Mutants: On Genetic Variety and the Human Body

The clinical and developmental genetic literature is both vast and growing. However, I have attempted to give a guide to where the major results can be found and, occasionally, further details on particular topics. Beyond these notes, the most important source for those seeking further information about particular genetic disorders is Mendelian Inheritance in Man, an on-line database authored and edited by Victor A. McKusick and his colleagues at Johns Hopkins University, Baltimore, and supported by the National Center for Biotechnology Information, USA. It contains continually updated descriptions of each disorder, the mutations that cause them, and the clinical genetic literature. MIM can be found at http://www3.ncbi.nlm.nih.gov/Omim/. To assist those who wish to penetrate this difficult literature I give, for each syndrome and gene, the MIM numbers in bold so: achondroplasia (100800) is caused by mutations in FGFR3 (134934). Neither MIM nor this book should be used for self-diagnosis.

PROLOGUE

xiii Genetics, to quote one popular writer. Steve Jones, whose book The language of the genes (1993), HarperCollins, London, remains the best popular account of human population genetics.

xiv On 15 February 2001. The sequence of the human genome (International Sequencing Consortium 2001).

xv To learn from animals. See Gilbert (2000) p.361 for Leonardo’s cow placenta and Needham (1959) p.65 for Cleopatra’s alleged studies of human development.

CHAPTER I: MUTANTS

3 We had heard that a monster had been born at Ravenna. The monster of Ravenna has been much discussed. See Landucci (1542; 1927) pp.249–50 for a contemporary account of the monster. Jean Céard dicusses its evolution in his edition of Pare’s Des monstres (1573; 1971) PP.153–5; Niccoli (1990) pp.35–51 its political meaning; see also Fischer (1991) pp.54–6 and Daston and Park (1998) pp. 177–82. I suggest that the monster’s disorder is Roberts’s syndrome (268300), but others (Walton et al. 1993; Martinez-Frias 1993) have suggested cyclopia, sirenomelia or else hydrocephalus. All these diagnoses are guesses – which one you favour is a matter of which depiction of the monster you use, and which of its many odd features you believe are real.

6 In the sixteenth and seventeenth centuries. My description of Renaissance teratology is indebted to Park and Daston (1981) and Daston and Park (1998), though it has perhaps a more Whiggish flavour than theirs. See Boaistuau (1560, 2000) for a reproduction of an unusually beautiful teratological manuscript of the time, Melancthon and Luther (1523; 1?23) for the Monk-calf, and Paré (1573; 19?2) pp.3–4 for his list of the causes of deformity. For nineteenth-century views on maternal impressions see Gould and Pyle (1?97) and Bondeson (1997) pp.144–69. For the seventeenth-century teratologists see Aldrovandi (1642); the first edition of Liceti’s De monstrorum was published in 1616, but I have worked from the second (1634) edition, a synopsis and French translation of which is given by Houssay (1937). A brief account of Liceti’s life and work is given by Bates (2001).

8 There is a moment in time. For Sir Thomas Browne’s views on deformity in Religio medici (1654) see his Works (1904) volume 1, pp.26–7. For the shift in opinion of monsters from wrath to wonders of God see Park and Daston (19?1). For William Harvey’s writings on generation see Anatomical Exercises on the generation of animals; to which are added, essays on parturition; on the membranes, and fluids of the uterus; and on conception (1650) in his Works (1965). The quote, however, is from the 1653 translation by Martin Llewellyn as given by Needham (1959) p.134.

10 It was, however, a contemporary of Harvey’s. For Bacon’s division of science see Bacon (1620, 2000) pp.14?–9 and 223–4. For an account of Renaissance collections of marvels see Daston and Park (199?) pp.255–301. Harvey, John Aubrey tells us, thought little of Bacon as either a philosopher or a writer, but Harvey espoused very Baconian sentiments when he wrote: ‘Nature is nowhere accustomed more openly to display her secret mysteries than in cases where she shows tracings of her workings apart from the beaten path.’

13 Most of these people have mutations. On-line Mendelian Inheritance in Man lists about a thousand genes that cause phenotypic variation, be it pathological or not (e.g. brown eyes).

16 If there is no such thing as a perfect or normal genome. The estimate of how many times each of the genome’s base-pairs have mutated in the last generation alone is given by Kruglyak and Nickerson (2001). The estimate of 65 per cent of genes as having polymorphisms applies to alleles defined by non-synonymous polymorphisms only. Conversely, my claim that most genes have an overwhelmingly common variant comes from the observation that 35 per cent of genes are monomorphic, and that in the known polymorphic ones, the minor alleles usually have a frequency below 5 per cent. Again, this applies to non-synonymous polymorphisms only (Cargill et al. 1999; Stephens et al. 2001).

18 Each embryo has about a hundred mutations. Eyre-Walker and Keightley (1999) estimate the rate of production of deleterious mutations in humans. Their estimates are consistent with those from Cargill et al. (1999) and Stephens et al. (2001) obtained by other means. Crow (2000) reviews the fitness effect of novel mutations.

CHAPTER II: A PERFECT JOIN

25 The Parodis arrived in Paris. Contemporary accounts of Ritta and Christina Parodi are given by Anon (1829 a; b; c); Saint-Ange (1830); Janin (1829) and Danerow (1830). Later accounts by Thompson (1930; 1996) p.84 and Bondeson (2000) pp.168–73.

26 The first cut exposed the ribcage. The major anatomical monograph on Ritta and Christina is Serres (1832). É. Geoffroy Saint-Hilaire (1829) considers the girls in a small paper and I. Geoffroy Saint-Hilaire (1832–37) in volume 3 pp. 161–74 of his synoptic teratology.

27 The oldest known depiction. The Anatolian statue, from the Catal Hüyük site, dates from around 6500 BC; the Australian rock carving from 3–4000 BC. For the Molionides brothers, and a more general discussion on con joined twins in ancient Greek art, science and myth, see Dasen (1997; 2002). For Renaissance teratologies see Paré (1573; 1982) and Boaistuau (1560; 2000) pp.134–7. For the Montaigne quote see Montaigne (1603; 1998), and for their intellectual context Daston and Park (1998) pp.205–7. The conflict between Duverney and his rivals is discussed by Fischer (1991) pp.71–4 and Wilson (1993) pp.150–9. For the intellectual context of preformationism and epigenesis see Needham (1959) chapters 3 and 4, and Pinto-Correia (1997).

32 What makes twins conjoin? Conjoined twins occur at a frequency of 1 in 100,000 live births; monozygotic separate twins occur at a frequency of 1 in 300 live births. For Aristotle on conjoined twins see The generation of animals in his Complete works volume 1 pp.1192–1996. See Friedman (1981) pp.180–1 on baptising conjoined twins.

33 Until recently, the origin of conjoined twins. For a typical medical embryology textbook account of conjoined twinning see Sadler (2000) p. 155. Although most conjoined twins seem to be monozygotic (they are nearly always of the same sex) there is at least one case that has been shown, by genetic tests, to be the result of a fusion between dizygotic embryos (Logroño et al. 1997). For the sex ratios of conjoined twins see Steinman (2001 a; b). Spencer (2000 a; b; 2001) gives a detailed critique of the fission model of conjoined twinning based on the geometry of the joins. See Martin (1880) pp. 153–69 for the evolution of theories of the causes of conjoined twins.

35 On the seventh day. For a description of early human embryogenesis see Beddington and Robertson (1999) and Sadler (2000).

37 In the spring of 1920. For the Hilda Mangold (nee Pröscholdt) paper see Spemann and Mangold (1924); for a translation and commentary see Willier and Oppenheimer (1964); for her biography see Hamburger (1988) and Fässler and Sander (1996).

39 For seventy years. For a brief history of the search for the organiser molecules see Gilbert (2002) A selective history of induction. http://zygote.swath-more.edu/. Spemann quoted in Gilbert (2000).

40 It would be tedious to recount. ‘Noggin’ is slang for ‘head’. For the initial identification of noggin (602991) see Lamb et al. (1993); for contemporary commentary see Baringa (1993); for a textbook survey of the organiser see Gilbert (2000) pp.303–38; and for a recent technical review see Beddington and Robertson (1999). The number of molecules involved in cell–cell communication includes both signalling molecules and their receptors (International Sequencing Consortium 2001). For the antagonism between BMP4 and chordin (603475) and noggin see Zimmerman et al. (1996) and Piccolo et al. (1996). For the noggin-defective mouse see McMahon et al. (1998); for the noggin and chordin double-defective mouse see Bachiller et al. (2000).

45 When Eng and Chang. The ‘two organiser’ theory is sometimes called the crowding model to distinguish it from the fission and fusion models (J.-F. Oostra, pers. comm.). Most fusion models postulate separate embryonic discs. My model is very similar to that of Hamburger (1947). It also seems similar to that of Spencer (2000 a; b; 2001) though she is ambiguous as to whether conjoined twins arise from one or two embryonic discs. Although, as stated, the vast majority of conjoined twins have a single amnion and placenta, there is apparently evidence that some have two amnions or even – truly strange this – two placentas. The ‘two organiser’ model would not apply to such twins. For chemical induction of conjoined twins (and mono-amnion monzygotic twins) see Kaufman and O’Shea (1978).

46 One man who thought deeply. For a history of artificial incubation see Needham (1959) pp.22–5 and 203–4. For Geoffroy’s attempts to artificially create monstrous chickens see Geoffroy Saint-Hilaire (1825); Fischer (1972) and Appel (1987) pp.121–9. See also his son’s account of the influence of these experiments (Geoffroy Saint-Hilaire 1847). Étienne’s major teratological work is Geoffroy Saint-Hilaire (1822) Philosophie Anatomique des monstruosités humaines. The classificatory work was carried on by Isidore in Geoffroy Saint-Hilaire (1832–37) Histoire générale et particulière des anomalies de l’organisation chez I’homme et les animaux. See Morin (1996) for a modern evaluation of Geoffroyean teratology. For some of Geoffroy’s intellectual background see Appel (1987) pp.121–9; for a paen by Geoffroy to Bacon see Geoffroy Saint-Hilaire (1825).

49 Geoffroy was deeply enamoured. Isidore Geoffroy Saint-Hilaire (1847) claimed that his father first conceived of the soi pour soi in 1826, several years before Ritta and Christina appeared. A full, late statement of the law’s implications can be found in É. Geoffroy Saint-Hilaire (1838); see also Appel (1987).

51 The ability of disparate organ primordia. For a textbook review of cell adhesion molecules see Alberts et al. (1994) pp.950–1006. Spina bifida (182940) and anencephaly (206500) are both neural defects caused by unknown genetic factors and many environmental ones (Corcoran 1998). For bifid heart see Gilbert (2000) p.474.

52 The power of cell–cell adhesion. Parasitic conjoined twins have an occurrence of one in a million live births. See Serres (1832) for a description of these cephalothoracoileopagus conjoined twins. See Spencer (2000 a; b; 2001) for a review of conjoined twins and parasites. Park and Daston (1981) and Bondeson (2000) pp.vii–xix tell the story of Lazarus Colloredo; see Thompson (1930; 1994) p.93 for Laloo; Ta-Mei et al. (1982) for the Chinese man and Rodriguez (1870) for the multiple parasites. Spencer (2001) persuasively argues the foetus-in-foetu-teratoma theory. Naudin ten Cate (1995) describes the twenty-one-foetus teratoma.

56 In recent years, much has been learned. The story of the old soldier is told by Geoffroy Saint-Hilaire (1832–37) volume 3 pp. 8–11; Martin (1880) p.147, and Fischer (1991) p-74. Geoffroy and Martin claim that the old soldier was also the inspiration for Molière’s he Médicin malgré lui (Doctor in spite of himself), a comedy which turned on a doctor who diagnosed a patient’s heart as being on the right side of her chest. Appealing though this is, it cannot be true since Molière composed his comedy in 1666–67, Les Invalides was built in 1671, and the old soldier died in 1688 (J.-L. Fischer, pers. comm.).

56 It is a diagnosis that allows. Kartagener’s syndrome or primary ciliary dyskinesia (244400) is caused by recessive mutations in the DNAI1 (604366) and DNAH5 (603335) genes (Guichard et al. 2001 and Olbrich et al. 2002). See Afzelius (1976) for the causal link to cilia; see Kosaki and Casey (1998), Casey and Hackett (2000) and Brueckner (2001) for recent reviews.

58 I said earlier that the organiser is. For ciliary flow around the node see Nonaka et al. (1998). Of course, in a sense the discovery of asymmetrical ciliary movement doesn’t absolutely solve the problem of the embryo’s handedness, since it merely raises the question of why cilia should beat from right to left rather than the other way around. There’s no good answer to this question in mammals at least, but it is worth noting that the embryo is a three-dimensional object, and once two axes (head to tail and back to belly) have been established, the last (left to right) obtains automatically. So a cell that wishes to tell left from right merely needs to know which way its cilia protrude with respect to the rest of the embryo’s geometry. For some conceptual models of symmetry-breaking see Lander et al. (1998).

59 There is a lovely experiment that proves this. The original path-breaking paper defining the signals that determine left–right asymmetry in chicks is Levin et al. (1995). Since then a large literature has accumulated on left–right signalling in all major vertebrate model systems – and it seems clear that although the precise distribution of the various signalling molecules differs between mice and chickens, the general principles are the same (Tsukui et al. 1999; Meyers and Martin 1999; Casey and Hackett 2000). The differences between mice and chickens also affect the interpretation of Levin et al. (1996) who gave the original explanation for situs inversus in conjoined twins. Their explanation, which is surely correct in essence, was based on the earlier chicken data and now cannot be right in detail.

62 For Étienne Geoffroy Saint-Hilaire. The quote is from his Philosophic anatomique (1822) and runs in full: ‘les Monstres ne sont plus des jeux de la Nature; leur organisation est soumise à la loi commune; les Monstres sont d’autres êtres normaux; ou plutôt il n’y a pas de Monstres et al Nature est une’.

CHAPTER III: THE LAST JUDGEMENT

65 In 1890 the citizens of Amsterdam. For a history of teratology in the Netherlands and a re-evaluation of Vrolik’s specimens see the magnificent series of articles by Vrolik’s successors at the University of Amsterdam: Baljet and Oostra (1998) and Oostra et al. (1998a-e). Baljet and Öjesjö (1994) have suggested that Hieronymus Bosch’s demonic creatures were inspired by human malformations. Most scholars point instead to medieval illustrations or else the gargoyles of Sint Jan’s cathedral, s’Hertogenbosch. Although Bosch surely cannot have seen many deformed infants in the flesh, the correspondence between many of his grotesques and known deformities is certainly striking. See also Bos and Baljet (1999).

67 Of Willem Vrolik’s published writings. Vrolik’s first major work on cyclopia was a long article in Dutch (Vrolik 1834), followed a decade later by his Tabulae (Vrolik 1844–49). See Baljet (1990) and Baljet et al. (1991) for an account of Vrolik’s work on cyclopia and his collections.

68 Hesiod says that there were three Cyclopes. Homer The Odyssey (trans. E.V. Rieu. 1946. Penguin Books, Harmondsworth, UK); Ovid Metamorphoses (trans. A.D. Melville. 1986. Penguin Books, Harmondsworth, UK). For classical Cyclopean iconography see Touchefeu-Meynier (1992).

68 Many teratologists have linked the deformity to the myth. Aetiological explanations of myths have the delightful property that they are more or less unfalsifiable, but that hasn’t stopped many from proposing them. For arguments that the forms of the various monsters of Greek mythology (the Cyclops, the Hydra, Typhon, the Harpies etc.) are derived from foetuses that show various abnormalities see Schatz (1901). For the claim that the Cyclopes were inspired by the Pleistocene remains of dwarf elephant skulls that have been found on Sicily see Mayor (2000). Far more sophisticated discussions of deformity in ancient Greece and Egypt by a classicist can be found in Véronique Dasen’s numerous papers. The relationships and symbolic meanings of the Cyclopes are discussed by the pre-eminent scholar of comparative mythology G.S. Kirk (1974) pp-85, 207.

69 Homer to Vrolik. Pliny the Elder Natural history: a selection, (trans. J.F. Healy. 1991. Penguin, Harmondsworth, UK). For an early and important outline of the origin of the Plinian races and their fate in medieval literatures see Wittkower (1942). For much further detail see the very readable Friedman (1981) as well as Kappler (1980), Williams (1996) and Daston and Park (1998).

71 The first illustration of a cyclopic child. Liceti (1634) also identifies an additional two eyes on the back of the Firme child’s head, but these must be fictitious (he never saw the child himself, but worked from a sketch). The original sketch of the Janus-headed twins with cyclopia is now lost. Some, but not all, features of the presentation and subject in Liceti’s engraving are consistent with it having been derived from an original by Leonardo. No other sketches of teratologies by him are known (Martin Kemp, pers. comm.). The report of a Janus-headed twin with cyclopia is given in Abbott and Kaufmann (1916).

72 Looking at his bottled babies. For a historical account of causal theories of cyclopia in the nineteenth and twentieth centuries see Adelmann (1936). For the modern definition of holoprosencephaly see Cohen (2001); for the incidence of the disorder see Muenke and Beachy, (2000); for a review of the teratology, Cohen and Shiota (2002); and for the role of veratrum in lamb cyclopia, Incardona et al. (1998).

74 Most cases of cyclopia. Holoprosencephaly (HPE) consists of at least seven distinct inherited syndromes (HPE1 to HPE7). HPE3 (142945) is due to heterozygosity for mutations or deletions in sonic hedgehog (600725) (Roessler et al. 1996). For the spectrum of Shh mutations see Nanni et al. (1999, 2001). For the Shh knockout mouse see Chiang et al. (1996); for a general revew of the other HPE genes see Muenke and Beachy (2000).

76 An embryo’s face. See Hu and Helms (1999) for an elegant study of the role of Shh in craniofacial formation and Ditto the pig. GH3 (165240) mutations cause Greig’s cephalopolysyndactyly (175700) which is characterised by hypertelorism among other things. Hypertelorism with nasal bifurcations is characteristic of craniofrontonasal syndrome (CFNS) (304110). The causal gene underlying has not yet been identified.

78 Among the disorders that appear. For a history of sirenomelia and siren-like iconography see Gruber (1955); for a more recent review see Valenzano et al. 1999). For the CYP26A1 (602239) deletion mice see Sakai et al. (2001) and Abu-Abed et al. (2001). For pigs without eyeballs see Hale (1933); for isotretinoin in humans see Lammer et al. (1985); and for a review of retinoic acid function and gradients in the embryo see Maden (1999).

83 The consequences of cells. The claim that the lumps on the necks of Pans are supernumerary auricles is made by Sutton (1890); Cockayne (1933) discusses goats. See Boardman (1997a) pp.36–7 for the Hellenistic Pans bearing supernumerary auricles. See Boardman (1997a; b) for a technical synopsis of the history of Pan in Greek and Roman art and a charming essay on his iconography by the famous scholar of Greek art. The girl with four auricles is discussed by Birkett (1858); a more general discussion of supernumerary auricles is given by Bateson (1894) pp.177–80 and Cockayne (1933) pp.339–41.

86 Homeosis was first identified. Bateson’s (1894) purpose was not, as now, to use homeotic variation as a means for studying development. He was instead struggling towards a theory of inheritance; that is, attempting to fill the gap left by Darwin’s account of evolution. He failed, but he was among the first to retrieve Gregor Mendel’s experiments on peas from the fathomless obscurity of the Verhandlungen des naturforschenden Vereines in Brünn. As such, he is recognised today as one of the fathers of modern genetics.

87 Over the last eighty-odd years. See Lawrence (1992) for an account of the homeotic genes in Drosophila. Strictly speaking these genes specify parasegments, divisions in the embryo that contribute to, but are out of register with, the segments visible in the larva. The seminal paper on the homeotic genes is Lewis (1978). In 1995, Ed Lewis shared the Nobel Prize with two other Drosophilists, Christiane Nüsslein-Volhard and Eric Wieschaus.

91 Extra ribs have always caused trouble. See Sir Thomas Browne’s Pseudodoxia epidemica, or, enquiries into very many recieved tenents and commonly presumed truths (1646) in Works (1904) volume III chapter II pp.5–8. Estimates on the variation in rib number come from Bornstein and Peterson’s 1966 study of 1239 skeletons. They found that 9 per cent of their skeletons had thirteen pairs of ribs. Of this fraction, just under 1 per cent were due to the seventh cervical vertebra gaining ribs, about 5 per cent were due to the first lumbar vertebra gaining ribs and 3 per cent were due to a simple increase in the total number of vertebrae, that is, were not due to a homeotic transformation of vertebral type. Other studies, reviewed in Galis (1999), put the incidence of seventh cervical vertebrae ribs somewhat lower, at around 0.2 per cent. Cervical ribs can cause ‘thoracic outlet syndrome’, a compression of the nerves and blood vessels of the neck. Galis (1999) also addresses the fascinating question of why all mammals have just seven cervical vertebrae where the numbers of other vertebrae vary greatly among species. She argues that mutations that cause cervical ribs may be far more pathological than is generally appreciated and so under strong stabilising selection.

92 It is no surprise, then, that the identity of each vertebra. This account of C7 specification is based on the results of deletions for particular Hox genes in mice. It is probably incomplete and some of the AND statements should be OR (since partial transformations are common, suggesting that other Hox genes can compensate) or else couched in more quantitative terms – but it is a reasonable start. Deletions in the following genes cause C7->T1 transformations: Hoxa4 (Horan et al. 1994); Hoxa5 (Jeanotte et al. 1993); H0XD5 (Rancourt et al. 1995); Hoxa6 (Kostic and Capecchi, 1994); Hoxb6 (Rancourt et al. 1995). Other genes, such as HoXC4, may be affect this vertebra as well, but results disagree (Boulet and Capecchi 1996; Saegusa et al. 1996; Horan et al. 1995a; b)

93 Distinguishing one vertebra from another. These disorders are caused by deletions in the following genes: anteriorised limbs: Hox5b (Rancourt et al. 1995); partly missing hindbrains: Hoxai (Lufkin et al. 1991; Mark et al. 1993; Carpenter et al. 1993); hernias: Hoxd4 (Ramirez-Solis et al. 1993); no thymus: Hoxa3 (Chisaka and Capecchi 1991); unable to walk: Hoxc8 (Le Mouellic et al. 1992).

94 The Hox gene calculator. For Hox gene expression in human embryos see Vielle-Grosjean et al. (1997). The embryos used in this study were ‘collected with full ethical permission’. For a discussion of what this means and legislation of such studies in various countries see Burn and Strachan (1995).

95 Writing of the ‘calculator of fate’. For supernumerary eyelashes or distichiasis (126300) see Cockayne (1933) P-330 who notes, incidentally, that hedgehogs normally have two rows of eyelashes. For the seven-hearted chicken see Taussig (1988) following a 1904 report by the pathologist Verocay who happened to be staying at the inn and who managed to secure the viscera, but not the rest of the chicken, for study. Isidore Geoffroy Saint-Hilaire (1832–37) volume 1 pp.723–9 discusses various putative cases of heart duplications in humans but can come up with only one possibly authentic example, an early-eighteenth-century case of a grossly deformed infant. See Lickert et al. (2002) for extra hearts in ?-catenin-conditional null mutant mice.

96 And then there is Disorganisation. Disorganisation (223200). The Disorganisation mutation was first studied by Hummel (1958, 1959), then by Crosby et al. (1992). The possibility of a human homologue of Disorganisation was mooted by Winter and Donnai (1989) and Donnai and Winter (1989), who proposed its existence to explain children whose malformations had been previously attributed to a miscellany of other cause. Disorganisation is caused by a dominant mutation on mouse chromosome 14.

97 The power of the homeotic genes. The evolution of snake limblessness is discussed by Cohn and Tickle (1999). This explanation for the loss of fore-limbs in snakes does not account for the loss of the hind-limbs – which is due to a failure of the limb-buds to grow. Many studies have shown comparable changes in Hox gene expression patterns, particularly in arthropods. And some very recent studies have actually demonstrated that mutations in Hox genes are directly responsible for evolutionary changes in morphology (as a change in Hox gene expression, the snake example implies a mutation in some upstream regulatory factor). Mutations in Emx2 (600035), a human homologue of ems, is responsible for schizencephaly (269160). For a discussion of ems and other conserved brain genes see Reichert and Simeone (2001). Mutations in Pax6 (607108), a human homologue of eyeless, mutations cause aniridia (106210) (Ton et al. 1991). For a review of the conservation of eyeless/Pax6 in eye evolution see Gehring and Ikeo (1999).

99 In the cyclical way of intellectual fashion. Geoffroy’s major ideas on what we now call homology can be found in his Philosophic anatomique. Des organes respiratoires sous de la déermination et de l’dentité de leurs pièces osseuses (1818) and Considerations generates. Sur la vertèbre (1822) which have been collected by Le Guyader (1998) who also discusses the dispute with Cuvier as does Appel (1987). The revival of the dorso-ventral inversion hypothesis is due to, among others, De Robertis and Sasai (1996); for a sceptical update see Gerhart (2000).

CHAPTER IV: CLEPPIES

105 For the mark of Cain see Friedman (1981) pp.87–107. The football coach (p. 106) was Glenn Hoddle. He was sacked (The Times, London, 1 February 1999).

107 As recently as 1900. The story of the Cleppies is told by the British geneticist Karl Pearson (1908) in one of the first studies of a ‘lobster-claw’ family. Most British historians, Macaulay among them, accept that the Wigtown martyrs existed, but some such as Irving (1862) have noted that there are no eyewitness accounts and doubts that the whole thing happened, the graves notwithstanding. The best account is Fraser (1877). Irving and Fraser also note that another legend has another officer, the Provost, saying to maid Wilson, ‘Hech, my hearty! tak anither drink,’ only to find himself evermore afflicted with an unappeasable thirst. The uncertain nature of the story is made even clearer by the reference to ‘Good King Charlie’ – Charles II – who at the time of the execution, 11 May 1685, had been dead for two months. Historians generally blame his successor, the Roman Catholic James II, for unleashing the army on the Scottish Lowlands. The etymology of ‘clep’ is also confusing. W.A. Craigie, in his Dictionary of the older Scottish tongue(1931), gives ‘clep’ as ‘call’, but Pearson’s story suggests that ‘clepped’ also means to have a limb deformity. The clepped families themselves are described by Pearson (1908), McMullen and Pearson (1913) and Lewis and Embleton (1908). Pearson’s papers are of particular historical interest for he uses them to advance the agenda of the biometricians against the Mendelians by showing that this apparently dominant gene does not segregate in Mendelian ratios. While his campaign against Mendelianism proved futile, he was partly right about this trait: it looks as though at least one ectrodactyly allele is over-represented in male progeny, an apparent case of meiotic drive, the only one known in a human pathology (Jarvik et al. 1994).

109 The fragments of myth. Euterpe Bazopoulou-Kyrkanidou (1997) argues persuasively that Hephaestus’ lameness was usually represented as club-feet. Aterman (1999) proposes that Hephaestus’ deformity is related to the achondroplasia of the Egyptian deity Ptah – on which more in Chapter V – and, later, to arsenic neuritis, an acquired disease associated with smiths. These points of view are not necessarily inconsistent as the iconography clearly evolved over time. For the origins of the story of the Ostrich-Footed Wadoma see Gelfland et al. (1974); Roberts (1974); articles resurrecting the myth (e.g. Barrett and McCann 1980) and genetic investigations (Farrell 1984, Viljoen and Beighton 1984). Limb defects are second only to congenital heart defects in frequency (Bamshad et al. 1999).

111 One of the strange things about limbs. Pearson (1908) and Lewis and Embleton (1908) recount the manual dexterity of ‘lobster-claw’ families. See Hermann Unthan’s (1935) memoirs for an edifying account of armlessness. The goat is described by Slijper (1942).

113 What induces a limb-bud to grow out into space? The original description of the apical ectodermal ridge (or AER) and its experimental removal is described by Saunders (1948). Acheiropody (200500) is described by Freire-Maia (1975, 1981).

115 The apical ectodermal ridge is the sculptor of the limb. ‘Lobster-Claw syndrome’ and ‘ectrodactyly’ are both now less commonly used than ‘split-hand-split-foot-malformation’ syndrome (SHFM). The disorder occurs in 1 in 18,000 newborns; inheritance is usually dominant. There are at least three distinct SHFM loci in humans: SHFM1 at 7q21.3-q22.1 (183600); SHFM2 at Xq26 (313350), SHFM3 at 10q24 (600095), and we can add a fourth, ectrodactyly, ectodermal dysplasia and cleft lip syndrome (EEC) at 3q27 (129900) (Celli et al. 1999). There are many other related syndromes besides. Celli et al. (1999) identify the EEC gene as p63, a close relative of the tumor suppressor gene, P53; Yang et al. (1999) and Mills et al. (1999) study its function in mice. Another ectrodactyly gene in mice, Dactylplasia, encodes an F-box/WD40 family protein thought to be involved in protein destruction, and although the human homologue of this gene maps near to SHFM3, it has not yet been shown to be causually involved (Crackower et al. 1998; Sidow et al. 1999). The same is true for two distal-less related genes, DLX5 and DLX6, thought to be responsible for SHFM1 (Merlo et al. 2002). Both P63 and Dactylplasia are involved in the maintenance of the AER; among their many other skin defects, p63-homozyous mice have no limbs at all.

116 Action at a distance in the embryo. Developmental biologists will notice that the account given here, which focuses on the AER’s role in promoting the growth of the limb-bud, is not that given in textbooks. There is no mention of how the AER patters the proximo-distal axis of the limb-bud via the ‘Progress-Zone clock’ (Wolpert 1971). This is because a pair of recent papers (Sun et al. 2002; Dudley et al. 2002) have convincingly shown that the Progress-Zone clock model is wrong. This is fascinating, but a bit upsetting, since it seems to throw the question of proximo-distal patterning open again. Niswander et al. (1993) describe how beads soaked in FGF can replace the apical ectodermal ridge. Sun et al. (2002) also give the most recent account of what is now a plethora of engineered FGF mutations in mice which have shed light on how they work.

116 Ridge FGFs not only keep mesodermal cells proliferating. The role of FGFs in regulating cell death is shown by Dudley et al. (2002). See Zou and Niswander (1996) for the role of cell death in eliminating inter-digital webbing in chickens but not ducks. Webbing in humans, more precisely syndactyly, is sometimes the result of an excess of FGF signalling caused by gain-of-function mutations in the FGF receptor, FGFR2, as in Apert syndrome (101200; 176943) (Wilkie et al. 1995).

118 This account of the making of our limbs. The role of thalidomide in phocomelia was first reported by McBride (1961) and Lenz (1962). Phocomelia appears in Roberts’s syndrome (268300) and SC Phocomelia syndrome A (269000), which may be the same disorder and are known as ‘pseudothalidomide’ syndromes; the genetic basis of neither is known. Goya’s sketch of a phocomelic infant is in the Louvre; Vrolik (1844–49) depicts Pepin; a brief account of his life is given in Gould and Pyle (1897) p.263.

120 How does thalidomide have its devastating effects? Stephens et al. (2000) reviews some of the voluminous literature on thalidomide. He firmly discounts recent sensationalistic claims that thalidomide-induced phocomelics (who are now in their late thirties) are giving birth to phocomelic children – which, if true, would imply the existence of some form of Lamarkian inheritance. In principle, however, thalidomide might be a general mutagen causing high frequencies of all sorts of genetic disorders in second-generation infants. Exhaustive studies have failed to show that this is so. Until recently, the best account of the action of thalidomide on limb formation was given by Tabin (1998). His explanation, which he convincingly defended against others (Neubert et al. 1999; Tabin 1999), rested on the idea that thalidomide causes a disassociation between proliferation and proximal-distal specification of limb-buds. In other words, it was couched in terms of the ‘Progress Zone’ model of limb specification. With the demise of that model (Sun et al. 2002; Dudley et al. 2002) the specificity of thalidomide becomes a little more difficult to explain but still probably depends on the abnormal inhibition of proliferation in particular populations of bone-precursors. It is striking that FGF8-conditional limb mutants in mice have phocomelia (Lewandoski et al. 2000; Moon and Capecchi 2000).

121 Metric, with its base 10 units. Until recently it was held that all modern vertebrates (living or not) have no more than five digits (Shubin et al. 1997). True, some creatures such as pandas and moles appeared to have six, but they could be dismissed as not being true fingers, but rather modified wrist bones (the radial sesamoid in pandas and falciform bone in moles). Polydactyly can, however, evolve in flippers such as the paddles of the icthyosaur, Opthalmosaurus, which appear to conceal eight digits (Hinchliffe and Johnson 1980 p.56), and those of the vaquita dolphin, which have six (Ortega-Ortiz and Villa-Ramirez 2000). Alberch (1986) discusses polydactylous dogs; Lloyd (1986) does so for cats; and Wright (1935) for guinea pigs. Galis (2001) reviews the question of why, despite the frequency of polydactylous mutations, so few species exist with more than five digits per limb. Polydactly in humans (603596) and many other entries). Frequencies and kinds of Polydactyly from Flatt (1994); in the Ruhe family (Glass 1947); in the Scipion family (Manoiloff 1931).

122 If the apical ectodermal ridge. For the discovery of the zone of polarising activity see Saunders and Gasseling (1968); for its interpretation see Tickle et al. (1975). Sonic hedgehog (600725) was first identified as the gene encoding the morphogen by Riddle et al. (1993). Since then, some (Yang et al. 1997) have argued that it is not the morphogen since it does not form a gradient in the limb. More recent evidence suggests that it does (Zeng et al. 2001).

126 This catalogue of mutations. Many polydactyly genes have been identified in mice and humans, and many are transcription factors. For example, mutations in GH3 (165240), a zinc-finger transcription factor, cause Greig’s cephalopolysyndactyly (175700), Pallister-Hall syndrome (146150) and postaxial polydactyly (174200; 174700). See Manouvrier-Hanu et al. (1999) for a brief review of others. On-line Mendelian Inheritance in Man (August 2002) lists ninety-seven disorders with polydactyly in the clinical synopsis. How many of these are genuinely different is an interesting question, but the suggestion is certainly that more than ten genes are involved in correctly determining Shh activity. The Shh regulatory mutation causes extra thumbs and index fingers, more broadly, preaxial polydactyly (190605; 174500). The genetics are complicated. Zguricas et al. (1999) mapped the mutations, deletions and translocations to 7q36, close to the Shh gene. Clark et al. (2001) showed that these mutations deleted a portion of Lmbri (605522), a gene near sonic hedgehog, and inferred that Lmbri was causal. Lettice et al. (2002), whose interpretation I follow here, provide evidence that 7q36 Polydactyly mutations are due to deletions of sonic hedgehog cis-acting regulatory elements that lie within a Lmbri intron rather than Lmbri itself. Achieropody (200500), which also maps to 7q36, has a similarly complex history. Achieropody mutations also delete Lmbri and, again, this gene was thought to be causal (Ianakiev et al. 2001; Clark et al. 2001), but is also probably due to a Shh regulatory mutation – though the jury is still out (Lettice et al. 2002). Certainly, the similarity of acheiropody to the pawless limbs of Shh-null mice is striking (Chiang et al. 1996; Chiang et al. 2001).

127 Around day 32 after conception. For the gross development of the human limb see Hinchliffe and Johnson (1980) p.75 and Ferretti and Tickle (1997). The condensations are described by Shubin and Alberch (1986). For a Hoxa13 mutation in man causing hand-foot-genital syndrome, (142959; 140000) see Mortlock and Innes (1997). Mouse models: Fromental-Ramain et al. (1996) and Mortlock et al. (1996). For a Hoxan mutation causing radioulnar systosis (142958; 605432) see Thompson and Nguyen (2000); Hoxd13 (gain of function) (142989), Muragaki et al. (1996); Hoxd cluster deletion, Del Campo et al. (1999). For the most comprehensive attempt at determining what the Hox genes are doing in the limb see Zákány et al. (1997), who report the effects of knocking out a variety of Hoxa and Hoxd genes in combination.

128 Limbs are not the only appendages. For Hox mutations that cause both limb and genital defects in humans see the preceeding note and Kondo et al. (1997). Penis size and foot length (Siminoski and Bain, 1993). For the roles of FGFs and sonic hedgehog in genitals see Perriton et al. (2002) and Haraguchi et al. (2000).

131 The result is rather puzzling. On the homology of lobe-finned fish fins to tetrapod limbs see Shubin et al. (1997) for a review. Sordino et al. (1995) describe Hox gene expression patterns in zebrafish compared to tetrapods. Cohn and Bright (2000) review zebrafish fin development. Dollé et al. (1993) report the Hox d13 knockout in mice; Zákány et al. (1997) argue for the successive accretion of Hox genes in evolution. The first edition of Darwin’s The variation of animals and vegetables under domestication was published in 1868; Gegenbauer’s critique in 1880. The Darwin quote is from the second (1882) edition of The variation pp.457–8 where he retreats. Coates and Clack (1990) describe Acanthostega’s limbs.

CHAPTER V: FLESH OF MY FLESH, BONE OF MY BONE

137 Around 1896, a Chinese sailor named. Arnold and his descendants are described by Jackson (1951) and Ramesar et al. (1996). Their disorder was cleidocranial dyplasia (119600) caused by a dominant haploinsufficient mutation in the osteoblast transcription factor CBFA1 gene (600211). See Komori et al. (1997), Mundlos et al. (1997) and Mundlos (1999) for the identification of the mutation, its function in mice, and a review of the disorder. One of the minor puzzles of this disorder is the absence of apparent homozygous infants in South Africa. With so many carriers living in a small community, two carriers must surely have occasionally married. If the mutation works in humans as it does in mice (and there is every indication it does), one quarter of the children from such a marriage would be completely boneless and stillborn (and half would be partly boneless and one quarter would be normal).

138 Perhaps because they are the last of our remains to dissipate. For a more general review of bone growth see Olsen et al. (2000). For a review of the bone morphogenetic proteins see Cohen (2002). The emphasis placed here on the role of BMPs in making condensations is a little controversial; the evidence from mouse mutations tends to support a role for BMPs in patterning rather than osteoblast and chondrocyte differentiation or condensation formation (Wagner and Karsenty 2001). I suspect that this is due to redundancy among BMPs.

140 By one of those quirks of genetic history. Sclerosteosis (269500) is caused by recessive mutations in sclerostin, a secreted protein (605740). Until recently it was thought that the South African families (who are all Afrikaaners), a family in Bahia, Brazil, and Dutch families with a similar disorder called Van Buchem’s disease or hyperostosis corticalis generalisata (239100) were all related, however remotely. However, the Afrikaaner, Bahia and Dutch families have now all been shown to carry different mutations in or near the SOST gene, so they cannot be related, and the presence of a similar disorder in all three populations is merely a coincidence (Brunkow et al. 2001; Balemans et al. 2001). The fused-finger disorder is proximal symphalangism syndrome (185800) caused by dominant haploinsufficient mutations in noggin (602991) (Gong et al. 1999). For null noggin mutations in mice see Brunet et al. (1998).

141 The disorder is known as. Fibrodysplasia ossificans progressiva or FOP (135100), caused by dominant mutations in an unknown gene. In 2001, a French group reported noggin mutations in FOP patients (Sémonin et al. 2001), but this could not be replicated (Cohen 2002). For Harry Eastlack’s clinical history see Worden (2002) pp.185–6. For a nice essay about FOP, the people afflicted by it, and the search for its cure, see Maeder (1998).

144 A newly born infant has. Baker (1974) notes that the brain case of most adults is about 5 millimetres; of Australids it can be 10 millimetres; Kohn (1995) briefly discusses the head-beating ritual.

144 What makes bones grow to the size that they do? For an account of Victor Twitty’s experiments see Twitty and Schwind (1931) and Twitty (1966). These experiments were carried out in the laboratory of the great developmental biologist Ross Harrison at Yale who initiated them (Harrison 1924). See Brockes (1998) for a review of salamander limb regeneration.

147 The man whose name. For a biography of Mengele see Posner and Ware (1986).

148 Among those spared. The account is partly based on Elizabeth Ovitz’s memoir (Moskovitz 1987) which is also the source of quotations. A careful study of the family and their experience at Auschwitz-Birkenau (Koren and Negev 2003) has, however, shown numerous inaccuracies in the memoir. Detailed accounts of the medical experiments on human subjects carried out in the Third Reich can be found in Lifton (1986).

154 Pseudoachondroplasia – the disorder that afflicted. Pseudoachondroplasia (177170) is caused by dominant gain-of-function mutations in the cartilage oligomeric matrix protein gene (600310) (Briggs and Chapman 2002). The diagnosis of the Ovitzes as having this disorder rather than achondroplasia (as is often stated) is given in Koren and Negev (2003) and is consistent with their attractive facial features.

154 Achondroplasia is caused. Achondroplasia (100800) is caused by dominant gain-of-function mutations in the fibroblast growth factor receptor 3 gene (134934) Rousseau et al. (1994); Bonaventure et al. (1996). For a history of the iconography of dwarfism see Dasen (1993; 1994) and Aterman (1999).

156 If an excess of FGF signalling. For the role that FGFs play in limb growth see Naski et al. (1996; 1998) and Chen et al. (2001). Colvin et al. (1996) study the FGFR3 knockout mouse; De Luca and Baron (1999) review FGFR3 function.

156 Achondroplasia is a relatively mild disorder. Thanatophoric dysplasia (187600) is caused by severe dominant gain-of-function mutations in FGFR3 (Rousseau et al. 1995; Tavormina et al. 1995). Oostra et al. (1998b) describes the Vrolik skeletal dysplasia specimens.

157 FGF must be only one molecule among many. The authoritative review of overgrowth syndromes is Cohen (1989). Myostatin (601788) McPherron et al. (1997) for the mouse mutation; McPherron and Lee (1997) for cattle. The original myostatin mutation occurred naturally on a Flemish farm and so Belgian Blue meat is made ubiquitously into hamburger. Had the same animal been engineered by Monsanto it would have been surely rejected by a public ever suspicious of ‘genetically modified foods’.

158 Mutations that disable bone collagens. There are several types of osteogenesis imperfecta. The most common type that is not lethal at birth is osteogenesis imperfecta type 1 (166200) caused by dominant hapoloinsufficient or gain-of-function mutations in the collagen 1A2 or collagen 1A1 genes (120150; 120160) (Olsen et al. 2000).

160 Even once our growth plates. See Blair (1998) and Günther and Schinke (2000) for reviews on osteoclast function and specification.

160 There are many ways to upset the balance. Malignant autosomal recessive osteopetrosis (259700) is caused by recessive mutations in genes that encode part of the vacuolar proton pump need for hydrochloric acid transport (Kornak et al. 2000).

161 The shortness of Henri de Toulouse-Lautrec. The biographical material and anecdotes come largely from Frey’s (1994) authoritative biography. See Lazner et al. (1999) for the relationship between osteopetrosis and osteoporosis. Maroteaux and Lamy (1965) diagnosed Lautrec with pycknodysos-tosis and review older diagnoses; see Frey (1995 a; b) and Maroteaux (1995) for the exchange concerning his malady. Pycnodysostosis (265800) is caused by recessive mutations in the Cathepsin K gene (601105) (Gelb et al. 1996).

CHAPTER VI: THE WAR WITH THE CRANES

169 From the walls of the Prado. For the iconography of dwarfing see Tietze-Conrat (1957) and Emery (1996). Tanner (1981) pp.120–1 discusses Geoffroy and Buffon on dwarfs.

170 Were all the court dwarfs unhappy. See Boruwlaski’s memoirs (1792) and Heron (1986) for a modern account of his life.

175 At the base of our brains. See Laycock and Wise (1996) for the regulation of growth by the hypothalamus-pituitary pathway. Primary growth-hormone deficiency (262400) is caused by recessive mutations in the growth-hormone gene (139250). There is an enormous literature on this group of syndromes; see López-Bermejo et al. (2000) for a brief review.

176 Joseph Boruwlaski has all the signatures. The Ecuadorean dwarfs have Laron- or growth-hormone resistance syndrome (262500) caused by recessive mutations in the growth-hormone receptor gene (600946); see Rosenfeld et al. (1994) and Rosenbloom and Guevara-Aguirre (1998).

177 In 1782 Joseph Boruwlaski met. Frankcom and Musgrave (1976) write about Patrick Cotter; Bondeson (1997) about Charles Byrne; Thompson (1930, 1996) tells of both men as well as other famous eighteenth-century giants.

179 Charles Byrne had a pituitary tumor. See Keith (1911) for original diagnosis.

179 An old photograph shows a triptych of skeletons. See Schnitzer (1888) (Emin Pasha’s given name) for an account of how he obtained the skeletons. See Schweinfurth (1878) for an account of meeting Akadimoo. The Homeric quote is from The Iliad (1950 trans. E.V. Rieu, Penguin Books, Harmondsworth, UK). Tyson (1699, 1966); Schweinfurth (1878); de Quatrefages (1895); Cavalli-Sforza (1986) all give accounts of Greek and Roman writings on pygmies, but the authoritative work on pygmies in antiquity is Dasen (1993). See Addison (1721) for his verses on pygmies.

182 Addison’s poem. See Tyson (1699,1966) for his dissection of a ‘pygmie’; de Quatrefages (1895) says that the word ‘Aka’ can be found inscribed on the frescos of a fifth-century Egyptian tomb beneath a depiction of a dancing pygmy; Cavalli-Sforza (1986) repeats the story. This would be remarkable if true, but sadly Véronique Dasen assures me it is not.

183 The French anthropologist Armand de Quatrefages. See Schebesta (1952) and Weber (1995–99) for the history of the negritos; Diamond (1991) discusses theories about pygmy smallness. For the most recent study on the genetic relationships of Andaman Island negritos see Thangaraj et al. (2003).

185 The diagnosis of achondroplasia. The Attic vase is just one of many examples of Greek pygmy iconography given by Dasen (1993). Gates (1961) asserted thatpygmies have achondroplasia.

185 That pygmy proportions. The genetics of pygmy smallness are obscure, but the evidence seems to exclude a single locus with substantial dominance. See Shea and Bailey (1986), Shea and Gomez (1988) and Shea (1989) for an analysis of pygmy proportions.

187 The geographers, entranced by their acquisition. The history of Chair-Allah and Thibaut is given in de Quatrefages (1895) and Schweinfurth (1878); their growth curve is given in Cavalli-Sforza (1986) p.366.

188 A newborn infant grows. See Tanner (1990) p.12 for the pubertal growth spurt and Tanner (1981) pp.104–5 for the history of its study; see Bogin (1999) for pubertal spurts in other primates.

189 The pubertal spurt is driven. There is some disagreement as to whether pygmies show low IGF-1 serum titres or whether they have a less effective IGF-1 receptor relative to taller people. In any event, what we know about short stature in pygmies is based on endocrinological studies (Merimee et al. 1981; 1987; Geffner et al. 1995). The following note also bears on the interpretation of these data.

189 The proof of this is the mini-mouse. The account of the relationship between IGF and GH given here differs from the ‘somatomedin hypothesis’ given in most textbooks. Recent experiments have suggested that: (1) IGF-I’S primary role is as a paracrine (short-range) growth factor rather than an exocrine hormone; (2) that liver IGF-1 is responsible for most serum IGF, and that it contributes little to post-natal growth; (3) that IGF-1’s effects on growth are therefore to a considerable degree – though not entirely – inde pendent of GH’s. See Le Roith et al. (2001) for a review of these matters, and Lupu et al. (2001) for an account of the GHR; IGF-1 double knockout mouse.

190 Schweinfurth’s discovery set off a global hunt. See Haliburton (1891; 1894)for ‘pygmies’ in Spain, the Atlas Mountains and Switzerland. See Johanson and Edgar (1996) for the stature of fossil hominids and Bogin (1999a) p.3 for the ‘Maya in Disneyland’.

192 It is even possible that the most recent. Exploration accounts can be found in Kingdon-Ward (1924; 1937) and many other books by the same author. I first learned about the Burmese ‘pygmies’ from Prof. Harry Saing, who saw them while travelling in Kachin State in 1964. The Burmese government reports are given in Mya tu et al. (1962, 1966).

194 It is not a pretty word. For a general review of iodine deficiency diseases and cretinism see Delange and Hetzel (2000). Several reports from western China show a form of myxedmatous cretinism that has many of the features seen in the Taron. Various co-factors have been suggested for the extreme form of cretinism seen in the Congo such as selenium deficiency or a fondness for eating goitrogenic plants.

195 Cretinism is a global scourge. See Merke (1993) for Napoleon’s investigations into Swiss cretins and the iconography of cretinism in Aosta Cathedral. See Delange and Hetzel (2000) for the discovery of the thyroid and Laycock and Wise (1996) pp.203–40 for how it works. The hormone called here ‘thyroxine’ is, more formally, two hormones, tri and tetraiodothyronine or T3 and T4. Williams et al. (1998) discusses the cellular role of thyroxines on the growth plate. Among the many genes with known mutations causing a deficiency of thyroxine production or function are: thyroid peroxidase, an enzyme that is involved in thyroid hormone production (188450) (Abramowicz et al. 1992); thyroglobulin (606765), or the thyroid hormone receptor (190160) (Refetoff et al. 1996); as well as several other genes involved in the synthesis, transport or storage of thyroid hormone (de Vijlder et al. 1999; Vassart 2000). Mutations in all these genes tend to cause cretinism with goitre.

197 There is also a class of mutations more vicious by far. Among the genes with known mutations that affect the pituitary’s stimuation of the thyroid are those that encode: thyroid stimulating-hormone (TSH) itself (188540) (Hayashizaki et al. 1989) and those that cause combined pituitary hormone deficiency (CPHD) due to mutations in transcription factor genes such as PROP-1 (601538) and PIT-1 (173110) required for the specification of somatotroph, thyrotroph and lactotroph cells (Tatsumi et al. 1992; Voss and Rosenfeld 1992; Sornson et al. 1996; Wu et al. 1998). These mutations cause cretinism without goitre and with or without dwarfism.

199 Nearly fifteen hundred years ago. For Aristotle on the effects of castration see his Historia animalium in Complete works volume 1 pp.981–2. For a more recent view of the same subject see Wilson and Roehrborn (1999).

203 We think of the estrogens. For estrogen-receptor deficiency (133430) in men see Smith et al. (1994). The enzyme that converts testosterone to estrogen is aromatase cytochrome P450 (107910) For loss-of-function mutations that cause continued growth in men see Sharpe (1998); Lee and Witchel (1997); for gain-of-function mutations that cause excess in women see Stratakis et al. (1998).

203 Growth hormone and IGF are extremely powerful. One of the several cancer-predisposition syndromes caused by mutations in PTEN (601728) is Cowden syndrome (158350). For the evidence that Proteus syndrome (176920) is also caused – at least sometimes – by PTEN loss-of-function mutations see Zhou et al. (2000, 2001). The Ovid quote is from Metamorphoses (trans. A.D. Melville. 1986. Penguin Books, Harmondsworth, UK). Seward (1992) re-examines Joseph Merrick’s skeleton in detail and upholds the traditional diagnosis of neurofibromatosis type 1 (162200). Tibbies and Cohen (1986), Cohen (1988) and Cohen (1993), however, argue for Proteus syndrome.

206 The intimate relationship between growth and cancer. For the relationship between IGF titres and dog body-size see Eigenmann et al. (1984, 1988); Eigenmann (1987). The association between osteosarcoma and size in dogs was detected by Tjalma (1966), in children by Fraumeni (1967). The latter result has been confirmed by three out of four studies since. See Leroi et al. (2003) for a general discussion on the causes of pediatric cancers. Jenkins (1998) reviews increased propensity of acromegalics to a variety of cancers. The causal role of IGF is reviewed by Holly et al. (1999). In his classic book on ageing, Comfort (1964) first noted that big dogs do not live as long as small dogs. See also Patronek et al. (1997) and Miller and Austad (1999). The best data on ageing rate, from the Swedish pet health insurance scheme, is given in Egenvall et al. (2000).

208 I am fascinated by these findings. Krzisnik et al. (1999) discuss the dwarfs of Krk who are homozygous for recessive mutations in PROP-I. Samaras and Elrick (1999) and Samaras et al. (1999) give a partisan account of the evidence for a negative association between human height and longevity. See Waaler (1984) and Power and Matthews (1997) for the general positive association between health and longevity. There is, however, some evidence from the Finnish studies of a U-shaped mortality distribution in women, possibly associated with skeleto-muscular problems in the tallest women (Läärä and Rantakallio 1996; Silventoinen et al. 1999).

208 These results seem to tell us. The first dwarf mice which were shown to be long-lived were Snell dwarfs (dw) which are deficient in a number of their pituitary gland lineages because of a mutation in PIT-I, a pituitary specific transcription factor. Ames dwarf (dj)has the same phenotype, but has a mutation in another transcription factor, PROP-I. Both these mice are long-lived, but since they lack both somatotrophs and thyrotrophs, they fail to produce both growth hormone and thyroid-stimulating hormone, making it impossible to distinguish the effects of lacking either (Brown-Borg et al. 1996; Bartke et al. 2001 a; b). However, the Little mouse (lit) is also long-lived (Flurkey et al. 2001). Since this dwarf mouse has a mutation in its growth-hormone releasing-hormone receptor (GHRHR), it is very likely that it is growth-hormone deficiency, or its sequelae, that cause the longevity of these strains.

209 To be poor is to be both short and at higher risk. See Mansholt (1987) on Dutch growth; Mackenbach (1991) on the socio-economic causes of height differences in Holland and Didde (2002) for height-related activism in the Netherlands. Also see Cavelaars et al. (2000) for a fascinating comparision of the secular increase in all European countries which shows that although all countries show a secular increase in height, they are all getting taller at more or less the same rate. There is much evidence that milk consumption is responsible for a good deal of environmental variation in height, for example, in the Japanese; Takahashi (1984); Bogin (1999a) p.268.

210 The poverty and short stature of the north’s people. See Rosenbaum et al. (1985) and Mascie-Taylor and Boldsen (1985) for regional differences in height in England and Townsend et al. (1992) for the authoritative survey of health inequalities in Britain. Tanner (1981) p.147 discusses Chadwick and his surveys of height.

211 It is precisely the antiquity of the positive association. There is a huge literature on the attractions of height. Some of it reviewed by Bogin (1999a) pp.326–7. See Sandberg et al. (1994), Guyda (1998) and Root (1998) on growth-hormone therapy for short children.

CHAPTER VII: THE DESIRE AND PURSUIT OF THE WHOLE

217 In February 1868, a Parisian. For the journal and other relevant papers see Barbin (1980). Confusingly, Herculine Adélaïde (Alexina) refers to herself as ‘Camille’.

221 Is Alexina a woman? Chesnet (1860) Annales d’hygiène et de médecine légale 2e série, XIV: 206 quoted in pp.124–8 of Barbin (1980). See Goujon (1869) for her autopsy. See Dreger (1998) for a social history of hermaphrodites.

223 Anatomists, however, have other tastes. Laqueur (1990) claims that Vesalius’ and Galen’s homologies are confirmed by modern, or rather nine-teeth-century, embryology, but this is not so. Thiery and Houtzager (1997) p.51 describe the background of Vesalius’ analysis of the vagina and note that Vesalius so loved his homology between the uterus and scrotum that he depicted the former with a dividing cleft comparable to the raphe of the scrotum where there is none.

225 It was another Paduan anatomist. Laqueur (1989) delves deeply into the history of the identity of male and female genitalia. O’Connell et al. (1998) redescribe the vestibular bulbs as the clitoris; Williamson and Nowak (1998) add further details about the discovery; Kobelt (1844) gives an earlier view.

228 By day 28. Descriptions and timing of embryological events are from McLachlan (1994). I have omitted a description of the internal genitalia (fallopian tubes, uterus and upper vagina, epididymus, vasa deferentia and seminal vesicles), all of which have other embryological origins and are controlled by other hormones.

230 To develop as a female is to travel. For an account of the discovery of the Y see Mittwoch (1973); XX[n]Y males are also often mentally retarded and sterile. The condition is known as Kleinfelter syndrome and occurs a frequency of 1 in 1000 male births (Conner and Ferguson-Smith 1993).

231 The search for the source of the Y’s power. The papers that describd SRY (480000) in humans and mice are Sinclair et al. (1990); Gubbay et al. (1990); McLaren (1990) gives a contemporary commentary.

233 Perhaps SRY activates a few critical genes. For an update of the genes known to be regulated by SRY in the gonad see Graves (1998). For Alexandre Jost’s experiments see Jost (1946–47). The results described are true for rabbits castrated before day 22 post coitem.

234 Or at least it needs its Leydig cells. The testosterone synthesis pathway can be upset at many points. Luteinising hormone is needed for proper Leydig cell growth. Mutations in the luteinising hormone receptor gene (152790) cause Leydig cell hypoplasia and so pseudohermaphroditism (Kremer 1995 and Laue et al. 1996). Then various mutations can disrupt the testosterone biosynthetic pathway (Besser and Thorner, 1994). Some of these cause a group of syndromes known as the congenital adrenal hyperplasias (CAH) (e.g. 201910) since they affect not only testosterone synthesis but the synthesis of other steroids by the adrenal gland as well and have, accordingly, widespread physiological effects. Good examples of testoterone synthesis mutations are those in the 17-? hydroxysteroid dehydrogenase gene (605573) (Russell et al. 1994; Geissler et al. 1994).

235 Such girls are, it is often said, exceptionally feminine. Androgen insensitivity syndrome (300068) caused by mutations in the testosterone receptor gene (313700). For the height of testosterone receptor-null people see Quigley et al. (1992). For the identical twin flight attendants see Marshall and Harder (1958); for the French model see Netter et al. (1958).

236 Alexina/Abel and Marie/Germain were both isolated cases. Montaigne, the humanist, attributed Marie/Germain’s sex change to sublimated sexual desire; Pare, the (evidently Vesalian) surgeon, thought that the exertion of the chase had caused Marie’s genitals to fall out. Montaigne (1580; 1958) p.38; Paré (1573; 1982) p.31. For the Dominican Republic guevedoche see Imperato-McGinley et al. (1974); for the Papua New Guinea kwolu-aatmwol see Imperato-McGinley et al. (1991). The mutated gene in all these cases has either been shown, or else is presumed to be, 5-?-reductase (264600; 607306).

238 When I said that the route to femininity. For infant female pseudohermaphroditism caused by deficiency in aromatase (107910) see Shozu et al. (1991); for the adult aromatase deficiency see Conte et al. (1994) and Morishima et al.(1995). For aromatase excess (shortness, gynecomastia in boys, large breasts in girls) due to dominant gain-of-function mutations, see Stratakis et al. (1998).

240 Spotted hyenas are unsympathetic creatures. See Neaves et al. (1980); Glickman et al. (1992); Licht et al. (1992); Holekamp et al. (1996) and Frank (1997) for spotted hyena endocrinology, genitalia and social structure. Moles (Talpa) also have a kind of female pseudohermaphroditism – although they actually have ovotestes, so could be said to be true hermaphrodites.

242 In The symposium. Plato, The symposium pp.59–65 (trans. W. Hamilton. 1951. Penguin Books, Harmondsworth, UK). For sexual relations of the guevedoche and kwolu-aatmwol see Imperato-McGinley et al. (1991) and Herdt (1994).

CHAPTER VIII: A FRAGILE BUBBLE

247 Our species has, since 1758. Bendyshe (1865) gives a summary and English translation of Linnaeus’ anthropological works; Pearson et al. (1913) and Broberg (1983) discuss Homo troglodytes. Lindroth (1983) discusses Linnaeus’ intellectual roots in medieval thought.

251 His French rival Buffon. For an account of Geneviève see Buffon (1777)Addition à Particle qui a pour titre, Variétés dans l’espèce humaine, Supplement à l’histoire naturelle volume 4 pp.371–454.

253 We are a polychrome species. For a general review of pigmentation genetics see Sturm et al. (1998). The most common form of albinism is oculocutaneous albinism type 1 or OCA1 (203100), which is due to recessive mutations in the tyrosinase gene (606933). Albinism with grey eyes is oculocutaneous albinism type 2 or OCA2 (203200), due to recessive mutations in the P gene (Durham-Pierre et al. 1994; Stevens et al. 1997).

254 In 1871, en route to his encounter with the Aka. See Schweinfurth (1878) volume 2 pp.100–1 for his account of albinos in Africa, and Woolf and Dukepoo (1969) for albinos among the Hopi.

255 Those children would have fascinated Buffon. For the history of Marie Sabina see Buffon (1777) p.557. Pearson et al. (1913) and Dobson (1958). For some of the other eighteenth-century piebalds see Blanchard (1907). For Lisbey’s history see Pearson (1913). Pearson’s insistence on this rather forced account of the inheritance of piebaldism stems, again, from his opposition to the Mendelian theory of inheritance.

261 Molecular devices are required. Piebald trait, white forelock and bilateral hypopigmentation of the limbs and trunk (172800) is caused by dominant mutations in c-Kit (164920) which encodes a receptor tyrosine kinase. c-Kit’s ligand is steel (Sl) in mouse, but no human disorder has been identified with mutations in this gene. c-Kit and its ligand are thought to help in guiding the migration of the presumptive melanocytes. The other piebald syndromes often only cause white forelock but are associated with deafness or megacolon. These are Waardenburg’s syndromes types I through IV (193500, 193510, 602229), caused by dominant mutations in Pax3, Sox10 and MITF (Tassabehji et al. 1992; Watanabe et al. 1998). These genes are transcription factors needed for specification of melanocyte lineages (Goding 2000). All these syndromes manifest variably, all are caused by dominant mutations; homozygotes are probably lethal.

261 What gives us our skin colours? Africans differ from Europeans, and East Asians are known to differ in the structure and density of their melanosomes (Szabó et al. 1969; Toda et al. 1972), but very little is known about the genetics – except for a hint it might have something to do with the P gene (Sturm et al. 1998). See Linnaeus (1758) pp.20–1 for his diagnosis of the human species; a translation is given by Robins (1991) p.171.

262 For nearly half a century. The history of race-classification in South Africa is discussed by Posel (2001). Rita Hoefling’s story is told by Joseph and Godson (1988).

265 Life growth hormone, melanotropins. Red hair and obesity are caused by mutations in POMC (176830), a gene that encodes the ?-MSH and ACTH precusor (Krude et al. 1998).

266 Yet not all redheads are fat. Red hair (266300) caused by recessive mutations in the MC1R gene (155555) (Robbins et al. 1993; Valverde et al. 1995; Smith et al. 1998; Flanagan et al. 2000; Healy et al. 2001). Besides the general plausibility arguments that I have given as to whether red hair has been selected or not (Darwin 1871, 1981 volume 2 pp.316–405; Robins 1991 pp.59–72) and is therefore properly thought of as a mutation or polymorphism, there are also elaborate statistical tests which can sometimes detect historical patterns of selection. Such tests have been applied to MC1R, but they are inconclusive (Rana et al. 1999; Harding et al. 2000).

268 Pale, and proud of it. For a detailed discussion of the pre-PRC history of Chinese anthropology and eugenic thought see Dikötter (1992, 1997, 1998). For a study of Ainu hairiness see Harvey and Broth well (1969).

269 In the collection of the Capodimonte. For the history and iconography of the Gonsalvus family see Aldrovandi (1642); Siebold (1878); Zapperi (1995); Haupt et al. (1990) pp.92–7 and, especially, Hertel (2001).

273 In 1826 John Crawfurd, British diplomat and naturalist. For the history of Shwe-Maong and his family see Crawfurd (1827); Yule (1858) and Bondeson and Miles (1996)M.

276 We are born with about five million hair follicles. For a general review of hair (and feather) specification see Oro and Scott (1998). For the role of BMPs and FGFs, Jung et al. (1998) and Noramly and Morgan (1998). Reynolds et al. (1999) carry out the trans-gender transplantation experiment.

280 The one thing that many of us. Most of the anecdotal material here comes from Segrave (1996) – a delightful social history of balding. Male pattern balding or androgenetic alopecia (109200). See Cotsarelis and Millar (2001) for a general biology of the dying hair follicle, and Kuester and Happle (1984) for a review of the genetics of the androgenetic alopecia.

282 One fact is, however, known: to go bald you need testosterone. See Aristotle Historia animalium in Collected works pp.983–4. Hamilton (1942) recounts the experiments with testosterone. Knussmann et al. (1992) discuss the relationship between testosterone levels, virility and balding.

283 Is there any hope for the bald? Trotter (1928) discusses the relationship between hair growth and shaving. Sato et al. (1999) and Callahan and Oro (2001) discuss the role of sonic hedgehog in rejuvenating hair follicles; Huelsken et al. (2001) discuss (?-catenin.

285 One can still, occasionally. The portraits of the Ambras family were first described in the modern scientific literature by the physiologist C. Th. Siebold (1878). He proposed that they were atavistic, a claim echoed by Brandt (1897), who points out that the Burmese family have the same disorder. Both men recognised that the surplus hair in the two families was lanugo (Siebold explicitly compares Petrus Gonsalvus’s hair to that of a foetal orangutan), but suppose that lanugo is more ‘primitive’ – a conflation between phylogeny and ontogeny that is typical of German workers of the time, who were deeply influenced by Haeckel. Felgenhauer (1969) gives a summary of nineteenth-century views on hairy people. More recently, there has been a great deal of debate about just how many surplus-hair syndromes there are, and who had what (see Garcia-Cruz et al. 2002 for one point of view). I argue that Petrus Gonsalvus’s and Shwe-Maong’s families both have the same condition: hypertrichosis lanuginosa (145700), the mutant gene of which may reside on chromosome 8. The hair of at least one man with this syndrome (a Russian named Adrian Jewtichjew) has been examined microscopically and seems to have been lanugo. The most famous modern pedigree of hairy people, the Gomez family of Mexico, have another, unrelated, disorder: X-linked hypertrichosis terminalis (145701); Figuera et al. (1995). See this paper and Hall (1995), recent – and perhaps reasonable – claims that this latter kind of hairiness is indeed atavistic.

286 Darwin himself knew of the Burmese hairy family. See Darwin (1871; 1981) volume 2. p.378 for his account of sexual selection and hairiness of the Burmese family; see Darwin (1859; 1968) pp.183–4 and Darwin (1882) volume 2, pp.319–21 for the homology between skin organs, the Burmese family and the ‘Hindoos of Scinde’. See Thadani (1935) for a later account of the same pedigree (the ‘Bhudas’) who have a syndrome called ectodermal dysplasia 1, anhydrotic or ED1 (305100) caused by a mutation in ectodysplasin (EDA) (Kere et al. 1996). The Mexican hairless dog’s mutation is still unknown (Schnaas 1974; Goto et al. 1987) but is probably this gene or its receptor, EDAR (224900; 604095) (Headon and Overbeek 1999; Monreal et al. 1999). The scaleless variety of Medaka has a mutation in the EDAR gene (Kondo et al. 2001). For ectodysplasin’s proposed role in establishing hair papillae see Barsh (1999). See Sharpe (2001) on the evolutionary history of the hair follicle.

288 The use of a single molecule in the making. For hens’ teeth see the classic experiments by Kollar and Fisher (1980), a commentary by Gould (1983) pp.177–86, and recent experiments showing that chicken mandibles are BMP4-defective (Chen et al. 2000).

289 Perhaps it is also the retrieval of an ancient signalling system. Nipples, supernumerary or polymastia (163700). For a review see Cockayne (1933) pp.341–5; Japanese polymastia, Iwai (1907). I thank Alan Ashworth and Beatrice Howard for telling me about Scaramanga.

290 Breasts bring us back to Linnaeus. The ancient iconography of Artemis Ephesia is discussed by Fleischer (1984) and Linnaeus’ use of it by Gertz (1948) – for the translation of which I am indebted to Lisbet Rausing. Nosce te ipsum – the slogan that meant so much to Linnaeus is rarely attributed to Solon, but rather (as in Plato) to the seven wise men of Protagorous who wrote it on the temple of Apollo at Delphi. The Oxford dictionary of quotations gives its source as ‘Anonymous’.

CHAPTER IX: THE SOBER LIFE

297 Huntington disease is one of the nastier. Huntington disease, also Huntington’s Chorea or HD (143100), is caused by dominant mutations in the huntingtin gene. Rubinsztein (2002) reviews the molecular basis of the pathology; Bruyn and Went (1986) review the history and spread of the disease.

298 How can so lethal a disorder? See Haldane (1941) pp.192–4.

300 Were it not for ageing’s pervasive effects. Ricklefs and Finch (1995) give estimates of longevity in the absence of ageing.

302 But it was another British scientist. See Medawar (1952) and Williams (1957) for the seminal papers on the evolutionary theory of ageing. Rose (1991) gives an incisive historical review. Albin (1988) discusses the fecundity of women with Huntington’s based on data collected by Reed and Neel (1959).

304 In his declining years, flush with cash and fame. Alexander Graham Bell (1918) analyses the Hyde family; Quance (1977) discusses Bell’s interests in the genetics of longevity.

306 In the 1980s the evolutionary account of ageing. See Rose (1984) for the original experiment; Rose (1991) for a review; and Sgrò and Partridge (1999)for a more detailed analysis of a similar experiment.

308 Since Aristotle. Aristotle On length and shortness of life in Complete Works volume 1 p.743. See Diamond (1982) for the cost of reproduction in marsupial mice and Westendorp and Kirk wood (1998) for the cost of reproduction in British aristocrats. See Leroi (2001) for a sceptical treatment of cost of reproduction data.

309 Is there a recipe for long life? See Cornaro (1550,1903) for a translation of the Vita sobria, and Gruman (1966) for a review of Cornaro’s thought and its influence.

311 The worst of it is that there is an element of truth. See Finch (1990) pp.506–37 for a review of the earlier literature on caloric restriction; ibid. pp.20–1 for mortality rates of the Dutch during the Hongerwinter. See Holliday (1989) and Chapman and Partridge (1996) on reproductive costs and caloric restriction. Several experiments in flies and mice have been done to look at the effects of caloric restriction on ‘whole genome expression profiles’. The best is a study on flies (Pletcher et al. 2002); the mouse studies (Lee et al. 1999) are more difficult to interpret.

313 We term sleep a death. See Beckman and Ames (1998) and Ames et al. (1993) for a review of the free radical theory of ageing. See Rose (1991) for SOD in gerontocratic flies. Parkes et al. (1998) for overexpression of superoxide dismutase in Drosophila motorneurons; Finch and Ruvkun (2001) for a general review of SOD and ageing.

316 Our genomes contain three genes. Familial amyotrophic lateral sclerosis or ALS1 (105400) is caused by dominant mutations in Cu/Zn superoxide dismutase or SOD1 (147450) (Rosen et al. 1993). Deleting this gene in mice seems to have little obvious phenotypic effect, although longevity does not seem to have been examined (Reaume et al. 1996). For the experiments excluding free radicals and hydrogen peroxide as a cause of ALS see Subramaniam et al. (2002); for a review, Orr (2002). For a more general discussion on the causes of ALS see Newbery and Abbott (2002). For the role of SOD1 in Down’s syndrome see Epstein et al. (1987) and Reeves et al. (2001).

319 Wrinkling is a manifestation. Werner’s syndrome (277700) caused by recessive mutations in RECQL2 (also known as WRN) helicase (604611) (Yu et al. 1989) reviewed by Martin and Oshima (2000).

319 As we age. For two reviews of the proposed role of cellular senesence (or Hayflick’s limit) in ageing see Rose (1991) pp.126–36 and Shay and Wright (2000). Bodnar et al. (1997) show that overexpression of telomerase in human cell lines confers cellular immortality. There are some reports that the neuronal cells of mice do not undergo cellular senesence in vitro (Tang et al. 2001; Mathon et al. 2001). There are also strong suggestions that the proliferation of mouse cells in vitro is not telomere limited (Shay and Wright 2000).

321 Mice, it seems, can get by without telomerase. See Blasco et al. (1997), Lee et al. (1998) and Rudolph et al. (1999) for telomerase-deficient mice. One worry about these results is that laboratory mice seem to have much longer telomeres than wild mice (Weinstein and Ciszek, 2002).

322 One way to prove the point would be to clone a human. The original report on cloning Dolly was Wilmut et al. (1997). She died on 14 February 2003. Shiels et al. (1999) reported Dolly’s short telomeres. Cloned cattle appear to have perfectly normal, indeed rather long, telomeres (Lanza et al. 2000; Betts et al. 2001). There is a controversy about the healthiness of cloned animals (Cibelli et al. 2002; Wilmut, 2002). Six generations of mice have been cloned with no sign of rapid ageing – but then, they do seem to have very long telomeres.

323 Telomerase-mutant humans. Hutchinson-Gilford syndrome (progeria)(176670) is caused by a mutation in the gene encoding Lamin A and C.

323 In the last ten years there has been a revolution. See Kenyon et al. (1993) for a pioneering paper in C. elegans ageing studies, and Leroi (2001), Finch and Ruvkun (2001) and Partridge and Gems (2002) for recent reviews.

326 One of the first longevity genes to be identified. Alzheimer’s disease (104300). Late onset (AD2) is associated with particular polymorphisms in the apolipoprotein E gene (107741). For the relative risk of the ?4 allele see Corder et al. (1993); for its rarity in French centenarians see Schächter et al. (1994) and Charlesworth (1996).

327 All this seems to matter less if you are black. For the worldwide distribution of APOE alleles and discussion of relative risk of Alzheimer’s among ethnic groups see Fullerton et al. (2000). There are two ideas why Africans may not feel the deleterious effects of the ?4 allele. First, haplotype analysis shows that their ?4 alleles are somewhat different from those in European populations. Perhaps it simply lacks the pathogenic effect. Second, perhaps it has exactly the same effect, but Africans have, at high frequency, a variant at another locus that protects them against ?4. There is no reason to favour one idea over the other. For African APOE allele frequencies see Zekraoui et al. (1997).

327 In Europeans, at least, the genetics of Alzheimer’s provide. The early onset Alzheimer’s genes are: AD1, ?APP (104760); AD3, Presenilin 1 (104311) and AD4, Presenilin 2 (600759) (Charlesworth 1996).

327 These kinds of findings are only the beginning. Heijmans et al. (2000) review the state of the centenarian gene hunt.

329 In 1994 a remarkable thing happened. Much of the discussion on late-life mortality trends is based on Wilmoth (2000) and Wilmoth et al. (2000).

CHAPTER X: ANTHROPOMETAMORPHOSIS

335 The authors of books. Steve Jones, in the concluding chapter of his The language of the genes (1993) HarperCollins, London gives a classic Utopian account of humanity’s future. Mark Ridley, in the concluding chapter of his Mendel’s demon (2000) Weidenfeld and Nicolson, London suggests the wacky, but interesting, idea that we might evolve huge genomes and fantastically complex life-cycles. For the ethical views of some of the less inhibited scientists see the writings of Richard Dawkins and the late William Hamilton; for the opposition see the New York Review of Books (New York) and the Sunday Times (London).

337 Race has long been under siege. Steve Jones gives a good, if dated, account of these issues in The language of the genes. More recently, see Barbujani et al. (1997) and Rosenburg et al. (2002) for studies based on microsatellite loci; and Stephens et al. (2001) for single nucleotide polymorphisms.

339 The variants are known as AIMS. For an account of the search for AIMS see Collins-Schramm et al. (2002) and Shriver et al. (2003). For an account of the molecular genetics of FY (also known as Duffy) see Li et al (1997).

340 Skull measuring has a long history. See Bindman (2002) pp.201–21 for Camper on skull measurement (from which the quotes as well).

340 Sadly, Camper’s iconography. See Gould (1981) for the classic debunking work on craniometry and IQ. See Lahr (1996), Hanihara (2002) and Hennessy and Stringer (2002) for recent major craniometric studies, all of which build on the work of Bill Howells.

341 Human skulls are wonderfully diverse. See Lahr (1996) for an authoritative treatment of recent human skull diversity. The relative prognathism of Eskimos and Australian Aborigines is calculated from Hanihara (2002) Table 3.

342 My claim that we will soon be able. Boas published several studies on his immigrant data set, the most important of which was Boas (1912). The Rose quote is from Boyd (1955) p.299. Two recent papers, Sparks and Jantz (2002) and Gravlee et al. (2003), have reanalysed Boas’s data. The analysis done by each is somewhat different and they draw somewhat different conclusions. Sparks and Jantz (2002), however, do the critical analysis of variance – with ancestry, birthplace and their interaction as the effects. They show that there is a significant effect of birthplace and – just as one would expect from Boas’s hypothesis – a strong interaction effect. Contrary to Boas, however, the plasticity is small compared to the persistence of ancestral effects and the interactions are not of the sort that would necessarily cause skull shape to converge. They do not accuse Boas of fraud, but one cannot help but suspect that he presented those results that favoured his hypothesis and ignored those that did not. Gould (1981) p.108 cites Boas with approval.

347 They are only the latest casualties. See notes to Chapter VI for the history of the negritos as well as the various essays in McEwan et al. (1997) for a history of the Selk’nam and their legend and their fate.

348 ‘Beauty,’ says the philosopher. See Scarry (2000) p.4 for beauty and the impulse to reproduce. See Plato, The symposium (trans. W. Hamilton. 1951. Penguin Books, Harmondsworth, UK) p.87 for the same. See Darwin (1871, 1981) Vol. 2 p.92 for the Argus pheasant. See Bindman (2002) for a survey of eighteenth-century aesthetic theory with respect to race. See Darwin (1871, 1981) V0I.2 pp.342–54 on the particularity of beauty.

351 The universality of beauty’s standard. See Thornhill and Gangestad (1999) for a survey of the recent literature on facial attractiveness. See Perrett et al. (1994) for a classical study on the perception of female beauty. What Brazilians say is recorded (with delight) by the late William Hamilton in The narrow roads of gene land (2002, Oxford University Press, Oxford) Vol. 2 p.677. Many of the ideas about the meaning of beauty expressed in this chapter can be traced to Hamilton’s writings.

353 The effects of poor childhood nutrition. For the genetics of the face see Winter (1996). For spontaneous abortion as an adaptation to eliminate defective embryos see Forbes (1997).

354 Mutation is a game of chance. See Crow (2000) for the number of deleterious mutations and a model of truncation selection.

356 Beauty, Stendhal says. Stendhal, De l’amour (Folio, Paris) p.59.