the roughly 20,000 pairs of a human, more combinations than there are atoms in the body. It holds strictly only for genes on different chromosomes, or far apart on the same chromosome; genes close together on one chromosome tend to be inherited together, a phenomenon called linkage that was discovered after Mendel. Mendel's seven characters happened not to show linkage, which was his good fortune.

Mendel spoke of factors without knowing what they were. In the decades after his rediscovery the factors were located, named and finally analysed chemically, and heredity became a science of molecules.

Chromosomes. Inside the nucleus of every cell are thread-like bodies that take up dyes strongly and so are called chromosomes (coloured bodies). They are visible only when the cell divides. Every species has a fixed number: the garden pea has 14, the fruit fly 8, the onion 16, rice 24, the dog 78 and the human 46. In body cells the chromosomes occur in pairs, the two members of a pair being alike in size and shape and carrying genes for the same traits; the human has 23 pairs, and the paired condition is called diploid (2n). One chromosome of each pair came from the father's sperm and one from the mother's egg. When gametes are formed by meiosis, the members of each pair separate, so that a gamete has only one chromosome of each pair, 23 in humans, and is haploid (n). Fertilisation restores the diploid number. This is precisely the behaviour Mendel deduced for his factors, and in 1902 Sutton and Boveri pointed out that the factors must be carried on the chromosomes. The independent separation of different chromosome pairs at meiosis is the basis of independent assortment, and the separation of the two members of a pair is the basis of segregation.

Genes. A gene is a segment of a chromosome that controls one trait, or more precisely that carries the instructions for making one protein, and it is the protein that produces the trait. Genes are arranged along the chromosome in a fixed order, each at its own position, its locus, like beads on a string; a human chromosome carries hundreds to thousands of genes, and the whole set of about 20,000 to 25,000 is the genome. The two alleles of a gene, such as T and t, occupy the same locus on the two chromosomes of a pair. Thomas Hunt Morgan, working with the fruit fly Drosophila from 1910, proved that genes lie on chromosomes and mapped their positions, and showed that genes on the same chromosome tend to be inherited together (linkage).

DNA. The chromosome is made of protein and a long molecule of deoxyribonucleic acid, DNA, and it is the DNA that carries the genetic information. In 1953 Watson and Crick showed that DNA is a double helix, two strands twisted round each other like a spiral staircase, each strand a chain of units called nucleotides. Each nucleotide carries one of four bases, adenine (A), thymine (T), guanine (G) and cytosine (C), and the bases of the two strands pair across the helix, A always with T and G always with C. The sequence of bases along a strand is the information; a gene is a stretch of a few hundred to many thousand bases, and the order of its bases, read in groups of three, specifies the order of amino acids in a protein. Because A pairs only with T and G only with C, each strand is a template for the other: when a cell divides, the helix unzips and each strand builds its partner, giving two identical copies, one for each daughter cell. This replication is how genes are copied for every cell and every gamete, and an error in it is a mutation.

Mendel's abstract factor is thus a length of DNA at a fixed place on a chromosome, whose base sequence dictates a protein. A dominant allele typically makes a working protein; a recessive allele often makes a faulty or absent one, so that a single dominant copy is enough for the trait to appear, which is why Tt is tall. The recessive pea allele for wrinkled seeds, for instance, is a broken gene for an enzyme that makes starch; without enough starch the seed loses water and wrinkles.

Of the 23 pairs of human chromosomes, 22 pairs are alike in males and females and are called autosomes. The 23rd pair is different in the two sexes and its members are the sex chromosomes. A woman has two similar sex chromosomes, called X, and her chromosome set is written 44 + XX. A man has one X and a much smaller, differently shaped chromosome called Y, and his set is 44 + XY. The X chromosome is large and carries many genes, including some for characters that have nothing to do with sex, such as colour vision and blood clotting; the Y is small and carries few genes, but among them is the gene that, when present in the embryo, switches on the development of testes. An embryo with a Y becomes male; without it, it becomes female.

How sex is decided. When a woman forms ova by meiosis, the two X chromosomes separate and every ovum receives one X; a woman therefore produces only one kind of gamete with respect to sex, 22 + X. When a man forms sperms, the X and Y separate and half the sperms receive the X and half the Y; he produces two kinds of gamete, 22 + X and 22 + Y, in equal numbers. At fertilisation, if an X-bearing sperm fertilises the ovum the zygote is XX and develops into a girl; if a Y-bearing sperm fertilises it the zygote is XY and develops into a boy. Since the two kinds of sperm are equal in number and have an equal chance of reaching the ovum, the probability of a boy or a girl at each conception is one-half, and a Punnett square of XX × XY gives XX, XX, XY, XY, a 1 : 1 ratio. In practice about 105 boys are born for every 100 girls, but the numbers are nearly equal, as this mechanism predicts.

Two consequences deserve emphasis. First, the sex of the child is determined by the father's sperm, not by the mother, who can give only an X. A woman blamed by her family for bearing daughters is being blamed for something that depends entirely on the husband's gamete, and this simple biological fact is one of the strongest arguments against the prejudice that lies behind female foeticide. Second, the sex of each child is a matter of chance at each conception, independent of the previous children: a family with three daughters has exactly the same one-in-two chance of a son at the next birth as any other family. No diet, ritual, timing or medicine can change which sperm reaches the egg.

Other animals. The XY system is found in all mammals and in the fruit fly. In birds, butterflies and some reptiles and fish it is reversed: the female has two different sex chromosomes (ZW) and the male two alike (ZZ), so that the mother's gamete decides the sex. In grasshoppers and some other insects there is no Y at all: the male has one X (XO) and the female two. In honeybees sex depends on fertilisation itself: fertilised eggs, with two sets of chromosomes, become females (queens and workers), and unfertilised eggs, with one set, become males (drones). In many reptiles, including crocodiles and turtles, sex is decided not by chromosomes but by the temperature at which the egg is incubated, warm nests giving one sex and cool nests the other. The human XY system is thus one of several solutions, and in all of them the ratio of the sexes is set by the mechanism, not by the wishes of the parents.

Heredity explains why offspring resemble their parents; the second half of this chapter asks what happens to inherited variation over many generations. The answer is evolution: the gradual change in the characters of a population of organisms over generations, which over long periods produces new species and has produced, from the first simple cells some 3,500 million years ago, every living thing on Earth. Evolution needs two things: variation to work on, and a process that changes which variations become common.

Where variation comes from. As seen at the start of the chapter, inherited variation arises by mutation and recombination. A mutation is a change in the DNA: a base substituted, deleted or added in a gene, a piece of chromosome lost or duplicated, or a whole chromosome gained or lost. Mutations happen spontaneously through errors in replication, at a rate of about one in a thousand million bases per copying, and are increased by radiation (X-rays, ultraviolet light, radioactivity) and by certain chemicals (mustard gas, the chemicals of tobacco smoke). A mutation in a body cell affects only that individual and may cause cancer; a mutation in a gamete is passed to the offspring and enters the population. Most mutations are neutral or harmful, since a random change in a working machine seldom improves it, but occasionally one confers an advantage, and every allele that now exists began as a mutation. Recombination in sexual reproduction, through the independent assortment of chromosomes, the crossing over of segments between homologous chromosomes at meiosis, and the random meeting of gametes, shuffles existing alleles into new combinations at every generation, so that the same set of alleles can produce an almost unlimited variety of individuals. Mutation supplies the raw material; recombination spreads it into every possible combination.

What the environment does with it. In every population more offspring are produced than can survive: a pair of mice could fill the earth in a few years if all their descendants lived. Which individuals survive and reproduce is not entirely a matter of chance. Those whose inherited variations fit them better to their environment, to find food, escape enemies, resist disease, tolerate cold or attract mates, leave more offspring, and those offspring carry the same variations. Generation by generation the favourable alleles become commoner in the population and the unfavourable rarer. This is natural selection, the process that Charles Darwin identified as the main cause of evolution, and it is described in detail in the next topic. Its result is adaptation: the fit between an organism and its way of life, the streamlined body of a fish, the long neck of a giraffe, the camouflage of a stick insect, the drought-resistance of a cactus, each the outcome of many generations of selection among inherited variations.

Chance also acts. In small populations the frequency of an allele can change from generation to generation by accident, regardless of its usefulness, simply because the few individuals who happen to breed carry more or less of it; this is genetic drift. A storm that kills most of a small island population, or a few individuals that found a new colony, can fix or eliminate an allele by chance alone. Drift produces differences between isolated populations that have nothing to do with adaptation, and it is why small populations of endangered species lose their variation and become vulnerable.

The examination often asks for the distinction between variations that are inherited and those that are not: evolution acts only on inherited variation, on alleles in gametes, never on acquired characters, however useful.

That species change over time was suspected by several naturalists before 1800, but the first full theory was proposed by the French biologist Jean-Baptiste Lamarck in 1809, and the theory accepted today was proposed by Charles Darwin in 1859. Comparing them is a standard examination question, and it shows why the inheritance of acquired characters matters.

Lamarck's theory. Lamarck held that organisms change in response to their needs. An organ that is used more becomes stronger and larger, and one that is not used weakens and shrinks: the law of use and disuse. He then supposed that these changes, acquired during the animal's life, are passed on to its offspring: the inheritance of acquired characters. His famous example is the giraffe: its ancestors had short necks, but as they stretched to browse on higher branches their necks lengthened a little in each lifetime, the lengthening was inherited, and over generations the long neck evolved. Similarly he explained the webbed feet of water birds by the stretching of the skin in swimming, and the blindness of moles by the disuse of eyes underground. Lamarck deserves credit for seeing that species are not fixed and that the environment shapes them, but his mechanism is wrong. Acquired characters are not inherited: as we saw, the sons of blacksmiths are not born with strong arms, and August Weismann cut off the tails of mice for twenty-two generations and every mouse was still born with a full tail. A character reaches the next generation only if it is in the DNA of the gametes, and stretching the neck does not alter the DNA of the sperm or egg.

Darwin's theory. Darwin sailed round the world on HMS Beagle from 1831 to 1836 as the ship's naturalist, and what he saw, especially the finches of the Galapagos Islands, each island with its own species differing in beak, convinced him that species change. Back in England he read Malthus on the tendency of populations to outgrow their food, and saw how breeders improve pigeons and cattle by selecting the individuals they want. He spent twenty years assembling evidence and, prompted by a letter from Alfred Russel Wallace who had reached the same idea, published On the Origin of Species in 1859. His theory of natural selection has four steps. (1) Overproduction: every species produces far more offspring than can survive. (2) Variation: the individuals of a species vary, and much of the variation is inherited. (3) Struggle for existence: because resources are limited, the offspring compete with one another, with other species and with the environment, and most die before reproducing. (4) Survival of the fittest: individuals whose variations suit them better to their conditions are more likely to survive and reproduce, and they pass those variations to their offspring; over generations the favourable variations accumulate and the species changes. Applied to the giraffe: the ancestral population varied in neck length; when food was scarce the longer-necked animals reached more leaves, survived and bred, while the shorter-necked starved; their offspring inherited the longer necks; and the process repeated until the neck reached its present length. The neck was not stretched by effort; the long-necked were selected by hunger.

The difference is that Lamarck makes the environment cause the variation, in a direction that suits the animal, whereas Darwin has the variation arise first, at random, and the environment then select among it. Darwin did not know the source of variation, which was the weakness of his theory in his own day; the rediscovery of Mendel and the discovery of mutation supplied it, and the combined theory, natural selection acting on Mendelian variation from mutation and recombination, is called neo-Darwinism or the modern synthesis. It is the foundation of all modern biology.

A theory about events over millions of years cannot be tested by direct observation; it is tested by asking what it predicts and looking for the traces it should have left. Evolution predicts that the organisms of the past were different from those of today, that they became more like modern forms as time went on, and that transitional forms once existed between major groups. The rocks confirm all three.

What fossils are. A fossil is any preserved remains or trace of an organism that lived in the past: bones, teeth and shells turned to stone, impressions of leaves and feathers, footprints, burrows, insects trapped in amber, and whole mammoths frozen in ice. Most organisms decay and leave nothing; fossils form only when an organism is buried quickly in sediment, mud, sand or volcanic ash, before it rots, and the hard parts are slowly replaced by minerals as the sediment hardens into rock. Fossils are therefore commonest in sedimentary rocks laid down under water, and hard-bodied animals are far better represented than soft-bodied ones.

What fossils show. Sedimentary rocks form in layers, the older below and the younger above, so a cliff or a quarry is a record of time from bottom to top. Reading the record, one finds the simplest organisms, single-celled bacteria and algae, in the oldest rocks, about 3,500 million years old; the first animals with shells about 540 million years ago; fish before amphibians, amphibians before reptiles, reptiles before birds and mammals, and flowering plants late; and humans only in the topmost layers, within the last two million years. No fossil has ever been found out of place, a rabbit in rocks of the age of fishes, for example, though such a find would overturn the theory at once. The age of a fossil is estimated from its position in the layers, and measured precisely by radioactive dating: certain elements in rock decay at a known rate, so the proportion remaining gives the time since the rock formed; carbon-14 dates organic remains up to about 50,000 years, and uranium and potassium isotopes date rocks of millions of years.

Transitional forms. Evolution predicts intermediate forms, and they have been found. Archaeopteryx, from rocks 150 million years old in Germany, had feathers and wings like a bird but teeth, a long bony tail and clawed fingers like a reptile, a link between reptiles and birds. The evolution of the horse is traced through a series of fossils from Eohippus, a fox-sized animal with four toes on the front feet living 55 million years ago in forests, through forms of increasing size with fewer toes and longer legs and teeth suited to grass, to the modern one-toed horse of open plains. The whale is traced from four-legged land mammals through forms with shrinking hind legs to the fully aquatic animal, whose skeleton still has tiny useless hip bones. Fossils of early humans and their relatives, described later, trace our own line. The fossil record is incomplete, since fossilisation is rare, but everything found in it agrees with evolution and nothing contradicts it.

In India, the Siwalik hills of the Himalayan foothills have yielded fossils of ancient elephants, hippopotamus, giraffes and apes from the last 20 million years, the Narmada valley has given one of the oldest human fossils of the subcontinent, and the Deccan traps preserve dinosaur eggs and bones from about 66 million years ago, the time when the dinosaurs died out.

Besides fossils, evolution has left its evidence in the bodies of living organisms, which show the marks of common ancestry beneath the differences of adaptation.

Homologous organs. The forelimb of a human, the foreleg of a horse, the wing of a bat, the flipper of a whale and the wing of a bird are used for utterly different purposes, grasping, running, flying, swimming, and look different, yet each is built on the same plan: one long bone (humerus), then two (radius and ulna), then a cluster of wrist bones, then five digits, sometimes reduced or fused. Organs with the same basic structure and origin in the embryo, but different functions, are homologous. There is no reason a wing and a flipper should share this pattern unless they were inherited from a common ancestor that had it and were then modified in different lines for different uses. Homology thus shows divergent evolution: one ancestral structure becoming many. Plants show it too: the thorn of bougainvillea and the tendril of the cucumber are both modified branches; the leaves of a pea, the spines of a cactus and the scales of an onion bulb are all leaves.

Analogous organs. The wing of a bird and the wing of a butterfly both serve for flight, but the bird's wing is a limb with bones, muscles and feathers and the insect's is a fold of the body wall with no bones at all; they have nothing in common but their function. Organs with the same function but different structure and origin are analogous. They show convergent evolution: unrelated organisms, subjected to the same conditions, evolve similar solutions. The streamlined bodies of a shark (a fish), an ichthyosaur (an extinct reptile) and a dolphin (a mammal) are analogous; so are the eyes of an octopus and a man, and the underground stem of the potato and the root of the sweet potato. Analogy warns that resemblance alone does not prove relationship; homology reveals it.

Vestigial organs. Many organisms carry structures that are reduced and useless to them but were functional in their ancestors. The human vermiform appendix is a shrunken remnant of the large caecum in which our plant-eating ancestors digested cellulose; the coccyx is the stump of a tail; the muscles that move the ears, the wisdom teeth, the nictitating membrane in the corner of the eye and the goose-flesh reflex are others. The python has tiny hind-limb bones under its skin, the whale has hip bones, flightless birds have wings, and cave fish have eyes under skin. Such organs make no sense as designs but are exactly what descent with modification predicts.

Embryology. The early embryos of a fish, a frog, a tortoise, a chicken, a rabbit and a human are so alike that an expert has difficulty telling them apart: all have a tail, gill slits in the neck region, and a similar heart and blood vessels. In the fish the gill slits become gills; in the others they close and are remodelled into parts of the jaw, ear and throat, and the human tail is reabsorbed. These shared stages are the inheritance of a common ancestral development, modified late in each line. Ernst Haeckel expressed this as the idea that the development of an individual repeats the evolution of its group; stated so strongly it is not correct, but the similarity of early embryos is real and strong evidence of common descent.

Molecular evidence. The strongest evidence of all is the newest. All living things use the same DNA, the same genetic code and the same twenty amino acids, and the same basic proteins, which could only be so if all descend from one ancestral line. The degree of difference in the DNA or in a protein such as haemoglobin or cytochrome between two species matches the relationship expected from other evidence: human and chimpanzee DNA differ by little more than one percent, human and monkey by more, human and mouse by more again, and the sequence differences accumulate at a roughly steady rate, a molecular clock that dates the branching of lineages. The classification of living things that Linnaeus built from structure and the family tree that molecular biology builds from DNA agree, and this agreement between independent lines of evidence is what makes evolution one of the most secure theories in science.

Speciation. A species is a group of organisms that can breed among themselves and produce fertile offspring, and cannot do so with members of other groups. Evolution within a population makes it better adapted; the formation of a new species, speciation, requires that a population split into two that can no longer interbreed. It usually happens in steps. First a population is divided by a barrier, a river, a mountain range, a sea, a stretch of desert, or the rising of an island, so that the two parts no longer exchange genes; this is geographical isolation. Then the two parts diverge: natural selection fits each to its own conditions, which differ, and genetic drift adds chance differences, while different mutations accumulate on either side. After many generations the two populations differ so much in appearance, behaviour, breeding season or chromosomes that, even if the barrier disappears and they meet again, they cannot or will not interbreed, or their hybrids are sterile like the mule; this is reproductive isolation, and at that point they are two species. The Galapagos finches, the different monkeys on either side of the Amazon and the fish species of isolated crater lakes all show speciation by isolation. In plants, a doubling of the chromosome number can produce a new species in a single generation, since the doubled plant cannot breed with its parents; bread wheat, cotton and tobacco arose this way. Speciation, repeated over millions of years, is how the two million or more species now living arose from a single origin, and the branching tree of species that classification describes is the family tree of life.

Artificial selection. Humans have caused evolution deliberately, by choosing which animals and plants breed. This is artificial selection, and Darwin took it as the model for natural selection. All the breeds of dog, from the Great Dane to the Chihuahua, were derived from the wolf in a few thousand years by selecting pups with the desired size, coat and temperament. All the varieties of wild cabbage were produced by selecting different parts of one Mediterranean plant: selection for leaves gave cabbage and kale, for swollen stem gave kohlrabi, for flower buds gave cauliflower and broccoli, and for lateral buds gave Brussels sprouts, so that a single species has been shaped into vegetables that look nothing alike. The high-yielding rice, wheat and maize of the green revolution, the milk yield of the Jersey cow and the Murrah buffalo, the fine wool of the Merino sheep and the size of the domestic hen's egg are all products of selection over generations. Artificial selection works fast because the selector is single-minded, and it demonstrates on a human time scale what natural selection achieves over a geological one: that selecting among inherited variation changes a species without limit.

Human evolution. Humans belong to the primates, the order that includes lemurs, monkeys and apes, and share a common ancestor with the chimpanzee about six to seven million years ago; we did not descend from chimpanzees, and they are our cousins, not our ancestors. The fossils of our line, found mostly in Africa, show a sequence. Australopithecus, living in Africa four to two million years ago, walked upright but had a brain little larger than a chimpanzee's, about 450 cubic centimetres; the famous skeleton called Lucy is one. Homo habilis, about two million years ago, had a brain of 600 to 700 cubic centimetres and made the first crude stone tools. Homo erectus, from 1.8 million years ago, had a brain of about 900 to 1000 cubic centimetres, made hand axes, used fire, and spread out of Africa to Asia, where fossils were found in Java and China. The Neanderthals, living in Europe and western Asia until about 40,000 years ago, had brains as large as ours, buried their dead and made fine tools, but were a separate branch. Homo sapiens, our own species, appeared in Africa about 300,000 years ago with a brain of about 1350 to 1400 cubic centimetres, and spread across the whole earth from about 70,000 years ago, reaching India early on the way east; molecular evidence shows that all living humans descend from that African population, and that the differences between human groups are trivial and recent. The trends in the human line are an upright gait, a larger brain, a smaller face and jaw, smaller teeth, a more skilful hand, and the growth of tool-making, fire, language and culture. There is only one human species today, and every human being belongs to it.