J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from Review Article

J. med. Genet. (I967). 4, 44. Human Polymorphism J. PRICE*

F-rom the Nuffield Unit of Medical Genetics, Department of Medicine, University ofLLiverpool Ford (I940) has defined polymorphism as the If this were the case, any attempt to relate this occurrence together in the same habitat of two or polymorphism to differences in function of the more discontinuous forms or phases of a species various forms of acid phosphatase would plainly not in such proportions that the rarest of them cannot succeed. be maintained merely by recurrent mutation. So few selective mechanisms (similar to that In his Genetic Polymorphism (I965) he states relating sickle cell trait to falciparum malaria) (p. I4) that 'A unifactorial character must be poly- have been discovered, that the present position is morphic if found even in I% of a considerable that most polymorphisms described are, so to population, amounting perhaps to 5oo individuals speak, 'in search of a disease', which they seem on or more, when random genetic drift may reasonably present evidence unlikely to find. The selective be excluded as unimportant'. This figure of I% factors may, of course, not be concerned with has been used in this paper as a rough dividing disease, but be connected with fertility, ability to line between what needs to be mentioned and what gather food, and other forms of fitness.

does not. This fruitless searching for selective mechanismscopyright. Ford's definition excludes continuous variation, which will give a clear-cut explanation of known as exemplified by human height and skin colour, polymorphisms seems to have deterred a great and it excludes rare disadvantageous recessive con- many investigators from making any attempt to ditions, such as albinism and haemophilia. determine how a polymorphic situation has come Two types of polymorphism are included in the about. Many publications give lists of data for the balanced and the transient. A definition: poly- many human polymorphic systems from every http://jmg.bmj.com/ morphism may be balanced, that is, with the ratio corner of the globe, without referring to possible between two (or more) different characters remain- selective factors that may be maintaining them. ing constant throughout many generations, through So often the data collected are wholly genetic- the operation of one of several mechanisms of which thus we may be given the haptoglobin and trans- heterozygous advantage is the best known (see the ferrin phenotypes, but we are seldom given the section on blood groups). Altematively, poly- serum iron or the haemoglobin level. morphisms may be transient; if two discontinuous The situation of polymorphisms in search of forms (A) and (B) exist, then in time one or other of disease has a corollary: disease in which neither on September 27, 2021 by guest. Protected the transitions (A) -* (A + B) -- (A) or (A) -* (A + B) the number of genes involved nor their precise -÷ (B) will take place. If a certain form has a io% effects are known, but in which various facets incidence in a population can we be sure whether suggest that a polymorphic situation may exist. it is increasing, decreasing, or stable in that popu- Two such diseases may be mentioned here- lation? The answer in many cases seems to be no, diabetes and schizophrenia: both are common; because of lack of knowledge of frequencies in the both are inherited conditions; both have interfered past and of the selective forces involved. markedly (and still do, to an extent which is greater Genes always have multiple effects. It therefore for schizophrenia) with biological fitness. Further- follows that we may discover a polymorphism re- more, in each condition mechanisms may exist to lating to one effect, for example red cell acid phos- maintain the frequency of the responsible gene or phatase, in which it is another property of the genes: in diabetes the increase in prediabetic genes responsible for acid phosphatase production matemal fertility; in schizophrenia, increased which is bringing about the polymorphic situation. resistance to stress (this does not seem to be estab- ished, but it remains a possibility). *Bates Research Fellow, Mental Health Research Fund. Particular attention has been paid in this review 44 J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from Human Polymorphism 45 to those polymorphisms in which selective factors would rise steadily. One disadvantage could occur are known to exist, or in which hypotheses have in compatible heterozygote matings (e.g. AO x AO been constructed to explain a polymorphic situa- or AB x AB) if heterozygous offspring survived tion. Data on some of the newly-discovered poly- better than homozygous offspring. morphisms, not yet widely known, have also been Both Chung and Morton (I96I) and Matsunaga included. (i962b) found an excess of AB children in AB x AB mating. The same might be true of AO x AO A: BLOOD GROUP POLYMORPHISMS matings, but these are difficult to study because the Three aspects of these complex polymorphisms AO phenotype cannot be established by simple are discussed here. testing as can the AB phenotype. Chung and Morton (I96I) stress that 'the genetic evidence I: Selective Forces Acting on ABO and Rh strongly supports the assumption of Brues (1954) Blood Groups that the ABO polymorphism is maintained because The subject of blood groups and natural selection heterozygous advantage in compatible matings is discussed (principally with reference to ABO and more than compensates for selection against hetero- Rh) by Sheppard (1959), Reed (I96I), Clarke (I964), zygotes in incompatible mating'. They further hold Chung and Morton (I96I) among many others. that 'the overall effect of maternal/foetal incompati- Among the selective factors the following call for bility is virtually the same for 0, A, and B mothers, consideration. despite the fact that clinical haemolytic disease is almost restricted to children of 0 mothers'. This Prezygotic Selection. Three possible mechan- should, perhaps, bring home the importance of not isms exist: the first two mechanisms (which cannot overestimating clinical haemolytic disease in con- be distinguished by testing) are confined to the sidering selection and ABO blood groups. heterozygous male. Reed, Gershowitz, Soni, and Napier (I964) divide Meiotic drive-unequal numbers of sperms carry- the reproductive indicator into several independent cate- ing each type of allele. gories: the number of pregnancies, the proportion of copyright. Sperm competition-unequal opportunity for each pregnancies terminating in abortion or in stillbirth; type of sperm in the act of fertilization. and the proportion of liveborn children dying non- Thus with an AO father meiotic drive would accidentally under the age of 5. These data were mean more A sperms than 0 sperms (or vice versa) obtained from several hundred couples where the wife i.e. a was over 40, so that reproduction had virtually been instead of equal numbers, quantitative differ- completed. Six blood groups were studied in these ence. Sperm competition would mean some couples (ABO, Rh, MN, Kell, Duffy, P) as was the http://jmg.bmj.com/ intrinsic property favouring A sperms at the secretor status. expense of 0 sperms (or vice versa), i.e. a qualita- An effect found to be significant at the o-ooi level tive difference. was that K females w-re found to have m re pregnancies; Sperm incompatability-meaning, for example, at o oos level Group N females were more likely to be that a sperm carrying an A gene coming in contact sterile (taken from amended data: Reed, I965). Further with cervical secretions containing anti-A would, studies of this type will plainly be required, both to thereby, be eliminated. substantiate or refute associations already found, and to Matsunaga (i962b) provides evidence to support extend the work to other blood groups. The task of on September 27, 2021 by guest. Protected the determining the selective factors operating in the blood hypothesis that prezygotic selection operates group polymorphisms is formidable indeed, particu- throughout so as to favour group 0 sperms: Morton larly as interactions such as those between ABO and (I965) does not consider the case proven. Rh may occur. Zygotic Selection. By this is meant the elimina- The Rh System. Newcombe (I963) and (I965) tion of zygotes whose blood group is not that of the (amended data) has studied the ABO and Rh mother, by spontaneous abortion or by clinical blood groups ofmothers experiencing foetal death in haemolytic disease. That this occurs has been New York, and has found that, with the exception of established by Chung and Morton (I96I), and AB Rh- mothers (who fare worst, being most likely Matsunaga (i962b). The loss is always of hetero- to become Rh sensitized), the risk of foetal death zygotes. to mothers over 30 rises as the number of anti- genically active ABO and Rh alleles unrepresented ABO Heterozygotes. The tendency for group in the mother's genotype rises (this number is o 0 sperms to be favoured must be balanced by dis- for AB Rh +; I for A Rh + and B Rh +; 2 for advantages, otherwise blood group 0 frequency ARh -,BRh -,andORh +;and3forORh -). J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from 46 J. Price In connexion with Rh incompatibility, a theoretical Livingstone (I960) and Vogel, Pettenkofer, and point that does not seem to be widely understood may Helmbold (I960) discuss this subject. be mentioned. If three genotypes appearing with The effect of differences in incidence of duodenal frequencies p2, 2 pq, and q2 (corresponding to genes p ulcer, gastric carcinoma, and pernicious anaemia and q) have relative selective advantages of a, b, and c, on the ABO polymorphism is infinitesimal compared respectively, then for equilibrium: P ap2 bpq with prenatal and early postnatal influences in main- q bpq +cq2 taining the polymorphism, because these diseases p _b-c whence: affect biological fitness hardly at all. This is because q b-a mortality from these diseases chiefly occurs after This means that equilibrium can be achieved only if the reproductive period is over. either b - c and b -a are both positive, or if they are both negative (because p/q must be positive), i.e. if the Influence of Blood Groups on Disease Pro- heterozygote is either better or worse adapted than both cesses. Considerable interest has been taken in the homozygotes. Now the baby afflicted by Rh haemo- this subject since I953 when Aird, Bentall, and lytic disease is always a heterozygote. It is, therefore, Fraser Roberts showed an association between tempting to suggest that the disadvantage of this heteroz- gastric carcinoma and blood group A. gygote maintains a stable equilibrium. Unfortunately A number of further associations have been for such a simple hypothesis, this type of equilibrium is established: pernicious anaemia and rheumatic not stable. Stability in this type of situation (invol- ving only a single locus) occurs only where there is fever with group A, and duodenal ulcer (and, to a heterozygous advantage. Li (I955) gives the mathematical lesser extent, gastric ulcer) with group 0. proof of this. Group 0 subjects possess H-antigen in the highest concentration, and Eichner, Finn, and Krevans The compensation phenomenon whereby mothers (i963) suggest that this may explain the relation who have lost babies through Rh haemolytic disease between this blood group and the titres ofantibodies produce more living babies than average seems against the organism Esch. coli o86B7, which 0

established beyond doubt (for Caucasian mothers). subjects possess at lower levels than A, B, and ABcopyright. Yet, as Li (1953) has pointed out, this cannot lead subjects. If Esch. coli o86B7 itself carries an 'H-like- to a stable equilibrium, nor, of course, can genetic antigen, human subjects with group H antigen may drift produce such a state of balance. Li (I955) fail to produce an antibody against the Esch. coli suggests that a small heterozygous advantage (which antigen and, hence, have lower antibody titres. A he computes as o oog) hardly detectable in ordinary clear-cut disease association in this connexion has genetic data will be sufficient to counterbalance the not yet been established. http://jmg.bmj.com/ loss of those infants dying through Rh incompati- Edwards (I965) emphasizes that achievements to bility and that a stable equilibrium could then be date in this field should be kept in perspective. He maintained. The nature of this advantage does not calculates that if duodenal ulcer were purely the result seem to have been determined as yet. of susceptibility inherited through the equal actions of a large number of genes of the potency and frequency of those at the ABO locus, then I00 loci should be im- II: Blood Groups and Disease plicated. He goes on: 'before hereditary influences can

be regarded as sufficiently marked to merit such atten- on September 27, 2021 by guest. Protected Two aspects call for consideration: tion a substantial contribution to total variance should be demonstrated'. It appears that ABO and secretor Influence of Disease on Polymorphic Blood status contribute together about one-fortieth part of Group Systems. Does disease as it occurs today the total variance. Wiener (I965) in stronger vein writes influence the ABO polymorphisms to an important 'studies on blood groups and disease have led to no degree? Available evidence seems to indicate that useful result to date. No one has used blood-grouping it does not. Has disease ever influenced the poly- in differential diagnosis.... No one has offered a morphisms to an important degree? This is more satisfactory explanation how one's blood group could difficult to answer, but the evidence, as far as such directly affect one's susceptibility to duodenal ulcer or diseases as plague and smallpox (which have in the to carcinoma, nor how one's blood group could affect past been pandemic with a very high mortality) are one's personality or temperament' (one can infer that concerned, is not convincing (Harris, 'disease' in this quotation excludes haemolytic disease). I963). One might answer these arguments with a quotation The decreasing likelihood of pandemic lethal from Ford (I965)-'The association of polymorphisms infectious disease seems to make it less likely that with disease of any kind provides an example of the the opportunity of relating blood groups to survival fact that controlling genes have important effects in or death from such diseases will present itself. addition to those by which they are normally recognised'. J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from Human Polymorphism 47

III: Excess of Heterozygotes in MN Blood observe other Amazonian groups for this pheno- Group System menon. On the other hand, Das and Mukherjee This occurrence led Morton and Chung (I959) (I964), and Das, Mukherjee, and Bhattacharjee and Chung, Matsunaga, and Morton (I96I) (I963) have described remote tribes of the state of to conclude that data were consistent with Orissa (India) where the proportion of tasters to heterozygote advantage apparently limited to chil- non-tasters is approximately i to i. Other studies dren of MN mothers. This hypothesis would include those of Barnicot (I950) in African Negroes require that the MN locus should have properties and Chinese: in both groups, non-tasters are less other than that of producing erythrocyte antigen. common than in European populations. This Hiraizumi (I964) proposed a different model to fit applies also to the Japanese survey of Lugg (I966) the data-whereby N has the lowest viability in the who criticizes the classical allelomorphic gene pair zygotic stage but has selective advantage in the (T and t) hypothesis on the grounds that it inade- prezygotic stage. Wiener, Gordon, and Wexler quately explains available data, particularly a (I963) did not find the excess of heterozygotes, and group of subjects with extremely high activities of criticized reports showing such excess on technical taste which he has found. grounds. Levene, Hermoni, and Cividalli (I965) Mechanism of Differing Taste Sensitivity also failed to confirm the excess of heterozygotes in a study from Israel. There is evidence to show that the saliva of non- tasters can metabolize PROP and other thioureas B: PHENYLTHIOCARBUTAMIDE more rapidly than the saliva of tasters (Griffin and TASTING Fischer, I960; Fischer and Griffin, I960). These investigations have further shown that this effect If one determines the highest dilution in which a can be abolished by incubation of saliva with beef test substance can just be tasted, one obtains a catalase, and the authors conclude that hydrogen bimodal distribution showing that a polymorphism peroxide is largely responsible for it. They point exists for the tasting of phenylthiocarbutamide out that such an effect can only account for a two- copyright. (PTC), and a variety of other related compounds. fold difference in the threshold for PROP These include 6-n-propylthiouracil (PROP) and whereas the non-taster would require something like other antithyroid substances. Individuals able to a sixteenfold lowering in the concentration which taste weak solutions are called tasters and those he can just appreciate to be converted to taster who require strong solutions are called non- status. A physiochemical difference between the tasters. A pair of autosomal alleles T and t deter-

receptor sites on the tongue in tasters and non- http://jmg.bmj.com/ mine this effect. Subjects of the genotypes TT and tasters seems to be necessary to explain the experi- Tt are tasters, and of genotype tt, non-tasters. mental findings. Many aspects of taste perception Testing does not distinguish between TT and Tt. are dealt with by Fischer, Griffin, and Pasmanick Population Studies (I963). In European populations approximately 30% of Possible Selective Factors Maintaining people are non-tasters, which gives a gene fre- Polymorphism quency for 't' of slightly over 50°o%. The finding by Various publications have sought to establish or on September 27, 2021 by guest. Protected Fisher, Ford, and Huxley in I939 that 30% of refute association between taster status and goitre, chimpanzees are non-tasters led them to suggest tuberculosis, diabetes, paralytic poliomyelitis, and that a balanced polymorphism had existed since the duodenal ulcer, as well as in less common conditions, separation of anthropoid and human stocks. Un- e.g. mucoviscidosis and mongolism. fortunately (for this hypothesis) large human popu- lations with a very much lower percentage of (a) Goitre. Goitre has attracted most attention, non-tasters have been discovered (Barnicot, 1950). and rightly so, for PTC and PROP are goitrogens. Junqueira, Kalmus, and Wishart (I957) found no Clearly, if either consumption or metabolism of non-tasters in the Carajas Indians who live on the naturally-occurring goitrogens related to these sub- Island of Bananal in the Arguaia river, and Mon- stances differ between taster subjects and non- tenegro (I964) has also found low values for taster subjects, differences in goitre incidence may Brazilian Indians, with a marked sex difference, be expected. Harris, Kalmus, and Trotter (1949) female non-tasters being much less common than showed an association between non-taster status males. This difference is statistically significant and nodular goitre; Kitchin, Howel-Evans, Clarke, (p = o-oi), and the author suggests the need to McConnell, and Sheppard (I959) have confirmed J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from 48 J. Price this association, particularly for English males, and data strongly supporting the hypothesis that there Azevedo, Krieger, Mi, and Morton (I965) have was a differentiaUy increased mortality from juvenile confirmed it in Brazil. The work of Kitchin et al. tuberculosis in 'non-tasters'. Ifthis were confirmed, (I959) showed also a marked deficiency of non- it would provide evidence of a selective mechanism tasters among subjects with diffuse parenchymatous operating against non-tasters. This work has been goitre. Covarrubias, Barzelatto, Stevenson, Boba- criticized by Akesson (i959) on the grounds that the dilla, Pardo, and Beckers (I965) could not confirn population studied by Saldanha was not homo- these associations for the Pewenche tribe of Chilean geneous so that socio-economic stratification factors Indians, nor could Luca and Cramarossa (I965) in were not avoided. Akesson's own work does not Central Italy, nor Memoria (1959) in Brazil, nor show any significant correlation between the Fraser (I963) in England. However, in an endemic occurrence of tuberculosis (or of diabetes) and the goitre area of Israel (non-toxic nodular goitre being taster phenotype. the type of goitre occurring), a significant increase in A paper on the increased incidence of taster -numbers of non-tasters was found among goitrous status in patients with non-fatal paralytic polio- as compared with non-goitrous children (Brand, myelitis (Brand, I964) is based on a study carried I963). It seems that present evidence does not lead out in Israel. The population studied was a mixture -to the establishment of simple rules relating taster of European, Asian, and North African immigrants, status to thyroid disease. or the first generation offspring of such, and is open to the same criticism as that of Saldanha Where an association has been established, how can mentioned above. Kaplan, Fischer, Glanville, -such an association come about? Clements (I960) has Powell, Kamionkowski, and Fleshler (I964) have described a sequence of events to account for the produced evidence of an increase in the proportion appearance of goitre in newly-settled land. According of tasters in patients with duodenal (but not with to this, cows first graze on indigenous grasses, while gastric) ulcer. This cannot be related to differences other land is cultivated for crops. When this is allowed in age, smoking habits, race, or diet, between ulcer to lie fallow, infestation with weeds such as Raphanus raphanistrum and Brassica campestris rapidly follows, patients and non-ulcer controls. For reasonscopyright. and these weeds contain significant concentrations of mentioned in the blood group section of this review, bitter-tasting goitrogenic compounds which may be such an association is not, of itself, likely to operate present in milk. If 'taster' children are deterred from on biological fitness to a marked extent. drinking such milk by its taste, they would, thereby, avoid consumption of goitrogenic substances and be less likely to develop this type of goitre. (There is

C. COLOUR http://jmg.bmj.com/ evidence (Fischer, Griffin, England, and Gain, I96I) BLINDNESS to suggest that 'tasters' have far more food dislikes POLYMORPHISM than non-tasters, and tasters, by rejecting a variety of Colour blindness is of interest both to the animal foods, may relatively reduce their consumption of bitter and to the human geneticist. The former will be and pharmacologically active substances.) fascinated by the uneven distribution of colour The development of a goitre is not, per se, evidence vision in the animal kingdom (e.g. men and butter- that selective forces are operating against its possessor's ffies have colour vision whereas dogs do not). The biological fitness, though obvious cosmetic disadvantage human geneticist will find it of special interest, is one simply understood factor that could operate. on September 27, 2021 by guest. Protected Non-tasters are also at a disadvantage in that a higher because colour blindness was one of the earliest proportion of their children will suffer athyreotic creti- sex-linked anomalies to be documented. nism than would be expected by chance (Fraser, I96I). Definitions. When red, green, and blue light are all projected on a single screen, the triple mixture of these (b) Other Diseases. A clear relation between colours will make a white light. A person with normal diabetes and PTC 'non-taster' status does not seem colour vision will require a particular intensity ratio of to have been established. Harris et al. (I949) failed red: green: blue, in order to see white; he is a normal to confirm a relationship put forward by Terry and trichromat. Some people need a more than usual Segall (I947). A recent paper showing a slight, but intensity of one particular colour of light to see white significant, hyposensitivity of diabetics to the taste light-these are 'anomalous' trichromats (hence the of names protanomaly and deuteranomaly, see below). dextrose has been published by Schelling, A further group of people need only two of the three Tetreault, Lasagna, and Davis (I965). This sensi- colours to match white, and these people are dichromats. tivity is said to be controlled by a locus different Deuteranopia and protanopia (see below) are the com- from that of PTC tasting (Fischer et al., I963). monest forms of dichromatic vision. In I956, Saldanha, working in Brazil, produced Finally, a few people will match a single colour against J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from Human Polymorphism 49 white light-these are monochi omats: they perceive no of the propositi was a heterozygote for genes at the hues seeing only white and grey. protan and deuteran loci, so that the two normal sons had to be considered as recombinants. A A useful classification is as follows: further family study giving additional support to the two loci theory is given by Dreyer and Gold- Trichromatic vision Dichromatic Monochromatic vision vision schmidt (I965). (not subdivided) Comparison of Data for Males and Females Normal colour vision Protanomaly Protanopia on a Single Gene and a Double Gene Hypo- Deuteranomaly Deuteranopia Tritanopia thesis. If we take x as the frequency of the protan defect in males and y as the frequency ofthe deuteran defect in males, the total protan/deuteran defect Protanomaly and protanopia are often grouped frequency in females is given by (x2 + y2) on the together under the title 'protan colour blindness', and 'two gene loci' hypothesis. the primary dysfunction is inability to recognize red we treat the data as that for light. Deuteranomaly and deuteranopia are similarly If simply red/green grouped as deuteran colour blindness, the difficulty colour blindness, the frequency for males is x + y; being primarily with the recognition of green colours. for females (x + y)2. The lower value x2 + y2 Deuteranomaly is the commonest abnormality account- agrees better with population study findings. ing for (very approximately) 6o% of all cases. In the The scarcity of double hemizygotes for red-green female, deuteranomaly is dominant to deuteranopia, defects (which has been calculated to be about i in should the unusual circumstance of inheritance of an 8oo of all males) is worth noting. Pickford (I962) abnormal X chromosome from each parent occur. illustrates this problem and discusses it further in Tritanopia, which is manif-sted by difficulty in blue/ a Ciba Foundation symposium (i965), p.238, yelow discrimination, is very rare. which is entirely devoted to the subject of colour Kalmus' monograph (I965) gives a comprehensive account of the methods used currently in the investiga- vision. In this latter article, Pickford distinguishes tion of colour blindness. three different subtypes of defect for each of the copyright. protan and deuteran types-protanopia, extreme Two Polymorphisms protanomaly, and simple protanomaly, with inheri- Red/green colour blindness includes deutera- tance true to type. In the discussion following his he says 'there must be many more than nomaly, deuteranopia, protanomaly, and protano- article, that, six genes . . .' with influence on colour pia. However useful the term red/green colour (i.e. red/green vision), and if this view is correct, linkage studies blindness is in everyday usage, the expression over- http://jmg.bmj.com/ may well show discrepancies due simply to hetero- simplifies the position when one is considering geneity between the populations being studied. polymorphisms, for it seems increasingly likely that there are two polymorphisms-normal/deuteran Population Studies and normal/protan with independent loci on the X chromosome. That there are at least 2 genetic Post (I962) lists a large number of studies up to loci for these two polymorphisms is suggested by I962. More recent data include those of Thuline evidence from a variety of sources. (I964) from U.S.A.; Kalmus, De Garay, Rodarte, and Cobo (I964) from rural Mexico; Plato, Cruz, on September 27, 2021 by guest. Protected Linkage Studies. Siniscalco, Filippi, and Latte and Kurland (I964) from the Western Pacific; and (I964) carried out linkage studies on a Sardinian Dutta (I966) from India. population and concluded that the sequence for the Post divided the data in his paper as applying to following 3 X chromosome genes was: three categories. A: hunters and food gatherers; deuteran: G6PD: protan, C: populations farthest removed from hunters in though this does not agree with the findings of time and habitat; and B: an intermediate category. Jackson, Symon, and Mann (i964) who have the Summing the data from all populations in each colour vision loci close together and not separated subgroup, he finds the frequency of colour blindness by the G6PD locus. in Group A to be 0-02; in Group B, O0O33; and in group C, around o-o8. He considers that the evi- Family Studies. Siniscalco et al. quote a dence fits in with the hypothesis that there has been family studied by VanderDonck and Verriest (ig60) a relaxation ofselection against the 'colour blindness' in which two deuteranopic, one protanomalous, and genes in the populations of group C. Dutta's two normal sons occur in the same family. This analysis (I966) of Indian data gives general support seemed best explained by assuming that the mother to this hypothesis. Post (I965) calculates the J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from 50 J. Price selection rate for category A subjects to be the Northern Chinese, the frequency of W equals same for protans and deuterans, namely o-o6. Post 002II; at the other extreme, W equals 0o9306 for considers that selection against young boys who American Negroes. have not learned to compensate for their defect An association between wet cerumen and the and who have not had reproductive opportunity be- development of apocrine sweat glands has been cause of their age may account for this high figure, noted (Matsunaga (i962a) gives references). It has as may the rewarding of successful (normal-sighted) been suggested that differing reactions to environ- hunters and warriors in primitive societies with mental changes may be associated with possession more wives (Neel and Post, I963). Ford (I965) has of these glands. Presumably this would mean criticized Post's hypothesis on the grounds that selection favouring the gene W where possession the gene could not have reached the status of a of the sweat glands conferred advantage. No sug- polymorphism unless it were associated with gestions as to possible advantages of w have been advantages as well as disadvantages (though the made. nature of these advantages is quite unknown). Although it is well known that persons with colour E: P AMINO ISOBUTYRIC ACIDURIA blindness may show superiority in recognizing There do not appear to have been many contri- camouflage, any effect resulting from this is likely butions to the genetics of this condition since the to adjust (rather than maintain) the polymorphism review by Gartler (I96I). A percentage ofthe popu- in any given habitat, for it has yet to be shown that lation (about i0% for English white individuals) are a habitat exists in which the colour-blind person high excretors of this amino acid in the urine, these has over-all visual advantage over a normal person. subjects being homozygotes for a recessive gene. Post (I962) suggests that relaxation of selection in Low excretors are either heterozygotes or homo- European communities may have been in operation zygotes for the corresponding dominant allele. since the introduction of domestic animals and agri- Malignant disease, and treatment thereof, and culture I20 generations ago. impaired liver function are causes of increased levels Selection in Civilized Communities of urinary BAIB which do not depend on geneticcopyright. control. Gramberg-Danielson (I96I) states that the theore- tical implication that colour defective persons should Armstrong, Yates, Kakimoto, Taniguchi, and Kappe be confused by the colour of a traffic light is practi- (I963) put forward the interesting hypothesis that high cally insignificant, but he 'welcomes a decision of BAIB excretion may enable essential amino acids (or German traffic authorities according to which the other essential food factors) to be conserved by a renal

tubular counterbalancing mechanism. They suggesthttp://jmg.bmj.com/ green of a traffic light will tend in the future towards that this conservation might be beneficial, particularly the more conspicuous green-blue'. under conditions of famine. This advantage, which One wonders whether the selection of people would apply to the homzygote for 'high excretor' gene, with normal colour vision for jobs where this is would not lead to a stable polymorphism, but a transient important (e.g. combatant servicemen, airline pilots, polymorphism could be produced if famine conditions engine drivers etc.), but where more risk is entailed were to persist in a population long enough for this than in the average job, may not constitute a slight gene (introduced by mutation) gradually to increase in frequency. Armstrong has investigated his hypothesis advantage for the colour-blindness genes. on September 27, 2021 by guest. Protected by comparing the urinary output of various amino acids D: HUMAN EAR-WAX in high and low BAIB excretors but has not detected any significant differences. Little appears to have been published on the subject of ear-wax since a review by Matsunaga High excretors of BAIB appear to be deficient in (i962a). Two types, wet and dry, differ in physical an enzyme which further metabolizes BAIB and properties, the former having a sticky consistency. they simply excrete it unchanged. Gartler (I96I) Chemical differences plainly exist but full charac- suggests that, as BAIB formation is essentially a terization of these differences does not seem to have waste pathway, high urinary excretion may simply been achieved. represent a shortening of the pathway, and may, If two autosomal alleles W and w are postulated, thereby, have some selective advantage and be most reports are consistent with the hypothesis that increasing in frequency. A state of 'enzyme wet cerumen corresponds to genotype WW or Ww; deficiency' must not automatically be considered dry cerumen to ww. It is remarkable that in various as indicating a biological deficit: proteins evolve human populations W and w vary in frequency by and those that cease to confer any benefit will tend almost as much variation as is possible: thus, in the to become extinct. J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from Human Polymorphismrt 5I F: HAEMOGLOBINOPATHIES zygote advantage is balanced against the loss of large numbers of the HbhS/Hb s homozygotes, so Huehns and Shooter (I965) have recently re- that the fitness of the latter is very low in Africa. viewed in this Journal a great deal of data on the However, Anderson, Went, MacIver, and Dixon structure and biosynthesis of haemoglobin and its (I960) and Jonxis (I965) report that some West variants and the genetics thereof. Only data relating Indian sickle cell homozygotes may survive to to the polymorphisms which involve variant haemo- reproductive age. globins will be touched on here. Additional evidence for the malaria hypothesis is: The major part of haemoglobin consists of the (a) the demonstration of lower P. falciparum para- A1 portion. The haemoglobin A1 molecule con- site rates and counts in heterozygote children (Hb 3S tains, in the globin fraction, four polypeptide chains, HbOA) than in non-sickling children; (b) the rarity of two a and two P. Each pair is under the control of sickling in fatal malarial infection; and (c) the dis- a pair of alleles. Abnormality of the a chain occurs tribution of the sickle cell gene, which does not in a-thalassaemia and of the P-chain in (-thalas- attain high frequency outside areas where malaria is saemia, sickle cell anaemia, and Hb C and Hb E hyperendemic or holoendemic, except in a small disease. area near Lake Chad (Nigeria) where a frequency Five abnormal haemoglobins are present in of 30% was reported by Roberts and Boyo (I960), various human populations in a frequency so high and where malaria is only epidemic. as to constitute examples of unquestionable poly- As far as the effects of the trait on fertility are morphism. The genes responsible are the following. concerned, Allison summarizes the situation thus: (i) Sickle cell gene-up to 40%O trait carrier in 'increased female fecundity is far from adequate to some Eastern African tribes (Allison, I954). explain the persistence of the sickle cell gene at the (2) The (-thalassaemia gene-up to 20% trait observed frequencies in Africa'. However, Firschein carrier frequency in Rhodes (Barnicot, Alli- (I96I) has reported that in the Black Caribs ofBritish son, Blumberg, Deliyannis, Krimbas, and Honduras (who are of African Negro origin), the Ballas, I963), and in Sardinia and Italy fertility roati of sickle cell trait mothers to normal copyright. (Carcassi, Ceppellini, and Pitzus, I957). is sufficient to maintain the gene in the absence of (3) Haemoglobin C gene-30% in Northern differential mortality due to malaria in infancy. Ghana (Edington and Lehmann, 1956). This effect is suggested as being prenatal, due to (4) Haemoglobin E gene-almost 20% in Thai- interruption ofpregnancy in non-sicklers by malaria. land (Kruatrachue, Charoenlarp, Chong- However, other reports (from Africa), e.g. those of suphajaisiddhi, and Harinasuta, I962). Burke, De Bock, and De Wolf (1958), Roberts and http://jmg.bmj.com/ (5) The a-thalassaemia gene. Boyo (I960), and Delbrouck (1958), have not given In each of these first four cases the polymor- such clear-cut results. phisms have been reported in several areas, the percentages quoted being at the upper end of the H: P-Thalassaemia frequency range. As far as f-thalassaemia is concerned, Fraser, Recent publications which include data on the Stamatoyannopoulos, Kattamis, Loukopoulos, De- population genetics of the haemoglobinopathies faranas, Kitsos, Zannos-Mariolea, Choremis, Fessas, include those of Allison (I965) and Lehmann and and Motulsky (i964) conclude that 'although it is on September 27, 2021 by guest. Protected Huntsman (I966). likely that falciparum malaria also plays a role in the maintenance of thalassaemia frequencies the I: Sickle Celi Polymorphism evidence is not as good as for sickling and G-6 Pd With the sickle cell gene (HbOs), the protection deficiency'. Thus, though Carcassi et al. (I957) afforded to the heretozygote Hbas/Hb5A against noted a higher incidence of P-thalassaemia in the falciparum malaria, increasing biological fitness, low-lying marshy areas of Sardinia than in the provides an example of heterozygous advantage mountainous areas where malaria has been less which few would question. Allison (I965) has common, Choremis, Fessas, Kattamis, Stamatoy- tabulated data showing that the ratio of sickling annopoulos, Zannos-Mariolea, Karaklis, and Belios trait in older children and adults (together) to that (I963) found a non-significant gene frequency in young children exceeds unity. This rise is ex- difference between schoolboy populations grouped plained as being due to the loss of Hb5A homo- according to the altitude at which they lived in a zygotes from malaria, an event most likely to happen region of Greece which had been malarious. in the first four years of life before acquired im- Increased iron absorption from the intestine munity affords substantial protection. This hetero- found in thalassaemia homozygotes by Witzleben J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from 52- J. Price and Wyatt (I96I) is not present in the heterozygote IV: Haemoglobin E (Bannerman, I962), and the suggestion of Haldane Data on haemoglobin E do not seem to be ade- (1949) that enhanced absorption might protect quate for the assessment of a 'malaria hypothesis'. heterozygotes against iron-deficiency anaemia and, Protection against malaria has yet to be shown. thus, confer advantage cannot be substantiated. The gene has attained high frequency in South- East Asia (Lie-Injo Luan Eng (I964) tabulates this), III: Haemoglobin C an area in which Allison (I957) has suggested that Homozygotes for the Hb5C gene, unlike those for malaria might be involved in maintaining a poly- the Hb 5S gene with which Hb 5C is widely associated morphism. Further data on this subject and on the in West Africa, are not afflicted by a disease which interaction between the genes for 3 thalassaemia is lethal before the reproductive period. Their and haemoglobin E (similar to that between Hbas condition will only rarely interfere with biological and HbaC discussed above) in the Thai population fitness. This abnormal gene is found in West Africa. are given by Flatz, Pik, and Sringam (I965a, b) Cabannes (I965) tabulates the frequencies of the gene which does not appear to spread to the East V: a-Thalassaemia of the Niger river. No rule of thumb can be given The a-thalassaemia gene is probably associated for the occurrence together of Hbas and Hb lC genes with intrauterine death in the homozygous state, in populations-there is neither a direct nor an in- and in Negro and Oriental populations the carrier verse relation over-all. Cabannes provides an state may reach levels of 5%. However, difficulty in analysis of available data. It seems likely that a recognizing the heterozygous state in adults has state of stable equilibrium has not been reached in precluded accurate population studies. Weatherall all the populations where both these genes occur. (I965) discusses this problem. The following data on haemoglobin C are of interest. It is clear that in the case of the sickle cell gene (a) Evidence in favour of protection of the heterozy- substantial evidence exists for the operation of gotes for Hb5C against P. falciparum does not seem copyright. completely conclusive, though Thompson (I962) pro- malaria as a selective force. In other haemoglobino- vides evidence showing a lower mean parasite density pathies (and in G6PD deficiency) selection due for AC than for AA phenotypes in Ghanaian children to malaria may be less intense and, therefore, aged I-6 years, which is significant at the o-oi level. more difficult to demonstrate. Nevertheless the (b) Using data from North Ghana, Edington and indirect evidence provided by distribution studies Laing (1957) calculate fitness ratios for Hb5A/Hb5A and suggests malaria as a selective force for these con- Hb,A/Hb C as o-976 and I I03, respectively. ditions also. http://jmg.bmj.com/ (c) Penrose, Smith, and Sprott (1956) reach the con- clusion that a stable equilibrium between Hbjs, Hb5C and Hb5A could be reached, provided HbOA/Hb5S and G: ENZYME POLYMORPHISMS Hb5A/HboC heterozygotes are at advantage. (d) Went (I96I), in an unconfirmed report, has drawn The known number of polymorphisms relating attention to a departure from expected Mendelian ratios to human enzymes has increased rapidly over the among offspring of Jamaican HbOA/HboC mothers mated past few years. Where polymorphic enzymes are with HbOA/Hb5A fathers; 24 Hbj3A/Hb5C offspring, as in the red or on September 27, 2021 by guest. Protected compared with 7 Hb5A/Hb5A (p < o-oi), were produced. present cells serum, direct measure- As this might provide an example of differential gametic ment of enzyme activity coupled with separation selection or su±-vival, further data would be worth of the individual components by electrophoresis collecting to see whether or not this observation can be affords a convenient method of study. The number substantiated. of such serum and red cell enzymes which have been shown to be polymorphic suggests that some Fullerton, Hendrickse, and Watson Williams of the parenchymal enzymes may be polymorphic (I965) report a io% mortality for haemoglobin also. Investigation of these is hampered by diffi- SC disease (genotype HbasHb5c) in pregnancy, from culty in obtaining them for study from organs Ibadan, Nigeria, mortality being much higher in that are inaccessible except at operation or necropsy. emergency admissions. Megaloblastic anaemia due Pharmacogenetics (which may be defined as the to folic acid deficiency was the largest single factor study of genetically determined variations in animal associated with this mortality. It is presumed that species which are revealed by the effects of drugs) dietary intake cannot cope with the double stress of offers one solution to this difficulty (see Evans, extra erythropoiesis and extra requirements due to I962, I963). This approach has met with useful the pregnancy. results: for example, the acetylation polymorphism J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from Human Polymorphism 53 has been investigated largely by means of a pharma- for the two sexes. If X1 is a chromosome bearing a cogenetic approach. gene determining deficiency and X a chromosome Electrophoretically-delineated polymorphismsmay with the normal gene, then males can be normal be divided into two types: those in which there (XY) or affected (X1Y = the affected hemizygote); is a quantitative alteration in enzyme activity and females can be normal (XX), heterozygote (X1X), those in which enzyme activity remains at a normal or affected (X,Xl = the affected homozygote). level. In the latter case, selective forces may still be able to exert a direct effect, e.g. naturally-occurring Clinical Picture. G6PD deficient people (X1Y enzyme inhibitors often have altered properties of and X1X1), particularly if non-Negroes, are subject inhibition when the structure of the enzyme on to neonatal jaundice and kernikterus; Rh haemolytic which they act is altered. disease may be more lethal when combined with Indirect effects, via some other property deter- G6PD deficiency. Haemolytic episodes may occur mined by the polymorphic genes, may also explain with drugs, e.g. antimalarials, sulphonamides, some of the enzyme polymorphisms in which, with sulphones, antipyretics, nitrofurans, etc., and with the exception of G6PD, knowledge of selective foods such as the fava bean (or with fava pollen), forces is at present virtually non-existent. certain peas, and several other vegetables, and during Eight enzyme systems are dealt with in this viral or bacterial infections. These more severe section. haemolytic episodes may be set against a background of chronic haemolytic anaemia, particularly in I: Erythrocyte Glucose-6-Phosphate Northern European and Japanese populations. Dehydrogenase Deficiencies The Malaria Hypothesis. It seems likely that General Considerations. decreased morbidity from malaria in the heterozy- Biochemistry. Red cell glucose-6-phosphate gote female has become balanced against loss of dehydrogenase (G6PD) is an enzyme concerned in hemizygote affected males and homozygote affected the oxidative metabolism of red cell glucose. This females. copyright. oxidative reaction forms part of the pentose- The 'malaria' hypothesis in relation to G6PD monophosphate shunt, a relatively minor portion deficiency has been extensively reviewed. Recent of glucose being dealt with by this shunt normally. reviews include those of Motulsky (i964, i965). It is thought that the presence of a certain amount Motulsky and Campbell-Kraut (I962) have maps of reduced glutathione may render haemoglobin showing that the distribution of G6PD deficiency more resistant to oxidant compounds (e.g. the has marked similarity to the distribution of fal- http://jmg.bmj.com/ sulphonamides), and that G6PD is necessary for ciparum malaria. There are a number of reports the production of this reduced glutathione. A showing a high incidence of G6PD deficiency in simple reaction scheme is as follows: low-lying malarious areas with a much lower Glu. 6 phosphate+ TPN G6PD Gluc. phosphogluconate+ TPNH+ H incidence of malaria and G6PD deficiency in high- Oxidized glutathione TPNH land areas only a few miles away. Kidson and Gor- reduced glutathione man (i962) have challenged this concept. One fact Reduced glutathione is necessary for optimal against the 'malaria' hypothesis used by them, namely metabolism of the Plasmodium falciparum parasite the low incidence of G6PD deficiency in populations on September 27, 2021 by guest. Protected (Motulsky, Kraut, Thieme, and Musto, 1959). such as the Malays and Indonesians where malaria Techniques. Motulsky (i965) gives an account of has been holoendemic, has prompted a reply from test systems for G6PD activity. Allison (i963) who points out that the only con- Distribution. Many populations exist in which vincing argument against the hypothesis would be the frequency of a single type of G6PD deficiency the presence ofhigh frequencies of G6PD deficiency attains a frequency (often well over 20%) so high in populations known to have lived in an environ- that a genetic polymorphism must be postulated to ment long free of malaria. Kidson and Gorman, occur. The chief areas of distribution of the defi- on the basis of data obtained from New Guinea, ciency are West and Central Africa, India, parts and using additional genetic parameters also suggest of South East Asia, the Philippines, and some of that spread of a genetic trait into the uplands may the countries around the Mediterranean. have occurred from a 'large gene pool on the coast near the river mouth with dilution and diminution Genetics. The loci determining normal and of gene frequency along the valley and into the variant enzymes are on the X chromosome. There- highlands'. This may indeed be so, but one would fore, the situation has to be considered separately wish to know how such a gene pool originally arose. J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from 54 J. Price Allison points out that though other parasitic deficiency gene may have occupied a time amount- diseases and nutritional inadequacies might account ing to some 200 generations or so. for different G6PD deficiency rates, no specific alternative hypothesis to the malaria one has yet Heterogeneity of G6PD Deficiency. There is been put forward. considerable evidence to indicate that the deficiency Although data regarding malaria parasite counts is not the result of an identical mutation in various in children with enzyme deficiency have been con- affected populations. This heterogeneity is mani- flicting, the bulk of evidence suggests that malarial fest in variations in degree of decrease of enzyme infection may be less severe in small children (who activity in different cell types; in differences in will have incomplete acquired immunity) with chemical properties of the enzyme; and in dif- G6PD deficiency. Powell and Brewer (1965) in- ferences in susceptibility to various agents capable fected I6 adult Negro males (8 being G6PD deficient of inducing haemolysis. Marks and Banks (I965) but symptom-free, and 8 normal) with P.falciparwn, tabulate data relating to the properties of eight dif- and could detect no difference in the attacks of fering variant enzymes with reduced activity. malaria so produced. As they point out, such experiments do not shed light on what obtains Variation without Deficiency. Kirkman and when there are very high levels of parasitaemia Hendrickson (I963) give further details of a system attended by substantial mortality. The perfor- described by Porter, Boyer, Schulze, and McKusick mance, under laboratory-controlled conditions, of (I96i), in which an electrophoretic variant unac- experiments designed to derive data on differential companied by diminished activity is described. human mortality has certain ethical difficulties. Data from Davidson et al. (i964), showing cor- relation between erythrocyte G6PD levels in Ascertainment of the Heterozygote. In normal male sibs, provide further evidence sug- detecting the heterozygote female, practical dif- gestive of a number of normal alleles. ficulties are encountered if an assay method is used.

The methaemoglobin reduction test will identify copyright. approximately 8Wo of heterozygotes (Tarlov, II: Erythrocyte 6-Phosphogluconate Brewer, Carson, and Alving, i962), but errors Dehydrogenase Polymorphism may be made in both directions, genetically-proven 6-phosphogluconate dehydrogenase (6PGD) is heterozygotes giving results in the normal or the an enzyme on the pentose monophosphate shunt, a deficient ranges instead of in an intermediate property which it shares with G6PD (q.v.) category (Trujillo, Fairbanks, Ohno, and Beutler, http://jmg.bmj.com/ I96I). The situation is technically easier in non- Glucose-6-phosphate G6PD 6-phosphogluconate Negro populations (e.g. in Sardinia), than in Negroes, because the variety of deficiency en- 6-phosphogluconate 6PGD 6-phospho-3-ketogluconate countered in the former is such as to cause greater Values of6PGD can be determined by the method deviation of G6PD activity from normal in the of Zinkham and Lenhard (I959). The common heterozygous female. Davidson, Childs, and variant to be discussed here does not confer a state Siniscalco (I964) give further data on this subject. of deficiency (though true deficiency can occur). There are, thus, practical difficulties of ascertain- In this respect, of course, 6PGD differs from on September 27, 2021 by guest. Protected ment in investigating heterozygote advantage in G6PD. G6PD deficiency. Data on malarial mortality and There has been a recent survey on 6PGD by fertility in heterozygotes do not yet appear to be Dern, Brewer, Tashian, and Shows (i966). They available (Motulsky, i964). examined haemolysates by starch gel electrophoresis in American subjects, and 6-i% of an abnormal Haemoglobinopathies and G6PD Deficiency. phenotype called AB was found among Negro Livingstone (i964) discusses the population dyn- subjects; this was also found in I of 58 Caucasian amics ofabnormal haemoglobin and G6PD deficiency subjects. Fildes and Parr (i963), working in genes. He points out that, in contrast to the London, found io examples of AB in 150 blood haemoglobinopathy genes, G6PD 'deficiency' genes samples from adults 'mostly of European descent'. do not have a great influence on biological fitness. Parr and Fitch (I964) found the common variant in Therefore, it takes much longer for the incidence I in every 27 of a sample of the population of of the latter genes to increase. His conclusions, London E.i, so that no gross differences seem to evidently somewhat conjectural, suggest that the exist between American and British data. Both state of attainment of equilibrium for the G6PD sexes are affected. Fildes and Parr (I964) suggest a J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from Human Polymorphism 55 two allele system with two genes X and P. Geno- among humans by Bonicke and Reif (I953) and types mm and oc thus correspond to the phenotypes by Hughes (I953). Distribution histograms pre- A and AB (AB subjects possess both A and B pared by Biehl (I956, I957) of the percentage of protein bands). [Note added in proof: Further details free urinary isoniazid showed a bimodal pattern. of electrophoretic properties of 6PGD, including Mitchell, Riemensnider, Harsch, and BeU (1958) phenotype B, can be found in a paper by J. E. considered that the polymorphism for isoniazid Bowman, P. E. Carson, H. Frischer, and A. L. de might be genetically determined, and studies by Garay in Nature (Lond.) (i966) 2I0, 8II.] Knight, Selin, and Harris (I959) and Evans, Manley, and McKusick (I960) showed that slow m: Erythrocyte Phosphoglucomutase inactivation of isoniazid was 'recessive' to rapid Polymorphism inactivation and that the genes concerned were autosomal. Phosphoglucomutase (PGM) plays an important part in carbohydrate metabolism. It is a phospho- Further work by Sunahara (I96I), using a more transferase which catalyses the transfer of a phos- refined microbiological technique, showed that phate group between the i- and the 6-position of the genotype could be directly determined, the glucose. PGM is widely distributed in the body. homozygote 'rapids' inactivating distinctly more Although the polymorphism to be described is rapidly than the heterozygotes, so that the latter studied most conveniently in red blood cells it can group formed an intermediate category. On this be shown to exist in a variety of tissues. basis neither character is dominant to the other. Spencer, Hopkinson, and Harris (I964) have Sunahara supports his hypothesis by studies on studied this enzyme in human red cell lysates by families. means of starch gel electrophoresis. At least seven Dufour, Knight, and Harris (I964) have studied enzymes having PGM activity can be distinguished the genetics of isoniazid metabolism in Caucasian, in human tissues. These have been labelled 'a' to Negro, and Japanese populations, and calculated 'g' according to the rate of migration on starch gel, gene frequencies for each population. The gene 'a' being the slowest running component of the frequency for the gene with 'rapid' effect is approxi- copyright. system, 'g' the fastest. (Rare abnormalities of 'e', mately o025 in both Negroes and Caucasians. In 'f', and 'g' have been described by Hopkinson and Japanese, on the other hand, the frequencies are Harris (I965).) Spencer et al. have shown that a reversed, 0o7I being the frequency for the 'rapid' polymorphism exists, characterized by the presence effect gene. A study of Eskimos by Armstrong and of 3 phenotypes, I, 2-I, and 2. Subjects with pheno- Peart (I960) shows that these subjects are more frequently 'rapid' inactivators. Sunahara, Urano, type i possess PGM in the 'a' and 'c' positions (as http://jmg.bmj.com/ well as 'e', 'f', and 'g'); with phenotype 2-I, PGM and Ogawa (I96I) state that in Japan the more in all positions 'a' to 'g'; and with phenotype 2, southerly the region, the higher the frequency of PGM in positions 'b' and 'd', as well as 'e', 'f', and the slow allele, though this is not without exception. 'g'. Data obtained from 338 unrelated English Hughes, Biehl, Jones, and Schmidt (I954) and adults show frequencies for the three types I, 2-I, Hughes, Schmidt, and Biehl (I955) showed that and 2 to be 0o550, o-376, and o0o74, respectively. there was a wide range of the percentages of a Family studies are consistent with there being two dose of isoniazid, which could be recovered from autosomal allelic genes, PGM' and PGM2 (frequen- the urine as the actetylated derivative in different on September 27, 2021 by guest. Protected cies PGM'= o074, PGM2= 0-26), with homozy- subjects. It was thought likely that the meta- gotes PGM1/PGM' and PGM2/PGM2, and hetero- bolic distinction between the phenotypes lay in zygotes PGM'/PGM2, representing electrophoretic their possession of different speeds of acetylation. types I, 2, and 2-I, respectively. Evans (I962) has produced evidence to support On present evidence, the amount of PGM this hypothesis. Other drugs acetylated in the activity seems to be approximately equal in the body, for example sulphadimidine, hydrallazine, three phenotypes so that gross deficiency is present and, in all probability, phenelzine (nardil), are in none of them. dealt with in a 'polymorphic' fashion, whereas The selective factors maintaining this poly- para-amino-salicylic acid and para-amino benzoic morphism are quite unknown. acid, which are also icetylated, are not dealt with polymorphically. Red cells contain only the latter acetylating enzyme, which is further IV: Acetylation Polymorphism evidence that two separate acetylating enzymes A large variation in the metabolism of the anti- exist (Motuisky and Steinmann, I962; Jenne, tuberculous drug isoniazid was found to exist Macdonald, and Mendoza, I96I; Evans, I965a). J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from 56 J. Price Side-effects from drugs used in clinical practice acid phosphatase was studied using starch-gel have been reported as being more frequent in slow electrophoresis. In a subsequent paper by the than in rapid inactivators: this has been shown to same authors in I964, their motives for selecting apply in the case of the polyneuritis of isoniazid- this particular enzyme for study are given. First, treated tuberculous patients (Hughes et al., 1954; that it was present in red cells and, therefore, Devadatta, Gangadheram, Andrews, Fox, Rama- available (whereas parenchymal proteins require krishnan, Selkon, and Velu, I960), and in patients biopsy or autopsy for their collection); and secondly, with depression treated with phenelzine (Evans, because it was thought likely to be susceptible to Davison, and Pratt, I965). examination by starch gel electrophoresis. Although the side-effects are increased in slow The fact that a protein, chosen at random (from inactivators of these drugs, there does not seem to the genetical standpoint), proves to be a striking be any reduction of the efficacy of the therapy in example of genetical polymorphism may be fortui- rapid inactivators, either of isoniazid in tubercu- tous, or it may be regarded as providing support losis (Harris, I96I; Gow and Evans, I964), or of for the view that such molecular polymorphisms phenelzine in depression (Evans et al., I965). may be much commoner than is generally sup- Perhaps the latter finding is not so surprising when posed. one considers that an investigation by the Medical Research Council into the effects of mono-amine Population and Family Studies. Population oxidase inhibitors in depression found that the studies are very incomplete as yet. Data from drugs were no more effective than placebo (Brit. family studies are consistent with the suggestion of med.J., I965). Hopkinson et al. (I963) of three alleles pA, pB, and There is evidence suggesting that the acetylation Pc at a single autosomal locus without dominance. polymorphism exists in rabbits (Frymoyer and There is a gene frequency of o036, o-6o, and O004, Jacox, i963a, b), and in Cercopithecus aethiops (an respectively, in an English population. Giblett African monkey) (Goedde, Schoepf, and Fleiscb- and Scott (1965) found slightly different figures

mann, I964). in Seattle Caucasians (o0394, o0547, and o0oso);copyright. The part played, if any, by these polymorphic their data for Seattle Negroes give frequencies of enzymes in normal intermediarymetabolism remains o-226, 0-759, and O-OI5. Pc was zero for Seattle to be demonstrated. Such a demonstration might Orientals and Northern Japanese. In neither of pave the way towards an understanding of the the two studies has the phenotype CC been selective mechanisms involved in maintaining the identified, but this is not surprising in view of the polymorphism, which are at present quite un- small number of individuals investigated to date known. and the expected rarity of this phenotype. Lai, http://jmg.bmj.com/ Nevo, and Steinberg (I964) in Brazilian families of mixed ancestry (Caucasian, Negro, and South V: Erythrocyte Acid Phosphatase American Indian) have obtained frequencies of o 2o, Acid phosphatase is an enzyme which is widely o077, and o0o3 for pA, pB, and pc, respectively. distributed throughout the human body: it has the property of cleaving phosphoric monoesters at Enzyme Kinetics. Hopkinson and his co-

an optimal pH of 5. A variety of different tissues workers (I964) have shown, using p-nitrophenyl on September 27, 2021 by guest. Protected contributes to the normal serum level of this phosphate as substrate, that the activity of the enzyme. A raised serum acid phosphatase is most acid phosphatase derived from the various pheno- commonly associated with prostatic carcinoma, types varies, with least activity in A, more in BA, particularly if the total bulk of malignant tissue is more still in B and CA, and greatest activity in CB great, as may happen if there are distant metastases. (phenotype CC was not available). Their data That there are different structural forms of the support the hypothesis of a simple additive effect acid phosphatase molecule is suggested by the fact (e.g. the activity of BA is the mean of the activity that enzymes derived from different tissues have of B and A separately). They point out that if a different inhibitors; certainly red cell and prostatic random sample of individuals were taken and red acid phosphatases have distinctly different proper- cell acid phosphatase activity estimated, the ties in this respect. resulting conventional histogram-activity level A human polymorphism for red cell acid phos- on the abscissa; number of subjects at each activity phatase was first described by Hopkinson, Spencer, level on the ordinate-would show a continuous and Harris (I963), and 5 different patterns (labelled unimodal distribution. In spite of this, the variation A, BA, B, CA, CB) were detected when red cell appears to be explicable on a simple three allele J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from Human Polymorphismr 57 hypothesis. It seems possible that some examples allele as calculated from these data agreed with the of so-called quantitative multifactorial inheritance observed value of 0o25. may have a similar uncomplicated underlying basis. Beckman et al. (I966) also describe 4 further modified types, determined by electrophoresis of placental extracts. VI: Serum Alkaline Phosphatase Polymorphism The Situation Excluding Pregnancy. The Alkaline phosphatase is a widely-distributed significance of findings with regard to alkaline enzyme which has the property of cleaving many phosphatase polymorphism in non-pregnant, heal- phosphoric monoesters, with an optimal pH of 9 3. thy human subjects has yet to be fully elucidated. Serum alkaline phosphatase is heterogeneous, Starch gel electrophoresis of sera from such sub- consisting of components of bone, liver, and intesti- jects shows either one or two zones as a rule. nal origin. Phosphatases derived from different Rarely a third band derived from bone is present. tissues can be differentiated because they have The first two zones mentioned have been described differing sensitivities to a variety of chemical by Arfors, Beckman, and Lundin (I963) as zone A agents. (possessors being phosphatase (Pp) group i), and Two distinct situations need to be distinguished zone B (possessors of zones A and B being Pp 2). in discussing alkaline phosphatases: (a) the situa- All of 89 monozygotic twin pairs were concordant tion in pregnant women, and (b) the situation in for these groupings-6i pairs type I, 28 pairs type non-pregnant human subjects. In both there is 2, strongly suggesting genetic control. evidence of polymorphism; in both there is evidence Further studies by Beckman (I964) in Brazil have of genetic control, but it is only in the former situa- shown that possession of zone B (Pp 2 status) has a tion that genotypes can be deduced. positive correlation with positive ABH secretor status and Duffy blood group Fya+ and negative Pregnancy Enzymes. Boyer (I96I) found that correlation with Lewis (a+b-) and blood group A1. copyright. in the sera of all women in late pregnancy, extra This last-mentioned association has been confirmed alkaline phosphatase components could be distin- in British subjects by Evans (I965b), and by Bam- guished by starch gel electrophoresis. Two bands, ford, Harris, Luffman, Robson, and Cleghorn A and B (in his terminology), were distinguishable in (I965) (who have also confirmed the Lewis associa- Caucasians, either A or B, or both, being present. tion), and in America by Shreffler (I965). Shref-

A third pregnancy band was found only in Negroes. fler, who has carried out semi-quantitative classi- http://jmg.bmj.com/ Complete agreement between bands in sera and fication of phosphatase bands, points out that, on placental extracts was found, except that examina- his data, not all blood group Al individuals show tion of placental extracts showing both A and B complete absence of zone B protein, indicating zones revealed an intermediate zone AB, a distinc- that this association is not absolute. He also shows tion not possible with serum at that time. that secretor status (rather than Lewis grouping) Beckman, Bjorling, and Christodoulou (I966), is the primary factor (non-secretors of ABH sub- using a modified technique described by Ashton stances being usually Le (a+b-)) when analysis is and Braden (I96I), have succeeded, in some cases, made of the correlations mentioned above. on September 27, 2021 by guest. Protected in distinguishing 3 zones in sera from subjects Thus, the appearance of this second alkaline classified as AB. The distribution of the main phosphatase (i.e. zone B) is apparently determined types, A, AB, and B, corresponds approximately to by at least three loci. Its origin is considered to be a Hardy-Weinberg distribution, but with deficiency the intestinal mucosa (Hodson, Latner, and of type AB amounting to a loss of I5%. Further Raine, I962; Weiser, Bolt, and Pollard I964). data on this deficiency of AB types will be awaited Zone A alkaline phosphatase originates in liver. with interest, for it may represent an example of It is certainly a remarkable thing that an enzyme intrauterine selection against a heterozygote. Boyer of intestinal origin should in any way be controlled (I96I) using presence or absence of the A band as by loci associated with blood group substances for the only safe classification (at that time) showed a the interrelation is without rhyme or reason in our significant variation in the presence of A between present state of knowledge. Nigerians (I2%) and Caucasians (54%). In From Beckman's data (I964) on family studies, American Negroes (who are considered to be 70% it is evident that no simple explanation can be African and 30% Europeans in genetic origin) found that will enable genotypes to be deduced, the expected frequency of the hypothetical 'A' using Ppi and Pp2 as phenotypes. J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from 58 J. Price VII: Erythrocyte Adenylate Kinase their genotype is, therefore, E2+E2-, where Adenylate kinase, also known as myokinase, is E2- is the gene usually present at the E2 locus a phosphotransferase which has been principally and E2+ the variant gene. The technique used studied in muscle, though it also occurs in other (starch gel electrophoresis) does not distinguish the tissues, including red blood cells. It catalyses the phenotype corresponding to the homozygote E2+ following reaction: 2ADP --ATP +AMP. E2+. The finding of C5- parents with C5+ off- Fildes and Harris (I966) report the examination spring on several occasions does not agree with of haemolysates of 960 unrelated English people this hypothesis, but the authors suggest that 'a few for this enzyme, using starch gel electrophoresis. of the heterozygotes may not be demonstrable by Three patterns emerge, designated AKi, AK2-I, the electrophoretic and staining methods at present and AK2. Family studies are consistent with a employed'. hypothesis of two autosomal alleles without domi- nance, so that, for the 3 phenotypes given above, E1 Variants. Whereas subjects phenotyped there are three genotypes: AKl/AK', AK'/AK2, as C5 + (corresponding to a variant at E2) have and AK2/AK2, respectively. Only one example of increased serum cholinesterase, subjects who have the AK2 phenotype was found in 960 unrelated atypical phenotypes with respect to the independent individuals mentioned (though three further ex- E1 locus have diminished serum cholinesterase amples were discovered during family studies). 95 activity. Heterozygotes for an atypical gene Ela individuals were AK 2-I, giving a gene frequency are E,a/E,u, where Eiu is the usual gene; such of o os for AK2 in English populations. subjects form 4% of a Toronto population and have Disease associations have not yet been studied, about 75%0 of normal serum cholinesterase activity. nor is anything known of other selective forces Subjects with both E,u and E2+ (usually ElUlE, which must be operating here. E. ± /EE.-) have serum cholinesterase activity somewhat above normal. Among subjects with genotype ElaEla selected by the occurrence of VIII: Serum Cholinesterase prolonged apnoea from suxamethonium, no certaincopyright. This enzyme is sometimes called pseudocholines- example of C5+ phenotype was detected (Robson terase. Red cell cholinesterase (true cholinesterase) and Harris, I966). This suggests that the effect of is a different enzyme having different properties. E2+ may afford protection against the sensitivity Cholinesterases catalyse the hydrolysis of choline to suxamethonium brought about by ElalEla geno- esters, e.g. Acetyl choline -+choline + acetic acid. type effect.

Serum cholinesterase production appears to be The atypical enzymes I (corresponding to E1u/http://jmg.bmj.com/ under the control of at least two loci, E1 and E,. E2 Ela) and A (coiresponding to ElalEla) are known to variants, though less well known than E1 variants, be inhibited to a lesser degree by dibucaine and the will be discussed first because the percentage solanine alkaloids (found in potatoes) than is the frequency of these is considerably higher than that usual enzyme U (corresponding to Elu/Elu). This of E, variants. might mean that subjects possessing the Ela gene could resist solanine poisoning (an exceedingly rare E2 Variants: C5+. Robson and Harris (I966) condition at the present time, which produces have recently produced further population and diarrhoea and vomiting). Harris and Whittaker on September 27, 2021 by guest. Protected family studies on the serum phenotype C5 + (I962) conclude that it is not clear how far the toxic which was reported in earlier papers by Harris, effects of solanine can be attributed to the inhibition Hopkinson, and Robson (I962) and Harris, Hop- of serum cholinesterase. kinson, Robson, and Whittaker (I963). The inci- To know what advantage such a gene as Ela dence of C5 + in the British sample of I941 persons might confer, it would be useful to know if there is o0O97. It is 029 in one of two Greek village are any circumstances in which the variant cholines- populations studied. terase might be directly advantageous. Naturally Simpson (I966) reports that the frequency of occurring dietary substrates are unlikely to pass the C5+ phenotype is about 7.9°0 in a Brazilian the intestinal wall unchanged, but naturally occur- population, and gives details of serum cholinester- ring inhibitors (which inhibit the atypical form less ase activity (which is greater than normal in C5+ than the usual enzyme) can do so. The calabar- subjects) in subjects genotyped at both E1 and E, bean ordeal used as a judicial procedure in certain loci. primitive tribes to separate the guilty (who die as It seems likely that most C5+ individuals are a consequence of eserine poisoning) from the inno- heterozygotes for a gene at a locus designated E2; cent (who do not) is an example of a selective force J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from Human Polymorphism 59 that should favour E,a; but the geneticist would detect variation of mobility during immuno- probably hesitate before accepting this as an electrophoretic analysis of human sera. To these example of pleiotrophy in which Ela effect deter- 2 globulins he gave the name Group-specific mines both an atypical cholinesterase and moral component (Gc). Three immunoelectrophoretic rectitude. Naturally occurring inhibitors have been patterns could be distinguished: a fast-moving described in sugar-beet and oranges (Menn, Mc- pattern, Gc i-I; a slow-moving pattern, GC 2-2; Bain, and Dennis, I964), and in apples (Orgell, and an intermediate pattern, GC 2-I which behaved I963). in a manner similar to that of an admixture of approximately equal parts of Gc i-i and GC 2-2. H: PLASMA PROTEIN Further study showed that the mode of inheri- POLYMORPHISMS tance for this system is that of a two-allele system without dominance. Genotypes Gcl/Gc1, Gc2/Gcl, The protein polymorphisms to be described are and Gc2/Gc2 correspond to the phenotypes Gci-i, the following: (i) prealbumin, an oca globulin; (2) Gc, GC2-I, and GC2-2 (Hirschfeld, I960). No excep- an o2 globulin; (3) haptoglobin, an °X2 globulin; tions have been found to this rule, a fact that has (4) transferrin, a globulin; (5) lipoprotein systems, contributed to the value of the system in investi- globulins; and (6) Gm+Inv systems, y globulin. gating cases of disputed paternity (Reinskou, I966; The names given above (i.e. a1 .C2, (3, and y) Nerstrom, I963; Hirschfeld and Heiken, I963). refer to the zones which the proteins occupy after Population studies have been carried out. Hirsch- serum is subjected to paper electrophoresis. Such feld (I962) and Salzano and Shreffier (I966) give a form of electrophoresis is too to altogether crude tables of data. Frequencies of the Gc2 gene vary separate satisfactorily the components of the from 0-025 in Navajo Indians and O-OI in Australian individual com- proteins listed above; methods aborigines to o-67 in Xavante Indians of the monly used for study of these proteins include agar Brazilian Mato Grosso studied by Neel, Salzano, gel and starch gel electrophoresis (which give much and better resolution of the various components), and Junqueira, Keiter, Maybury-Lewis (I964). In Europeans, values for Gc2 are around 0-25 copyright. immunoelectrophoresis and gel diffusion precipita- and in African Negroes somewhat below oI.v tion. The serum enzyme protein polymorphisms Kirk, Cleve, and Bearn (I963) give population and the blood group polymorphisms are not in- cluded here, being considered separately in view studies for the Gc genes from South East Asia of their different properties and the differences in and Australian aborigines. In 45 subjects from technique used in assessing them. New Guinea there were five heterozygotes

and two homozygotes for the variant allele Gc http://jmg.bmj.com/ Aborigine, first described for aboriginal Australians. I: Prealbumin These data suggest that the variant gene has at- Fagerhol and Braend (I965) report the results tained polymorphic status in this locality, and this from 390 Norwegian blood donors of examination may apply also in the New Hebrides (Kirk, I965). of the prealbumins, so-called because they migrate A further variant, Gc Chippewa, has been described in front of the albumins in zone electrophoresis by Cleve, Kirk, Parker, Bearn, Schacht, Kleinman, of serum in starch gel. Five types were distin- and Horsfall (i963a) in Chippewa Indians, with guishable and designated MM, MS, SS, FM, FS. 5 of 59 sera abnormal; other rarer variants have on September 27, 2021 by guest. Protected Only i6 donors were not MM, and of these, 9 also been reported. The subject of Gc variant were MS and 2 were SS. Studies on two families phenotypes has been discussed recently by Kitchin indicate that the prealbumins are inherited charac- and Beam (I966). teristics. The authors suggest a system of 3 Chemical differences between Gci-i and Gc2-2 codominant alleles PrF, Prm, Prs. Using their data, (Cleve, Prunier, and Beam, I963) appear to he the frequency of PrS is about 0-02. More studies very slight and the structural differences respon- will be needed before genetic drift can be excluded sible for the variation in electrophoretic mobility as being responsible for this situation. may well reside in a single amino acid substitution. II: Gc: Group Specific Component Human a2 globulin is a complex mixture of mI: Haptoglobin Polymorphism proteins having similar electrophoretic mobilities. Between I938 and I940 Polanovski and Jayle Haptoglobin, caeruloplasmin, and a2 macro- described, in a series of papers, an a2 globulin in globulin are major components; among several human serum, which had the unique property of minor components Hirschfeld (I959) was able to binding haemoglobin. This was named hapto- J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from 60 J. Price globin, and in I955 Smithies established (using This hypothesis has been further debated by Sutton starch gel technique) that haptoglobins were in- and Karp (I964), and by Parker and Bearn (I965). herited as a two allele system, with no dominance. On neither hypothesis, as it stands, can a genetic- ally-based explanation be found for the presence Genetics. The two alleles determine 3 common ofHp 0 in Australian and New Guinean populations phenotypes Hp i-i, Hp 2-I, and Hp 2-2, corres- in the virtual absence of Hp 2-IM in these popula- ponding to the genotypes Hp'/Hpl, Hp'/Hp2, Hp2/ tions. Curtain, Gajdusek, Kidson, Gorman, Champ- Hp2. Chemical reduction of purified haptoglobin ness, and Rodrigue (I964) have found low serum followed by starch gel electrophoresis (using 8 M haptoglobin levels in the lowlands of New Guinea urea and formic acid buffer) has enabled the compared with the highlands, and a higher frequency Hp i-i phenotype to be separated further into of Hp 0 in the lowlands than in the highlands. three subphenotypes, Hp iF-iF, Hp iF-iS, Hp iS- They suggest that these findings may be due to iS, these being determined by two alleles, HpiF removal of haptoglobin by increased vascular and Hpi S, at the haptoglobin locus (Connell, haemolysis, which is the commonest form of ac- Dixon, and Smithies, I962). quired ahaptoglobinaemia and which is related Further analysis by Smithies, Connell, and Dixon to the higher incidence of malaria in the lowland (i962a, b) has shown that haptoglobin can be split areas. into a and 3 polypeptide chains; the genetic It is worth emphasizing that Hp 0 (ahapto- determinants of haptoglobin phenotype reside in globinaemia) is the most difficult phenotype to be the a chain. Hp2 units represent the fusion of two sure of. There are a number of reports, including Hp' units (a,) into one molecule. Hp i-i consists those of Galatius Jensen (I958) and Blumberg and of a single molecule with two oal and one 3 chain. Gentile (i96i), of individuals who are ahapto- Hp 2-2 consists of a series of polymers of structure globinaemic at one time, but have detectable (P.20C,)2n;-1 while Hp 2-I consists of polymers which haptoglobin at another. Severe liver disease may contain a unit of Hp I-I in combination with one also distort phenotyping, as may infectious diseases of the Hp 2-2 polymers. These Hp 2-I polymers, other than malaria, and renal disease, etc. (seecopyright. because of their different structure, behave quite Shinton, Richardson, and Williams, I965). differently from the Hp 2-2 polymers on starch gel electrophoresis (see Allison, I959; Javid, I964; Haptoglobins and Disease. A possibility, Smithies, I964). suggested by Curtain et al. (I964) from his data, is The additional discovery of two further relatively that there is a selective pressure favouring Hp'

common phenotypes, HP2-IM (Connell and in lowland areas for, 'since sera of type i-i and http://jmg.bmj.com/ Smithies, I959; Giblett and Brooks, I963) and type 2-I individuals appear to bind more haemo- Hp 0 (hereditary ahaptoglobinaemia), with 25 globin per I00 ml than does those of type 2-2, possible phenotype matings [(3+2)2], presents a this could place the latter at a disadvantage in the formidable analytical problem. maintenance of the delicate iron balance in the Parker and Bearn (I963) have postulated control presence of haemolysis, intestinal iron loss due genes C' and C2 for the two structural genes to parasitic infection and dietary iron deficiency'. Hp' and Hp2. These control genes may be active or However, it appears that the haptoglobin passive (the latter being represented by Cl-, C2-). system in all its phenotypes is quite unable to on September 27, 2021 by guest. Protected There are, thus, 8 possibilities: C'Hpl, ClHp2, cope with massive haemolysis; indeed, it can Cl-Hpl, Cl-Hp2, C2Hpl, C2Hp2, C2-Hp', C2-Hp2 hardly bear the responsibility for the recovery of iron from There are 36 genotypes (N8.) The basic tenet defunct erythrocytes under normal circumstances. However, a further report by made in setting up the hypothesis is that the control Buettner-Janusch and Buettner-Janusch (I964) element C' produces a substance with a higher showing a high Hpi-i incidence in the coastal affinity for Hp' than Hp2. Thus, in genotypes areas of Madagascar (which have a high malaria where Hp2 is present without C2, the structural incidence) compared with that of the highlands gene (Hp2) lacks controUing substances so that a (where malaria is much less common) suggests modification, leading to phenotype 2-IM, is pro- that the hypothesis that Curtain has put forward duced (unless both controlling genes are absent, warrants further investigation. An association in which case the phenotype is Hp 0). between the thalassaemia gene and Hpi has This hypothesis is an alternative to that of been described in Greece (Blumberg, Murray, Giblett and Steinberg (i960), who have suggested Allison, Barnicot, Hirschfeld, and Krimbas, I964). a third allele Hp m at the haptoglobin locus. Blackwell, Chen, Chen, and Uylangco (I964) J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from Human Polymorphism 6I have found the incidence of cholera unrelated to two daughters being phenotyped as B.C and CD, Hp phenotype among Fillipinos. showing that three transferrin types are segregating in this family. Population Studies. There are numerous In only a few instances is the heterozygote population studies on haptoglobin frequency in frequency percentage of variant transferrins so different populations. Shim and Beam (I964) high that maintenance by mutation is unlikely: summarize these (including data for Hp1F and Navajo Indians BO-, 8% HpIS) up to mid-I964. Salzano and Sutton (I965) Caucasians B2 : I% summarize data for the American continent and Negroes D1 I2% point out that the apparent cline of increasing gene Chinese Dchi : 6% frequency for Hpi noted by Sutton, examining Kirk, Parker, and Beam (I964) have reviewed early data (up to i96o) from North to South in the data on D1 and Dchi The D1 variant is found American continent, has become blurred with both in African Negroes and in Aborigines (from further study, so that very large variations in gene Australia and Melanesia) who, on other evidence frequency are found within small areas. It has been (blood group and serum protein gene frequencies) pointed out that clines are ultimately the product of are of different genetic origin. Wang and Sutton two conflicting forces-selection (leading to unique (I965), using chymotrypsin digests, present evi- local adaptation) and gene flow (leading to uni- dence that C and D1 proteins differ consistently, formity). Since cultural patterns evolve indepen- in that a glycine residue on the D1 polypeptide dently and population movements take place chain replaces an aspartic residue on the C chain. independently of these forces, clines are not likely Dchi is found in Malaya, Thailand, Formosa, to be found in human populations (Mayr, I963). North East India, and the Veddah of Ceylon. Hp phenotyping has proved useful in the study It has also been found in the Yupa Indians of of a hybrid population in Chile (Nagel and Soto, Venezuela (Arends and GaUango, I964) where I964) and in supporting the Indian origin of a it attains a high frequency: CIDchi = 418%; group of Swedish gypsies (Beckman, Takman, DchilDchi= I6'4%. Salzano and Sutton (I965) copyright. and Arfors, I965). Arends and Gallango (I965) give a table of frequencies for the American conti- have studied Guiana Indians, and give a table of nent. There is also a summary of data by Giblett South American data. The absence of Hp 2-IM (I962). is taken as indicating the absence of Negro admix- The Navajo, Caucasian, Negro, and Chinese ture. HpO occurring in the same population suggests populations may represent states of balanced poly- the acquired form of ahaptoglobinaemia, due to its

morphism. So far, studies have shown no significant http://jmg.bmj.com/ commonest cause - malaria. difference in the rate of disappearance of iron from the serum containing the variant proteins. One function possessed by transferrin by virtue of its IV: Transferrin Polymorphism iron-binding function is as an antibacterial agent, Transferrin, a glycoprotein, is the serum protein its mode of action depending on the unavailability to which iron is bound. It is a n-globulin with a of iron to enable bacteria to grow. This effect, molecular weight close to go,ooo and it is thought it has been claimed, might convey differing anti- to contain a single polypeptide chain. The type bacterial resistance to possessors of variant proteins, on September 27, 2021 by guest. Protected most commonly found is designated C, and this, but as the capacity of these proteins for binding according to Jeppsson and Sjoquist (I963), is not iron appears to be normal, it is difficult to see how one protein but two, the two fractions differing in this can work. their carbohydrate composition. Ashton (I965) has studied the polymorphism Seventeen variants have been described; 9 of serum transferrins in cattle. He gives data show- of these move more slowly on starch gel electro- ing that the reproductive performance of cattle is phoresis than Type C and these are designated D influenced by several effects of the serum transferrin group; 8 move more rapidly than C-the B group. locus: (a) advantage in utero; (b) maternal-foetal Each member of groups B and D is numbered or incompatibility; and (c) fertility which differs named. Data on the inheritance of these groups between phenotypes in both cows and buUs. Ashton have not yet been established in all cases, but strong has constructed a model for these effects, which support is to be found for an autosomal allelic shows a comparison between expected and observed system without dominance, with alleles at a single breeding efficiencies with analysis for each of the locus. Beckman (I962) has reported an informative effects. mating (B2D1 x CC), from a Swedish population, The estimated population excess of heterozygotes J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from 62 J. Price is very close to that observed. It will be of great rheumatic fever or coronary artery disease between interest to see whether a similar model can be the two groups. No consistent differences have constructed for the human polymorphisms which been found in diabetics for, though Blumberg, exist for the transferrin proteins. Ledbetter, and Visnich (Alison and Blumberg, I965) found a higher proportion of Ag(a+) subjects (than expected) in a group of diabetic V: Lipoprotein Polymorphisms American Negroes, this was not the case in a The Ag System. Allison and Blumberg (I96I) Norwegian and Brazilian population studied by found that serum from a patient (C de B) who Berg, Blumberg, and Leon (quoted by Alison and had received multiple transfusions gave well- Blumberg). The selective factors maintaining the defined precipitation reactions in agar with some, polymorphism thus remain to be established. but not all, sera from other individuals. The antigen in these sera was shown by family studies The Lp and Ld Systems. Further lipoprotein to be inherited independently of other systems polymorphisms have been described by Berg such as haptoglobin, transferrin, and y-globulin (I963) (the Lp system) and Berg (I965) (the types. Possessors of it were designated Ag (a+); Ld system). Gedde-Dahl and Berg (I965) give non-possessors Ag(a-). Ag(a-) subjects were further data on the first of these systems. shown to be homozygous for a recessive gene Ag: Ag(+) subjects homozygous or heterozygous for the allelic gene designated AgA (Allison and VI: Gm and Inv Systems Blumberg, I96I; Blumberg and Allison, I96I; Sera from children who have received muttiple Blumberg, Bernanke, and Allison, I962). transfusions have also been shown to contain The antigen has been identified as a low density agglutinating antibodies against y-globulin absent protein which migrates as an a2 globulin in agar from their own y-globulin (Allen and Kunkel, gel electrophoresis but as a globulin on paper I963a, b). The antigens responsible for this effect electrophoresis. Further studies have shown that it are due to genetically determined factors-thesecopyright. is a 5-lipoprotein (Blumberg, Dray, and Robinson genes being present as multiple alleles at a genetic (I962), and it seems possible that a carbohydrate locus (Gm) or at several closely linked loci. component (which all lipoproteins contain) contri- If 0 positive DD (Rhesus) cells are (i) coated butes to its antigenicity. with incomplete anti-D antibody containing Gm(a), The subject has become complicated by the fact these cells can be (ii) agglutinated by the serum of that various antibody-containing sera studied certain with rheumatoid arthritis which subjects http://jmg.bmj.com/ (of which serum C de B was the first) contain contains anti Gm(a). This agglutination can be antibodies to more than one antigen (such anti- (iii) inhibited by y globulin from possessors of bodies are not found at all until more than 20 Gm(a) (Grubb and Laurell, 1956). transfusions have been given (Blumberg et al., Thus, a system showing this reaction must I964)). Serum C de B contains at least three contain:- (i) an appropriate coating or sensitizing antibodies (reacting with 3 distinct antigens) serum; (ii) an appropriate agglutinating serum; and so that, though the classification of subjects as of (iii) an appropriate inhibiting serum. phenotype Ag(a-) remains satisfactory, it now Further Gm factors (factors in category (iii)) on September 27, 2021 by guest. Protected indicates that they are homozygous for the absence have been detected by finding agglutinating sera of at least three genetic factors (determining which have reacted with some, but not all, of specificities al, x, and z). Hirschfeld (I964) has inhibiting serum known to show the Gm (a) factor reviewed this complex subject. (or other known Gm factors). Examples of this It should be clear from the foregoing that popula- are Gm(b) (Harboe, I959); Gm(X) (Harboe and tion studies are not meaningful if the phenotypes Lundevall, I959); Gm(r) (Fudenberg, I96I); Gm- are simply designated Ag(a+) and Ag(-), for like (Steinberg, Giles, and Stauffer, I960); Gm(D) Ag(a+) can indicate Ag(al+; x-, z-) or Ag (Thomas and Kampf, I96I); and Gm(e) (Ropartz, (al+; x+-; z-) etc. Allison and Blumberg (I965) Rivat, and Rousseau I962). All Gm factors have have tabulated the population study data. in common the property of being yG-globulins No differences have been found in total choles- (otherwise 7S y-globulins), and differences between terol, total phospholipid, or cholesterol/phos- them are determined by differences in the heavy pholipid ratio, between Ag(a+) and Ag(-) (H) polypeptide chains of the y-globulin molecule. subjects (Blumberg et al., I962); furthermore, A second group of factors, independent of Gm, no differences have been found in the incidence of associated with all three main classes of proteins J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from Human Polymorphism 63 possessing antibody activity (yG, yM, and yA), (4) Anilysis of adaptive trends in genotype and determined by light polypeptide chain (L frequency with age or among successive chain) differences in the y-globulin, has been generations. described: this is the Inv system. Descriptions (5) Detection of different genotype frequencies include those of Inv(a) (Ropartz, Lenoir, and in two sexes. Rivat, I96I); Inv(b) (Steinberg, Wilson, and (6) Measurement of fertility and mortality Lanset, I962); and Inv(l) (Ropartz et al., I962). differentials among genotypes. The genetics of the Inv locus have been discussed (7) Analysis of departures from Mendelian by Ropartz (I963). segregation frequencies. The subject of Gm systems has become very Examples of the use of these various methods complex, particularly in considering Negro and are included under the individual polymorphisms Mongoloid populations, and diversity of nomen- which have been described above. clature has not made understanding of these com- The subject of human polymorphism may con- plexities any easier. No genetic hypothesis seems fidently be expected to proliferate over the next to exist, which will explain all the rare phenotypes few years. It is to be hoped that attention will encountered. In Caucasians Gm (a) and Gm (b) continue to be paid to the part played by poly- appear to behave as the products of different alleles. morphic situations in human disease, and that the This is not true in Negroes, for 95% of American search for further polymorphic systems will be Negroes are Gm (a+b+). With mating of 2 sub- related to the need to find an explanation for those jects of genotype Gma/Gmb, only 5000 of offspring diseases in which inheritance plays an important will be Gm (a + b +), and, as Steinberg et al. aetiological role. (I960) have pointed out, 50% of conceptions would have to end in death to maintain heterozygosity- I wish to acknowledge my indebtedness to Dr. D. A. an untenable hypothesis. Therefore, a further Price-Evans for assistance with the biochemical sections allele Gmab has to be postulated. Gmar, Gmax, of this paper and to Dr. D. J. Weatherall for help with Gmarx, and Gmabc have also been described. the section on the haemoglobinopathies. The paper was read and criticized by Professor and Mrs C. A. copyright. Further data on this subject (which seems likely Clarke and by Dr. R. B. McConnell, and I am most to become a great deal more complex or a great grateful to them for their advice. I also wish to thank deal simpler if some current ideas prove ultimately Mrs. K. M. Cogley for secretarial assistance. untenable) can be found in Deicher, Wendt, Theile, and Kirchberg (I963), Ropartz, Rivat, Rousseau, Baitsch, and Van Loghem (I963), and REFERENCES

Aird I., Bentall, H. H., and Fraser Roberts, J. A. (I 953). A relation- http://jmg.bmj.com/ Steinberg and Goldblum (I965). ship between cancer of stomach and the ABO blood groups. Attempts to relate variations in the frequencies Brit. med. J., 1, 799. ofinherited y-globulin groups to certain pathological Akesson, H. 0. (I959). Taste sensitivity to phenyl-thio-urea in tuberculosis and diabetes mellitus. Ann. hum. Genet., 23, 262. conditions have not, as yet, proved successful. Allen, J. C., and Kunkel, H. G. (I963a). Antibodies to genetic types Data on population studies are to be found in of gamma globulin after multiple transfusions. Science, 139, 4I8. , and (I963b). Characteristics of the antibodies against papers by Ropartz et al. (I963), Steinberg (I962), gamma globulin in patients with multiple transfusions. J. clin. and Steinberg and Goldblum (I965). Invest., 42, 913. Allison, A. C. (1954). The distribution of the sickle-cell trait in East Africa and elsewhere, and its apparent relationship to the on September 27, 2021 by guest. Protected incidence of subtertian malaria. Trans. roy. Soc. trop. Med. Hyg., SUMMARY 48,3I2. (1957). Malaria in carriers of the sickle-cell trait and in new- Morton, Krieger, and Mi (I966) discuss several born children. Exp. Parasit., 6,4I8. aspects of polymorphism and give a list of methods (1959). Genetic control of human haptoglobin synthesis. Nature (Lond.), 183, 1312. (and the errors thereof) used to study the selective (1963). Malaria and glucose-6-phosphate dehydrogenase forces involved in maintaining polymorphisms. deficiency. ibid., s97, 609. (i) Identification of the (i965). Population genetics of abnormal haemoglobins and relevant environmental glucose-6-phosphate dehydrogenase deficiency In Abnormal differences between populations with high Haemoglobins in Africa, ed. J. H. P. Jonxis, p. 365. Blackwell and low gene frequencies. Scientific Publications, Oxford. , and Blumberg B. S. (i96i). An isoprecipitation reaction (2) Detection of systematic departures of geno- distinguishing human serum-protein types. Lancet, I, 634. type frequencies from Hardy-Weinberg and (i965). Serum lipoprotein allotypes in man. Progr. med. Gznet., 4, 176. equilibrium. Anderson, M., Went, L. N., MacIver, J. E., and Dixon, H. G. (3) Association between genotype and a specific (1960). Sickle-cell disease in pregnancy. Lancet, 2, 5i6. type of morbidity or response to a specific Arends, T., and Gallango, M. L. (i964). Transferrins in Venezuelan Indians: high frequency of a slow-moving variant. Science, 143, agent. 367. J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from 64 J. Price

and (I965). Haemoglobin types and blood serum factors Brit. med. J. (I965). Clinical trial of the treatment of depressive in British Guiana Indians. Brit. J. Haemat., II, 350. illness. Report to the Medical Research Council by its Clinical Arfors, K. E., Beckman, L., and Lundin, L. G. (I963). Genetic Psychiatry Committee. I, 88I. variations of human serum phosphatases. Acta genet. (Basel), Brues, A. M. (1954). Selection and polymorphism in the A-B-O 13, 89. blood groups. Amer. J. phys. Anthrop., 12, 559. Armstrong, A. R., and Peart, H. E. (I960). A comparison between Buettner-Janusch, J., and Buettner-Janusch, V. (I964). Hemo- the behavior of Eskimos and non-Eskimos to the administration globins, haptoglobins and transferrins in the peoples of Mada- of isoniazid. Amer. Rev. resp. Dis., 8I, 588. gascar. ibid., 22, I63. Armstrong, M. D., Yates, K., Kakimoto, Y., Taniguchi, K., and Burke, J., De Bock, G., and De Wulf, 0. (1958). La drepanocy- Kappe, T. (I963). Excretion of ,l-aminoisobutyric acid by man. traemie simple et l'anaemia drepanocytaire au Kwango (Congo J. biol. Chem., 238, I447. Belge). Mem. Acad. roy. Sci. colon. Cl. Sci. nat. med., 7, I. Ashton, G. C. (I965). Cattle serum transferrins. A balanced poly- Cabannes, R. (I965). Repartition des hemoglobines anormales morphism? Genetics, 52, 983. dans la partie ouest du continent africain (with English Summary). , and Braden, A. W. H. (I96I). Serum ,3-globulin polymorphism In Abnormal Haemoglobins in Africa, ed. J. H. P. Jonxis, p. 29I. in mice. Aust. J. biol. Sci., 14, 248. Blackwell Scientific Publications, Oxford. Azevedo, E., Krieger, H., Mi, M. P., and Morton, N. E. (I965). P.T.C. taste sensitivity and endemic goiter in Brazil. Amer. Carcassi, U., Ceppellini, R., and Pitzus, F. (I957). Frequenza della J. hum. Genet., 17, 87. talassemia in quattro popolazioni sarde e suoi rapporti con la K. E. and distribuzione dei gruppi sanguini e della malaria. Boll. ISt sieroter Bamford, F., Harris, H., Luffman, J. E., Robson, B., milan., 36, 206. Cleghorn, T. E. (I965). Serum-alkaline-phosphatase and the Choremis, C., Fessas, Ph., Kattamis, C., Stomatoyannopoulos, G., ABO blood groups. Lancet I, 530. Zannos-Mariolea, L., Karaklis, A., and Belios, G. (i963). Three Bannerman, R. M. (I962). Selection in thalassaemia. In Haemo- inherited red-cell abnormalities in a district of Greece. Thalas- globin-Colloquium, ed. H. Lehmann and K. Betke, p. 93. Thieme, saemia, sickling, and glucose-6-phosphate dehydrogenase de- Stuttgart. ficiency. Lancet, I, 907. Barnicot, N. A. (1950). Taste deficiency for phenylthiourea in Chung, C. S., Matsunaga, E., and Morton, N. E. (i96i). The African Negroes and Chinese. Ann. Eugen. (Lond.), 15, 248. MN polymorphism in Japan. Jap. J. hum. Genet., 6, i. , Allison, A. C., Blumberg, B. S., G., Deliyannis, Krimbas, , and Morton, N. E. (I96I). Selection at the ABO locus. C., and Ballas, A. (I963). Haemoglobin types in Greek popula- Amer.J. hum. Genet., 13, 9. tions. Ann. hum. Genet., 26, 229. Ciba Foundation (I965). Symposium on Colour Vision, ed. A. V. S. Beckman, L. (I962). Slow and fast transferrin variants in the same Reuck and J. Knight. Churchill, London. pedigree. Nature (Lond.), 194, 796. (I964). Associations between human serum alkaline phos- Clarke, C. A. (I964). Genetics for the Clinician, 2nd ed., p. 76. phatases and blood Acta 286. Blackwell Scientific Publications, Oxford. groups. genet. (Basel), 14, Clements, F. W. (1g60). Naturally occurring goitrogens. Brit. med. , Bjorling, G., and Christodoulou, C. (I966). Pregnancy Bull., I6, 133. enzymes and placental polymorphism. I. Alkaline phosphatase. R. W. A. L. ibid., I6, 59. Cleve, H., Kirk, L., Parker, C., Beam, G., Schacht, E.,

Kleinman, H., and Horsfall, W. R. (I963a). Two genetic variantscopyright. , Takman, J., and Arfors, K. E. (I965). Distributions of blood of the group-specific component of human sera-Gc Chippewa and serum groups in a Swedish Gypsy population. ibid., IS, and Gc Aborigine. Amer. J. hum. Genet., IS, 368. 134. Berg, K. (I963). A new serum type in man-the Lp system. Acta -, Prunier, J. H., and Bearn, A. G. (I963b). Isolation and partial path. microbiol. scand., 59, 369. characterization of the two principal inherited group-specific (i965). A new serum type in man-the Ld system. Vox Sang. components of human serum. 7. exp. Med., II8, 711. (Basel), 10, 513. Connell, G. E., Dixon, G. H., and Smithies, 0. (i962). Subdivision , Blumberg, and Leon (quoted by Allison and Blumberg, of the three common haptoglobin types based on 'hidden' dif- I965). ferences. Nature (Lond.), 193, 505. Biehl, J. P. (I956). The role of the dose and the metabolic fate of , and Smithies, 0. (0959). Human haptoglobins: estimation http://jmg.bmj.com/ ioniazid in the emergence of isoniazid resistance. Trans. Conf. and purification. Biochem.J.7, 72, 115. Chemother. Tuberc. 15, 279. Covarrubias, E., Barzelatto, J., Stevenson, C., Bobadilla, E., Pardo, (1957). Emergence of drug resistance as related to the dosage A., and Beckers, C. (i965). Taste sensitivity to phenyl-thio- and metabolism of isoniazid. ibid., I6, Xo8. carbamide and endemic goitre among Pewenche Indians. ibid., Blackwell, R. Q., Chen, H. H., Chen, H. C., and Uylangco, C. V. 205, 1036. (I964). Haptoglobin distribution in a Filipino population. Nature Curtain, C. C., Gaidusek, D. C., Kidson, C., Gorman, J., Champ- (Lond.), 202, 814. ness, L., and Rodrigue, R. (1964) (quoted by Kirk, 1965). Blumberg, B. S., and Allison, A. C. (I96I). Study on the isoprecipitin Das, S. R., and Mukherjee, D. P. (1964). P.T.C. taste sensitivity determined human serum polymorphism. Proc. 2nd int. Congr. survey among the Pareng Gadaba, the Ollaro Gadaba and the

Hum. Genet. Rome, 2, 733. Konda Paroja of Koraput District, Orissa. Acta genet. (Basel), on September 27, 2021 by guest. Protected , Bernanke, D., and Allison, A. C. (i962). A human lipoprotein I4, i68. polymorphism. J. clin. Invest., 41, 1936. - ,.1 ,and1 Bhattacharjee, P. N. (i963). P.T.C. taste threshold , Dray, S., and Robinson, J. C. (i962). Antigen polymorphism distribution in the Bado Gadaba and the Bareng Paroia of Koraput of a low-density beta-lipoprotein. Allotypy in human serum. District in Orissa. ibid., I3, 369. Nature (Lond.), 194, 656. Davidson, R. G., Childs, B., and Siniscalco, M. (i964). Genetic , and Gentile, Z. (i96i). Haptoglobins and transferrins of two variations in the quantitative control of erythrocyte glucose-6- tropical populations. ibid., x89, 897. phosphate dehydrogenase activity. Ann. hum. Genet., 28, 6i. -, Ledbetter, and Visnich (quoted by Allison and Blumberg, Deicher, H., Wendt, G. G., Theile, U., and Kirchberg, G. (i963). I965). Familienuntersuchungen uber die Gammaglobulingruppen Gm(a), Murray, R. F., Allison, A. C., Barnicot, N. A., Hirschfeld, Gm(b), Gm(x) und Gm(r). Acta genet. (Basel), 13, 124. J., and Krimbas, C. (i964). Serum protein polymorphisms in Delbrouck, J. (1958). Contribution i la genetique de la sicklemie. Greek populations. Ann. hum. Genet., 28, I89. Maintien de la frequence evelee de sicklemie au Congo belge. Bonicke, R., and Reif, W. (i953). Enzymatische Inaktivierung von Ann. Soc. belge Mid. trop., 38, I03. Isonicotinsaure-hydrazid in menschlichen und tierischen Organis- Dern, R. J., Brewer, G. J., Tashian, R. E., and Shows, T. B. (1966). mus. Naunyn-Schmiedeberg's Arch. exp. Path. Pharmak., 220, Hereditary variation of erythrocytic 6-phosphogluconate dehydro- 321. genase. J. Lab. clin. Med., 67, 255. Boyer, S. H. (i961). Alkaline phosphatase in human sera and Devadatta, S., Gangadheram, P. R. J., Andrews, R. H., Fox, W., placentae. Science, 134, 1002. Ramakrishnan, C. V., Selkon, J. B., and Velu, S. (ig60). Peri- Brand, N. (i963). Taste sensitivity and endemic goitre in Israel. pheral neuritis due to isoniazid. Bull. Wld Hlth Org., 23, 587. Ann. hum. Genet., 26, 321. Dreyer, V., and Goldschmidt, E. (x965). The offspring of two (1964). Taste response and poliomyelitis. ibid., 27, 233. colour-blind parents. Acta genet. (Basel), 15, 97. J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from Human Polymorphism

Dufour, A. P., Knight, R. A., and Harris, H. W. (I964). Genetic Fullerton, W. T., Hendrickse, J. P. de V., and Watson Williams, of isoniazid metabolism in Caucasian, Negro, and Japanese E. J. (i965). Haemoglobin SC disease in pregancy. In Abnormal populations. Science, 145, 391. Haemoglobins in Africa, ed. J. H. P. Jonxis, p. 4Ii. Blackwell Dutta, P. C. (I966). A review of the inherited defective colour- Scientific Publications, Oxford. vision variability and selection relaxation among the Indians. Galatius-Jensen, F. (i958). Rare phenotypes in the Hp system. Acta Genet. (Basel), x6, 327. Acta genet. (Basel), 8, 248. Edington, G. M., and Laing, W. N. (1957). Relationship between Gartler, S. M. (i96i). Comments on the population dynamics of haemoglobins C and S and malaria in Ghana. Brit. med. J., 2, 53-aminoisobutyric aciduria. In Proceedings of the Conference on 143. Genetic Polymorphisms and Geographic Variations in Disease, ed. , and Lehmann, H. (I956). The distribution of Haemoglcbin C B. S. Blumberg, p. 192. Grune and Stratton, New York. in West Africa. Man, 56, 34. Gedde-Dahl, T., Jr., and Berg, K. (r965). Linkage in man: the Edwards, J. H. (1965). The meaning of the associations between Inv and the Lp serum type systems. Nature (Lond.), 208, I I26. blood groups and disease. Ann. hum. Genet., 29, 77. Giblett, E. R. (i962). The plasma transferrins. Progr. med. Eichner, E. R., Finn, R., and Krevans, J. R. (i963). Relationship Genet. (N.Y.), 2, 34. between serum antibody-levels and the ABO blood group poly- -, and Brooks, L. E. (i963). Haptoglobin types. Haptoglobin morphism. Nature (Lond.), I98, i64. sub-types in three racial groups. Nature (Lond.), 197, 576. Evans, D. A. P. (i962). Pharmacogenetique. Mid. et Hyg. (Geneve), , and Scott, N. M. (i965). Red cell acid phosphatase: racial 20, 905. distribution and report of a new phenotype. Amer.J_. hum. Genet., (i963). Pharmacogenetics. Amer. J. Med., 34, 639. 17, 425. - (I965a). Conjoint clinic on drugs and genes. J. chron. Dis., ,,and Steinberg, A. G. (ig60). The inheritance of serum hapto- s8, 59. globin types in American Negroes: evidence for a third allele (I965b). Confirmation of association between ABO blood Hp2m. ibid., 12, i60. groups and salivary ABH secretor phenotypes and electrophoretic Goedde, H. W., Schoepf, E., and Fleischmann, D. (i964). Studies patterns of serum alkaline phosphatase. J. med. Genet., 2, 126. on pharmacogenetics. I. The enzymic acetylation of isonicotic , Davison, K., and Pratt, R. T. C. (i965). The influence of acid hydrazide (INH). Biochem. Pharmacol., I3, i671. acetylator phenotype on the effects of treating depression with Gow, J. G., and Evans, D. A. P. (1964). A study of the influence phenelzine. Clin. Pharmacol. Ther., 6, 430. of the isoniazid inactivator phenotype on reversion in genito- , Manley, K. A., and McKusick, V. A. (ig60). Genetic control urinary tuberculosis. Tubercle (Edinb.), 45, I36. of isoniazid metabolism in man. Brit. med. 7., 2, 485. Gramberg-Danielson, B. (i961). Untersuchungen uber die Unfall- Fagerhol, M. K., and Braend, M. (i965). Serum prealbumin: haufigheit von Farbenuntuchtigen im Strassenverkehr. Klin. polymorphism in man. Science, 149, 986. Mbl. Augenheilk., I39, 677. Fildes, R. A., and Harris, H. (i966). Genetically determined varia- Griffin, F., and Fischer, R. (1960). Differential reactivity of saliva tion of adenylate kinase in man. Nature (Lond.), 209, 26i. from 'tasters' and 'non-tasters' of 6-N-propylthiouracil. Nature , and Parr, C. W. (i963). Human red cell phosphogluconate (Lond.), 187,417. dehydrogenases. ibid., 200, 890. Grubb, R., and Laurell, A. B. (1956). Hereditary serological human , and (i964). Various forms of human erythrocyte phos- serum groups. Acta path. microbiol. scand., 39, 390.

phogluconate dehydrogenase. Proc. 6th int. Congr. Biochem., Haldane, J. B. S. (1949). Disease and evolution. Ricerca sci., I9, copyright. New York, p. 229. (Suppl.), 68. Firschein, I. L. (I96I). Population dynamics of the sickle-cell Harboe, M. (x959). A new hemagglutinating substance in the Gm trait in the Black Caribs of British Honduras, Central America. system, anti-Gmb. Acta path. microbiol. scand., 47, 19I. Amer. J. hum. Genet., 13, 233. -, and Lundevall, J. (1959). A new type in the Gm system. Fischer, R., and Griffin, F. (ig60). Factors involved in the mechan- ibid., 45, 357. ism of 'taste-blindness'. J. Hered., 51, i82. Harris, H., Hopkinson, D. A., and Robson, E. B. (i962). Two- , -, England, S., and Gain, S. M. (i96i). Taste threshold dimensional electrophoresis of pseudocholinesterase components and food dislikes. Nature (Lond.), 191, 1328. in normal human serum. Nature (Lond.), I96, 1296.

, -, and Pasmanick, B. (i963). The perception of taste: -,1 ,.1 ,.1 and Whittaker, M. (i963). Genetical studies on http://jmg.bmj.com/ some psychophysiological, pathophysiological, pharmacological a new variant of serum cholinesterase detected by electrophoresis. and clinical aspects. Proc. Amer. psychopath. Ass., 53, 129. Ann. hum. Genet., 26, 359. Fisher, R. A., Ford, E. B., and Huxley, J. S. (1939). Taste-testing , Kalmus, H., and Trotter, W. R. (1949). Taste sensitivity to the anthropoid apes. Nature (Lond.), I44, 750. phenylthiourea in goitre and diabetes. Lancet, 2, 1038. Flatz, G., Pik, C., and Sringam, S. (i965a). Haemoglobin E and 3- and Whittaker, M. (i962). The serum cholineterase variants. thalassaemia: their distribution in Thailand. Ann. hum. Genet., A study of 22 families selected via the 'intermediate' phenotype. 29, 151. Ann. hum. Genet., 26, 59. ,-, and (I965b). Haemoglobinopathies in Thailand. Harris, H. W. (196i). High-dose isoniazid compared with standard- i. A study of patients with the thalassaemia syndrome and their dose isoniazid with PAS, in the treatment of previously un- families. Brit.J. Haemat., II, 2i6. treated cavitary pulmonary tuberculosis. Trans. Res. Conf. Pulm. Ford, E. B. (1940). The New Systematics, ed. Julian Huxley, p. 493. Dis., 39. on September 27, 2021 by guest. Protected Clarendon Press, Oxford. Harris, R. (i963). Infectious diseases, the ABO blood groups and (i965). Genetic Polymorphism. Faber and Faber, London. human evolution. Eugen. Rev., 54, 201. Fraser, G. R. (x96x). Cretinism and taste sensitivity to phenyl- Hiraizumi, Y. (5964). Are the MN blood groups maintained by thiocarbamide. Lancet, I, 964. heterosis? Amerl. hum. Genet., I6, 375. (i963). A genetical study of goitre. Ann. hum. Genet., 26, Hirschfeld, J. (I959). A simple method of determining haptoglobin 335. groups in human sera by means of agar-gel electrophoresis. , Stamatoyannopoulos, G., Kattamis, C., Loukopoulos, D., Acta path. microbiol. scand., 47, I69. Defaranas, B., Kitsos, C., Zannos-Mariolea, L., Choremis, C., (ig60). Immuno-electrophoresis procedure and application to Fessas, P., and Motulsky, A. G. (i964). Thalassaemias, abnormal the study of group specific variations. Sci. Tools, 7, I8. hemoglobins and Glucose-6-phosphate dehydrogenase deficiency (1962). The Gc-system. Progr. Allergy, 6, i55. in the Arta area of Greece: diagnostic and genetic aspects of (I964). Human lipoprotein polymorphism. (Proc. ioth complete village studies. Ann. N.Y. Acad. Sci., 119, 415. Congr. int. Soc. Blood Transf.). Bibl. haemat. (Basel), 23, 365. Frymoyer, J. W., and Jacox, R. F. (I963a). Investigation of the , and Heiken, A. (I963). Application of the Gc system in pater- genetic control of sulphadiazine and isoniazid metabolism in the nity cases. Amer. J. hum. Genet., x5, I9. rabbit. J. Lab. clin. Med., 62, 89I. Hodson, A. W., Latner, A. L., and Raine, L. (i962). Iso-enzymes , and- (i963b). Studies of genetically controlled sulpha- of alkaline phosphatase. Clin. chim. Acta, 7,255. diazine acetylation in rabbit livers: possible identification of the Hopkinson, D. A., and Harris, H. (I965). Evidence for a second heterozygous trait. ibid., 62, 905. 'structural' locus determining human phosphoglucomutase. Fudenberg, H. (I96I). Quoted in Brandtzaeg, B., Fudenberg, H., Nature (Lond.), 208, 410. and Mohr, J. The Gm(r) serum group. Acta genet. (Basel), is, , Spencer, N., and Harris, H. (I963). Red cell acid phosphatase I70. variants: a new human polymorphism. ibid., I99, 969. J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from 66 J. Price

, , and (i964). Genetical studies on human red cell Li, C. C. (1953). Is Rh facing a crossroad; a critique of the com- acid phosphatase. Amer. J. hum. Genet., I6, 14I. pensation effect. Amer. Nat., 87, 257. Huehns, E. R., and Shooter, E. M. (i965). Human ha2moglobins. (I955). Population Genetics. University of Chicago Press, J. med. Genet.,2,48. Chicago. Hughes, H. B. (i953). On the metabolic fate of isoniazid. J. Phar- Lie-Injo Luan Eng (I964). Haemoglobinopathies in East Asia. macol. exp. Ther., 109,444. Ann. hum. Genet., 28, IOI. , Biehl, J. P., Jones, A. P., and Schmidt, L. H. (1954). Meta- Livingstone, F. B. (ig60). Natural selection, disease, and ongoing bolism of isoniazid in man as related to the occurrence of peri- human evolution, as illustrated by the ABO blood groups. pheral neuritis. Amer. Rev. Tuberc., 70, 266. Hum. Biol., 32, 17. , Schmidt, L. H., and Biehl, J. P. (I955). The metabolism of -(i964). Aspects of the population dynamics of the abnormal isoniazid: its implications in therepeutic use. Trans. Conf. Chemo- hMemoglobin and glucose-6-phosphate dehydrogenase deficiency ther. Tuberc., 14, 217. Washington, D.C., 1955, Veterans Admini- genes. Amer. J7. hum. Genet., I6, 435. stration. Luca, F. de, and Cramarossa, L. (I965). Phenylthiourea and endemic Jackson, C. E., Symon, W. E., and Mann, J. D. (i964). X chromo- goitre. Lancet, I, I399. some mapping of genes for red-green color blindness and Xg. Lugg, J. W. H. (I966). Taste thresholds for phenylthiocarbamide Amer. 7. hum. Genet., I6, 403. of some population groups. III. The thresholds of some groups Javid, J. (i964). The nature of the difference between haptoglobin living in Japan. Ann. hum. Genet., 29, 2I 7. polymers in the phenotypes Hp 2-I and Hp 2-2. Proc. nat. Acad. Marks, P. A., and Banks, J. (I965). Drug-induced hemolytic anemias Sci. (Wash.), 52, 663. associated with glucose-6-phosphate dehydrogenase deficiency: a Jenne, J. W., MacDonald, F. M., and Mendoza, E. (i96i). A study genetically heterogeneous trait. Ann. N.Y. Acad. Sci., 123, I98. of the renal clearances, metabolic inactivation rates, and serum Matsunaga, E. ( I962a). The dimorphism in human normal cerumen. fall-off interaction of isoniazid and paramino-salicylic acid in Ann. hum. Genet., 25, 273. man. Amer. Rev. resp. Dis., 84, 37I. (i962b). Selective mechanisms operating on ABO and MN Jeppsson, J. O., and Sjoquist, J. (i963). Separation of normal blood groups with special reference to prezygotic selection. human transferrin into two fractions. Biochim. biophys. Acta Eugen. Quart., 9,36. (Amst.), 78, 658. Mayr, E. (i963). Animal Species and Evolution. Harvard University Jonxis, J. H. P. (X965). Haemoglobinopathies in West Indian Press, Cambridge, Massachusetts. groups of African origin. In Abnormal Haemoglobins in Africa, Memoria, J. M. P. (X959). Endemic goitre and sensitivity to P.T.C. ed. J. H. P. Jonxis, p. 329. Blackwell Scientific Publications, Rev. bras. Malar., xi, 5. Oxford. Menn, J. J., McBain, J. B., and Dennis, M. J. (i964). Detection of Junqueira, P. C., Kalmus, H., and Wishart, P. (1957). P.T.C. naturally occurring cholinesterase inhibitors in several crops by thresholds, colour vision and blood factors of Brazilian Indians. II paper chromatography. Nature (Lond.), 202, 697. Carajas. Ann. hum. Genet., 22, 22. Mitchell, R. S., Riemensnider, D. K., Harsch, J. R., and Bell, J. C. Kalmus, H. (I965). Diagnosis and Gsneticsof DefectiveColour Vision. (i958). New information on the clinical implications of individual Pergamon Press, Oxford. variations in the metabolic handling of antituberculous drugs, , De Garay, A., Rodarte, U., and Cobo, L. (i964). The frequency particularly isoniazid. Trans. Conf. Chemother. Tuberc., 17, 77. of PTC tasting, hard ear wax, colour blindness and other genetical Montenegro, L. (i964). P.T.C. tasting among Tucano Indians.copyright. characters in urban and rural Mexican populations. Hum. Biol., Ann. hum. Genet., 28, I 85. 36, I34. Morton, N. E. (i965). Models and evidence in human population Kaplan, A. R., Fischer, R., Glanville, E., Powell, W., Kamion- genetics. In Genetics Today (Proceedings of the XIth International kowski, M., and Fleshler, B. (i964). Differential taste sensitivi- Congress of Genetics, The Hague, 1963), ed. S. J. Geerts, Vol. 3, ties in duodenal and gastric ulcer patients. Gastroenterology, 47, p. 935. Pergamon Press, Oxford. 604. , and Chung, C. S. ( I 959). Are the MN blood groups maintained Kidson, C., and Gorman, J. G. (i962). A challenge to the concept by selection? Amer. J. hum. Genet., II, 237. of selection by malaria in glucose-6-phosphate dehydrogenase , Krieger, H., and Mi, M. P. (i966). Natural selection on poly- deficiency. Nature (Lond.), I96, 49. morphisms in North Eastern Brazil. ibid., I8, 153. http://jmg.bmj.com/ Kirk, R. L. (i965). Population genetic studies of the indigenous Motulsky, A. G. (i964). Hereditary red cell traits and malaria. peoples of Australia and New Guinea. Progr. med. Genet., 4, Amer. J. trop. Med. Hyg., 13, 147. 202. (i965). Theoretical and clinical problems of G 6 PD deficiency. , Cleve, H., and Bearn, A. G. (i963). The distribution of the In Abnormal Haemoglobins in Africa, ed. J. H. P. Jonxis, p. I43. group specific component (Gc) in selected populations in South Blackwell, Oxford. and South East Asia and Oceania. Acta genet. (Basel), 13, 140. , and Campbell-Kraut, J. M. (i962). Population genetics of , Parker, W. C., and Bearn, A. G. (1964). The distribution of glucose-6-phosphate dehydrogenase deficiency of the red cell. the transferrin variants Di and Dehi in various populations. ibid., In Proceedings of the Conference on Genetic Polymorphisms and Geographic Variation in Disease. Grune and Stratton, New 14,41. York. Kirkman, H. N., and Hendrickson, E. M. (i963). Sex-linked on September 27, 2021 by guest. Protected electrophoretic difference in glucose-6-phosphate dehydrogenase , Kraut, J. M., Thieme, W. T., and Musto, D. F. (1959). deficiency. Amer. J'. hum. Genet., 15, 241. Biochemical genetics of glucose-6-phosphate dehydrogenase Kitchin, F. D., deficiency. Clin. Res., 7, 89. and Beam, A. G. (1966). The electrophoretic patterns , and Steinmann, L. (i962). Aryl amine in of normal and variant phenotypes of the group specific (Gc) acetylation human components in human red cells. J. clin. Invest., 4x, 1387. serum. ibid., I8. 201. Nagel, R., and Soto, 0. (I964). Haptoglobin types in Native I Howel-Evans, W., Clarke, C. A., McConnell, R. B., and Chileans: a hybrid population. Amer. J. phys. Anthrop., 22, 335. Sheppard, P. M. (I959). P.T.C. taste response and thyroid disease. Neel, J. V., and Post, R. H. (1963). Transitory 'positive' selection Brit. med.J7., I, Io6g. for colorblindness?. Eugen.Quart., 10, 33. Knight, R. A., Selin, M. J., and Harris, H. W. (1959). Genetic , Salzano, F. M., Junqueira, P. C., Keiter, F., and Maybury- factors influencing isoniazid blood levels in humans. Trans. Lewis, D. (I964). Studies on the Xavante Indians of the Brazilian Conf. Chemother. Tuberc., 18, 52. Mato Grosso. Amer.J_. hum. Genet., I6, 52. Kruatrachue, M., Charoenlarp, P., Chongsuphajaisiddhi, T., and Nerstr0m, B. (i963). On the inheritance of the Gc-system. A Harinasuta, C. (1962). Erythrocyte glucose-6-phosphate dehydro- Danish family study. Acta genet. (Basel), 13, 30. genase and malaria in Thailand. Lancet, 2, II83. Newcombe, H. B. (i963). Risk of fetal death to mothers of dif- Lai, L., Nevo, S., and Steinberg, A. G. (i964). Acid phosphatases ferent ABO and Rh blood types. Amer. J. hum. Genet., I5, 449. of human red cells: predicted phenotype conforms to a genetic (1965). Risks in ABO and Rh incompatibility. ibid., 17, 97. hypothesis. Science, 145, II87. Orgell, W. H. (I963). Inhibition of human plasma cholinesterase in Lehmann, H., and Huntsman, R. G. (i966). Man's Haemoglobins. vitro by plant extracts. Lloydia, 26, 59. North-Holland, Amsterdam. Parker, W. C., and Beam, A. G. (5963). Control gene mutation as Levene, C., Hermoni, D., and Cividalli, L. (i965). MN heterozy- a possible explanation of certain haptoglobin phenotypes. Amer. gotes among newborn in Israel. IsraelJ. med. Sci., I, 448. J. hum. Genet., 15, 159. J Med Genet: first published as 10.1136/jmg.4.1.44 on 1 March 1967. Downloaded from Human Polymorphism 67

, and (i965). Variations in heterozygous expression at in respect of the G6PD locus. Nature (Lond.), 204, io62. the haptoglobin locus. ibid., 17, 370. Smithies, 0. (1955). Zone electrophoresis in starch gels: group Parr, C. W., and Fitch, Linda, I. (I954). Hereditary partial defi- variations in the serum proteins of normal human adults. Biochem. ciency of human-erythrocyte phosphogluconate dehydrogenase. 1, 6I, 629. Biochem.J_., 93, 28c. (i964). Starch-gel electrophoresis. Metabolism, I3, 974. Penrose, L. S., Smith, S. M., and Sprott, D. A. (I956). On the , Connell, G. E., and Dixon, G. H. (I962a). Inheritance of stibility of allelic systems, with special reference to haemoglobins haptoglobin subtypes. Amer. J. hum. Genet., 14, 14. A, S and C. Ann. hum. Genet., 21, 90. ,-,and (I962b). Chromosomal rearrangements and Pickford, R. W. (i962). Compound hemizygotes for red-green colour the evolution of haptoglobin genes. Nature (Lond.), I96, 232. vision defects. Vision Res., 2, 245. Spencer, N., Hopkinson, D. A., and Harris, H. (i964). Phospho- Plato, C. C., Cruz, M. T., and Kurland, L. T. (i964). Frequency of glucomutase polymorphssm in man. ibid., 204, 742. glucose-6-phosphate dehydrogenase deficiency, red-green colour Steinberg, A. G. (i962). Progress in the study of genetically deter- blindness and Xga blood-group among Chamorros. Nature mined human gamma globulin types (The Gm and Inv groups). (Lond.), 202, 728. Progr. med. Genet., 2, I. Polonovski, M., and Jayle, M. F. (I940). Sur la preparation d'une , Giles, B. D., and Stauffer, R. (i960). A Gm-like factor present nouvelle fraction des proteines plasmatiques, l'haptoglobine. in Negroes and rare or absent in Whites: its relation to Gma and C. R. Acad. Sci. (Paris), 211, 517. Gmx. Amer. 7. hum. Genet., 12, 44. Porter, I. H., Boyer, S. H., Schulze, J., and McKusick, V. A. (i96i). Genetic , and Goldblum, R. (1965). A genetic study of the antigens control of glucose-6-phosphate dehydrogenase produc- associated with the Gm(b) factor of human gamma globulin. tion. Proc. 2nd int. Conf. hum Genet., Rome 1961, I, 6i8. ibid., 17,133- Post, R. H. (I962). Population differences in red and green color -, Wilson, J. A., and Lanset, S. (i962). A new human gamma vision deficiency: A review and a query on selection relaxation. globulin factor determined by an allele at the Inv locus. Vox Eugen.Quart., 9, 131. Sang. (Basel), 7, I 5 I. -- (i965). Selection against 'colorblindness' among primitive Sunahara, S. (i96i). Genetical, geographical and clinical studies populations. ibid., 12, 28. on isoniazid. (Proc. i6th int. Tuberc. Conf., Toronto, Canada.) Powell, R. D., and Brewer, A. G. (i965). Glucose-6-phosphate Excerpta med. (Amst.) int. Congr. ser., no. 44, 2, 5I3- dehydrogenase deficiency and falciparum malaria. Amer. J. , Urano, M., and Ogawa, M. (i96i). Genetical and geographic trop. Med. Hyg., 14, 358. studies on isoniazid inactivation. Reed, T. E. (i96i). Polymorphism and natural selection in blood Science, 134, 1530. groups. In Proceedings of the Conference on Genetic Polymorphisms, Sutton, H. E., and Karp., G. W. (i964). Variations in the heterozy- ed. B. S. Blumberg, p. 80. Grune and Stratton, New York. gous expression at the haptoglobin locus. Amer. 7. hum. Genet., i6, 419. (I965). Magnitude of binomial variation in family data. Tarlov, A. R., Brewer, G. J., Carson, P. E., and Alving, A. S. (i962). Amer. 7. hum. Gnmet., 17, 95. Primaquine sensitivity: Glucose-6-phosphate dehydrogenase --, Gershowitz, H., Soni, H., and Napier, J. (i964). A search deficiency: an inborn error of metabolism of medical and bio- for natural selection in six blood group systems and ABH secre- logical significance. Arch. intern. Med., I09, 209. tion. ibid., I6, I6I. M. and copyright. Reinskou, T. (i966). Application of the Gc system in 1338 paternity Terry, C., Segall, G. (i947). The association of diabetes cases. Vox and taste-blindness. J. Hered., 38, 135; Postscript (by Terry) Sang. (Basel), II, 59. ibid., 38, 378. Roberts, D. F., and Boyo, A. E. (i96o). On the stability of haemo- globin gene Thomas, K., and Kampf, D. (i96i). Die verebbaren Gm-Serumei- frequencies in West Africa. Ann. hum. Genet., 24, weissgruppen: ein neuer Faktor, Gm Dresden, in diesem System. 375. Dtsch. Robson, E. B., and Harris, H. (i966). Further data on the incidence Gesundh.-Wes., I6, II85. and genetics of the serum cholinesterase Co +. ibid., 29, 403. Thompson, G. R. (i962). Significance of haemoglobins S and C in Rooartz, C. (i963). Ghana. Brit. med. J., I, 682. Donnees actuelles sur les groups hereditaires Thuline, H. (i964). Color-vision defects in American school

do gamma-globulines. http://jmg.bmj.com/ Transfusion (Paris), 6, 5. children. J. Amer. med. Ass., I88, 514. Lenoir, J., and Rivat, L. (I96I). A new inheritable property of human sera: the InV Trujillo, J., Fairbanks, V., Ohno, S., and Beutler, E. (I96i). Chro- factor. Nature (Lond.), 189, 586. mosomal constitituion in glucose-6-phosphate dehydrogenase -, Rivat, L., and Rousseau P.-Y. (I962). Deux nouveaux facteurs deficiency. Lancet, 2, 1454. dans les systemes hereditaires de gamma-globuline: le Gm(e) et VanderDonck, R., and Verriest, G. (ig60). Femme protanomale et l'Inv. (Proc. 9th Congr. int. Soc. Blood Transf., Mexico.) heterozygote mixte (Genes de la protanomalie et de la deutera- Bib!. haemat. (Basel), 19, 455. nopie en position de repulsion), ayant deux fils deuteranopes, un Baitsch, H., and Van Loghem, J. (I963). Les fils protanomal et deux fils normaux. Biotypologie, 21, I 0. systemes Gm et Inv en Europe. Acta genet. (Basel), 13, 109. Vogel, F., Pettenkofer, H. J., and Helmbold, W. (ig60). Uber die Saldanha, P. H. (I956). Apparent pleiotrophic effect of genes Populationsgenetik der ABO Blutgruppen. 2. Mitteilung. Gen- determining taste thresholds for phenylthiourea. Lancet, 2, 74. haufigkeit und epidemische Erkrankungen. Acta genet. (Basel), on September 27, 2021 by guest. Protected Salzano, F. M., and Shreffler, D. C. (I966). The Gc polymorphism 10, 267. in the Caingang Indians of Brazil. Acta genet. (Basel), I6, 242. Wang, A.-C., and Sutton, H. E. (1965). Human transferrins C and , and Sutton, H. E. (I965). Haptoglobin and transferrin types Di: chemical differences in a peptide. Science, 149, 435. of Indians from Santa Catarina, Brazil. Amer. J. hum. Genet., 17, Weatherall, D. J. (i965). The Thalassaemia Syndromes. Blackwell 280. Oxford. Schelling, J.-L., Tetreault, L., Lasagna, L., and Davis, M. (I965). Weiser, M. M., Bolt, R. J., and Pollard, H. M. (i964). Isozyme Abnormal taste threshold in diabetes. Lancet, I, 508. patterns of the human gastrointestinal tract in the normal state Sheppard, P. M. (I959). Blood groups and natural selection. Brit. and in nontropical sprue. J. Lab. clin. Med., 63, 656. med. Bull., I5, 134. Went, L. N. (i96i). Segregation-ratio advantage of abnormal Shim, B. S., and Beam, A. G. (I964). The distribution of hapto- haemoglobins. Nature (Lond.), 192, 382. globin subtypes in various populations including subtype patterns Wiener, A. S. (1965). Blood groups and personality traits. Amer. in some nonhuman primates. Amer.J. hum. Genet., i6, 477. J. hum. Genet., 17, 369. N. Shinton, K., Richardson, R. W., and Williams, J. D. F. (I965). , Gordon, E. B., and Wexler, J. P. (i963). The M-N types, Diagnostic value of serum haptoglobin. J. clin. Path., I8, I I4. with special reference to the mating MN x MN. Exp. Med. Shreffler, D. C. (I965). Genetic studies of blood group-associated Surg., 21, 89. variations in a human serum alkaline phosphatase. Amer. J. Witzleben, C. L., and Wyatt, J. P. (196i). The effect of long hum. Genet., 17, 71. survival on the pathology of thalassaemia major. J. Path. Bact., Simpson, N. E. (I966). Factors influencing cholinesterase activity 82, I. in a Brazilian population. ibid., I8, 243. Zinkham, W. H., and Lenhard, R. E., Jr. (I959). Metabolic ab- Siniscalco, M., Filippi, G., and Latte, B. (I964). Recombination normalities of erythrocytes from patients with congenital non- between protan and deutan genes; data on their relative positions spherocytic haemolytic anemia. J. Pediat., 55, 319.