Differential fertility as a mechanism maintaining balance polymorphisms in Sardinia
Haldane’s (1949) theory of a selective advantage of heterozygotes in the presence of malaria has been largely tested and confirmed, principally for the sickle cell gene and glucose-6-phosphate dehydrogenase (G6PD) deficiency, in areas where malaria is still endemic (Africa, tropical zones). A wide range of studies have pointed out the concomitant presence of high frequencies of these polymorphisms and Plasmodium falciparum malaria and have researched the physiological, biochemical, and, recently, molecular bases of the infection-resistant mechanisms (Allison 1954a, b, 1955, 1961, 1964; Firschein 1961; Livingstone 1957, 1964, 1971; Luzzatto et al. 1969, 1970; Luzzatto and Bienzle 1979; Luzzatto 1979; Motulsky 1964, 1975; Pasvol et al. 1978, 1982; Stamatoyannopoulos et al. 1966; Usanga and Luzzatto 1985; Vulliamy et al. 1992).
A similarity between the geographic distributions of thalassemia and past incidence of malaria in the Mediterranean basin has also been shown in many studies [see Weatherall and Clegg (1981)], but the protection mechanisms of thalassemic individuals are not yet completely understood.
Mechanisms maintaining these balanced polymorphisms have been studied in endemic areas by directly analyzing the fertility and survival of the different genotypes (Allison 1956, 1964; Custodio and Huntsman 1984; Delbrouk 1958; Edington 1955; Firschein 1961; Fleming et al. 1979; Garlick 1960; Hamilton et al. 1972; Madrigal 1989; Motulsky 1964; Molineaux and Gramiccia 1980; Rucknagel and Neel 1961). However, the first demonstration of the higher fitness (in terms of number of children) of heterozygotes comes from a study on thalassemia in Italy, when Haldane’s hypothesis had only just been formulated. Silvestroni et al. (1950) posed the problem of how the high genetic frequencies for thalassemia in a specific area (the district of Ferrara) could be “preserved through many generations, although the gene is being constantly eliminated” (p. 682). The hypothesis of a “higher fertility of the families in which both parents are heterozygous” (p. 682) was supported by experimental results. Subsequently, a study in the same area confirmed this tendency, although not in a statistically significant way (Aguzzi et al. 1978).
No direct studies of the fertility of different genotypes were made on the island of Sardinia (the region studied here) when malaria was still endemic. The analyses of Siniscalco et al. (1961, 1966), followed by many other studies (Cao et al. 1978; Carcassi et al. 1957; Modiano et al. 1986; Piazza et al. 1972, 1985; Terrenato et al. 1971; Terrenato 1973; Workman et al. 1975), have concentrated on revealing a high frequency of heterozygotes for thalassemia and G6PD deficiency in areas of past malaria morbidity.
Nevertheless, an indirect study at the population level to test fertility as a mechanism maintaining balanced polymorphisms in an unfavorable environment has been performed by Zei et al. (1990); women’s fertility and malarial morbidity rates, both concerning all villages on Sardinia during a period in which malaria was endemic, were compared. The results demonstrated that women living in hyperendemic malarial areas had more children than women living in mesoendemic malarial areas. However, Zei’s analysis did not comprehensively confirm Haldane’s theory, because in each homogeneous area of malaria incidence the villages for which genetic data were also available showed high heterogeneity with respect to heterozygote frequencies for thalassemia and G6PD deficiency.
Therefore we studied the relation between heterozygote frequency and fertility rates, disregarding the characterization for incidence of malaria. A sample of Sardinian villages is classified according to the estimates of heterozygote frequency obtained by Siniscalco et al. (1966), and the reproductive behavior of the women who lived in these villages is reexamined.
The aims of this study are to test the hypothesis that differential fertility is a selection-induced mechanism that maintains balanced polymorphisms and to investigate the reasons for and the effects on reproduction of the incomplete correspondence of heterozygote frequency and malarial morbidity levels on Sardinia.
MATERIALS AND METHODS
Demographic Data. Demographic data come from the special investigation of female fertility carried out as part of the 1961 Italian population census. Information regarding residence, age, and education of all married or widowed Sardinian women and number, sex, viability, and birth dates of their children is available. Married women of postreproductive age (over 45 years) were selected. These women had been born between 1861 and 1915 and had had children from 1876 onward during a period when malaria was endemic on Sardinia. Widowed women were excluded from the analysis because of the lack of information regarding the end of their reproductive period.
Two different aspects of women’s fecundity were considered: fertility rate, measured as the mean number of live-born children per woman who had given birth to at least one child, and sterility rate, measured as the percentage of women without children.
Many studies of the Italian population have pointed out that reproductive behavior is influenced by socioeconomic and cultural factors (Livi Bacci 1977; Golini 1967; Zei and Cavalli-Sforza 1977). On Sardinia the variation in fertility was strongly associated with the educational level of the women (the less education, the more children) and to the degree of urbanization of the villages (fewer children in urban areas) (Zei et al. 1990). To control these factors and to clarify the relationship under study, we selected illiterate women (those with no education or with less than three years of schooling). During the period considered, illiterate Sardinian women represented the majority of the female population, ranging from 81.5% for the women born before 1880 to 44.6% for women in the 1911-1915 birth cohort [see Zei et al. (1990)]. Moreover, in this education group the decrease of birth rate characteristic of the demographic transition process had not yet started in the period considered (Zei and Cavalli-Sforza 1977; Zei et al. 1990).
It was not necessary to check the degree of urbanization because the villages considered were all rural.
Genetic Data. The genetic data are derived from the study of Siniscalco et al. (1966) on 52 villages distributed throughout Sardinia (Figure 1). Siniscalco and colleagues collected the frequencies of heterozygotes for thalassemia (Th+) and G6PD deficiency (GD-) in random samples of schoolboys from each village in the 1950s, a few years after the eradication of malaria. The average percentage of sampling on the 1951 total population for each village was 5.96% 0.46. We assumed that these observed frequencies are a good estimate of the number of adult heterozygous individuals.
Two groups of villages were selected. They belong to the first and the last quartiles of the standardized frequency distributions for all the villages and had been separately classified according to their frequency of heterozygotes for thalassemia and G6PD deficiency.
The frequencies of the two polymorphisms in these villages were correlated (Siniscalco et al. 1966), as expected, because they had been subjected to the same selective pressure. In fact, only minor differences, resulting from sampling error, were observed in the location of the villages in the two distributions. However, because the data on G6PD deficiency were complete and the data on thalassemia were missing for seven villages, the G6PD distribution was considered initially, and two areas of fifteen villages each were compared. Moreover, two restricted samples of villages with high or low heterozygote frequencies for both systems were selected.
Fertility. The average number of live-born children was significantly higher for the women living in the villages with a higher frequency of heterozygotes (Table 1, column 1). (Table 1 omitted)
The hypothesis underlying this study–that differential fertility was a mechanism that maintained the high frequency of heterozygotes–would seem to be confirmed. However, besides cultural and socioeconomic factors (kept constant by selecting illiterate women living in rural villages), a demographic factor (women’s age at marriage) seemed to influence fertility rates.
The difference between the average age at marriage of the women living in the two areas (Table 1, column 2) can be explained. In Figure 1 the concentration of the villages in two geographically distinct zones is clear. (Figure 1 omitted) The villages with low heterozygote frequency, save Carloforte, are located in the mountains of the Nuoro and Sassari provinces; villages with a high frequency lie in the plain of Cagliari province. Different geographic environments are often linked to different jobs or customs. The isolation of the small mountain villages and the consequent limited opportunity for social interchange may have been additional causes of later age at marriage (Golini 1967). The difference in fertility rates could be merely the result of this difference in mean age at marriage. But when this demographic factor is considered as a covariate, a significant difference in the average number of children between the women of the two zones is still observed (r = 6.60, p
Malaria was always present, even if at differing levels of morbidity, in the two areas under study. Thus the observed differential fertility could be interpreted as the result of different numeric ratios within each area between heterozygous and normal homozygous women. These women, according to Haldane’s theory, should be more and less resistant, respectively, to malarial infection. Hence, where the proportion of heterozygous women was lower, the complementary proportion of less resistant homozygous women was higher; the risk of fever, anemia, and placental parasitization and hence fetal loss resulted in a lowering of the fitness (fertility) in that area.
The other demographic parameters of Table 1 seem to support this hypothesized effect of the different proportion of heterozygotes on fertility rates. The average interval elapsing between age at marriage and first childbirth (be it live-born or stillborn) was significantly greater where the heterozygote frequency was low (Table 1, column 4) because of the high risk of early fetal loss in the first pregnancy (McGregor 1984; McFalls and McFalls 1984) for the numerous less resistant homozygous women. Moreover, for the same women the consequences of malarial infection seem to be evident also in subsequent pregnancies, as shown by the high average birth interval (see Table 1, column 5).
Similar results were obtained with a more rigorous selection of villages in which heterozygotes for thalassemia and G6PD deficiency were both present with a low or high frequency. Despite the marked reduction of the samples, the different heterozygote frequency seems to have an even more highly significant influence on the number of children, on the marriage–first child interval, and on the average birth interval (Table 2). (Table 2 omitted)
Sterility. Women with no children have been considered separately in this study. In the two areas corresponding to the lower and upper tails of the heterozygote frequency distribution, a significantly different proportion of women without children was found, namely, a higher sterility rate where the heterozygote frequency was low (Table 3). (Table 3 omitted)
The causes of sterility may be different from those of subfertility, and the possible effects of the different frequency of heterozygotes need to be compared with the numerous biological and physiological age-dependent factors that can produce the incapacity to conceive. The data of this study did not allow us to distinguish complete infecundity from early fetal loss. However, because no significant difference between age at marriage in the two areas was demonstrated (Table 3), the differential sterility rates can be ascribed to the different number of heterozygote malaria-resistant women.
When the reduced samples were analyzed, the percentage of sterile women was still higher in the low heterozygote frequency area, even if significance was not reached (G = 3.08, 0.05
Joint Effect of Genetic and Epidemiological Factors. The results of this study show a higher fertility rate in areas of higher heterozygote frequency for thalassemia and G6PD deficiency. The results of a previous study in the same area (Zei et al. 1990) demonstrated a higher fertility rate in areas of higher malarial morbidity rates. However, heterozygote frequency and morbidity levels did not correspond completely. Thus the joint effect of genetic (GD- and Th+ heterozygote frequencies) and epidemiological (malarial morbidity) factors on reproduction were analyzed to clarify their respective and combined roles in maintaining balanced polymorphisms.
All the villages were distributed in 3 x 3 tables according to malarial morbidity degree and heterozygote frequencies for both GD- and Th+. The medium genetic level was obtained by grouping the two central quartiles of the heterozygote frequency distributions. Malarial morbidity levels were based on a survey made by Fermi (1934, 1938, 1940), and they refer to the period for which data on female reproduction are available [see Zei et al. (1990)]. The fertility and sterility of the women belonging to these nine areas were analyzed.
The mean number of live births per woman in areas where both heterozygote frequency and malaria levels were low was much lower than that in areas where heterozygotes and malaria were both at high levels: 4.70 versus 6.05 (t = 9.09, p
Moreover, a decreasing percentage of sterile women was observed (Table 5), as expected, when progressing from low to high levels of joint epidemiological and genetic effects (G = 17.82, p
The variation in the genetic frequencies caused by natural selection is based on the assumption that selection operates through a differential fitness of the various genotypes. Resistance to malarial infection favors the heterozygotes for some genetic systems, increasing both their viability between birth and reproductive age and their relative fertility.
In Sardinia malarial infection was eradicated soon after the end of World War II, but no direct studies of the survival and fertility of heterozygous people for thalassemia and G6PD deficiency are available. However, fertility rates from census data and estimates of heterozygote frequencies, both concerning times and places in which malaria was endemic, have allowed us to study at the population level one of the mechanisms that may have maintained the stability of polymorphisms at that time.
The results obtained from this analysis support the hypothesis of a higher fertility rate where Th+ and GD- allele frequencies are higher. In areas with high heterozygote frequencies the mean number of children per woman was 10-20% higher than in areas with low heterozygote frequencies.
Haldane’s theory would be fully confirmed from this result if a direct positive correlation of the heterozygote levels with the malaria levels could be demonstrated. This is not the case for the villages of Sardinia in which the correlation between the degree of malarial morbidity and heterozygote frequency for thalassemia or G6PD deficiency is weak.
This observation allowed Brown (1981, 1984) to hypothesize that malaria was not the selective factor of an autochthonous mutation (which would thus be present only in malarial zones) but the maintaining factor of a mutation brought about by external gene flows during the Phoenician and Carthaginian conquests.
Nevertheless, an explanation for the lack of association between morbidity levels and heterozygote frequencies may be that the two sets of data do not refer to the same periods. Malarial morbidity reflects the situation of the Sardinian villages at the time of Fermi’s survey (the first decades of the twentieth century), whereas the proportion of heterozygotes is the outcome, by then stable, of a past selective process that occurred under environmental conditions that were probably different from those in the early twentieth century.
Thalassemia may well have had Phoenician origins, as stated by Brown (1981). In fact, the most prevalent thalassemia mutation found in Sardinia is in haplotypes similar to those found in other Phoenician and Carthaginian colonies (Cao et al. 1989). Malaria could also have been introduced in the same period as a result of ecological changes deriving from large-scale wheat production carried out by the Carthaginians (Brown 1984). Only the simultaneity of the two events can explain the increase and subsequent stabilization of thalassemia, which could not have survived without the selective effect exercised by malaria. From the low frequencies with which the mutation (which gives rise to a lethal gene, such as thalassemia) was presumably introduced into the population, fewer than 60-80 generations were necessary to reach equilibrium, if conditions for the balanced polymorphism (malaria) existed (Cavalli-Sforza and Bodmer 1971). Hence, in the period considered in this study, the gene frequency had been steady for about 1000 years, during which malaria operated as the maintaining factor.
Malaria should also have been steady during this long period, although presumably varying in space. Some constant ecological characteristics, such as high altitude, maintained a low level of infection in some villages because of the difficulty for the vector (the Anopheles labranchie) to conclude its life cycle. But many changeable factors, both ecological and cultural (Brown 1986), could have intermittently affected the incidence of malaria in the villages. However, these fluctuations could not have been large enough to modify the condition of equilibrium for these polymorphisms. Malaria was eradicated on Sardinia by the end of the 1940s as a result of the complete disinfestation of the vectors of infection and the consequent relaxation of selection; furthermore, in this drastic case the process of heterozygote reduction remains slow (Cavalli-Sforza and Bodmer 1971).
Hence, for the period to which this study refers, the genetic picture of a village can be considered the same as that of many years ago, whereas the epidemiological situation is different from the one that led to polymorphism equilibrium.
Acquired immunity is stressed when reproductive performance is studied in relation to morbidity levels (Zei et al. 1990); inherited immunity is emphasized when the effect of differing heterozygote frequencies is analyzed. The general fertility rate in a village may be influenced by both acquired and inherited immunity. Our data seem to confirm this hypothesis. It is worth noting that the difference between the levels of fertility and sterility (Tables 4 and 5), which correspond to two completely opposed situations (low heterozygotes + low malaria versus high heterozygotes + high malaria), was much greater than the differences for fertility and sterility rates relating to all degrees of malaria (Tables 1 and 3).
In highly endemic and high heterozygote frequency areas, the great number of women resistant to the consequences of malarial infection through immunity acquired during childhood should be added to the substantial number of women resistant because of genetic heredity. For both groups, at the time of reproduction the probability of fetal loss was low and the total fertility rate was near that of a normal population of women. On the contrary, in less endemic and low heterozygote frequency areas, the low acquired immunity and low inherited immunity increased the probability of fetal loss and hence decreased the total fertility rate.
No other factors apart from differential survival and differential fertility could have maintained these balanced polymorphisms over such a long period. From studies of the sickle cell gene, differential survival in the first five years seems to be the most effective of the two mechanisms through which natural selection acts, by favoring heterozygote sickle-cell-affected individuals (Allison 1956; Motulsky 1964; Molineaux and Gramiccia 1980).
Nevertheless, some direct studies on the fertility of different genotypes have pointed out that differential fertility favors heterozygous women over normal homozygous ones (Allison 1956, 1964; Fleming et al. 1979) with varying fertility ratios. The highest observed fitness advantage was 45% (Firschein 1961), but this value can be justified only if differential viability and differential fertility acted in opposite directions (Bodmer 1965). The two mechanisms, survival and fertility, should act together (but in different periods of an individual’s life) to maintain the balanced polymorphism.
Data from Sardinia, which refer to the female reproductive period only, support the hypothesis that heterozygous women enjoy a higher chance of bringing a pregnancy to conclusion than normal homozygous women. The result was displayed by the greater number of surviving children in areas where the frequency of heterozygotes was higher, independent of malaria morbidity levels. The observed difference in fertility rates was so great that a considerable heterozygote fitness advantage could be supposed. A part of this advantage might be ascribed to the reproductive compensation of the children who died (at birth or in early childhood) from their thalassemia homozygote condition. This particular form of reproductive behavior might be considered natural when a pregnancy resulting in a stillborn child is immediately followed by another pregnancy; however, no differential stillbirth related to differential heterozygote frequency was observed. Alternatively, reproductive compensation might be cultural, a result of the consciousness of the high probability of infant mortality due to the pathological condition of the family (Aguzzi et al. 1978; Cowan and Kerr 1986). In any case, such an attitude is unlikely to be shared by the whole population, stable over time, and is of such importance as to overcompensate for the losses caused by thalassemia.
Hence differential fertility rates could be considered partly responsible for the differential frequency of heterozygotes for thalassemia and G6PD deficiency. Acquired immunity, which depends on the level of endemicity in an area, might contribute to lowering or increasing the fitness that was already influenced by these balanced polymorphisms.
Aguzzi, S., C. Vullo, and 1. Barrai. 1978. Reproductive compensation in families segregating for Cooley’s anemia in Ferrara. Ann. Hum. Genet. 42:153-160.
Allison, A.C. 1954a. The distribution of the sickle-cell trait in East Africa and elsewhere, and its apparent relationship to the incidence of subtertian malaria. Trans. R. Soc. Trop. Med. Hyg. 48:312-318.
Allison, A.C. 1954b. Protection afforded by the sickle-cell trait against subtertian malarial infection. Br. Med. J. 1:290-294.
Allison, A.C. 1955. Aspects of polymorphism in man. Cold Spring Harbor Symp. Quant. Biol. 20:239-255.
Allison, A.C. 1956. The sickle cell and hemoglobin C genes in some African populations. Ann. Hum. Genet. 21:67-89.
Allison, A.C. 1961. Genetic factors in resistance to malaria. Ann. NY Acad. Sci. 91(3):710-729.
Allison, A.C. 1964. Polymorphism and natural selection in human populations. Cold Spring Harbor Symp. Quant. Biol. 29:137-149.
Bodmer, W.F. 1965. Differential fertility in population genetics models. Genetics 51(3):411-424.
Brown, P.J. 1981. New considerations on the distribution of malaria, thalassemia, and glucose-6-phosphate dehydrogenase deficiency in Sardinia. Hum. Biol. 53:367-382.
Brown, P.J. 1984. Malaria in Nuragic, Punic, and Roman Sardinia: Some hypotheses. In Studies in Sardinian Archeology, M.S. Balmouth and R.J. Rowland, eds. Ann Arbor, MI: University of Michigan Press, 209-235.
Brown, P.J. 1986. Cultural and genetic adaptations to malaria: Problems of comparison. Hum. Ecol. 14(3):311-332.
Cao, A., M. Gossens, and M. Pirastu. 1989. beta-thalassaemia mutations in Mediterranean populations. Br. J. Hematol. 71:309-312.
Cao, A., R. Galanello, M. Furbetta et al. 1978. Thalassaemia types and their incidence in Sardinia. J. Med. Genet. 15:443-447.
Carcassi, U., R. Ceppellini, and F. Pitzu. 1957. Frequenza della talassemia in quattro popolazioni sarde e suoi rapporti con la distribuzione dei gruppi sanguigni e della malaria. Boll. 1st. Sieroterap. Milanese 36:207-218.
Cavalli-Sforza, L.L., and W.F. Bodmer. 1971. The Genetics of Human Populations. Oxford, England: W.H. Freeman.
Cowan, J., and C. Kerr. 1986. Reproductive patterns and thalassemia major. J. Biosoc. Sci. 18:285-295.
Custodio, R., and R.G. Huntsman. 1984. Abnormal hemoglobin among the Black Caribs. In Black Caribs: A Case Study in Biocultural Adaptation, M. Crawford, ed. New York: Plenum Press, 335-343.
Delbrouk, J. 1958. Contribution a la genetique de la sicklemie. Ann. Soc. Belge Med. Trop. 38:103.
Edington, G.M. 1955. The pathology of sickle-cell disease in West Africa. Trans. R. Soc. Trop. Med. Hyg. 49:253-267.
Fermi, C. 1934, 1938, 1940. Regioni malariche Decadenza risanamento e spesa. Sardegna. Sassari, Italy: Stamperia della Libreria Italiana e Straniera, vols. 1-3.
Firschein, I.L. 1961. Population dynamics of the sickle-cell trait in the Black Caribs of British Honduras. Am. J. Hum. Genet. 13:233-254.
Fleming, A.F., J. Storey, L. Molineaux et al. 1979. Abnormal hemoglobins in the Sudan savanna of Nigeria. I. Prevalence of hemoglobins and relationships between sickle cell trait, malaria, and survival. Ann. Trop. Med. Parasitol. 73(2):161-172.
Garlick, J.P. 1960. Sickling and malaria in southwest Nigeria. Trans. R. Soc. Trop. Med. Hyg. 54:146-154.
Golini, A. 1967. Aspetti demografici della Sardegna. Quaderni del Seminario di Scienze Politiche dell’Universita degli Studi di Cagliari. Milan, Italy: Giuffre Editore.
Haldane, J.B.S. 1949. Disease and evolution. Ric. Sci. 19(suppl. 1):3-10.
Hamilton, P.J.S., D.A.M. Gebbie, N.E. Wilks et al. 1972. The role of malaria, folic acid deficiency, and hemoglobin AS in pregnancy at Mulago Hospital. Trans. R. Soc. Trop. Med. Hyg. 66:594-602.
Livi Bacci, M. 1977. A History of Italian Fertility. Princeton, NJ: Princeton University Press.
Livingstone, F.B. 1957. Sickling and malaria. Br. Med. J. 1:762-763.
Livingstone, F.B. 1964. Aspects of the population dynamics of the abnormal hemoglobin and glucose-6-phosphate dehydrogenase deficiency genes. Am. J. Hum. Genet. 16:435-450.
Livingstone, F.B. 1971. Malaria and human polymorphisms. Annu. Rev. Genet. 5:33-64.
Luzzatto, L. 1979. Genetics of red cells and susceptibility to malaria. Blood 54:961-976.
Luzzatto, L., and U. Bienzle. 1979. The malaria/G-6-PD hypothesis. Lancet 1:1183-1184.
Luzzatto, L., E.S. Nwachukes-Jarret, and S. Reddy. 1970. Increased sickling of parasitized erythrocytes as mechanisms of resistance against malaria in sickle cell trait. Lancet 1:319-322.
Luzzatto, L., E.A. Usanga, and S. Reddy. 1969. Glucose-6-phosphate dehydrogenase deficient red cells: Resistance to infection by malarial parasites. Science 164:839-842.
Madrigal, L. 1989. Hemoglobin genotype, fertility, and the malaria hypothesis. Hum. Biol. 61(3):311-325.
McFalls, J.A., and M. Harvey McFalls. 1984. Disease and Fertility. Orlando, FL: Academic Press.
McGregor, I.A. 1984. Epidemiology, malaria, and pregnancy. Am. J. Trop. Med. Hyg. 33(4):517-525.
Modiano, G., L. Terrenato, R. Scozzari et al. 1986. Population genetics in Sardinia (with a historical account of the birth of the Haldane “malaria hypothesis”). Atti Accad. Naz. Lincei Mem. 18(3):257-330.
Molineaux, L., and G. Gramiccia. 1980. The Garki Project: Research on the Epidemiology and Control of Malaria in the Sudan Savanna of West Africa. Geneva, Switzerland: World Health Organization.
Motulsky, A.G. 1964. Hereditary red cell traits and malaria. Am. J. Trop. Med. 13:147-158.
Motulsky, A.G. 1975. Glucose-6-phosphate dehydrogenase and abnormal hemoglobin polymorphisms: Evidence regarding malarial selection. In The Role of Natural Selection in Human Evolution, F.M. Salzan, ed. Amsterdam. The Netherlands: North-Holland, 271-291.
Pasvol, G., and R.J.M. Wilson. 1982. The interaction of malaria parasites with red blood cells. Br. Med. Bull. 38:133-140.
Pasvol, G., D.J. Weatherall. and R.J.M. Wilson. 1978. Cellular mechanism for the protective effect of hemoglobin S against P. falciparum malaria. Nature 274:701-703.
Piazza, A., M.C. Belvedere, D. Bernoco et al. 1972. HL-A variation in four Sardinian villages under differential selective pressure by malaria. In Histocompatibility Testing 1972, J. Dausset and J. Colombani, eds. Copenaghen, Denmark: Munksgaard, 73-84.
Piazza, A., W.R. Mayr, L. Contu et al. 1985. Genetic and population structure of four Sardinian villages. Ann. Hum. Genet. 49:47-63.
Rucknagel, D.L., and J.V. Neel. 1961. The hemoglobinopathies. In Progress in Medical Genetics, G. Steinberg, ed. New York: Grune and Stratton, 158-260.
Silvestroni, E., I. Bianco, G. Montalenti et al. 1950. Frequency of microcythemia in some Italian districts. Nature 165:682-683.
Siniscalco, M., L. Bernini, G. Filippi et al. 1966. Population genetics of hemoglobin variants, thalassaemia, and glucose-6-phosphate dehydrogenase deficiency, with particular reference to the malaria hypothesis. Bull. WHO 34:379-393.
Siniscalco, M., L. Bernini, B. Latte et al. 1961. Favism and thalassaemia in Sardinia and their relationship to malaria. Nature 190:1179-1180. Stamatoyannopoulos, G., A. Panayotopoulos, and A.G. Motulsky. 1966. The distribution of glucose-6-phosphate dehydrogenase deficiency in Greece. Am. J. Hum. Genet. 18(3):296-308.
Terrenato, L. 1973. beta-and non-beta-thalassaemia in Sardinia and their frequencies. Ann. Hum. Genet. 36;285-295.
Terrenato, L., E. Van Loghem, L. Bernini et al. 1971. Preliminary data on the genetic heterogeneity among the Sardinian isolates. Rend. Accad. Naz. Lincei 51:249-253.
Usanga, E.A., and L. Luzzatto. 1985. Adaptation of Plasmodium falciparum to glucose-6-phosphate dehydrogenase deficient host red cells by production of parasite-encoded enzyme. Nature 31:793-795.
Vulliamy, T., P. Mason, and L. Luzzatto. 1992. The molecular basis of glucose-6-phosphate dehydrogenase deficiency. Trends Genet. 8(4):138-143.
Weatherall, D.J., and J.B. Clegg. 1981. The Thalassemia Syndromes. Oxford, England: Blackwell Scientific.
Workman, P.L., P. Lucarelli, R. Agostino et al. 1975. Genetic differentiation among Sardinian villages. Am. J. Phys. Anthropol. 43;165-176.
Zei, G., and L.L. Cavalli-Sforza. 1977. Education and birth control. Genus 33(1-2):15-42.
Zei, G., A. Lisa, and P. Astolfi. 1990. Fertility and malaria in Sardinia. Ann. Hum. Biol. 17(4):315-330.
Acknowledgments Ornella Fiorani’s assistance is gratefully acknowledged. A. Lisa was supported by a fellowship from the Fondazione Adriano Buzzati-Traverso. C. Di Pasquale undertook this work with the support of the International Centre for Theoretical Physics (ICTP) Programme for Training and Research in Italian Laboratories, Trieste. This research was partly supported by an Italian Ministero Pubblica Instruzione (MPI) 60% fund.
Received 5 April 1993; revision received 30 September 1993.
Copyright Wayne State University Press Aug 1994
Provided by ProQuest Information and Learning Company. All rights Reserved